Provider characteristics are presented in Table 2. Providers were primarily female (87%), and from diverse disciplines a majority of whom (78.3%) had six or more years of experience in HIV care.
Inner Setting for Adherence Support Approaches
Providers weighed the effect of a particular approach on patient-provider interactions in and outside of the clinic visit and the possible added burden to providers or clinic staff, including having to act on information once it was known. Providers also thought that added burdens could extend to patients; for example, patients might need to store the signaling pills in a new location as the signal emitted might attract unwanted attention for individuals who have not disclosed their status. Patients might also feel that they are expected to respond to text messages. Several providers described that having a low number of patients who are pregnant or postpartum, compared to the larger clinic volume, could serve as a facilitator to the integration of a new technology for this population, by reducing the overall time burden required to respond to output. Others remarked that the data provided by these approaches could easily be integrated on an existing electronic health record (EHR) and to other EHR-based initiatives they had adopted to assess practice-wide adherence metrics. Providers across sites consistently described human effort and burden (time and tasks) as key factors they would weigh in when considering which adherence support approach to select.
Outer Setting for Adherence Support Approaches
Providers situated adherence among many challenges their patients navigate, including housing instability, food insecurity, and legal difficulties. “I have not found adherence itself to be the major problem, but more the steps before it. […]” (Physician). One payer with prior experience as a HIV physician described housing as the “fifth vital sign.” Discrimination and hardship based on race, immigration status, and socioeconomic status were cited as consistent contributors to nonadherence. Payers understood adherence as subject to rapid changes: “Somebody can be completely 100% adherent for six months and then things can happen in their lives that drop off”(Payer). Case managers observed that adherence declines when patients lack the basic security of stable shelter, and ability to pay bills and feed oneself and one’s family: “It’s usually not a medication access thing. It's, ‘Oh, I take the medication, and I need to eat with it, but I didn't have any food, so I missed the med because I didn't have any food’” (Case Manager). For providers, connection to care was inextricable from medication adherence. Actions taken to identify and ameliorate the upstream, outer context causes for non-adherence were understood by providers to be intrinsic to, rather than separate from, their clinical responsibilities. Providers tended to view “human connection” as integral to identifying such factors and to developing collaborative plans to address them. Providers consistently identified the postpartum period as particularly challenging for retention in care and ART adherence, citing comparatively fewer resources available than during pregnancy, as one physician put it, “there is loss of insurance, pregnancy [coverage],, and sometimes their source of payment for the medication changes.” Dominant external setting themes in provider explanations for declining adherence postpartum included increased financial, cognitive and physical demands in the setting of sometimes loosing access to financial supports available during pregnancy; direct competition between care for self and care for newborn; declining risk of avoiding perinatal transmission; and postpartum depression. Providers did not view technology-based approaches as a solution to addressing social determinants but as one of many tools that could be used to better support women’s ART adherence in the postpartum period.
Intervention Characteristics and Implementation Process
Providers explicitly prioritized approaches that they felt would enhance patients’ overall wellness and promote patient-provider trust. Text messaging was the most popular approach because it was familiar to providers, easily accessible to patients, and could enhance patient-provider relationships. The text message approach was also perceived as less invasive compared with other approaches. There were greater concerns about privacy and surveillance for the signaling pill, signaling pill bottle, video check with provider, and automated video check. Table 3 provides an overview of associations providers made between adherence support approaches and factors they deemed important. The most consistently cited facilitators were enhancing patient-provider relationship, predictable reminder mechanisms, and options for customization based on patient preference. Payers anticipated regulatory hurdles with unfamiliar approaches, particularly the signaling pill and signaling pill bottle. Below, we discuss barriers and facilitators to each approach in detail and strategies that can be used to integrate use of the approach in clinical care.
Providers saw text reminders as an opportunity to develop better and more frequent interactions with patients. They believed text message reminders and report could be helpful for most patients, as most have text-capable devices, check them frequently, and know how to text. Providers presumed younger patients would text more often and more seamlessly utilize the intervention. Many providers had successfully used text messages to contact their patients in the past.
