Study design

A pragmatic qualitative study design underpinned our approach to the study. Pragmatic studies, a category of generic qualitative descriptive research designs, seek to generate findings which are immediately translatable and applicable to the context where the study occurred [15]. The hospital’s administration was a stakeholder in the study’s design, which aimed to generate results to inform improvements to care delivery in this setting.

Theoretical framework

We employed the social-ecological model, which provides a conceptual framework for understanding the complex and interrelated effects of an environment on individual behavior [16, 17]. It comprises five nested levels – the individual, interpersonal, institutional, community and socio-cultural – corresponding to layered systems of influence (Figure 1). The framework provided a structure from which we could explore how providers’ hypertension practices are influenced by their context.

Sample

We conducted our study at a large, 110-bed urban private hospital in Uganda. Study participants were recruited from the group of providers working in the outpatient setting between June and August 2017. Study locations included the outpatient department, specialist clinic, and cardiology clinic, and providers sampled included nurses, general practitioners (physicians with one year of post-graduate internship), general physicians (physicians with post-graduate specialist training in internal medicine), and cardiologists.

Recruitment and data collection

We selected participants based on their degree of contact with hypertensive patients in the outpatient setting. Inclusion criteria designated English proficiency as a requirement for practical reasons (neither interviewer spoke other languages fluently), but this did not limit participation because all providers approached were fluent English speakers. We purposively sampled to ensure that participants had a diversity of work experience, gender, and training level. We recruited participants both by public announcement at a continuing medical education meeting and then individually based on staff lists or word of mouth. We sought to recruit between 15 to 20 providers, as that sample size would allow us to achieve data saturation [18].

The semi-structured interview guides were designed to assess providers’ approach to hypertension management: specifically, screening, diagnosis, treatment, follow-up, and use of the electronic medical record (EMR). We also asked about perceived barriers to hypertension care and hospital-specific challenges. The content of our qualitative guides was informed by initial findings from concurrent quantitative research, which investigated rates of hypertension prevalence, diagnosis, and control in the same setting. For example, observed inconsistencies in electronic recordings of blood pressures guided our questions about screening protocols and utilization of the EMR. We pilot-tested the interview guide with a Ugandan physician employed by the hospital to ensure that questions were clear, consistent with cultural norms, and appropriate for the clinical setting. We designed four research guides tailored to each provider type (e.g. general practitioner versus cardiologist); each was specific to the providers’ responsibility in managing hypertensive patients and their role in the context of an interdisciplinary hospital system. Following our first interview with a specific clinician type, we updated the guide informed by their responses and suggestions. We continued to iteratively update the guides’ content throughout the interview period based on emerging trends and provider feedback. The final interview guides are included as appendices.

Two researchers [ASG and HML] conducted participant interviews in English after obtaining written consent. Interviews were 30–50 minutes in length and were digitally recorded. Research team members transcribed the interviews, with transcription quality checked by a separate team member.

Analysis

We developed themes and codes for the study by directed content analysis, a method that uses predefined categories and concepts to structure the analytic process [19]. We based our codes and themes on the social-ecological model as described above, which allowed us to explore the relationships between individual providers and their broader interpersonal, cultural, and societal context. The iterative process facilitated a holistic analysis of hypertension management from the patient through the system level.

To conduct the coding process, three researchers [ASG, HML, DGK] used a sample of selected transcripts to independently generate codes, and then shared these codes with colleagues [DJH, APS, RNC] to iteratively generate a standardized codebook. Using this refined codebook, one researcher [ASG] utilized NVIVO software (version 11) to re-code all interviews and generate reports. We then analyzed these reports for patterns and emerging themes, with final themes and categories generated by team consensus.

Ethical Review

The study design, data collection tools, and consent forms were approved by the Mount Sinai Institutional Review Board, Clarke International University Research Ethics Committee, and the Uganda National Council of Science and Technology.

Patient self-management

All providers cited patient-centered factors, particularly medication and lifestyle non-adherence, as central challenges to hypertension management. Many physicians explained that patients do not feel sick, and thus see no reason to take medications: ‘[It] is within their means to adhere. But for something that is not causing any symptoms, becomes very tricky to enforce’ (C2). Relatedly, patients struggled to accept that chronic conditions require long-term medications. One provider explained, ‘So, that’s a problem we always get. Them to accept that they have a problem that is going to need life-long treatment. And they are very resentful to that’ (GP4).

Providers also described cultural expectations about weight and dieting as challenges to management. As one provider described:

Some people, for example, believe that when you’re obese it’s a sign of money. So they can’t adjust despite you advising them that we can’t manage your pressure as much as we give you drugs when you’re still overweight…and so they will think, when I lose weight, people know me by this size, when I lose weight people will think I’m, either HIV positive or bankrupt (GP3).

