Factors shaping participant reporting of health and treatments
Memory When discussing discrepancies between their responses to the different questioning (elicitation) methods, participants often said they forgot to report things in response to the general enquiry and, to a minimal extent, the checklist. Minor, intermittent and resolved health issues were more likely to be forgotten compared to those more severe, persistent or current. There were no consistent categories of treatments recalled more easily, except for ARVs, which could never be forgotten as they are considered so important to sustaining health:
“I only used it [diclofenac, an anti-inflammatory] to reduce the tooth pain. When the pain stopped I [forgot] them. But with these ones [ARVs] I have no time to [forget] them, maybe God takes my spirit [fear of ill health/death from non-adherence]” [In-depth interview (IDI) 11, Tanzania].
Participants predominantly attributed forgetting to report to the burden of having too much to remember, although two South Africans said it was due to significant memory decline since starting ARVs. The detail of the checklist, and to some extent the in-depth interview, was strongly viewed as helpful in overcoming memory problems.
Significance to participants When faced with a general enquiry, participants in both sites described conscious decisions to report current medical histories and AEs that were more bothersome, severe or persistent as opposed to those experienced as intermittent or less severe. Several South African participants also described decisions to first spontaneously report a new deterioration in health during the trial only when the symptoms worsened further. South African participants often recounted medical history in relation to when they initiated ARVs, when health was at a low point from AIDS-related illness or initial side-effects. Illnesses experienced since this period (major or minor) were commonly under-reported,
“Ever since 2006 I don’t want to lie but I don’t feel anything… I always just get sick like ordinary people” [IDI 4, South Africa].
This extended into reporting about AEs during the trial period, with declarations of good health followed by mention of health issues “No, I was not sick at all. Just only coughing, flu, only that” [IDI 8, South Africa].
Participant definitions or normalising of ill-health influenced when and how they revealed information in the in-depth interview. Categorising illness as 'major’ or 'bad’ was frequently associated with hospitalisation; tuberculosis treated as an outpatient was not being 'badly sick’, and only revealed in response to a specific question about tuberculosis. One participant did not mention her blindness in one eye until specifically asked.
Non-significance amongst Tanzanian participants sometimes reflected a slightly different concept. Rather than relating to classification of an experience as 'normal’ and therefore not reported, participants seemed to shape their perception of significance more around expectations of what would be significant to the doctor. When asked about symptoms by way of a general enquiry, those deemed the main problem(s) were chosen with expectation that others would be treated by default:
I don’t think that you can tell the doctor one part after another that is in pain. … But you may decide to tell him the basic problem, or what is making me more sick. If I tell the doctor that I have fever he might give me the medicine … then all that I am feeling will calm down [Focus group discussion (FGD) 1 Tanzania respondent 5].
Treatment use phraseology meanwhile revealed a hierarchy; after ARVs or antimalarials, use of intermittent or over the counter substances (such as painkillers or vitamins) were mentioned, but qualified with 'only’ or 'apart from’. These perceptions of significance may intersect with the next factor shaping reporting behaviour: relevance to report.
Relevance to report Participants appeared to delay reporting experiences that they perceived irrelevant, with the checklists helping them to decide what was necessary. Manifestation of relevance to report was, largely, different between sites. In South Africa, decisions to report were sometimes related to the trial’s objectives, i.e. whether the information would contribute to the success of the trial:
“If I say after drinking the tablet I felt weak and tired … then they will write it … [as] the effects of the tablet. [To] them it gives …the wrong impression” [IDI 5 South Africa].
History of sexual dysfunction, and poor memory (not on the checklists), were only revealed in the in-depth interview by two participants, who described these as irrelevant to the trial, and also untreatable, and therefore not useful to report. In Tanzania, relevance was chiefly attributed to the medical context. If a symptom at baseline was considered by the participant to be due to something other than malaria, such as work-related activities, it might not have been reported unless specifically prompted for. Similarly, if medication was bought for something other than malaria there was no reason to report:
“I thought because I purchased it [pain killer] by myself because of the tooth pain I had no reason of telling him. But the Fansidar I purchased, I had a reason of telling him, because I can be tested and seen with malaria again. So they’ll understand, after the Fansidar being useless, what [they should] do” [IDI 11 Tanzania].
