Cervical
• The effectiveness of thoracic spine manipulation on immediate relief of mechanical neck pain.

Lumbar
• Change in perceived disability and pain is not significant between the treatment choice of exercise or manipulation for low back pain.
• Spinal stabilization exercises plus trunk strengthening versus general trunk strengthening alone in preventing recurrent low back pain

Shoulder
• Two-year Prognosis for Recurrent Shoulder Instability after a Conservatively Treated Primary Anterior Glenohumeral Dislocation

Elbow
• Short and Long term effectiveness of Manual Physical therapy vs Steroid injection

• Comparing the Performance Enhancing Effects of Squats on a Vibration Platform with Conventional Squats in Recreationally Resistance-Trained Men

Hip
• Return to independent ambulation after hip fracture in previously independent ambulatory elderly patients.
• Comparison of manual therapy and exercise therapy in the treatment of osteoarthritis of the hip

Knee
• Is Manual Therapy and Exercise Effective in Treating Patients with Knee Osteoarthritis?
• Patellofemoral pain syndrome (PFPS) still lacks good valid clinical tests.

Foot/Ankle
• Calcaneal taping decreases patient's pain in the short-term, but does not improve function.

• Can Removable Cast Walkers be Modified to Aid in Pressure Relief and Healing of Diabetic Foot Ulcers?
• Efficacy of MIST Ultrasound in the Treatment of Chronic Diabetic Foot Ulcers

• Prognosis for Recovery of Good Upper Extremity Use after Stroke in Adults
• Determination of the clinical usefulness of Tinnel's test, Phalen's test and sonographic measurement in the diagnosis of Carpal Tunnel Syndrome

• Use of the Alberta Infant Motor Scales in Screening for Motor Abnormality
• Positive effects from individualized developmental care on healthy, pre-term infants in NICU

• Determination of the predictive value of Body Mass Index, Waist Circumference and Waist-Hip Ratio in determining risk of getting Type 2 Diabetes Mellitus in men and women

ORTHOPEDICS:

The effectiveness of thoracic spine manipulation on immediate relief of mechanical neck pain.

Patients between the ages of 18-60 with purely mechanical neck pain appear to have immediate, clinically significant relief from thoracic spine manipulation. It appears from the authors report that there were no adverse side effects of thoracic spine manipulation. However, one should use caution when considering this treatment completely safe as the authors did not perform a formal follow-up for adverse events.

Level of Evidence: 1b

Citation/s:Cleland JA, Childs JD, McRae M, Palmer JA, Stowell T. Immediate effects of thoracic spine manipulation in patients with neck pain: a randomized clinical trial. Manual Therapy. 2005;10:127-135.

Lead author's name and fax: Joshua A. Cleland. Fax #: 603-785-5576

Three-part Clinical Question: For a 25 year old Navajo female with mechanical neck pain, is thoracic spine manipulation more effective than placebo in providing immediate reduction in the patient's neck pain?

Search Terms: Pub Med clinical queries search - Therapy with a narrow/specific focus: "Neck pain" and " Thoracic spine manipulation" yielded 1 article

The Study:Single-blinded concealed randomized controlled trial with intention-to-treat.

The Study Patients: Patients included in the study were 18-60 years old with primary complaint of mechanical neck pain which was defined by neck pain at the cervicothoracic junction exacerbated by neck movements. A total of 68 patients were referred by primary care to an outpatient orthopedic physical therapy clinic. Of the 68 initially referred, 16 patients were excluded due to one of the following exclusion criteria: any red flags for serious spinal condition (such as infection, tumors, osteoporosis, spinal fracture, etc.), pregnancy, signs or symptoms of nerve root impingement, history of cervical or thoracic surgery, hypermobility of the thoracic spine, or prior experience with spinal manipulation techniques. Of the 52 patients remaining, 16 additional patients were excluded because either they indicated that they preferred not to receive a spinal manipulation or they specifically asked for spinal manipulation. The 36 remaining patients in the Experimental (E) and Control Group (C) were not statistically different in age (E=36 years +/-8.5; C=35 years +/- 11.3), Gender (E= 14 females, 5 males; C= 13 females; 4 males), Neck Disability Index score (E=28.4 +/- 11.9; C=33.6 +/-14.2), duration of symptoms (E=12.2 weeks +/- 3.5; C=13.2 +/-4.2) pre-treatment reatment pain level (per Visual Analog Scale) (E=41.6 +/- 17.8; C=47.7 +/- 18.4). All 36 patients that qualified for the study completed the study.

Control group (N = 17; 17 analyzed): After initial physical therapy evaluation to included blinded segmental mobility testing of the thoracic spine, the control patients received a placebo manual therapy maneuver of the thoracic spine. Patients were placed in a supine position in a manner that modeled the same position for manipulation as the experimental group. However, unlike the experimental group, the control group did not receive high velocity low amplitude manipulation to the thoracic spine.

Experimental group (N = 19 ; 19 analyzed): Experimental Group - After initial physical therapy evaluation to include blinded segmental mobility testing of the thoracic spine, the experimental group received spinal manipulation of the thoracic spine in the supine position to areas identified on segmental mobility examination as limited in mobility. Patients received high velocity low amplitude manipula tion at each level that was deemed limited in the segmental mobility examination. If the initial manipulation did not produce an audible cavitation, a second manipulation was performed. Only two manipulation attempts were made at each level of restriction identified on the initial segmental mobility examination.

The Evidence:

Comments: Are the results valid? The results of this study appear to have very good validity. First, the patients were blinded to the group assignment and were randomly assigned to either the Experimental or Control group. In addition, to ensure patients were not aware of their group of assignment, the authors eliminated those patients who had experience with spinal manipulation from the study. Those patients familiar with spinal manipulation would be much more capable of detecting the placebo treatment and this knowledge would potentially confound the results of the study. In addition, the experimenters that administered the primary outcome measure (the Visual Analog Scale or VAS) before and after treatment were completely blinded to the patients' group assignment and thus unlikely to bias the results of the patients' scores on the VAS. Likewise, the treating therapists were blinded to the patients' group assignment when performing the physical therapy examination to include the initial segmental mobility examination. The only drawback to the validity of the results is that the treating therapist was not blinded to the patients' group assignment during treatment because the treating therapist had to perform the correct intervention. Although unlikely, the therapist's knowledge of group assignment after the segmental mobility examination (during treatment) could have possibly influenced each patient's knowledge to which group he or she was in by gaining clues from the treating therapist's behavior. For instance, therapists may not have been as enthusiastic when treating the control patients as opposed to the experimental group thus tipping them off to their group assignment. This is unfortunately unavoidable due to the fact that the therapist providing the treatment must know to either provide a true manipulation or a placebo manipulation in order to perform this study. I see no other way to conduct the study other than the methods used by the authors of this study.

The patients in each group were very similar in all key characteristics to include age, gender make-up, perceived disability as measured by the Neck Disability Index, pre-treatment pain level, and symptom duration prior to treatment. Therefore, the groups appeared to be sufficiently homogeneous at the beginning of the study. In addition, due to the short, 1 session duration of the study, all patients that were ultimately selected to participate completed the study and thus follow-up for the study was complete for the primary outcome variable (the VAS scores). Finally, the eliminating patients from the study that were unwilling to receive a spinal manipulation was not only reasonable but mirrored clinical practice. It is standard practice that spinal manipulation, and any other treatment, cannot be performed without the proper consent from the patient.

What are the results? The results of this short duration study are somewhat mixed. The authors report in a table format the pre- and post-treatment pain VAS scores as well as the overall pain VAS change score for both Experimental and Control groups. According to this table, the pre-treatment VAS scores are not significantly different between the two groups as mentioned above this lends itself to the homogeneity between the two groups prior to treatment. The post-treatment VAS scores are reported as statistically significantly different (p <.01) with means and standard deviations reported. However, the 95% confidence intervals (CI) for each of these means are relatively wide. In fact, if one assumes the upper bound of the 95% CI for the Experimental group's post-treatment VAS score of 34.4 mm and the lower bound of the 95% CI for the Control group's post-treatment VAS score of 33.5 mm, one could reasonably see that the possibility exists (although very small) that there is no difference in post treatment pain between the two groups. The authors do not focus on this statistic but rather choose to focus on the change in pain VAS score as their primary outcome measure.

When assessing the effectiveness of a given treatment, it is reasonable for the authors to focus on the change in pain level for each group as a primary outcome measure. The experimental gro15.5 mm a 15.5mm (95% CI = 11.8 - 19.2) decrease in VAS score after treatment. This represents a clinically significant 37% decrease in pain level in just a single treatment session. The change in the Control group's VAS score was a 4.2mm (95% CI = 1.8-6.6) decrease representing a clinically insignificant 9% decrease in pain level in a single treatment . Not only are these change scores statistically and clinically significantly different (p<.001) but they both have a relatively narrow 95% CI. If one assumes the a change of 11.8mm for the Experimental group this is clinically meaningful as it represents a 28% reduction in pain as apposed to the 14% reduction of pain for the Control group is one assumes the actual change is the upper limit of the 95% CI at 6.6mm.

Finally, the authors report that patients were encouraged to report any adverse side effects (defined as anything worse than soreness lasting 3 hours or less) to the principle investigator. The authors reported that none of the subjects reported an adverse event. However, the authors do not provide the time frame for which patients were allowed to contact the principle investigator. The authors also did not formally inquire with each patient to identify any possible adverse events or outcomes. Therefore, calculations of number needed to treat and harm are not possible to estimate. However, it appears that the treatment is safe given the report that no patient contacted the principle investigator about adverse events.

How do the results apply to my patient? Overall, the results support the use of thoracic spine manipulation for immediate relief of mechanical neck pain when indicated by a the patients history and physical examination. The results of this study apply well to my patient as the study was performed primarily on female patients with a mean age very close to that of my patient. In addition, given the single treatment reduction in pain reported in this study, it would be important for me to offer such a treatment to my patient given the remote nature of my practice on the Navajo Indian reservation. Most of my patients either do not have reliable transportation, live a great distance from the PT clinic, or live in areas where the roadways are greatly affected by changes in weather. Often these factors translate into the inability to regularly see my patients for physical therapy. Therefore, I am obligated to at least offer the option of a treatment that can provide a patient a mean of 37% decrease in pain in a single treatment. Finally, it is reported by the authors of this study that no patient contacted the authors about adverse events related to treatment via thoracic spine manipulation therefore it appears that the treatment is likely safe when using the same inclusion and exclusion criteria as the authors.

Appraised by: Matthew J. Armentano, Email: matthew.armentano@ihs.gov

Change in perceived disability and pain is not significant between the treatment choice of exercise or manipulation for low back pain.

This Critical Appraisal Topic's purpose is to help the reader decide that exercise therapy and manipulation therapy both improve perceived disability and pain for low back pain patients. The results of this study were clinically insignificant for choosing one therapy over the other for better relief of low back pain. It did illustrate that manipulation for some patients (those that fall into the upper end of the CI boundary for the pain scale) may improve pain scores by over 25% This randomized trial was not well protected from threats to validity. Therefore, this study's level of evidence is a 2B using Sackett's hierarchy. It is an individual low quality RCT with a greater than 20% loss to follow-up.

Citation/s:BEAM Trial Team. United Kingdom back pain exercise and manipulation (UK BEAM) randomized trial: effectiveness of physical treatments for back pain in primary care. BMJ. 2004;329:1377.

Lead author's name and fax: Martin Underwood, M.D., Professor of General Practice. Fax not available

Three-part Clinical Question: For a 45 year old female with low back pain with some left leg radiation (never past her knee) for 3 monthst 3months, is a prescribed exercise program or manipulation more effective in reducing back pain and perceived disability?

Search Terms: Back pain AND manipulation AND effectiveness AND randomized trial entered under Clinical Queries (broad, sensitive search). This yields 4 hits one of which is the UK BEAM article.

The Study:Non-blinded randomized controlled trial with intention-to-treat.The Study Patients: This study started with 1334 patients with low back pain who were randomly assigned to one of four interventions. The interventions consisted of a "Best Care" treatment (A UK national acute back pain guideline advising on continuing normal activities and avoiding rest and a booklet called The Back Book). All the subsequent interventions included the "Best Care" treatment. Treatments consisted of a guided exercise program, manipulation, and manipulation with a guided exercise program. The latter two interventions were further divided into private practices versus National Health Service (NHS) practices. There were 14 participating medical practices where research nurses identified patients who had come to their general practitioners for simple low back pain or through searching computerized records of these facilities. They assessed potential participants' eligibility and interest by brief postal questionnaires. They saw the participants twice. The first time to explain the trial and assess eligibility and the second time to confirm eligibility, collect practitioners' consent, seek participants' consent, collect baseline data, and randomize participants. The mean age at randomization was 43 +/-11 years and 56% were female. Over 58% had experienced low back pain for longer than 90 days. The mean for the Roland Disability Questionnaire (RDQ 0-24, 0=best)) was 9.0 +/-4.0. The six groups overall had very similar characteristics. Inclusion criteria was met if the patient was between 18-65 years, consulted for simple low back pain, had a score of 4.0 or more on the RDQ, had a history of consistent low back pain for the last month or longer, and agreed to avoid other physical treatments of their back pain other than the assigned trial treatment for 3 months. Exclusionary criteria included over age 65, possibility of serious spinal pathology, pain mainly below the knee, previous spinal surgery, RDQ score of 3 or less, other musculoskeletal disorder that was worse than their back pain, had previously attended a pain management clinic for their back pain, severe psychiatric or psychological disorder, moderate to severe hypertension, on anticoagulant treatment, taking long term steroids, unable to walk a 100 meters when free of back pain, unable to get up and down from floor independently, physical therapy in the last 3 months, and could not read and write fluently in English.

After all inclusion and exclusionary criteria had been met all participants completed questionnaires. Two were back specific instruments- the Roland Morris disability questionnaire (RDQ) and the modified Von Korff (VK) scales. Next they completed two measurements of belief questionnaires (the back beliefs questionnaire and the fear avoidance beliefs questionnaire) and finally two generic measures (the SF-36 and the EuroQol) that were used for a subsequent economic paper. These questionnaires were repeated at 1,3, and 12 months after randomization. Since, the question being asked is whether exercise or manipulation is more effective in decreasing low back pain and perceived disability this CAT is looking at only the Best Care and exercise group and the Best Care and manipulation group with private and NHS groups recombined. The exercise group had an N of 310, 170 females and 140 males with a mean age of 44 +/-11 years and the manipulation group had an N of 353, 212 females and 141 males with a mean age of 42.8 +/-11.3 years. The exercise group had an average baseline RDQ of 9.2 +/-4.3 and a baseline Von Korff pain score of 60.8 +/-17.6. The manipulation group had a baseline RDQ average of 8.9 +/-4.0 and a Von Korff (VK) pain score of 61.5 +/-19.0.Control group (N = ; 204 analyzed): The control group received the "Best Care" advice and The Back Book booklet and a guided exercise program. The exercise component consisted of an initial individual assessment followed by group classes provided by trained physical therapists with at least 2 years experience to deliver the program. There were up to 10 people in a class at local community facilities. The participants were invited to attend up to eight 60-minute sessions over four to eight weeks and a refresher class at 12 weerandomization. Experimentalerimental group (N = ; 275 analyzed): The treatment group received the "Best Care" advice and The Back Book booklet and spinal manipulation. A multidisciplinary group developed a package of techniques representative of those used by the UK chiropractic, osteopathic, and physical therapy professions. These were agreed upon by all three organizations for use in this trial. Following initial assessment, manipulators chose from the agreed manual and non-manual treatment options. They agreed to do a high velocity thrust at least once on most patients. Participants were invited to attend up to eight 20 minute sessions, if necessary, over 12 weeks.

The Evidence:

Comments: Low back pain, as pointed out in this article, is a common and costly medical problem. It is a major cause of missed workdays and causes a decrease in one's perceived quality of life. Finding a therapy that gives the greatest pain relief and quickest return towards normal function would be welcomed by anyone who has ever experienced back pain. The intent of this study was to determine if exercise, manipulation, or exercise with manipulation was any better than basic back care protocol in returning patients to pre-back pain levels of function.

Unfortunately, there were a number of threats to validity of this study both in its execution and methodology. This study recruited patients directly from a large number of clinics and through searching computerized records. After all eligibility requirements were met the subjects were randomized. These measures helped to ensure there was an equal prognosis between the groups and therefore help prevent bias in the studies' outcome. This study did use a remote randomization service for which group the subjects would be allocated too. All participants were analyzed in the group they were randomized to, which results in an intention-to-treat analysis. This action further helps keep the prognostic factors between groups relatively equal and therefore the result will be from the treatment assigned. It goes on to say that nurses notified the participants of their allocations; it is unclear if the participants were aware that there were multiple options for treatment. I suspect that the patients were aware as were the clinicians performing the different services. The article goes on to say that they thought that blinding of the participants and professionals was not desirable or possible. This introduces two more threats to this study's validity. First, if the patient is aware that he/she is getting additional intervention the participant might answer the questionnaires differently depending if they believe that their treatment is the treatment being assessed. This is similar to a placebo effect. The second threat was that the clinicians' were aware of the patient allocation and may have consciously or unconsciously tried a little harder with their manipulation therapies or exercise strategies than if they were unaware if the subject in front of them was part of the study. It would also appear that the outcome assessors were aware of the group allocations. This did not seem to be as much of a threat to validity since the subjects completed the questionnaires (presumably at home) and the questionnaires did not leave the assessors the leeway of a biased interpretation like some physiologic or clinical tests could. Another unclear threat to this studies' validity is loss to follow-up. At 3 months the exercise group had a loss to follow-up of 27% and the manipulation group loss was 19%. At the outset of the study the authors allowed for a 67% follow-up rate at 12 months and did not comment on any other time frames. They recruited 1350 participants based on this 67% follow-up rate. They did not mention any reason for their subsequent loss to follow-up. They did give the demographics of those who were most likely to respond as female, above average age, educated beyond 16, and to have severe low back pain at randomization. They reported that this trend was consistent across all the groups and therefore little risk of bias exists. When one looks at the two groups, the loss of the exercise group reflects a 22% greater loss in participants than the manipulation group (85 exercise subjects vs. 66 manipulation subjects). It should be considered that the reason that there was a greater loss in the exercise group is that there was an unreported adverse affect within that program or perhaps human nature plays a part because the commitment of exercise is more difficult of a commitment than being a passive recipient of manipulation. Considering the improvement of the RDQ and VK scores from baseline to 3 months is very similar in both groups loss to follow-up is probably a low threat to the validity of this studies' outcomes.

