Study design

Our principal approach was semi-structured interviews with people with hypertension and/or diabetes. We explored their experience of services across the different levels of service delivery at two time points over the study period. This qualitative study was nested within a larger quantitative survey [9], and we extracted survey data to describe socio-demographic characteristics of the qualitative study participants.

Study setting

The study was set in the mostly rural Busia and Trans Nzoia counties in western Kenya, with populations of between 900,000 and 1 million and high levels of poverty [5]. Each county has one county referral hospital (level 5), five (Trans Nzoia) to six (Busia) sub-county hospitals (level 4), seven (Trans Nzoia) to 12 (Busia) health centres (level 3) and between 38 and 47 dispensaries (level 2) serving up to 200 community units. PIC4C was implemented across 40 level 2–5 health facilities in Busia and 33 in Trans Nzoia.

Sampling strategy and recruitment

Study participants were identified from respondents to a survey of patients’ experiences of treatment burden (n = 301) (12/2020-2/2021) who had indicated willingness to be contacted to participate in an interview. Details of the survey methods are reported elsewhere [9]. In brief, survey respondents had been screened and hypertension and/or diabetes confirmed through PIC4C service efforts between 2018 and 2020. For this qualitative study, we used purposive sampling, seeking to recruit 30–40 participants across the two counties, include a range of perspectives of those experiencing care stratified by gender, health condition(s), age and facility level, and allow for some anticipated loss to follow-up over interviews. We reviewed the diversity of patients in the sample and the data generated periodically during the first round of interviews, concluding that a final sample size of 40 was appropriate as all identified categories of experience were included, and no new substantive themes were being identified.

Potential participants were contacted by telephone (May–July 2021 and February–March 2022) by trained bilingual research assistants who explained the purpose of the interview and what participation would involve. For those agreeing to take part, an interview date most convenient for the participant was set. Participants were provided with an information sheet, which was read out to those who were unable to read and time given for questions. Informed consent was sought, and was recorded in writing and verbally at the beginning of the interview recording.

Data collection

Interviews were held in person at a health facility closest to participants’ homes between 24 May and 8 July 2021 (round 1) and between 24 February and 4 March 2022 (round 2). This enabled exploration of patients’ experiences at different points in their trajectory since diagnosis and the setting-up and embedding of PIC4C services. Interviews were conducted by trained research assistants using a semi-structured interview guide, informed by literature on treatment and self-management burden [10–12], while also exploring people’s experiences of the health service (S1 File). The topic guide was adapted for the second interview to focus on changes in participants’ circumstances, and experiences of care and interactions with the health system since their first interview. We also explored patients’ recommendations for strengthening services, and for health workers introducing similar services in other counties. Interview guides were developed in English, translated into Swahili, piloted with four participants, reviewed by the field research team, adapted and finalised. Interviews were conducted in Swahili or English, audio recorded, transcribed and translated into English. Transcriptions and translations were reviewed for accuracy, and linked and cross-checked with survey data. Identifiable information (individual and health facility names, locations) was replaced with anonymous identifiers during transcription.

Analysis

Demographic characteristics of the sample were described using existing survey data [9]. Interview transcripts were analysed using a thematic approach. This included familiarisation of data through reading and rereading transcripts, organising data according to pre-defined and newly identified topics, and synthesising and comparing data to identify core themes [13, 14]. We used NVivo 12 software to support qualitative data analysis [15]. We shared findings from a preliminary analysis of round 1 interviews in a stakeholder workshop in May 2022 with representatives from patient groups, community health workers and health facility staff, regional decision makers and the national Ministry of Health; feedback regarding areas to explore in greater depth informed subsequent analysis.

Ethics review

The study received approvals from Moi University Institutional Research and Ethics Committee (FAN:0003586) and the London School of Hygiene & Tropical Medicine (17940), and a research permit from the National Commission for Science, Technology and Innovation (NACOSTI/P/20/4880).

Inclusivity in global research

Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the Supporting Information (S3 Checklist)

Participant characteristics

We interviewed 43 patients, of whom 37 participated in both rounds; three participated in round 1 only (S1 Table) and three in round 2 only. Ten participants (23%) reported having diabetes, 19 (44%) hypertension and 14 (33%) both diabetes and hypertension (Table 1). Over half of participants (n = 27) were diagnosed within two years preceding the interview. Twenty-three were female and 19 male. The majority (n = 25) were educated to primary level. Income was low, with half of participants (n = 21) reporting average monthly household income of <3,000 KES (USD27 in January 2021 (1 USD/110 KES)) [16] during the previous 12 months. One quarter (n = 11) reported coverage by the National Health Insurance Fund (NHIF).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Demographic and health characteristics of study participants.

https://doi.org/10.1371/journal.pgph.0003245.t001

Cross-cutting contextual factors

We report findings by each stage of patients’ interactions with PIC4C integrated services from screening and diagnosis to referral and follow-up and ongoing treatment. We identified two inter-related contextual factors that dominated participants’ accounts across their diagnosis and treatment journey. These were: (i) individual financial resources and (ii) participants’ social situation. We first introduce these contextual factors, which informed interpretation of subsequent findings.