Consequences – attempts to control a situation Participants in both sites described themselves or others in the trial withholding information in response to general questions and checklists for fear of negative consequences. These negative consequences were starkly different between sites. In South Africa, participants feared exclusion from the trial; there were three explicit, independent, second-hand reports of self-treatment with laxatives, non-adherence with ARVs, and gastro-intestinal illness that were withheld for this reason. When such scenarios were presented during the focus group discussion, participants denied that this happened, saying that misreporting could undermine the trial’s objectives. In contrast, Tanzanians considering consequences of reporting (or not) did not mention the trial; they were worried about going against hospital 'rules’. One said she did not report an anti-inflammatory taken during the trial for this reason. Most discussion on this topic in Tanzania, however, was in the focus group discussions and concerned fear of reporting traditional medicines:
We who are living with [HIV], .. we are highly advised to refrain from traditional medicine issues because there is a difference between traditional medicines and these drugs. ….Yesterday I felt ill and went to dig muarobaini [neem tree root used for malaria] but without feeling any relief after drinking it. If I come to [doctor] today and he asks me what kind of medicine you used yesterday, telling him muarobaini….I must lie because already that is wrong. …That is due to fear of telling the doctor because we were already prohibited that thing. …There is some sort of [difficulty] telling the truth though it’s true, but the truth remains within me. [FGD 1 Tanzania, respondent 3].
Articulating a response South African participants, who generally did not take long-term medications apart from ARVs, appeared to have a small personal formulary of named over-the-counter medications, but names of ad-hoc prescription medications were largely unknown. Though literate, one man displayed little interest in knowing what his prescription was. By contrast they had impressive detailed knowledge of their previous and current ARV regimens. In Tanzania, names of prescription medicines, including ARVs, were not known. As for HIV itself, ARVs were mentioned euphemistically, even when reverence appeared as high. In focus group discussions, Tanzanian participants initially cited their lack of education and an inability to read English labels, though it became clear that names of ARVs and other long-term prescription medicines are seldom verbalised during prescribing and dispensing. Meanwhile, equally complex names of antibiotics and antimalarials are known, because they are talked about in public.
Respondent: But I have not kept those [ARVs] in my mind. I have their container but I have never read it.
Interviewer: How come you knew amoxycillin?
Respondent: [laughs] They are simple to pronounce
Interviewer: How did you get that name?
Respondent: Because they are mentioned in the streets … like aspirin, simple names.
[IDI 4 Tanzania].
The blood test tells doctors everything they need to know When South African participants narrated what happened in the trial, this was largely about blood tests and examinations, with little mention of the verbal discussion with the clinician until they were prompted. While this could relate to the numerous blood samples taken in the trial ward and knowledge about the trial’s pharmacokinetic objectives, it persisted in their descriptions of follow-up visits, when tests were far fewer. When asked about the relative importance of blood tests versus being asked about their health and use of medicines, several declared the former as equivalent or superior.
“[The test is] like they are asking about your health. … it’s the same thing” [FGD 01 South Africa respondent 01].
Others said blood could reveal non-adherence to ARVs and use of any substance, including traditional medicines. Tanzanian participants did not display such a marked pattern, possibly because they expected to be asked about malaria symptoms and what antimalarials they had tried in order to inform treatment. However, when questioned about the importance of blood tests versus being asked about their health and use of medicines their discourse reflected that the former were considered a far superior source of information for the doctor. One participant indicated that questions were asked merely for the doctor to identify the more important test required.
Social context of the trials
While there may also be influence from psychological factors, conversations with participants in this study reveal social constructs that may underpin reporting behaviour. A minor construct “Being a subject” was common to both sites, though much stronger in South Africa; participants indicated they were beholden to certain trial-dictated behaviours, mostly regarding concomitant medication use. However, two more dominant constructs, “Trial citizenship” and “Deferred responsibility” may explain some of the differences observed between the sites.
Trial citizenship in South African inpatient volunteers South African participants described their important role, indeed sometimes their work, in facilitating the trial’s success. They were largely knowledgeable of, and aligned their work as trial participants with, researchers’ objectives. On hearing a participant was withdrawn due to a contra-indicated medication prescribed by her own doctor, they felt she was correct in reporting it even though she could not play her trial role anymore:
Interviewer: So, how did you feel about [another participant being withdrawn from the trial for taking a contra-indicated medication]?
Respondent 3: We feel bad
Respondent 5: We feel very bad
Respondent 4: But [on] Saturday, the doctor told us [about] the side-effect [of] the tablets that the girl was taking for the skin. It will affect what? [Respondent 2: kidneys]. ... Then [that] girl, she can’t work with us. [FGD 02 South Africa].