The between group differences were shown homogeneous at the outset of the study by a comparison of the mean Roland Disability Questionnaire scores and the Von Korff pain scale scores. The exercise group averaged almost 4 points lower on the RDQ at 12 weeks. This reflects a 41% drop (P< 0.01) in perceived disability. They also had a 16-point drop in the VK pain scale, which represents a 23% improvement (P< 0.05) in pain after 12 weeks of exercise intervention. The manipulation group also averaged almost 4 points lower on the RDQ and an average decrease of 21 points on the VK pain scale at 12 weeks. This is a 43% improvement (P<\0.001) on the RDQ at 12 weeks and a 33% (P< 0.001) improvement in pain after 12 weeks of manipulation intervention. To further reinforce that the two groups were homogenous at the outset and therefore these improvements at 12 weeks reflect the interventions we need to look at the confidence intervals for the difference between the two means. If the two groups were from the same population we would expect that the confidence interval to include zero 95% of the time, and if it excluded zero then we would suspect that they were from different populations. The RDQ confidence interval between the means of exercise and manipulation was -0.42 to 1.18 and the VK confidence interval between the means of exercise and manipulation was -0.65 to 8.31. These CI's both include zero so we can confidently assume that the samples were from the same population. In addition, these confidence intervals provide us some insight to the magnitude of the treatment effect of exercise versus manipulation. The RDQ CI of -0.42 to 1.18 reflects that at the lower boundary of the CI that exercise would improve the RDQ by reducing its score by 0.42. The upper boundary shows us that manipulation would improve the RDQ by reducing the score by 1.18. Given the likelihood that 95% of the subjects would not be at either boundary the change in scores would not be significant enough to choose one therapy over the other for a patient. The VK confidence interval of -0.65 to 8.31 illustrates that at the lower boundary of the CI the exercise group would improve (by reducing it) their pain score by 0.65 and at the upper end of the CI boundary the manipulation group would improve (by reducing it) their pain score by 8.31 points. At first glance this also looks to be insignificant. However if you were the subject at the upper boundary your 8-point reduction actually translates into a greater than 25% reduction in one's pain score. This is an improvement that most back patients would find significant and may be worth considering manipulation. This could be the minimal clinically important difference for a patient especially, if they have already tried exercise therapy. This is a drawback to this study as it does not give details on how the VK scale works and the reader would not readily ascertain this information.

Clinically, this study presented that exercise and manipulation therapies can both improve low back pain and perceived disability. It does not significantly prove that one therapy is better than the other. It does provide the possibility that if exercise does not improve the patient's outcome then a trial of manipulation may provide an improved outcome that the patient finds important such as possibly reducing their back pain by 25 percent. Perhaps a larger study or an improved loss to follow-up would provide more concrete answers to which treatment is more beneficial. In applying my patient to this study, this article would not change my approach with the exception that if a reasonable trial of exercise had been truly undertaken without much improvement then a course of manipulation may be the intervention that would make a difference in this patient's low back pain outcome.

Appraised by: Heather E. Khan; Wednesday, February 07, 2007 Email: ehenfeld@comcast.net

Spinal stabilization exercises plus trunk strengthening versus general trunk strengthening alone in preventing recurrent low back pain

This study does not conclusively demonstrate that a program of specific stabilization exercises added to general trunk strengthening is or more or less effective than general trunk strengthening alone in preventing recurrent low back pain in patients without spinal instability. Patient follow-up was poor and there were not sufficient subjects who completed the study to protect against Type 2 error.

Level of evidence: 2b

Citation/s: Koumantakis GA, Watson PJ, Oldham JA. Trunk muscle stabilization training plus general exercise versus general exercise only: randomized controlled trial of patients with recurrent low back pain. Phys Ther. 2005;85:209-225.

Lead author's name and fax: GA Koumantakis, Drosopoulou 6, Kypseli, Athens 112 57, Greece gak4@otenet.gr

Three-part Clinical Question: For a 53 year-old male with chronic, recurrent episodes of low back pain, is there a difference in function or low back pain recurrence when specific lumbar stabilization exercises are added to general lumbar and abdominal strengthening exercises?

Search Terms: PubMed Clinical Queries (narrow, specific search) with the following search terms: "lumbar stabilization exercises." The search resulted in 11 articles, one of which was the study by Koumantakis et al.

The Study: Single-blinded concealed randomized controlled trial with intention-to-treat.

The Study Patients: Fifty-five patients with recurrent, non-specific low back pain were randomized to either a general exercise group (n=26) or a stabilization-enhanced general exercise group (n=29). Subjects were excluded if they had a prior history of spinal surgery, displayed any spinal "red flags," or had signs and symptoms of spinal instability. The descriptive statistics of the two groups are as follows (with standard deviations for each measurement):

General exercise-only group: mean age 35.2 years (9.7), height 174.4 cm (9.1), body mass 80.5 kg (12.0), BMI 26.4 kg/m2 (3.2), time since first onset of LBP 44.2 months (51.6), and current duration of LBP 12 weeks (8.0-12.0). Stabilization-enhanced group: mean age 39.2 years (11.4), height 170.1 cm (7.5), body mass 75.9 kg (12.8), BMI 26.2 kg/m2 (4.2), time since first onset of LBP 57.1 months (48.1), and current duration of LBP 12.0 weeks (7.3-22.0).

Control group (N = 26; 26 analyzed): The control group first performed warm-up exercises for 10-15 minutes consisting of stretching and stationary bicycling. The subjects then performed exercises for lumbar and abdominal strengthening. These exercises were progressed each week and are all listed in the Appendix of the study. They include exercises such as crunches, prone back extensor strengthening, abdominal oblique strengthening, supine lumbar bridging, arm and leg extension in quadruped, and side lying oblique isometric hold. Subjects had clinical visits twice per week for eight weeks. Subjects also received illustrated handouts on how to perform the exercises at home as well.

Experimental group (N = 29; 29 analyzed): The experimental group (stabilization-enhanced group) also performed a warm-up of stretching and stationary bicycling for 10-15 minutes at the beginning of each therapy session. They also attended therapy twice per week for 8 weeks. This group, like the control (general exercise) group, also had progressive exercises that were modified each week during the eight-week period. However, these exercises were initially focused on specific spinal stabilization, (as demonstrated by co-contraction of the lumbar multifidus and transversus abdominis muscles.) Repeated attempts were made to avoid any muscle substitution strategies: those in which trunk movement muscles were activated instead of local stabilization muscles. During the final three weeks, exercises that incorporated larger trunk movements were also performed.

The Evidence:

Comments:

Are the results valid? Most characteristics of this study give credence to its validity. Patients were randomized to either the control or the experimental group via a computerized random number generator, and the randomization codes were concealed in sealed envelopes. The assessing physical therapist was blinded to group allocation. All patients were analyzed in the groups to which they were randomized, and patients in each of the two groups were similar with regard to prognostic factors. Clinicians and patients were aware of group allocation, although an attempt was made to disguise the purpose of the differences between the two exercise groups from the patients. They were each told that the purpose of the study was "to identify any differential effect between 2 exercise regimens for the trunk muscles, which have a role in protecting the spine from further injury."

The validity issue that was a glaring scar on the face of this study was the question of whether or not follow-up was complete. According to the Figure in the study, there were 9/26 (35%) patient drop-outs from the control group and 8/29 (28%) drop-outs from the experimental group by the 5-month (20 week follow-up). Although the authors attempted to explain the statistical corrections they made for these drop-outs (inserting group means in the place of missing values - "a relatively conservative approach," the high drop-out rate may alone represent a fatal flaw for this study. Half of the statistical analyses/conclusions for this study were done at the 20-week time frame, and by that time the large number of drop-outs had occurred. If one-third of the patients (the number of drop-outs) in either group had had the opposite response to treatment (more or less pain or disability,) the statistical results of the study could have been strongly in favor of or against either exercise program.

What are the results? The authors of this study did not find a statistically significant benefit of adding specific spinal stabilization exercises to general trunk strengthening. The authors claimed that at the 8-week follow-up there was a statistically significant difference on the RMDQ in favor of the general exercise-only group, but that this difference was lost by 20 weeks. As mentioned previously, the follow-up for this study was poor; at 8 weeks, there were already 10 drop-outs, giving a follow-up of 82%, while at 20 weeks there were 7 more drop-outs, decreasing follow-up to 69%. This poor follow-up makes the data and conclusions for the study unreliable. The authors claimed to have recruited sufficient subjects to have a power of 80% and therefore reduce the chance of a Type 2 error. However, this calculation was based on 28 subjects per group, or 56 subjects total. Initial statistics were collected on only 55 subjects. This number dropped to 45 at 8 weeks and 38 by 20 weeks. What may have been a power of 80% at the beginning of the study was no longer the case by the end.

I focused on three primary outcome measures from the study for my particular patient: the VAS B, VAS C, and RMDQ, since they were the more direct measures of pain and disability used by the authors. Between group means and 95% confidence intervals were listed on Table 3 in the study. Even though the 95% CI's for 5 of the 6 measures (2 each for VAS B, VAS C, and RMDQ, one at 8 weeks and 20 weeks) included 0 and therefore were not statistically significant, a further evaluation of the CI's is warranted. The authors stated earlier in the study that a difference of 2.5 points on the RMDQ was a meaningful clinical difference. The 95% CI's for the RMDQ at both 8 weeks and 20 weeks included 2.5 (the upper end of the 95% CI at 8 weeks was 4.81 and that at 20 weeks was 3.64). Given this data, one cannot conclusively claim that there was no significant difference between the two groups, because one may well have existed. Also, the 95% CI for VAS C went from -15 to +15 or 16; a difference of 16 points between group means on the VAS may well have been a significant difference that was not conclusively ruled out.

How can I apply the results to patient care? I believe that the premise behind this study was a valuable clinical research question, but the poor patient follow-up and failure to eliminate the chance of a Type 2 error give no basis for any of the authors' conclusions. In the Guest Editorial for this issue of Physical Therapy Journal, Dr. Anthony Delitto appears to not agree with the "magic bullet approach" of this study, which is giving stabilization exercises to patients without signs and symptoms of instability. I believe this was a valid research question, however, because there are articles in the published literature which give credence to this very idea (stabilization exercises to prevent recurrent low back pain in any low back pain patient) without sufficient data to back up the conclusion. This study could have helped debunk the proclaimed value of specific stabilization exercises for all low back pain patients.

The subjects in this study were younger than my patient, but if the results were valid, which they are not, they would probably be applicable to my patient despite the age difference. The results of this study do not add any useful information to the body of published literature, but if this study was repeated and had better follow-up and protection against Type 2 errors, it might be beneficial when deciding which exercises to prescribe to prevent recurrent low back pain.

Appraised by: Scott Mitchell, DPT -- Tuesday, March 27, 2007 Email: scottm_pt@hotmail.com

Two-year Prognosis for Recurrent Shoulder Instability after a Conservatively Treated Primary Anterior Glenohumeral Dislocation

This evidence suggests that 55.7% (95% CI = 49.1% to 62.2%) of patients, 15 to 35 years old, who have sustained a primary anterior glenohumeral dislocation are likely to develop recurrent instability within two years. Patients of greater age and of female gender had a better prognosis for not developing recurrent instability. Minimal threats to validity, strong study design, and validation against multiple previous studies strengthens the clinician's confidence in using age and gender as prognostic factors for recurrent glenohumeral instability. This study has been assigned a level of 1b evidence using Sackett's Hierarchy.

Citation/s:Robinson CM, Howes J, Murdock H, Will E, Graham C. Functional Outcome and Risk of Recurrent Instability After Primary Traumatic Anterior Shoulder Dislocation in Young Patients. J Bone Joint Surg Am. 2006 Nov;88(11):2326-36.

Lead author's name and fax: C. Michael Robinson, FRCSEd(Orth). E-mail: c.mike.robinson@ed.ac.uk. Fax not given.

Three-part Clinical Question: For a 26 year-old male with a recent anterior glenohumeral dislocation, what is the probability of recurrent instability within the first two years after the primary dislocation?

Search Terms: I used PubMed Clinical Queries to conduct a narrow, specific search for Prognosis with the following search string: (shoulder dislocation AND instability) AND (prognos*[Title/Abstract] OR (first[Title/Abstract] AND episode[Title/Abstract]) OR cohort[Title/Abstract]). The search yielded 24 hits, the second of which was the Robinson et al. article.

The Study:The Study Patients: Subjects were comprised of 252 consecutive patients between the ages of 15 and 35 (mean age = 23.4, 95% CI = 22.7 to 24.1) who had sustained a radiographically confirmed, primary anterior glenohumeral dislocation. The patients were first treated in an emergency department and immediately referred to an orthopedic shoulder clinic within one week of the primary dislocation. Primary dislocations were treated conservatively. There were 225 (89%) males and 27 (11%) females. The main exclusion criterion was atraumatic instability with a dislocation occurring without a discrete external injury. Those patients requiring surgery due to unstable orthopedic injuries were excluded. Various mechanisms of injury included falling from 2 meters or less [n=46 (18%)], falling from higher than 2 meters [n=9 (4%)], sports related injuries [n=148 (59%)], motor-vehicle accidents [n=12 (5%)], assault [n=25 (10%)], seizure [n=10 (4%)], and other [n=2 (<\1%)].

Prognostic Factor: Only age-group and gender were independently predictive of recurrent instability. Other prognostic factors considered significant on univariate survival analysis but failing to remain independently predictive of recurrent instability included: presence of general ligamentous laxity, participation in sports, level of sport played, return to full activity or work at six weeks, absence of a greater tuberosity fracture, and presence of a nerve injury.The Outcome: Recurrent shoulder instability, operationally defined as the presence of either an additional radiographically confirmed anterior dislocation or symptoms of recurrent anterior subluxa tion or dislocation accompanied by positive findings on both an anterior apprehension test and an anterior load-atest .Thereest.There was a well-defined sample at a uniform (early) stage of illness. Follow-up was long enough; can't tell if follow-up was complete. There were not blind, objective outcome criteria. Adjustment was made for other prognostic factors. There was validation in an independent test-set of patients.

The Evidence:

Comments:Are the results valid?This study was a prospective longitudinal cohort study with data collected from a consecutive, unselected series of patients, ages 15 to 35, residing locally and initially presenting to an emergency department with subsequent follow up at an orthopedic shoulder clinic within one week of initial injury. This is a strong study design for prognosis but does suffer from a few minor threats to validity. Based on demographic and objective information available, it appears that the study subjects were adequately homogenous at baseline. The mean time for the recurrence of instability in those who developed it was 13.3 months (95% CI = 11.0 to 15.7 months). The mean duration of follow-up for those subjects that did not have recurrence was 46.7 months (39.7 to 53.8 months). Since the upper boundary of the 95% CI for recurrence does not overlap the lower boundary of follow-up we can be confident that follow-up was of adequate duration. It appears that follow-up may have been complete. The authors report that 34 subjects (13.5%), not known to have recurrent instability, were withdrawn at the two-year follow-up. An additional 23 patients who would have been lost to follow-up were contacted by telephone or mail and episodes of recurrent dislocation were recorded.

Perhaps the most significant threat to validity was the lack of blinding in determining recurrent instability. Radiographic evidence of subsequent dislocation is objective and of little concern. However, physical examination of patients with subjective complaints of instability was carried out by the senior author, who was not blinded to the prognostic factors. The authors adjusted for multiple prognostic factors that would logically have made subjects more prone to instability. However, it may have strengthened the study to include additional prognostic factors such as a history of previous shoulder trauma that did not result in dislocation or previous subjective complaints of subluxation. The authors validated and compared their findings to multiple previous studies that had determined age and gender to be independent prognostic factors.

What are the results? Of the 252 subjects initially enrolled in the study, 127 (50.4%) were known to develop recurrent instability by the end of two years. If the number of subjects lost from the study during this period was taken into account, 58.3% were known to develop instability. On survival analysis, 55.7% (95% CI = 49.1% to 62.2%) of the cohort had recurrent instability develop during the same time period. Further strength is added to this point estimate in that it is very close to the cumulative recurrence rate (58%) based on multiple previous studies reported in the appendix of the article. The operational definition of recurrent instability seems adequate, reasoning that some individuals alter their lifestyle based solely on subjective symptoms of instability and apprehension. For a young, active population, the prospect of achieving shoulder stability may outweigh the risks of surgical intervention. Of the 150 subjects that had recurrent instability develop during the entire length of this study, only 16 had subjective complaints with positive clinical signs; 134 actually had repeat dislocations. These numbers strengthen the sufficiency of the operational definition, considering that the clinical examination was not blinded. Although the authors report significance of multiple prognostic factors on univariate analysis, only gender and age remained independently predictive of recurrent instability on multivariate analysis. No P values were reported for the prognostic factors failing to achieve independence under multivariate analysis. Some of the prognostic factors, such as the presence of a greater tuberosity fracture, participation in sports, or generalized ligamentous laxity, may still influence clinical decision making if they were to approach statistical significance. Based on the Cox Regression Model, males were 2.53 (95% CI = 1.329 to 4.818) times more likely to develop recurrent instability than females during the length of the study. The regression coefficient (B) for the prognostic factor of age is negative. This means that the likelihood of recurrent instability decreases as age increases. Exponential B for age is less than 1.0. This is interpreted as the prognostic factor, age, being protective against recurrent instability. Since the confidence intervals surrounding both exponential B values do not cross the null value of 1.0, we can be confident that age and gender are truly prognostic for the development of recurrent instability. How can I apply the results to patient care?Age and gender are both very easily determined prognostic factors in a clinical setting. A large percentage of subjects recruited into this study were young, active males. This allows the results of the study to be easily generalized to my patient population. Additionally, I would consider the duties of a soldier very similar to the contact and non-contact sports participation reported for many of the study subjects.

