1. Introduction
In Canada, the HIV epidemic is on the rise, with a 24.9% increase in reported cases across the country in 2022, amounting to 1833 new diagnoses [
1]. The national rate of new HIV diagnoses was 4.7 per 100,000 people, with Québec slightly exceeding this average at 4.9 per 100,000 [
2]. Montréal saw a particularly alarming spike, with a 120% increase in HIV diagnoses between 2021 and 2022, resulting in approximately 310 new cases [
3]. This surge is attributed in part to the resumption of HIV testing post-pandemic and the arrival of immigrants from regions where HIV is more prevalent [
4].
Managing HIV requires interprofessional efforts, and pharmacists play a significant role in the care of people with HIV (PWH) [
5,
6,
7]. They often serve as first, regular, and trusted points of contact for people seeking medical treatment and advice [
8,
9,
10,
11,
12], especially in community pharmacies [
13]. They support PWH, notably, by providing crucial assistance in the regular refill and management of their antiretroviral therapy (ART).
Generalist pharmacists in Canada can struggle to stay current with HIV treatments and make informed decisions [
14]. They can face challenges related to rapidly evolving drug landscapes, constrained time resources, and expanded care roles [
15]. These difficulties can be compounded by infrequent interactions with PWH and significantly impair their ability to acquire specialized knowledge and expertise in HIV care, which can, in turn, impact the quality of care they provide.
A range of digital tools are available to support pharmacists in medication management, patient education, and professional communication, including mobile health applications and online clinical references [
16,
17,
18]. These resources provide access to drug information, clinical guidelines, and patient care tips [
16,
17,
18]. Nevertheless, the abundance of online health information complicates the retrieval of specific, reliable data [
19], which can make it difficult for pharmacists to provide informed care and navigate treatment options effectively.
Chatbots that leverage artificial intelligence (AI) offer a promising solution to these challenges by facilitating the retrieval of medical information [
20,
21,
22,
23]. Chatbots, also known as “conversational agents”, are computer programs designed to mimic human conversation through various platforms, including websites, texting apps, and computer software [
22,
23,
24,
25,
26,
27,
28,
29,
30,
31,
32,
33]. As such, they could provide pharmacists with instant and streamlined access to the most current, validated information on HIV treatments, potentially outperforming web- or app-based tools in these functions. Chatbots could generate personalized advice, aligning with demands from healthcare providers for comprehensive information resources on medication, including, for example, drug–drug interactions and dietary supplements [
34,
35].
Given pharmacists’ challenges in HIV care delivery, our overarching goal is to adapt an existing AI-based chatbot and connect it with a constantly updated and verified source of information to help pharmacists serve PWH. Since 2020, a team of experts, including physicians, pharmacists, PWH, researchers, and engineers at the McGill University Health Centre (MUHC), developed an innovative AI chatbot called MARVIN for Minimal AntiretRoViral Interference [
36]. Running on Facebook Messenger, MARVIN was created as a retrieval-based independent component analysis (ICA) trained to chat with PWH on various aspects of self-management of HIV in both English and French [
36]. To configure MARVIN for pharmacists (MARVIN-Pharma), a formal assessment of local pharmacists’ needs is required. With this needs-assessment survey, our objectives are to describe the pharmacists’ knowledge, attitudes, involvement, and barriers in HIV care, as well as their perceptions of MARVIN-Pharma’s usability. Ultimately, this needs assessment aims to inform the adaptation of MARVIN-Pharma to ensure it effectively addresses the identified gaps and challenges in HIV care, enhancing pharmacists’ ability to deliver high-quality care and improving the overall usability and acceptance of the chatbot in clinical practice.
3. Results
We recruited 41 pharmacists, and all participants provided complete data.
3.1. Sociodemographic and Professional Characteristics
Table 1 presents descriptive sociodemographic information on respondents. Over two-thirds (68%) worked in the Montreal area. The average age of respondents was 36. In terms of ethnic backgrounds, approximately one half (49%) identified as French Canadian while the other half (51%) identified as East Asian, South Asian, Arab, or North African. The gender distribution showed that slightly more women than men were recruited. As to sexual orientation, most (85%) identified as heterosexual.
Professionally, respondents had an average of 10 years of experience. Over two-thirds (68%) worked in community pharmacies, with the remainder working in hospital settings. As to their patient groups, approximately half served women (54%) and LGBTQ+ community members (46%). Over a third provided care to people who use injection drugs (39%) and migrants from HIV-endemic countries (37%). Fewer attended to such groups as the elderly (12%), Indigenous people (12%), and sex workers (10%). Respondents’ patients were approximately 8% PWH, with an additional 12% deemed at high risk of acquiring HIV.
