Research Briefs

Curns, A.T., Steiner, C., Sejvar, J.J., and Schonberger, L.B. (2008, June). "Hospital charges attributable to a primary diagnosis of infectious diseases in older adults in the United States, 1998 to 2004." Journal of the American Geriatric Society 56, pp. 969-975.

This study found that hospital charges for the elderly rose nearly 40 percent—from $31 billion to $46 billion—between 1998 and 2004. Charges rose from $865 to $1,160 for each elderly person hospitalized for an infectious disease. During the same period, the infectious disease hospitalization rate remained relatively stable at 503 hospitalizations for every 10,000 elderly persons. Sustained cost increases of this magnitude have major implications for Medicare, note Claudia Steiner, M.D., M.P.H., of the Agency for Healthcare Research and Quality (AHRQ), and colleagues. The elderly population is expected to double by 2030 from 35 million to 72 million, and by then, 1 in 5 Americans will be age 65 or older. By 2035, spending by Medicare is projected to increase to 7.5 percent of the U.S. Gross Domestic Product from 2.6 percent in 2004. The four leading causes of infectious disease hospitalizations during the 6-year period studied were: pneumonia, acute bronchitis, and other lower respiratory tract infections (45 percent); kidney, urinary tract, and bladder infections (14 percent); septicemia or blood poisoning (13 percent); and cellulitis, a bacterial skin infection (9 percent). The findings were based on data from AHRQ's Nationwide Inpatient Sample.

Reprints (AHRQ Publication No. 08-R073) are available from the AHRQ Publications Clearinghouse.

Glasgow, R.E., Peeples, M., and Skovlund, S.E. (2008, May). "Where is the patient in diabetes performance measures?" (AHRQ grant HS10123). Diabetes Care 31(5), pp. 1046-1050.

Care quality measures, an important strategy to enhance quality of care, are often used by employers or consumers to select health plans or in pay-for-performance initiatives. Diabetes performance measures currently do not include behavioral, quality of life, and other patient-centered measures, which have been successfully incorporated in other countries and by the American Association of Diabetes Educators (AADE), note the authors of this paper. They suggest that seven patient-centered and self-management performance measures be included in assessments of diabetes care quality and as a standard part of diabetes quality indicators. The proposed indicators include: patient self-management goal(s), measures of health behaviors (for example, healthy eating, taking medication, physical activity, and smoking status), quality of life, and patient-centered collaborative care.

Hsu, J., Fung, V., Price, M., and others (2008, April). "Medicare beneficiaries' knowledge of Part D prescription drug program benefits and responses to drug costs." (AHRQ grant HS13902). Journal of the American Medical Association 299(16), pp. 1929-1936.

The study authors interviewed 1,040 community-dwelling elderly Medicare Advantage beneficiaries in California about their knowledge of cost sharing in Part D. They also asked beneficiaries about their cost-related responses to drug coverage such as cost-coping behaviors and reduced drug adherence. Only 40 percent of beneficiaries were aware that their drug plan in 2006 included a coverage gap, with those who reached the gap during the year more likely to know of its existence. More than one-third (36 percent) of beneficiaries reported at least one of the following responses to drug costs: cost-coping behavior, such as switching to lower-cost medications (26 percent); reduced drug adherence (15 percent); or suffering a financial burden due to drug costs (7 percent). After accounting for other factors, those with lower household income more often reported a cost response to drug coverage (difference of 14.5 percentage points for those making less than $40,000 as compared with those making more). Individuals who were unaware of having a coverage gap more often reported a cost response than those who were aware of the gap (difference of 11.3 percentage points), but made fewer reports of borrowing money or going without necessities to pay for medication (difference of 5.5 percent).

Jasti, H., Mortensen, E.M., Obrosky, D.S., and others, (2008, February 15). "Causes and risk factors for rehospitalization of patients hospitalized with community-acquired pneumonia." (AHRQ grants HS08282 and F32 HS00135). Clinical Infectious Diseases 46, pp. 550-556.

