PURPOSE
To organize primary care providers, and their older patients with polypharmacy, in a way that permits them to collaboratively determine the lowest effective dose of all amenable medications. And, in so doing, to provide randomized controlled trial evidence as to whether minimizing medication doses improves health outcomes in this population.
HYPOTHESIS
1. Primary care providers can easily reduce medication doses in their own patients when organized appropriately to do so.
2. Determining and utilizing the minimum effective dose of all medications will prolong independence, reduce hospitalization, and potentially prolong life.
JUSTIFICATION
A) THE PROBLEM
The number of medications used to treat chronic conditions increases with age. "Polypharmacy" is said to be present when the number of daily medications is high, although the threshold to meet that definition (e.g. ≥6 daily medications) is not universally agreed upon.
According to data provided by Alberta Health Services (AHS), 62.2% of community-dwelling Albertans ≥80 yrs (74,525 individuals as of Dec 31, 2016) renew 6 or more medications annually. Many such individuals are frail, and virtually all have multiple chronic conditions that place them at high risk of adverse health events - events that can threaten their independence.
The potential for adverse drug-drug and drug-disease interactions increases exponentially with each additional co-morbidity, and each additional medication. Although prescribed for good reasons, all medications have potential for adverse effects. This is especially true for seniors, who are more susceptible to adverse drug effects, in whom adverse drug effects may go unrecognized, and in whom these drugs are seldom studied (only 7% of cardiovascular trial participants are over 75).
Polypharmacy is associated with disturbances of balance, and deterioration of gait (walking ability). Observational studies that control for comorbidity (the presence of various health conditions) additionally show the number of daily medications to be associated with falls, fractures, adverse drug events, all-cause hospitalization, cognitive impairment, diminished physical function (e.g. ability to cook or clean for yourself), and death. Conceivably, if such associations are causal, reducing the number of medications (termed "deprescribing") might reduce the cost of care while simultaneously improving quality of care, potentially prolonging independence, delaying mortality, and improving a sense of well-being.
B) THE SOLUTION
BACKGROUND: Efforts to reduce polypharmacy typically 1) focus on stopping drugs deemed potentially inappropriate for seniors, and 2) employ external experts (e.g. pharmacists or geriatricians) to evaluate medications and make recommendations to the primary care provider. Although these are undoubtedly valuable reviews, the number of medications stopped following such an intervention is not large. Cochrane systematic review suggests five patients must be reviewed to stop one potentially inappropriate drug, in one patient. However, the vast majority of emergency room visits for adverse drug events are for COMMONLY USED MEDICATIONS that are NOT deemed potentially inappropriate (e.g. glucose lowering medications, opioids, anticoagulants, antihypertensives). Potentially inappropriate drugs make up only 3.4% of such presentations.
A NEW PERSPECTIVE: Most adverse drug reactions are dose-related, the optimal dose of most mediations is lower in seniors, and 2/3 of older adults want to reduce their medications. A simple evidence-based approach that appropriately MINIMIZES (e.g. reduces by 50%) doses of ALL amenable medications in older adults may do more to reduce adverse drug events than targeting the subset of potentially inappropriate medications for discontinuation.
AN UNRECOGNIZED OPPORTUNITY: Relying on external experts to review medication lists is not readily scalable without significant cost. It also produces little change in medications since the usual care provider often declines the changes, citing patient context and history not available to the reviewer. If a mechanism existed to enable primary care providers to rationalize/minimize THEIR OWN PRESCRIBING, it would be far more cost-effective and spreadable.
PILOT DATA: To determine if family physicians can reduce their own prescribing, the investigators organized 2 large clinics to present medication profiles for 44 randomly selected patients (≥80 yrs with ≥6 daily medications) to each patient's provider (14 family physicians, 1-5 patients/physician). These physicians identified 25.2% of medications as potentially reducible or stoppable if patients wanted that and came in for medication review. This compares to 2.5% of medications being stopped (albeit potentially inappropriate medications) in conventional deprescribing trials.
THE TRIAL: The investigators intend to implement organizational changes that will enable primary care providers to focus on minimizing/rationalizing the prescribing of all medications in their older polypharmacy patients - something widely recommended as good geriatric prescribing practice. The intention is to do this in over 150 primary care practices spread widely across Alberta, and use administrative claims data to assess differences in patient-oriented outcomes (e.g. all-cause death, nursing home admission) between those receiving early versus delayed targeting for medication minimization.
OBJECTIVES
1. To determine, for older adults with polypharmacy, whether a medication review by their own primary care provider, will lead to lower medication dosages.
2. To determine, for older adults with polypharmacy, whether lower medication doses will result in better health outcomes.
