Secondary Logo

Journal Logo

Implementation Science

Integrated HIV-Care Into Primary Health Care Clinics and the Influence on Diabetes and Hypertension Care: An Interrupted Time Series Analysis in Free State, South Africa Over 4 Years

Rawat, Angeli MPH, PhD*; Uebel, Kerry MBBS, MFamMed, PhD; Moore, David MD, MHSc, FRCPC; Yassi, Annalee MD, MSc, FRCPC*

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: April 15, 2018 - Volume 77 - Issue 5 - p 476-483
doi: 10.1097/QAI.0000000000001633



Low- and middle-income countries (LMICs) with high HIV burdens face unique challenges, where health systems must address the confluence of twinning epidemics of noncommunicable diseases (NCDs) and HIV. NCDs account for 70% of deaths globally; 3 quarters of which (31 million) occur in LMICs.1 Although impressive gains have been made in expanding access to antiretroviral therapy (ART) in Eastern and Southern Africa, merely 50%–58% of those needing treatment were accessing it in 2015, with one-third of ART enrollees seeking care at advanced disease stages.2,3 Given that the number of people on ART in LMICs is expected to rise to 28.5 million by 2025,4 comorbid and multimorbid conditions are also predicted to rise. In 2013, 62% of adults aged 20–64 years in sub-Saharan Africa were experiencing sequelae of multimorbidities, necessitating improvements to health system organization and prioritization.5

One widely promoted strategy to facilitate scale-up of ART is to integrate or decentralize HIV care to the primary health care (PHC) level.6–10 Integration has been described as the “management and delivery of health services so that clients receive a continuum of preventive and curative services, according to their needs over time and across different levels of the health system.”11 While evidence exists supporting integrated NCD and HIV care within PHC settings in Kenya,12 Zambia,13 and South Africa,14 little empirical evidence exists evaluating how integrating HIV care into PHC clinics (PHCCs) influences PHC service delivery, especially for NCDs.8,15,16 Addressing this research gap has been challenging because of weak study designs, poor data quality, and outcomes that are difficult to compare.8 Evidence that examines the effect of integration on non-HIV services has been heavily focused on outcomes related to the integration of HIV care with tuberculosis,17,18 sexual/reproductive health and maternal/child health,19 and in low-income countries with large investments from international donors.19–21 With the recognition that HIV can be managed within PHC-based health systems and the growing shift of HIV funding toward national health systems, HIV care demonstrates a large chronic disease care model that could be leveraged to address impending noncommunicable disease epidemics while ensuring quality NCD care.6,22,23

South Africa currently has the largest ART programme in the world with 7 million people living with HIV, 3.4 million of whom are on ART.24,25 In 2012 during the middle of the study, diabetes prevalence in South Africa for those older than 15 years old was 9.5%, whereas the hypertension prevalence was 31.8% with only a 38.4% hypertension treatment coverage.26 South Africa is a country with one of the highest rates of HIV globally, with rates of diabetes and hypertension in people older than 50 years which are higher than any other African country.27

Multimorbidity and comorbidity present challenges in countries such as South Africa where populations on ART will be aging with HIV. Oni et al27 found that 45% of total prescription visits in a PHCC in South Africa constituted 4 conditions: HIV, tuberculosis, diabetes, and hypertension, where younger patients with HIV (under 45) had more multimorbidities than their noninfected counterparts. While the evidence is limited, lifelong ART may be associated with increased diabetes, dyslipidaemia, and myocardial infarction, which is especially challenging for HIV-burdened countries experiencing an epidemiological transition in the determinants of cardiovascular diseases.28 Understanding the potential benefits and vulnerabilities to providing quality NCD care within PHCCs when HIV care is integrated is especially important for countries where a large proportion of the population are HIV infected, such as in South Africa.

This study aims to identify changes in NCD care for diabetes mellitus and hypertension before and after HIV-care integration in public-sector PHCCs and identify if changes may relate to the numbers of patients with HIV in PHCCs.

