The economic costs and cost-effectiveness of HIV self-testing among truck drivers in Kenya.

HIV status awareness is critical for ending the HIV epidemic but remains low in high-HIV-risk and hard-to-reach sub-populations. Targeted, efficient interventions are needed to improve HIV test-uptake. We examined the incremental cost-effectiveness of offering the choice of self-administered oral HIV-testing (HIVST-Choice) compared with provider-administered testing only [standard-of-care (SOC)] among long-distance truck drivers. Effectiveness data came from a randomized-controlled trial conducted at two roadside wellness clinics in Kenya (HIVST-Choice arm, n = 150; SOC arm, n = 155). Economic cost data came from the literature, reflected a societal perspective and were reported in 2020 international dollars (I$), a hypothetical currency with equivalent purchasing power as the US dollar. Generalized Poisson and linear gamma regression models were used to estimate effectiveness and incremental costs, respectively; incremental effectiveness was reported as the number of long-distance truck drivers needing to receive HIVST-Choice for an additional HIV test-uptake. We calculated the incremental cost-effectiveness ratio (ICER) of HIVST-Choice compared with SOC and estimated 95% confidence intervals (CIs) using non-parametric bootstrapping. Uncertainty was assessed using deterministic sensitivity analysis and the cost-effectiveness acceptability curve. HIV test-uptake was 23% more likely for HIVST-Choice, with six individuals needing to be offered HIVST-Choice for an additional HIV test-uptake. The mean per-patient cost was nearly 4-fold higher in HIVST-Choice (I$39.28) versus SOC (I$10.80), with an ICER of I$174.51, 95% CI [165.72, 194.59] for each additional test-uptake. HIV self-test kit and cell phone service costs were the main drivers of the ICER, although findings were robust even at highest possible costs. The probability of cost-effectiveness approached 1 at a willingness-to-pay of I$200 for each additional HIV test-uptake. HIVST-Choice improves HIV-test-uptake among truck drivers at low willingness-to-pay thresholds, suggesting that HIV self-testing is an efficient use of resources. Policies supporting HIV self-testing in similar high risk, hard-to-reach sub-populations may expedite achievement of international targets.

Access to Federally Qualified Health Centers and HIV Outcomes in the U.S. South.

INTRODUCTION: Federally Qualified Health Centers may increase access to HIV prevention, care, and treatment for at-risk populations. METHODS: A pooled cross section of ZIP Code Tabulation Areas from cites in the U.S. South with high HIV diagnoses were used to examine Federally Qualified Health Center density and indicators of HIV epidemic control. The explanatory variable was Federally Qualified Health Center density-number of Federally Qualified Health Centers in a ZIP Code Tabulation Areas' Primary Care Service Area per low-income population-high versus medium/low (2019). Outcomes were 5-year (2015-2019 or 2014-2018) (1) number of new HIV diagnoses, (2) percentage late diagnosis, (3) percentage linked to care, and (4) percentage virally suppressed, which was assessed over 1 year (2018 or 2019). Multiple linear regression was used to examine the relationship, including ZIP Code Tabulation Area-level sociodemographic and city-level HIV funding variables, with state-fixed effects, and data analysis was completed in 2022-2023. Sensitivity analyses included (1) examining ZIP Code Tabulation Areas with fewer non-Federally Qualified Health Center primary care providers, (2) controlling for county-level primary care provider density, (3) excluding the highest HIV prevalence ZIP Code Tabulation Areas, and (4) excluding Florida ZIP Code Tabulation Areas. RESULTS: High-density ZIP Code Tabulation Areas had a lower percentage of late diagnosis and virally suppressed, a higher percentage linked to care, and no differences in new HIV diagnoses (p<0.05). In adjusted analysis, high density was associated with a greater number of new diagnoses (number or percentage=5.65; 95% CI=2.81, 8.49), lower percentage of late diagnosis (-3.71%; 95% CI= -5.99, -1.42), higher percentage linked to care (2.13%; 95% CI=0.20, 4.06), and higher percentage virally suppressed (1.87%; 95% CI=0.53, 2.74) than medium/low density. CONCLUSIONS: Results suggest that access to Federally Qualified Health Centers may benefit community-level HIV epidemic indicators.

