This research employed a qualitative descriptive research design.¹⁸ This design intends to provide a comprehensive description of participants’ experiences of a phenomenon, without being restricted to a specific interpretative or theoretical framework.¹⁹

In line with this design, the data collection involved short semi-structured interviews conducted at two points during the research: (1) before commencement of the CBRP and (2) at the end of the rehabilitation, whether this occurred because of completion or attrition.

Two semi-structured interview guides were developed. The pre-intervention guide included questions about the participants’ expectations and concerns with regard to the CBRP and perceptions of potential facilitators and impediments to their engagement with the programme. The post-intervention questionnaire was more comprehensive, including questions eliciting participants’ experiences and perceptions of the content of the program (in terms of the interface, potential cognitive and emotional effects, and levels of difficulty and motivation throughout the process).

Field notes were produced by the researchers during the entire intervention in order to keep track of significant comments made by the participants during the coaching calls, as well as general observations that were deemed potentially significant by the researchers.

Convenience and purposive sampling methods were used. Psychiatrists and nurses at a government mental health hospital and directors of three halfway houses (non-governmental organisations) were approached in order to inform them about the project. These professionals then extended invitations to potential participants and allowed the posting of recruitment ads. Those interested in participating contacted the researchers or requested (via communication with the health professionals) to be contacted telephonically. Meetings were arranged in order to inform them about the research project and to ascertain their suitability for the study. Times and venues for the cognitive assessment and installation of the program were arranged with those who met the inclusion criteria and signed the consent form.

Two diagnostic groups were formed: (1) participants living with HIV and on antiretroviral treatment and (2) participants previously diagnosed with paranoid schizophrenia in full remission. These groups were selected because HIV¹⁹ and schizophrenia²⁰ patients are known to display cognitive deficits associated with the pathology and pharmacological treatment specific to each diagnosis, with WM deficits common to both.^21,22 Although the neurobiological bases of the WM deficits in HIV and schizophrenia are different,^23,24 and variances in therapeutic outcomes of CBRP have been linked to the underlying diagnosis,⁷ the implementation of the programme was the same in both groups, thus allowing for the qualitative exploration of the participants’ experiences of the CBRP. Furthermore, participants of both groups were healthcare users of the same government hospital, and therefore it was assumed that the socio-economic status of the participants was relatively homogenous.

A different researcher was assigned responsibility for managing each group in terms of recruiting, interviewing, monitoring and coaching. Those with co-morbid medical and psychiatric diagnoses not related to the main diagnosis were excluded. Only participants with WM impairment (maximum performance in the low average range) and access to a computer and an Internet connection were included. Participants’ demographic characteristics are presented in Table 1.

The CBRP used was the CogMed Working Memory Training Program, a computerised and Internet-based rehabilitation tool. This training consists of 25 daily sessions of approximately 30–40 min each, to be completed in 5 weeks. Each participant carries out the training in their own home. The coach is responsible for providing information about the best training conditions.

The program offers multiple indices that can be used to ascertain the levels of functioning in WM, as well as progress measures including daily total span, daily training and pause times, and performance rate. The interface provides access to detailed quantitative information easily available for the researcher to track the performance on a daily basis. During the program, technical support was provided by the researchers when needed. The researchers monitored daily progress scores, which are automatically registered by the rehabilitation programme. The researchers contacted the participants telephonically once a week in order to provide feedback. If particular problems (such as low or absent engagement in activities) were identified by any of the program’s performance indices, the researcher called more often. These calls also aimed to identify and troubleshoot difficulties and to encourage and congratulate the participant for persistence and achievements.

Thematic analysis²⁵ was used to analyse the data. Themes were defined as pieces of information that were relevant in terms of the research question. Therefore, only references made by the participants (or present in the field notes) in relation to the barriers or obstacles that hindered their participation in the CBRP were extracted. These themes were captured and coded by using an inductive and semantic approach.²⁶ This approach is data driven and focuses on meanings explicitly available in the data. Thematic categories were identified based on all the available data, which was produced without establishing an a priori principle of saturation. All themes that emerged within the topic were noted and grouped according to their explicit similarities. Subsequently, descriptive labels were developed and then integrated into a flowchart to reflect the interactions between the themes and the period within which they were relevant.

The salience of the cognitive deficit and the awareness of cognitive loss were pivotal factors in participants’ engagement with the programme. Participants who were able to clearly identify cognitive symptoms and take ‘ownership’ of a complaint were more amenable to the suggestion of treatment and perceived the programme as an opportunity for improvement. In contrast, participants who declined the invitation or dropped out of the programme in the first week were typically referred by a healthcare worker who reported cognitive issues in the referral, but the participants themselves did not spontaneously indicate cognitive complaints (despite having shown deficits in previous cognitive assessments). These cases were typically coupled with reports of a lack of understanding about cognition in general and what rehabilitation is (e.g. P1: ‘To tell you honestly, I don’t even know what cognitive training is about!’).

For some participants, rehabilitation was for those with high levels of impairment. This suggests that patients may not have been seeking help with cognitive deficits because they did not see the cognitive deficit as a symptom that could potentially be treated unless it was incapacitating (e.g. P1: ‘I know that my friend went for cognitive training, but she doesn’t really have a memory’).

Expectations such as worsening of anxiety levels because of increased awareness of cognitive symptomatology, or the development of anxiety symptomatology because of exposure to the exercises, were recurrently reported by participants who chose not to participate or who dropped out early. Specifically, participants from the schizophrenia group appeared to take into consideration their abilities to cope with challenges and stress when deciding to do the rehabilitation process, while participants from the HIV group did not seem to have this concern, for example:

Two elements of this theme are thus apparent: (1) it is important for a person living with chronic illness to be able to identify emotional triggers in order to avoid relapses; (2) the programme is seen as a potential threat to self-esteem and hence a potential precipitating factor of emotional stress.

