Assessing the acceptability and usability of MedSafer, a patient-centered electronic deprescribing tool

Abstract

Background

Deprescribing is the clinically supervised process of stopping or reducing medications that are no longer beneficial. MedSafer is an electronic decision support tool that guides healthcare providers (HCPs) through the deprescribing process. We recently developed a novel patient-facing version of the software, allowing patients and caregivers to generate a personalized deprescribing report to bring to their prescriber.

Objective

The study aimed to evaluate the usability and acceptability of MedSafer among older adults, caregivers, and community HCPs (physicians, nurse practitioners and pharmacists).

Method

A mixed-methods feasibility study was conducted with a convenience sample of 100 older adults/caregivers, and 25 healthcare practitioners. Participants were invited to test MedSafer and answer telephone or electronic surveys via RedCap. The Extended Technology Acceptance Model (TAM2) and System Usability Scale (SUS) were used for evaluation. A semi-structured interview was also conducted with a subset of participants (5 per group) who were selected on a volunteer basis, and thematic analysis was used following Braun & Clarke’s approach.

Results

Healthcare providers scored more favorably on TAM2 constructs such as perceived usefulness (PU) (median: 4.25 for HCPs; 3.75 for caregivers; 3.00 for patients), and SUS compared to patients and caregivers (mean: 79.50 for HCPs; 52.95 for caregivers; 55.75 for patients). Thematic analysis revealed that participants recognized MedSafer as an empowering tool but noted the need for some usability improvements.

Conclusion

MedSafer is a promising tool to support deprescribing conversations. Enhancing usability, accessibility, and patient education may improve adoption.

Introduction

Polypharmacy, often defined as the simultaneous use of at least 5 medications, is common in older adults aged 65 years and older.^1,2 It can be accompanied by negative effects when the actual or future harms of taking multiple medications outweigh the actual or future benefits. Examples of harmful effects include adverse drug events (ADEs) such as falls, fractures, impaired cognition, loss of autonomy, hospitalization, and death.^3–5

Deprescribing is a promising solution to tackle polypharmacy. It is the clinically supervised process of tapering, reducing, stopping, or changing classes of medications that have been identified as potentially inappropriate.^1–3 However, deprescribing is challenging with multiple barriers identified at the patient level (eg, patients may fear stopping a medication that is treating a symptom, caregivers may perceive that stopping a medication equates to withdrawal of care), at the provider level (eg, how to deprescribe is not emphasized in the medical curriculum, the process is time-consuming, there is a lack of support and tools in place to facilitate the process), and at the system level (eg, it can be challenging to obtain an accurate medication list, electronic medical records lack interconnectivity).^6–8

MedSafer is an electronic decision support tool that helps overcome some of these barriers.⁹ The tool helps guide healthcare providers (HCPs) through the process of deprescribing by cross-referencing patient medical conditions and medications with guidelines for safer prescribing in older adults.^10–14 MedSafer was previously tested in a 5698-patient cluster randomized controlled trial and found to be safe and effective, having no demonstrable impact on ADEs while producing an absolute increase in deprescribing of 22% when used during an acute care setting.⁹ This version of the tool was designed for use by clinicians, with an interface that can be accessed by physicians, pharmacists, and nurse practitioners. In this study, data were entered by research assistants and deprescribing reports were generated for the treating team to review.

Another intervention previously shown to be effective for deprescribing is the so-called patient EMPOWER brochures.^15–18 EMPOWER brochures are handed directly to the consumer and provide an overview of why a medication may be harmful, how to approach a discussion with the prescriber about deprescribing, and provide non-pharmacologic options and a tapering regimen.¹⁹ For example, a brochure on reducing sleeping pills may include strategies for developing good sleep habits and managing daily stress, address common misconceptions about sleep, and provide a structured tapering program.²⁰ Brochures on proton pump inhibitors may help patients assess whether they should continue the medication by suggesting lifestyle and diet changes to prevent heartburn, including patient testimonials, and providing a tapering regimen.²¹ The approach has been tested in prospective controlled clinical trials, with patient-facing materials being shown to increase deprescribing by 10%-30% by nudging the patient to act.^15–18