Study Setting

This study was conducted in the Free State province, South Africa—a province with a total population of 2.8 million,29 comprising 5.2% of South Africa's population. In 2012, the HIV prevalence was 20.4% for 15–49-year olds, higher than the national prevalence of 18.8%.30 The ART treatment programme began through a vertical system in 2004 where patients accessed HIV testing, treatment, and care at discrete sites.31 In April 2010, a national policy32 to integrate HIV care into all PHC facilities was implemented. In this study, we defined integration as the month and year that patients could initiate HIV treatment within PHCCs. This enabled patients to access comprehensive HIV care, from prevention to treatment initiation and follow-up, as part of the services provided at the public-sector PHCCs (a.k.a., “mainstreaming33”). Nurse Initiated Management of ART (NIMART)—a process of training PHC-based professional nurses in the comprehensive management of patients with HIV—was a key component of integration. Nurses were trained through the Practical Approach to Lung Health in South Africa (PALSA plus) guidelines34 before the integration of HIV care into PHCCs. Also before integration, clinics underwent a readiness assessment, and additional pharmacy assistants and data entry staff were employed to strengthen ART drug delivery systems and data monitoring in PHCCs. In practice, the manner in which integrated HIV service delivery was offered varied by site. For example, some sites had disease-specific nurses in separate consultation rooms within the same clinics (ie, co-location), whereas at others, nurses providing comprehensive care for HIV and other illnesses were in 1 consultation room. Where possible, the Free State Department of Health (FSDOH) attempted to remedy space concerns through modular structures and pharmacy renovations. The study period encompassed 4 years from April 2009—12 months before the implementation of the integration policy until April 2013—36 months from when the policy was first introduced. Clinics integrated at different months and year across the study period.

Ethics approvals were obtained from the Behavioural Review Ethics Board at the University of British Columbia in Vancouver, Canada (Certificate number H11-02259) and the Health Sciences Ethics Board at the University of the Free State in Bloemfontein, South Africa (Rec reference number 230408-011, IRB number 00006240). This manuscript is part of a larger mixed-methods study examining the impact of integrated HIV care in PHCCs on the health system.


Clinic Selection

All PHCCs that had at least 1 professional nurse in April 2009 were identified within the District Health Information Software (DHIS) (n = 224) and matched to clinics in the HIV data set from 3 Interlinked Electronic Register ( The following clinics were excluded: (1) “priority sites” identified by key informants as clinics that had been designated as ART clinics under the previous vertical programme—excluded because the clinics received additional staff or financial resources to provide HIV care (n = 45); (2) clinics that integrated HIV care on or after October 2012—excluded for not contributing a minimum of 6 months of data postintegration or for not having integrated (n = 25); (3) atypical clinics—excluded because the clinic catered to specific populations were co-located with a hospital, run by nongovernmental organizations, or used elements of the previously vertical system for HIV treatment or assessment (n = 8); and (4) missing HIV dataexcluded for missing more than 4 months of consecutive HIV data (n = 15). A total of 131 (56%) clinics met the inclusion criteria (Fig. 1), including 37 urban, 64 rural, and 30 from a former homeland PHCCs (Supplemental Digital Content Table 1,

Clinic selection criteria—analysis of noncommunicable disease indicators.

Noncommunicable Disease Indicators

NCD indicators were selected based on the South African National Department of Health PHC package.35 In an iterative process over the study period, FSDOH managers were interviewed to understand data availability, reliability, potential confounders, and indicators that were most likely to be indicative of, and sensitive to, changes in clinic functioning. Preference was given to indicators that were (1) monitored by the FSDOH, and that prompted action and inquiry when an indicator changed; (2) available and relatively complete over the study period; (3) robust (ie, resilient to outliers); and (4) important to the national goals of PHC service provision. We used Jenkins box plots and interquartile range to visualize trends and to identify outliers, respectively. Four NCD indicators and 2 HIV-specific indicators were selected for analysis (Table 1).

Noncommunicable Disease (NCD) and HIV Indicators

The 4 NCD indicators represent workloads related to new diabetes mellitus and hypertension patients. These were examined relative to the clinic catchment population older than 30 years as designated annually by Statistics South Africa and to patients older than 5 years who attend study clinics monthly. Two HIV-related service indicators, which were explanatory variables of interest, captured workload from new and existing patients with HIV on ART at the PHCCs.

Data Sources and Collection

Administrative data were obtained from the FDSOH and collected from and the DHIS. Indicators were collected in the source data monthly, except for those that were calculated once per year. For these, a mid-year population estimate was divided by 12 and used as the denominator for each month. Data were aggregated to the clinic level with patient identifiers removed before acquisition by the research team. HIV service indicators were extracted from for some clinics (20%, n = 26) and DHIS for most clinics (80%, n = 105). Information on the month and year each clinic integrated was collected by the Provincial Nurse Initiated Management of ART Mentoring and Support Coordinator.