Disparities in access to primary care, a key site for HIV prevention services, among gay and bisexual men in the United States.

U.S. HIV diagnoses disproportionately affect Non-Hispanic Black (NHB) and Hispanic gay and bisexual men. Using data from the National Health Interview Survey (2013-2018), we examined race and ethnicity and primary care access, an HIV prevention resource, among gay and bisexual men. The explanatory variable was NHB, Hispanic or Non-Hispanic White (NHW). Outcomes were primary care-specific usual place of care (potential access) and saw general doctor <12 months (realized access). We used multivariable logistic regression, adjusting for individual sociodemographic characteristics, health status, and care barriers. In sensitivity analysis, we examined general access (any place/doctor) and subgroups (1) NHB (2) has usual place of care. The sample included 1,858 adult, gay and bisexual men (unweighted). Nearly one-third self-identified as NHB or Hispanic. Compared to NHW men, NHB and Hispanic men were younger, with lower household income, and more care barriers (p < 0.05). NHB and Hispanic men had lower realized access (aOR 0.7058, p = 0.030) than NHW men. Potential access was lower for NHB only (versus NHW) and, among those with any usual place of care, NHB and Hispanic men versus NHW men. Lower primary care access for NHB or Hispanic, rather than NHW, gay and bisexual men, may reduce HIV prevention access.

Economic costs and cost-effectiveness of conditional cash transfers for the uptake of services for the prevention of vertical HIV transmissions in a resource-limited setting.

BACKGROUND: Prevention of mother-to-child transmission (PMTCT) is critical for halting the HIV epidemic. However, innovative approaches to improve PMTCT uptake may be resource-intensive. We examined the economic costs and cost-effectiveness of conditional cash transfers (CCTs) for the uptake of PMTCT services in the Democratic Republic of Congo. METHODS: We leveraged data from a randomized controlled trial of CCTs (n = 216) versus standard PMTCT care alone (standard of care (SOC), n = 217). Economic cost data came from multiple sources, with costs analyzed from the societal perspective and reported in 2016 international dollars (I$). Effectiveness outcomes included PMTCT uptake (i.e., accepting all PMTCT visits and services) and retention (i.e., in HIV care at six weeks post-partum). Generalized estimating equations estimated effectiveness (relative risk) and incremental costs, with incremental effectiveness reported as the number of women needing CCTs for an additional PMTCT uptake or retention. We evaluated the cost-effectiveness of the CCTs at various levels of willingness-to-pay and assessed uncertainty using deterministic sensitivity analysis and cost-effectiveness acceptability curves. RESULTS: Mean costs per participant were I$516 (CCTs) and I$431 (SOC), representing an incremental cost of I$85 (95% CI: 59, 111). PMTCT uptake was more likely for CCTs vs SOC (68% vs 53%, p < 0.05), with seven women needing CCTs for each additional PMTCT service uptake; twelve women needed CCTs for an additional PMTCT retention. The incremental cost-effectiveness of CCTs vs SOC was I$595 (95% CI: I$550, I$638) for PMTCT uptake and I$1028 (95% CI: I$931, I$1125) for PMTCT retention. CCTs would be an efficient use of resources if society's willingness-to-pay for an additional woman who takes up PMTCT services is at least I$640. In the worst-case scenario, the findings remained relatively robust. CONCLUSIONS: Given the relatively low cost of the CCTs, policies supporting CCTs may decrease onward HIV transmission and expedite progress toward ending the epidemic.

Urban-rural disparities in geographic accessibility to care for people living with HIV.