Even though participants conveyed this sense of responsibility towards the integrity of the trial, there were second-hand reports of others not playing their role properly, deliberately withholding information for fear of being withdrawn. Though loss of reimbursement was cited as a reason for this, one participant expressed it in relation to a worry about having to leave because participants had “come [to] enjoy the trial” [IDI South Africa, respondent 11]. This South African trial was conducted under stricter conditions than the trial in Tanzania in terms of, for example, admission for intensive pharmacokinetic sampling periods, and the foods and concomitant medications allowed. Being with each other and the staff for 72 hours offered a liberating experience; there was much mention of the opportunity to speak freely with others who understood what it meant to be HIV positive, phone numbers were swapped and focus group discussions were happy reunions. This may have nurtured the trial citizenship underpinning some decisions regarding relevance to report, which were sometimes made by consensus. Thus, for instance, an AE was considered unrelated to the South African trial if others were not experiencing it:
“Somebody will wake up and say 'Guys I am feeling this, is anyone feeling it?’ And then because that one said no… you will also think 'Ah maybe it’s me. It’s only me” [IDI South Africa, respondent 05].
Over and above these participants’ roles in this trial, however, there was recognition that they were monitored to avoid personal risk. Perhaps in addition to trial citizenship, the fact that the antimalarial trial drug was being used experimentally in participants without malaria meant that they had a heightened vigilance for possible side-effects:
“I was busy trying to look for side-effects from the trial.” [FGD 01 South Africa respondent 4] “They asked me how did I feel.. and what the medicine [did] to me” [FGD 02 South Africa, respondent 3].
Deferred responsibility in sick Tanzanian outpatients In Tanzania, particularly in the in-depth interviews, there was little discussion, or sometimes understanding, of the trial. Some participants showed how the locus of responsibility for knowing relevant information fell with trial staff as clinicians, rather than with themselves; the doctor had the knowledge to prompt them to reveal whatever additional information was required over and above what they respond to a general enquiry:
“So I think the doctor understands more, that’s why he went on to probe. So I didn’t forget or wasn’t careless, but it’s my knowledge that is low, that if the stomach aches then even diarrhoea may be there, 'what about diarrhoea?” [FGD 1 Tanzania, respondent 1].
Tanzanian participants were recruited to the trial individually when they presented to routine services with malaria symptoms for which they sought a cure. The trial was largely incidental to them achieving good health. There were few references to researchers’ priorities; it was personal:
“I didn’t just join because of the name malaria, no. But it’s because there are tests and examinations that are being done on me so as to know how my health is. …So I was being researched. So I know how my state is, I was examined” [FGD 1 Tanzania, respondent 1].
The experience predominantly described was of feeling ill, joining a project (to get optimal management), and thus feeling better. Participants considered the questions the doctor asked about health and treatments were for their personal benefit:
“When he asked what kind of drug have you used, I suspect that he asked so that we don’t make it a habit buying drugs from drug shops, we should come and get tested first” [FGD 1 Tanzania, respondent 4].
In the in-depth interview, responses to questions about change in health since baseline overwhelmingly concerned gratitude for improvement, malaria-related or not.
Trial clinicians' reflections on the limits of sensitivity
Participants in both sites had overwhelmingly recommended that more detailed questioning (checklists or in-depth interviews) helped them to report, and Tanzanian participants said that the focus group discussion taught them the importance of maintaining their own detailed illness and treatment record, for its personal health benefits. Trial clinicians, meanwhile, spoke of the challenge of eliciting comprehensive but relevant data when they will never know everything. For well-studied drugs, the focus of more detailed questioning could be on known or anticipated risks, combined with general enquires to detect anything else. But it was a quandary whether to probe for AEs that are perhaps insignificant or irrelevant to both clinicians and participants:
“If they choose not to tell me about their headache when I’m asking them how they are and how they [feel] in their body…how severe could it be or… how important [is it] to them?“ [Trial clinician 2].
For trials of drugs with known safety profiles:
“Isn’t the grading such a key issue? Because people’s lives are full of minor mishaps, and minor symptoms all the time. And if you really, really wring it out of people, you could generate a lot of irrelevant grade 1 stuff” [Trial clinician 4].
There was, however, concern that selective detailed questioning could miss minor illness that impacts on adherence, and thus efficacy and an increased risk of malaria resistance at a population level. These clinicians, reflecting how patients may be intimidated by their role as doctors, despite them being at pains to be otherwise, suggested that other cadres of staff be involved in order to overcome barriers to reporting. This involvement, they said, could include designing elicitation strategies (social scientists), questioning participants (nurses or social scientists) or interpreting safety results (anthropologists).