Interrupted Time Series Analysis

The primary analysis used interrupted time series analysis36 to understand changes in the trends of indicators pre- and post-integration. We measured and analyzed baseline trends, as well as trends before, after, and at the time of intervention. Baseline trends were defined as the predicted trend that would have been expected to occur through time if integration had not taken place. The following formula described by Lagarde37 was used to model the changes since integration:where Yt is the outcome variable for an indicator of PHC across the study period denoted as “t” for Time. The trend at baseline was represented by β0(intercept) and existing or secular trend estimated by β1(time). A change at integration was estimated by β2(integration), and the change in the postintegration trend when compared to its preintegration trend was estimated by β3(post slope).

One interrupted time series model was applied for the each of the 4 indicators at the clinic level. For each clinic the month and year of integration was coded as 0, with months leading up to the integration coded sequentially negative for each month before integration (ie, −1 for the month preceding integration, −2 for 2 month preceding integration, etc.) and months postintegration were coded sequentially in the positive direction (ie, 1 for the month directly after integration, 2 for 2 months after integration, etc.). Hypertension indicators spanned from April 2009 to March 2013 (a 48-month period) (Table 1) while diabetes indicators data were only available from April 2009 to February 2012. After this date these indicators were discontinued. Since the month and year clinics integrated (ie, time zero) varied across the study period, we assumed that seasonal trends were negligible. Results are presented for 2 periods: ±18 and ±30 months since integration and are adjusted for first level autocorrelation (AR1). We chose these periods to maximize the available data across the study period. It is known that behavior before and immediately after an intervention may influence the outcomes.36 Therefore, as a sensitivity analysis, we broadened the definition of integration to include a 5-month period (ie, 2 months before and after the month of integration) to account for lead and lag times as recommended by key informants.

Linear Mixed-Effect Regression Models

We investigated whether HIV-related patient workload may influence a change in the trends when comparing pre- and post-integration. Using linear mixed-effect regression analyses, we observed changes based on the addition of 2 HIV-related explanatory variables. A total of 3 models were created for each outcome of PHC: (1) without HIV-related indicators, (2) with the addition of total new patients initiated on ART per month per facility, and (3) with the addition of total HIV patients in care at the end of the month, per month, and facility. Further explanation of the data sources, management and models used including assumptions can be found in the supplement.


The 131 clinics included in the study represent a total catchment population of 54% of the province and 1.5 million people (2013). The distribution of the months and years that clinics integrated across the study period was not uniform (Supplemental Digital Content Figure Supplement-A, Of the 131 clinics that met the inclusion criteria, integration (ie, the month the clinic provided comprehensive HIV care) occurred between March and December 2010 for 21% (n = 27), January and December 2011 for 42% (n = 55), and January and September 2012 for 37% (n = 49).

A dramatic increase in the number of patients accessing ART in PHCCs was identified among the 131 clinics over time (Supplemental Digital Content Table 2, Figure Supplement-B, In the first month of the study (April 2009), 1614 patients (1495 adults and 119 children younger than 15 years) were in on ART in PHCCs. By the last month of the study (end of March 2013), a total of 57, 958 patients (54,070 adults and 3888 children) were on ART. During the study period 44,596 treatment naive adults and children were initiated on ART accounting for 77% of total patients in care at the 131 PHCCs in our study. The number of patients remaining on ART was higher than the number of newly initiated patients on treatment presumably because patients initiated on ART in the previously vertical model were referred to these PHCCs after integration.

Total HIV Patients Remaining in Care and Initiated on ART at Primary Health Care Clinics (n = 131)

From the interrupted time series analysis, despite an increase of new hypertensive patients on treatment at the population level at the time of integration (seen from the data at ±18 months) of 99 per 100,000 (SE = 44, P < 0.03), decreases were identified at ±30 months postintegration (Fig. 2). A decrease of 6 per 100,000 (SE = 3, P < 0.02) new hypertension patients placed on treatment per population older than 30 year was identified (Table 3). No changes were observed at ±18 months since integration. New initiations on ART influenced both indicators and a decrease of 1 per 100,000 (P > 0.002) new hypertension patients placed on treatment per PHC client older than 5 years (P < 0.001) was identified (Table 3). No changes in the number of new patients with diabetics on treatment were observed pre- versus post-integration. Data to assess the influence of new initiations on ART or existing patients with HIV on ART in PHCCs on diabetes indicators were insufficient because the diabetes indicators were not available for the entire study period.

