We assigned 112 eligible HIV patients to the intervention, of which 110 agreed to participate. Of the 110 enrolled patients, 104 had at least one recorded EPC clinical encounter with their assigned clinician. Of the 6 patients missing an EPC encounter, 5 either became lost-to-follow-up or refused to follow up with the EPC intervention, and one died before interacting with their clinician. The average number of visits per patient was 4 (range 1–9).
All the four clinicians were present to offer the EPC intervention to their assigned patients for the entire 6-month period. There was an equal distribution of patients among the four providers (called P, Q, R and S to maintain confidentiality) with two having 27 patients and the rest 28 patients (Table 2). During the 6-month follow-up, there were 511 clinical encounters, for an average of 128 encounters per clinician (range 115–140).
Integration of research and clinic staff
Prior to the onset of the intervention, the research team held three meetings with the clinic management team to gain their buy in. The management team appreciated that the intervention was in line with their strategic plan of enhancing patient engagement and adherence, and was highly supportive. An office space was quickly created for the intervention coordination, and the research team was invited to attend all biweekly meetings held at the facility. During these meetings, various aspects of clinical care were discussed. In addition, the research team, together with the clinic staff, shared their experiences including the progress, successes and setbacks. The sense of shared goals allowed the research and management teams to collaborate and problem-solve in ways that did not compromise the fidelity of the intervention.
Training of health providers
Clinicians identified the motivational interviewing training as the most useful aspect of the training. They reported that motivational interviewing provided practical solutions on how to deal with different groups of patients presenting at the clinic. It also helped them understand how to de-escalate when there was tension and how to manage their own stress levels during a clinical encounter. Common questions from clinicians during the training focused on age dynamics, uncooperative patients, stress management, and health system challenges (Table 3).
In addition, we noted that the involvement of study peers in the trainings was valuable since they were able to share their past experiences (positive and negative) with clinicians at the facility. They articulated what patients expect of their clinicians which facilitated fruitful discussions during the sessions.
Continuity of clinician-patient relationship
During the initial random assignment of patients to clinicians, patients frequently need to have the concept of randomization explained. Patients often had a preference for certain clinicians over others based on past experiences. We found that it was helpful to assure patients that all clinicians including all the providers at the health facility had received training on appropriate patient engagement techniques.
At the end of the 6-month follow-up period, the majority (66 %) of patients’ experienced full clinician-patient continuity, by maintaining the same clinician assigned to them for the entire 6 month study period. Temporary switching of clinicians, whereby a clinician stepped in to attend to patient(s) assigned to another clinician, happened to 38 patients. These occasions were inevitable and occurred when clinicians were required to attend impromptu facility-related meetings or address other emerging personal issues such as death of a family member, illness or annual leave. The majority (32/38 = 84 %) of switching happened once except for 6 cases where it happened more than once among three clinicians. In only 2 instances did the switching happen due to patient-related reasons such as them not adhering to the clinic appointment day and/or time. Upon return of the clinician who was away, a briefing was provided by the covering clinician and on subsequent clinic visits the clinician would continue to see their assigned patients. There were, however, occasions where the patients experienced conflict with their assigned clinician. In the event that the conflict could not be resolved, a complete switch of clinician was necessitated. This happened only twice during the study period.
Treatment dialogue
Clinicians completed treatment dialogues for each of their assigned patients during each clinic visit. Only 7 % (36) of the treatment dialogues were incomplete. Among all four clinicians, the average time spent with a patient during an encounter was 16 min. We however noted time variation among the clinicians with the highest average time being 22 min and the lowest 13 min (Table 2). In addition, the lowest time a clinician recorded as having spent with a patient was 2 min and the highest was 65 min (not shown).
In terms of the treatment dialogue content, clinicians found it difficult to complete the treatment dialogues in Swahili even though they spoke to patients in Swahili. Furthermore it was sometimes difficult to read content of the completed treatment dialogues since most of the clinicians writing was illegible. This problem of illegibility continued despite encouraging clinicians to write more clearly in Swahili. As a result, we observed that patients did not make use of these summaries and heavily relied on their memory to recall what was discussed during their clinical encounter. The treatment dialogues were mainly utilized by the study peers who assessed the level of patients’ understanding of the clinical encounter at the end of every session.
We also noted that clinicians focused their discussion on medication adherence, mainly on the correct timing of ART. Clinicians recommended that patients identify reminders that would prompt them to take their ART medications at the appropriate time. The treatment dialogue did not have content that addressed the patients’ financial, social and psychological aspects that are critical to patients’ adherence. On the contrary, we observed that the study peers addressed more issues related to the patients’ social and psychological health, in addition to promoting medication adherence. In fact, peers sometimes referred patients to other social support services within the facility.
Scheduling of clinic visits
Out of the 563 scheduled appointments, we reported a 91 % adherence to the clinic appointment date by patients. Clinicians adopted the use of Google calendar to schedule clinic appointments for their assigned patients. During the first two months, adapting to Google calendar was difficult for clinicians since they were not accustomed to scheduling appointments while considering the potential patient load on a particular day or any future events (e.g., meetings or annual leave). At first, clinicians scheduled a number of patients on the same day without giving ample time between patients. This meant that they got overwhelmed with having to attend to their scheduled study patients as well as their regular patients. Over time, clinicians learned how to spread out their patient appointment dates and times, easing the burden of patients they had to see on one day. They also considered any upcoming events and did not book their patients on those days.
We however noted that clinicians did not initiate their scheduled clinic appointment on time at 73 % of clinic visits. The delays reported were on average 1 h 27 min. (Table 3). The reasons for delays in initiating clinic appointments on time were mainly due to the health system factors including: electronic medical record system malfunction, impromptu meetings that clinicians had to attend, delays in patient care procedure. On the other hand, the majority of patients kept their appointment times, with the majority coming to the clinic early. In only 22 % of the cases was the delay linked to patients being late for their appointment.