Search

Background: Cardiovascular diseases (CVD) such as ischaemic heart disease and cerebrovascular incidents cause up to 12% of the total deaths in all ages for both males and females, and one in four of all hospital admissions in Kenya. Further, risk factors for these diseases have increased by a third in the last ten years (2005-2016). The use of simple and efficient interventions such as CVD risk assessment and scoring tools is largely suboptimal in primary care settings. Current practice in the prevention and management of CVD risk in Kenyan primary care settings remain poorly characterised for clinicians, patients and policy makers. Therefore, the aim of this thesis was to examine the prevention and management of CVD risk in Kenyan primary care settings by proactively identifying and characterising important elements relevant to optimising practice in these settings. Objectives: 1) Assess the potential burden of CVD in the Kenyan population by modelling CVD risk and factors underpinning this burden; 2) Systematically synthesise evidence on the implementation of CVD risk assessment in primary care and how factors that influence implementation relate to and impact on Kenyan primary care settings; 3) Understand current practice by characterising the experiences of key stakeholders in the prevention and management of CVD risk in Kenyan primary care; and identifying factors influencing the adoption of recommended practices; 4) Explore the inaccessibility of essential CVD medicines in Kenyan primary care as a significant challenge in the assessment and management of CVD risk in these settings. Methodology: This thesis consists of an epidemiological study providing the most upto-date information on the potential burden of CVD in the Kenyan population,illustrating the importance of the growing burden and highlighting potential drivers of this burden (Chapter 3). It also includes a mixed-methods systematic review (Chapter 4) that synthesises the evidence on the implementation of CVD risk assessment in primary care (as a means to address the burden identified in Chapter 3) and how factors that influence implementation relate to and impact on Kenyan or similar primary care settings. A qualitative study (Chapter 5) identifies key stakeholder practice experiences in CVD risk prevention and management in Kenyan primary care and the factors influencing the adoption of recommended practices. Lastly, a cross-sectional observational study of retrospective data, i.e. latest national essential medicine lists to compare the recommended levels of use of essential medicines to treat the leading causes of mortality in Eastern sub-Saharan Africa (CVD, TB and HIV/AIDs) (Chapter 6). Results: Study 1: Using data from 1627 participants from the Kenya STEPwise survey carried out between April and June 2015 as part of the international WHO approach to surveillance of chronic disease risk factors, this study provided the most up-to-date information on the potential burden of CVD in the Kenyan population by modelling CVD risk and factors underpinning this burden. This risk modelling was based on the updated CVD risk prediction model for Eastern sub-Saharan Africa which was developed by the WHO with the International Society of Hypertension (ISH). The median CVD risk score in the study sample was 3.0% (IQR 0.3, 46.2), and the majority (69.1% [95% CI 66.8-71.3]) of the subjects were grouped in the lowest CVD risk category (< 5% CVD risk). According to the national guidelines on CVD prevention and management, up to 11.9% of the study population (moderate, high and highest-risk categories) are eligible for both pharmacological and lifestyle interventions. Women recorded a higher mean CVD risk score than men: 5.0% (SD 5.5) vs. 4.6% (SD 4.7). This is in addition to more women (14%) than men (9%) in the study population having a CVD risk score of more than 10%. The mean CVD risk scores differed significantly across socio-demographic and behavioural groups. The data for the subjects showed that their educational levels, marital statuses, ethnicities, their frequency of consumption of processed food that was high in sugar, the number of days on which they ate vegetables in a typical week, whether they had received lifestyle advice on reduction of their salt intake in the previous three years and the number of days in a typical week on which they undertook vigorous-intensity activities as part of their work were significantly associated with their CVD risk scores. Factors such as ethnicity, level of education, marital status, dietary practices and physical activity could be used to refine further the entry points for targeted CVD risk screening, prevention and management in Kenyan primary care settings. This study suggests that public primary care facilities are a critical entry point for targeted screening and the delivery of interventions aimed to prevent and manage CVD risk. This work will be submitted for peer review shortly. Study 2: A mixed methods systematic review was conducted in accordance with PRISMA guidelines to synthesise evidence on the implementation of CVD risk assessment in primary care and how factors influencing implementation relate and impact on Kenyan or similar primary care settings. An extensive search was conducted through