Potentially inappropriate medications (PIMs) can lead to adverse drug events (ADEs) among older adults and especially those classified as frail. ADEs are responsible for nearly 27,000 hospital admissions in Canada annually. Within 30 days of discharge, medications contribute to two-thirds of adverse events, with nearly 60% being preventable or ameliorable. MedSafer is software product that guides patients and physicians in the community through the process of deprescribing. MedSafer electronically cross-references patient comorbidities with the most recent evidence-based PIMs, as of the study date, in order to generate a patient-tailored deprescribing care plan. This study will evaluate whether this application, when applied at hospitalization, leads to a reduction in the proprotion of patients with PIMs prescribed at discharge, by highlighting harmful medications for deprescription to treating physicians. This study will take place on the clinical teaching units (CTUs) at four hospitals. Based on historical records, the investigators estimate enrolling 480 patients aged 65 or older over three months. A trained research assistant will identify eligible patients, and will enter their medications, comorbidities, and an estimate of frailty into MedSafer. A deprescribing plan will be generated for the CTU team containing rationale for suggested medication changes and strategies for safe and successful deprescription. The CTU team will then decide in conjunction with the patient or proxy, and with relevant consultants, which medications can rationally be stopped or tapered at discharge.

Deprescribing during hospitalization can be achieved through the concerted effort of expert clinicians. A single-centre randomized controlled trial (400 patients) demonstrated that by having a dedicated geriatrician apply the STOPP/START criteria [screening tool of older people's prescriptions (STOPP) and screening tool to alert to right treatment (START)] to screen for PIM use on admission could improve prescribing appropriateness by providing recommendations for medication management; these changes were shown to persist for at least six months post discharge. Recommendations were made in 58% of intervention patients, and these were accepted in greater than 90% of cases. At discharge the proportion of patients taking "unnecessary" therapy was 20% in the control group and 5.2% in the intervention arm. This study was underpowered to show meaningful differences in ADEs and was entirely dependent on a dedicated expert physician whose full time job was to provide this service; despite this, medications were discontinued in nearly 1 in 5 patients with a sustained post-discharge effect.

This catalyst program will build on the Canadian Frailty Network-funded Catalyst Research Program Grant (MEdication RAtionalization or MERA) by Co-Investigator James Downar. MERA identifies medication for deprescribing by combining STOPP/START, the Beers criteria, and Choosing Wisely Canada recommendations. A multidisciplinary case team then guides elderly patients and their caregivers through deprescription during their hospitalization. Both the Gallagher and MERA approaches are intensive in terms of human resources and are challenging to generalize, implement and sustain nationwide. Furthermore, the current lack of firm evidence on the risks and benefits of systemic deprescribing has limited the implementation of these concepts into routine medical practice. A technological solution that facilitates deprescribing and with proven effectiveness and safety would be a tremendous step forward for the deprescribing movement.

Research Objectives: The investigators will adapt and implement an electronic tool called MedSafer to identify PIMs within 72 hours of hospitalization and prioritize medications for deprescription. The tool will use three key criteria for deprescription: the potential for a drug to improve symptoms, the potential to reduce the risk of future illness, and the likelihood of causing harm. The recommendations from MedSafer will cue the treating team to consider deprescribing PIMs after taking into consideration their knowledge of the individual patient comorbidities, along with an understanding of patient values and goals of care.

The primary goal of this pilot study is to demonstrate that MedSafer can facilitate deprescription (beyond usual care) in a diverse population of elderly patients. This pilot will serve as the proof of concept for a larger multicenter RCT looking at the impact of this process on adverse drug events. This pilot will facilitate: 1) estimates of the frequency of PIMs; 2) adverse drug event rates; 3) rates of recruitment, refusal and loss to follow up; 4) refinement of MedSafer based on real world user feedback; and 6) refinement of the adverse drug event endpoint adjudication and health-related quality of life interview process. The study will be conducted at the Ottawa Hospital, the Toronto General Hospital and two hospitals at the McGill University Health Centre, in Montreal (Royal Victoria and Montreal General Hospitals).

Primary objective: The primary outcome is the proportion of patients for whom 1 or more PIMs are stopped at discharge. We hypothesize that MedSafer will facilitate deprescription of PIMs compared to usual practice.

