|
Medication Appropriateness Index (MAI)
Please note that this section is an archive (last updated in June 2006). [disclaimer]
Sections:
Overview |
Instrument Reviews |
Construct Overviews |
Book Compendium Reviews |
Internet Site Reviews
|
|
|
|
Practical Information
|
Instrument Name:
|
Medication Appropriateness Index (MAI)
|
Instrument Description:
|
The MAI employs implicit criteria to judge the appropriateness of medication prescribing, using the criteria to evaluate each medication individually. (Ref: 1) It measures the magnitude of inappropriate prescribing for most dimensions of drug use that are clinically relevant. (Ref: 7) The MAI was developed using a) the clinical experience of a clinical pharmacist and internist-geriatrician, along with b) a background literature review regarding medication assessment measures and drug-related problems, and c) a study using a sample of ten elderly veterans. The developers independently identified key areas of desirable medication use in order to create an instrument that could “address multiple elements of drug therapy prescribing, applicable to a variety of medications, clinical conditions and settings.” From their gathered information, they created the ten criteria of the MAI, which are presented in the instrument as worded as questions. (Ref: 1)
The criteria include: 1) medication indication, 2) effectiveness, 3) dosage, 4) correct directions, 5) drug-drug interactions, 6) drug-disease interactions, 7) expense, 8) practical directions, 9) therapeutic duplication, and 10) duration. (Ref: 1-3,5,8) Each criterion includes specific instructions, an operational definition, and examples of appropriate and inappropriate ratings. The MAI does not address the areas of adverse drug reactions and patient medication compliance. (Ref: 1,8) Most of the research on the MAI has been performed by the same set of researchers using the same cohorts of elderly veterans. (Ref: 1-5)
|
Price:
|
Free; article appendices (instructions and materials available from Dr. Hanlon). (Ref: 1)
|
Administration Time:
|
Ten minutes per drug assessed. (Ref: 1)
|
Publication Year:
|
1992
|
Item Readability:
|
No information found.
|
Scale Format:
|
Three-point Likert scale (a “don’t know” response is included as a fourth response option)
|
Administration Technique:
|
Trained rater, using medical charts or chart abstractions; patient interview with simultaneous interviewer access to medical chart is also possible; training is essential. (Ref: 7)
|
Scoring and Interpretation:
|
1= “appropriate medication use,” 2= “marginally appropriate medication use,” and 3= “inappropriate medication use.” Each item/criterion also contains the response option “don’t know.” (Ref: 1) Researchers have also formulated a weighted summated MAI score, with items clustered into three groups: A=indication and effectiveness items, B=dosage, correct directions, drug-drug interaction and drug-disease interaction items, C=practical directions, expense, duplication and duration items. Each “inappropriate” rating in group A receives a weight of 3, each “inappropriate” rating in group B receives a weight of 2, and each “inappropriate” rating in group C receives a weight of 1. Therefore, the possible value of the summated MAI score for a single medication ranges from 0-18. (Ref: 2-3,5)
|
Forms:
|
No information found.
|
|
|
Research Contacts
|
Instrument Developers:
|
J.T. Hanlon, K.E. Schmader, G.P. Samsa, et al.
|
Instrument Development Location:
|
Center for the Study of Aging and Human Development
Box 3003
Duke University Medical Center
Durham, NC 27710
FAX: 919-684-8569
|
Instrument Developer Email:
|
|
Instrument Developer Website:
|
|
|
|
Annotated Bibliography
|
1. Hanlon JT, Schmader KE, Samsa GP, Weinberger M, Uttech KM, Lewis IK, Cohen HJ, Feussner JR. A method for assessing drug therapy appropriateness. J Clin Epidemiol. 1992 Oct;45(10):1045-51. [PMID: 1474400]
|
Purpose:
To report on the development of the new MAI and evaluate its reliability as determined from the study’s sample.
|
|
|
Sample:
1) Reliability testing random sample: ten elderly ambulatory male patients attending the GMC of the Durham VAMC and each using five or more regularly scheduled medications and as a group using 60 total medications; 2) generalizability testing random sample: ten different elderly ambulatory male patients attending the GMC of the Durham VAMC, using five or more regularly scheduled medications and as a group using 105 total medications.