In the United States, people living with HIV (PLWH) in rural areas fare worse along the HIV care continuum than their urban counterparts; this may be due in part to limited geographic access to care. We estimated drive time to care for PLWH, focusing on urban-rural differences. Adult Medicaid enrollees living with HIV and their usual care clinicians were identified using administrative claims data from 14 states (Medicaid Analytic eXtract, 2009-2012). We used geographic network analysis to calculate one-way drive time from the enrollee's ZIP code tabulation area centroid to their clinician's practice address, then examined urban-rural differences using bivariate statistics. Additional analyses included altering the definition of rurality; examining subsamples based on the state of residence, services received, and clinician specialty; and adjusting for individual and county characteristics. Across n = 49,596 PLWH, median drive time to care was 12.8 min (interquartile range 26.3). Median drive time for rural enrollees (43.6 (82.0)) was nearly four times longer than for urban enrollees (11.9 (20.6) minutes, p < 0.0001), and drive times exceeded one hour for 38% of rural enrollees (versus 12% of urban, p < 0.0001). Urban-rural disparities remained in all additional analyses. Sustained efforts to circumvent limited geographic access to care are critical for rural areas.

Brief Report: Physician Reimbursement and Retention in HIV Care: Racial Disparities in the US South.

BACKGROUND: Retention in HIV care remains a national challenge. Addressing structural barriers to care may improve retention. We examined the association between physician reimbursement and retention in HIV care, including racial differences. METHODS: We integrated person-level administrative claims (Medicaid Analytic eXtract, 2008-2012), state Medicaid-to-Medicare physician fee ratios (Urban Institute, 2008, 2012), and county characteristics for 15 Southern states plus District of Columbia. The fee ratio is a standardized measure of physician reimbursement capturing Medicaid relative to Medicare physician reimbursement across states. Generalized estimating equations assessed the association between the fee ratio and retention (≥2 care markers ≥90 days apart in a calendar year). Stratified analyses assessed racial differences. We varied definitions of retention, subsamples, and definitions of the fee ratio, including the fee ratio at parity. RESULTS: The sample included 55,237 adult Medicaid enrollees with HIV (179,002 enrollee years). Enrollees were retained in HIV care for 76.6% of their enrollment years, with retention lower among non-Hispanic Black (76.1%) versus non-Hispanic White enrollees (81.3%, P < 0.001). A 10-percentage point increase in physician reimbursement was associated with 4% increased odds of retention (adjusted odds ratio 1.04, 95% confidence interval: 1.01 to 1.07). In stratified analyses, the positive, significant association occurred among non-Hispanic Black (1.08, 1.05-1.12) but not non-Hispanic White enrollees (0.87, 0.74-1.02). Findings were robust across sensitivity analyses. When the fee ratio reached parity, predicted retention increased significantly overall and for non-Hispanic Black enrollees. CONCLUSION: Higher physician reimbursement may improve retention in HIV care, particularly among non-Hispanic Black individuals, and could be a mechanism to promote health equity.

Human Immunodeficiency Virus-Experienced Clinician Workforce Capacity: Urban-Rural Disparities in the Southern United States.

BACKGROUND: Human immunodeficiency virus (HIV)-experienced clinicians are critical for positive outcomes along the HIV care continuum. However, access to HIV-experienced clinicians may be limited, particularly in nonmetropolitan areas, where HIV is increasing. We examined HIV clinician workforce capacity, focusing on HIV experience and urban-rural differences, in the Southern United States. METHODS: We used Medicaid claims and clinician characteristics (Medicaid Analytic eXtract [MAX] and MAX Provider Characteristics, 2009-2011), county-level rurality (National Center for Health Statistics, 2013), and diagnosed HIV cases (AIDSVu, 2014) to assess HIV clinician capacity in 14 states. We assumed that clinicians accepting Medicaid approximated the region's HIV workforce, since three-quarters of clinicians accept Medicaid insurance. HIV-experienced clinicians were defined as those providing care to ≥ 10 Medicaid enrollees over 3 years. We assessed HIV workforce capacity with county-level clinician-to-population ratios, using Wilcoxon-Mann-Whitney tests to compare urban-rural differences. RESULTS: We identified 5012 clinicians providing routine HIV management, of whom 28% were HIV-experienced. HIV-experienced clinicians were more likely to specialize in infectious diseases (48% vs 6%, P < .001) and practice in urban areas (96% vs 83%, P < .001) compared to non-HIV-experienced clinicians. The median clinician-to-population ratio for all HIV clinicians was 13.3 (interquartile range, 38.0), with no significant urban-rural differences. When considering HIV experience, 81% of counties had no HIV-experienced clinicians, and rural counties generally had fewer HIV-experienced clinicians per 1000 diagnosed HIV cases (P < .001). CONCLUSIONS: Significant urban-rural disparities exist in HIV-experienced workforce capacity for communities in the Southern United States. Policies to improve equity in access to HIV-experienced clinical care for both urban and rural communities are urgently needed.