Considering technology usage, on average, respondents used Facebook Messenger regularly for over 10 years, indicating familiarity with the MARVIN chatbot’s current communication platform. Moreover, pharmacists reported, on average, “sometimes” using health or work-related applications on their mobile devices.
3.2. Knowledge
Perceived Knowledge
Our results indicate a moderate level of perceived knowledge on HIV and its treatment among respondents (
Table 2). The average score across the items was 3.74 on 6, almost equivalent to being “somewhat prepared” to advise PWH in the areas evaluated. The highest average item scores (over 4.0) were for preparedness to counsel PWH on HIV transmission and HIV prevention. Meanwhile, the lowest individual item scores (with means between 3.5 and 3.6) concerned adverse drug effects and interactions; HIV treatments; and HIV symptoms and disease progression. These values corresponded to being “somewhat unprepared” to “somewhat prepared” to address these topics.
3.3. Attitudes
For the most part, pharmacists indicated positive attitudes towards PWH and HIV care in the context of their professional work, with a global mean of 4.02 on 6, where a higher mean indicates more favorable attitudes (
Table 3). They strongly disagreed with stigmatizing statements (e.g., PWH deserve to have HIV or HIV treatments are not worthwhile) and agreed or strongly agreed with inclusive statements (e.g., I am willing to work with an HIV-positive colleague or providing information about HIV is a professional responsibility). However, pharmacists’ mean attitude toward mandatory continuing education on HIV/AIDS was tepid (M = 3.66; between “Somewhat Disagree” and “Somewhat Agree”).
3.4. Practice
3.4.1. Involvement
On average, respondents indicated being “little” involved in HIV care across the services considered (M = 2.08) (
Table 4) and were most involved in ART adherence counseling, refilling ART, treatment and monitoring of ART, and referring patients for postexposure prophylaxis (PEP), with mean involvement ratings ranging from 2.3 to 2.8 on 5 (which refer to being “little” to “moderately” involved in these areas).
3.4.2. Competence
Pharmacists’ global mean score on items related to their perceived competence in HIV care delivery was 3.78/6 (
Table 5). On average, they agreed they were sometimes hesitant to dispense ART to PWH and somewhat agreed they were unable to spend enough time counseling PWH. They also somewhat disagreed that they were able to confidently manage the complexities faced by PWH and that they were satisfied with the amount of care they provided them.
3.4.3. Barriers
Table 6 shows perceived barriers to providing HIV care. Most mentioned barriers were lack of time, of staff resources, of clinical tools, and of information or training on HIV/AIDS services. On average, they at least “somewhat” agreed that they experienced these barriers.
3.5. Usability
Table 7 presents the results on acceptability, compatibility, and self-efficacy. On average, pharmacists were “undecided” about the acceptability of the proposed MARVIN-pharma chatbot (M = 4.34/7) as well as about its compatibility with their work (M = 4.13). In addition, pharmacists showed tempered agreement with their self-efficacy to use Internet health services (M = 5.65).
4. Discussion
Drawing on a conceptual framework and existing instruments, this study aimed to better understand the needs of pharmacists in HIV care and their perceptions of the MARVIN-Pharma chatbot (in development) as a tool to reduce their challenges in HIV care. Specifically, it sought to characterize their knowledge, attitudes, and practices tied to HIV care, as well as the chatbot’s usability, in terms of its acceptability, its compatibility, and pharmacist self-efficacy.
Our findings indicate that pharmacists perceive their HIV knowledge as moderate and face significant barriers, such as lack of time and resources, which hinder their involvement in HIV care. Despite these challenges, there is a generally favorable attitude towards providing HIV care, suggesting a readiness to engage more deeply if provided with the right support tools. However, pharmacists were generally undecided about the acceptability and compatibility of MARVIN-Pharma. Nevertheless, this study highlights the potential of the chatbot to address the identified gaps and barriers, but also the importance of involving pharmacists in its development to ensure its future uptake and usability.
4.1. Knowledge
Our results showed that pharmacists perceived their knowledge to be moderate across the various aspects of HIV care evaluated. As found in past research [
50,
52], pharmacists felt most prepared to advise patients about HIV transmission and prevention. However, they did not feel fully prepared to advise on HIV treatments [
50,
52,
59]; symptoms and progression of HIV; and adverse drug effects and interactions with ART. Pharmacists’ moderate HIV knowledge, as observed in our study, given their generally limited specialization in this area, reflects the broader trend towards increased HIV knowledge among healthcare professionals in general [
50,
52,
59,
60,
61,
62,
63,
64,
65].
