Curbing the non-communicable disease epidemic: an evaluation of integrating rapid testing for NCD risk factors and navigated linkage to care into a standard HIV testing service platform for adults in Soweto, South Africa

Hopkins, Kathryn L
Journal Title
Journal ISSN
Volume Title
BACKGROUND: The top causes of death and disability worldwide are chronic non-communicable diseases (NCDs), such as cardiovascular disease (CVD) and diabetes mellitus (DM), which contribute to more than three in five deaths. South Africa is also the epicentre of the human immunodeficiency virus / acquired immune deficiency syndrome (HIV/AIDS) epidemic, containing the world’s highest prevalence of HIV-1. People living with HIV (PLHIV), now living longer while on treatment, are becoming increasingly at-risk for non-HIV-related chronic conditions similar to the HIV-uninfected population. In understanding the expected impact of NCDs on both the general population and amongst PLHIV, it becomes critical to pursue healthcare strategies that effectively and efficiently prevent, diagnose, treat, and control NCDs for both HIV-infected and uninfected persons. AIMS: This study aimed to provide an overall evaluation of integrating counsellor task-shifted rapid diagnostic testing (RDT) for NCD risk factors with optional peer-navigated linkage to care within an HIV Testing Services (HTS) centre in Soweto, South Africa. Specifically, this study sought to, 1) determine the prevalence patterns of NCD risk factors amongst adult HTS clients by age, sex, HIV-status and ART-use; 2) describe the client level of satisfaction with integrated NCD-HTS, as compared to standard of care HTS; 3) assess the quality and feasibility of task-shifting RDT for NCD risk factors to counsellors; 4) determine the proportion of integrated NCD-HTS clients linked to care and initiated on treatment using peer-navigated linkage to care, as compared to standard of care HTS clients using passive referral; and 5) investigate the barriers to integrated NCD-HTS client linkage to care and initiation of treatment across chronic diseases. METHODS: This study was a prospective evaluation which consisted of a two-phase longitudinal study design to compare targeted study measures between the three-month standard of care HTS phase utilising passive referral for HTS clients with abnormal/positive results (19 February to 14 June, 2018) against a subsequent nine-month integrated NCD-HTS phase with optional peer-navigated referral (18 June 2018 to 28 March 2019). Standard of care HTS provided counsellor-led height, weight and BP measurements; symptoms screening for sexually transmitted infections (STIs) and tuberculosis (TB); and HIV rapid testing. Integrated NCD- xi HTS further included counsellor-led BMI categorisations and RDT for blood glucose (both RPG and average HbA1c) and TC services. Walk-in, adult clients at the PHRU Zazi HTS Centre who consented to the health screening programme were enrolled and had HTS client files opened. These paper files collected data on socio-demographics; health risk behaviour (alcohol and tobacco use); previous health screening history and treatment use for hypertension (HT), DM, hypercholesterolemia, and HIV; health screening results; clinic flow start and stop times; number of correct and incorrect counsellor-provided referrals; and peer-navigated referral uptake. These clients were then able to consent into an embedded study, comprised of HTS client clinic exit surveys and linkage to care follow-up surveys (only for clients with at least one referral for an abnormal result with follow-up conducted up to three month post-clinic visit). Surveys included both open and closed-ended questions using a five-point Likert scale to investigate HTS client satisfaction with HTS clinic flow time and services provided; time to linkage to care and treatment initiation; and reasons for not choosing peer-navigated referral, linking to care, and/or linking to care but not initiating on treatment. All quantitative statistical analyses were conducted in SAS Enterprise Guide 7.1 (SAS Institute, Cary, NC) using SAS/STAT procedures. Descriptive statistics (e.g. medians, means, interquartile ranges [IQR], and standard deviation) were reported for continuous variables. Frequencies and associated percentages were determined for categorical variables, stratified where appropriate by age group (18-24, 25-34, 35-44 and ≥45 years), sex, HIV status, ART use, and HTS phase. Fisher’s exact or chi-square tests were conducted to test statistical significance for categorical measures stratified by HTS phase. Comparisons for descriptive statistics by HTS phase were determined using the Kruskal-Wallis test and Student’s T-Test. Survey responses to close-ended and open-ended questions (once thematically sorted) were tabulated. RESULTS: There were 325 and 780 clients with aggregate data collected within standard of care HTS and integrated NCD-HTS, respectively. Of the 780 enrolled integrated NCD-HTS clients who were screened, 19.2% (n=149/775) were HIV-infected, 37.5% (n=289/770) were overweight/obese, 18.0% (n=139/772) had high BP, 10.8% (n=83/766) had high HbA1c (four clients had high random glucose), and 8.1% (n=62/768) had high TC. Women had significantly more overweight/obese BMI than men (46.8% [n=237/506] vs 19.7% [n=52/264]; p<0.0001); and 39.8% (n=127/319) of 18-34 year old women were overweight/obese. Males had significantly higher BP than women (23.9% [n=63/264] vs. 15.0% [n=76/508]; p=0.0023). Of 18-34 year xii olds, 7.2% (n=31/433) had high blood glucose; and there were over three times as many women aged 25-34 years with high TC as compared