|
|
|
Methods:
This research was undertaken as part of a randomized controlled health services research trial. A clinical pharmacist and an internist-geriatrician independently identified key elements of appropriate medication use, based on their clinical experience and a background review of the literature. From this information they developed the MAI. The MAI was pilot tested and then a formal evaluation of its reliability was performed on a random sample of ten elderly VAMC Internal Medicine patients. The clinical pharmacist and internist made blinded assessments of the patients’ medication use at baseline and two to four months later based on chart reviews and the MAI. To determine generalizability, two different clinical pharmacists performed independent and blinded assessments of drug therapy appropriateness using chart reviews and the MAI in a separate random sample of ten elderly patients at the same VAMC clinic. |
|
|
Implications:
The MAI may be able to reliably detect drug therapy appropriateness. It may also have potential use as a quality of care outcome measure in research applications and in institutional quality assurance programs. However, study sample sizes are small (N=10), study population is restricted (elderly male ambulatory patients who are veterans), and convincing validity evidence is not presented, all suggesting that more psychometric evaluation is warranted. Raters use of a single set of prepared patient chart abstractions may have elevated the reported level of rater agreement.
|
2. Samsa GP, Hanlon JT, Schmader KE, Weinberger M, Clipp EC, Uttech KM, Lewis IK, Landsman PB, Cohen HJ. A summated score for the Medication Appropriateness Index: development and assessment of clinimetric properties including content validity. J Clin Epidemiol. 1994 Aug;47(8):891-6. [PMID: 7730892]
|
|
Purpose:
To describe the development and validation of a weighted summated scoring scheme that produces a single MAI score per medication prescribed. |
|
|
Sample:
1) Survey item weighting sample: convenience sample of ten academic healthcare professionals; 2) polypharmacy study: two hundred eight elderly veterans 65 years or older, enrolled in a randomized trial, who used five or more prescription medications and received primary care at the Durham VAMC (mean age=69.8 years, 99% male, 76% white, mean chronic conditions=10.3 and mean regularly prescribed medications=7.9); 3) reliability testing sample: ten elderly veterans from a GMC at the Durham VAMC (see Ref: 1: Generalizability sample).
|
|
|
Methods:
A survey was designed and distributed to a convenience sample of ten healthcare professionals to determine how MAI items should be weighted. Healthcare professionals were asked to rate the importance of each MAI item. On the basis of their ratings, a weighting scheme was derived from which a single, summated score could be computed. To assess the clinimetric properties of the MAI, the researchers referred to baseline medications prescribed to patients in a randomized trial. A clinical pharmacist completed MAIs on these patients’ 1644 prescribed medications and MAI scores were then examined for their distributional properties. To measure reliability, MAI scores were developed from two clinical pharmacists’ ratings of 105 medications from ten patients, using the new weighting scheme. MAIs were completed twice for each medication, over a period of four months. Inter- and intra-rater reliability were calculated.
|
|
|
Implications:
MAI summated scores can range from 0 (no items deemed inappropriate) to 18 (all items deemed inappropriate). Studied distribution had a mean summated MAI score of 2.2, SD=2.1, range=0-10, and skew=positive. Content validity of the summated MAI is supported by clinician ratings of all MAI items as either definitely or moderately important. Inter-rater reliability estimated to be .74. However, these findings are limited by the use of a restricted population and the minimal number of raters to date (two) used in the assessment of reliability.
|
3. Schmader K, Hanlon JT, Weinberger M, Landsman PB, Samsa GP, Lewis I, Uttech K, Cohen HJ, Feussner JR. Appropriateness of medication prescribing in ambulatory elderly patients. J Am Geriatr Soc. 1994 Dec;42(12):1241-7. [PMID: 7983285]
|
|
Purpose:
To assess the quality of medication prescribing using the MAI with a group of elderly patients on multiple medications. Researchers sought to describe the prevalence and nature of prescribing appropriateness and to determine if medication and/or patient factors influenced appropriateness. |
|
|
Sample:
Two hundred eight elderly outpatients at the Durham VAMC’s GMC, each using five or more regularly scheduled medications for a total of 1,644 medications. (See Ref: 2: polypharmacy study). |
|
|
Methods:
Data were collected from patients enrolled in a randomized controlled trial and prior to their assignment to groups. A clinical pharmacist prepared a problem list from abstracted medical charts and medication lists and evaluated all patients and medications using the MAI. Medications were then classified as being either high- or low-risk, depending on their likelihood of causing adverse drug reactions.
|
|
|
Implications:
Medication prescribing for elderly outpatients using multiple medications was generally appropriate. Physicians tended to prescribe high-risk drugs more appropriately than low-risk drugs. Areas in possible need of prescribing improvement were exact directions, cost of drugs, and practical directions. Patient characteristics assessed (including age, race, marital status, education, number of chronic conditions, self-perceived general health, number of scheduled medications, patient-rated medication helpfulness, and multiple prescribers) were not predictive of average patient MAI score. Physician characteristic influence on average patient MAI score was not investigated.