Interrupted time series trend ±30 months pre- and post-integration (noncommunicable disease).
Beta Coefficients From Interrupted Time Series Analysis at ±18 and ±30 Months From HIV Integration for Noncommunicable Disease Indicators (Per 100,000)


Our study provides evidence of potential decreases in new hypertension patients on treatment at the population level during a rapid and massive scale-up of treatment and care for patients with HIV in an integrated PHC model in Free State, South Africa (Fig. 2 and Table 3). The reduction could be due to the fact that hypertensive patients were not diagnosed; were diagnosed and not placed on anti-hypertensive medications; or were diagnosed and treated but not recorded; or a combination of these factors within an underresourced health system. Our qualitative findings from health care workers and managers38 suggest the possibility that increased strain on resources and higher provider workloads resulting from adding HIV services may have compromised the quality of hypertensive care or reporting. More than 91% of South Africa's hypertensive population is not receiving screening, treatment, diagnosis, or care.39 These findings are important, given the challenges of the South African health system in addressing the quadruple burden of disease within changing mortality patterns where NCDs account for more deaths than HIV and tuberculosis combined.40

Despite ambiguities on how to define and operationalize integration,41 evidence from other studies suggest that integrating HIV care into PHC on health systems leads to enhanced efficiencies between HIV and non-HIV service provision. For example, increased access to ART through integration has resulted in improved survival and greater engagement in care.6,9,33 Furthermore, increased non-HIV service utilization,20 improved clinic infrastructure and sharing of resources,13,20 and strengthened referral systems and laboratory capacity23 have been observed since integration. Conversely, other studies have shown that patients' wait times may have increased13 and that patients may experience reduced access to specialists.42 In addition, quality of care may be compromised because of increased HIV-related workloads for health care workers, resulting in a reduced focus on non-HIV conditions,43 especially in contexts with prevailing high workloads and health care workers' shortages.23,44 Hyle et al45 suggested that implementation of integrated strategies could have unintended consequences that require further research to ensure health systems are strengthened.

Further research from multiple settings on integrated HIV and NCD care is urgently needed, especially with a health system lens. Systematic reviews have highlighted the need for understanding micro- to macro-level health system factors that influence NCD integration including strong communication, governance, policy, and leadership; well-equipped facilities; trained, coordinated, and motivated health care workers; and patient-centered health systems.46–48 In addition, cascade analysis and mixed-methods studies could identify where health system resources could be maximized to meet patient care needs and determinants of changes in care. This is especially important with a need to maintain high quality of care with integrated service delivery models.

The strengths of this study are the analysis of longitudinal population-level data over a critical period of 4 years during the implementation of integrated HIV care in PHCCs. To our knowledge, it is the first study using population-level data examined during multiyear implementation, especially in a context where investment from international donors was largely absent. Although the generalizability of the study may be limited because the unique context of the South African health system and high burden of HIV, the implementation of integration in this setting occurred as part of a national implementation scheme; therefore, the lessons learned could be useful for other countries considering national policy changes.

There are also some limitations to this study, particularly around the accuracy of data and the availability of indicators. Many shortcomings exist in the data collection system in South Africa, which have been discussed previously,49 as well as the use of administrative data broadly.50 To ensure that the most robust indicators were collected, selection was prioritized if the data were monitored by a programme manager, which facilitated verification and explanation when changes occurred during the study period. Although we were unable to account for cross-catchment service utilization, the denominators were purposefully chosen conservatively where populations older than 5 and 30 years were used. Therefore, any significant changes detected were likely underestimated, making us believe that changes are not an artifact of data collection. Because of changes in the administrative data set during our study period, ideal diabetes indicators were discontinued in 2012, and alternate data elements were used which may not reflect the workload related to diabetes care because most diabetes patients are initiated on treatment outside PHCCs. Therefore, we expect an underestimation for this indicator. The improved standardization of administrative data, more indicators on patient flow, readiness of clinics to implement integrated services, and how integrated care is delivered at the clinic level could allow for comparability across settings to assist in formulating a larger picture of the impact of integration on NCD care. In addition, these indicators could allow implementers to identify challenges in clinic function as implementation progresses.