The results of this study are important to clinical decision making within my current patient population in that recurrent instability may occur during physically stressful situations such as combat. Not only would recurrent instability decrease readine ss and put the soldier at risk, but may also put fellow soldiers at risk during combat situations. As reported by the authors, 2 years may be a good marker in a patient's rehabilitation. Nearly 87% of all recurrent instability occurred within the first two years. After that milestone, the likelihood of instability decreased significantly. Due to the strong study design, minimal threats to validity, and good precision of the prognostic factor estimates, I plan on keeping a copy of the estimated probabilities in my exam room to educate patients on their likely prognosis.

Appraised by: Chad M. Rodarmer; Saturday, March 03, 2007 Email: chad.rodarmer@us.army.mil

Short and Long term effectiveness of Manual Physical therapy vs Steroid injection

For patients 18-65 years of age, steroid injection for lateral epicondylitis shows no benefit over manual physical therapy in providing relief of symptoms at 6 weeks post treatment. However, it does appear that receiving a steroid injection poses significant risk for failing to experience complete or much improved relief at 1 year when compared to manual physical therapy alone.

Level of evidence: 1b

Citation/s: Bisset L, Beller E, Jull G, Brooks P, Darnell R and Vicenzino B. Mobilization with movement and exercise, corticosteroid injection or wait and see for tennis elbow: randomized trial. British Medical Journal. 2006; 333; 939-944.

Lead author's name and fax: Leanne Bisset; correspondence to B Vicenzino: b.vincenzino@uq.edu.au

Three-part Clinical Question: In a 35 y/o Navajo male laborer with right lateral epicondylitis is steroid injection superior to manual physical therapy in eliminating the patient's pain in the short term and in providing long term pain relief.

Search Terms: PubMed Clinical Queries for Therapy (narrow and specific) with search terms "Manual therapy" and "Lateral epicondylitis" which yielded 18 results. The article chosen was most relevant to my clinical question and most recently released as well.

The Study:Single-blinded randomized controlled trial with intention-to-treat analysis reported by the authors as performed but not actually performed.

The Study Patients: The study included 198 patients in Brisbane Australia age 18-65 years old with a clinical diagnosis of lateral epicondylitis of at least 6 weeks duration who had not had other previous treatment by another health care practitioner in the past 6 months. Patients were volunteers recruited via advertisement over a period of two years. Inclusion criteria were age 18-65 and at least 6 weeks of pain over the lateral elbow that increased with palpation of lateral epicondyle, gripping activity, resisted wrist or finger extension (2nd and 3rd fingers). Exclusion criteria were any treatment of the elbow pain by healthcare practitioner within the preceding 6 months, bilateral elbow symptoms, cervical radiculopathy, any other elbow joint pathology, peripheral nerve involvement, pervious surgery to the elbow, a history of dislocation, history of elbow fracture, history of tendon rupture, systemic or neurological disorders, any other pathology of the UE joints, and contraindication to steroid injections.

Experimental group (N = 66; 63 analyzed): For the purpose of answering my clinical question, I have chosen to look at only the Steroid Injection group as an experimental group despite the fact the this study included a manual physical therapy group as well. The participants assigned to the corticosteroid injection group received at least one but up to two standardized corticosteroid injections over the course of two weeks based on the patient's symptoms and treating practitioner's clinical discretion. In addition, these patients received an education pamphlet on the disease process and practical advice for self management techniques.

Control group (N = 65; 65 analyzed): The participants assigned to the physical therapy group received a total of eight 30 minute treatments over the course of six weeks. I have chosen to use the physical therapy group as the control in this review despite the fact that there was a wait and see group in this study as well. The treatments consisted of elbow manipulation and therapeutic exercise as well as a home exercise of self treatment of manipulation and resistance therapeutic exercises. The treatments were administered by a total of 6 physical therapists trained in the specific physical therapy regimen in order to standardize the treatment intervention. In addition, these patients received an education pamphlet on the disease process and practical advice for self management techniques.

The Evidence:

Comments: Are the results valid? The authors performed a large amount of data analysis comparing many outcome measures between three groups. For the purposes of this review, I focused on two of the three groups (Manual physical therapy and Steroid injection groups) and only one outcome measure (patient reported Global improvement). Overall, the results of this study that were applicable to answering my clinical question were valid. The authors randomized the participants to their respective groups [three total: Wait and See (WS), Manual Physical therapy (PT), and Steroid injection (SI)] in a concealed manner. Normally, blinding of participants is important to ensure the validity of the results of a randomized trial. However, in this study participants were aware of their group assignment but this did not seem important because this study did not use a placebo treatment. All patients, despite group assignment, were given a reasonable treatment to include the true control group (wait and see group) who received ergonomic education, education on stretching, education on use of thermal modalities and braces to help control the pain. Due to the lack of placebo and the fact that all patients received legitimate treatment approaches, I do not think that knowing group assignment would have biased the results of the study. Equally important, the authors report that the group allocation was concealed from all study personnel throughout the entire study including data analysis. This is very important in regard to the collection of outcome measures data because of the potential for bias when administering the outcomes tools and collecting outcomes measures data. Details on the process by which the outcome measures were obtained are lacking in the manuscript of this study. This poses a small potential threat to the validity of the data because of bias that may have occurred during the outcome measures data collection. If the authors' assertion that all study personnel were blind to group allocation is true then this potential bias in data collection is not present. I found their blanket statement regarding blindness to group allocation to be very difficult to believe as the clinicians providing treatment had to be aware of group allocation in order to provide the appropriate treatment. Due to the lack of a placebo group, I do not think that the treating clinician's knowledge would have biased the results because the patients were also aware of the treatments they were receiving and thus their group of assignment.

Regarding the similarity of patients between the groups, all groups were similar in age (WS = 47.3 +/- 8.1 years; PT = 47.9 +/- 7.2 years; Steroid Injection (SI) = 47.8 +/- 8.2 years. Participants were also similar in gender allocation (WS= 36% female; PT = 32% female; SI = 38% female), mean pain free grip force ration (WS = 48 +/-21.5%; PT= 42.4+/- 20.8% ;SI = 38.4 +/- 17.1%), mean pain severity rating out of 100 in the last week (WS = 61.3+/-22.6; PT = 57.5 +/-25.0; SI = 53.5+/- 23.0), and Pain free function rating out of 100 (WS = 76.7+/-21.3 ; PT = 75.4 +/- 18.9; SI = 80.8 +/- 16.1). The other characteristics were reported and seem similar but standard deviations were not reported for involvement of dominant hand, previous episodes of lateral elbow pain, alleged cause of pain, and employment status. Therefore, I am unable to determine if the patients were truly similar in these characteristics. Inability to determine the similarity between groups regarding these characteristics is a small threat to the validity of the results.

Regarding follow-up for this study, the follow-up was fairly complete for the two groups that were of interest in answering my clinical question. Participant data was analyzed according to the group which they were assigned except for 3 patients who were eliminated from the PT group due to failure to follow protocol. It seems reasonable to eliminate patients if they did not adhere to the study protocol however this fails to follow and intention to treat analysis reportedly performed by the authors. I will address the implications of this in the following section. I did not see any threat to the validity of the results in the methods of administering the treatment regimens. The physical therapy group used six different therapists which would add some validity to the results because the treatments were not administered by one individual. It is therefore less likely that any treatment effect from the Physical therapy group would be as a result of only the skills of one very talented manual therapist. Conversely, there is a potential threat to the validity of the results due to sampling bias. In other words, all patients were gathered via the participants' response to an advertisement. It is reasonable to assume that these patients might be somewhat different from the average patient presenting to a primary care physicians office with complaints of elbow pain. However, I do not think that the differences between these two groups are such that it destroys the validity of this study's results.

What are the results? The outcome of interest for answering my clinical question was whether or not patients reported complete or much improved relief on the Global improvement scale. The results of this study indicate that manual physical therapy (PT) is an equally effective treatment when compared to steroid injection (SI) at 6 weeks post treatment. The relative risk reduction for failure to achieve complete or much improved relief via steroid injection when comparing it to manual physical therapy is 38% (95% CI = -6 to 83%). Due to the fact that the confidence interval was very wide to include a 0% risk reduction, one must assume there is no difference between the two treatments. This is equally true when we consider the number needed to treat (NNT) for this particular comparison between SI and PT (NNT = 7, 95% CI= -49 to 3). Essentially the confidence interval for the NNT actually includes a negative range which means that the range of possible outcomes includes the possibility that treating patients with a steroid injection could result in more patients failing to achieve complete or much improved recovery (or the NNT when negative = number needed to harm or NNH).

The authors reported that the study performed an intention to treat analysis of the data. Essentially this means that they were supposed to include all patients despite any failure to adhere to the protocol. All patients from the SI group were included in analysis of Global improvement data but not all patients for the PT group were included in that analysis. I therefore performed an analysis by placing the data points for the three patients eliminated back into the data pool. Assuming a worse case scenario for comparison to the SI group (those PT patients did get better), the data did not change the conclusion that there was no significant treatment effect of steroid injection when compared to manual therapy ( RRR= 32%, 95% CI = -15% to 80%; NNT = -21 to 4). Therefore, the authors' assertion that the exclusion of those patients only minimally changed the results of the analysis was true for the 6 weeks data.

The results of the comparison of PT and SI groups indicated that steroid injection is potentially harmful to patients in the long term because more patients reported failure to achieve long term complete or much improved relief of their pain (Relative Risk Exacerbation = 413%, 95% CI = 209% to 619%; NNH = 4, 95% CI = 3 to 8). The confidence intervals on the data for relative risk exacerbation are fairly wide but the low end is still fairly large in magnitude indicating that the risk of exacerbation at one year is clinically significant. In addition, the NNH is fairly low with a fairly narrow confidence interval such that a physician does not need to treat many individuals with the steroid injection to have one patient experience failure to have complete or much improved relief of symptoms. Performance of the intention to treat analysis does not significantly change this conclusion (Relative Risk Exacerbation = 430%, 95 % CI = 220% to 639%; NNH = 4, 95% CI = 4 to 7). How do these results apply to my patient? In regard to my patient, it appears there is no benefit to receiving a steroid injection especially when considering that a patient has a good chance of being worse at 1 year post treatment than if they receive manual physical therapy alone. Given that my patient is in the age range and only different than the patients in this study by ethnicity, I would recommend to my patient not to have a steroid injection due to the risk of harm at one year and the fact that they are not likely to experience more relief than manual physical therapy at 6 weeks.

Appraised by: Matthew J. Armentano; Review started Saturday, February 24, 2007 Email: matthew.armentano@ihs.gov

Comparing the Performance Enhancing Effects of Squats on a Vibration Platform with Conventional Squats in Recreationally Resistance-Trained Men

No conclusions about the relative efficacy of whole body vibration on the 1 rep max squat strength or counter movement jump height can be made from this study.

Citation:Ronnestad BR. Comparing the Performance Enhancing Effects of Squats on a Vibration Platform with Conventional Squats in Recreationally Resistance-Trained Men. J Strength Cond Res, 2004, 18(4),839-845.

Lead author's name and fax: Bent R. Ronnestad, bent.r.ronnestad@hit.no

Three-part Clinical Question: For a 30 year old recreational weight lifting male, is a resistance training program combining squats on a whole body vibration platform more effective than performing squats alone for improving 1 rep max (1RM) squat strength and countermovement jumping height (CMJ)?

Search Terms: whole body vibration AND resistance training AND strength

Treating patients who have hip osteoarthritis (OA) with specific manual therapy techniques for 5 weeks produced significant improvement over patients treated with individualized exercise therapy (number needed to treat was 3). Level of evidence is 1b.

Citation/s: Hoeksma H, Dekker J, Ronday H, Heering A, Van der Lubbe N, Vel C, Breedveld F, Van Den Ende C. Comparison of manual therapy and exercise therapy in osteoarthritis of the hip: A randomized clinical trial. Arthritis and Rheumatism. 2004: 51(5):722-729.

Lead author's name and fax: Hugo L. Hoeksma, PT. E-mail is h.hoeksma@antonius.net

Three-part Clinical Question: In a 75 year old woman with hip osteoarthritis, would a individualized physical therapy program of exercise and manual therapy be more effective in reducing pain and increasing function than rest?

Search Terms: Using PubMed's clincal queries I entered the following narrow, specific search: (hip AND osteoarthritis AND exercise) AND (randomized controlled trial[Publication Type] OR (randomized[Title/Abstract] AND controlled[Title/Abstract] AND trial[Trial/Abstract])). I had 32 hits, one of which was the Hoeksma et al article.

The Study: Single-blinded concealed randomized controlled trial with intention-to-treat.

The Study Patients: Fifty six patients in the manual therapy group had a mean age of 72 (+/-) 7 years. Thirty eight were female, 18 were male. The exercise group had 53 patients with a mean age of 71(+-) 6 years. Thirty eight were female, 15 were male.

Control group (N = 53; 50 analyzed): The exercise treatment group was established as the control group as it has already been reported to be effective in patients with OA of the hip. The exercise therapy program was planned by a physical therapist according to the individual symptoms of the patient. Exercises focused on increasing muscle function, improving range of motion, decreasing pain, and improving walking ability, All patients were treated twice weekly for a period of 5 weeks with a total of 9 treatments, with the first session used to design the treatment protocol to the individual patient.

Experimental group (N = 56; 53 analyzed): The manual therapy treatment group was established as the experimental group in order to compare it to the exercise, or control group. The manual therapy sessions started with stretching techniques of identified shortened muscles associated with the hip joint. Secondly, a traction technique was performed on the hip, followed by a high velocity thrust technique. All patients were treated twice weekly for 5 weeks with a total of 9 treatments, with the first session used to design the treatment session according to the individual patient.

The Evidence:

Comments: Are the results valid? The patients were randomized and the randomization was concealed using numbered and sealed envelopes. Patients were analyzed in the groups to which they were randomized, including those who ended up getting a hip arthroplasty during the follow up time period. The patients in the manual therapy group and in the exercise group were similar with respect to their baseline characteristics. All these factors strengthen the validity of the study. Both patients and clinicians were aware of group allocation, with separate therapists treating the exercise group and the manual therapy group. The study does not indicate if the patients were aware of the other group's intervention, nor does it address the skills, education, length of time in practice or general expertise of the various therapists that were assigned to each treament arm; only that they were trained in the duties required for each group. If the study participants in one of the groups thought they were being assigned to the "less desirable" group, this bias could have been reflected in their 5 week assessment of improvement of their main complaint. Additionally, if all the manual therapists had more experience, education, etc, than the exercise therapists, a bias could have been introduced as well. This bias would have resulted in the exercise group patients possibly not having all their individual needs met during assessment by a less experienced therapist, thus altering the improvement assessments at the 5 week mark. The outcome assessor was blinded to the patient's group assignment and follow-up was done as completely as possible, with a total of 3 patients being lost to follow-up in each group at 5 weeks.

What are the results? Treatment of the patient with hip OA using specific manual therapy techniques reduced the risk of no improvement in the patient's main complaint by 62% relative to that occurring among the exercise group participants. The precision of this estimate is moderate - we can be 95% confident that the true relative risk reduction for this population lies between 27% and 98%. The RRR for hip arthroplasty between the 2 groups was negligible, with a very wide confidence interval. Secondary outcomes included quality of life, hip function and walking speed. In the quality of life assessment, no significant differences were noted between groups for pain and physical function, and the exercise group had a significantly higher group mean (11.3% higher) for the role functioning component. The manual therapy group had significantly higher hip function scores (11.2% higher) and significantly faster walking speeds (8.2% faster) than the exercise group. The precision of these secondary outcomes varied: for the role functioning component the CI was wide with an interval from 1.1 - 21.5, for the hip function score the CI was narrower with an interval from 6.1 - 16.3, and for the walking speed the CI was wide, with an interval from 0.5 to 16.7. These results do not include the null hypothesis, but only the hip function score is precise enough to use with confidence in a clinical setting.

How can I apply the results to patient care? The baseline demographics of the study participants matched the characteristics of my patient quite well. In particular the majority of the patients were either classified as having moderate or severe OA and willing to participate in a study to improve upon the chief complaint, instead of resorting immediately to surgery. All clinically important outcomes were considered, including the long term outcome of surgery. Pain, quality of life, hip function, walking speed, improvement of chief complaint and surgery were outcomes that covered the spectrum needed, including issues that interest both the patient and the clinician. The treatment benefits are worthwhile in that the number of patients needed to treat is 3 to avoid a negative outcome (no improvement in primary complaint). While manual therapy showed benefit over exercise in this study, both showed a positive improvement in different measures. In addition, the long term follow-up showed no difference in the number of resulting hip arthroplasties. The specialized training required for the manual therapy component would be worth the time and cost if it was of a reasonable length of time and easy to learn. If the training involved numerous courses that built upon each other at a considerable cost, I would be tempted to focus on the stretching and joint mobilizations without the high velocity thrust and utilize on the individualized exercise component as well. Knowing that the positive outcome of each of the groups could be attributed to the placebo effect, one must weigh the time and effort of coming in to the clinic against the perception of improvement on the part of the patient. If the patient is willing and the clinic can accommodate, then this treatment is worthwhile. If the visits are inconvenient for the patient and the clinic cannot accommodate, then the treatment, although promising, may not be worthwhile. This study looked at a subjective measure of improvement for its primary outcome. If the patient's perception of improvement (including improved hip function) is of primary importance, then I would recommend the treatment. If the outcome of interest for the patient is avoidance of surgery, I cannot recommend this treatment with confidence.

Appraised by: Amy Eschenberg, MPT. October 17, 2007, Email: Amy_Eschenberg@baylor.edu

        Back to CAT Express Ortho         Back to Top

Is Manual Therapy and Exercise Effective in Treating Patients with Knee Osteoarthritis?