Screening and diagnosis

Sixteen participants (37%) reported having initiated their patient journey at community level, generally through local screening events delivered by community health volunteers (CHV) linked to PIC4C. Prompts for attending screening varied. Some reported having experienced unexplained symptoms and seeing local screening events as an opportunity to seek health advice, while others responded to community health worker invitations to attend events. These events helped people who could not easily afford to pay for services to be diagnosed (‘let’s go and be tested, it is free’ [ID42; hypertensive]), reducing the need to travel as screening services were ‘brought to the villages’ and enabling patients to be ‘found’ by community health workers:

If it weren’t for them [CHVs], I would still be chilling out there, and perhaps my life would be over [ID33; hypertensive].

Study participants were generally accepting of their diagnoses. Those who had experienced unexplained symptoms were grateful to have an explanation for this and access to treatment. Several noted that they had been diagnosed in the past and not ‘accepted’ or ‘followed up’ on their diagnosis, but re-diagnosis following community screening had enabled active engagement in self-management and treatment. For some, this was specifically related to information provided by CHV and referral to primary care, as described below. Study participants recognised the importance of patient counselling by community health workers to support acceptance of diagnosis and treatment:

As they screen people, they can attend to them…if they go somewhere where these services have not been introduced, those people don’t understand what diabetes is, they have not understood what hypertension is, unless they screen and tell them. They should also be counselled so they can accept they have these conditions, and then start taking medications. [ID21, hypertensive and diabetic]

The role of counselling provided by CHV was seen to be especially important for people who do not experience symptoms:

It is about explaining to them politely so that they may understand their condition, because there are some people who refuse medicines because they don’t believe, because they don’t feel any pain’ [ID41; hypertensive]

Referral and follow-up

Participants described their appreciation of simple, specific procedures for referral to primary and secondary level services. Clear instructions, for example being given a specific date and location to attend hospital, helped patients to follow up and receive care. Some were also encouraged by community health workers, emphasising the importance and urgency of seeking care, for example: ‘She told me not to delay in going to the hospital, [as] I had been found with hypertension and it was elevated’ [ID04; hypertensive].

Patients were instructed to attend regular NCD clinics so that healthcare providers could monitor their condition(s) and prescribe medication. Interview participants reported that they endeavoured to attend regularly, and some explained that they found information provided during their monitoring visits useful to track their progress in managing their condition. Three measurements were highlighted by patients; blood sugar level (diabetes), blood pressure (hypertension), and weight (for those who had been advised to lose weight). For asymptomatic patients, monitoring visits provided knowledge about their health status which patients did not otherwise have access to:

If you are at home your pressure can rise and you don’t know because you don’t have the measuring machine. When you come here you are measured and then they tell you your state, that helps us [ID17; hypertensive]

Making services available at lower level health facilities was widely seen to encourage regular clinic attendance, as was being given a choice of clinic, because it minimised travel and associated costs for some participants who now were able to access services closer to their home:

I’m happy because it is close. [Location] was very far. I’m happy because I come here at 7am while in [location] we used to go at 5am….I’m happy here. I even walk from home to this place. [ID08; diabetic]

However, those living further away from health facilities continued to experience major challenges to attend monitoring appointments regularly, in particular those with very limited financial resources; consultation fees charged by level 4 and 5 facilities also continued to pose a major barrier to access services for some participants.

Similarly, clinic wait times were experienced as challenging, with reports ranging from 4–6 hours to entire days, due to the volume of patients waiting. Long waits were especially problematic where they interfered with income-generating activities, with participants saying that they would delay or miss follow-up visits altogether. For frail people and those with diabetes, who had to attend clinics fasting, long waits frequently caused them to feel unwell: ‘we will take long….we always come when we have not eaten anything. When it reaches 1 or 2 pm you start to feel dizzy’ [ID10; diabetic and hypertensive].

While long waits could cause people to drop out of care, positive interactions with healthcare providers tended to support retention in care. Participants generally felt well treated by health care providers, which they identified as a strength of the care they received: ‘you’re given a warm welcome and you’re not laughed at. It feels nice’ [ID10; diabetic and hypertensive]. Participants gave examples of staff listening and responding to challenges they encountered in condition management, for example changing medication prescriptions in response to side effects and adapting dietary advice where first line alternatives were not available or affordable to patients. Participants also described receiving individually tailored advice, which enabled them to incorporate exercise into their daily routine, for example while travelling to work, doing farm work or playing with grandchildren, depending on the individual’s circumstances. Respondents reported that they themselves understood the advice they were given, but they commented that some older patients who spoke languages other than Kiswahili might need additional support:

There are those who don’t understand, like the really old women who don’t understand. They only understand their language. Many a times they have problems. [ID18; hypertensive]

In these cases, fellow patients sometimes helped to translate.