A limitation, as discussed by Atun et al,41,51 is that integration of an intervention in a health system is not a binary function. Rather, integration comprises a complicated process that occurs over time within the setting of a complex adaptive system. Therefore, understanding the interconnectedness and dynamic complexity of the system would allow for a deeper understanding of the short- and long-term consequences of integrated interventions. By defining integration as having occurred at 1 discrete point in time, we were unable to account for changes leading up to integration or directly after integration to understand how these influenced the trend. Presumably, in the first few months after integration, patient numbers would increase slowly until a peak, at which time clinic function may begin to deteriorate (or strengthen) and would impact PHC. To account for this, we conducted a sensitivity analysis by broadening the window period that was defined as the time of integration to 5 months (ie, 2 month lead and lag time) which agreed with the primary analyses. As with other large data sets, the relevance of the statistical significance versus the clinical significance must be acknowledged. As with any observational study, the changes observed may be due to unmeasured confounding effects. However, during the indicator selection, we were able to understand potential confounders, and the analysis allowed us to control for autocorrelation and identify the secular trends that could be influencing results. Last, because of the high multicollinearity of the HIV explanatory variables with covariates in the models (eg, integration and time), the magnitude and direction of the influence must be interpreted with caution.


Integration of HIV care into primary health care is a viable and necessary strategy in high HIV prevalence settings through which to expand access to ART and to begin to reach universal coverage. The gains in infrastructure and investment in HIV care could be leveraged to strengthen, not erode, NCD care. Rates of NCDs are projected to increase dramatically in the coming years, and by 2025 it is expected that 75% of adults in LMICs will be living with hypertension.52 By harmonizing preventive efforts to reduce and treat NCDs within the context of HIV care, countries can synergistically advance health and social benchmarks. South Africa has made strides toward implementing sustainable, comprehensive HIV care within the national health care system; a critical step for the HIV epidemic. However, the growing burden of other NCDs combined with the greater life expectancy of people living with HIV demand that health systems remain strong to ensure that comprehensive HIV care does not come at the expense of screening and treatment of NCDs, especially in PHC settings. Adequately resourced health systems are needed to ensure that integration is effective and for implementation, whereas empirical evidence should guide best practices for integration for high-quality, patient-centered primary health care.


The authors would like to acknowledge the support of the Free State Department of Health in sharing the data, Dr. Michael Law for contributing comments on the analysis and interpretation and Corine Bond for the data visualization in the supplement. The authors appreciate the insights from reviewers that enriched the manuscript.