seven electronic databases, hand-searching references, and in grey literature. The quality of studies was assessed through the MMAT tool, the critical appraisal checklist by NICE and the Axis tool, depending on the study design of the studies included; mixed-methods, qualitative and cross-sectional studies, respectively. Of the 3010 studies screened, only 25 studies were included in the review. This review broadly conceptualises factors influencing implementation of CVD risk assessment as those relating to CVD risk scoring tools, users, clinical settings and the healthcare system. This review explains that the implementation of CVD risk assessment interventions is complex and is often affected by the interaction of the interventions with the real world: the users and the environment in which they are being used. Notably, this study illuminates a critical gap in CVD related implementation research from LMICs. This work is published in BMC Implementation Science. Study 3: This study qualitatively 1) examined current practice by understanding how prevention and management of CVD risk was experienced by clinicians, patients, and policymakers in everyday clinical practice; and 2) identified contextual factors that influenced the adoption of recommended practice in Kenyan primary-care facilities. This involved conducting 560 hours of observations, 15 interviews with clinicians and policymakers, 30 exit interviews with patients and a documents analysis across seven healthcare facilities in Kilifi County, Kenya. The use of both thematic and framework analysis helped to 1) sufficiently identify a critical gap in knowledge and practice partly due to a lack of familiarity with clinical guidelines related to CVD risk and to better understand the use of clinical judgement in CVD risk assessment by clinicians in these facilities; 2) categorise factors that influence the adoption of recommended practice in CVD risk prevention and management in Kenyan primary care into three levels: level 1 (clinician and patient), level 2 (clinical setting and local government), and level 3 (global and national influences); and 3) explore the intricate relationships across these three levels. As such, findings from this study can be used to develop and support meaningful and pragmatic change in CVD prevention and management in these settings. This work will be submitted for peer review shortly. Study 4: This study is follows on from one of the main findings of the qualitative study: inaccessibility of essential CVD medicines, which poses a significant challenge in the assessment and management of CVD risk in Kenyan primary care settings. By comparing the recommended level of use of 42 essential medicines extracted from nine National Essential Medicines Lists (NEMLs): CVD (n=29), HIV (n=8), and TB (n=5), this cross-sectional observational study of retrospective data confirmed the hypothesis that there is a disproportionate percentage difference in the recommended levels of use of essential medicines used in communicable diseases compared with those for NCDs in NEMLs. The findings from this study show that at level 1 (proxy for primary care), more essential medicines were recommended for HIV and TB treatment than for CVD risk prevention and management.

Background. HbA1c monitoring to guide diabetes management usually requires more than one clinic appointment. Point-of-care (POC) testing allows testing and treatment changes in a single appointment, but trials of POC HbA1c testing have not been shown to benefit patients and the healthcare system. This work explores how HbA1c testing in primary care could be streamlined using POC testing. It comprises four pieces of work and is guided by complex intervention and test-evaluation frameworks. Methods. Database analyses were used to describe current practice in HbA1c testing and identify factors associated with testing schedules not recommended by guidelines. Systematic review and meta-analysis was used to determine the most accurate POC device. A device was selected to use in a feasibility study to determine how POC HbA1c testing can be used in primary care consultations. Data on recruitment, care pathways, care delivery, satisfaction and behaviour were gathered using data collection forms and questionnaires. Qualitative interviews with patients and clinicians explored perceptions and barriers to POC testing. Results. Database analyses showed that high or low HbA1c testing rates are associated with previous HbA1c levels and use of incentive schemes. Most POC HbA1c devices have a mean negative bias and large variability compared to laboratory methods. When POC HbA1c testing was used in GP practices, nurses used the result to make clinical decisions during the consultation. Nurse consultations were longer than GP consultations (22mins versus 11mins) and POC results were successfully uploaded to patient records. Patients liked POC testing and found it convenient. It helped some link their HbA1c result to recent behaviour. Clinicians found the tests easy to use but were concerned about cost. Conclusion. POC testing can be effectively delivered in primary care but funding mechanisms are needed to encourage adoption. This work could inform the design of a randomised trial to test the effect on patient outcomes.