Author Query Form Journal Title Article Number
: :
MDM 342752
Dear Author/Editor, Greetings, and thank you for publishing with SAGE Publications. Your article has been copyedited, and we have a few queries for you. Please respond to these queries when you submit your changes to the Production Editor. Thank you for your time and effort.
Please assist us by clarifying the following queries:
Sl. No.
Query
1
PLEASE PROVIDE KEY WORDS
2
PLEASE PROVIDE VOLUME AND PAGES
Remarks
Increasing the Detection and Response to Adherence Problems with Cardiovascular Medication in Primary Care through Computerized Drug Management Systems: A Randomized Controlled Trial Robyn Tamblyn, PhD, Kristen Reidel, BSc, Allen Huang, MDCM, Laurel Taylor, PhD, Nancy Winslade, PharmD, Gillian Bartlett, PhD, Roland Grad, MDCM, Andre´ Jacques, MD, Martin Dawes, MD, Pierre Larochelle, Alain Pinsonneault, PhD
Background. Adherence with antihypertensive and lipidlowering therapy is poor, resulting in an almost 2-fold increase in hospitalization. Treatment side effects, cost, and complexity are common reasons for nonadherence, and physicians are often unaware of these potentially modifiable problems. Objective. To determine if a cardiovascular medication tracking and nonadherence alert system, incorporated into a computerized health record system, would increase drug profile review by primary care physicians, increase the likelihood of therapy change, and improve adherence with antihypertensive and lipid-lowering drugs. Methods. There were 2293 primary care patients prescribed lipid-lowering or antihypertensive drugs who were randomized to the adherence tracking and alert system or active medication list alone to determine if the intervention increased drug profile review, changes in cardiovascular drug treatment, and refill adherence in the first 6 months. An intention to treat analysis was conducted using generalized estimating equations to account for clustering within
physician. Results. Overall, medication adherence was below 80% for 36.3% of patients using lipid-lowering drugs and 40.8% of patients using antihypertensives at the start of the trial. There was a significant increase in drug profile review in the intervention compared to the control group (44.5% v. 35.5%; P < 0:001), a nonsignificant increase in drug discontinuations due to side effects (2.3% v. 2.0%; P = 0:61), and a reduction in therapy increases (28.5% v. 29.1%; P = 0:86). There was no significant change in refill adherence after 6 months of followup. Conclusion. An adherence tracking and alert system increases drug review but not therapy changes or adherence in prevalent users of cardiovascular drug treatment. Targeting incident users where adverse treatment effects are more common and combining adherence tracking and alert tools with motivational interventions provided by multidisciplinary primary care teams may improve the effectiveness of the intervention. Key words: [AQ: 1]. (Med Decis Making XXXX;XX:xx–xx)
C
controlling hypertension and hyperlipidemia, 2 risk factors for cardiovascular disease.2 Indeed, cardiovascular morbidity has declined by almost 50% in the last decade, and it is estimated that almost half of the reduction in morbidity is attributable to drug treatment.2 While substantial benefits have accrued, the full benefits of drug treatment are compromised
ardiovascular disease is the leading cause of mortality in developed countries.1 Drug treatments can reduce morbidity and mortality by
Received 4 August 2008 from the Department of Epidemiology & Biostatistics, McGill University (RT, KR), the Department of Medicine, McGill University (RT, AH, NW, GB), the Faculty of Management, McGill University (LT, MD, AP), the Department of Family Medicine, McGill University (RG), the College of Physicians of Quebec (AJ), and the Department of Medicine, University of Montreal (PL), Montreal, Quebec, Canada. Source of support from the Canadian Institutes of Health Research and Pfizer Canada Inc. Revision accepted for publication 27 May 2009.