1. WHO. WHO noncommunicable diseases. Available at: Accessed June 4, 2017.
2. WHO. WHO 17 million people with access to antiretroviral therapy. Available at: Accessed June 4, 2017.
3. Auld AF. Trends in prevalence of advanced HIV disease at antiretroviral therapy enrollment—10 countries, 2004–2015. MMWR Morb Mortal Wkly Rep. 2017;66:558–563.
4. Gupta A, Juneja S, Vitoria M, et al. Projected uptake of new antiretroviral (ARV) medicines in adults in low-and middle-income countries: a forecast analysis 2015-2025. PLoS One. 2016;11:e0164619. Available at: Accessed June 27, 2017.
5. Atun R. Transitioning health systems for multimorbidity. Lancet. 2015;386:721–722.
6. Frenk J. Reinventing primary health care: the need for systems integration. Lancet 2009;374:170–173.
7. World Health Organization. Global update on HIV treatment 2013: results, impact and opportunities. 2013. Available at: Accessed February 27, 2014.
8. Vasan A, Ellner A, Lawn SD, et al. Integrated care as a means to improve primary care delivery for adults and adolescents in the developing world: a critical analysis of integrated management of adolescent and adult illness (IMAI). BMC Med. 2014;12:6.
9. Walton DA, Farmer PE, Lambert W, et al. Integrated HIV prevention and care strengthens primary health care: lessons from rural Haiti. J Public Health Policy. 2004;25:137–158.
10. Gilks CF, Crowley S, Ekpini R, et al. The WHO public-health approach to antiretroviral treatment against HIV in resource-limited settings. Lancet. 2006;368:505–510.
11. World Health Organization. Integrated health services-what and why? 2008. Available at: Accessed January 31, 2012.
12. Edwards JK, Bygrave H, Van den Bergh R, et al. HIV with non-communicable diseases in primary care in Kibera, Nairobi, Kenya: characteristics and outcomes 2010–2013. Trans R Soc Trop Med Hyg. 2015;109:440–446.
13. Topp SM, Chipukuma JM, Giganti M, et al. Strengthening health systems at facility-level: feasibility of integrating antiretroviral therapy into primary health care services in Lusaka, Zambia. PLoS One. 2010;5:e11522.
14. Ameh S, Klipstein-Grobusch K, Musenge E, et al. Effectiveness of an integrated approach to HIV and hypertension care in rural South Africa: controlled interrupted time-series analysis. J Acquir Immune Defic Syndr. 2017;75:472–479.
15. Dudley L, Garner P. Strategies for integrating primary health services in low- and middle-income countries at the point of delivery. Cochrane Database Syst Rev. 2011:CD003318. doi: 10.1002/14651858.CD003318.pub3.
16. Briggs CJ, Garner P. Strategies for integrating primary health services in middle- and low-income countries at the point of delivery. Cochrane Database Syst Rev. 2009:CD003318. doi: 10.1002/14651858.CD003318.pub2.
17. Howard AA, El-Sadr WM. Integration of tuberculosis and HIV services in sub-Saharan Africa: lessons learned. Clin Infect Dis. 2010;50:S238–S244.
18. Coetzee D, Hilderbrand K, Goemaere E, et al. Integrating tuberculosis and HIV care in the primary care setting in South Africa. Trop Med Int Health. 2004;9:A11–A15.
19. Matsubayashi T, Manabe YC, Etonu A, et al. The effects of an HIV project on HIV and non-HIV services at local government clinics in urban Kampala. BMC Int Health Hum Rights. 2011;11:1–11.
20. Price JE, Leslie JA, Welsh M, et al. Integrating HIV clinical services into primary health care in Rwanda: a measure of quantitative effects. AIDS Care. 2009;21:608–614.
21. Rasschaert F, Pirard M, Philips MP, et al. Positive spill-over effects of ART scale up on wider health systems development: evidence from Ethiopia and Malawi. J Int AIDS Soc. 2011(14 suppl 1):S3. Available at: Accessed April 26, 2014.
22. Schneider H, Blaauw D, Gilson L, et al. Health systems and access to antiretroviral drugs for HIV in Southern Africa: service delivery and human resources challenges. Reprod Health Matters. 2006;14:12–23. Available at: Accessed March 7, 2014.
23. Pfeiffer J, Montoya P, Baptista AJ, et al. Integration of HIV/AIDS services into African primary health care: lessons learned for health system strengthening in Mozambique-a case study. J Int AIDS Soc. 2010;13:1–9.
24. South Africa UNAIDS. Available at: Accessed January 12, 2017.
25. UNAIDS. UNAIDS announces 2 million more people living with HIV on treatment in 2015, bringing new total to 17 million. Available at: Accessed January 12, 2017.
26. Health Systems Trust.South African health review. 2016. Available at: Accessed January 12, 2017.
27. Oni T, Youngblood E, Boulle A, et al. Patterns of HIV, TB, and non-communicable disease multi-morbidity in peri-urban South Africa- a cross sectional study. BMC Infect Dis. 2015;15:20.
28. Keates AK, Mocumbi AO, Ntsekhe M, et al. Cardiovascular disease in Africa: epidemiological profile and challenges. Nat Rev Cardiol. 2017;14:273–293.