For a 60 year old male with knee osteoarthritis it does appear an 8 visit course of physical therapy treatment, consisting of specific manual therapy and exercise to the knee, hip, ankle, and spine as needed, can reduce the risk of total knee surgery, increase walking distance, and reduce pain compared to placebo. However, it is difficult to ascertain the strength of the treatment effect, due to wide confidence intervals and loss of subjects to follow-up in dependent variable calculations. This study is likely valid and has strong methodology, but future conduction of studies with larger sample sizes and longer follow-up duration are necessary to provide additional support for the original question and determine if the patient would benefit significantly over a longer time period. 1b- level of evidence according to Sackett's hierarchy.

Citation/s:Deyle GD, Henderson NE, Matekel RL, Ryder MG, Garber MB, Allison SC (2000.)

Effectiveness of manual physical therapy and exercise in osteoarthritis of the knee: A randomized, controlled trial. Annals of Internal Medicine, 132(3), 173-181.

Lead author's name and fax: Gail D. Deyle, MPT.

Three-part Clinical Question: For a 60 year old male with bilateral osteoarthritis of the knees, is a physical therapy program of manual therapy plus guided exercises more effective than placebo in reducing pain, improving function, and reducing the likelihood of surgery?

Search Terms: "knee osteoarthritis" AND manual AND therapy

The Study:Single-blinded randomized controlled trial without intention-to-treat.

The Study Patients: Patients were randomly divided into two groups. The treatment group consisted of 42 subjects (15 men and 27 women) and the placebo group had 41 subjects (19 men and 22 women). The study began with 83 participants, however 14 (9 in the treatment group, 5 in the placebo group) were lost to follow-up. The mean age in the treatment group (n=33) was 60 +/- 11 years and in the placebo group (n=36) was 62+/- 10 years. Body mass index was similar between groups at 31.1 +/- 6.7 kg/m squared and 30.4 +/- 5.0 kg/m squared respectively. In the treatment group the mean duration of symptoms was 81.7 +/- 88.2 months and in the placebo group it was 57.2 +/- 96.1 months. The mean WOMAC score in millimeters was comparable in both groups at 1046.7 +/- 455.0 and 1093.5 +/- 497.1 respectively. The mean distance on the 6-minute walk test again was comparable with the treatment group's mean distance at 431.2 +/- 120.2 meters and the placebo group at 402.9 +/- 104.5 meters at baseline. 33% of the treatment group and 36% of the placebo group had bilateral knee symptoms and 83% and 81% utilized medication to aid in reducing their symptoms related to osteoarthritis respectively. The treatment group did appear to be more active with 46% participating in vigorous physical activity 3 or more days per week, with only 17% of the placebo group having a similar activity level. The severity of radiographic findings at baseline were similar overall. The treatment group demonstrated Grade 0 for 9% of subjects, Grade 1 for 31%, Grade 2 for 22%, Grade 3 in 31%, and Grade 4 in 6%. The placebo group had 9% of subjects who demonstrated Grade 0 osteoarthritis, 18% with Grade 1, 38% with Grade 2, 29% with Grade 3, and 6 % with Grade 4 severity. The patients who did not complete the study were also compared for homogeneity and were similar on most prognostic factors except demonstration of higher WOMAC scores. In the treatment group, 4 subjects withdrew due to unrelated medical reasons, 1 sustained a knee injury in an altercation, 2 had transportation difficulties, 1 needed to care for terminally ill husband, and 1 withdrawal was for unknown reasons. In the placebo group, 1 patient developed cardiac problems, 1 acquired plantar fasciitis, 1 was disqualified after receiving a cortisone injection to the knee, and 2 had transportation difficulties.

Inclusion criteria: diagnosis of osteoarthritis of the knee based on fulfillment of the clinical criteria developed by Altman and Colleagues, eligible for military health care, sufficient English skills, no surgical procedure on either lower extremity in the past 6 months, no physical impairment unrelated to the knee that would prevent safe participation, and live within a one-hour drive of the physical therapy clinic.

Exclusion Criteria: Could not attend the necessary number of visits or received a cortisone injection within the previous 30 days.

Control group (N = 41; 41 analysed target outcomes): The control (placebo) group after receiving a thorough examination of the knee, hip, ankle, and lumbar spine, received treatment by the physical therapist consisting of sub therapeutic ultrasound for 10 minutes at an intensity of 0.1 W/cm squared and 10% pulsed mode to the area of the knee with symptoms. The amount of time directly spent with a therapist was 30 minutes. The placebo group received treatment twice weekly for 4 weeks, for a total of eight clinic treatments.

Experimental group (N = 42; 42 analysed target outcomes): The Experimental Group after receiving a thorough examination of the knee, hip, ankle, and lumbar spine, received manual physical therapy and a supervised exercise program. The manual therapy was performed by therapists with formal training in manual therapy. The treatment techniques consisted of passive physiologic and accessory joint movements, muscle stretching, and soft tissue mobilization, primarily to the knee. The same treatments were also administered to the lumbar spine, hip, or ankle if these areas demonstrated limitation in active or passive movement, were symptomatic, or were contributing to overall lower limb dysfunction. The treatment group also performed a closely supervised standardized knee exercise program at each of the sessions. The exercise program consisted of active range-of-motion exercises for the knee, strengthening exercises for the hip and knee, muscle stretching for the lower limbs, and riding a stationary bike. Each subject was seen twice weekly for 4 weeks, for a total of eight clinic treatments. The time spent in direct therapist care each session was 30 minutes with an additional 30-45 minutes dedicated to performing their exercises in the clinic.

The Evidence:

Outcome	Time to Outcome	CER	EER	RRR	ARR	NNT
Total Knee Surgery	1 year	0.195	0.048	75%	0.147	7
	95% Confidence Intervals:
5% to 100%	0.010 to 0.284	4 to 105
Steroid injection	1 year	0.146	0.048	67%	0.098	10
	95% Confidence Intervals:			-19% to 100%	-0.028 to 0.224	NNT = 4 to INF; NNH = 36 to INF
Non-Event Outcomes		Time to outcome/s	Control group	Experimental group	P-value
WOMAC scores (mm)		4 weeks	921.2 (730.8-1112.1)	505.2 (438.0-572.4)	<.05
6 minute walk distances (m)		4 weeks	402.1 (359.9-444.3)	484.0 (442.7-525.3)	<.05

Comments:This critical appraisal topic is designed to aid the reader in determining whether their patient with osteoarthritis of the knee may benefit from a conservative physical therapy program involving manual and exercise therapy. The first step in determining if the therapy is a robust choice of treatment is to determine if the study evidence is valid. Did the subjects in the placebo (non-therapeutic ultrasound) and the treatment group (focused manual therapy and exercise) have a similar prognosis? Overall, the patients who completed the study were very similarly matched, through the randomization process, within the treatment and placebo groups. Age, body mass index, duration of symptoms, WOMAC scores, 6 minutes walk test values, bilateral symptoms, and medication use variables were all similar at baseline. A few minor differences did exist. The placebo group was more sedentary than the treatment group and the groups differed in the severity of radiographic findings. However, only Grades 1 & 2 differed significantly with the treatment group having 4 more subjects with Grade 1 osteoarthritis and 7 less subjects with Grade 2 findings. This could potentially bias the results in favor of the treatment group, however the effect is likely to be small.

Subjects were randomized to the treatment and placebo groups by a random number generator. Randomization was concealed by using folders, in which the therapist was required to draw the next folder from the file that determined the group of assignment. Outcome assessors were blind to the patient's treatment groups which allowed inherent objectivity in measuring dependent variables. However, blinding in this study was not perfect as the therapist was aware of the treatment the patient received, thereby allowing for potential bias of differential administration and reinforcement in regards to the treatment. In addition, the patients knew which treatment they were receiving, although they were not aware of the treatment the other group was undergoing. In light of time and effort placed upon their particular therapy, bias may have been introduced. It is concerning and a potential threat to validity that 21% of the treatment group and 12% of the placebo group disenrolled from the study. However, they were not lost to follow-up entirely as target outcomes were assessed in all 83 study participants. But, it does leave an open door as the subjects disenrolled from the study were not included in analysis of the dependent variable. This could greatly affect the unbiased comparison provided by randomization. Returning to the question of validity, this study has strong potential to be a valid study, despite some potential risks. It likely provides an unbiased look at the effectiveness of specific manual therapy and exercise versus placebo in the treatment of knee osteoarthritis.

To better analyze the results of this study, we will look at three facets of the outcomes: size of the treatment effect, the precision of the treatment effect estimate, and confidence intervals. The results of this study demonstrated positive treatment effects in all analysis variables to include: distance walked, WOMAC scores, knee surgery and steroid injection reduction. However, the results only reveal point estimates of the true treatment effect. How sure can we be that the point estimates of the results are plausible representations of the true effect? To answer that question let us look at the confidence intervals. Confidence intervals reveal the range of possibilities within which, 95% of the time, the true effect results lie. The treatment group demonstrated a reduction in WOMAC scores and an increase in distance walked in 6 minutes at the 4 and 8 week periods. Unfortunately, the confidence intervals about these dependent variables are wide, demonstrating a possible less substantial effect. Once again, another confounding variable potentially affecting the results of WOMAC scores and distance walked was the fact that participants who disenrolled from the study were not evaluated on these measures.

This study also demonstrates a significant treatment effect size in relative risk reduction. At one year, the treatment had reduced the risk of total knee surgery by 75% and the risk of steroid injection by 67%. Absolute risk reduction point estimates also point to a large treatment effect size with 15% of patients being spared a total knee surgery and 10% being spared a steroid injection. In addition, the results demonstrate a powerful statement that a clinician would need to treat only 7 patients in order to prevent one total knee surgery and 10 patients to prevent one steroid injection. The treatment effect size point estimates demonstrate the treatment is superior to the placebo in reducing adverse target outcomes. However, returning to the treatment's relative risk reduction of total knee surgery and confidence interval, the true reduction of knee surgery likely lies between 5% and 100%. Similarly in reducing the relative risk of a steroid injection, the true estimate likely lies between -19% to 100%. Therefore, potentially favoring the placebo (the placebo could potentially reduce the risk of steroid injection by 19%). Absolute risk reduction point estimate for steroid injection reduction also has a widely varying confidence interval. The treatment's true estimate likely lies between -0.028 - .224. Meaning it is likely the true effect of the treatment could spare up to 22% of the subjects from a steroid injection, but it is possible that the truth could lie in favor of the placebo (lower boundary indicates 3% of subjects spared a steroid injection by the placebo). Confidence intervals for absolute risk reduction in regards to total knee surgery are promising. The true absolute risk reduction likely spares 1% to 28% of the patients from total knee surgery by receiving the treatment, thereby demonstrating a definitive positive effect. Confidence intervals for the number needed to treat for preventing one steroid injection and total knee surgery were also wide. Therefore, the true number needed to treat likely lies between 4-105 patients to prevent one total knee surgery and 4-infinite number of patients necessary to treat in order to prevent one steroid injection. Based on the results and confidence intervals, it is safe to infer that the treatment does reduce the risk of total knee surgery and possibly steroid injections, however more research needs to be done to improve the strength of these inferences and enhance clinical significance.

The confidence intervals for relative risk reduction, number needed to treat, and absolute risk reduction are wide and variable, therefore was there a large enough sample size? Keep in mind, that many subjects must participate in order for a research trial to generate narrow precision estimates. This study does demonstrate definitive evidence in reducing the risk of a total knee surgery, but the precision of the point estimates could be improved by increasing the sample size. Most clinicians would agree that a relative risk reduction of even 5% (lower boundary of relative risk reduction confidence interval) of subjects receiving an expensive total knee surgery with an eight treatment course is worth the effort. The evidence to support the reduction of steroid injections is not so promising, therefore a larger sample size is necessary to provide greater support for the precision of the estimate. Based on the Deyle et al randomized clinical trial it is safe to infer that the 8 visit treatment does reduce the risk of total knee replacement at one year, but the results for reducing steroid injects is not definitive. In the future, a study which possesses a larger sample size is necessary to further support the findings found in this study.

Returning to the clinical question, "For a 60 year old male with bilateral osteoarthritis of the knees, is a physical therapy program of manual therapy plus guided exercises more effective than a placebo in reducing pain, improving function, and reducing the likelihood of surgery? ", it is likely with an 8 visit treatment course of specific manual and exercise therapy will reduce pain, improve function, and reduce the risk of a total knee surgery in comparison to placebo, but by how much is the question.

Appraised by: Jessica Feda, PT, OCS; Friday, October 20, 2006 Email: jfeda@hotmail.com

Patellofemoral pain syndrome (PFPS) still lacks good valid clinical tests.

This Critical Appraisal Topic's purpose is to help the reader to decide that the clinical testing done in this research was not sensitive enough for a clinician to use any one test on its own. Its level of evidence is 2B. It is an exploratory cohort study with good reference standards applied with the exception of blinding. The examiner and the subjects were not blinded. The clinical tests were evaluated in a small population of young adult males thus limiting its usefulness for any other population. Therefore, better-designed research is needed with a larger and more diversified patient population for clinical assessment of PFPS.

Citation/s:Haim A, Yaniv M, Dekel S, Amir H. Patellofemoral pain syndrome: validity of clinical and radiological features. Clin Orthop Relat Res. 2006;451:223-228.

Lead Author's name and fax: Amir Haim, MD. Fax: 972-2-5322925 Israel

Three-part Clinical Question: For a 22 year old male runner is there a clinical test that is valid for determining patellofemoral pain syndrome?

Search Terms: Patellofemoral pain syndrome and validity. My search yielded 7 hits, one of which was the Haim, Yaniv, Dekel, and Amir article.

The Study:The Study Patients: A total of 86 patients were enrolled in this study. They prospectively examined and followed 61 male soldiers with anterior knee pain (mean age, 19.4 +/- 1.2 years; range 18.5-20.2) and 25 control subjects (mean age 24.1 +/- 6.5 years; range 19-27 years). The control subjects consisted of male infantry soldiers who had served in the army for at least 2 years and were without a history of knee pain, injury, or surgery. The knee pain group was made up of infantry soldiers who had a prior history of knee pain and had been referred to the orthopedic clinic for further evaluation. The inclusion criteria consisted of anterior knee pain after increased physical activity, aggravated by walking up and down stairs, squatting, or sitting with knees flexed and accompanied by crepitus, giving way, and catching; persistent pain despite rest and conservative treatment for at least 3 months; and military discharge because of knee pain. Exclusion criteria were: no history of knee injury or surgery and no medical condition that would require discharge from the infantry unit. Duration of symptoms for the knee pain subjects was 19 months +/- 9.9 with a rangemonths .Independentdependent, non-blind comparison with a reference (gold) standard. There was an appropriate spectrum of patients. The gold standard was applied regardless of the test result.

Target disorder and Gold Standard: Anterior knee pain in young (18-29 years) males serving in army infantry units. Gold standard established in patients with previously diagnosed patellofemoral pain syndrome by a Medical Corps physician that were referred for additional evaluation by a central military orthopedic clinic. Control group also referred to the central military orthopedic clinic for other than knee orthopedic disorders.

Diagnostic test: All subjects completed a 9-question questionnaire at the time of their examination. It included demographics and army service details. The questionnaire established whether or not there was knee pain prior to military duty, if pain was bilateral or not, any knee injury to either knee during service time, what aggravates their knee pain, does their knee pain improve with rest, if they have ever experienced a locking sensation, and a family history of knee pain. All were clinically examined to rule out any ligament or cartilage tears that could cause anterior knee pain. In addition, all subjects underwent testing to rule out referred pain from the hip with supine hip internal rotation test and anteroposterior radiographs of the knees were checked to rule out other pathologic conditions. They used four tests to evaluate the sensitivity and specificity of patellofemoral pain syndrome; the patellar tilt test, the patellar apprehension test, the patella alta test, and the active instability test. In addition, they performed qualitative tests that looked at lower limb alignment (genu valgum and genu varum), foot posture (pes planus and pes cavus), quadriceps angle in standing (the Q-angle), presence of knee effusion assessed with the ballottement test, patellofemoral joint crepitus, presence of tenderness over the medial and lateral retinacula, medial and lateral patellar glide (as percentage of patellar width), and presence of squinting patella (inward tilt). They also documented all subjects' plain radiographic findings. Lateral and axial patellar radiographs were taken with subject's knees in 30 degrees flexion. They measured the Insall-Salvati index for patella alta, sulcus angle, and Merchant angle and Laurin angle. The presence of subchondral sclerosis and patellar subluxation was determined qualitatively from the axial radiographs.

The Evidence:

Comments: This was a prospective cohort study of army infantry soldiers who had persistent knee pain on average for 19 months (19 +/_9.9, range 4-60). This study did not blind their subjects. They had been referred to the orthopedic clinic because of their knee pain. A motivational bias by the subjects should be considered here. Some soldiers may magnify their symptoms while others may downplay their symptoms depending on their motivation to stay in the military service or receive a military discharge. The examiner, also, was not blinded to the group assignments and this could have introduced detection bias. That is the tendency to look more carefully for an outcome in one of two groups being compared. However, none of the tests were new so knowledge of the reference standard would have less influence of interpretation than say a new diagnostic clinical test. The reference standard was applied to all subjects regardless of the test results thus preventing a work-up bias.