|
4. Hanlon JT, Weinberger M, Samsa GP, Schmader KE, Uttech KM, Lewis IK, Cowper PA, Landsman PB, Cohen HJ, Feussner JR. A randomized, controlled trial of a clinical pharmacist intervention to improve inappropriate prescribing in elderly outpatients with polypharmacy. Am J Med. 1996 Apr;100(4):428-37. [PMID: 8610730]
|
|
Purpose:
To evaluate the effect of ongoing clinical pharmacist intervention on appropriate medication prescribing for elderly outpatients with polypharmacy while remaining in full association with their PCPs. |
|
|
Sample:
Two hundred eight elderly veteran patients aged 65 years or older from the Durham VAMC, using five or more regularly scheduled medications. (See Ref: 2: polypharmacy study).
|
|
|
Methods:
Randomized trial assessing the following outcomes: prescribing appropriateness, health-related quality of life, adverse drug events, medication compliance and knowledge, number of medications, patient satisfaction and physician receptivity. Patients were randomized into control (receiving usual care) or intervention (receiving usual care and clinical pharmacist care) groups by computer generation. Patients were followed for one year, then received a closeout telephone interview. A clinical pharmacist assessed appropriateness of medication prescribing using the MAI at randomization and again three and 12 months later; other study outcomes were measured using instruments such as the SF-36, patient self-report, and the Health Care Attitude Questionnaire (four items only). Physician receptivity to pharmacist intervention was also assessed via a recommendation enactment measure and interview. |
|
|
Implications:
Inappropriate medication prescribing was reduced in the intervention group to a greater extent than in the control group. Fewer intervention group patients than control group patients experienced adverse drug events. Prescribing appropriateness increased, in large part due to physician enactment of the recommendations of the clinical pharmacist. A clinical pharmacist intervention can reduce inappropriate medication prescribing and adverse drug events, though it did not improve patient-reported quality of life.
|
5. Schmader KE, Hanlon JT, Landsman PB, Samsa GP, Lewis IK, Weinberger M. Inappropriate prescribing and health outcomes in elderly veteran outpatients. Ann Pharmacother. 1997 May;31(5):529-33. [PMID:9161643]
|
|
Purpose:
To determine the relationship of inappropriate prescribing to health outcomes in the elderly; in particular, 1) to measure the association between MAI scores for a patient’s drug regimen and health services use, and 2) to determine the association between MAI scores for a patient’s antihypertensive medications and blood pressure control. |
|
|
Sample:
Two hundred eight elderly veterans from the Durham VAMC’s GMC involved in an intervention trial. They were aged 65 years or older and were using five or more medications. Mean age=69.8 years, 76% white, 76% married, 99% male, mean chronic conditions=10.3, mean prescribed medications=7.9, mean years of education=10.0. (See Ref: 2: polypharmacy study). |
|
|
Methods:
A summated patient MAI score was determined by a clinical pharmacist in order to assess prescribing appropriateness. Health outcomes included hospitalization, unscheduled ambulatory or emergency visits, and blood pressure control (for study patients that were hypertensive). The relationship between MAI scores and health outcomes, regardless of intervention or control group status, was examined. |
|
|
Implications:
Inappropriate medication prescribing was related to adverse health outcomes. Patients with less appropriate prescribing were more likely to be admitted to the hospital or to have unscheduled ambulatory or emergency care visits. MAI scores for antihypertensive medications were higher for patients with inadequate blood pressure control, compared with those with adequate blood pressure control. |
6. Fitzgerald LS, Hanlon JT, Shelton PS, Landsman PB, Schmader KE, Pulliam CC, Williams ME. Reliability of a modified medication appropriateness index in ambulatory older persons. Ann Pharmacother. 1997 May;31(5):543-8. [PMID:9161645]
|
|
Purpose:
To evaluate the reliability of a modified MAI in elderly, non-VA outpatients. |
|
|
Sample:
Ten community dwelling elderly in a university-based health services intervention study, aged 65 years or older, using five or more regularly scheduled medications. Mean age=72.1 years, 60% male, 90% white, mean number of medications=6.5. |
|
|
Methods:
A previous study approach was replicated, using a modified MAI in this group of non-VA elderly. Subjects in the intervention study were interviewed by telephone to assess medication history, while their medical records were simultaneously available to the interviewer for review. Ten subjects were randomly selected to receive a medication appropriateness assessment from two clinical pharmacists using the modified MAI. Inter-rater reliability was determined. Modifications to the MAI included a revised definition for “ineffective” and refined directions for instructions, procedures to assess drug interactions, and methods to determine medication expense. |
|
|
Implications:
The modified MAI displayed reasonable reliability characteristics with non-VA elderly patients. It could be useful in assessing the pharmaceutical care of the elderly. |
7. Murray MD. Medication appropriateness index: putting a number on an old problem in older patients. Ann Pharmacother. 1997 May;31(5):643-4. [PMID:9161665]
|
|
Purpose:
This reference refers to an opinion letter that discusses the MAI.
