Person:

Salomon, Joshua

Background: This study characterizes the historical relationship between coverage of measles containing vaccines (MCV) and mortality in children under 5 years, with a view toward ongoing global efforts to reduce child mortality. Methodology/Principal Findings: Using country-level, longitudinal panel data, from 44 countries over the period 1960–2005, we analyzed the relationship between MCV coverage and measles mortality with (1) logistic regressions for no measles deaths in a country-year, and (2) linear regressions for the logarithm of the measles death rate. All regressions allowed a flexible, non-linear relationship between coverage and mortality. Covariates included birth rate, death rates from other causes, percent living in urban areas, population density, per-capita GDP, use of the two-dose MCV, year, and mortality coding system. Regressions used lagged covariates, country fixed effects, and robust standard errors clustered by country. The likelihood of no measles deaths increased nonlinearly with higher MCV coverage (ORs: 13.8 [1.6–122.7] for 80–89% to 40.7 [3.2–517.6] for ≥95%), compared to pre-vaccination risk levels. Measles death rates declined nonlinearly with higher MCV coverage, with benefits accruing more slowly above 90% coverage. Compared to no coverage, predicted average reductions in death rates were −79% at 70% coverage, −93% at 90%, and −95% at 95%. Conclusions/Significance: 40 years of experience with MCV vaccination suggests that extremely high levels of vaccination coverage are needed to produce sharp reductions in measles deaths. Achieving sustainable benefits likely requires a combination of extended vaccine programs and supplementary vaccine efforts.

Background: Despite the existence of effective drug treatments, tuberculosis (TB) causes 2 million deaths annually worldwide. Effective treatment is complicated by multidrug-resistant TB (MDR TB) strains that respond only to second-line drugs. We projected the health benefits and cost-effectiveness of using drug susceptibility testing and second-line drugs in a lower-middle-income setting with high levels of MDR TB. Methods and Findings: We developed a dynamic state-transition model of TB. In a base case analysis, the model was calibrated to approximate the TB epidemic in Peru, a setting with a smear-positive TB incidence of 120 per 100,000 and 4.5% MDR TB among prevalent cases. Secondary analyses considered other settings. The following strategies were evaluated: first-line drugs administered under directly observed therapy (DOTS), locally standardized second-line drugs for previously treated cases (STR1), locally standardized second-line drugs for previously treated cases with test-confirmed MDR TB (STR2), comprehensive drug susceptibility testing and individualized treatment for previously treated cases (ITR1), and comprehensive drug susceptibility testing and individualized treatment for all cases (ITR2). Outcomes were costs per TB death averted and costs per quality-adjusted life year (QALY) gained. We found that strategies incorporating the use of second-line drug regimens following first-line treatment failure were highly cost-effective compared to strategies using first-line drugs only. In our base case, standardized second-line treatment for confirmed MDR TB cases (STR2) had an incremental cost-effectiveness ratio of $720 per QALY ($8,700 per averted death) compared to DOTS. Individualized second-line drug treatment for MDR TB following first-line failure (ITR1) provided more benefit at an incremental cost of $990 per QALY ($12,000 per averted death) compared to STR2. A more aggressive version of the individualized treatment strategy (ITR2), in which both new and previously treated cases are tested for MDR TB, had an incremental cost-effectiveness ratio of $11,000 per QALY ($160,000 per averted death) compared to ITR1. The STR2 and ITR1 strategies remained cost-effective under a wide range of alternative assumptions about treatment costs, effectiveness, MDR TB prevalence, and transmission. Conclusions: Treatment of MDR TB using second-line drugs is highly cost-effective in Peru. In other settings, the attractiveness of strategies using second-line drugs will depend on TB incidence, MDR burden, and the available budget, but simulation results suggest that individualized regimens would be cost-effective in a wide range of situations.

Background: Neonatal intensive care improves survival, but is associated with high costs and disability amongst survivors. Recent health reform in Mexico launched a new subsidized insurance program, necessitating informed choices on the different interventions that might be covered by the program, including neonatal intensive care. The purpose of this study was to estimate the clinical outcomes, costs, and cost-effectiveness of neonatal intensive care in Mexico. Methods and Findings: A cost-effectiveness analysis was conducted using a decision analytic model of health and economic outcomes following preterm birth. Model parameters governing health outcomes were estimated from Mexican vital registration and hospital discharge databases, supplemented with meta-analyses and systematic reviews from the published literature. Costs were estimated on the basis of data provided by the Ministry of Health in Mexico and World Health Organization price lists, supplemented with published studies from other countries as needed. The model estimated changes in clinical outcomes, life expectancy, disability-free life expectancy, lifetime costs, disability-adjusted life years (DALYs), and incremental cost-effectiveness ratios (ICERs) for neonatal intensive care compared to no intensive care. Uncertainty around the results was characterized using one-way sensitivity analyses and a multivariate probabilistic sensitivity analysis. In the base-case analysis, neonatal intensive care for infants born at 24–26, 27–29, and 30–33 weeks gestational age prolonged life expectancy by 28, 43, and 34 years and averted 9, 15, and 12 DALYs, at incremental costs per infant of US$11,400, US$9,500, and US$3,000, respectively, compared to an alternative of no intensive care. The ICERs of neonatal intensive care at 24–26, 27–29, and 30–33 weeks were US$1,200, US$650, and US$240, per DALY averted, respectively. The findings were robust to variation in parameter values over wide ranges in sensitivity analyses. Conclusions: Incremental cost-effectiveness ratios for neonatal intensive care imply very high value for money on the basis of conventional benchmarks for cost-effectiveness analysis.