Inspired by the dual success of the MedSafer app and the EMPOWER patient-facing deprescribing interventions, we recently developed a version of the MedSafer software with a new patient-facing portal. The portal allows older adults and/or their caregivers to generate personalized deprescribing reports to bring to their prescriber. Users can enter information on their medications and medical conditions into the secure web-based portal, which then cross-references this information with the existing MedSafer algorithms. The system generates a personalized report, which HCPs can access through a secure 2-step authentication token.

Successful uptake and scaling-up of the MedSafer patient portal hinges on its usability and acceptability by older adults and/or their caregivers, as well as the acceptability of accessing and actioning such a report by community prescribers.²² Ensuring MedSafer’s accessibility can empower patients to initiate deprescribing discussions with their HCPs, potentially leading to reductions in the use of potentially inappropriate medications.²³

The primary objective of this study was to evaluate the acceptability and usability of MedSafer among (1) members of the public (older adults and/or their caregivers) and (2) community HCPs (doctors, pharmacists, and nurse practitioners). Secondary objectives included identifying challenges and areas for improvement in the design and functionality of MedSafer that could be addressed for version 2.0 of the software.

Methods

Study design

We conducted an exploratory, sequential mixed-methods feasibility study by administering quantitative surveys and conducting qualitative interviews. We selected the Extended Technology Acceptance model (TAM2) and the System Usability Score (SUS) surveys,^24,25 because they are widely validated, reliable tools (Cronbach alpha ranging from 0.80 to 0.90) for assessing usability and acceptability of a digital health tool and have been tested in older adults.^26–31 Interview questions were developed to further explore the user perspective.

Study population

Patients were eligible for participation if they were aged 60 years or older, taking at least 5 prescribed medications, residing in Canada, able to provide informed consent, able to read and understand French or English, and had access to the internet. Individuals were excluded if they had a limited prognosis (expected death within 3 months as judged by the study team), residing in long-term care, or were currently hospitalized in an acute care setting. Eligible caregivers were self-identified people in the community who were involved in the care of an older adult who met the patient eligibility criteria. HCPs were required to work in Canada, speak and write in French or English, have an active license to practice (medicine, nurse practitioner, or pharmacy), be able to provide informed consent, and were providing care for older adults taking multiple medications (self-identified).

Recruitment

The study information was shared via a pamphlet to patients, caregivers, and HCPs through the Canadian Medication Appropriateness and Deprescribing Network’s (CADeN) contacts via email, the public awareness committee and the HCP committee communication channels and on social media.¹⁹ Additionally, caregivers and patients were recruited through the Contactivity Centre, a non-profit community center for active seniors aged at least 60 years old based in Westmount, Quebec. The study recruitment poster can be found in Section S1.

Testing MedSafer

Eligible patients and caregivers received an email inviting them to access MedSafer and to subsequently complete the surveys. They tested the app by inputting their own or their care-recipient’s medications (generic or brand name) and answered subsequent questions about relevant medical conditions (in a public friendly language), allowing them to generate a personalized deprescribing report. Medication entry only required the name, and not the dose, route or frequency. HCPs were invited to schedule an appointment with a study member (J.L.) to access a test patient report. During this session, the study member (J.L.) provided the HCP with a report code and a 2-factor authentication code required to access a report. Figure 1 illustrates how each group tested MedSafer. Section S2 provides a detailed illustration of the MedSafer interface and how each group used the tool.

Figure 1.

Flow diagram of MedSafer use by different groups.

Data collection methods

The surveys were administered to all groups through REDCap.^32,33 Baseline data was collected on the participant’s province of residence, age group, sex, gender, language spoken, language written, highest level of education, occupation, ethnicity. For HCPs, information on the number of years in practice and specialty was also collected.