Providers also felt that patients were skeptical about “Western” medicine, and preferred alternative, herbal-based treatments found in the community. Other factors included lack of understanding about hypertension and its consequences, denial about their condition, and the inability to afford medications and other treatments.

Providers’ practices were consistent internally and with international guidelines

Participants were asked about their approach to screening, diagnosis, patient education, follow up, monitoring for complications, and treatment of hypertension. All interviewed nurses, who were responsible for taking and recording pressures, reported that every patient regardless of presenting complaint had their blood pressure taken at triage before visits. This was confirmed by general practitioners and physicians, who would not see patients until vitals had been obtained and logged. The one exception was the cardiology clinic, in which physicians took blood pressure measurements themselves. Providers across specialties reported similar thresholds and criteria for a hypertension diagnosis, and generally agreed that a diagnosis of hypertension required an elevated blood pressure reading on more than one occasion. Most considered assessment for secondary causes of hypertension (e.g., renal disease) a routine part of diagnosis:

Our next step [after diagnosis] is to evaluate the cause. Much as we know that 95% the cause is not identified, but we will evaluate for the cause… by doing some laboratory investigations, further physical examinations, and then appropriate radiological investigations (GP3).

Providers similarly described evaluations for complications of hypertension as standard-of-care. Serum lipids, glucose, renal and liver function tests, complete blood count, echocardiograms, and electrocardiography were cited as standard follow-up tests for new diagnoses.

Nurses and physicians perceived patient counseling as crucial to effective hypertension management and made patient education a priority of care. All clinicians mentioned providing diet and exercise counseling, and several regularly educated their patients about weight reduction, smoking cessation, reduction of alcohol intake, medication adherence and the long-term risks of hypertension. Most providers recommended lifestyle modification as initial hypertension therapy before starting medications, except in severe cases. Physicians prescribing practices were highly consistent: they chose medications based on patients’ comorbidities and end-organ damage, as well as drug cost and medication availability. Several practitioners and specialists additionally identified patient ethnicity as key factor in decision-making. When prompted about utilization of hypertension medications in different patient populations, one provider noted, ‘The only difference I find is that beta-blockers are more effective in Caucasians. than in our natives, our Africans… the beta-blocker alone will hardly make an impact on Africans with hypertension’ (GP4).

Practitioners and specialists generally believed hypertension control rates in their patient population were high but differed in their perception of what constituted a “controlled” patient. Most providers estimated control rates of 65–95% among patients in their care. Only two providers estimated lower control rates: 40% and 50% respectively. The participants who estimated high control rates more often cited patient adherence as a major limitation to blood pressure control.

Numerous providers, particularly specialists, relied on international guidelines to inform their diagnostic and treatment practices. Five participants reported utilizing the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC) guidelines, and other mentioned guidelines included those created by the United Kingdom’s National Institute for Health and Care Excellence (NICE), the American Association for Family Physicians (AAFP), the American Heart Association, the Uganda Ministry of Health, and Uganda Clinical Guidelines.

Challenges with care coordination, interprofessional collaboration and communication

Providers were confident and competent at individual-level hypertension management but lacked clarity on how to operate efficiently within the hospital system. For example, providers disagreed on which personnel and departments should manage hypertensive patients:

GP3: ‘Though the protocol is every patient with hypertension should be transferred to a physician.’

PH2: ‘I personally do not like chronic patients to be followed up in the OPD [outpatient department]. A lot of patients still come to the OPD with hypertension, diabetes, and so on because of convenience, but then what tends to happen is that they end up seeing different doctors, which to me is not an ideal situation. So I tend to channel them to the specialist center.’

GP10: ‘When controlled, there’s no complications… we see them, then we will ask them to come back after one month. So in case there are complications that we are not able to handle, we refer them to specialist.’

PH3: ‘So, if the patient’s blood pressure is okay and they [General Practitioners] don’t see any big issue going on, they don’t have to call me.’

There was also no uniform system for transferring referred patients’ information between departments. General practitioners in the outpatient department relied on the hospital’s electronic medical record in conjunction with a paper-based system; other physicians and specialists used a different paper chart in clinic. As such, providers lacked a single platform to track each patient’s progress, as articulated by a physician member of the hospital administration:

Sometimes we get patients who come in and they say they want a drug refill. So, somebody has not taken care to see, when was their last renal function, what’s happening with all that, what’s happening with eGFR and then before you know it in 2 years’ time, you look and you see somebody’s renal function actually deteriorated. So, that’s my main concern. Just making sure that we are not having that end organ damage as a result of no proper follow up. Because when they come to the [specialist] clinics, they have the book…yeah they have the follow up book so you can easily look back and see, when did this person have their lipid profile? When did they have their renal function? While when they come to the [outpatient department], they have some pieces of paper which are then taken away and stored, and the patient may not come back with their copy from the last time. So, you ask them when was the last test done. And then you have to go and check, it takes time. So people will omit some of those things because of the time (PH2).