The physical examination tests and radiographs taken were commonly used tests to evaluate for the presence of patellofemoral pain syndrome. The researchers primary purpose was to determine how specific and sensitive these main tests are for the diagnosis of patellofemoral pain syndrome in their routine patient population. The spectrum of patients were appropriate for this treatment facility since it primarily sees young military aged males and could readily use the data from the research. However, it could not necessarily be applied to all patient populations seen at orthopedic clinics, but perhaps generalized to young adults complaining of knee pain. Sensitivity was poor in the 4 main clinical tests the authors evaluated. They were all less than 50%. This means that less than 50% of the time the subjects with the disorder of interest would have a positive test. In the apprehension test their results indicated a sensitivity of 7%, which would mean 93% of the patients with the disorder of interest would be missed if this were the only test performed. In addition the positive likelihood ratio was .82 and the negative likelihood ratio was 1.02. Likelihood ratios in these ranges alter probability to a small and rarely important degree. In addition the CI's of both the negative and positive likelihood ratios cross over 1 indicating no real change of probability. The specificity of the apprehension test was 92% indicating that the test was better at detecting patients without patellofemoral pain syndrome. The sensitivity for the patellar tilt test, the active instability test, and the patella alta test were 43%, 25%, and 49% respectively. Ideally, a diagnostic test should have a relatively high sensitivity to be considered practical because the clinician is trying to rule out the disorder of interest. Therefore, none of these 4 tests should be used alone to rule out patellofemoral pain syndrome. The patella alta test had the highest sensitivity at 49% with only moderate specificity of 72%. This means that 51% of the subjects with the disorder of interest will be missed and that 28% will be false positives. These numbers would not lend a lot of confidence to the tester for any diagnostic conclusions. In addition, the +LR of the patella alta test was 1.76 with a CI of 0.89-3.46 and the -LR was 0.71 with a CI of 0.50-1.00. These ratios indicate probability being altered to a small and rarely important degree and their CI's either crossed over 1 or included 1 indicating no real change of probability. The specificity of the patellar tilt test and the active instability test were 92% and 98%. Since specificity is the proportion of people without the target disorder in whom a test result is negative it could be said a positive active instability test effectively rules in the disorder. At first glance the patellar tilt test showed the greatest statistical promise. Its sensitivity was low at 43% and its specificity was high at 92% but the likelihood ratios gave the greatest insight to this test. It had a +LR of 5.33 (CI 1.37-20.78) and a -LR of 0.62 (CI 0.49-0.80). This meant if we had a pretest probability of 50% that a positive test would increase it to 84%. Even at the lower limit of the CI there is a positive posttest probability change (though small) from 50% to 55%. This change should allow a clinician to feel more confident that a positive test is more likely to mean a diagnosis of PFPS. A negative test, with a pretest probability of 50%, is lowered to 38%. This small decrease is not enough of a posttest change to feel confident about a negative result. Even at the lowest end of the CI limit the decrease is only to 29%. Again, leaving a significant amount of room for diagnostic error based on this test. The active instability test also had a good positive likelihood ratio of 13.00 but its CI was so large (.81 to 209.98) that it would not change or assist in deciding on a clinician's test or treat threshold. Its negative likelihood ratio was .75 (CI of 0.65-0.87). This is very close to 1 and therefore has little impact in posttest probability even at the lower limits of the CI. At the very best a positive result would allow the clinician, with some confidence, of ruling in patellofemoral pain. A larger study (with the same high specificity) may help make the CI's more precise and cross that test to treat threshold with a positive test result.

The four main clinical tests addressed here are all easily reproducible by a variety of clinicians and do not require a lot of training or expertise. This makes these tests useful in most clinical settings. However, the validity of these tests is not high enough to use as the sole determining factor of patellofemoral pain syndrome. Further research using a larger and more varied patient population may yield better validity for one or more of these tests. As it stands, a patient with a constellation of positive PFPS tests, symptoms, and history appears to be the best evaluative tool that is available to a clinician. Based on this research there is no single or even group of tests that would solidly diagnose PFPS. The results of this research would best be applied to young active adult male populations because a more varied group (especially in age) are more likely to have competing symptoms seen with other knee disorders. Again, this research would not change my management strategy of PFPS. Finally, a patient will not be better off as a result of these tests but they do provide a stepping-stone towards getting the patient into the correct rehabilitation program.

Appraised by: Heather E. Khan; Wednesday, January 24, 2007Email: Ehenfeld@comcast.net

Calcaneal taping decreases patient's pain in the short-term, but does not improve function.

Clinical Bottom Line: The calcaneal taping technique is an easy immediate treatment with very little risk that can be used to alleviate pain for patients with plantar heel pain in the short term. VAS decreased significantly more with calcaneal taping technique than with both the control group (p<.001) and sham-taping group (p<.001). None of the treatment methods demonstrated improvements in the patient specific functional scale (PSFS). There were very few threats to validity in this study, but the utility of the results is limited due to the short-term follow-up, lack of other treatment included and no comparison to other taping methods or orthotics. This study is level 1b evidence with 100% follow-up.

Citation/s: Hyland MR, et al. Randomized Controlled Trial of Calcaneal Taping, Sham Taping, and Plantar Fascia Stretching for the Short-Term Management of Plantar Heel Pain. Journal of Orthopedic and Sports Physical Therapy. 2006; Vol 36-6: 364-371.

Lead author's name and fax: Matthew R. Hyland, Fax unavailable

Three-part Clinical Question: For a 41 year old female with unilateral plantar heel pain will plantar taping provide more relief in the short term than placebo treatment?

Search Terms: Plantar heel pain, plantar fasciitis, taping The Study: Single-blinded concealed randomized controlled trial with intention-to-treat.

The Study Patients: A total of 42 subjects were randomly assigned to one of four groups: stretching alone, calcaneal taping, control group and sham taping. The inclusion criteria was appropriate to target those individuals who would best benefit from the treatment. This criteria included: age between 18-65, pain with first steps upon walking, pain located at the heel or plantar surface of mid-foot and presence of and everted calcaneus greater than or equal to two degrees. The stretching group (N=10, 8 males) had a mean age of 34.1 (5.9). The calcaneal-taping group (N=11, 5 males) had a mean age of 45.5 (12.0). The control group (N=10, 3 males) had a mean age of 37.6 (10.1). The sham-taping group (N=10, 5 males) had a mean age of 40.Descriptivescriptive statistics were provided and their were no significant differences between groups at the start of the study for anything except gender. Control group (N = 10 & 10; 10 & 10 analyzed): The control group (N=10) was provided no treatment for the one week duration of the study. The sham-taping group (N=10) had their heels taped with cover roll and Leukotape but it was applied in a way that did not attempt to control the position of the calcaneous. Subjects in the sham-taping group returned half-way through the week to get their foot retaped. Experimental group (N = 10& 11; 10 & 11 analyzed): There were two different treatment groups: stretching alone and calcaneal taping alone. The stretching group (N=10) received passive stretching of the ankle plantar flexors and fascia on day one, and day 3 or 4 in a physical therapy office. A description of the stretches was included in the article. Subjects in this group were instructed not to perform the stretches at home. The calcaneal taping group (N=11) received calcaneal taping alone. Cover roll was applied to the plantar heel and foot and then covered with Leukotape. The Leukotape layer was applied laterally to medially in an attempt to pull the calcaneus medially. Specific directions and photos were included in the article. Both taping groups were instructed to keep the tape on for 24 hours a day. Subjects in the calcaneal-taping group also returned to get their foot taped again in the middle of the week.

The Evidence:

VAS scores with 95% Confidence Intervals

Non-Event Outcomes

Comments: The results to this study are valid and there are very few threats to validity. The subjects were randomly assigned to one of the four groups and were blinded to their group assignment. The authors did not mention if the individuals doing the statistical analysis were blinded to treatment received. All subjects were accounted for at the conclusion of the study and were analyzed in the groups they were initially assigned to. Groups were treated relatively equal aside from their different treatments. The main difference is that all but the control group had some sort of hands on treatment, either taping or stretching, done in the clinic at two different times. This additional time being seen could increase the placebo affect taping and stretching could have. All four groups were homogenous in regards to most of the descriptive statistics at the start of the trial. Groups only varied in the amount of males and females they had in them. For instance the stretching group had 8 males and 2 females while the control group had 3 males and 7 females. The two outcome measures, Visual Analog Scale (VAS) and Patient Self-Reported Functional Scale (PSFS) have been shown to be reliable and capture the results were are interested in for this study.

The statistical analysis done for this study was appropriate. The authors performed Tukey post hoc analysis to account for the multiple groups. There were significant differences both within and between groups at the conclusion of the study. The study found significant improvements for pain in stretching (-1.7 decrease on VAS, CI 4.08-5.1, P<.001), sham-taping (-0.4 decrease on VAS, CI 5.48-6.92, P<.05) and calcaneal taping groups (-4.3 decrease on VAS, CI 1.49-3.91, P<.001) on VAS scale. Calcaneal taping also significantly improved pain compared to sham taping and stretching (p<.05). The change in VAS score for calcaneal taping is clinically significant because if a patient can improve by 50% on the VAS scale they will definitely notice a positive change. The 95% confidence interval of calcaneal taping, which was not recorded by the authors, is larger than preferred, however, the interval does not overlap the Pre VAS CI therefore the results are still of value. An improvement of 1.7 for the stretching group is borderline clinically significant and gives pause to the question of how much improvement could have happened if subjects were told to stretch on their own at home. The confidence interval for stretching is tight and does not overlap pre VAS CI which increases the value of this finding. The sham taping group improved by less than 0.5 which is not clinically significant. No improvements were found in the Patient Self-Reported Functional Scale (PSFS). I agree with the authors' conclusion that the short duration of the study may have contributed to not finding a change on this scale. The authors believe this tool may not have been sensitive enough to pick up changes in the short term even if they had occurred. There were no events recorded and therefore number needed to treat and relative and absolute risk reduction could not be calculated. The authors could have used an increase in subjects' VAS as an event, but it would not have added much weight to the article.

The patient population used in the trial is very broad and therefore the results can be applied to a large amount of the patients with plantar heel pain that are seen in a normal outpatient PT clinic. Most patient's values and preferences would be satisfied if they were treated with taping, stretching or both. The results of one of those treatments alone, however, would not be sufficient to satisfy patients' desires to return to pre-injury function or resolve pain. Both techniques require patients to return to the clinic at least twice a week for treatment. Some patients may not be able to do this based on their schedule. Both the calcaneal taping technique and the plantar stretching are inexpensive easy techniques to learn and apply to your patients. A downfalls of this study is that they did not give patients in one of the groups both the calcaneal taping and the stretching. This might have shown a larger treatment effect. They also did not have the plantar stretching group do any stretching on their own, which does not resemble what a normal PT treatment regiment would include. This study demonstrated either a simple taping technique or passive stretches twice a week can significantly improve a patient's self reported pain level in the short-term. Although this study has limitations, which the authors acknowledged, the results are significant and the treatments used are simple enough to apply to my patients immediately. The taping technique is especially useful when treating patients in remote locations where orthotics would not be available. Future studies should compare low dye taping to calcaneal taping.

Appraised by: Kara Weigel; Saturday, April 07, 2007 Email: kara.weigel@us.army.mil

        Back to CAT Express Ortho         Back to Top

WOUND CARE:

• Can Removable Cast Walkers be Modified to Aid in Pressure Relief and Healing of Diabetic Foot Ulcers?
• Efficacy of MIST Ultrasound in the Treatment of Chronic Diabetic Foot Ulcers

Can Removable Cast Walkers be Modified to Aid in Pressure Relief and Healing of Diabetic Foot Ulcers?

Instant Total Contact Casting (iTCC) does improve neuropathic wound healing over Removable Cast Walkers (RCW). Patient compliance is crucial to epithelialization of the wound as demonstrated in this study. Instant total contact casting demonstrated a 57% reduction in non-healed wounds at 12 weeks. iTCC also demonstrated statistical significance in the percentage of wounds healed at 12 weeks (82.6% iTCC vs. 51.9% RCW ). However, future studies with larger sample sizes are needed to better determine treatment effect size. This study is 1b- level of evidence according to Sackett's hierarchy.

Citation/s:Armstrong DG, Lavery LA, Wu S, Boulton AJM: Evaluation of removable and irremovable cast walkers in the healing of diabetic foot wounds. Diabetes Care 28:551-554, 2005.

Lead author's name and fax: David G Armstrong

Three-part Clinical Question: What is the best treatment strategy to reduce pressure and promote complete wound healing in a 54-year old man with a chronic forefoot diabetic ulcer: a removable or irremovable cast?

Search Terms: Feet AND ulcer* AND cast* (Cochrane Database)

The Study:Non-blinded randomized controlled trial with intention-to-treat. The Study Patients: The patients were randomly divided into two groups. The Removable Cast Walker (RCW) group consisted of 27 subjects (24 men and 3 women) and the "instant" total contact cast group (iTCC) consisted of 23 subjects (20 men and 3 women). The study began with 50 participants, however four were lost to follow-up due to discomfort/weight of the device (2 subjects, one in each treatment group) or failure to return for follow-ups visits (2 subjects). The mean age in the RCW group (n=27) was 64.6 +/- 9.8 and 66.9 +/- 10.1 in the iTCC group (n=23). Body mass index was similar in both groups with 33.5 +/- 6.2 kg/m squared and 33.3 +/- 6.8 kg/m squared respectively. In the RCW group the baseline wound size was 2.0 +/- 1.1 cm squared and in the iTCC group it was 2.7 +/- 1.3 cm squared. Baseline vibration perception thresholds and HgbA1C results were also similar.

Inclusion criteria: Included those with a diagnosis of diabetes with a loss of protective sensation (>25V), at least one palpable foot pulse, and a neuropathic diabetic foot ulcer corresponding to Grade 1A (superficial, not extending to tendon, capsule, or bone). Exclusion criteria included active infection, inability to walk without a wheelchair, or wound locations on the heel, rearfoot, or a location other than the plantar foot, or with severe vascular disease.

Control group (N = 27; 27 analysed): All subjects received an evaluation consisting of wound measurements, assessing the location of the wound, and analysis of protective sensation and pulses. Subjects were instructed in use of a removable cast walker and were told to wear it at all times while ambulating. All subjects were followed on a weekly basis for device inspection, wound care, and debridement. All wounds were debrided as required on each visit. Subjects were followed on a weekly basis for 12 weeks or until complete wound healing (complete epithelialization) whichever came first.

Experimental group (N = 23; 23 analysed): All subjects received an evaluation consisting of wound measurements, assessing the location of the wound, and analysis of protective sensation and pulses. Testers applied a removable cast walker with Plaster of Paris to secure the walker and transition it to an instant total contact cast. All subjects were followed on a weekly basis for device inspection, wound care, and debridement. All wounds were debrided as required on each visit. Subjects were followed on a weekly basis for 12 weeks or until complete wound healing whichever came first.

The Evidence:

Comments: This study possesses strong methodology and a rigorous design thereby generating strong potential for a valid study. Prognostic factors between groups were well balanced to include age, gender, body mass index, vibration perception threshold, and Hemoglobin A1C values. The only factor that the iTCC and RCW group differed significantly on was the size of the wound (iTCC = 2.7 +/- 1.3 cm squared, RCW = 2.0 +/-1.1 cm squared). The experimental group began the study with a larger average wound size, thereby further strengthening the effect of treatment. Subjects were randomized into the two groups through a computerized randomization schedule and randomization was concealed and provided to the tester by telephone. Potential risks to randomization did exist as the patients and clinicians were aware of group allocation. However, group allocation awareness likely did not affect the study as the outcomes measured are objective measurements without subjective participant outcomes. Outcome assessors were also aware of group allocation and this may have led to surveillance bias. The study design reduced the risk of potential bias by focusing on a dichotomous outcome of complete wound healing rather than wound reduction. Complete epithelialization was necessary to deem the wound healed, thereby minimizing potential surveillance bias and subjectivity. The study design ensured complete follow-up of all subjects. Although, four subjects were lost to follow-up, all were included in an intention-to-treat analysis. Therefore, all subjects who did not complete the study were deemed treatment failures for result analysis.

The primary outcome of interest was the number of patients who achieved complete wound healing within twelve weeks. The intent-to-treat analysis showed that a higher proportion of patients were healed at 12 weeks in the iTCC group than in the RCW group ( 82.6% vs. 51.9 %) and that it was statistically significant at P=0.02. The results are also clinically meaningful and demonstrate that this study was indeed a positive study. Diabetic wounds are time-consuming and costly to treat. Therefore, a treatment that provides a 10% relative risk reduction in delayed wound healing is substantial (minimal clinically important difference). The treatment consisting of a removable cast walker made irremovable by applying Plaster of Paris reduced the risk of a non-healed wound (12 weeks) 57% compared to the RCW group. The treatment demonstrated an absolute risk reduction of .273, meaning the treatment would spare 27% of subjects with a neuropathic diabetic wound from a non-healing ulceration. In addition, four patients would need to be treated in order to prevent one wound from delayed healing. Unfortunately, the precision of the estimate is not narrow as demonstrated by the confidence intervals. The true value of the relative risk reduction has a 95% chance of lying within 5% to 100%. While this does enhance confidence that the treatment does indeed have a positive effect, it does little to strengthen our trust in the precision of the estimate. The same rings true for the number needed to treat and absolute risk reduction as the confidence intervals are wide. However the analysis and confidence intervals demonstrate a definite positive effect of the treatment over a removable cast walker. Not only did a greater proportion of the iTCC subjects heal, but they healed faster. The iTCC group's mean days to complete wound healing was 41.6 versus 58.0 days in the RCW group. In addition, the 95% confidence interval demonstrated a positive effect indicating that the true healing time is 7 to 26 days faster in the iTCC group. In wound care, each day added requires additional investment of time and money to properly care for and dress the wound. Therefore, healing five days faster is significant clinically and the treatment confidently outperforms a removable cast walker.

Soft tissue infection rates were similar between the iTCC and RCW groups. The results demonstrated a 29% reduction in soft tissue infection rates in those with the instant total contact cast and that 10% were spared from an infection. The number needed to treat statistically demonstrated that nine people would need to be treated with the instant total contact cast to prevent one soft tissue infection. However, the confidence intervals were wide for all results, thereby significantly reducing the precision of the estimate and making it nearly impossible to determine if the instant total contact cast had any effect on soft tissue infections. In fact, due to a negative confidence interval lower boundary, it is possible that the removable cast walker served better in reducing infection rates. These findings do not detract from the outcome of the study however, because the treatment was not designed nor ever intended to reduce the rate of soft tissue infection. The results of this study can be applied to patient care and in particular to the patient described within the clinical question. Instant total contact casting demonstrated a positive effect on wound healing over removable cast walkers and at a rate similar to total contact casting. Future research should include larger sample sizes to better determine the precision of estimates affecting clinical significance. In addition, future studies with further inquiry into the location of the plantar ulcer and effects on those patients with arterial disease would aid clinicians in further determining if their patient would benefit from instant total contact casting.