Implementation of "Treat-all" at adult HIV care and treatment sites in the Global IeDEA Consortium: results from the Site Assessment Survey.

INTRODUCTION: Since 2015, the World Health Organization (WHO) has recommended that all people living with HIV (PLHIV) initiate antiretroviral treatment (ART), irrespective of CD4+ count or clinical stage. National adoption of universal treatment has accelerated since WHO's 2015 "Treat All" recommendation; however, little is known about the translation of this guidance into practice. This study aimed to assess the status of Treat All implementation across regions, countries, and levels of the health care delivery system. METHODS: Between June and December 2017, 201/221 (91%) adult HIV treatment sites that participate in the global IeDEA research consortium completed a survey on capacity and practices related to HIV care. Located in 41 countries across seven geographic regions, sites provided information on the status and timing of site-level introduction of Treat All, as well as site-level practices related to ART initiation. RESULTS: Almost all sites (93%) reported that they had begun implementing Treat All, and there were no statistically significant differences in site-level Treat All introduction by health facility type, urban/rural location, sector (public/private) or country income level. The median time between national policy adoption and site-level introduction was one month. In countries where Treat All was not yet adopted in national guidelines, 69% of sites reported initiating all patients on ART, regardless of clinical criteria, and these sites had been implementing Treat All for a median period of seven months at the time of the survey. The majority of sites (77%) reported typically initiating patients on ART within 14 days of confirming diagnosis, with 60% to 62% of sites implementing Treat All in East, Southern and West Africa reporting same-day ART initiation for most patients. CONCLUSIONS: By mid- to late-2017, the Treat All strategy was the standard of care at almost all IeDEA sites, including rural, primary-level health facilities in low-resource settings. While further assessments of site-level capacity to provide high-quality HIV care under Treat All and to support sustained viral suppression after ART initiation are needed, the widespread introduction of Treat All at the service delivery level is a critical step towards global targets for ending the HIV epidemic as a public health threat.

The effects of community-based distribution of family planning services on contraceptive use: The case of a national scale-up in Malawi.

RATIONALE: Although many sub-Saharan African countries have scaled-up pilot projects of community-based distribution (CBD) of family planning services, the effects of the scaled-up CBDs on contraceptive use remain unclear. OBJECTIVE: We leveraged a national scale-up of Malawi's Learning and Innovation Population and Family Planning pilot to evaluate the effects of a scale-up of CBDs on contraceptive use. We also investigated whether education and income, two important determinants of contraceptive use behaviors, moderate the effects of the scaled-up CBDs. METHOD: We used the 2000/2004 and 2010/2016 Malawi Demographic and Health Surveys (N = 57,978) and difference-in-differences analyses to estimate the effects of the 2005 national scale-up of CBDs on modern contraceptive use. We used rural and urban communities as the intervention and comparison groups because the national CBDs were implemented only in rural communities. Contraceptive use is defined as the current use of any modern contraceptive method (e.g., pills) and was modelled using multilevel logistic regression. RESULTS: Prior to the implementation of the national scale CBDs (2000/2004), the probability of using contraceptives was 21.5% in rural communities and 26.3% in urban communities. After the scale-up (2010/2016), the probability of using contraceptives increased in both rural and urban communities but was greater in rural communities (44.9% vs. 42.9%). The effect attributable to the national scale CBDs was 6.8 percentage points (95% CI [3.3, 9.7]). The effects of the national CBDs were greater among uneducated and low-income women. CONCLUSIONS: These findings suggest that national CBDs increase overall contraceptive use, particularly in rural communities, and that poor and uneducated women benefit more from family planning interventions that reduce communication and financial barriers.

