Gaussian process emulation to improve efficiency of computationally intensive multidisease models: a practical tutorial with adaptable R code.

BACKGROUND: The rapidly growing burden of non-communicable diseases (NCDs) among people living with HIV in sub-Saharan Africa (SSA) has expanded the number of multidisease models predicting future care needs and health system priorities. Usefulness of these models depends on their ability to replicate real-life data and be readily understood and applied by public health decision-makers; yet existing simulation models of HIV comorbidities are computationally expensive and require large numbers of parameters and long run times, which hinders their utility in resource-constrained settings. METHODS: We present a novel, user-friendly emulator that can efficiently approximate complex simulators of long-term HIV and NCD outcomes in Africa. We describe how to implement the emulator via a tutorial based on publicly available data from Kenya. Emulator parameters relating to incidence and prevalence of HIV, hypertension and depression were derived from our own agent-based simulation model and other published literature. Gaussian processes were used to fit the emulator to simulator estimates, assuming presence of noise for design points. Bayesian posterior predictive checks and leave-one-out cross validation confirmed the emulator's descriptive accuracy. RESULTS: In this example, our emulator resulted in a 13-fold (95% Confidence Interval (CI): 8-22) improvement in computing time compared to that of more complex chronic disease simulation models. One emulator run took 3.00 seconds (95% CI: 1.65-5.28) on a 64-bit operating system laptop with 8.00 gigabytes (GB) of Random Access Memory (RAM), compared to > 11 hours for 1000 simulator runs on a high-performance computing cluster with 1500 GBs of RAM. Pareto k estimates were < 0.70 for all emulations, which demonstrates sufficient predictive accuracy of the emulator. CONCLUSIONS: The emulator presented in this tutorial offers a practical and flexible modelling tool that can help inform health policy-making in countries with a generalized HIV epidemic and growing NCD burden. Future emulator applications could be used to forecast the changing burden of HIV, hypertension and depression over an extended (> 10 year) period, estimate longer-term prevalence of other co-occurring conditions (e.g., postpartum depression among women living with HIV), and project the impact of nationally-prioritized interventions such as national health insurance schemes and differentiated care models.

Healthcare workers' experiences with COVID-19-related prevention and control measures in Tanzania.

The ability of a health system to withstand shocks such as a pandemic depends largely on the availability and preparedness of health-care workers (HCWs), who are at the frontline of disease management and prevention. Despite the heavy burden placed on HCWs during the COVID-19 pandemic, little is known regarding their experiences in low-income countries. We conducted a web-based survey with HCWs in randomly selected districts of Tanzania to explore their experiences with COVID-19-related prevention and control measures. The survey assessed implementation of COVID-19 control guidelines in health facilities, HCW perceptions of safety, well-being and ability to provide COVID-19 care, and challenges faced by frontline workers during the pandemic. We used multivariate regression analysis to examine the association between HCW and health facility characteristics, a score of guideline implementation, and challenges faced by HCWs. 6,884 Tanzanian HCWs participated in the survey between December 2021 to March 2022. The majority of respondents were aware of the COVID-19 guidelines and reported implementing preventive measures, including masking of both HCWs and patients. However, HCWs faced several challenges during the pandemic, including increased stress, concerns about infection, and inadequate personal protective equipment. In particular, female HCWs were more likely to report exhaustion from wearing protective equipment and emotional distress, while physicians were more likely to experience all challenges. While most HCWs reported feeling supported by facility management, they also reported that their concerns about COVID-19 treatment were not fully addressed. Notably, perceptions of protection and well-being varied widely among different HCW cadres, highlighting the need for targeted interventions based on level of exposure. In addition, various factors such as HCW cadre, facility ownership and COVID-19 designation status influenced HCWs' opinions about the health system's response to COVID-19. These findings highlight the importance of consistent implementation of guidelines and social and emotional support for HCWs.

Care Cascades for Hypertension in Low-Income Settings: A Systematic Review and Meta-Analysis.