MEDICAL DECISION MAKING/MON–MON XXXX
Address correspondence and reprint requests to Robyn Tamblyn, PhD, McGill University, Morrice House, 1140 Pine Avenue West, Montreal, Quebec, CAN, H3A 1A3; e-mail:
[email protected]. DOI: 10.1177/0272989X09342752
1
TAMBLYN AND OTHERS
by problems in medication adherence.3;4 Studies in the US, UK, and Canada have shown that only 40% to 55% of patients continue to use lipid-reducing therapy in the second year of treatment, even though 4 to 5 years of continuous therapy are needed to reduce the risk of cardiovascular events.5 7 A similar picture emerges for antihypertensive drug use. Only 36% to 40% of people prescribed drugs for hypertension are still using them after 1 year, and 32% to 38% will have switched to another class.8 13 The direct costs associated with nonpersistence are significant. In the UK, Hughes10 estimated that 26.9 million of the 76.5 million for treating hypertension was spent on patients who discontinue initial therapy. Of interest, these costs were predominantly related to more frequent physician visits and hospital admissions among patients switching and discontinuing therapy. A more recent population-based Canadian study confirmed that patients whose adherence level fell below 80% were twice as likely to be hospitalized for cardiovascular problems.4 Nonadherence with other medications was not predictive, suggesting that increased morbidity was not due to placebo biases related to ‘‘healthy adherence effects.’’3;4 To date, there are limited means of alerting physicians about potential adherence problems that would allow them to investigate potential reasons for underuse of medication. Physicians are unable to predict which patients will not take medications as prescribed.14 Indeed, the majority of information physicians obtain about adherence is from patient selfreport, and patients who underuse antihypertensive medication tend to substantially overestimate their adherence.15 Moreover, recent research suggests that adverse effects16;17 or an inability to pay for prescribed treatment16 are responsible for a substantial proportion of nonadherence problems in antihypertensive treatment. Yet only half of all patients will discuss these problems with their physician,18 possibly because limited time is devoted to a discussion of medication management during the clinical encounter.19 As a result, physicians may change medication or add to treatment regimens under the assumption that the medication regimen is ineffective when in reality the medication simply has not been taken as prescribed. Advanced computerized drug management systems provide new opportunities for monitoring drug adherence, particularly in jurisdictions where comprehensive information exists on dispensed medications.20 Increasingly, information on all community-based prescriptions is stored in regional and national repositories, or by third-party insurers, and can be used to assess adherence based on prescription refill rates. 2 • MEDICAL DECISION MAKING/MON–MON XXXX
Refill rates can be used to estimate the proportion of days in which medications are taken. This measure of medication use appears to be a better predictor of therapeutic outcomes in chronic conditions compared to self-report or pill count.21 23 Feedback to physicians on patient medication refill rates may be useful, particularly to primary care physicians as they are responsible for the majority of chronic disease management and could use tools that would enable them to identify and monitor patients who have treatment adherence problems. We had the opportunity to test the benefits of providing primary care physicians with a medication adherence tracking and alert system for patients using cardiovascular medication within a computerized medical record system. This system is integrated with comprehensive information on dispensed medication through the Quebec public insurance program. We tested the hypothesis that physicians who had access to adherence tracking for patients using cardiovascular medication would be more likely to review the drug profile, detect adverse treatment effects and change medication, and would be less likely to increase therapy due to ineffective treatment. We also determined if alerts for adherence problems improved compliance among patients who were noncompliant at the index visit. CONTEXT The MOXXI primary care research program, situated in the Canadian province of Quebec, was used to conduct the study. In this program, primary care physicians use a light electronic health record that comprises an electronic prescribing pad, drug discontinuation and change orders, an automated problem list, a drug profile of all prescriptions dispensed from community pharmacies, dates of emergency room visits and hospitalization, and automated decision support for prescribing problems.24 Information on all drugs and services received by each patient in the practice is retrieved, in real time, from the Quebec provincial health insurance databases. These databases provide information on all medical services for Quebec residents and all prescriptions for the 50% of the population covered by the government drug insurance program. At the time of the study, 59 physicians and 15,486 patients had consented to participate in the research program. Design and Population A single-blind randomized controlled trial was conducted to assess the benefits of providing an