The TAM2 survey assessed factors such as social influence, cognitive processes, and constructs like PU and ease of use to understand their impact on users’ intention to adopt MedSafer.^24,26 All the TAM2 constructs and their definitions can be found in Section S3. The survey was adapted by replacing the original technology references with “MedSafer.” The SUS survey was similarly adapted by replacing the generic “system” wording with “MedSafer.”^25,27 Both surveys were rated on a Likert scale (1-5) where 1 indicated strongly disagree and 5 indicated strongly agree. The full survey instruments and item wording are provided in Section S4.

Qualitative data were collected through semi-structured interviews with a subset of the participants who volunteered following completion of the surveys. The interview questions were developed based on the key constructs assessed in the TAM2 and SUS surveys as well as input from the research team. The guide was designed to explore participants’ experiences through open-ended, exploratory questions focusing on visual appeal, functionality and challenges with using MedSafer, and overall usefulness. The interview guide for all groups can be found in Section S5.

All data collection tools were piloted with an older adult, caregiver and HCP prior to administration.

Data analysis

Descriptive statistics was used to summarize the demographic characteristics of the study population. The TAM2 results were analyzed by averaging the responses for each construct and the median scores were examined for each group. For the computer anxiety (CA) construct, the positively worded items were reverse coded so that higher scores reflect greater anxiety. The SUS includes 10 items with 5 positively and 5 negatively worded statements. The negative items were reverse-scored and total SUS scores were calculated by summing the item scores and multiplying by 2.5 to yield a score of 100, with higher scores indicating better usability. Kruskal-Wallis tests and a post hoc Dunn’s tests were conducted to compare group differences for both surveys. The patient and caregiver TAM2 and SUS scores were pooled to together, and we performed a Spearman’s correlation test to examine the relationship between TAM2 constructs and behavioral intention to use (BIU) as well as the SUS scores. All statistical analyses were conducted using R, with a 2-tailed alpha of <0.05 considered significant.

The qualitative data were thematically analyzed using Braun and Clarke’s approach.^34,35 The 2 coders (J.L. and E.F.) initially coded the data independently using the MAXDAQ software before discussing similarities and differences to develop a unified codebook.³⁶ The refined codebook was then used to re-code the interview transcripts, from which themes and sub-themes were derived. Through this iterative process, a convergence of themes emerged, ensuring consistency and reliability in the analysis.

Results

Demographic information

We received survey responses from 72 patients, 28 caregivers, and 25 HCPs. Patients were primarily aged 65-69 years (n = 48, 66.7%), with the majority being female (n = 44, 61.1%) and retired (n = 32, 44.4%). Caregivers were predominantly male (n = 22, 78.6%) and in the age range of 70-74 years (n = 21, 75%). Spouse/partner relationships were the most common caregiving role (n = 12, 42.9%). HCPs were mostly pharmacists (n = 14, 56.0%), with 28% having over 21 years of practice (n = 7), and the majority were female (n = 23, 92%). All baseline characteristics can be found in Table 1 for patients and caregivers and Table 2 for HCPs.

Table 1.

Baseline patient and caregiver characteristics.