Appraised by: Jessica Feda, PT, OCS; Saturday, October 21, 2006 Email: jfeda@bop.gov

In a 54 year old patient with a chronic diabetic foot ulcer, the use of MIST ultrasound does not appear to provide significant advantages in closing the wound and conversely may actually impair the healing of the wound by some 11% as compared to the control group. Considering the astronomical expense of this treatment ranging from $55 to $75 per session, the clinician would be better served to use electrical stimulation which has been proven to be a safe cost-effective tool in treating chronic leg ulcers. Further research is needed to determine the role of this new technology in the treatment of chronic diabetic foot ulcers. This is Level 1B- evidence.

Citation/s:1. Ennis WJ, Formann P, Mozen N, Massey J, Conner-Kerr T, Meneses P. Ultrasound Therapy for Recalcitrant Diabetic Foot Ulcers: Results of a Randomized, Double-Blind, Controlled, Multicenter Study. Ostomy/Wound Management 2005; 51(8):24-39. 2. Houghton PE, Kincaid CB, Lovell M, Campbell KE, Keast DH, Woodbury MG, Harris KA. Effect of electrical stimulation on chronic leg ulcer size and appearance. Physical Therapy 2003; Jan. 83(1):17-28.

Lead author's name and fax: William J. Ennis, DO. Email: wct211@pol.net. No Fax number available.

Three-part Clinical Question: In a 54 year old patient with diabetes mellitus and a recalcitrant neuropathic ulcer on the dorsal aspect of his 5th toe, would the use of MIST ultrasound improve the closure rate of the wound as compared to the standard of care?

Search Terms: Pubmed Clinical Queries: narrow, specific search for Therapy with the following search string: (ultrasound AND ulcer* AND diabet*) AND (randomized controlled trial[Publication Type] OR (randomized[Title/Abstract] AND controlled[Title/Abstract] AND trial[Title/Abstract]))

The Study:Single-blinded concealed randomized controlled trial with intention-to-treat. In this study, the subjects and investigators were blind to the group assignment but the clinicians performing the ultrasound or sham treatments were aware of group assignment.

The Study Patients: Ultrasound Treatment Group (n=27): 13 male and 14 female; mean age 56 +/- 11 years; mean body mass index 34.57 +/- 1; mean baseline HbA1c 9.4%; race: 17 white, 8 black, and 2 Hispanic; smoking status: 18 never, 4 current, 5 prior; mean ulcer duration of 35 +/- 32 weeks; and mean ulcer area of 1.7 +/- 0.8 cm2. Control Group (n=28): 19 male and 9 female; mean age 54 +/- 12 years; mean body mass index 35.30 +/- 1; mean baseline HbA1c of 8.4%; race: 20 white, 6 black, and 2 Hispanic; smoking status: 19 never, 5 current, 4 prior; mean ulcer duration of 67 +/- 108 weeks; and mean ulcer area of 4.4 +/- 4.0 cm2. Control group (N = 63; 63 analyzed): All potential subjects were screened to ensure they met the study inclusion criteria of being a Type 1 or Type 2 diabetic, have a Wagner Grade 1 or 2 neuropathic wound for greater than 30 days, be at least 18 years old; have a glycosylated hemoglobin value of less than or equal to 12% within 30 days of the study onset; have an ankle-brachial index of between 0.65 and 1.2, and be ambulatory with weight bearing on the index foot.

Those meeting the criteria underwent a baseline debridement after which a quantitative culture biopsy was taken of the wound. Patients were fit with a standard fixed ankle walking boot and received dressing supplies and instructions on performing twice daily saline gauze dressing changes for seven days. Those subjects whose wounds had decreased in size by more than 30 % over the course of the seven day baseline wound care were excluded from the study. The Control Group received sham MIST ultrasound treatments three times a week until the wound closed or up to 12 weeks. The sham device looked exactly like the real unit and delivered the same pressure and volume of saline mist as the true MIST ultrasound unit. The investigator performed wound debridement as necessary prior to the sham ultrasound treatment. The patients underwent primary dressing changes during their sham treatment sessions. The patient or a family member performed secondary dressing changes twice daily on non-treatment days. Experimental group (N = 70; 70 analyzed): The treatment group received the exact care as described for the control group except that they received the actual MIST ultrasound treatment 3 times a week.

The Evidence:

Outcome	Time to Outcome	CER	EER	RRR	ARR	NNT
Failure of wound to close	Up to 12 weeks	0.778	0.743	4%	0.035	29
95% Confidence Intervals:	-14% to 23%			-0.110 to 0.180	NNT = 6 to INF; NNH = 9 to INF
Clinically significant adverse reactions such as wound enlargement, wound infection, blister formation or additional wound formation	Up to 12 weeks	0.016	0.100	-525%	-0.084	-12
95% Confidence Intervals:	-1005% to -45%			-0.161 to -0.007	-139 to -6

Comments: Are the results valid? On the positive side, the investigators really attempted to develop a highly clinically relevant study. Care was taken to select patients that met criteria that matched the patient population that a typical wound care specialist would see in a clinic. Once the pool of subjects had been determined, the investigators used a computer generated randomization table to determine which patients were placed in the treatment group (MIST ultrasound) and which were placed in the control group (sham ultrasound). In order to further control for the treatment effect a sham ultrasound machine was developed by the manufacturer that produced the same saline mist in terms of volume and pressure. Because the sham unit was louder when operating, a curtain was placed between the patient and the ultrasound units and both units were turned on so as to not give any clues to the patient as to which treatment they were receiving. To strengthen the study further, the investigators who assessed the wound and performed the wound debridements were blinded to the group randomization. Additionally, every patient received the same wound dressings, protective walking boot, and wound care instructions.

With all of these variables controlled, the investigators had the potential for conducting one of the strongest wound care studies that I have come across .Unfortunately, the actual conduction of the investigation and the reporting of the findings created some major threats to the validity of the study. The clinician who conducted the MIST or sham ultrasound was aware of the treatment groups and in fact had to use a slightly different technique for the sham instrument by holding the device 100 to 150 mm away from the wound as compared to 5 to 15 mm with the actual ultrasound device. This difference in operation became problematic when a mid-study investigation determined that clinicians in two institutions had reversed the distances for the two units. This investigation was initiated because the initial study results were not as expected. To the reader the mid-study determination seemed to smack of result hunting and indeed the study was sponsored by the ultrasound manufacturer. To correct for this improper application of both ultrasound and sham devices, the investigators chose to focus their findings on an "efficacy group" in which the treatments were done correctly. In fact, the majority of the data makes reference to the remaining 55 subjects, 27 in the treatment group and 28 in the control group. Baseline demographics were only reported on this "efficacy" population and the groups appeared homogenous. However, within this population, the control group had statistically significant larger wounds. To determine the effect of wound size, the investigators did perform a Cox proportional hazards regression and found that the size difference in wounds did not significantly impact their reported findings on the effect of the ultrasound. While the authors chose to focus on the "efficacy" population, they did include intent-to-treat data for wound closure and adverse events for the entire subject population of 133, all of whom received at least one ultrasound or sham treatment. Had the authors not chosen to include the data for the intent-to-treat analysis, the study would have had serious enough bias to create a major threat to the validity of the study. However, if the reader ignores the overly optimistic reports on the "efficacy" group that the authors focus on and analyzes the intent-to-treat data, the study seems to much more acceptable.

What are the results? If one were to focus on the "efficacy" population as the investigators did, the results for the use of MIST ultrasound are quite impressive with 40.7% of the treated wounds healed as compared to 14.3% in the control group. However, to fully assess the data, an intention-to-treat analysis is truly needed and when this is done there is no statistically significant difference in treatments .With the data from the original 133 subjects, the absolute risk reduction for wound non-closure was 3.5% with 95% CI of -11.0 to 18.0 which means that there is just about equal chance that the ultrasound treatment impaired the closing of the wound as it did to help it. The clinician is left with a quandary as to the true effect of the treatment and since the absolute risk reduction crosses the 0 point, care must be taken to understand that there is a good possibility that this treatment could actually slow healing. In terms of the Number Needed to Treat statistic, an NNT of 29 is not impressive but when coupled with the 95% CI of NNT= 6 to infinity and the Number Needed to Harm (NNH) = 9 to infinity, the clinician really has to consider that using this new treatment could cause harm in 1 out of 9 patients treated as well as aid in wound closure in 1 out of 6. These statistics truly give insight as to the possible value of this treatment which is nothing close to that reported by the manufacturer .. Additionally the wide range of the 95% CI for the ARR and NNT does show that the study truly lacked sufficient power to truly give the correct clinical picture. With a larger patient population, the effect of the treatment may truly be positive and serve as a new treatment tool for the wound care expert, but then again there is a good chance that the opposite could true as well. For the safety of the device, the investigators did provide data for the entire original subject population. The adverse events which were possibly, probably, or definitely caused by the ultrasound or sham device listed in the study were pain, erythema, ulcer enlargement, ulcer infection, additional ulcer development, blister formation, edema, and other. As a wound care provider, the most important of these were ulcer enlargement, ulcer infection, additional ulcer development and blister formation. Analyzing the data from these speci fic adverse events provides the clinician a different picture than the reported proven "safety" of the product as reported by the investigators. The ARR of -8.4% with 95% CI of -16.1% to -7.0% tells clinicians that the treatment ultrasound definitely caused between 0.7-16.1% more of these adverse events as compared to the control. Since the ARR CI does not cross 0, the clinician can be 95% positive in this possibility of potential harm. What makes this even more disturbing is that the intent-to-treat statistics include those ultrasound treatments that were done at a distance that was judged to be ineffective and consequently thrown out of the "efficacy" population. Would the rate of adverse events been even higher had the treatments been carried out as specified by the manufacturer? Perhaps in a wound which carries a higher bioburden of greater than 1x105 bacterial colonies/ gram of tissue, the use of the ultrasound may drive the bacteria deeper into the tissue in a type of phonophoretic effect resulting in higher rates of infection and additional ulcer development.

How can I Apply the Results to Patient Care? When faced with a new technology, the clinician has to ask the question: Is this treatment safe for my patients and more cost-effective than what is being offered currently? From the data provided, the clinician must first be very leery of this new form of ultrasound as it appears to increase the risk of ulcer enlargement, ulcer infection, ulcer development and blister formation. The overall impact on wound closure is equally unimpressive as there appears to be almost as likely that the treatment slowed wound closure as it did to aid closure. Couple this with the fact that this new piece of equipment costs some $25,000 to purchase or if leased costs range from $55 to $75 per treatment, the prohibitory costs and potential harm seem to make the decision to use this new device rather easy. While I was contemplating attempting a trial of this technology in the clinic, this study gives me more than enough cause for concern that I will wait for future studies with larger subject populations to provide a better picture of the effects of this new technology.

Appraised by:Eric Payne; Saturday, January 20, 2007 Email: epayne@bop.gov

        

        Back to CAT Express Wound Care         Back to Top

        

NEUROLOGY:

Prognosis for Recovery of Good Upper Extremity Use after Stroke in Adults

Citation/s:Katrak P, Bowring G, Conroy P, Chilvers M, Poulos R, McNeil D. Predicting upper limb recovery after stroke: the place of early shoulder and hand movement. Arch Phys Med Rehabil. 1998 Jul;79(7):758-761.

Lead author's name and fax: Pesi Katrak, MD, Department of Rehabilitation Medicine, Prince Henry Hospital, Anzac Parade, Little Bay, NSW 2036, Australia. (No fax or email was given)

Three-part Clinical Question: For a 70 year old male with right arm/hand hemiparesis after a stroke 6 days ago, what is the probability of good right hand/arm recovery at three months?

Search Terms: PubMed Clinical Queries: narrow, specific search for Prognosis with the following search string: (stroke AND shoulder function AND recovery) AND (prognos*[Title/Abstract] OR (first[Title/Abstract] AND episode [Title/Abstract]) OR cohort [Title/Abstract]). The search yielded 8 hits, one of which was the Katrak et al. article.

The Study: The Study Patients: 71 consecutive first-stroke patients (validated by CT) admitted for inpatient rehabilitation medicine service at three teaching hospitals. The mean time from stroke to initial assessment was 11days after stroke (range, 0 to 23 days). There were 43 men and 28 women with a mean age of 67, range of 22 to 91 years; 28 had a left hemiparesis and 43 had right hemiparesis. After initial assessment patients were reassessed at 1 month (28 - 45 days), 2 months (between 45 and 75 days), and 3 months (between 76 and 105 daystroke. Prognosticrognostic Factor: Shoulder shrug, shoulder abduction, and initial presence of synergistic hand movement.

The Outcome: Good hand movements, i.e., independent index finger extension or opposition of finger/s to thumb and ability to perform one of the specified hand function tasks, such as rotate a matchbox in the hand .Thereand.There was a well-defined sample at a uniform (early) stage of illness. Follow-up was long enough; follow-up was not complete. There were not blind, objective outcome criteria. Adjustment was not made for other prognostic factors. There was no validation in an independent test-set of patients.

The Evidence:

Prognostic Factor	Outcome	Result	Measure	Confidence Interval	Independent?
Ability to shrug affected shoulder	Good hand mvmt (HMS score of > than 3) @ 3 mos	6.0	Odds Ratio	1.4 to 25.6	yes
Initial shoulder abduction greater than 30 degrees	Good hand mvmt (HMS score of > than 3) @ 3 mos	1.4	Odds Ratio	0.2 to 10.8	no
Initial synergistic hand movement scale score of 2 or 3	Good hand mvmt (HMS score of > than 3) @ 3 mos	12.2	Odds Ratio	2.1 to 71.0	yes
Ability to shrug affected shoulder	Good hand function at 3 mos (1 or more hand fct tasks)	11.3	Odds Ratio	2.0 to 45.3	yes
Initial shoulder abduction greater than 30 degrees	Good hand function at 3 mos (1 or more hand fct tasks)	3.6	Odds Ratio	0.3 to 44.8	no
Initial synergistic hand movement scale score of 2 or 3	Good hand function at 3 mos (1 or more hand fct tasks)	9.0	Odds Ratio	0.9 to 93.8	no
UE hemiparesis following stroke and an HMS score of 3 or less	HMS score of 4 or more Able to do 1 or more hand fct tasks at 1 month	41.5%	Percent	30% to 53%	yes
UE hemiparesis following stroke and an HMS score of 3 or less	HMS score of 4 or more Able to do 1 or more hand fct tasks at 2 months	46%	Percent	33% to 60%	yes
UE hemiparesis following stroke and an HMS score of 3 or less	HMS score of 4 or more Able to do 1 or more hand fct tasks at 3 months	52%	Percent	38% to 66%	yes

Comments: Are the results valid? This study was a prospective cohort study that followed 71 consecutive first-time stroke patients. This study was able to assemble a group of patients at a fairly similar set point of time following their onset of stroke. The average time was 11 days post stroke with a range of 0 to 23 days. This should reflect that the group of patients was in a similar point of time for their disease process. This homogeneity of the time point lends strength to the study but the age range threatens it. The mean age of the subjects was 67 with a range of 22 to 91. One would not expect the outcome of the 22 year old or the 91 year old to be similar to each other. There may have been other prognostic factors that should have been considered as part of the inclusion or exclusion criteria. Follow up was not complete with a large loss to follow-up. The initial study started with 71 and at 3 months there was 46, representing a 35% loss to follow-up. In addition, there is no explanation for the loss and this further threatens this study's validity because often a different outcome of interest is possible for the group that is followed and the group that is lost. The three authors of this study and four advanced trainees gathered all data. This introduces bias into the study because all individuals were aware of the prognostic factors. The frequency and target outcomes were clearly defined and based on objective measures but this still allowed subjectivity to enter the study because the assessors were not blind to the prognostic factors. Finally, there has been no true validation of an independent study to see if these prognostic factors coincide with the results of another study. There appears to be only a hint or perhaps a hunch that these factors are indeed prognostic for UE recovery based on clinical observation of a small sample of patients.

What are the results? Of the 65 (though there is a discrepancy between this number from table 3 and the written text) patients assessed on the Hand Movement Scale (HMS) at 1 month, 27 (41.5%) met the author's criteria of an HMS score of 4 or greater. They characterize this as "good" hand movement and it is unclear if the specified hand function task was met or not. At 2 months 50 patients were reassessed on the HMS and 23 (46%) met the criteria and at 3 months 46 patients were reassessed on the HMS and 24 (52%) met the author's criteria. In reviewing the definitions of the scores on the HMS, a score of 4 indicates that the subject can extend his/her index finger while maintaining the other fingers in flexion. A score of 5 indicates that a person can oppose his/her thumb and index finger only and a score of 6 indicates that all fingers can oppose the thumb. I would not consider the score of 4 a reasonable "good" hand movement because it would not assist in eating, writing, or any kind of pinching movement associated with independent living skills. It would allow additional communication for the patient who has impaired verbal ability by fostering a pointing action but would not be the most helpful in activities of daily living. The ability to oppose allows for independence in most self care activities. In addition, the precision of these point estimates at the 1, 2, and 3-month time frame are moderate to somewhat wide. This means that at the one month assessment the true prognosis for having a score of 4 or greater on the HMS could be as low as 30% or as high as 53% (with a 95% level of confidence). These widths get even larger at the second and third month. The results from the logistic regression analysis further decrease one's confidence secondary to the large bordering on huge confidence intervals. A case in point of this is when you look at patients who were able to shrug their affected shoulder at initial assessment were 6 times as likely to have good hand movement at 3 months. However, the true population value might be as low as 1.4 times as likely, which is a very different clinical scenario to present to a patient or their family. The confidence intervals of 3 independent variables (initial shoulder abduction > than 30 degrees for hand movement and hand function and synergistic hand movement as a predictor of hand function) at 3 months all crossed the null value (1.0 for odds ratios) indicating that these factors are not significant predictors of good hand movement and/or function.

