Description

OBJECTIVES AND PURPOSE Our primary objective is to evaluate the effectiveness of provider-facing MedSafer reports, combined with patient-facing educational brochures, for deprescribing PIMs in adults aged ≥60 with rheumatic diseases who have polypharmacy (taking 5 or more regular medications).

HYPOTHESIS MedSafer reports combined with consumer-facing brochures could safely increase PIM deprescription above usual care, improving medication appropriateness.

STUDY DESIGN AND SETTING Prospective before-and-after quality improvement quasi-experimental study within the outpatient rheumatology clinic at the Montreal General Hospital, McGill University Health Centre (MUHC), Montreal, Canada. All participants will be followed during 'control' and 'intervention' periods, and all participants will receive the intervention, acting as their own control.

DESCRIPTION OF INTERVENTION

1. Physician-oriented individualized deprescribing opportunity reports are created for each patient within MedSafer by cross referencing a patients' medical conditions and medications (https://www.medsafer.org/public). Deprescribing opportunities are categorized as "high risk" for adverse drug events if the drug is continued, "intermediate risk" (harms must be weighted against the benefits) and "low risk" for harm (but with little to no added value to continue). Specific tapering instructions, including for sedative hypnotic medications, are provided to complement deprescribing reports. Deprescribing decisions remain at the treating physicians' discretion, and no physician will be obligated to act on a deprescribing suggestion.
2. Patient-oriented educational material, available to the public for routine use (https://www.deprescribingnetwork.ca/patient-handouts).
   • All participants receive a deprescribing fact sheet in either French or English.
   • Participants taking specific PIMs will also receive consumer-facing bilingual 'EMPOWER' educational brochures (French or English).

STUDY PROCEDURES

Pre-screening and recruitment The study will recruit approximately 100 participants during routine clinical visits. A research team member will pre-screen participants for eligibility by previewing clinic lists using the electronic medical record and reviewing medications lists contained within prior clinical notes.

During regular routine clinical visits, a member of the research team will approach consecutive eligible individuals (who have expressed interest in the study after introduction by their treating physician) and invite them to participate in the study. See Eligibility section for inclusion and exclusion criteria. Participants will be aware that the study will evaluate medication use and changes over time, but will be unaware of the specific details of the intervention or which medications will be studied.

Informed Consent Informed consent will be sought from participants. At the time of consent, potential participants will be informed that they can withdraw their participation from the study at any time, including during the intervention period.

Study visits:

All participants will participate in the 'control' and 'intervention' periods (each approximately 10 months duration), with 2 in-person study visits (occurring concurrently with routine care visits) and 2 virtual visits conducted with a study team member outside usual care visits. The timing of in-person visits will be pragmatic, according to usual care.

'Control' period (Visit 1 to Visit 2) Visit 1 (index control visit; in-person) will be the recruitment and enrolment visit. After informed consent, participants will complete a baseline questionnaire collecting self-reported demographics. Rheumatic disease diagnosis and year of diagnosis will be collected from medical records. A research team member will record age, sex, relevant medical history and co-morbidities, medications with doses, renal function, and glycated hemoglobin into the MedSafer application, and generate deprescribing opportunity reports. However, the treating physician will not receive the reports. Medication lists will be accessed as usual from clinical records as part of best practice for clinic visits, and deprescribing at the clinic visit will be left to 'usual care'.

The 5-level version of the EuroQol 5-dimension questionnaire (EQ-5D-5L(31)) will assess baseline quality of life.

Visit 2 (in-person or virtual) will take place approximately 4 months (+/- 1 month) following Visit 1. A research team member will review medical records for any medication changes (cessation, initiation, change in dose), and record emergency department visits or hospitalizations (diagnoses, outcomes of the visits). Participants will complete EQ-5D-5L questionnaires (on paper or electronically). If a routine clinical visit is already taking place during the Visit 2 study window, Visit 2 can also occur in-person at the time of this visit.

Participants may have additional clinical visits during the 'control period', as dictated by their treating physician and their clinical condition (usual care), during which no study procedures will take place.

Intervention period (Visits 3 and 4) Approximately 2 weeks prior to switching to the intervention period, treating physicians will be introduced to the intervention format and procedures.

Visit 3 (index intervention visit; in-person) will be participants' first returning visit to the rheumatology clinic for usual care during the intervention period. Prior to this visit, a research team member will update current medications within the MedSafer application to generate deprescribing opportunity reports. Treating physicians will then receive personalized deprescribing opportunity reports at the patient visit. Physicians will not be obligated to act on deprescribing opportunities, which require clinical judgement.

Participants will complete the EQ-5D-5L, and receive the intervention bundle.

Visit 4 (virtual/telephone) will be approximately 4 months (+/- 1) following Visit 3. As in Visit 2, a research team member will review medical records and contact participants to ascertain any medication changes (cessation, initiation, change in dose), record emergency department visits, and any hospitalizations, along with details (diagnoses, outcomes of the visits). Participants will complete EQ-5D-5L questionnaires. If a routine clinical visit happens to occur during the Visit 4 study window, this visit can also occur in-person.