Offering rehabilitation programmes in hospital contexts is often associated with particular services. Attendance at specific clinics may thus force diagnosis disclosure by some healthcare users. This raises important ethical concerns that can act as a barrier to treatment, as a face-to-face invitation compromises the person’s rights to anonymity and privacy. Two of the healthcare workers indicated that their difficulty in attracting participants may have been partly a result of issues of confidentiality in relation to HIV. Additionally, participants who enrolled in the programme suggested that it was difficult to find interested potential users because ‘other people are not as open about disclosing their HIV status’ (Field Notes 1). This was thus an apparent obstacle in the HIV clinic but not for the schizophrenia group.

Because of the immune suppression that is commonly associated with HIV, even for those patients on antiretrovirals (ARVs), the presence of frequent illnesses can severely interfere with the ability to commit to a time-intensive training programme. Some of the participants were thus unable to enrol for the study because of ill health during the recruitment and installation stage (Field Notes).

Although one of the inclusion criteria for this research was having access to a computer and Internet connection, in practice, access was problematic. Specifically, the poor quality and high cost of Internet connections interfered negatively with some participants’ motivation to engage in the sessions (e.g. P4: ‘Some days I struggled to connect to the Internet and it made me feel so, so demotivated’). Furthermore, some participants were not aware of the quantity of data that was involved in the use of the program and hence did not plan for the costs involved (e.g. P6: ‘My mom pays for the Internet. I just thought you were going to pay for it’).

The implications of having limited access are threefold. Firstly, the quality of the Internet connection may interfere with the exercises, feedback and monitoring of the program and consequently increase the frustration of participants. Secondly, the amount of data required by the program could become problematic after the participants were involved with the training, which could threaten the participation of some participants with initial high motivation. These factors are out of the control of the participant and appear to negatively affect those with high motivation but low tolerance for frustration, which is a common combination in psychiatric populations. Thirdly, limited access must also be considered as a financial issue, which is of particular concern in contexts of high inequality like South Africa.

Use of the materials was also reportedly affected by the relative ease of access, as participants reported that their motivation was negatively impacted by the effort required to set up the laptop (e.g. P1: ‘I could find the time but I was like, just getting my computer out’). Consequently, some participants reported that these types of interventions would be easier to follow or adhere to if they were easier to access (such as through cellular phones and tablets) and available at times that were more convenient (e.g. P3: ‘But if it was on cell phone, like if it was at a place like where I was at the clinic waiting, then I could’ve done it there’).

Participants who displayed low tolerance for frustration had difficulties in maintaining participation because of the increased difficulty of the exercises. For some participants, the increases in difficulty were challenging and motivating. However, participants with high levels of anxiety associated with low performance reported feeling discouraged and overwhelmed when they experienced failure during the tasks, thus concluding that the programme was beyond their personal capabilities and deciding to abandon it.

Because the program adjusts the level of difficulty according to the progress history of the user, what is initially experienced as successful and rewarding may later on be seen as failure (e.g. P6: ‘I started off doing very well, but I just got worse and worse and worse at it’). The anticipation of failure causes disengagement with the activity and a tendency to give up (e.g. P1: ‘The ones that I couldn’t do, especially near the end, I get like demotivated’).

The increasing difficulty of the exercises also reportedly elicited in some participants a fear of failing (e.g. P1: ‘I suppose in a way I was a bit threatened, in a way. Um, like of failing’) and mental exhaustion (e.g. P3: ‘It was quite draining. It was almost like it took a hell of a lot of energy from me … after I’d done it I felt tired; my mind was completely tired’). These were reported as reasons for dropping out during the middle section of the programme.

Participants with difficulties incorporating the programme into their daily routine did not continue the training after initial uptake. Common challenges reported were associated with family and work responsibilities (e.g. P4: ‘There was just no time … It’s difficult to find suitable time during the day at the office’).

Unpredictable schedules or the inability to establish a regular daily training time were recurrently reported as reasons for discontinuance. This issue was particularly present in the schizophrenia group (e.g. P5: ‘My schedule is quite erratic generally so I didn’t have a specific schedule that I kept’).

Participants who were aware of their planning and monitoring difficulties expressed the need for an external source to provide support in this regard (e.g. P2: ‘If I had someone giving me a set timetable’; P1: ‘Say to me: P1 you must do it at ten o’clock today, or P1 you must do it at three’).

An ideal environment for the intervention is a quiet, well-lit room without interruptions. In clinical settings the environment is controlled in order to reduce interference. However, all participants engaged in the training in their homes, with many citing frequent interruptions (e.g. noise, cell phone, family members) as hindrances to their motivation to engage with the rehabilitation (e.g. P4: ‘I always have to wait until the kids are sleeping; otherwise they interrupt me a lot and sometimes it’s more difficult to concentrate while they are watching TV’).

The challenge with this tendency arose when the training was dependent on the availability of a facilitative space, instead of implementing a regularly scheduled slot without any regard for when the participant had an optimal cognitive disposition. Hence, participants who dropped out of the programme tended to report that the training was controlled or affected by the environment(s) available to them.

In particular, participants from the schizophrenia group reported dissatisfaction with their training environment on the basis of being disturbed by the presence of other people and a house-related routine that they claimed interfered negatively with their motivation to engage with the programme, for example:

The environment was also understood as an emotional ambiance, as they reported being negatively affected by the lack of a supportive and encouraging atmosphere (e.g. P9: ‘I just felt this place has, like, breaking me down. People aren’t encouraged … it feels like encouragement or more positivity or more group work or more teamwork’).