RESEARCH METHOD/PROCEDURES
Setting: "Real-World" primary care clinics widely distributed across Alberta.
Population: All community dwelling adults (no exclusions) ≥80 years of age using ≥6 longterm medications.
Design: Parallel randomized controlled trial with patient level randomization. No stratification or blocking, however eligible patients with the same residence (likely to be spouses) will be allocated to the same study arm to minimize confounding.
Intervention: Workflow efficiencies enabling patients and their usual primary care provider to work together on minimizing medication dosages - i.e. advising providers of eligible patients, arranging minimization specific visits, "brown-bag" review of medications, symptom-driven minimization, and adding a lower limit to the target range for blood pressure (e.g. systolic BP 130-140 mmHg), and blood sugar (e.g. HbA1c 7.5-8.0).
Comparator: Usual care
Outcomes:
All outcomes are derived from administrative claims data that are collected, accessed, and analyzed by Alberta Health Services (Alberta Provincial Government). Specific outcomes are listed below.
STUDY FLOW:
1. IDENTIFYING/ENGAGING PROVIDERS: The principal investigator, and other family physician co-investigators, will continually recruit practices to participate in MinMed. Most participating practices will be affiliated with the Pragmatic Trials Collaborative (www.PragmaticTrials.ca), with new practices being recruited by word of mouth, and through presentations at continuing medical education events.
2. IDENTIFYING PATIENTS: AHS Provincial Research Data Services will identify all patients ≥80yrs with ≥6 longterm medications attending each participating provider, and give those providers, randomly determined, the names of HALF their eligible patients to work with ("early" minimization group), delaying the other half (comparison group) until study completion. Patients sharing a home address (often spouses) receive the same allocation. Longterm medications are defined as those dispensed ≥2 times in the preceding 200 days.
3. INVITING PATIENTS: Providers will receive, for each early-minimization patient, one un-addressed postage-paid envelope containing a letter of invitation to a medication minimization visit. Provider offices will attach mailing labels, and mail these envelopes. Interested patients make a provider appointment and bring with them all pill bottles. Provider offices make follow-up phone-calls to letter non-responders to confirm whether or not they are interested in participating.
4. MINIMIZING MEDICATIONS: Providers follow 2 simple rules: 1) For medications treating a symptom, work collaboratively with patients to find the lowest effective dose and consider reducing by ¼ or ½ dose every few weeks to get there; and 2) For drugs treating blood pressure or blood sugar - introduce lower limits to BP and HbA1c that are 10mmHg lower than the intended BP target (e.g. targeting 130-140mmHg instead of \<140mmHg), and 0.5 units lower than the intended HbA1c target (e.g. 7.5-8.0 instead of \<8.0). This permits compliance with guidelines but avoids unnecessarily low values. Booking visits specifically for drug review are critical to enabling this process, given such reviews seldom occurs during time-pressured problem-focused visits. Bringing all pill bottles to the visit ("brown bag" review) also allows providers to see what patients are actually taking, and clarifies discussions around specific medications and what to change.
5. FUTILITY ANALYSIS: After the patients of 20 practices have undergone 6 months follow-up investigators will analyze the process outcomes. If, on average, fewer than 10% of oral medications have a 50% or better reduction in the number of milligrams dispensed then we will 1) interview participating providers who have completed their reviews to gather insight into barriers and facilitators of medication minimization, 2) discuss amongst co-investigators whether there are process improvements which could be implemented to improve the minimization process, and 3) discuss whether the observed differences in medication use are clinically important enough to warrant continuation of the trial (if not, the trial will break early and report it's process outcomes).
6. EXAMINING PRIMARY / SECONDARY / SAFETY / COST OUTCOMES: The primary clinical outcome is a composite of all-cause death and admission to nursing home. Once 300 such events are observed, the study will break and evaluate implementation, clinical, and cost outcomes using administrative claims data. An Independent Data Safety Monitoring Board (organized by UBC's Therapeutics Initiative) will evaluate data mid-study (after 150 events) to ensure no signals of harm. If p is ≤ 0.001 for benefit (the Haybittle-Peto boundary - recommended to reduce the chance of stopping too early and magnifying benefit), or if p is ≤ 0.05 for harm, the DSMB will apply clinical judgement and make recommendations to the steering committee on whether the trial should break early.
PLAN FOR DATA ANALYSIS
All data analysis will be conducted by AHS data analysts. At no time will the investigators view or receive identifiable patient level data. Only the analysts/programmers working for the data steward, and the care providers, will know who the participants are.
Most primary / secondary / safety outcomes will be analyzed by Cox proportional hazards survival analysis or multiple linear regression utilizing covariates predictive of the outcome of interest. These covariates will be analysis specific, and pre-identified in an upcoming protocol publication.