29. Statistics South Africa. Mid-Year Population Estimates 2013. Pretoria, South Africa: Statistics South Africa; 2013. Available at: Accessed November 10, 2013.
30. Shisana O, Rehle T, Simbayi L, et al. South African National HIV Prevalence, Incidence and Behavoir Survey, 2012. Cape Town, South Africa: Human Sciences Research Council; 2014. Available at: Accessed June 25, 2014.
31. Van Rensburg D. The Free State's approach to implementing the comprehensive plan: notes by a participant outsider. Acta Acad Suppl 2006. 2006;1:44–93.
32. Colvin CJ, Fairall L, Lewin S, et al. Expanding access to ART in South Africa: the role of nurse initiated treatment. SAMJ South Afr Med J. 2010;100:210–212.
33. Uebel KE, Lombard C, Joubert G, et al. Integration of HIV care into primary care in South Africa: effect on survival of patients needing antiretroviral treatment. J Acquir Immune Defic Syndr. 2013;63:e94–e100.
34. Zwarenstein M, Fairall LR, Lombard C, et al. Outreach education for integration of HIV/AIDS care, antiretroviral treatment, and tuberculosis care in primary care clinics in South Africa: PALSA PLUS pragmatic cluster randomised trial. BMJ. 2011;342:d2022.
35. Department of Health. The primary health care package for South Africa—a set of norms and standards. 2000. Available at: Accessed March 16, 2014.
36. Wagner AK, Soumerai SB, Zhang F, et al. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther. 2002;27:299–309. Available at: Accessed March 17, 2014.
37. Lagarde M. How to do (or not to do)... Assessing the impact of a policy change with routine longitudinal data. Health Policy Plan. 2011;27:76–83.
38. Rawat A, Uebel K, Moore D, et al. Health system strengthening through integration of HIV-care into primary health care: a mixed method province-wide study in South Africa [333100]. 2015.
39. Berry KM, Parker W, Mchiza ZJ, et al. Quantifying unmet need for hypertension care in South Africa through a care cascade: evidence from the SANHANES, 2011-2012. BMJ Glob Health. 2017;2:e000348.
40. Pillay-van Wyk V, Dorrington RE, Bradshaw D. Guest editorial: rapidly changing mortality profiles in South Africa in its nine provinces. S Afr Med J. 2017;107:168–169.
41. Atun R, De Jongh T, Secci F, et al. Integration of targeted health interventions into health systems: a conceptual framework for analysis. Health Policy Plan. 2010;25:104.
42. Yu D, Souteyrand Y, Banda MA, et al. Investment in HIV/AIDS programs: does it help strengthen health systems in developing countries? Glob Health. 2008;4:8.
43. Lewin S, Lavis JN, Oxman AD, et al. Supporting the delivery of cost-effective interventions in primary health-care systems in low-income and middle-income countries: an overview of systematic reviews. Lancet. 2008;372:928–939.
44. Topp SM, Chipukuma JM, Chiko MM, et al. Integrating HIV treatment with primary care outpatient services: opportunities and challenges from a scaled-up model in Zambia. Health Policy Plan. 2012;28:347–357.
45. Hyle EP, Naidoo K, Su AE, et al. HIV, tuberculosis, and Noncommunicable diseases: what is known about the costs, effects, and cost-effectiveness of integrated care? J Acquir Immune Defic Syndr. 2014;67:S87–S95.
46. Watt N, Sigfrid L, Legido-Quigley H, et al. Health systems facilitators and barriers to the integration of HIV and chronic disease services: a systematic review. Health Policy Plan. 2017;32(suppl 4):iv13–iv26. doi: 10.1093/heapol/czw149.
47. Topp SM, Abimbola S, Joshi R, et al. How to assess and prepare health systems in low- and middle-income countries for integration of services—a systematic review. Health Policy Plan. 2017. doi: 10.1093/heapol/czx169.
48. Haldane V, Legido-Quigley H, Chuah FL, et al. Integrating cardiovascular diseases, hypertension, and diabetes with HIV services: a systematic review. AIDS Care. 2018;30:103–115.
49. Jacucci E, Shaw V, Braa J. Standardization of health information systems in South Africa: the challenge of local sustainability. Inf Technol Dev. 2006;12:225–239.
50. Gloyd S, Wagenaar BH, Woelk GB, et al. Opportunities and challenges in conducting secondary analysis of HIV programmes using data from routine health information systems and personal health information. J Int AIDS Soc. 2016;19(5 suppl 4):20847.
51. Atun R. Health systems, systems thinking and innovation. Health Policy Plan. 2012;27(suppl 4):iv4–iv8.
52. Gaps in hypertension guidelines in low- and middle-income versus high-income countries. Hypertension. Available at: Accessed August 15, 2017.

integrated service delivery; ART; noncommunicable diseases; NIMART; decentralization; primary health care

Supplemental Digital Content

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.