Secondary objectives: To determine if this intervention leads to demonstrable improvements in: a) Proportion of PIMs at 30-days post hospital discharge; b) adverse drug events within 30 days of hospital discharge; c) falls in hospital; d) length of stay; e) in-hospital mortality; f) the overall number of any medications the average patient receives at discharge and 30 days post; g) post-discharge self-reported health-related quality of life (HRQOL); h) emergency department visits and readmissions within 30 days; i) 30-day all cause mortality.

This intervention is a novel electronic tool that combines the newly refined version of the STOPP/START criteria, the 2015 Beers criteria, and applicable recommendations from Choosing Wisely Canada to guide and facilitate deprescribing. To the investigator's knowledge, this is the first electronic intervention to be implemented among heterogeneous at-risk, frail elderly, hospitalized patients. This trial will test a tool that could lead to a transformative change in medication appropriateness.

2.2 Methodological Approach Study design: The study design is a before-after intervention study. All hospitals will participate in six weeks of baseline measurement followed by six weeks of intervention. The primary outcome is the proportion of patients who have 1 or more PIM discontinued at discharge.

Study Setting: This study will take place on the clinical teaching units where patients are cared for by teams composed of medical students, junior and senior residents in internal medicine and other specialties, and attending physicians. The CTUs where this study will take place admits complex elderly patients with multiple comorbidities where the majority are non-surgical frail elderly and are most likely to benefit from a medication stewardship intervention. Most medical trainees will spend weeks to months on these units each year, yielding high exposure to the intervention. The investigators have shown that medication stewardship projects can be successful on the CTU and that changes taking place on the CTU can continue after discharge.

Planned trial interventions: Within 72 hours of admission (excluding weekends and holidays) to both control and intervention units, each patient's documented medical history and medication list will be manually extracted by a trained research assistant.

Intervention: Participants in the intervention phase will be electronically screened using MedSafer which will cross-reference this information to automatically identify and prioritize "deprescribing opportunities" for further expert evaluation by the CTU team.

In this intervention, PIMs that are electronically identified will be communicated to the treating medical team through MedSafer generated evidence-based recommendations including how to safely deprescribe or taper medications as necessary. In the case of multiple recommendations, they will be limited and prioritized in order to avoid overwhelming the treating team. The doctors caring for the hospitalized patient will receive the information. At that point, based on expert medical judgement, in collaboration with the patient/caregiver and other relevant physicians (i.e. consultants and the community physician), a decision will be made to deprescribe if appropriate.

Control: all participants will receive medication reconciliation by the unit pharmacist, a review of medical history and medication use, and identical follow up, but no prioritized list will be generated for the treating team to systematically trigger deprescribing.

The results of this catalyst study will ideally establish the safety and efficacy of MedSafer to electronically facilitate the inpatient application of criteria based on STOPP/START, Beers and Choosing Wisely recommendations to identity and prioritize PIMs for deprescribing. Based on a positive outcome of this study, the investigators would plan to implement the tool in a larger study at multiple hospitals across Canada to determine its effectiveness at reducing ADEs among a similar frail and elderly population.

Inclusion Criteria:

Participants will represent a consecutive series of persons.

• Eligible participants include those who are:

  • aged 65 years or older;
  • taking 5 or more medications; and
  • admitted to the inpatient medical ward at any of the study sites.

• Patients who are cognitively impaired or otherwise unable to provide consent will still be included as this sub-population of patients may be at greatest risk of ADEs because of their communication problems. Tri-council ethics guidelines (Canada) deem that this sub-population should not be excluded from interventions that potentially provide direct benefit to the participant due to the inability to provide informed consent (Article 4.5, Tri-Council Policy Statement 2008). For these patients, the family/proxy will authorize study participation pending the return of patient capacity to consent.

Patients who are ultimately discharged from non-study units during their hospitalization will be excluded unless that unit is a transitional care, rehabilitation, or post-acute care unit used to bridge the gap between acute medical hospitalization and community services.

Exclusion Criteria:

• patients expected to be transferred to another acute care hospital, a non-medical unit, or a palliative care unit;
• patients who normally live outside of the province of hospitalization;
• patients not enrolled in (or eligible for) the provincial drug plans;
• patients previously enrolled in the study;
• patients or proxy not able to speak English or French;
• no means of contacting patient or proxy post-discharge;
• patients expected to die within 30 days and whose level of care has been declared to be palliative will be maintained in the study but will be excluded from select secondary outcomes.