How can I apply the results to patient care? It is clear that the patient from the 3-part clinical question is similar to this study's patients. It would be very easy to administer the first two prognostic factors (shoulder shrug and shoulder abduction) by a variety of health care personnel. The third prognostic factor (Initial synergistic hand movement score of 2 or 3) is a little more difficult because the health care personnel would have to be familiar with the test and know its scoring system. Overall the tests are very quick, inexpensive, and not difficult to interpret. This makes these prognostic factors very attractive as possible predictors of arm/hand recovery. However, given the wide confidence intervals, the unprotected threats to validity (such as not blinded and no independent validation study), and the large, unaccounted for loss to follow-up we can not feel very confident that these prognostic factors actually predict the likelihood of a patient recovering his UE movement. Additional, more powerful studies that support these prognostic factors need to be done before a clinician should use these as predictors of hand/arm recovery. I also think that any clinician who has experience in stroke rehabilitation may already use these tests as part of their work-up and that early self-initiated movement would usually be a good sign. However, it would be nice if a clinician could give the patient or family members a true likelihood of recovering good upper extremity use based on valid evidence.

Overall, this study would not make a tremendous impact on my conclusions of what to tell my patient or family member. It does allow me to feel a bit more confident that early-identified hand and shoulder movement is more likely to mean a greater recovery of upper limb function but given the confidence intervals around the point estimates at 1, 2, and 3 months I would be reluctant to give a percentage to the patient. The large range may be discouraging for some and encouraging for others. Since all the patients would be undergoing rehabilitation, I would rather instill hopefulness in the rehab process than on tenuous numbers.

Appraised by: Heather E. Khan; Wednesday, March 21, 2007 Email: Ehenfeld@comcast.net

        

Determination of the clinical usefulness of Tinnel's test, Phalen's test and sonographic measurement in the diagnosis of Carpal Tunnel Syndrome

In adults in whom carpal tunnel syndrome is suspected, a positive Tinnel's test is highly predictive of neuropathy, as is a postive Phalen's test and a positve ultrasound test (resulting in at least one pathological measurement) of the median nerve. Level of evidence 1b.

Citation/s: Kotevoglu N, and Gulbahce-Saglam S. Ultrasound imaging in the diagnosis of carpal tunnel syndrome and its relevance to clinical evaluation. Joint Bone Spine. 2005: 72(2):142-145.

Lead Author's name and fax: Nurdan Kotevoglu, Dept of Physical Therapy and Rehabilitation, Sisli Etfal Teaching Hospital, Istanbul, Turkey

Three-part Clinical Question: In a 35 year old woman with exercise-induced nerve pain of the median nerve, can Tinnel's test diagnose neuropathy?

Search Terms: I used PubMed Clinical Queries to conduct a narrow, specific search for Diagnosis with the following search string: (Tinnel) AND (specificity[Title/Abstract]). The search yielded 1 hit, which was the Kotevoglu and Gubahce-Saglam article.

The Study Patients: In the Carpal Tunnel Syndrome (CTS) group, 23 patients were female and one was male with a mean age of 46.9 years and a range of 35-54 years. In the Control group, 14 "age and sex matched" healthy subjects were used that had a mean age of 38.8 years and a range of 30-48 years. Independent, blind comparison with a reference (gold) standard. Couldn't tell if there was appropriate spectrum of patients. The gold standard was not applied regardless of the test result.

Target disorder and Gold Standard: Carpal Tunnel Syndrome; EMG

Diagnostic test: Tinnel's test is performed by tapping over the median nerve with a broad head hammer while the wrist is resting in extension and is considered positive if the patient describes paresthesia radiating distally from the wrist. Phalen's test is performed by the patient holding the wrists in extreme but not forced flexion for 1 minute and is considered positive if paresthesia is reproduced during that time. The ultrasound testing performed for this study was done with real-time equipment with a high resolution 10 to 5MHz multi-D linear array transducer, with the wrist in a prone neutral position resting on a hard surface.

The Evidence:

	Target Disorder: CTS
Test: Tinnel's	Present		Absent
Test Result	Num	Prop	Num	Prop	Likelihood Ratios
Positive	32	a	4	B	4.91	1.95 to 12.35
Negative	12	c	23	D	0.32	0.19 to 0.53
Sensitivity:					73%	60 to 86
Specificity:					85%	72 to 99
Prevalence:					62%	51 to 73
Positive Predictive Value:					89%	79 to 99
Negative Predictive Value:					66%	50 to 81

	Target Disorder: CTS
Test: Phalen's	Present		Absent
Test Result	Num	Prop	Num	Prop	Likelihood Ratios
Positive	40	a	4	B	6.14	2.47 to 15.23
Negative	4	c	23	D	0.11	0.04 to 0.28
Sensitivity:					91%	82 to 99
Specificity:					85%	72 to 99
Prevalence:					62%	51 to 73
Positive Predictive Value:					91%	82 to 99
Negative Predictive Value:					85%	72 to 99

Comments: Are the results valid? The study included a series of patients who presented with typical signs and symptoms of CTS and whose diagnosis was confirmed by EMG, which is the typical gold standard measure of CTS. The authors did not indicate whether these patients were all collected from a certain geographical area at a certain time or if they were all consecutively seen at the author's institution. If patients for the study were picked and chosen according to severity of symptoms or occupation (for example), bias might have been introduced which inflated the results and the population might not have been representative of every patient who presents with CTS. The authors did exclude patients who presented with median neuropathy due to other (mostly systemic) causes, so they did rule out confounding factors, which helps with the validity of the study.. The definitive diagnostic standard was appropriate in that a well established and valid diagnostic test was used to compare results with. In addition, the authors compared the results of two clinical tests routinely administered by physical therapists (Tinnel's and Phalens test) as well as sonographic measures obtained for the median nerve and were able to find a positive correlation between these three tests as well as with EMG. The study's resulting finding was that the combination of positive clinical tests combined with at least one pathological measurement of the median nerve (via ultrasound) was highly predictive of CTS. The only concern with the findings of this diagnostic approach is that intertester reliability was not studied. From current searches in the literature, no other studies have been conducted on the Tinnel's test. As a result, such a study would increase the validity of these findings. In addition, the control subjects deemed as having normal median nerve function did not have an EMG performed. Of these control subjects, (involving 27 wrists) 4 wrists were positive for Phalen's test and 4 wrists were positive for Tinnel's test. The authors do not state if the 4 wrists were the same for both tests, but by performing an EMG on the control group the study could have provided further correlation between the tests, thus increasing the predictive value of Tinnel's or Phalen's test.

What are the results? The Tinnel's test had a positive likelihood ratio of 4.91, which would result in a significant change (about 30%) from pre- to posttest probability. The negative likelihood ratio was 0.32, which again would result in a signficant change (about 30%). If a patient presented to the clinic with a low pretest probability, a positive test would merit more tests while a negative test would most likely rule out CTS. If a patient presented to the clinic with a moderate to high pretest probability, a positive Tinnel's test would most likely result in a diagnosis of CTS while a negative test would merit more tests. The confidence interval (CI) for the positive likelihood ratio is wide (1.95 - 12.35) which decreases the strength of the evidence, however the lower value does not go below 1, therefore the evidence is still significant. The CI for the negative likelihood ratio is moderate (0.19 - 0.53) and again is significant as it does not cross 1. Phalen's test had a positive likelihood ratio of 6.14 resulting in a larger posttest probability than Tinnel's test. The CI interval was wide (2.47 - 15.23) but significant. The negative likelihood ratio was 0.11 resulting in a lower posttest probability than Tinnel's test. The CI interval was not large (0.04 - 0.28), increasing the predictive power of a negative test. The ultrasound test had 39 positive tests and 5 negative tests with the EMG-confirmed CTS group. The authors did not report how many positive and negative tests were present in the control group, thus likelihood ratios were not calculated. The authors reported the sensitivity and specificity for the ultrasound as 89% and 100%, respectively.

How can I apply the results to patient care? The occupations and recruitment methods of the study participants were not given, so I cannot determine if they would be similar to patients that I may treat. Thus the pretest probability of my patients may vary from the 62% given in the study. However, I would consider the test threshold for Tinnel's and Phalen's test as pretty low in that they are both quick and easy-to-administer tests with significant findings (the worst case scenario for predictive value according to the CIs for each test would still be useful). In addition, I would be inclined to use both tests together versus just the Tinnel's test due to the stronger findings for the Phalen's test. I would add the ultrasound test to these other two tests only to clarify inconclusive findings or to determine severity of the median nerve involvement (due to the increased time and effort in performing the test).

Appraised by: Amy Eschenberg, MPT. October 6, 2007, Email: Amy_Eschenberg@baylor.edu

        

        Back to CAT Express Neurology         Back to Top

        

PEDIATRICS:

Use of the Alberta Infant Motor Scales in Screening for Motor Abnormality

The Alberta Infant Motor Scales (AIMS) is a tool for assessing motor development. The AIMS has high sensitivity for ruling out motor abnormality and this study recommends the use of the 5th percentile as a cutoff score for ruling in motor abnormality. This study is level 4 evidence due to its non-independent reference standard and non-blinded design.

Citation/s: Campos D, Santos DC, Goncalves VM, Goto MM, Arias AV, Brianeze AC, Campos TM, Mello BB. Agreement between scales for screening and diagnosis of motor development at 6 months. J Pediatr (Rio J). 2006; 82(6):470-4.

Lead Author's name and fax: Vanda Maria Gimenes Goncalves. vandagg@uol.com.br

Three-part Clinical Question: What is the most appropriate cutoff score on the Alberta Infant Motor Scales (AIMS) to identify motor development abnormalities in 6 month old infants?

Search Terms: A broad, sensitive defined search on PubMed Clinical Queries in the category Diagnosis using the terms "Alberta Infant Motor Scales" brought up 5 citations. The title and a quick peek at the abstract on the cited article seemed right on target for my question.

The Study: The Study Patients: The sample population in this study were 43 infants, age 6 months (+/- 7 days) who were part of a cohort that had been enrolled as neonates all being born single-fetus pregnancies, at gestational ages of 37 to 41 week, in the standard maternity ward, requiring no special care, were discharged from the maternity unit 2 days after birth, residents of the metropolitan area of Campinas, Brazil. Infants were excluded if they had low birth weight, exhibited genetic syndromes, malformations, or congenital infections. Demographic data on this sample showed a healthy population with: birth weight (grams) mean 3065.93, SD 370.26, gestational age (weeks) 39.9, SD 1.03, 1-minute Apgar (1-10 scale) mean 8.18, SD 1.72, 5-minute Apgar mean 9.45, SD 0.55. The spectrum of healthy infants aged 6 months seemed appropriate. The study made comparison with a reference (gold) standard for each subject, but non-independently and non-blinded. Target disorder and Gold Standard: Target disorder was "motor abnormality". The "Gold Standard" was an index score less than 85 (or greater than one standard deviation below the mean) on the Bayley Scales of Infant Development II (BSID-II). The BSID-II is a developmental diagnostic tool that assesses children aged 1 to 42 months and requires about 40 minutes to administer. It is considered among the best scales available in the area of child development assessment, providing reliable, valid and precise results. Diagnostic test: The Alberta Infant Motor Scale (AIMS) is an assessment instrument that evaluates gross motor function from birth to 18 months. This instrument classifies infants against a development curve between the 5th and 90th percentiles, but does not establish a range of percentiles as cut-off points for further diagnostic assessment. Assessment with the AIMS is a process of scoring observed behaviors of an illustrated sequence of postural control development in the positions of prone, supine, sitting, and standing. Administration of the AIMS takes approximately 15 minutes and has excellent psychometric properties of test-retest and interobserver reliability and validity.

The Evidence: Note: Added 0.5 to the counts in all four cells of the observed tables below.

	Target Disorder: Motor Abnormality
Test: AIMS <=5%	Present		Absent
Test Result	Num	Prop	Num	Prop	Likelihood Ratios
Positive	6.5	a	8.5	b	4.15	2.22 to7.77
Negative	0.5	c	29.5	d	0.09	0.01 to 1.34
Sensitivity:					93%	74 to 100
Specificity:					78%	64 to 91
Prevalence:					16%	5 to 26
Positive Predictive Value:					43%	18 to 68
Negative Predictive Value:					98%	94 to 100

	Target Disorder: Motor Abnormality
Test: AIMS <=10%	Present		Absent
Test Result	Num	Prop	Num	Prop	Likelihood Ratios
Positive	6.5	a	19.5	b	1.81	1.25 to 2.62
Negative	0.5	c	18.5	d	0.15	0.01 to 2.16
Sensitivity:					93%	74 to 100
Specificity:					49%	33 to 65
Prevalence:					16%	5 to 26
Positive Predictive Value:					25%	5 to 42
Negative Predictive Value:					97%	90 to 100

Comments: Are the results valid? A significant threat to the validity of this study is the lack of blinding. Assessments of the infants in this sample were made simultaneously by an examiner and two observers with scores recorded by consensus among these three team members. Although the diagnostic test and the gold standard test were both administered for each subject, they were administered by the same team, and at the same time. The examiners knew the score on both the diagnostic test and the gold standard, and although there is no description of which assessment was administered first, the absence of blinding allowed for bias. In regard to the sample spectrum, I am not too concerned about the omission of the gender characteristics. Neither am I bothered about the possible cultural differences between the Brazilian infants and my American infants, after all, the AIMS normative values are based on Canadian infants. But my concern is that the study's healthy-at-birth group of infants had no description of their characteristics at 6 months, which was at the time of the testing. The study provides no information about the health status of the subjects between birth and age 6 months. Since the intent of the study was to ascertain agreement between the screening and diagnostic tool and define the best cutoff point for the screening, the lack of description of the current medical status of the population does not further question the validity of the study.

What are the results? The study reports by two cutoff point (5th and 10th percentile scores on the AIMS) the frequency of infants identified as true positive and negatives and as false positives and negatives, and the degree of agreement between the AIMS and BSID-II. The sensitivity of the AIMS at both cutoff points is high 93% (range 74 to100). Applying the SnNout concept, this indicates that the AIMS is very good at ruling out motor abnormality when the test is negative (or the score is at or above the 10th percentile score). The specificity of the AIMS at the 10th percentile cutoff is 49%, with a broad confidence interval 33-65%, which indicates that at this cut point that the AIMS is not very good at ruling in a motor abnormality. In addition, given the pre-test likelihood of 16% and the small positive likelihood ratios at 1.81 (range 1.25 to 2.62) a positive test does not provide a large change or convincing post-test likelihood (even though the PLR is significant as the range does not include the 1.0 null point). But the specificity of the AIMS at the 5th percentile cutoff is 78%, still with a broad confidence interval 64-91%, indicating that the 5% cutoff point is better than the 10% cut-point at ruling in a motor abnormality given a low (positive) test score. Looking at it another way, the lower cutoff score reduces the number of false positives, thus reducing the likelihood of inappropriately identifying an infant which may lead to unnecessary evaluation or bring unnecessary concern. The PLR at this lower cutoff point is 4.15 (range 2.22 to 7.77), and though still small is up one level per Jaeschke's taxonomy, as compared to the 10% cut-point PLR, and may generate small yet important changes in probability. Additionally the NLRs for both cut-points are not significant since the sample results include the null value, 1.0. In regard to the degree of agreement between the AIMS and the BSID-II, the authors report accuracy and kappa indices for the two cut points. But since they were all together, observing and scoring each infant after agreement had been reached, the biases introduced would influence these tests.

How can I apply the results to patient care? Since the purpose of my health promotion program is to promote parent-infant interaction to facilitate normal sensorimotor development, screening to identify infants as typically developing is important. The AIMS is an appropriate tool because of its high sensitivity making it appropriate for use in screening. The AIMS is already available in my practice, without additional cost, which is important in my small private business. After reviewing the sections on psychometric properties and norm-referencing of the AIMS in Motor Assessment of the Developing Infant by Piper MC and Darrah J. WB Saunders (1994), I have enough confidence to over-ride my concerns of the validity of the study, and I will plan to use the AIMS, with the 5% cutoff score for motor screening of the babies enrolling in Massage and Movement class. This study confirmed the sensitivity of the AIMS as a screening tool and statistically defined the specificity value of using the 5% cutoff score to ascertain ruling in motor abnormality.

Appraised by: Dawn Standley, November 17, 2007 Email: dawnstandley@msn.com

        

Positive effects from individualized developmental care on healthy, pre-term infants in NICU

Treatment using the Newborn Individual Developmental Care and Assessment Program (NIDCAP) as compared to standard care for low-risk pre-term infants during time in the neonatal intensive care nursery (NICU) makes a clinically important improvement in their neurobehavioral functional level of development at corrected ages of 2 weeks and 9 months. Level of Evidence 1b.

Citation/s: Als H, Duffy FH, McAnulty GB, Rivkin MJ, Vajapenyam S, Mulkern RV, Warfield SK, Huppi PS, Butler SC, Conneman N, Fischer C, Eichenwald EC. Early Experience Alters Brain Function and Structure. Pediatrics. 2004. 113(4):846-57.

Lead author's name and e-mail: Heidelise Als, PhD. heidelise.als@chilrens.harvard.edu

Three-part Clinical Question: A review of the cited article was completed with the focus on the techniques of neuroimaging used to study brain structural changes as related to very early experience. The article is now being critically analyzed from the standpoint of effectiveness of the therapeutic early experience on functional behavioral outcomes.

Is developmentally supportive care (NIDCAP) more effective than standard care for improving function in a healthy, preterm infant girl with a corrected age of 1 week?

Search Terms: This article had previously been located for purposes of neuroimaging review using the terms "MRI AND infants AND development" on PubMed Clinical Queries for therapy with a narrow, specific search, where the title "Early experience alters brain function and structure" was the most appealing title of 9 articles found on the search. Searching using the above clinical question with the terms "NIDCAP AND premature infants" on PubMed Clinical Queries with a narrow, specific search in the category of therapy yielded 9 articles including the Als article.

The Study: Single-blinded concealed randomized controlled trial with intention-to-treat.