COMPUTERIZED DECISION SUPPORT FOR CARDIOVASCULAR DRUG ADHERENCE
adherence tracking and alert system for patients receiving medications for cardiovascular disease. Physicians and patients were blind to the outcome assessed but not intervention status. Patients within a primary care physician’s practice were randomized to the intervention (complete drug profile, refill compliance calculation, and adherence alerts) or control group (medication list alone). Each patient was followed for 6 months after the index visit to assess the primary (drug profile review, change in therapy) and secondary study outcomes (medication adherence). The McGill Faculty of Medicine Institutional Review Board and the provincial privacy commission provided ethics approval. Patients were eligible for the study if they and their physicians had consented to participate in the MOXXI research program, they were insured with the provincial drug insurance program, and they had at least 1 active lipid-lowering or antihypertensive drug prescribed by the study physician in the 3 months prior to the index visit. The central database server conducted real-time assessment of patent eligibility at the first visit after the start of the study in April 2006, and eligible patients were randomized to intervention or control groups using a randomized block design with randomly selected block sizes of 6, 8, and 12. Intervention and Control In the intervention group, the primary care physician was provided with the drug profile: the patient’s list of current prescribed and dispensed drugs, total costs of medications dispensed each month, the amount of out-of-pocket expenditures paid by the patient (deductibles and copayments), graphic representation of unfilled prescriptions, and days of drug supply for each medication, based on the start and end dates of dispensed medications (Figure 1). Physicians could identify potential adherence problems by gaps in refill dates and/or by clicking on a drug to obtain numerical estimates of refill compliance based on standard algorithms.21;23;25 In addition, at each visit, patient adherence to lipid-lowering and antihypertensive drugs was calculated based on drugs dispensed in the past 3 months. If treatment adherence was less than 80%, the physician received an alert, when opening the chart, to check the drug profile for potential treatment adherence problems. In the control group, the primary care physician had access only to the current list of prescribed and dispensed drugs and did not receive alerts when patient adherence was less than 80% (Figure 2). All
Figure 1 Intervention group: comprehensive information on prescribed and dispensed medication and adherence calculations and alerts.
3
TAMBLYN AND OTHERS
all dispensed medication (drug name, route, strength, quantity, duration, date dispensed, prescribing physician, and dispensing pharmacy) was retrieved daily for each enrolled patient from the provincial drug insurance database. These data are updated, in real time, as part of the online adjudication process for payment with all provincial pharmacies. The MOXXI application databases provided a record of the date of each visit, including drugs that were prescribed, modified, and discontinued, the reason for stopping or modifying medication, and audit trails of access to the drug profile. The physician had to select the reason from standardized pick lists for each order to modify/ discontinue medication. Examples of reasons for changes in therapy included ineffective treatment, intolerance to drug or side effects, trial withdrawal, and incorrect medication dispensed. A prior study has shown that physician-recorded reasons for therapy are valid in that they are concordant with chart documented reasons in 92% of orders.26
Primary Outcomes Drug profile review. Audit trails recorded by the MOXXI application were used to determine whether the drug profile was reviewed on each visit for eligible patients. On average, physicians review the drug profile in 15 per 100 visits, and access is greater for patients on several medications,27 likely because treatment adherence problems are more likely to occur among patients who use many medications.28 We measured drug profile access rates for all patients in the intervention and control groups, and within subgroups, stratified by the number of cardiovascular medications used.
Figure 2 Control group: active medication list alone.