Researchers examined factors associated with rehospitalization for community-acquired pneumonia among 577 patients discharged from 7 Pittsburgh hospitals. Overall, 70 (12 percent) were rehospitalized within 30 days. About 74 percent of rehospitalizations were related to coexisting medical conditions and 20 percent were due to treatment failure for or worsening of pneumonia. Patients with underlying chronic obstructive pulmonary disease and coronary artery disease were two to three times more likely to be rehospitalized than patients without these conditions. The researchers recommend that discharge strategies include vaccination for flu and pneumonia when indicated, clear instructions on correct use of medications, review of signs or symptoms that may suggest worsening of the patient's underlying medical conditions, and an emphasis on appropriate outpatient followup within a week after discharge if there is a risk for rehospitalization. The study was supported in part by the Agency for Healthcare Research and Quality.

Kilbridge, P.M. and Classen, D.C. (2008, July/August). "The informatics opportunities at the intersection of patient safety and clinical informatics." (Contract No. 290-04-0016). Journal of the American Medical Informatics Association 15(4), pp. 397-407.

In this position paper, the authors recommend several ways to improve adoption of health information technology (IT) systems to enhance patient safety. They recommend that the Federal government, via the Agency for Healthcare Research and Quality (AHRQ), help fund an organization to manage health care data standards to make data and systems interoperable. Because of the cost of implementing health IT systems, they suggest the Federal Government should provide seed money to assist them in purchasing systems. Electronic health record systems should include agile, useful clinical decision support systems to improve currently available systems that frustrate users, causing them to override alerts or disable the systems altogether. These improvements could follow an AHRQ-funded Clinical Decision Support Roadmap that would pinpoint best practices of decision support applications. The authors recommend that health IT systems undergo a certification process by an agency like the Certification Commission for Health Information Technology to catch potential errors before they harm patients. They suggest that AHRQ further fund its research on informatics incentives in pay-for-performance programs to develop models that improve patient safety and fund safety-related research on personal health records. Finally, the authors recommend that Congress direct AHRQ to fund an aggressive research agenda on the intersection of patient safety and health informatics.

Matheny, M.E., Sequist, T.D., Seger, A.C., and others (2008 July/August). "A randomized trial of electronic clinical reminders to improve medication laboratory monitoring." (AHRQ grant HS11046). Journal of the American Medical Informatics Association 15(4), pp. 424-429.

The researchers studied 20 ambulatory care settings to see if adding electronic alerts about needed lab tests to an electronic health record affected the behavior of 303 primary care physicians for 1,922 patients. From January to June 2004, the study looked for patients who were overdue for at least one laboratory test for potassium or creatinine levels, liver or thyroid function, and therapeutic drug levels for medications such as phenobarbital or phenytoin. Only 5 percent of patients on targeted medications in the study were due for laboratory monitoring. Because compliance rates were already high, researchers found that the electronic alerts issued through an electronic health record did not cause an upswing in laboratory monitoring. High rates of laboratory monitoring may be explained by the fact that the health system used in the study's ambulatory care settings effectively shares information through a clinical data repository and electronic health records to give clinicians a comprehensive picture of patient care, the authors suggest.

Neuman, H.B., Brogi, E., Ebrahim, A., and others (2008, January). "Desmoid tumors (fibromatoses) of the breast: A 25-year experience." (AHRQ grant T32 HS00066). Annals of Surgical Oncology 15(1), pp. 274-280.

To chart the history of how breast desmoids are handled, researchers reviewed their 25 years of experience with the tumors, examining symptoms, diagnosis, and treatment. They found that all 32 patients who had breast desmoids from 1982 to 2006 had physical signs that were suspicious for cancer. Twenty-eight had lumps the physician could feel, and six had skin dimpling. Of the 16 patients who had mammograms, a mass was detected in 6, but no mass was found in 10. Magnetic resonance imaging, however, detected the mass for all eight patients who underwent it. To prevent damage of nearby tissue, surgery is the most common treatment for desmoids. Eight of 28 patients for whom followup information was available experienced repeat tumors, which also required surgery in most cases (6 of 8 patients). Therapies to thwart future tumors include nonsteroidal anti-inflammatory drugs (indomethacin and sulindac), hormone therapy (tamoxifen), and cytotoxic therapies, which kill specific cells.