Suboptimal geographic accessibility to comprehensive HIV care in the US: regional and urban-rural differences.

Achieving US state and municipal benchmarks to end the HIV epidemic and promote health equity requires access to comprehensive HIV care. However, this care may not be geographically accessible for all people living with HIV (PLHIV). We estimated county-level drive time and suboptimal geographic accessibility to HIV care across the contiguous US, assessing regional and urban-rural differences. We integrated publicly available data from four federal databases to identify and geocode sites providing comprehensive HIV care in 2015, defined as the co-located provision of core HIV medical care and support services. Leveraging street network, US Census and HIV surveillance data (2014), we used geographic analysis to estimate the fastest one-way drive time between the population-weighted county centroid and the nearest site providing HIV care for counties reporting at least five diagnosed HIV cases. We summarized HIV care sites, county-level drive time, population-weighted drive time and suboptimal geographic accessibility to HIV care, by US region and county rurality (2013). Geographic accessibility to HIV care was suboptimal if drive time was >30 min, a common threshold for primary care accessibility in the general US population. Tests of statistical significance were not performed, since the analysis is population-based. We identified 671 HIV care sites across the US, with 95% in urban counties. Nationwide, the median county-level drive time to HIV care is 69 min (interquartile range (IQR) 66 min). The median county-level drive time to HIV care for rural counties (90 min, IQR 61) is over twice that of urban counties (40 min, IQR 48), with the greatest urban-rural differences in the West. Nationally, population-weighted drive time, an approximation of individual-level drive time, is over five times longer in rural counties than in urban counties. Geographic access to HIV care is suboptimal for over 170,000 people diagnosed with HIV (19%), with over half of these individuals from the South and disproportionately the rural South. Nationally, approximately 80,000 (9%) drive over an hour to receive HIV care. Suboptimal geographic accessibility to HIV care is an important structural barrier in the US, particularly for rural residents living with HIV in the South and West. Targeted policies and interventions to address this challenge should become a priority.

The impact of an insecticide treated bednet campaign on all-cause child mortality: A geospatial impact evaluation from the Democratic Republic of Congo.

OBJECTIVE: To test the impact of a nationwide Long-Lasting Insecticidal Nets [LLINs] distribution program in the Democratic Republic of Congo [DRC] on all-cause under-five child mortality exploiting subnational variation in malaria endemicity and the timing in the scale-up of the program across provinces. DESIGN: Geospatial Impact Evaluation using a difference-in-differences approach. SETTING: Democratic Republic of the Congo. PARTICIPANTS: 52,656 children sampled in the 2007 and 2013/2014 DRC Demographic and Health Surveys. INTERVENTIONS: The analysis provides plausibly causal estimates of both average treatment effects of the LLIN distribution campaign and geospatial heterogeneity in these effects based on malaria endemicity. It compares the under-five, all-cause mortality for children pre- and post-LLIN campaign relative to children in those areas that had not yet been exposed to the campaign using a difference-in-differences model and controlling for year- and province-fixed effects, and province-level trends in mortality. RESULTS: We find that the campaign led to a 41% decline [3.7 percentage points, 95% CI 1.3 to 6.0] in under-5 mortality risk among children living in rural areas with malaria ecology above the sample median. Results were robust to controlling for household assets and the presence of other health aid programs. No effect was detected in children living in areas with malaria ecology below the median. CONCLUSION: The findings of this paper make important contributions to the evidence base for the effectiveness of large scale-national LLIN campaigns against malaria. We found that the program was effective in areas of the DRC with the highest underlying risk of malaria. Targeting bednets to areas with greatest underlying risk for malaria may help to increase the efficiency of increasingly limited malaria resources but should be balanced against other malaria control concerns.

Research priorities to inform "Treat All" policy implementation for people living with HIV in sub-Saharan Africa: a consensus statement from the International epidemiology Databases to Evaluate AIDS (IeDEA).