Characteristics	Patients (n = 72)	Caregivers (n = 28)
	N (%)	N (%)
Age group
18-24 years	–	5 (17.9)
25-34 years	–	1 (3.6)
45-54 years	–	1 (3.6)
60-64 years	2 (2.8)	0 (0.0)
65-69 years	48 (66.7)	0 (0.0)
70-74 years	10 (13.9)	21 (75.0)
75-84 years	10 (13.9)	0 (0.0)
85 and older	1 (1.4)	0 (0.0)
Sex
Male	28 (38.9)	22 (78.6)
Female	44 (61.1)	6 (21.4)
Ethnicity		505
Caucasian/Whiteian/white	36 (50.0)	7 (25.0)0)
African Canadian/Black	10 (13.9)	8 (28.6)
Hispanic/Latino	6 (8.3)	1 (3.6)
Asian/Pacific Islander	7 (9.7)	5 (17.9)
Native American	13 (18.1)	3 (10.7)
Other	2 (2.8)	4 (14.3)
Highest level of education
Less than high school	3 (4.2)	2 (7.1)
High school graduate	5 (6.9)	2 (7.1)
Some college or vocational training	21 (29.2)	1 (3.6)
Bachelor’s degree	17 (23.6)	11 (39.3)
Master’s degree	13 (18.1)	3 (10.7)
Doctoral degree	10 (13.9)	4 (14.3)
Prefer not to say	3 (4.2)	5 (17.9)
Employment status
Full-time	8 (11.1)	2 (7.1)
Part-time	23 (4.2)	2 (7.1)
Retired	32 (44.4)	23 (82.1)
Homemaker	24 (33.3)	1 (3.6)
Other	5 (6.9)	0 (0.0)
Nature of caregiver relationship
Spouse/partner	–	12 (42.9)
Child	–	2 (7.1)
Friend	–	10 (35.7)
Other family member	–	4 (14.3)

This table summarizes the demographic characteristics of patients and caregivers, including age, sex, ethnicity, highest level of education, and employment status.

Table 2.

Baseline HCP characteristics.

Characteristics	Patients (n = 72) N (%)
Age group
25-34 years	6 (24.0)
35-44 years	9 (36.0)
45-54 years	8 (32.0)
55-64 years	2 (2.8)
Sex
Male	2 (8.0)
Female	23 (92.0)
Ethnicity
Caucasian/whiteian/white	19 (76.0)
African Canadian/Black	0 (0.0)
Hispanic/Latino	0 (0.0)
Asian/Pacific Islander	5 (20.0)
Native American	0 (0.0)
Other	0 (0.0)
Years in practice
Less than 1 year	1 (4.0)
1-5 years	6 (24.0)
6-10 years	6 (24.0)
11-15 years	3 (12.0)
21 years or more	2 (8.0)
Doctoral degree	7 (28.0)
Profession
Medical Doctor	6 (24.0)
Nurse	4 (16.0)
Pharmacist	14 (56.0)
Other	1 (4.0)

This table summarizes the demographic characteristics of HCPs, including age, sex, ethnicity, years in practice, and profession.

Extended technology acceptance model survey

The TAM2 survey results revealed differences in user perceptions of MedSafer between HCPs and non-clinicians. Across most constructs, HCPs rated MedSafer more favorably, particularly in terms of PU (median= 4.25, IQR= 1.00) and perceived ease of use (PEU) (median= 4.25, IQR= 1.25). They also reported high confidence in their ability to use the tool, represented by computer self-efficacy (CSE) (median= 4.00, IQR= 0.67), and minimal CA (median= 1.50, IQR= 1.00). Patients and caregivers gave lower ratings overall, particularly for PU (median= 3.00 and 3.75, respectively) and PEU (median = 3.00 and 3.13, respectively). Table 3 shows a summary of the median scores of all groups for each construct. The Kruskal-Wallis test showed a significant difference across groups for all constructs (P < .01 for all comparisons). Post hoc Dunn’s test with Bonferroni correction revealed that HCPs differed significantly from both caregivers and older adults across multiple constructs. In contrast, caregivers and older adults exhibited more similar perspectives, with fewer significant differences (P-values mostly > .05) (Section S6).

Table 3.

Extended technology acceptance model scores.