Raab, S., Andrew-Jaja, C., Grzybicki, D.M., and others (2008). "Dissemination of lean methods to improve pap testing quality and patient safety." (AHRQ grant HS13321). Journal of Lower Genital Tract Disease 12(2), pp. 103-110, 2008.

Process redesign can reduce diagnostic errors in Pap smears comparable with some new technologies. In a 1-year case-control study, researchers compared the diagnostic accuracy of Pap tests procured by 5 clinicians prior to implementing a lean process redesign before (5,384 controls) and after its implementation (5,442 cases). The process redesign involved a checklist of procedures for the clinician to obtain and the laboratory technologist to process a cervical tissue sample. Following process redesign implementation, there was a significant decline in the mean proportion of Pap tests lacking the required 10 well-preserved endocervical or squamous metaplastic cells to interpret the cervical sample. Also, two of five clinicians showed a significant decrease in their unsatisfactory Pap test frequency, but results for the overall case group was not significantly lower. Finally, the case group showed a 114 percent increase in newly detected cervical intraepithelial cancer following a previous benign Pap test.

Sampson, M., Shojania, K.G., Garritty, C., and others (2008). "Systematic reviews can be produced and published faster." (AHRQ Contract No. 290-02-0021). Journal of Clinical Epidemiology 61, pp. 531-536.

The value of systematic reviews is influenced by many factors, especially the emergence of new data. The interval between the last date of searching and the availability of the review is a period when the evidence base is at risk of becoming outdated. It is also a time when the results are unavailable to prospective users. The authors studied the currency of systematic reviews at the time of publication to determine typical and achievable times to publication for reviews published in journals. The eligible reviews included in the study consisted of 156 quantitative systematic reviews published between 1995 and 2005. Most (59 percent) were journal-published reviews, 23 percent were Cochrane reviews, and 17 percent were technical reports. The median time from final search to publication was 61 weeks with an interquartile range of 33-87 weeks. The authors recommend that the first quartile of the best performing review type is a reasonable target for those who produce and publish reviews. Thus, the final search would be within 10 weeks of submission, acceptance within 11 weeks of submission, and publication within 12 weeks of acceptance.

Sampson, M., Shojania, K.G., McGowan, J., and others (2008). "Surveillance search techniques identified the need to update systematic reviews." (AHRQ Contract No. 290-02-0021). Journal of Clinical Epidemiology 61, pp. 755-762.

Updating systematic reviews is important for preserving their usefulness to clinicians and researchers. The authors sought to determine the performance characteristics of various search methods and to assess their feasibility as surveillance strategies for those interested in monitoring the biomedical literature for the purpose of updating systematic reviews. The 77 systematic reviews chosen for the study were drawn from major peer-reviewed journals. Of the five surveillance search approaches selected for study, three were subject searches: the optimized Clinical Query, the Core Clinical Journals subset together with the randomized controlled trials (RCT) publication type and the Cochrane Collaboration's Central Register of Controlled Trials. The other two methods were a "related articles" search using PubMed and a "citing reference" search for RCTs. Since none of these methods yielded a consistently high recall of relevant new evidence, the authors turned to a combination of search strategies. The most successful was a search algorithm based on PubMed's related article search combined with a subject search using clinical queries. It retrieved all relevant new records in 68 cases.

Stahl, J.E., Kreke, J.E., Malek, F.A.A., and others (2008, June). "Consequences of cold-ischemia time on primary nonfunction and patient and graft survival in liver transplantation: A meta-analysis." (AHRQ grant HS09694). PLoS ONE 3(6), p. e2468.

The ability to preserve organs prior to transplant is essential to the organ allocation process. The researchers performed a meta-analysis to determine the impact of cold-ischemia time (CIT) on patient and graft survival in liver transplant. CIT is the time interval that begins when an organ is cooled with a cold perfusion solution after organ procurement surgery and ends when the organ is implanted. After searching MEDLINE®, EMBASE, and the Cochrane database, the researchers identified 26 studies that met their criteria. Patient survival was measured at 1, 3, 6, and 12 months following liver transplant. At each survival interval, maximum patient and graft survival occurred with CITs between 7.5 and 12.5 hours. Organ and patient survival were worse for both high and low CITs. This could be due to a disadvantageous combination of patients and organs. Patients receiving organs less than 5 hours from harvest or more than 12.5 hours from harvest were more likely to be sicker than average.

