Deprescriptions, Inappropriate Prescribing, Medical Overuse
Conditions
Keywords
Physician-Patient Relations, Implementation Science, Decision Making, Shared
Brief summary
One mechanism to reduce potentially inappropriate medications is through deprescribing, a deimplementation-based approach to thoughtfully discontinue a medication a patient is currently prescribed. Many interventions to overcome deprescribing barriers target the provider, who is already overburdened. Although some believe providers have primary responsibility for deprescribing, patient-initiated discontinuation discussions can effectively facilitate deprescribing. In a single-site pilot study, the investigators successfully engaged VA Primary Care patients to facilitate deprescribing of select potentially inappropriate medications. The investigators now propose a multisite randomized controlled trial of engaging Veterans who may be deprescribing candidates. By study end, the investigators will have established the effectiveness of an innovative, low-tech, patient-focused intervention to promote deprescribing, thereby directly improving quality, safety, and value of VA care while also setting the stage for generalization of this approach to other potentially inappropriate medications.
Detailed description
Background - Despite multiple provider- and system-level interventions to reduce potentially inappropriate medications (PIMs), many Veterans are still prescribed drugs that provide little benefit, placing them at unnecessary risk of adverse drug events (ADEs). One mechanism to reduce PIMs is through deprescribing, a de-implementation-based approach to thoughtfully discontinue a medication a patient is currently prescribed. Many Choosing Wisely recommendations address PIMs. Specifically, proton pump inhibitors (PPIs), a medicine used to reduce gastric acid, should be de-escalated to the lowest dose necessary to provide relief. Many older patients with diabetes are over-controlled, with blood sugar levels lower than recommended, yet remain on multiple diabetes medicines and may be able to use fewer medicines. These patients are also at higher risk of low blood sugar from insulin and sulfonylureas, and should have limited use of these agents. Finally, gabapentin is often used off-label to treat pain, with greatly increased use over the past several years. There are many barriers to deprescribing PIMs. Many interventions solely target the prescribing provider. Although some believe providers have primary responsibility for deprescribing, patient initiation of discontinuation conversations can effectively facilitate deprescribing. In a single-site pilot study, the investigators successfully reduced PIMs by engaging VA Primary Care patients by providing them with Veteran-centric EMPOWER (Eliminating Medications through Patient Ownership of End Results) brochures. However, it is not known if this approach will be as successful for Veterans with other chronic conditions or at non-pilot sites. Aims - The investigators propose three aims. 1) Examine the impact of a patient-centered intervention to change provider prescribing (the primary outcome), as determined by the frequency with which medications are either deprescribed or de-escalated. 2) Examine the effect of a patient-centered intervention on engaging patients, via post-visit surveys of Veterans' interaction with the brochures and their influence on deprescribing discussions and deprescribing. 3) Using qualitative methods, identify key organizational contextual factors related to intervention fidelity, feasibility, acceptability, and appropriateness to support future implementation. Methods and Innovation - The investigators propose a multisite quasi-experimental trial using a Hybrid Type I Effectiveness-Implementation design of providing EMPOWER brochures directly to Veterans who may be deprescribing candidates for three cohorts of PIMs (PPIs, diabetes medications, and gabapentin). The investigators will mail brochures in advance of scheduled primary care visits, unlike distribution methods used in other studies. The primary outcome will be the composite of deprescribing and de-escalation of target medications, identified in pharmacy dispensing records of the Corporate Data Warehouse (Aim 1). Mail-based surveys sent after the scheduled primary care visit will assess patient engagement with the brochure and its impact on patient-provider communication (Aim 2). Finally, qualitative data from clinicians and staff addressing Proctor's Implementation Outcomes will provide the foundation for future implementation strategies (Aim 3). Significance and Next Steps - The study directly addresses multiple Veteran Care Priorities, including health care value, primary care practice, quality/safety, and Whole Health, and is aligned with current VA initiatives to prioritize patient preferences via individually-tailored, proactive care plans. The proposed work is strongly supported by Pharmacy Benefits Management and Office of Patient Centered Care and Cultural Transformation, which will facilitate the dissemination of findings to improve the quality and safety of medication use within VA. By study end, the investigators will have established the effectiveness of an innovative, low-tech, patient-focused intervention to promote deprescribing of commonly used medications for three populations, thereby directly improving quality, safety, and value of VA care while also setting the stage for wider implementation and generalization of this approach to other potentially inappropriate medications.
