Page 159 - SAHCS HIVMed Journal Vol 20 No 1 2019

Page 4 of 14 Original Research

capacity (35 patients per club) in the same year – seven clubs considered censored if they were transferred out to a different
were identified. We purposefully decided to select two clubs club or clinic, or died. For adherence to medication, when
to allow us to compare their retention in care and ability to a patient was not attending that club for whatever
enhance adherence to medication. These two clubs were reason, they were considered censored. The viral load of
randomly selected from the identified seven to conduct the the patients is used as a proxy indicator of adherence to
survival analysis using the fishbowl or lottery method – medication. Non-adherence was identified as any reading
without replacing. All the patients in the cohort of each of > 400 copies/cm and adherence was represented as LDL
30
3
the selected adherence clubs (35 per club) were included in (lower than detectable reading).
the cohort analysis.
We conducted four non-participant observations of the
31
Regarding the interview process, we included all the adherence club meetings, where we observed club sessions
operational staff working on the adherence club programme without interfering in any of the processes. These included
at the facility, which comprised an adherence club nurse, two sessions of exclusive medication collection and a blood
who heads the adherence club programme, and three lay sample collection plus medication collection. The goal of
counsellors. We applied a purposive sampling approach the non-participant observation was to obtain insights
to select six participants to be interviewed from the two into events and activities and the meanings that the club
adherence clubs sampled for the quantitative retrospective members attach to the sessions. We captured the dynamics of
arm of the study. Our goal was to obtain at least one male interactions of the group members with each other and with
and one female from each of the clubs. We also included care providers in our field notes. During each observation
two patients who were members of the adherence club, but session, we took detailed field notes.
had been asked to return to the standard care scheme at the
main facility because they failed to follow all the club rules. After the non-participant observations, we conducted realist
Table 1 elaborates on the characteristics of the participants interviews – a theory-driven approach to interviewing 32,33 –
who were interviewed. to uncover the causal relationship of aspects related to
the implementation of the adherence club intervention.
Data collection process The investigation looked at the relevant context, generative

We used the two sampled clubs as the focus of data collection mechanisms and emerging outcomes in relation to the
32
for the observation. The quantitative data were extracted patients (actors). Pawson advises that in applying the realist
from the adherence club registers at the clinic. In the Western interviewing approach, the researcher’s theory is the subject
Cape, information relating to retention in care is registered matter of the interview, and the subject is there to confirm or falsify
using the modalities outlined in Table 2. and, above all, to refine that theory. He also suggests that the
care providers are versed in issues around the context and
Concerning retention in care, the patients were considered outcome of the intervention, while the patients, being the
not retained in care if they did not attend a club session or principal actors in the intervention, can provide mechanism-
32
sent a ‘buddy’ and were sent back to the clinic. Patients were related attributes. The patient interviews were conducted
after the second non-participant observation.
TABLE 1: Characteristics of study participants for the qualitative interviews.
Stakeholder Number of participants Time on adherence club The quantitative data were captured using Microsoft Excel
Nurses 1 Nurse 1 – 2012 and prepared for analysis using the Statistical Package for
Counsellors 3 Counsellor 1 – 2010 Social Sciences (IBM SPSS) version 24. The field notes from
(club facilitators)
Counsellor 2 – 2010 the non-participant observations were also developed into
Counsellor 3 – 2014
Patients 4 Patient 1 (female) – 2014 transcripts. The audio-recorded interviews were transcribed
(club members) verbatim by a professional transcriber and prepared for
Patient 2 (male) – 2014
Patient 3 (female) – 2014 analysis. Atlas.ti version 7 was used to manage the field notes
Patient 4 (male) – 2014 and interview transcripts.
Patients (former 2 Ex-member 1 (female) – 2014
club members) Ex-member 2 (male) – 2014 Data analysis

TABLE 2: Modalities defining adherence club attendance. To identify and describe the outcome patterns of the adherence
Recorded outcome Outcome event club intervention regarding retention in care and adherence
DNA Did not attend club session or sent a buddy within 5 days to medication, we used the Kaplan–Meier method – the
after the club sessions probability of surviving in a given time while considering
BTC Back to Clinic – exiting the club for medical reasons and
34
reinstated in routine standard of care time in many small intervals. This method was suitable
TFOC Transferred out to a different club – patient is transferred because it allowed us to estimate the rate at which patients
out to another club in the same facility
TFO Transfer out – patient is leaving the facility completely and remained in care and the rate at which they maintained a viral
will attend a clinic elsewhere load lower than detectable (< 400 copies/mm³ of blood) at
RIP Rest in peace – patient has died 6-month intervals, covering a 24-month period.

http://www.sajhivmed.org.za 152 Open Access