Sussman, A.L., Williams, R.L., Leverence, R. and others (2008, July-August). "Self determination theory and preventive care delivery: A research involving outpatient settings network (RIOS Net) study." (AHRQ grant HS13496). Journal of the American Board of Family Medicine 21(4), pp. 282-291.

Traditional approaches to improving delivery of preventive care have met with limited success. The researchers conducted a study using self-determination theory (SDT) to clarify clinician-level factors that influence preventive care delivery by examining the psychology of clinician decisionmaking. Using obesity counseling as a case study, they gathered data through a three-stage process involving interviews, focus groups, and mailed surveys. More than 146 clinicians, all members of a primary care research network, took part in the study. Response data were compared to the three principal domains of SDT theory: autonomy, competence, and relatedness. Clinicians expressed a strong sense of autonomy (acting out of personal choice). However, factors both within and external to the clinical setting presented critical barriers to the clinician's sense of competency (ability to achieve a desired outcome) in obesity preventive counseling. The third domain, relatedness (the desire to achieve meaningful relationships and belongingness with one's colleagues) was generally lacking within the practice setting. The authors concluded that, absent effective tools to prevent obesity and opportunities for relevant clinician interactions with colleagues and community members, high levels of clinician autonomy and internal motivation are not sufficient to maintain counseling efforts.

Tate, J.E., Simonsen, L., Viboud, C., and others (2008, May). "Trends in intussusception hospitalizations among US infants, 1993-2004: Implications for monitoring the safety of the new rotavirus vaccination program." Pediatrics 121, pp. e1125-e1132.

In 2006, a new vaccine was recommended to immunize U.S. infants against rotavirus, the most common cause of severe gastroenteritis in young children. A previous rotavirus vaccine had been withdrawn in 1999 after it was associated with intussusception, a problem in which a portion of the bowel slides into the next, causing bowel obstruction. Claudia Steiner, M.D., M.P.H., of the Agency for Healthcare Research and Quality, and colleagues examined annual prevaccine intussusception hospitalization rates of bowel obstruction to establish a baseline with which to compare rates after introduction of the new rotavirus vaccine. Intussusception hospitalization rates declined 25 percent from 1993 to 2004, but have remained stable at about 35 cases per 100,000 infants since 2000. Rates varied nearly 12-fold by week of age during the 6- to 32-week age range for vaccination, and less so by race/ethnicity. Although the downward trend in hospitalization rates might reflect a true reduction in incidence of severe intussusception, it could also reflect changes in management practices, such as greater use of nonsurgical interventions that do not require hospitalization, note the researchers.

Reprints (AHRQ Publication No. 08-R071) are available from the AHRQ Publications Clearinghouse.

Trentham-Dietz, A., Sprague, B.L., Klein, R., and others (2008). "Health-related quality of life before and after a breast cancer diagnosis." (AHRQ grant HS06941). Breast Cancer Research 109, pp. 379-387.

Over 2 million women in the United States are living with breast cancer. Improved therapies mean that more women with a breast cancer diagnosis are surviving longer and that issues of quality of life (QOL) merit greater attention. These researchers examined QOL among women before and after diagnosis and whether those changes differed substantially from changes experienced by all women during aging. They compared QOL for 114 women with breast cancer with 2,527 women without breast cancer. QOL was measured by the Medical Outcomes Study Short Form 36 Health Status Survey, which was administered four times between 1991 and 2002. Compared with women without breast cancer, women with breast cancer reported lower scores on physical function, physical role function, bodily pain, general health, vitality, and social function scales. Although average scores, adjusted for age, were 4.5 points lower on the Physical Component Summary scale, there was no difference in the Mental Component Summary scale between women with or without breast cancer.