INTRODUCTION: "Treat All" - the treatment of all people with HIV, irrespective of disease stage or CD4 cell count - represents a paradigm shift in HIV care that has the potential to end AIDS as a public health threat. With accelerating implementation of Treat All in sub-Saharan Africa (SSA), there is a need for a focused agenda and research to identify and inform strategies for promoting timely uptake of HIV treatment, retention in care, and sustained viral suppression and addressing bottlenecks impeding implementation. METHODS: The Delphi approach was used to develop consensus around research priorities for Treat All implementation in SSA. Through an iterative process (June 2017 to March 2018), a set of research priorities was collectively formulated and refined by a technical working group and shared for review, deliberation and prioritization by more than 200 researchers, implementation experts, policy/decision-makers, and HIV community representatives in East, Central, Southern and West Africa. RESULTS AND DISCUSSION: The process resulted in a list of nine research priorities for generating evidence to guide Treat All policies, implementation strategies and monitoring efforts. These priorities highlight the need for increased focus on adolescents, men, and those with mental health and substance use disorders - groups that remain underserved in SSA and for whom more effective testing, linkage and care strategies need to be identified. The priorities also reflect consensus on the need to: (1) generate accurate national and sub-national estimates of the size of key populations and describe those who remain underserved along the HIV-care continuum; (2) characterize the timeliness of HIV care and short- and long-term HIV care continuum outcomes, as well as factors influencing timely achievement of these outcomes; (3) estimate the incidence and prevalence of HIV-drug resistance and regimen switching; and (4) identify cost-effective and affordable service delivery models and strategies to optimize uptake and minimize gaps, disparities, and losses along the HIV-care continuum, particularly among underserved populations. CONCLUSIONS: Reflecting consensus among a broad group of experts, researchers, policy- and decision-makers, PLWH, and other stakeholders, the resulting research priorities highlight important evidence gaps that are relevant for ministries of health, funders, normative bodies and research networks.

Mathematical modelling to inform 'treat all' implementation in sub-Saharan Africa: a scoping review.

OBJECTIVE: Despite widespread uptake, only half of sub-Saharan African countries have fully implemented the World Health Organization's 'treat all' policy, hindering achievement of global HIV targets. We examined literature on mathematical modelling studies that sought to inform scale-up and implementation of 'treat all' in sub-Saharan Africa. METHODS: We conducted a scoping review, a research synthesis to assess emerging evidence and identify gaps, of peer-reviewed literature, extracting study characteristics on 'treat all' policies and assumptions, setting, key populations, outcomes and findings. Studies were narratively summarised and potential gaps characterised. RESULTS: We identified 16 studies examining 'treat all' alone (n=12) or with expanded testing (n=7) and/or care continuum improvements (n=6). Twelve studies examined 'treat all' for Southern African countries, while none did so for Central Africa. Four included the role of resistance; one evaluated any key population. A range of health and economic outcomes were reported, although fewer studies formally assessed budget impact. Fourteen studies involved co-authors with any in-country affiliation; one study also had co-authors with local government affiliation. Overall, 'treat all' improves health outcomes and is cost-effective compared to deferred HIV treatment; 'treat all' with expanded testing or care continuum improvements may provide further health benefits. However, studies generally used optimistic assumptions about the implementation of expanded testing or care continuum improvements. CONCLUSIONS: The modelling literature demonstrates improved health and economic benefits of 'treat all'. Using mathematical modelling to inform real-world implementation of 'treat all' requires realistic assumptions about expanded testing and care continuum interventions across a wide range of settings and populations.

Structural barriers to comprehensive, coordinated HIV care: geographic accessibility in the US South.