The Study Patients: Medically healthy, pre-term (28-33 week gestation) infants (born to mothers who had prenatal care) with 5-minute Apgar score>=7; weight and head circumference at birth appropriate for gestational age (between 5th and 95th percentile; normal initial cranial ultrasound, MRI, and/or EEG; <=72 hours of ventilator support; <=72 hours of vasopressor medication; free of congenital or chromosomal abnormality, congenital or acquired infection, prenatal brain lesions, and neonatal seizures. Family selection criteria included residence in the greater Boston area; mothers >=14 years; absence of major maternal medical or psychiatric illness, or chronic medication treatment, or history of substance abuse at any time (including alcohol or tobacco), family accessibility by telephone, and some English-language facility. The infants were randomly assigned to control and experimental groups dependent on gender and ethnicity.

Control group (N = 14; 14 analyzed): These 15 infants received "standard care" practiced at the time of the study at the Brigham and Women's Hospital NICU, which included an effort at primary care nursing and staff-dependent inconsistent parent inclusion, uniform shielding of incubators with white hospital blankets, early use of dressing in T-shirts, and side and foot rolls: liberal provision of pacifier; and inconsistent nurse-dependent encouragement of skin-to-skin holding (kangaroo care) and breastfeeding. The control group received this standard care for the duration of their NICU stay.

Experimental group (N = 16; 16 analyzed): The 18 infants of the experimental group received care consistent with the Newborn Individualized Developmental Care and Assessment Program (NIDCAP) approach. This approach to care involved individualized intervention consisted of daily observation and weekly evaluations of the infants' behavior by developmental specialists (a trained psychologist and neonatologist), with suggestions for parents and staff in terms of ways to support each infant's development. The developmental specialist used their formal observations of 91 behaviors to formulate descriptive neurobehavioral reports and suggestions to structure care giving procedures to the infant's sleep/wake cycle, to maintain the infants' well-regulated autonomic, motor, and organization behavioral balance. They also had daily contact with the infants' caregivers to provided ongoing support for the care teams and parents in jointly planning and implementing individually supportive care. The care plans were facilitated by the NICU's nurse managers and medical directors. The nurses caring for the experimental group were trained in this developmental care approach. NIDCAP includes the use of the infant "tuck" by swaddling to position the infants in a more curled up position using soft blankets, bedding, individually sized buntings; supportive holding in the tucked position and cradling in the caregiver's hands; individually adjusted bedding including use of natural sheepskins, terry cloth buntings, soft, special-sized, appropriate body and hugging pillows, and soft special pacifiers; bathing in swaddle blankets in deep, arm water; weighing while swaddled, providing a second caregiver to support the infant during stressful procedures such as suctioning, chest radiographs, and cranial ultrasounds; increasing darkness and quiet for the infant; and supporting the parents in caring for their infant, nursing, and holding their infant in restful, skin-to-skin contact for prolonged periods of time in comfortable recliner chairs in which to relax and hold and sleep with their infants; encourage the parents to personalize their infant's bed area, provide muted, indirect lighting, encouraged parents and staff to make use of custom-made attractive, soft-colored "privacy screen" and crib canopies. The NIDCAP interventions started at the phase of the infant's initial stabilization and then every 7 days throughout hospital discharge and to 2 week's corrected age

The Evidence:

Outcome	Time to Outcome	CER	EER	RRR	ARR	NNT
Pneumothorax	2 weeks corrected age	0.143	0	100%	0.143	7
	95% Confidence Intervals:				-0.040 to 0.326	NNT = 3 to INF; NNH = 25 to INF
Intraventricular Hemorrhage	2 weeks corrected age	0	0.063	INF	-0.063	-16
	95% Confidence Intervals:				-0.182 to 0.056	NNT = 18 to INF; NNH = 5 to INF
Bronchopulmonary Dysplagia	2 weeks corrected age	0.071	0.188	-165%	-0.117	-9
	95% Confidence Intervals:			-494% to 100%	-0.351 to 0.117	NNT = 9 to INF; NNH = 3 to INF

Measure	Control Group		Experimental Group		Difference	95% CI
	Mean	SD	Mean	SD
Weight At Birth, g	1730	350	1648	232	82	-137 to 301
Head circumference	36.70	0.75	36.75	1.18	-0.050	-0.80 to 0.70
5 min. Apgar score	8.5	0.52	7.88	0.081	0.620	0.35 to 0.89
Weight at 2 weeks corrected age, g	4095	518	3999	400	96	-248 to 440
Head Circumference 2 weeks corrected age	36.70	0.74	36.75	1.12	-0.050	-0.77 to 0.67
Total Prechtl at 2 weeks corrected age	38.30	9.55	17.33	15.52	20.970	11.16 to 30.78
MDI at 9 months corrected age	94.85	9.22	109.55	7.23	-14.700	-21.82 to -7.58
PDI at 9 months corrected age	89.23	14.88	107.00	9.28	-17.770	-28.51 to -7.03
BRS, total score at 9 months corrected age	39	23	73	16	-34	-51.10 to -16.90

Comments: Are the Results Valid? This study was a controlled trial design with 2-group randomization. Subjects randomly assigned to either control or experimental group. Assignment was revealed by parental opening of prenumbered, sealed envelope drawn from the respective randomization box, dependent on the infant’s gender and ethnicity (blocked a priori). Allocation of group assignment remained concealed from the assessment staff, this included double-blinding of the pediatrician who derived the medical information from the charts, double-blinding of the psychologist who interviewed the parents in order to obtain any missing demographic or medical data, blinding of the evaluator who performed the neurobehavioral assessments at 2 weeks’ corrected age, and blinding the two independent examiners who performed the 9 months’ corrected age neurobehavioral assessments. Randomization of the groups showed to be equal with no significant differences noted in medical and demographic variable, note the birth weight and head circumference means with 95% CIs that include the null value for between-group mean differences. There was a statistical difference between the two groups based on ethnicity where 8/14 or 0.57 estimated population proportion with 95%CI (.33 to 79) were Caucasian and 15/16 or 0.97 estimated population proportion with 95% CI (.72 to .99) showing a no difference between groups (comparison of proportions or odds ratio of 11.25 with 95% CI (1 to 110), 1 is null and included for odds ratios. The only medical background variable that was statistically different between the control and experimental groups was the 5 minute Apgar, where the experimental group had a 0.62 point lower average being significant with 95%CI (.35 to .89). Intention to treat the experimental group was conducted with 100% consistency according to observation reports. Nurse work records which showed that the experimental group infants (mean 53%) were cared for by nurses rated as developmentally skilled for more hours/shifts (F=11.06, df=1,18; P=.004) in the course of their hospitalization than the control group (38%) . There is a threat to validity in that the intention to treat the control group with standard care had no formal effort to prevent spillover effects from the experimental group. The authors report that the experimental treatment effect on the control group as a conservative influence or not beyond the inevitable spillover where both infant groups were in the same NICU being cared for at the same time. Assigning the control infants to a separate nursery with nurses trained only in standard care would have strengthened the certainty of the standard care regime for the control group. Follow-up at the 2 weeks’ corrected age included all 30 infants of the study. At the time of the 9 months’ corrected age outcomes measures, 1 control (6%) and 5 experimental (36%) infants were lost from follow-up/did not return for assessment because of family scheduling conflicts.

What are the Results? The purpose of the study was to explore the effects of the NIDCAP intervention on low-risk preterm infants. The authors also reported some events to include pneumothorax, interventricular hemorrhage, and brochopulmonary dysplagia. There was significant improvement in the neurobehavioral tests done at both 2 weeks’ and 9 months’ corrected age. The size effect of the BRS total score at 9 months was large at 1.7 (mean score experimental group – mean score control group/standard deviation: 73-39/20). The size effect of the Prechtle was large at 1.7 (38.30-17.33/12.5). The scores of the experimental group on the MDI and PDI (Bayley II) were clinically important differences with improvements equal to twice the standard deviations.

How Can I Apply the Results to Patient Care? The infants in this study differ from my infant patient in that they were all treated in the NICU, and my patient, is in the home setting, 4 weeks after hospital discharge at beginning care at corrected age of 1 week. The study showed that there were significant differences in the neurobehavioral function of the infants that received the individual developmental care in those very early days and weeks in the NICU at periods of time before 36 weeks gestational age. The article also described other clinical outcomes related to significant correlation of improvements in brain function and brain structural maturity in the experimental group. Therapy intervention using some of the NIDCAP care techniques for this infant may be indicated if the infant is having autonomic (breathing, heart rate, color changes, and visceral signs), motor (postures, muscle tone fluctuations, or movements), or state organizational difficulties (levels of arousal, patterns of transitions between states, and clarity and robustness of sleep and awake states). Other studies would need to be found to answer whether or not it is too late in the brain development of my infant patient for NIDCAP to be effective.

Appraised by: Dawn Standley, November 03, 2007 Email: dawnstandley@msn.com

        

        Back to CAT Express Pediatrics         Back to Top

        

        

OTHER:

Determination of the predictive value of Body Mass Index, Waist Circumference and Waist-Hip Ratio in determining risk of getting Type 2 Diabetes Mellitus in men and women

The measures of Body Mass Index (BMI), Waist Circumference (WC) and Waist-Hip Ratio (WHR) are each strongly and independently related to the development of incident type 2 diabetes in both men and women after controlling for possible confounding factors. Level of evidence is 1b.

Citation/s: Meisinger C, Doring A, Thorand B, Heier M, Lowel H. Body fat distribution and risk of type 2 diabetes in the general population: are there differences between men and women? The MONICA/KORA Augsburg Cohort Study. American Journal of Clinical Nutrition. 2006; 84:483-489.

Lead author's name and fax: C Meisinger, Central Hospital of Augsburg, MONICA/KORA Myocardial Infarction Registry, Stenglinstrasse 2 D-86156 Augsburg, Germany. Email: christa.meisinger@gsf.dc.

Three-part Clinical Question: For a 40 year old woman with a BMI of 25, a waist circumference of 86 cm, and a waist-hip ratio of .89, what is her risk of getting Type 2 Diabetes Mellitus?

Search Terms: I used PubMed Clinical Queries to conduct a narrow specific search for Prognosis with the following search string: (Diabetes AND WHR AND incidence AND epidemiology) AND (prognos*[Title/Abstract] OR (first[Title/Abstract] AND episode{Title/Abstract]) OR cohort[Title/Abstract]). The search yielded 26 hits, one of which was the Meisinger et al. article.

The Study: The Study Patients: The study included 7814 subjects, (3916 men and 3898 women) who were aged 35-74 years, and who did not have the diagnosis of Type 2 diabetes mellitus (Type 2 DM) at the beginning of the study time frame. Prognostic Factor: Factors included gender, BMI, WHR (waist-hip ratio), and waist circumference.

The Outcome: Diagnosis of Type 2 Diabetes Mellitus. There was a well-defined sample at a uniform (early) stage of illness. Can't tell if follow-up was long enough; follow-up was not complete. There were blind, objective outcome criteria. Adjustment was made for other prognostic factors. There was validation in an independent test-set of patients.

The Evidence:

Prognostic Factor	Outcome	Result	Measure	Confidence Interval	Independent?
Men	Diagnosis of Type 2 DM	7.95%	Percent	6.96-8.94	yes
Women	Diagnosis of Type 2 DM	5.34%	Percent	4.53-6.15	yes
Men with BMI>29.4	Diagnosis of Type 2 DM	4.15	Hazard Ratio	2.58-6.66	yes
Women with BMI>29.5	Diagnosis of Type 2 DM	10.58	Hazard Ratio	3.81-29.33	yes
Men with WC>102cm	Diagnosis of Type 2 DM	3.40	Hazard Ratio	2.15-5.37	yes
Women with WC>90.5 cm	Diagnosis of Type 2 DM	10.70	Hazard Ratio	3.84-29.80	yes
Men with WHR>0.97	Diagnosis of Type 2 DM	2.84	Hazard Ratio	1.82-4.42	yes
Women with WHR>0.85	Diagnosis of Type 2 DM	3.52	Hazard Ratio	1.88 - 6.57	yes
Men with BMI>29.4 and WC>102cm	Diagnosis of Type 2 DM	3.27	Hazard Ratio	2.4 -4.1	yes
Women with BMI>29.5 and WC>90.5	Diagnosis of Type 2 DM	4.03	Hazard Ratio	2.7-5.3	yes
Men with BMI>29.4 and WHR>0.97	Diagnosis of Type 2 DM	3.7	Hazard Ratio	2.7-4.7	yes
Women with BMI>29.5 and WHR>0.85	Diagnosis of Type 2 DM	5.44	Hazard Ratio	3.6-7.2	yes

Comments: Are the results valid? The patients in this study were all of German nationality, so the results may not be as valid when attempting to apply to people of other ethnicities. Additionally a response bias cannot be excluded in this study, so the possibility exists that those who chose to respond were not as healthy, etc., compared to the population at large and may not be as representative as a result. The patients came from a heterogeneous group of people with a wide range of characteristics that could pose as risk factors, however, these factors were taken into account when performing the statistical analysis, so the adjusted risk of the body measurements can be considered valid. Follow-up was incomplete on all study participants with a 14% drop out rate. A selection bias could have been introduced (for example) with all the skinny people who ended up with diabetes being part of the dropout rate (although this is unlikely due to evidence from other studies), so we do not know the effect this has on the data. The outcome was objective in that the subject either had the diagnosis of Type 2 DM or did not, but the authors of the study relied on the subject's report of this information before confirming it via medical records. The possibility exists that many people had Type 2 DM and were unaware, thus not representing the true prevalence of the disease.

What are the Results? Overall, the likelihood of developing Type 2 DM for men was 7.95% (95% CI 6.96-8.94) and the likelihood of women developing Type 2 DM was 5.34% (95% CI 4.53-6.15). Overall, the incidence of Type 2 DM increased with increasing BMI, WC, and WHR in both sexes. The results were divided into 4 quartiles, with the highest, or 4th quartile indicating the highest risk compared to those in the 1st, or lowest quartile. After looking at the individual associations between the 3 body measures, the authors combined the three different measures to obtain different combined prognostic factors. The authors determined that men with both high BMI and high WC were 3.27 times more likely to get Type 2 DM over men with the lowest BMI and WC. (CI approx. 2.4 - 4.1 via graph) The authors indicated the results were "similar" (approx. 3.7 with CI 2.7 - 4.7 via graph) for the combined BMI and WHR. For women, those with a high BMI and WC were 4.03 times more likely to get Type 2 DM (CI approx. 2.7 - 5.3 via graph) and women with a high BMI and WHR were 5.44 times more likely to get Type 2 DM (CI approx. 3.6 - 7.2 via graph) compared to women with low body measures. The precision in these estimates is such that the confidence intervals are narrow enough to use this data with confidence in determining risk factors for Type 2 DM. Additionally, the lower ends of the confidence intervals in the analysis of the combined measures do not cross 1.0 (or the null value for HR) and are not close to 1.0, further strengthening the prognostic use of this data.

How can I apply the results to patient care? The results of this study can be applied to patients that are adults aged 35 -64, but I would use caution in applying these results to people who fall into high-risk categories for Type 2 DM, including African-American and Hispanic individuals. The possibility exists that this study would underestimate the risk for different ethnicities, as well as for individuals over the age of 64. The included body measures are relatively easy to obtain, so I would have no issue with performing the different body measures in a clinical setting with any patient I felt was at risk. Additionally, if this study overestimates the risk of Type 2 DM in any population (ie individuals that are muscular, resulting in a high BMI, or individuals with very slim hips resulting in a high WHR) the resulting advice to increase physical activity or change the quality of their food intake should not put the patient at harm. The follow up of this study ranged from 8-9 years, and the possibility exists that certain study participants had Type 2 DM and did not know it at follow up and other participants ended up getting Type 2 DM after the study was over, thus underestimating the overall risk. If I look at the possibility that the risks are underestimated and I also consider the cost/benefit of making a few body measurements, I am spending a few extra minutes to make measures that should have no harm to the patient and give them information that could greatly improve their quality of life if they take action. As a result I would be comfortable screening my patients for the risk factors indicated and emphasizing lifestyle modification habits.

Appraised by: Amy Eschenberg, MPT. September 22, 2007 Email: Amy_Eschenberg@Baylor.edu

        

        Back to CAT Express Other         Back to Top

        

EVIDENCE BASED MEDICINE LINKS - some sites may be blocked from government worksites. If so, please consider accessing from home.

Pubmed

NCBI PubMed Clinical Queries

National Library of Medicine

NIH Library - with Librarian consultant to assist with searches, if needed. She will also email full-text prn. cooperd@mail.nih.gov

PT Evidence in Motion

Evidence Based Practice from the University of Washington

Evidence-Based Medicine Health Sciences Library, UNC Chapel Hill

Evidence-Base Medicine Online

Centre for Evidence-Based Medicine

Center for Health Evidence

IHS Rehab Resources

IHS Clinical Questions

IHS Clinical Resources

IHS Online Journals

Virtual Naval Hospital

Physiotherapy Evidence Database

American Physical Therapy Association

Physical Therapy Journal of the American Physical Therapy Association

Journal of Orthopaedic and Sports Physical Therapy

Wheeless Textbook Orthopaedics

American College of Radiology

American Academy of Wound Management

The Cochrane Library

Medcyclopedia e-learning in medicine and medical imaging

 

Look for free full-text access to journal articles at these websites:

Highwire Free Full Text Articles

BioMed Central

Look Smart - Find Articles

Free Full Text

Free Medical Journals

Public Library of Science

Scientific Electronic Library Online

PubMed Central

        

         Evidence Based Medicine Links         Back to Top

        

CLINICAL CONSULTATION FORUMS/BLOGS

Rehab Edge

Soma Simple

My PT Space

NATA Blog

        

                 Back to Top

        

PHYSICAL THERAPY PRESENTATION POSTINGS

Screening for Disease

Evidence Based Therapy Practice

Clinical Presentation of Acute Charcot Arthropathy

Please provide presentations that have an impact on clinical practice. Submit presentation in power point or pdf format to CarlosEstevezPT@yahoo.com.

        

                 Back to Top

        

USER NOTICE: If you note a discrepancy or non-operational link, please contact CarlosEstevezPT@yahoo.com.

 

Therapist Category Home | CPO | TPAC | Administration | Information | Readiness | Strategic Planning | Professional Development | Links Comments or Questions Last updated January 2009