Objective: High blood pressure is the leading risk factor for cardiovascular disease. The hypertension care cascade (HCC) is increasingly being used to evaluate the effectiveness of interventions. This systematic review aims to examine HCC in low-income settings. Methods: The search strategy included articles published between January 2010 and April 2023. We excluded studies with incomplete HCC, on fragile patients or aged <18 years, reviews. We used the MOOSE guideline. Five researchers retrieved data on the survey year, country, population, HCC and diagnostic methods for hypertension. We used JBI Critical Appraisal Tools for quality assessment. Results: Ninety-five articles were analyzed. Average hypertension prevalence was 33% (95% CI: 31%-34%), lower in LICs than in LMICs (25% vs. 34%). The overall mean awareness of hypertension was 48% (95% CI: 45%-51%), its treatment was 35% (95% IC: 32%-38%) and its control 16% (95% CI: 14%-18%). In almost all steps, percentages were lower in LICs and in Sub-Saharan Africa. Conclusion: Trends in HCC vary between countries, with poorer performance in LICs. This review highlights the need for interventions tailored to low-income settings in order to improve hypertension care.

Investigating sustainability challenges for the National Health Insurance Fund in Tanzania: a modelling approach.

OBJECTIVE: This study aimed to model the long-term cost associated with expanding public health insurance coverage in Tanzania. DESIGN, SETTING AND PARTICIPANTS: We analysed the 2016 claims of 2 923 524 beneficiaries of the National Health Insurance Fund in Tanzania. The analysis focused on determining the average cost per beneficiary across 5-year age groups separated by gender, and grouped by broad health condition categories. We then modelled three different insurance coverage scenarios from 2020 to 2050 and we estimated the associated costs. OUTCOME MEASURES: Average cost per beneficiary and the projected financing requirements, projected from 2020 to 2050. RESULTS: The analysis revealed that the average per beneficiary cost for insurance claims was $38.58. Among males over 75 years, the average insurance claims costs were highest, amounting to $125. The total estimated annual cost of claims in 2020 was $151 million. Under the status quo coverage scenario, total claims were projected to increase to $415 million by 2050. Increasing coverage from 7% to 50% would result in an additional financing requirement of $2.27 billion. If coverage would increase by 10% annually, reaching 56% of the population by 2050, the additional financing need would amount to $2.84 billion. CONCLUSION: This study highlights the critical importance of assessing the long-term financial viability of health insurance schemes aimed to cover large segments of the population in low-income countries. The findings demonstrate that even without expansion of coverage, financing requirements for insurance will more than triple by 2050. Furthermore, increasing coverage is likely to substantially escalate the cost of claims, potentially requiring significant government or external contributions to finance these additional costs. Policymakers and stakeholders should carefully evaluate the sustainability of insurance schemes to ensure adequate financial support for expanding coverage and improving healthcare access in low-income settings.

Social-political and vaccine related determinants of COVID-19 vaccine hesitancy in Tanzania: A qualitative inquiry.

Vaccines have played a critical role in the response to the COVID-19 pandemic globally, and Tanzania has made significant efforts to make them available to the public in addition to sensitizing them on its benefit. However, vaccine hesitancy remains a concern. It may prevent optimal uptake of this promising tool in many communities. This study aims to explore opinions and perceptions on vaccine hesitancy to better understand local attitudes towards vaccine hesitancy in both rural and urban Tanzania. The study employed cross-sectional semi-structured interviews with 42 participants. The data were collected in October 2021. Men and women aged between 18 and 70 years were purposefully sampled from Dar es Salaam and Tabora regions. Thematic content analysis was used to categorize data inductively and deductively. We found that COVID-19 vaccine hesitancy exists and is shaped by multiple socio-political and vaccine related factors. Vaccine related factors included worries over vaccine safety (e.g., death, infertility, and zombie), limited knowledge about the vaccines and fear of the vaccine's impact on pre-existing conditions. Participants also found it paradoxical that mask and hygiene mandates are expected even after vaccination, which further exacerbated their doubts about vaccine efficacy and their hesitancy. Participants possessed a range of questions regarding COVID-19 vaccines that they wanted answered by the government. Social factors included preference for traditional and home remedies and influence from others. Political factors included inconsistent messages on COVID-19 from the community and political leaders; and doubts about the existence of COVID-19 and the vaccine. Our findings suggest that the COVID-19 vaccine is beyond a medical intervention, it carries with it a variety of expectations and myths that need to be addressed in order to build trust and acceptance within communities. Health promotion messages need to respond to heterogeneous questions, misinformation, doubts, and concerns over safety issues. An understanding of country-specific perspectives toward COVID-19 vaccines can greatly inform the development of localized strategies for meaningful uptake in Tanzania.