|
|
|
Sample:
non-applicable
|
|
|
Methods:
non-applicable
|
|
|
Implications:
Preliminary data on the MAI are encouraging for its use with elderly patients.
|
8. Buetow SA, Sibbald B, Cantrill JA, Halliwell S. Appropriateness in health care: application to prescribing. Soc Sci Med. 1997 Jul;45(2):261-71. Review. [PMID: 9225413]
|
|
Purpose:
To discuss how appropriateness has been defined and applied in prescribing and in health care studies in general.
|
|
|
Sample:
non-applicable |
|
|
Methods:
non-applicable |
|
|
Implications:
The MAI provides a foundation for identifying areas of prescribing appropriateness.
|
|
|
top
|
Factors and Norms
|
Factor Analysis Work:
|
No information found.
|
Normative Information Availability:
|
No normative information using large samples was found. Most research has been performed using elderly veterans, however, one study has reported generalizability of the scale in ten community dwelling elders. (Ref: 6)
|
|
|
Reliability Evidence
|
Test-retest:
|
Intra-rater reliability was assessed across time (from baseline to two-to-four months later). Agreement for drugs overall was 97%, with a kappa of 0.92. (Ref: 1)
|
Inter-rater:
|
In the developmental article, the authors report that inter-rater agreement for individual items ranged from 92%-100% and was 93% for drugs overall. Chance adjusted agreement was reported using the kappa statistic, which ranged from 0.71 to 0.96 for individual items and was 0.83 for drugs overall. In this study, inter-rater assessments were also made by independent researchers using a second sample. They found here that overall agreement was 89% and reported a kappa of 0.59 for all drugs overall. (Ref: 1). Using a modified MAI, inter-rater agreement for each criterion rated “appropriate” ranged from 87-100% and for each “inappropriate” rating ranged from 47-100%. Overall, the chance adjusted agreement was kappa=0.64. (Ref: 6) A study using a summated MAI score in veterans found that the intraclass correlation coefficient (ICC) was 0.74; another study using non-veterans found ICC=0.80. (Ref: 2,6)
|
Internal Consistency:
|
No information found.
|
Alternate Forms:
|
No information found.
|
|
|
Validity Evidence
|
Construct/ Convergent/ Discriminant:
|
No information found.
|
Criterion-related/ Concurrent/ Predictive:
|
One study demonstrated known groups validity by using the MAI to judge appropriateness in a group receiving clinical pharmacist intervention versus a control group. Covariate adjusted MAI scores were comparable at randomization (p=0.90), but a 24% improvement in prescribing was found at three months in the intervention group compared with 6% in the control group (p=0.0006). At 12 months, the intervention group showed 28% improvement, while the control group showed 5% improvement (p=0.0002). (Ref: 4)
|
Content:
|
Developers report that content validity is supported. Using baseline data, they examined the frequency distribution of scores to assess whether they adequately reflected the supposed heterogeneity in prescribing appropriateness among the elderly veteran population. Heterogeneity was shown: 25.5% of medication received a score of 0 (no “inappropriate” items), 38.5% had scores of 1-2, and 36% had scores of 3 or above. (Ref: 2) Identification of the MAI’s key criteria from the literature and from clinical experience supports content/construct validity. (Ref: 1,8) The development of the MAI’s summated weighting scheme by external clinical “experts” also supports content validity. (Ref: 2,8)
|
Responsiveness Evidence:
|
No information found.
|
Scale Application in VA Populations:
|
Yes. (Ref: 1-5)
|
Scale Application in non-VA Populations:
|
Yes. (Ref: 6)
|
|
|
Comments
|
This is a brief, promising scale, which could be an important tool in evaluating the quality of medication management. Limited psychometric measures are reported (primarily inter-rater reliabilities). These are generally supportive of the instrument’s stability across raters; however, the overall and item specific kappas reported range from modest to excellent. There is also supportive, but limited, evidence of responsiveness. Further, most work to-date has been done by the same group of investigators, using samples of elderly veterans from a limited set of studies. Thus, additional evidence would be helpful in establishing the stability and generalizability of the instrument in other settings and populations. The MAI is perhaps best considered as a promising medication prescribing appropriateness assessment tool that is still in the stages of development, with a comprehensive determination of the instrument’s external validity yet to be made.
|
Updates
No information found.
|
|
|
|