Criteria for discontinuing or modifying the intervention There will be no interim analysis. The MedSafer ruleset will remain unchanged throughout the study. In an individual participant, any medication can be re-started or dose escalated at a physicians' discretion on clinical grounds. Re-starting or re-escalation that occurs during the study period will be captured in the study outcomes, along with serious adverse drug withdrawal events (ADWEs). Feedback from treating physicians will be evaluated after study completion (a separate study) which will be used to inform future iterations of the intervention.

Strategies to promote uptake of the interventions At the start of the intervention period, the study team will send an introductory notice to the rheumatologists caring for study participants. This document will contain an overview of how study interventions will integrate with the existing workflow, and the expectations for the treating rheumatologist (to review the MedSafer report at the time of performing a medication review, perform any relevant deprescribing). We will provide sample MedSafer reports to physicians (including residents), so they can familiarize themselves with the report format prior to the intervention commencing.

OUTCOMES

1. Primary: proportion with ≥1 PIM deprescribed (stopped, dose reduced, switched to a safer medication class) during the control and intervention periods
2. Secondary 2.1 Change in median total number of medications during control and intervention periods 2.2 Proportion with stable or improved EQ-5D-5L and VAS scores during the control and intervention periods 2.3 Serious Adverse Events (SAEs): any unplanned emergency department visit or hospitalization, and any death 2.4 Serious Adverse Drug Withdrawal Events (ADWEs), defined as any SAE adjudicated to be due to deprescribing. Two independent reviewers, blinded to intervention period, will adjudicate each ASAE for the presence of a serious ADWE using the Leape and Bates Likert scale (Leape \& Bates 1995). A third blinded reviewer will resolve discordant rankings. Scores of 5 or 6 (probably or definitively caused by a medical withdrawal) will define ADWEs, and a score of 4 or greater will be explored in a sensitivity analysis.
3. Additional outcomes: Medication dose re-escalation or re-starts following deprescribing during control and intervention periods.

SAMPLE SIZE We expect that 5% of participants will undergo deprescribing during usual care (control period), based on prior studies. A sample size of 90 participants will permit detection of a 15% absolute increase in deprescribing during the intervention period, with 80% power to detect a 2-sided alpha of 0.05, accounting for up to 10% loss to follow-up.

STATISTICAL ANALYSES The primary analysis will be restricted to those who have clinical visits during both control and intervention periods. We will summarize the study population characteristics, including baseline demographics, disease characteristics, co-morbidities, number of medications, and PIMs (number and medication class).

Primary outcome: the proportion with ≥1 PIM deprescribed during the control and intervention periods will be compared with the McNemar test for paired proportions.

Secondary outcomes:

The paired Student's t-test or Wilcoxon signed rank test will compare the number of regular medications taken at the end of the control and intervention periods.

Event rates for SAEs and ADWEs will be calculated as the number of events during the control or intervention periods divided by the person-time of observation during each period.

We will convert raw EQ-5D-5L scores into index values using a Canadian value set. Change in index scores and VAS during the control and intervention periods will be categorized as deterioration, no change, or improvement using established minimally important differences. The proportion with unchanged or improved EQ-5D-5L values and VAS will be compared between control and intervention periods using the McNemar test for paired proportions.

Additional analyses:

1. A per-protocol analysis will exclude participants who were not exposed to any part of the intervention (e.g. physician did not receive deprescribing reports, participant did not receive deprescribing brochures).
2. Stratified analyses of the primary outcomes by sex (female and male), gender-related variables (e.g. education, employment), age (\<75 vs \>75), rheumatic condition (i.e., inflammatory arthritis vs systemic autoimmune rheumatic disease), baseline number medications (\<10 vs ≥10), and will be performed if sufficient group sizes exist. Exploratory analyses will also stratify secondary outcomes according to these characteristics.
3. Exploratory logistic regression analyses will evaluate factors associated with deprescribing ≥1 PIM, considering age, sex, other demographics, physician, rheumatic condition, number of medications for inclusion, and intervention period in models.
4. Medication re-starts or dose re-escalation following deprescribing will be reported descriptively for the control and intervention periods.

METHODS TO HANDLE PROTOCOL NON-ADHERENCE AND MISSING DATA The primary analysis will be conducted according to the intention-to-treat principle: participants returning for a clinical visit during the intervention period who do not receive the intervention (ie, the physician does not receive or review the electronic decision support, a participant does not receive a relevant educational brochure, if indicated, or the participant is no longer taking a PIM) will be included in the analysis. However, participants who do not return at all for clinical follow-up (ie no virtual or in-person visit) during the intervention period will be excluded from the analysis, as they would not have an opportunity to receive the intervention or have medication data collected. In a sensitivity analysis (above), we will exclude participants who attended a visit during the intervention period but did not receive the intervention (per-protocol analysis). Every effort will be made to reach participants and/or to ascertain reason for not returning to clinic during the intervention period. In study participants do not return for clinical follow-up during the intervention period but can still be contacted, medications and SAEs will still be collected. We will not perform any imputation of missing data.