other features of the electronic prescribing and drug management system were available for patients in the intervention and control groups. Outcome Assessment Data sources. The Quebec provincial beneficiary and prescription databases were used to identify and link information on all drugs dispensed with MOXXI electronic health record information through a unique common patient identifier. Age and sex were retrieved from the health beneficiary database. Information on 4 • MEDICAL DECISION MAKING/MON–MON XXXX
Change in therapy. An increase in therapy was defined as an increase in the dose of an active drug, the addition of another drug in the same therapeutic category, or a switch to a new therapy because of ineffective treatment within 6 months after the index visit. Active drugs at the index visit were determined by retrieving all antihypertensive or lipid-lowering drug prescriptions dispensed in the previous 3 months. The daily dose for each active drug was calculated by multiplying the strength per unit dose by the number of doses per day (doses per day = quantity dispensed/prescription duration) for the most recently dispensed prescription. Increase in drug dose was determined by calculating the difference in the dose for the same drug for all prescriptions dispensed in the next 6 months. The addition of drug(s) in the same therapeutic category was
COMPUTERIZED DECISION SUPPORT FOR CARDIOVASCULAR DRUG ADHERENCE
determined by inspecting all electronic and dispensed prescriptions in the 6-month follow-up for new drugs that had been added since the index visit. A switch in therapy due to ineffective therapy was determined by the discontinuation of a drug that was active at the index visit for the reason ‘‘ineffective treatment’’ coupled with the prescription of a new drug in the same therapeutic class. A drug discontinuation due to adverse effects was defined as a discontinuation of a lipid-lowering or antihypertensive drug within the 6 months after the index visit where the reason provided was ‘‘intolerance to drug/side effects.’’ Secondary Outcome Adherence with treatment. Refill adherence, defined as the proportion of days in which an individual had a supply of prescribed medication on the basis of prescription refills, was calculated for each active drug at the index visit.21;23 Refill adherence was estimated in the 3 months prior to the index visit and the 6 months after. Over 97% of prescriptions dispensed in Quebec are for ≤ 30 days’ duration. A drug by day matrix was constructed for each drug, using the dates that prescriptions were dispensed and the recorded duration. To avoid left-sided censoring, all prescriptions dispensed in the 6 months prior to the index visit were retrieved. Only drug supply days covered by prescriptions filled prior to or during the 3 months before and 6 months after a visit were included. When a patient took more than 1 antihypertensive or lipid-lowering drug, the mean adherence of all drugs was used. Adherence for newly prescribed therapy was calculated from the first dispensing date to the end of follow-up, and for discontinued therapy, adherence was calculated to the date of drug discontinuation. Refill adherence was treated as a continuous variable as well as a binary outcome where nonadherence was defined as less than 80% of prescribed treatment. Covariates Patient age and sex were collected at the time of consent. To determine whether the effectiveness of the intervention varied as a function of the number of cardiovascular medications, we measured the number of different lipid-lowering and antihypertensive medications for each patient at the index visit using records of dispensed medications. We also assessed whether the extent of the primary care physician’s involvement in a patient’s management
influenced the effectiveness of the intervention. For each patient, we measured their continuity of drug management, defined as the proportion of all of his/ her lipid-lowering and antihypertensive medication prescriptions that were prescribed by the study physician in the year prior to the index visit. Analysis Descriptive statistics were used to compare characteristics of patients randomized to the intervention and control groups and to summarize baseline compliance. Logistic regression within a generalized estimating equation (GEE) framework was used to test the hypothesis that the availability of the drug profile and alerts for adherence problems would increase drug profile review, increase the likelihood of drug discontinuation due to adverse drug events, and reduce the likelihood that therapy would be increased. The patient was the unit of analysis, the physician was the clustering variable, and an exchangeable correlation structure was used to account for dependence among patient observations for the same physician. Multivariate models that included adherence status at the visit, number of cardiovascular medications, and drug therapy group (lipid-reducing medication, antihypertensive medication) were used to control for differences in patient characteristics between the intervention and control groups that were not balanced by randomization. The same approach was used for the secondary outcome, change in adherence rate for cardiovascular medication, except multivariate linear regression was used to model individual differences in adherence rates between the baseline and follow-up periods. To determine if the effect of the intervention was modified by adherence status at the index visits as well as the number of active cardiovascular medications, we fit interaction terms between adherence status (1 cardiovascular medication) and intervention status. We used the same approach to determine if continuity of drug management modified the effectiveness of the intervention. RESULTS Among the 15,486 patients enrolled in the MOXXI research program, 6372 patients had public
5
TAMBLYN AND OTHERS
Table 1
Demographic Characteristics of Patients Randomized to the Intervention (Adherence Alert + Comprehensive Drug Profile) and Control Group (Medication List)
Patient Characteristics
No. of patients Sex
Female Male Age