Strengthening primary care for diabetes and hypertension in Eswatini: study protocol for a nationwide cluster-randomized controlled trial.

BACKGROUND: Diabetes and hypertension are increasingly important population health challenges in Eswatini. Prior to this project, healthcare for these conditions was primarily provided through physician-led teams at tertiary care facilities and accessed by only a small fraction of people living with diabetes or hypertension. This trial tests and evaluates two community-based healthcare service models implemented at the national level, which involve health care personnel at primary care facilities and utilize the country's public sector community health worker cadre (the rural health motivators [RHMs]) to help generate demand for care. METHODS: This study is a cluster-randomized controlled trial with two treatment arms and one control arm. The unit of randomization is a primary healthcare facility along with all RHMs (and their corresponding service areas) assigned to the facility. A total of 84 primary healthcare facilities were randomized in a 1:1:1 ratio to the three study arms. The first treatment arm implements differentiated service delivery (DSD) models at the clinic and community levels with the objective of improving treatment uptake and adherence among clients with diabetes or hypertension. In the second treatment arm, community distribution points (CDPs), which previously targeted clients living with human immunodeficiency virus, extend their services to clients with diabetes or hypertension by allowing them to pick up medications and obtain routine nurse-led follow-up visits in their community rather than at the healthcare facility. In both treatment arms, RHMs visit households regularly, screen clients at risk, provide personalized counseling, and refer clients to either primary care clinics or the nearest CDP. In the control arm, primary care clinics provide diabetes and hypertension care services but without the involvement of RHMs and the implementation of DSD models or CDPs. The primary endpoints are mean glycated hemoglobin (HbA1c) and systolic blood pressure among adults aged 40 years and older living with diabetes or hypertension, respectively. These endpoints will be assessed through a household survey in the RHM service areas. In addition to the health impact evaluation, we will conduct studies on cost-effectiveness, syndemics, and the intervention's implementation processes. DISCUSSION: This study has the ambition to assist the Eswatini government in selecting the most effective delivery model for diabetes and hypertension care. The evidence generated with this national-level cluster-randomized controlled trial may also prove useful to policy makers in the wider Sub-Saharan African region. TRIAL REGISTRATION: NCT04183413. Trial registration date: December 3, 2019.

Care cascades for hypertension and diabetes: Cross-sectional evaluation of rural districts in Tanzania.

BACKGROUND: Noncommunicable diseases (NCDs), especially hypertension and diabetes, are rapidly rising in sub-Saharan Africa, necessitating health systems transformations. In Tanzania, current policies aim to improve control of hypertension and diabetes, but information is still needed to assess the gaps in treatment. METHODS AND FINDINGS: We conducted a cross-sectional household survey of 784 adults in two districts in Tanzania from December 2020 to January 2021, capturing the cascade-of-care for hypertension and diabetes. The ages of the respondents ranged from 18 to 89 years. Of those screened positive for these conditions, we measured the proportion in each step of the cascades: awareness, care engagement, treatment, and control. We conducted multivariable logistic regression analyses for all four steps along the hypertension care cascade with the independent variables of social health protection schemes, and prior diagnosis of comorbid diabetes, and demographic information. In our sample, of the 771 who had their blood pressure measured, 41% (95% confidence interval (CI): 38% to 44%) were screened positive for hypertension, and of the 707 who had their blood sugar measured, 6% (95% CI: 4% to 8%) were screened positive for diabetes. Of those with hypertension, 43% (95% CI: 38% to 49%) had a prior diagnosis, 25% (95% CI: 21% to 31%) were engaged in care, 21% (95% CI: 3% to 25%) were on treatment, and 11% (95% CI: 8% to 15%) were controlled. Of the 42 respondents with diabetes, 80% (95% CI: 69% to 93%) had a prior diagnosis. The diabetes care cascade had much less drop-off, so 66% of those with diabetes (95% CI: 52% to 82%) were engaged in care and on treatment, and 48% (95% CI: 32% to 63%) had their diabetes controlled at the point of testing. Healthcare fee exemptions were independently associated with higher odds of being previously diagnosed (OR 5.81; 95% CI [1.98 to 17.10] p < 0.005), engaged in care (OR 4.71; 95% CI [1.59 to 13.90] p 0.005), and retained in treatment (OR 2.93; 95% CI [1.03 to 8.35] p < 0.05). Prior diagnosis of comorbid diabetes was highly associated with higher odds of being engaged in care for hypertension (OR 3.26; 95% CI [1.39 to 7.63] p < 0.005). The two primary limitations of this study were reliance on screening at a single time point only of people available at the village at the time of the sample and dependence on self-report for to inform the three cascade steps of prior diagnosis, healthcare visits for engagement in care, and treatment use. CONCLUSIONS: The high burden of hypertension and low levels of control in our study underscores the importance of improving the awareness and treatment of hypertension. The differences in the care cascades for hypertension and diabetes demonstrates that chronic NCD treatment is possible in this setting, but efforts will be needed across the entire care cascade to improve hypertension control.