Uronis, H.E., Currow, D.C., McCrory, D.C., and others (2008). "Oxygen relief for dyspnoea in mildly- or non-hypoxaemic patients with cancer: A systematic review and meta-analysis." (AHRQ grant T32 HS00079). British Journal of Cancer 98, pp. 294-299.

A review of studies on the use of palliative oxygen to relieve breathlessness toward the end of life shows that oxygen failed to relieve the sensation of refractory dyspnea in cancer patients who would not otherwise qualify for home oxygen therapy (they were mildly- or non-hypoxemic). However, limitations in the data make it difficult to come to firm conclusions on such an important issue. The researchers conducted a systematic review and meta-analysis of studies comparing oxygen therapy (delivered by nasal cannula, mouthpiece, or face mask) with medical air (clean, compressed air) in cancer patients not qualifying for home oxygen therapy. The studies examined the ability of these interventions to improve symptoms of dyspnea at rest or on exertion (6-minute walk). Based on a review of 204 citations and 54 articles, the researchers found only 4 studies and a total of 134 patients with cancer that qualified for analysis of the benefit of palliative oxygen. Three studies evaluated oxygen versus medical air for relief of dyspnea, and the fourth study evaluated use of Heliox28, a novel agent containing 72 percent helium and 28 percent oxygen versus oxygen and medical air. The overall quality of the studies was poor.

Weech–Maldonado, R., Fongwa, M.N., Gutierrez, P., and Hays, R.D. (2008, April). "Language and regional differences in evaluations of Medicare managed care by Hispanics." (AHRQ grants HS09204 and HS16980). HSR: Health Services Research 43(2), pp. 552-568.

Hispanics enrolled in Medicare managed care programs are less positive about their care experiences than are non-Hispanic whites, according to a study that used data from the Consumer Assessment of Healthcare Providers and Systems (CAHPS®) Medicare managed care survey. The 2002 survey included 125,369 respondents enrolled in 181 Medicare managed care programs nationally. More than half of Hispanics (52 percent) insured through Medicare were enrolled in managed care Medicare programs in 2002, as opposed to standard fee-for-service Medicare. The researchers used survey items about timeliness of care, provider communication, office staff helpfulness, getting needed care, and health plan customer service to compare Hispanics with non-Hispanic whites, and Hispanics who answered the survey's English questionnaire with those who answered the Spanish questionnaire. Additional analyses compared respondents by geographic region, and according to other socioeconomic factors such as education, income, gender, age, and health status. English-speaking Hispanics viewed all aspects of their care worse than whites did, except for provider communications. Their Spanish-speaking counterparts reported more negative care experiences than whites with timeliness of care, provider communications, and office staff helpfulness, but were more satisfied with getting needed care.

Werner, R.M., Goldman, L.E., and Dudley, R.A. (2008, May). "Comparison of change in quality of care between safety-net and non-safety-net hospitals." (AHRQ grant HS16117). Journal of the American Medical Association 299(18), pp. 2180-2187.

Safety-net hospitals that predominantly treat poor and underserved patients typically lack the resources necessary to invest in quality improvement or even to ensure accurate data collection for performance measurement. Researchers examined trends in disparities of care quality between hospitals with high percentages of Medicaid patients (safety-net hospitals) and low percentages of Medicaid patients (non-safety-net hospitals), using publicly available data on hospital performance between 2004 and 2006. Of the 3,665 hospitals studied, safety-net hospitals had worse performance in 2004 and significantly smaller improvement over time than non-safety-net hospitals. Over time, the safety-net hospitals had a lower probability of achieving high performance status. In addition, based on a simulation model, these hospitals were more likely to incur financial penalties due to low performance and were less likely to receive bonuses.

Wu, A.C., Smith, L., Bokhour, B., and others (2008, March/April). "Racial/ethnic variation in parent perceptions of asthma." (AHRQ grant T32 HS00063). Ambulatory Pediatrics 8(2), pp. 89-97.