Barriers included the ease of ignoring a text message or responding dishonestly which were seen as a diminishing return for investment. The frequency of text messages was also concern. One physician shared, “my general experience is that after two-to-four weeks, they began to ignore [text reminders]. You never want to get to that place that the patient is now avoiding the interaction.” Providers expressed concerns about HIV status disclosure through a text message, especially if a patient shared a phone. Several case managers described how even a discrete message could become “a potential outing for that patient” (Case Manager). Providers suggested that customizing reminder texts could resolve issues around disclosure, potentially facilitating use of the approach.
Video Check with Providers
Facilitators to this approach included a sense that video checks could improve the therapeutic alliance, as a “social approach” seen to enable “human connection”, a factor thought to mitigate outer context barriers providers emphasized. Providers framed the video check as best for patients who enjoy person-to-person contact. Providers suggested newly diagnosed patients, patients switching medications, and postpartum women for this approach. Overall, providers conceptualized video calls as a temporary tool for establishing or strengthening an adherence routine.
Barriers to this intervention included high demand on providers, requirement of video-calling capable devices, and the possibility of becoming cumbersome to patients: “It would take a lot of time and a lot of resources to make sure that patients have access to a smartphone and can Skype or FaceTime” (Case Manager). Providers feared a video call would make patients feel invaded: “With our population with the stigma and all of it, I don't think that'll work because [if] somebody is watching [a patient take a pill], the [patient] probably will feel violated” (Case Manager).
Automated Video Check
Providers were less enthusiastic about the automated video check, and few thought it would appeal to patients. Barriers included perceiving this option as a less personalized approach and one whose facial recognition technology risked patient discomfort. Generally, providers felt the automated call incurred the same effort of a live video call without the benefits of human connection with a provider.
“This method doesn't offer any of the support that the video potentially could. The video that we usually do is like a quick check in. "How are you feeling? Are you ready to take your pills today? Great, let's take it. Oh, you did such a good job." There's a positive reinforcement as opposed to a video just recording.” (Physician)
However, some thought it might help non-adherent patients who found personal contact burdensome. Providers viewed the automated video call as less resource-intensive for providers and less intrusive for patients and saw it as a possible method for stepping down from personal contacts.
Signaling Pill Bottle
Facilitators to this approach included an appreciation of the novelty of this visual reminder (e.g., remarking that a blinking pill bottle was an unheard of and exciting way to help patients remember whether they had taken their medications on a given day) and a sense that it would not disrupt routines for picking up and taking pills, even though it could change how patients store pills.
However, providers remarked that flashing pill bottles are not discreet for those trying to keep their status private who, for example, conceal their pills in a vitamin bottle. “Anything that would draw attention to the medication would be something they would want to avoid” (Physician). In addition to disclosure concerns, providers noted that the signaling pill bottle could be ineffective for patients who do not store their pills in the original prescription bottle or in a visible location (for example, keeping pills in a pill organizer or drawer). One physician worried the signaling bottle could communicate an “assumption that you don't trust the patient being able to take their meds without being monitored.” The most common concern among providers was that the data from the signaling pill bottle could be misleading if a patient opened the bottle but did not take their pill.
Many providers thought patients would dislike swallowing a sensor due to feelings of being watched or having their privacy invaded. Additionally, some providers worried that relying on the signaling pill rather than patient report to assess adherence could threaten the patient-provider relationship. “In some ways, it’s signaling a lack of trust to the patient” (Case Manager). Providers believed the signaling pill would be ideal for patients who were chronically non-adherent and dishonest in reporting missed pills, though they speculated those patients would not accept the approach. Facilitators included an appreciation of the accuracy of information from the signaling pill, even if few expressed willingness to use it.
Mixed reactions to detailed record keeping on medication adherence
Overall, providers differentiated supporting from verifying adherence, despite the potential for each approach to combine both functions. When asked whether they would like a detailed report of whether and when their patients take their medications, providers offered mixed responses. A plurality of providers (N=12/26) speculated that information would be “a nice tool” or could enable them to locate specific causes for missed doses. One provider felt positively about this option but added, “I’d definitely question how that information is collected and the validity and the [re]liability[…]” (RN). Others thought a detailed report would contribute little to their preexisting clinical practice. Overall, providers emphasized that having data does not lead directly to having the resources and capacity to address the problem(s) the data reveal.