Awareness and utilization of hospital guidelines was also highly variable. Often, one provider would cite a specific requirement or protocol – around diagnosis, treatment, referral or follow up practices – that other providers would contradict or not mention. Multiple providers, when prompted, denied any knowledge or use of guidelines specific to their hospital setting. However, most participants demonstrated enthusiasm for the development and implementation of hospital-wide protocols: ‘maybe if we could have protocols, SOPs [standard operating procedures], that would help…So that if I saw a patient or any other person saw a patient, we should come to the same conclusion about how to go about it’ (C2).

Generally, these observed inconsistencies in care coordination were not perceived as major impediments to effective care delivery. For example, when providers were probed specifically about challenges related to medical record-keeping, they reported frustrations with the EMR and acknowledged its limited potential for exchanging patient information. However, when asked more generally about barriers to hypertension care, no providers volunteered the EMR system as a barrier.

Existing systems challenge efficient hypertension management

Providers differed regarding which factors within the hospital system hindered hypertension care. Some providers believed that treatment cost rarely compromised care, but others considered the price of care (particularly medications and lab testing) to be a central barrier to effective treatment of hypertensive patients. One provider reported that his patients could often afford a visit but not the follow-up labs and tests, limiting his ability to optimize their management.

Similarly, some providers believed that the pharmacy stocked the most important hypertension medications adequately but lacked fixed-dose combination or specialty medications. Some complained that even basic medications were not available while others felt the pharmacy supply was entirely satisfactory. Several providers stated that generic, imported drugs stocked by the pharmacy were not effective:

But the challenge is the majority of the drugs, Indian drugs, Chinese drugs, the majority of the drugs stop. Sometimes you give, [and] you think reading of the blood pressure would be brought down by nifedipine alone… and the [blood pressures] are still elevated. You are forced to introduce a second drug. Yet, one drug would be enough. When they start to swallow the drug from Europe, they end up taking just half a dose and their pressures are no longer elevated. So one of the challenges is fake drugs. And we’ve been complaining [about that] (GP3).

Providers also noted that their patients shared the same concern: some patients even refused treatment with non-European medications.

The intersection of cultural norms, care seeking behaviors, and NCDs

Providers also cited socio-cultural factors as barriers to hypertension care delivery. Providers frequently complained about the absence of “wellness checks” in Uganda. Their patients were often only identified as hypertensive incidentally at unrelated visits, frequently late into their disease course and sometimes with end-organ complications. Many providers attributed this absence to a lack of effort from the media to educate the public about non-communicable diseases:

You have to be aware that you need the health checkup. Radio stations help in trying to advertise about health issues, sensitizing people. They are mostly, they are more like right now into cancers, which are prostate cancer, breast cancer, cervical cancer, is the campaign that they are really having now on health issues. But the issues about blood pressure and diabetes are not information tackled by the radio station. I think there’s a gap there, maybe you could try and make use of that (GP4).

Similarly, providers accused the government of failing to prioritize education and screening for chronic illness: ‘Our government has not taken up the initiative of educating people. It’s the NGOs… the churches, there are hospitals doing outreach… but the government does not. It’s not doing it the way it does it for HIV and other things’ (GP8). When prompted for suggestions to improve hypertension care, several providers mentioned societal solutions: increase public awareness, screening opportunities, and wellness checks and pressure the government to provide education and funding towards these ends.

Limitations

The generalizability of our findings is limited by the single site of the study: a large, private, urban hospital with a unique patient population (majority insured, middle upper class), a diversity of provider types (nurses, general practitioners, and specialists) and readily accessible services (in-hospital laboratory, full pharmacy). Our data collection and coding team comprised mostly non-Ugandan researchers, and all interviews were conducted in English. Thus, the research team’s prior beliefs regarding appropriate hypertension care – including the perceived efficacy and necessity of protocols and EMRs – may have biased our interpretation of the findings. As mentioned previously, interviews conducted by external observers may also have influenced providers’ responses, particularly those about their colleagues’ and their own clinical practices and decision-making. Nevertheless, our findings are consistent with prior data from high-income and limited low-and-middle-income settings as detailed above.