PAUSE: Sick Day Medication Management Mobile App Study

Our previous research surveyed healthcare providers from Alberta on "factors affecting clinician's decision to provide sick day medication guidance to patients with diabetes and CKD to prevent adverse events." Our results identified 75% of primary health providers were aware of sick day medication guidance, but just 56% knew where to find guidelines and resources. An overwhelming majority of respondents (97%) were supportive of enrolling patients in a study evaluating alternative innovations for providing sick day medication guidance.

In a recent scoping review summarizing existing interventions, our research team found the majority of published SDMG documents were aimed towards healthcare providers, with few patient-targeted documents. These were mainly in the form of handouts, wallet-sized cards, webpages, or telephone support. There is limited primary research on the development, implementation, or evaluation of current SDMG interventions. Most were reported to be challenging to follow and identification of sick days or qualifying medication without error was low. This survey and review highlight the need to develop and to evaluate new solutions for providing SDMG to patients. Our previous work also found that seniors in Canada were receptive to the use of electronic means of communication and several patients have expressed interest in using electronic health (eHealth) tools for sick day self-management.

Participants receiving the intervention will receive access to the PAUSE App, a self-management tool for SDMG intended for patients to use during an acute illness. Users' Loblaw/Shoppers Drug Mart pharmacy records are electronically linked to the PAUSE App within the President's Choice (PC) Health app allowing for up-to-date recommendations based on current prescribed medications. The app asks users a series of questions regarding signs and symptoms that identify a qualifying sick day illness, and screens for 'red flags' that would require emergency, or healthcare provider or urgent care referral, and help patients identify which of their medications they should temporarily withhold or adjusted, tailored to a patients' current medication list. This aims to provide patients with interactive support for managing medication during a sick day event. As part of the intervention, patients will also receive a SDMG patient handout. The intervention addresses the previously identified challenges of identifying qualifying signs and symptoms that warrant SDMG and which medications qualify via an interactive and individualized electronic application designed to facilitate provision of SDMG. The usual care group will receive a SDMG patient handout which outlines SDMG and addresses which medications qualify for SDMG.

Based on preliminary data, the investigators assume an absolute difference of 30% (50% with the PAUSE app vs. 20% without the PAUSE app) in the proportion of participants who complete a simulated sick day without error. Using a two-sided alpha of 0.05, 80% power, and an interclass correlation coefficient of 0.1 between pharmacy clusters, a sample size of 280 participants will be required. To account for a 10% loss to follow-up, the investigators will aim to recruit a total of 320 participants in the trial. The investigators plan to recruit 16 pharmacies that will recruit 20 participants each.

Data Analysis Participant baseline data, including sociodemographics, comorbidities, and active prescriptions will be analyzed using descriptive statistics. Feasibility and fidelity outcomes will be reported using descriptive statistics with numbers and percentages. Comparisons of outcomes between groups (e.g., PAUSE App vs. usual care) will be reported using unadjusted and adjusted generalized estimating equations to determine mean differences and risk differences between groups. Descriptive statistics will be used as appropriate to evaluate group differences following the follow-up period. Associations between key variables and study outcomes will be analyzed using appropriate univariate, multivariate, and mixed model analyses. Exploratory analyses of Google Analytics data will be performed to report user behaviour insights. Analyses of routinely collected health data over a 5-year extended follow-up period will be used to determine the effect, if any, of the intervention on health outcomes.

The simulated sick day evaluations will be scored and analyzed according to predefined scorecards based on scenarios used by Doerfler et al. measuring correct usage of SDMG during acute illness. Log-binomial regression models will be used to directly estimate the risk ratios (RRs) and 95% confidence intervals for the outcome of error free completion of the simulated sick day, as well as for correct completion of each of the 3 individual components of the sick day simulation. Random effects will be used to account for clustering by pharmacies. Unadjusted and adjusted models will be fit, including fixed effects for individual participant characteristics including age, sex, demographics, diabetes, other comorbidities, number of qualifying medications and any other significant confounding variables from univariate analyses. Additionally, data collected from participants on the usefulness of the PAUSE App and/or SDMG patient handout in managing a simulated sick day and overall acceptability of the interventions will be used to further assess the fidelity of the intervention. All statistical analysis will be completed in R.

Selected participants (patients and pharmacists) will be invited to be interviewed following their simulated sick day scenario evaluation based on the purposive sampling strategy. One-on-one semi-structured interviews will be conducted with participants and pharmacists ranging from 30-60 minutes in duration. Interview questions and analysis will be iterative throughout the study to allow for emerging or irregular themes to be examined in later interviews. Qualitative interviews will be audio-recorded, transcribed verbatim and examined using multiple phases of inductive thematic analysis. Collected field notes and transcriptions from interviews will be analyzed using NVIVO qualitative analysis software. Analysis of data will begin immediately following the conclusion of the first participant interview. Data will be coded by two researchers independently and then codes will be compared after the first interview to draft the coding manual for subsequent interviews.