Interventions
An EMPOWER medication brochure, adapted to the VA, designed to educate and activate patients. These brochures provide detailed medication information, allow self-testing of indications for use, prompt reflection of experiences with potential side effects, discuss alternative therapies (medication and non-pharmacologic options), and provide a vignette of a patient who successfully stopped the medicine. They were designed for a 6th grade reading level and were based upon theories of patient activation, adult learning, and cognitive dissonance. The visually appealing brochure repeatedly emphasizes that patients should not make any medication changes without first consulting their health care providers.
Sponsors
Study design
Intervention model description
The investigators will target Veterans at three primary care sites who meet eligibility criteria for one of the PIM cohorts and are prescribed the target medication at the time of their scheduled primary care visit. Each PCP was assigned to receive three medication groups (i.e., PPIs, diabetes medications, and gabapentin) in a randomized order, and during each period of time for each medication/medication group assignment, prescriptions for that specific medication/medication group were assessed. Patients that received the specific medication or medication group from the PCP during the time of assignment were enrolled in the study. The study will begin with a 13-month retrospective baseline period, with one month for baseline subject identification and 12 months for observation of baseline subjects' deprescribing outcomes. The primary comparisons are between brochure-intervention patients and baseline patients (matched in terms of eligibility) from the same provider.
Eligibility
Inclusion criteria
* Veteran with a Primary Care appointment at one of three Veteran Affairs Medical Centers (including Community Based Outpatient Clinics) * PPI Cohort: * \>90 consecutive days of PPI at any dose * Diabetes Cohorts: * HbA1c \<7% * At least one of Age \>65 years * Renal impairment * Cognitive impairment * either \>90 consecutive days insulin or sulfonylurea or \>90 consecutive days of \>2 DM medications (neither of which is insulin or sulfonylurea) * Gaba Cohort: * \>90 consecutive days with total daily dose \>1800mg
Exclusion criteria
* PPI Cohort Exclusions: * Diagnosis warranting PPI treatment * Medication warranting PPI treatment * Gaba Cohort Exclusions: * Neuropathic pain * Seizure disorder * and/or Cancer-related pain
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Deprescribing | 6 months post-index date | Non-refill of the medication in the 6 months following the primary care appointment (i.e., cessation) or any reduction in the total daily dose (i.e., de-escalation). There is an exception to the de-escalation rule for insulin, where only complete cessation will qualify since dose changes are less likely to be reflected in the order compared to oral medications. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Deprescribing Conversations | Index visit until survey completion (up to 8 weeks post-index visit) | Patient report of a conversation about deprescribing during the index primary care visit. Note that this Outcome Measure was only assessed in a subsample of the Intervention Group Arm/Group participants and not at all in the Historical Control Group Arm/Group subjects. |
Countries
United States
Participant flow
Participants by arm
| Arm | Count |
|---|---|
| Intervention Group Patients who received an EMPOWER brochure two weeks before their primary care appointment.
Direct to patient medication brochure: An EMPOWER medication brochure, adapted to the VA, designed to educate and activate patients. These brochures provide detailed medication information, allow self-testing of indications for use, prompt reflection of experiences with potential side effects, discuss alternative therapies (medication and non-pharmacologic options), and provide a vignette of a patient who successfully stopped the medicine. They were designed for a 6th grade reading level and were based upon theories of patient activation, adult learning, and cognitive dissonance. The visually appealing brochure repeatedly emphasizes that patients should not make any medication changes without first consulting their health care providers. | 3,206 |
| Historical Control Group The historical control cohort were seen by the same PCPs at the same sites as our intervention cohort from October 2019 to April 2020 (eighteen months to one year before the intervention timeframe).
The PPI group included adult patients of any age with an active prescription for any PPI with at least 90 consecutive days on the medication in the prior year, excluding diagnoses for which PPI continuation would be appropriate (e.g., Barrett's esophagus, or those prescribed chronic anti-inflammatory drugs associated with peptic ulcers).