Layering and scaling up chronic non-communicable disease care on existing HIV care systems and acute care settings in Kenya: a cost and budget impact analysis.

INTRODUCTION: Like many countries in sub-Saharan Africa, Kenya is experiencing a rapid rise in the burden of non-communicable diseases (NCDs): NCDs now contribute to over 50% of inpatient admissions and 40% of hospital deaths in the country. The Academic Model Providing Access to Healthcare (AMPATH) Chronic Disease Management (CDM) programme builds on lessons and capacity of HIV care to deliver chronic NCD care layered into both HIV and primary care platforms to over 24,000 patients across 69 health facilities in western Kenya. We conducted a cost and budget impact analysis of scaling up the AMPATH CDM programme in western Kenya using the International Society for Pharmacoeconomics and Outcomes Research guidelines. METHODS: Costs of the CDM programme for the health system were measured retrospectively for 69 AMPATH clinics from 2014 to 2018 using programmatic records and clinic schedules to assign per clinic monthly costs. We quantified the additional costs to provide NCD care above those associated with existing HIV or acute care services, including clinician, staff, training, travel and equipment costs, but do not include drugs or consumables as they would be paid by the patient. We projected the budget impact of increasing CDM coverage to 50% of the eligible population from 2021 to 2025, and compared it with the county budgets from 2019. RESULTS: The per visit cost of providing CDM care was $10.42 (SD $2.26), with costs at facilities added to HIV clinics $1.00 (95% CI: -$2:11 to $0.11) lower than at primary care facilities. The budget impact of adding 26,765 patients from 2021 to 2025 to the CDM programme was 3,088,928 under constant percent growth, and 3,451,732 under steady-state enrolment. Scaling up under the constant percent growth scenario resulted in 12% cost savings in the budget impact. The county programmatic CDM cost in 2025 was <1% of the county healthcare budgets from 2019. CONCLUSIONS: The budget impact of scaling up AMPATH's CDM programme will be driven by annual growth scenarios, and facility/provider mix. By leveraging task shifting, referral systems and partnering with public and non-profit clinics without NCD services, AMPATH's CDM programme can provide critical NCD care to new, rural populations with minimal financial impact.

Hypertension Control and Retention in Care Among HIV-Infected Patients: The Effects of Co-located HIV and Chronic Noncommunicable Disease Care.

BACKGROUND: As the noncommunicable disease (NCD) burden is rising in regions with high HIV prevalence, patients with comorbid HIV and chronic NCDs may benefit from integrated chronic disease care. There are few evaluations of the effectiveness of such strategies, especially those that directly leverage and extend the existing HIV care system to provide co-located care for NCDs. SETTING: Academic Model of Providing Access to Healthcare, Kenya, provides care to over 160,000 actively enrolled patients in catchment area of 4 million people. METHODS: Using a difference-in-differences design, we analyzed retrospective clinical records of 3603 patients with comorbid HIV and hypertension during 2009-2016 to evaluate the addition of chronic disease management (CDM) to an existing HIV care program. Outcomes were blood pressure (BP), hypertension control, and adherence to HIV care. RESULTS: Compared with the HIV standard of care, the addition of CDM produced statistically significant, although clinically small improvements in hypertension control, decreasing systolic BP by 0.76 mm Hg (P < 0.001), diastolic BP by 1.28 mm Hg (P < 0.001), and increasing the probability of BP <140/90 mm Hg by 1.51 percentage points (P < 0.001). However, sustained control of hypertension for >1 year improved by 7 percentage points (P < 0.001), adherence to HIV care improved by 6.8 percentage points (P < 0.001) and retention in HIV care with no gaps >6 months increased by 10.5 percentage points (P < 0.001). CONCLUSION: A CDM program that co-locates NCD and HIV care shows potential to improve BP and retention in care. Further evaluation of program implementation across settings can inform how to maximize hypertension control among patients with comorbid HIV, and better understand the effect on adherence.