Constructs	Patient (n = 72)		Caregiver (n = 28)		HCP (n = 25)
	Median	IQR	Median	IQR	Median	IQR
PU	3.00	1.00	3.75	1.13	4.25	1.00
PEU	3.00	1.50	3.13	1.56	4.25	1.25
CSE	3.33	1.00	3.17	1.25	4.00	0.67
CAa	2.75	1.31	2.08	3.00	1.50	1.00
PE	3.00	0.67	3.17	1.67	4.00	0.67
SN	3.00	1.00	3.67	1.42	2.67	0.67
Image (I)	3.00	1.33	3.00	1.08	3.33	1.00
Job Relevance (JR)	3.33	1.33	3.33	1.33	3.67	1.33
Output Quality (OQ)	3.00	1.00	3.33	1.42	4.00	0.67
Result Demonstrability (RD)	3.00	1.00	3.33	1.00	4.00	0.67
BIU	3.25	1.00	3.00	2.13	4.00	1.00

This table presents the scores for various constructs of the TAM2 as reported by patients (n = 72), caregivers (n = 28), and HCPs (n = 25). The scores represent the median and IQR for each construct. These constructs assess various aspects of technology acceptance and self-reported attitudes toward the MedSafer tool.

^aFor the CA construct, higher scores indicate higher levels of anxiety.

System usability scale survey

The SUS revealed differences in usability perceptions across the groups. HCPs rated usability highest, with a mean SUS of 79.50 (SD= 10.63; 95% CI: 74.97, 84.02), which is in the benchmark “acceptable” range. Patients and caregivers had lower mean SUS of 55.75 (SD= 14.55; 95% CI: 52.36, 59.15), and 52.95 (SD= 14.55; 95% CI: 47.49, 58.41), respectively, falling into the “low marginal” acceptability ranges. Figure 2 illustrates the scores in relation to the different benchmarks. The Kruskal-Wallis test showed a significant difference across groups for the SUS score [H(2) = 40.573, P < .001]. The post hoc pairwise comparisons of Dunn’s test with Bonferroni correction revealed a significant difference between HCPs and patients (P < .001) and HCPs and caregivers (P < .001). However, there was no significant difference observed between patients and caregivers (P > .05) (Section S6).

Figure 2.

System usability scale scores in patients (older adults), caregivers and HCPs in relation to different established benchmarks.

Correlation between TAM2 constructs and behavioral intention to use and SUS scores

The Spearman’s correlation revealed significant associations between TAM2 constructs and intention to use MedSafer for patients and caregivers. PU (ρ = 0.29, P < .001) showed the strongest association with BIU. Perceived ease of use (ρ = 0.21, P < .001) and subjective norm (SN) (ρ = 0.027, P < .01) also demonstrated positive associations with intention to use. Significant positive correlations were observed between TAM2 constructs and SUS scores. Notably, PEU (ρ =  0.43, P < .001) showed moderate correlations with SUS. Result demonstrability (ρ =  0.36, P < .001), PU (ρ =  0.29, P < .05), and perceived enjoyment (PE) (ρ =  0.26, P < .05) showed low associations. CA was negatively associated with SUS (ρ= −0.28, P < .05). Table 4 shows a summary of the Spearman’s correlation coefficients.

Table 4.

Spearman’s correlation for TAM2 constructs with SUS and BIU.

Construct	BIU (ρ, P-value)	SUS score (ρ, P-value)
PU	.29a	.29a
PEU	.21	.43b
CSE	.17	.16
Computer Anxiety (CA)	−.12	−.28a
PE	.21	.26a
SN	.027	.078
Image (I)	.069	−.095
Job Relevance (JR)	.19	.31a
Output Quality (OQ)	.23	.19
Result Demonstrability	.37c	.36a

Legend: This table presents the Spearman’s rank correlation coefficients (ρ) and P-values for the relationships between the TAM2 constructs with the SUS scores and BIU. Asterisks indicate the level of statistical significance.

^a P < .05;

^b P < .001;

^c P < .01.

Thematic analysis of MedSafer use: perspectives from patients, caregivers and healthcare providers

The thematic analysis of interview transcripts from 5 patients, 5 caregivers and 5 HCPs revealed 4 primary themes across groups: trust and understanding, usability and accessibility, communication and collaboration, PU and impact. Section S7 presents the themes and sub-themes, along with illustrative quotes for all groups.