Researchers interviewed parents of 739 children with persistent asthma in a Medicaid health plan and multispecialty provider group in Massachusetts. Overall, 75 percent, 84 percent, and 89 percent of Latino, black, and white children's parents, respectively, believed their children could be symptom-free most of the time. Also, 43 percent, 44 percent, and 55 percent of Latino, black, and white children's parents, respectively, expected their children should have no emergency room visits or hospitalizations for asthma. These differences held, even after controlling for parental age, gender, and household income. Black (32 percent) and Latino (38 percent) parents were more likely than white (23 percent) parents to agree that their children did not need as much medicine as the doctor prescribed. Finally, parents of black (18 percent) and Latino children (23 percent) were more likely to have competing family priorities "all of the time" or "most of the time" in addition to their children's asthma compared with parents of white children (8 percent), even after adjusting for income, education, insurance, and other factors.

Xie, X-J., Titler, M.G., and Clarke, W.R. (2008). "Accounting for intraclass correlations and controlling for baseline differences in a cluster-randomised evidence-based practice intervention study." (AHRQ grant HS10482) Worldviews on Evidence-Based Nursing 5(2), pp. 95-101.

Cluster-randomized designs have been increasingly used in health care and community-based intervention studies due to administrative convenience, a desire to minimize treatment contamination, and the need to avoid ethical issues that might arise. Yet this approach presents challenges for data analysis. These authors use a pain management intervention study to present two strategies that can be used when analyzing data from a cluster-randomized design—both of which account for baseline differences. One approach involves use of a mixed model, and the other involves a marginal model (using generalized estimating equations). Although the parameter estimates and their standard errors might be comparable for both strategies for certain link functions, the interpretations are quite different and each of the two approaches is suitable for answering different questions. The choice of strategy should be dictated by whether the primary interest is a population or individual.

Zaydfudim, V., Stover, D.G., Caro, S.W., and Phay, J.E. (2008, July). "Presentation of a medullary endocrine neoplasia 2A kindred with Cushing's Syndrome." (AHRQ grant T32 HS13833). The American Surgeon 74(7), pp. 659-661.

Medullary thyroid cancer (MTC) is rare, comprising three to five percent of all thyroid cancers, and fewer than 1,000 cases annually. Even rarer is the occurrence of Cushing's syndrome (CS) resulting from ectopic adrenocorticotropic hormone (ACTH), which is found in only 0.6 percent of all patients with MTC. The authors discuss a case of a 51-year-old woman diagnosed with ectopic ACTH production from an inoperable metastatic MTC to the liver. This is the first case of a medullary endocrine neoplasia (MEN) 2A kindred presenting with CS from ectopic ACTH production by metastatic medullary thyroid carcinoma. Genetic testing revealed a germ line RET proto-oncogene mutation at codon 609. This same mutation was also found in six other family members. The patient received palliative bilateral laparoscopic adrenalectomies with a significant improvement in her major comorbidities (weakness, exertional dyspnea, hypertension, striae, and hirsutism). This type of surgery allows for symptom reduction in patients that may survive for years even with widely metastatic disease. Five other family members have undergone thyroidectomy with central lymph node compartment dissection; the sixth family member will receive treatment after he reaches the age of five.

Zhou, X. and Cheng, H. (2008). "A computer program for estimating the re-transformed mean in heteroscedastic two-part models." (AHRQ grant HS13105). Computer Methods and Programs in Biomedicine 90, pp. 210-216.

The population of health care costs is typically skewed, heteroscedastic, and may include zero costs. Without proper accounting for these special distributional features, resulting cost predictions may be biased, and wrong inferences about the distribution of patients' health care costs may be made. The researchers developed a computer program for modeling skewed, heteroscedastic data with zero observations. The program is an implementation of the two-part regression model proposed by Welsh and Zhou. It computes two nonparametric estimators of the mean, their asymptotic standard derivation, estimated confidence interval, and optional bootstrap confidence interval. It can run in both user-friendly interactive mode and more efficient batch mode. It also provides flexibility for users to extend the program to a more general context. The two-part regression model can be generalized by recoding some of the modularized functions. For parameters in the two-part regression model, users can choose from different estimates and/or different numerical approaches.

Return to Contents
Proceed to Next Article