Structural barriers to HIV care are particularly challenging in the US South, which has higher HIV diagnosis rates, poverty, uninsurance, HIV stigma, and rurality, and fewer comprehensive public health programs versus other US regions. Focusing on one structural barrier, we examined geographic accessibility to comprehensive, coordinated HIV care (HIVCCC) in the US South. We integrated publicly available data to study travel time to HIVCCC in 16 Southern states and District of Columbia. We geocoded HIVCCC service locations and estimated drive time between the population-weighted county centroid and closest HIVCCC facility. We evaluated drive time in aggregate, and by county-level HIV prevalence quintile, urbanicity, and race/ethnicity. Optimal drive time was ≤30 min, a common primary care accessibility threshold. We identified 228 service locations providing HIVCCC across 1422 Southern counties, with median drive time to care of 70 min (IQR 64 min). For 368 counties in the top HIV prevalence quintile, median drive time is 50 min (IQR 61 min), exceeding 60 min in over one-third of these counties. Among counties in the top HIV prevalence quintile, drive time to care is six-folder higher for rural versus super-urban counties. Counties in the top HIV prevalence quintiles for non-Hispanic Blacks and for Hispanics have >50% longer drive time to care versus for non-Hispanic Whites. Including another potential care source-publicly-funded health centers serving low-income populations-could double the number of high-HIV burden counties with drive time ≤30 min, representing nearly 35,000 additional people living with HIV with accessible HIVCCC. Geographic accessibility to HIVCCC is inadequate in the US South, even in high HIV burden areas, and geographic and racial/ethnic disparities exist. Structural factors, such as geographic accessibility to care, may drive disparities in health outcomes. Further research on programmatic policies, and evidence-based alternative HIV care delivery models improving access to care, is critical.

Implementing parallel spreadsheet models for health policy decisions: The impact of unintentional errors on model projections.

BACKGROUND: Spreadsheet software is increasingly used to implement systems science models informing health policy decisions, both in academia and in practice where technical capacity may be limited. However, spreadsheet models are prone to unintentional errors that may not always be identified using standard error-checking techniques. Our objective was to illustrate, through a methodologic case study analysis, the impact of unintentional errors on model projections by implementing parallel model versions. METHODS: We leveraged a real-world need to revise an existing spreadsheet model designed to inform HIV policy. We developed three parallel versions of a previously validated spreadsheet-based model; versions differed by the spreadsheet cell-referencing approach (named single cells; column/row references; named matrices). For each version, we implemented three model revisions (re-entry into care; guideline-concordant treatment initiation; immediate treatment initiation). After standard error-checking, we identified unintentional errors by comparing model output across the three versions. Concordant model output across all versions was considered error-free. We calculated the impact of unintentional errors as the percentage difference in model projections between model versions with and without unintentional errors, using +/-5% difference to define a material error. RESULTS: We identified 58 original and 4,331 propagated unintentional errors across all model versions and revisions. Over 40% (24/58) of original unintentional errors occurred in the column/row reference model version; most (23/24) were due to incorrect cell references. Overall, >20% of model spreadsheet cells had material unintentional errors. When examining error impact along the HIV care continuum, the percentage difference between versions with and without unintentional errors ranged from +3% to +16% (named single cells), +26% to +76% (column/row reference), and 0% (named matrices). CONCLUSIONS: Standard error-checking techniques may not identify all errors in spreadsheet-based models. Comparing parallel model versions can aid in identifying unintentional errors and promoting reliable model projections, particularly when resources are limited.

Comprehensive Ryan White Assistance and Human Immunodeficiency Virus Clinical Outcomes: Retention in Care and Viral Suppression in a Medicaid Nonexpansion State.