Trust and understanding

Overall, participants viewed MedSafer as a credible starting point for initiating deprescribing discussions, though the level of trust varied. Patients and caregivers generally found the recommendations helpful for making more informed decisions about their medications, but some questioned the reliability of the recommendations (especially for complex conditions), and the evidence behind the reports. HCPs viewed the tool as a trustworthy resource to support deprescribing but emphasized that the accuracy depended on the quality of patient-entered medication lists and it may not fully capture nuances in complex cases. This underscored a shared view that MedSafer should complement, rather than replace, clinical judgement.

Usability and accessibility

Most participants found MedSafer was straightforward and easy to navigate, but some barriers remained. Group comparisons of MedSafer’s usability should be interpreted with caution as HCPs tested MedSafer in a different context and with different tasks than patients and caregivers. While many patients and caregivers found the platform intuitive, some struggled with the instructions and medication entry, suggesting a need for clearer guidance. HCPs found the tool user-friendly and felt it would integrate well into clinical workflows, especially in settings lacking geriatric pharmacists.

Patient-centeredness

MedSafer was widely seen as an empowering tool to allow patients and caregivers to initiate conversations about their medications. All groups appreciated having a tool that supports shared decision-making but still emphasized the importance of in-person conversations with prescribers.

Perceived usefulness and willingness to use

Participants recognized the benefits of using MedSafer but also identified areas for improvement. Many patients and caregivers would recommend MedSafer to others, appreciating its role in clarifying the safety of their medications and providing opportunities for deprescribing. However, they also suggested adding more educational context and resources to the reports to make them easier to understand and act on. HCPs unanimously described MedSafer as a valuable tool for guiding deprescribing conversations, offering a structured, evidence-based rationale that could help mitigate defensiveness from other prescribers and patients around medication changes.

Discussion

This study was the first to evaluate the acceptability and usability of a patient-initiated electronic deprescribing tool, integrating survey-based quantitative assessments (TAM2 and SUS) with qualitative interviews. By triangulating these data, we identified key factors shaping adoption and highlighted actionable opportunities to refine MedSafer’s design and implementation.

HCPs rated MedSafer highly for PU, PEU and result demonstrability, which aligned with their greater intention to use the tool. This is consistent with prior digital health research identifying PEU, clinician support, and workflow integration as facilitators of digital health technology adoption.^2,23,37–41 Qualitative findings confirmed this with HCPs viewing MedSafer as an evidence-based tool that could fit well into clinical workflows, particularly for facilitating complex deprescribing conversations, while patients reported usability challenges. This gap was reflected in the SUS scores: HCPs rated the tool’s usability well above the benchmark SUS score for a digital health tool,²² while patients and caregivers scored in the low marginal acceptability range.

These findings suggest that design refinements, such as clearer instructions and simplified data entry, may improve usability for non-clinician users. Importantly, the moderate correlation between PEU and SUS scores indicates that improving usability could meaningfully increase acceptance among patients and caregivers. The negative association between CA and SUS scores also highlights a barrier to adoption, especially for older patients who may be less confident with digital health tools. However, a recent study found that most older adults (81%, CI = 78.9%-83.8%) already use digital health technologies, such as patient portals, telehealth or mobile health apps.⁴¹ This suggests that adoption barriers may be less about digital literacy alone and more about the specific design and benefits of using the tool.

Trust emerged as a key theme in shaping MedSafer’s acceptance among participants. HCPs largely trusted MedSafer as a decision-support tool for deprescribing, though they remained cautious about patient-entered data and its applicability for more complex patients. In contrast, patients and caregivers were more cautious, with some comparing MedSafer’s output to generic online advice while others felt more confident in the recommendations because of their prior knowledge of deprescribing or their trust in the clinicians behind the tool. Importantly, patients and caregivers did not view MedSafer as a replacement for in-person conversations with their HCPs but rather as a tool to initiate the deprescribing conversation. This reinforces that digital tools should support and not replace trusted patient-provider relationships. Findings from the EMPOWER study also suggested that patient trust may increase when deprescribing tools are introduced directly by their HCP.¹⁶

The strengths of this study included its mixed-methods approach, allowing for a comprehensive evaluation of MedSafer’s usability and acceptability. The triangulation of both quantitative and qualitative data provided nuanced insights into user experiences across different groups, reducing the risk of isolated interpretations from either method alone. Additionally, the study captured perspectives from diverse stakeholders in deprescribing, capturing insights from patients, caregivers and HCPs, which provided a holistic assessment of MedSafer.