Epidemiological And Health Systems Implications Of Evolving HIV And Hypertension In South Africa And Kenya.

Health systems in sub-Saharan Africa are facing an ongoing HIV epidemic and increasing burden of noncommunicable disease. With the focus shifting to the development of comprehensive primary health care and chronic disease treatment, multidisease modeling is integral to estimating future health care needs. We extended an established agent-based model of HIV transmission to include hypertension in two rural settings: KwaZulu-Natal, South Africa, and western Kenya. We estimated that from 2018 to 2028 hypertension prevalence would increase from 40 percent to 46 percent in KwaZulu-Natal and from 29 percent to 35 percent in western Kenya, while HIV prevalence is stabilizing and predicted to decrease. As the health system burden in sub-Saharan Africa is changing, innovative chronic disease treatment and the broadening of successful programs, such as integrated HIV and noncommunicable disease care, are necessary to reach universal health care coverage.

Costs and cost-effectiveness of HIV/noncommunicable disease integration in Africa: from theory to practice.

: The current article reviews economic aspects of selected HIV/noncommunicable disease (NCD) service delivery integration programs to assess the efficiency of integration in limited capacity settings. We define economies of scope and scale and their relevance to HIV/NCD integration. We summarize the results of a systematic review of cost and cost-effectiveness studies of integrated care, which identified 12 datasets (nine studies) with a wide range of findings driven by differences in research questions, study methods, and health conditions measured. All studies were done in Africa and examined screening interventions only. No studies assessed the cost of integrated, long-term disease management. Few studies estimated the cost-effectiveness of integrated screening programs. The additional cost of integrating NCD screening with HIV care platforms represented a 6-30% increase in the total costs of the programs for noncancer NCDs, with cervical cancer screening costs dependent on screening strategy. We conducted 11 key informant interviews to uncover perceptions of the economics of HIV/NCD integration. None of the informants had hard information about the economic efficiency of integration. Most expected integrated care to be more cost-effective than current practice, though a minority thought that greater specialization could be more cost-effective. In the final section of this article, we summarize research needs and propose a 'minimum economic dataset' for future studies. We conclude that, although integrated HIV/NCD care has many benefits, the economic justification is unproven. Better information on the cost, cost-effectiveness, and fiscal sustainability of integrated programs is needed to justify this approach in limited-resource countries.

Predictors of completed childhood vaccination in Bolivia.

This project examines how access issues, ethnicity, and geographic region affect vaccination of children by two years of age in Bolivia. Bolivia's rich variation in culture and geography results in unequal healthcare utilization even for basic interventions such as childhood vaccination. This study utilizes secondary data from the 2008 Demographic and Health Survey for Bolivia to examine predictors of vaccination completion in children by two years of age. Using logistic regression methods, we control for health system variables (difficulty getting to a health center and type of health center as well as demographic and socio-economic covariates). The results indicated that children whose parents reported distance as a problem in obtaining health care were less likely to have completed all vaccinations. Ethnicity was not independently statistically significant, however, in a sub-analysis, people from the Quechua ethnic group were more likely to report 'distance as a problem in obtaining healthcare.' Surprisingly, living in a rural environment has a protective effect on completed vaccinations. However, geographic region did predict significant differences in the probability that children would be fully vaccinated; children in the region with the lowest vaccination completion coverage were 80% less likely to have completed vaccination compared to children in the best performing region, which may indicate unequal access and utilization of health services nationally. Further study of regional differences, urbanicity, and distance as a healthcare access problem will help refine implications for the Bolivian health system.