Treatment and self-management support

PIC4C integrated services aimed to provide treatment and support patient’s management of their condition(s) through education and ongoing support. Interview participants showed clear understanding of strategies to self-manage their condition through diet and exercise, and of the importance of medication adherence. However, incorporating these into their daily lives was often seen to be challenging because of associated costs and variable availability of medicines. We explore each factor in turn.

Change over time

Reflecting on whether or how things had changed between interviews (an interval of 8–10 months), people with hypertension and/or diabetes commonly highlighted the importance of ‘taking time’, in some cases significant periods of over a year, to adjust to and implement lifestyle changes and establish optimum medication regimes. Participants described instances of struggling to cope with the various demands that a diagnosis of chronic disease had placed on them and their families, such as being able to secure reliable sources of income to support their family and pay for medications. Patients cited the support from the treating doctor that had enabled them to cope with and address challenges to manage their condition, for example, difficulties to follow dietary advice, and how this had improved their situation over time:

I have seen a lot of changes this year. I feel different not like before. I do my work the one I trained [in]. I continue with the drugs and I’m not as sick as I was last year.

[Interviewer: You are saying that this year your health has improved?]

My health has improved and I have lost weight. This is because I was obese. After I got advice from the doctors and I reduced and changed the food that I take, I have seen great change. I’m doing well. [ID03, diabetic and hypertensive]

Additionally, we identified changes over time in some participants’ capacity to engage with services and adhere to treatment regimens, which was frequently shaped by their households’ vulnerability to financial shocks. One example includes unexpected illness of a family member, causing reprioritisation of household financial resources, which meant reduced clinic attendance and gaps in purchasing medication.

A key change at the system level that patients identified at their second interview was an increased volume of people accessing services (although this was not universally reported), for example:

The blood pressure population has gone high. When you come to the dispensary the number of hypertensives is more [ID15, hypertensive]

In some cases, this resulted in longer waiting times for patients unless staffing was expanded to meet the increased demand, although this was not widely observed, except in some cases (‘the number [of health centre staff] increased and they became free to see us’, meaning that ‘right now you don’t stay for too long waiting’ [ID24; diabetic and hypertensive]). Participants also reported changes in clinic organisation, which separated NCD patients from general patients, and this was seen to reduce waiting times:

Nowadays it has improved because we have our own room. When you arrive on clinic day, you just go to your room and you will be helped [ID28; diabetic]

Strengths and limitations of study

Key strengths of our study were the training and experience of the field team, the in-depth nature of interviews conducted at two time points within a nested study design, which allowed analysis of temporal changes and use of qualitative and quantitative data. Drawing on different data sources particularly strengthened our understanding of the range and the impact of individual financial resources and affordability of services.

Our study participants were people who engaged with care, had participated in a survey and were willing to be interviewed for further research. We therefore did not include people who did not engage with the PIC4C model after diagnosis. We cannot be certain about the nature of bias this has introduced into our data but it is likely that our sample has lower presentation of people with poorer experiences of diagnosis and linkage into care, and/or those who may have experienced greater barriers to accessing care.

More importantly perhaps, we anticipated that there was a risk of social desirability bias, meaning participants reporting what they believe to be expected positive behaviours and a reluctance to voice concerns or be critical about the service and service providers. We aimed to address this possibility by explaining that there was a clear separation between the researcher/s undertaking the interview and health care workers involved in PIC4C service delivery. We also explained that the purpose of the study was to understand how services work for patients. Study tools, developed with input from senior and junior staff across the team were designed to include open-ended questions and probes for positive and negative experiences.

Our sample included people with diverse conditions who were recruited across two counties (total 22 facilities, 11 in each county). This enabled triangulation of interview and previously collected survey data to ensure a robust data set, and inclusion of demographic and income data.

The study was a partnership between Kenyan researchers with extensive experience in implementation and evaluation of interventions, and clinical experience of service delivery for chronic conditions in Western Kenya, and UK based researchers with experience of implementation research in care for chronic conditions in diverse settings. Regular debrief meetings were held during data collection. The field team was led by a Kenyan medical sociologist (VN) with extensive experience of collecting, analysing and reporting qualitative research. Data analysis was led by a UK based researcher (RW) with experience in primary data collection and translation, and in conducting qualitative analysis with international teams. This combination enabled both in-depth understanding of the context in which the data was generated, and analytical distance. Discussion of preliminary findings with stakeholders was valuable to discuss interpretation and highlight themes of policy interest for exploration in greater depth.