The Gabapentin group included adult patients of any age with an active prescription of gabapentin \>1800mg/day with at least 90 consecutive days on the medication in the prior year, excluding patients with potential indications for use, including neuropathic pain, seizure disorders, and cancer-related pain.
The hypoglycemia-risk group included patients with diabetes based upon ICD-10-CM diagnosis codes, most recent HbA1c \<7%, as well as one or more of the following criteria: 1) age 65 years or older, 2) renal insufficiency defined as creatinine \>2 mg/dL, and/or 3) a diagnosis of cognitive impairment or prescription for an acetylcholinesterase inhibitor (e.g., donepezil). Patients in the hypoglycemia-risk group had an active prescription for insulin or sulfonylurea with at least 90 consecutive days on the medication in the prior year. | 2,740 |
| Total | 5,946 |
Withdrawals & dropouts
| Period | Reason | FG000 | FG001 |
|---|---|---|---|
| Overall Study | Death | 86 | 110 |
| Overall Study | Provider did not have historical control patients | 581 | 0 |
| Overall Study | Provider did not have intervention patients | 0 | 98 |
Baseline characteristics
| Characteristic | Historical Control Group | Total | Intervention Group |
|---|---|---|---|
| Age, Continuous | 70.61 years STANDARD_DEVIATION 12.59 | 70.68 years STANDARD_DEVIATION 12.3 | 71.15 years STANDARD_DEVIATION 12.03 |
| Medication Group Gabapentin | 112 Participants | 233 Participants | 121 Participants |
| Medication Group Hypoglycemic | 471 Participants | 932 Participants | 461 Participants |
| Medication Group Proton Pump Inhibitor | 2157 Participants | 4781 Participants | 2624 Participants |
| Race/Ethnicity, Customized American Indian or Alaskan Native | 36 Participants | 78 Participants | 42 Participants |
| Race/Ethnicity, Customized Asian | 53 Participants | 89 Participants | 36 Participants |
| Race/Ethnicity, Customized Black | 297 Participants | 756 Participants | 459 Participants |
| Race/Ethnicity, Customized Missing | 292 Participants | 647 Participants | 355 Participants |
| Race/Ethnicity, Customized Native Hawaiian or Pacific Islander | 33 Participants | 60 Participants | 27 Participants |
| Race/Ethnicity, Customized White | 2029 Participants | 4316 Participants | 2287 Participants |
| Sex: Female, Male Female | 141 Participants | 336 Participants | 195 Participants |
| Sex: Female, Male Male | 2599 Participants | 5610 Participants | 3011 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk |
|---|---|---|
| deaths Total, all-cause mortality | 86 / 3,206 | 110 / 2,740 |
| other Total, other adverse events | 0 / 2,539 | 0 / 2,532 |
| serious Total, serious adverse events | 25 / 2,539 | 26 / 2,532 |
Outcome results
Deprescribing
Non-refill of the medication in the 6 months following the primary care appointment (i.e., cessation) or any reduction in the total daily dose (i.e., de-escalation). There is an exception to the de-escalation rule for insulin, where only complete cessation will qualify since dose changes are less likely to be reflected in the order compared to oral medications.
Time frame: 6 months post-index date
Population: It is pre-specified in the Study Protocol and Statistical Analysis Plan to assess all medication groups combined within the Intervention and Control Arms/Groups
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Intervention Group | Deprescribing | 748 Participants |
| Historical Control Group | Deprescribing | 653 Participants |
Deprescribing Conversations
Patient report of a conversation about deprescribing during the index primary care visit. Note that this Outcome Measure was only assessed in a subsample of the Intervention Group Arm/Group participants and not at all in the Historical Control Group Arm/Group subjects.
Time frame: Index visit until survey completion (up to 8 weeks post-index visit)
Population: Note that this Outcome Measure was only assessed in a subsample of the Intervention Group Arm/Group participants who responded to the survey. It was not assessed at all in the Historical Control Group Arm/Group subjects.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Intervention Group | Deprescribing Conversations | 12 Participants |
| Historical Control Group | Deprescribing Conversations | 37 Participants |
| Proton Pump Inhibitor (PPI) | Deprescribing Conversations | 353 Participants |