Efficiently bridging pharmacists’ knowledge gaps is essential, as their guidance may significantly enhance PWH’s understanding of their treatment [
66], potentially boosting ART adherence and lowering HIV transmission rates [
67]. MARVIN-Pharma could therefore prioritize content that helps pharmacists provide informed HIV care, such as HIV treatment guidelines and ART regimens, drug interactions, management of medication side effects, and considerations during pregnancy. The chatbot’s connection to a regularly updated ART regimen database like the Canadian HIV/AIDS Pharmacists Network (CHAP) or the Canadian AIDS Treatment Information Exchange (CATIE) seems essential to address the challenge of keeping pace with current information and to streamline access to it. Indeed, Ada Health, Buoy Health, Woebot, and Your.MD are real-world examples of chatbots linked to constantly updated medical databases that demonstrate the feasibility and benefits of such an approach in healthcare [
68].
Additionally, MARVIN-Pharma’s eventual ability to deliver just-in-time training and educational content could further enhance pharmacists’ confidence and competence, addressing knowledge gaps as they arise during patient interactions. This practical application ensures the chatbot not only fills existing knowledge gaps but also supports continuous learning, directly linking back to the study’s objective of improving pharmacists’ knowledge in HIV care.
4.2. Attitudes
Pharmacists’ attitudes towards PWH and aspects of HIV care were favorable and non-stigmatizing, consistent with previous studies among pharmacists [
50,
51,
62,
64]. On average, the participants had been pharmacists for 10 years, which may partially explain these results. Healthcare personnel with greater work experience, especially related to PWH care, are found to have fewer stigmatizing attitudes towards PWH [
69].
Our results show that some pharmacists were reluctant towards mandatory continuing education on HIV. This finding increases the relevance of a tool like MARVIN-Pharma for practice [
70]. Indeed, AI-driven educational tools have been shown to improve healthcare outcomes by keeping practitioners updated with the latest guidelines and best practices in a user-friendly manner [
71,
72]. The chatbot, however, is not intended to teach pharmacists the latest guidelines directly. Rather, it will be designed to answer specific questions and provide responses based on these guidelines. Thus, pharmacists will be able to easily retrieve up-to-date information and apply it to real-life situations, enhancing their ability to make informed decisions quickly and effectively during patient care.
This focus on integrating the chatbot seamlessly into pharmacists’ workflows, rather than adding additional educational burdens, directly supports the objective of understanding and addressing pharmacists’ attitudes towards HIV care and ensuring that MARVIN-Pharma aligns with their professional practices.
4.3. Involvement
Our findings indicate that pharmacists, overall, were only a “little” involved in HIV care. Their most common activities concerned ART counseling, monitoring, and refilling, as well as PEP referrals. Notwithstanding, pharmacists’ delivery of HIV care remains crucial for extending services beyond specialized settings; it is both feasible and well-received and can help reach underserved populations [
73], to enhance antiretroviral adherence, viral suppression, and retention in care [
74]. MARVIN-Pharma could assist pharmacists in their important HIV care roles and help compensate for any limited involvement in this area. For example, an adherence counseling module could enable them to perform detailed, interactive assessments of patient adherence, providing personalized advice and interventions based on patients’ responses. Studies have shown that pharmacist-led adherence interventions significantly improve medication adherence and clinical outcomes in HIV patients [
75,
76,
77,
78,
79]. Additionally, a PEP referral feature could streamline and simplify the referral process, complete with checklists and easily submittable forms to health services. This approach is supported by research indicating that pharmacist-initiated post-exposure prophylaxis (PEP) services increase timely access to PEP and improve adherence to PEP regimens [
80,
81,
82].
These practical features demonstrate how MARVIN-Pharma could directly enhance pharmacists’ involvement in HIV care, fulfilling the study’s objective of assessing and improving their engagement in this area.
4.4. Competence
Our results indicate that pharmacists experience limited competence in HIV care delivery when managing complex issues faced by patients living with HIV, dispensing ART, and trying to provide sufficient care, underscoring the need for enhanced support and training.
The competence challenges identified align with those highlighted by past research [
52]. Given their findings, some investigators [
83] have emphasized the importance of continuous professional development to address knowledge gaps and improve clinical confidence among pharmacists, as well as the need for comprehensive training programs [
84] to enhance their ability to manage complex patient needs and improve overall care delivery.