to men of the same age group (18.2% [n=20/110] vs 5.6% [n=4/71]; p=0.0151). The majority of HIV-infected clients were female (72.5% [n=108/149]). HIV-infected clients had significantly more high HbA1c than HIV-uninfected clients (16.1% [n=24/149] vs 9.6% [n=59/614]; p=0.0223). HIV-infected individuals on ART had significantly higher TC than those not on ART (21.7% [n=10/46] vs. 4.9% [n=5/103]; p=0.0016). Of the HIV-infected participants, 48.3% (n=72/149) already knew their HIV-status, and of those, 63.9% (n=46/72) were on ART. Of all participants screened with elevated BP, high HbA1c and high TC, 72.4% (n=97/134), 96.1% (n=73/76), and 93.3% (n=56/60) were newly diagnosed for each condition, respectively. Of the clients who self-reported as being previously diagnosed, 83.8% (n=31/37) of hypertensive clients, two (n=2/3) clients with high HbA1c, and one (n=1/4) client with high TC self-reported they had already been initiated on treatment. All clients were found to have uncontrolled disease after screening. There were 284 and 333 HTS client clinic exit surveys analysed for standard of care HTS and integrated NCD-HTS, and therefore clinic flow time was analysed for the same sample size. Standard of care HTS clients spent significantly shorter time in minutes in the clinic from start to finish as compared to integrated NCD-HTS clients (86.0, IQR: 72.0-108.0, [n=283/284] vs 102.0, IQR: 87.0-136.0, [n=331/333]; p<0.0001). Of the integrated NCD-HTS clients, 97.9% (n=320/327) were ‘very satisfied’ with integrated NCD-HTS, overall, despite standard of care HTS having significantly shorter median time for counsellor-led HTS. Regarding quality of task-shifting RDT of NCDs to counsellors, there was one (n=1/3, 33.3%) and five (n=5/59, 8.5%) incorrect referrals made by counsellors for low and high BP, respectively. There were 2.4% (n=2/83) and 6.5% (n=4/62) incorrect referrals given for high blood glucose and blood cholesterol, respectively. There were 82 standard of care HTS and 238 integrated NCD-HTS client study participants with at least one abnormal health screening result enrolled for follow-up. Of the 320 referrals across both HTS phases, 40.0% were HIV-infections, 11.9% STIs, 6.6% TB, and 28.8% high/low BP. Of referred integrated NCD-HTS study participants, 29.4% were referred for glucose and 23.5% cholesterol. Integrated NCD-HTS had significantly more clients linked to care for HIV (76.7% [n=66/86] vs 52.4% [n=22/42], p=0.0052) and clients linked to care within a shorter average time (6-8 days [IQR:1–18.5] vs 8–13 days [IQR:2–32]) as compared to standard of care HTS. Integrated NCD-HTS clients initiated HIV/STIs/BP treatment on average more quickly as compared to standard of care HT (5 days for STIs [IQR:1-21], 8 days xiii for HIV/BP [IQR:5-17 and 2-13, respectively] vs 10 days for STIs [IQR: 4-32], 19.5 days for HIV [IQR:6.5-26.5], 8 days for BP [IQR: 2–29]). The majority of participants chose passive over peer-navigated referral (89.1% vs 10.9%; p<0.0001). Participants rejecting peer-navigated referral preferred to go alone (55.7% [n=39/70]). Non-linkage to care was due to being busy (41.1% [n=39/95]) and not being ready/refusing treatment (31.6% [n=30/95]). Non-initiation of treatment post-linkage to care was due to normalised results assessed at clinic (49.7% [n=98/196]), prescribed lifestyle modification/monitoring (30.9% [n=61/196]), and poor clinic flow/congestion and/or further testing required (10.7% [n=21/196]). CONCLUSIONS: There is a chronic NCD health transition plaguing South Africa, and amongst relatively young adults (18-34 years), especially women. The current risk-factor based, gated-approach to chronic disease screening, as recommended within South African national guidelines, do not adequately allow for the most effective screening for and prevention of DM or hypercholesterolemia, or their pre-cursors amongst the general population or PLHIV. It is feasible to integrate quality counsellor-led NCD rapid testing into standard of care HTS within historical HTS timeframes, yielding client satisfaction with HTS services and clinic flow. The integration of universal rapid HbA1c and TC screening amongst adults into standard of care practice within the public health sector, regardless of BMI categorization, HIV status or ART use, is recommended. While optional peer-navigated linkage to care did connect significantly more patients to the point of entry to care and more efficiently than standard of care passive referral, it did not translate to increased treatment. Same-day treatment initiation was not achieved across chronic diseases. While the implementation of integrated NCD-HTS within the study site was found to be feasible and clients also deemed the clinic flow time and services provided to be highly satisfactory, the transferability and scale-up of such healthcare systems strengthening strategies within the public sector may not be as successfully. A more robust, multi-site, nation-wide study is required to determine if similar results would occur within government HTS clinics. There is a need for linkage strategies to also address various, complex levels of psychosocial and health systems barriers at the point of entry into care and treatment. A revision of current health screening and management guidelines and the implementation of cost-effective, multi level interventions to assist in identifying individuals with unknown disease and both engaging and retaining patients in care is recommended
A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy to the Faculty of Health Sciences, School of Public Health, University of the Witwatersrand, Johannesburg, 2021