Background: Knowledge gaps remain about how the Ryan White human immunodeficiency virus (HIV)/AIDS Program (RW) contributes to health outcomes. We examined the association between different RW service classes and retention in care (RiC) or viral suppression (VS). Methods: We identified Virginians engaged in any HIV care between 1 January and 31 December 2014. RW beneficiaries were classified by receipt of ≥1 service from 3 classes: Core medical, Support, and insurance and/or direct medication assistance through the AIDS Drug Assistance Program (ADAP). Receipt of all RW classes was defined as comprehensive assistance. We used multivariable logistic regression to compare the odds of RiC and of VS by comprehensive assistance and by RW classes alone and in combination. Results: Among 13104 individuals, 58% received any RW service and 17% comprehensive assistance. Comprehensive assistance is significantly associated with RiC (adjusted odds ratio [aOR], 8.8 [95% confidence interval {CI}, 7.2-10.8]) and viral suppression (aOR, 3.3 [95% CI, 2.9-3.8]). Receiving any 2 RW classes or Core alone is significantly associated with RiC and VS, with the strength of association decreasing as the number of classes decreases. Recipients of Support alone are significantly less likely to have VS (aOR, 0.75 [95% CI, .59-.96]). For ADAP recipients also receiving Core and/or Support, insurance assistance is significantly associated with VS compared to receiving direct medication only (aOR, 1.6 [95% CI, 1.3-1.9]); this relationship is not significant for those who receive ADAP alone. Conclusions: Receiving more classes of RW-funded services is associated with improved HIV outcomes. For some populations with insurance, RW-funded services may still be required for optimal health outcomes.

Clinical outcomes of HIV care delivery models in the US: a systematic review.

With over 1 million people living with HIV, the US faces national challenges in HIV care delivery due to an inadequate HIV specialist workforce and the increasing role of non-communicable chronic diseases in driving morbidity and mortality in HIV-infected patients. Alternative HIV care delivery models, which include substantial roles for advanced practitioners and/or coordination between specialty and primary care settings in managing HIV-infected patients, may address these needs. We aimed to systematically review the evidence on patient-level HIV-specific and primary care health outcomes for HIV-infected adults receiving outpatient care across HIV care delivery models. We identified randomized trials and observational studies from bibliographic and other databases through March 2016. Eligible studies met pre-specified eligibility criteria including on care delivery models and patient-level health outcomes. We considered all available evidence, including non-experimental studies, and evaluated studies for risk of bias. We identified 3605 studies, of which 13 met eligibility criteria. Of the 13 eligible studies, the majority evaluated specialty-based care (9 studies). Across all studies and care delivery models, eligible studies primarily reported mortality and antiretroviral use, with specialty-based care associated with mortality reductions at the clinician and practice levels and with increased antiretroviral initiation or use at the clinician level but not the practice level. Limited and heterogeneous outcomes were reported for other patient-level HIV-specific outcomes (e.g., viral suppression) as well as for primary care health outcomes across all care delivery models. No studies addressed chronic care outcomes related to aging. Limited evidence was available across geographic settings and key populations. As re-design of care delivery in the US continues to evolve, better understanding of patient-level HIV-related and primary care health outcomes, especially across different staffing models and among different patient populations and geographic locations, is urgently needed to improve HIV disease management.

Performance of a mathematical model to forecast lives saved from HIV treatment expansion in resource-limited settings.

BACKGROUND: International guidelines recommend HIV treatment expansion in resource-limited settings, but funding availability is uncertain. We evaluated the performance of a model that forecasts lives saved through continued HIV treatment expansion in Haiti. METHODS: We developed a computer-based, mathematical model of HIV disease and used incidence density analysis of patient-level Haitian data to derive model parameters for HIV disease progression. We assessed the internal validity of model predictions and internally calibrated model inputs when model predictions did not fit the patient-level data. We then derived uncertain model inputs related to diagnosis and linkage to care, pretreatment retention, and enrollment on HIV treatment through an external calibration process that selected input values by comparing model predictions to Haitian population-level data. Model performance was measured by fit to event-free survival (patient level) and number receiving HIV treatment over time (population level). RESULTS: For a cohort of newly HIV-infected individuals with no access to HIV treatment, the model predicts median AIDS-free survival of 9.0 years precalibration and 6.6 years postcalibration v. 5.8 years (95% confidence interval, 5.1-7.0) from the patient-level data. After internal validation and calibration, 16 of 17 event-free survival measures (94%) had a mean percentage deviation between model predictions and the empiric data of <6%. After external calibration, the percentage deviation between model predictions and population-level data on the number on HIV treatment was <1% over time. CONCLUSIONS: Validation and calibration resulted in a good-fitting model appropriate for health policy decision making. Using local data in a policy model-building process is feasible in resource-limited settings.