In addition to real-time support for pharmacists’ questions, MARVIN-Pharma could significantly enhance pharmacists’ competence in HIV care by offering tailored interactive learning modules, such as on ART management, drug interactions, and counseling techniques. Connected to a database like CHAP or CATIE, it could incorporate real-time updates on treatment guidelines and emerging data, thus reducing hesitation when dispensing ART. Additionally, it could facilitate just-in-time training, offering educational content and clinical support as pharmacists encounter specific challenges, thereby building confidence and improving the overall quality of HIV care [
70].
By addressing these competence challenges through tailored features, MARVIN-Pharma aligns with the study’s ultimate objective of informing the development of tools to enhance pharmacists’ ability to deliver high-quality HIV care, ensuring the practical application of the findings.
4.5. Barriers
Mitigating barriers to integrating HIV services into pharmacy practice is crucial for enhancing care. Our study suggests that the pharmacists’ most prominent challenges were limited time, clinical tools, and staff resources, which are known to impact the provision of comprehensive HIV care [
51,
73,
85,
86,
87]. Importantly, pharmacists also somewhat agreed that they lacked information or training on HIV/AIDS services, which can hinder effective patient counseling and care [
14,
88,
89]. Lesser barriers, including a lack of collaboration with other healthcare professionals, inadequate pharmacy design, patient disinterest in preventive services, and the absence of financial incentives for HIV service provision, have also been reported in prior literature [
74,
85,
90,
91].
MARVIN-Pharma could be designed to help address the key barriers identified. By incorporating such features as educational modules [
92,
93,
94,
95,
96,
97], time-management tools [
94,
96,
98,
99,
100,
101], resource access [
94,
96,
100,
101,
102], collaboration facilitation [
99,
102,
103,
104], patient engagement strategies [
95,
97,
100,
105], and incentive guidance [
95,
96,
100,
101], MARVIN-Pharma could significantly enhance HIV care delivery in pharmacy practice. The literature supports the potential of chatbots to incorporate these features effectively, demonstrating their capacity to overcome various barriers in healthcare.
Addressing these barriers directly relates to the objective of identifying challenges in HIV care and informing the design of MARVIN-Pharma to overcome them, ensuring the chatbot is practically useful in real-world settings.
4.6. Usability
Concerning the acceptability and compatibility of MARVIN-Pharma, pharmacists demonstrated a relatively neutral stance. This may be attributed in part to the hypothetical nature of the chatbot at this stage. Nevertheless, these findings emphasize that, to foster adoption, pharmacists must be engaged in the chatbot’s design to ensure it meets their specific needs and resonates with their professional responsibilities and workflow requirements. Stakeholder involvement and co-design are crucial for the success of mHealth tools. Studies have shown that involving end-users in the design process leads to higher acceptance and more effective implementation [
106,
107]. By collaborating with pharmacists during the development of MARVIN-Pharma, the tool can be tailored to their circumstances, enhancing its usability and integration into daily practice.
Encouragingly, our study also points to a level of comfort with digital technologies among Quebec pharmacists, aligned with broader research showing openness to digital tools among pharmacists [
16,
17,
18,
108,
109,
110,
111,
112,
113,
114]. The sample’s self-efficacy with using Internet health services, their average 10 years of using Facebook Messenger, and their occasional use of health or work-related applications on their mobile devices indicate a potential facility to take up a tool like MARVIN-Pharma.
Emphasizing the importance of user involvement and the need for further development based on pharmacists’ feedback highlights how the findings align with the objective of assessing MARVIN-Pharma’s usability to ensure its successful implementation in practice.
5. Limitations of This Study
This study has several limitations. Despite employing multiple recruitment strategies, its sample size was small (n = 41 pharmacists). We were thus unable to examine the effects of potential confounders (e.g., age, gender, and professional experience level). While snowball sampling may have helped reach a broader range of participants, self-selection and sampling bias may still affect the generalizability of the obtained results. For instance, the predominance of pharmacists from Montreal may not accurately represent the broader population of pharmacists in the province, and the sample’s average of 10 years of pharmacy experience and slight overrepresentation of females suggest a relatively experienced and gender-skewed group. We did, however, largely attain our sample size target to help ensure adequate statistical precision.
While this study relied on participant self-report, potentially introducing risks of response, recall, social desirability, and confirmation bias, the consistency of results obtained with respect to previous literature is encouraging in this regard.
This study’s quantitative design and limited scope mean the nuances and depth of pharmacists’ experiences and needs were not explored. A more comprehensive perspective will be achieved with the analysis of qualitative interview data collected as part of the larger mixed methods research project in which this study is embedded.