However, there are a few limitations to consider. Patients and caregivers were mostly recruited through the CADeN’s social media and subsequently, participants likely had higher levels of digital literacy and education and may have had some prior knowledge of deprescribing. As such, the study population may not fully represent older adults or caregivers nationally. Additionally, while patients and caregivers were asked to use MedSafer independently, as intended in real-world practice, HCPs tested MedSafer in a controlled environment with a test patient report and access facilitated by a study team member, rather than in-clinic with a patient bringing in their own report. Thus, the impact on HCP workflow from a patient’s ability to bring in the reports and share the report code was not fully assessed. Importantly, this study did not address the accuracy of patient-entered medication lists or how providers might respond if a patient’s self-reported list differs from the medication recorded in their chart. Additionally, because patients and caregivers performed different tasks compared to HCPs, direct statistical comparisons of usability and acceptability scores across all 3 groups should be interpreted with caution as the tasks and interfaces were not equivalent for all users. Overall, this study did not fully mimic the real-world setting, highlighting the need for future studies to test MedSafer in routine clinical workflows. Furthermore, this feasibility study focused on assessing basic usability and acceptability rather than broader implementation planning, which could be strengthened by applying a formal implementation framework in future research. Lastly, while TAM2 and SUS surveys provided valuable insights, self-reported data may be subject to social desirability bias, particularly for CA and intention to use, as respondents may want to present themselves as being more confident to use technology than they are.

Future directions and conclusion

To enhance MedSafer’s usability and acceptability, future work should focus on iterative design improvements by including patients and caregivers in the design process. Potential areas for refinement include enhancing usability and accessibility by improving navigation, mobile compatibility, and clarity in the instructions and strengthening patient education and trust by providing clearer explanations of how recommendations are generated and integrating more patient-centered educational materials. Additionally, scanning the pill bottles rather than manually entering the medication names may improve usability for patients and caregivers. Finally, expanding usability testing to more diverse patient populations and settings, including rural areas where access to specialized care is limited, could reveal additional benefits and challenges in real-world implementation.

Overall, the study findings highlight MedSafer’s potential as a valuable deprescribing tool but also demonstrate the need for design improvements to enhance patient and caregiver engagement. By addressing usability challenges and fostering greater trust in patient-directed deprescribing, MedSafer can play a key role in promoting safer medication use among older adults.

Author contributions

Jimin J. Lee (Conceptualization, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review & editing), Eva Filosa (Formal analysis), Tiphaine Pierson (Project administration), Ninh Khuong (Project administration), Camille Gagnon (Project administration), Jennie Herbin (Project administration), Claire Godard-Sebillote (Writing—review & editing), Soham Rej (Writing—review & editing), Robyn Tamblyn (Writing—review & editing), Todd Lee (Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Supervision, Writing—review & editing), and Emily McDonald (Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Supervision, Writing—review & editing)

Supplementary material

Supplementary material is available at JAMIA Open online.

Funding

This work was supported by Health Canada grant number 2324-HQ-000047 and by a Master’s Training Award from the Fonds de recherche du Québec – Santé. The views herein do not necessarily represent the views of Health Canada.

Conflicts of interest

E.G.M. and T.C.L. hold the copyright for the MedSafer software in conjunction with McGill University. S.R. receives a provincial clinician-scientist award from the Fonds de Recherche Quebec Sante, has received grant funding from Mitacs to fund a graduate student, and is a shareholder of Aifred Health. The remaining authors have no conflicts to report.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.