Despite extraordinary global progress in increasing coverage of antiretroviral treatment (ART), the majority of people needing ART currently are not receiving treatment. Both the number of people needing ART and the average ART price per patient-year are expected to increase in coming years, which will dramatically raise funding needs for ART. Several international organizations are using interventions in ART markets to decrease ART price or to improve ART quality, delivery and innovation, with the ultimate goal of improving population health. These organizations need to select those market interventions that are most likely to substantially affect population health outcomes (ex ante assessment) and to evaluate whether implemented interventions have improved health outcomes (ex post assessment). We develop a framework to structure ex ante and ex post assessment of the population health impact of market interventions, which is transmitted through effects in markets and health systems. Ex ante assessment should include evaluation of the safety and efficacy of the ART products whose markets will be affected by the intervention; theoretical consideration of the mechanisms through which the intervention will affect population health; and predictive modelling to estimate the potential population health impact of the intervention. For ex post assessment, analysts need to consider which outcomes to estimate empirically and which to model based on empirical findings and understanding of the economic and biological mechanisms along the causal pathway from market intervention to population health. We discuss methods for ex post assessment and analyse assessment issues (unintended intervention effects, interaction effects between different interventions, and assessment impartiality and cost). We offer seven recommendations for ex ante and ex post assessment of population health impact of market interventions.

Background: The Xpert MTB/RIF test enables rapid detection of tuberculosis (TB) and rifampicin resistance. The World Health Organization recommends Xpert for initial diagnosis in individuals suspected of having multidrug-resistant TB (MDR-TB) or HIV-associated TB, and many countries are moving quickly toward adopting Xpert. As roll-out proceeds, it is essential to understand the potential health impact and cost-effectiveness of diagnostic strategies based on Xpert. Methods and findings: We evaluated potential health and economic consequences of implementing Xpert in five southern African countries—Botswana, Lesotho, Namibia, South Africa, and Swaziland—where drug resistance and TB-HIV coinfection are prevalent. Using a calibrated, dynamic mathematical model, we compared the status quo diagnostic algorithm, emphasizing sputum smear, against an algorithm incorporating Xpert for initial diagnosis. Results were projected over 10- and 20-y time periods starting from 2012. Compared to status quo, implementation of Xpert would avert 132,000 (95% CI: 55,000–284,000) TB cases and 182,000 (97,000–302,000) TB deaths in southern Africa over the 10 y following introduction, and would reduce prevalence by 28% (14%–40%) by 2022, with more modest reductions in incidence. Health system costs are projected to increase substantially with Xpert, by US$460 million (294–699 million) over 10 y. Antiretroviral therapy for HIV represents a substantial fraction of these additional costs, because of improved survival in TB/HIV-infected populations through better TB case-finding and treatment. Costs for treating MDR-TB are also expected to rise significantly with Xpert scale-up. Relative to status quo, Xpert has an estimated cost-effectiveness of US$959 (633–1,485) per disability-adjusted life-year averted over 10 y. Across countries, cost-effectiveness ratios ranged from US$792 (482–1,785) in Swaziland to US$1,257 (767–2,276) in Botswana. Assessing outcomes over a 10-y period focuses on the near-term consequences of Xpert adoption, but the cost-effectiveness results are conservative, with cost-effectiveness ratios assessed over a 20-y time horizon approximately 20% lower than the 10-y values. Conclusions: Introduction of Xpert could substantially change TB morbidity and mortality through improved case-finding and treatment, with more limited impact on long-term transmission dynamics. Despite extant uncertainty about TB natural history and intervention impact in southern Africa, adoption of Xpert evidently offers reasonable value for its cost, based on conventional benchmarks for cost-effectiveness. However, the additional financial burden would be substantial, including significant increases in costs for treating HIV and MDR-TB. Given the fundamental influence of HIV on TB dynamics and intervention costs, care should be taken when interpreting the results of this analysis outside of settings with high HIV prevalence. Please see later in the article for the Editors' Summary

Objective: United Nations Programme on HIV/AIDS reports regularly on estimated levels and trends in HIV/AIDS epidemics, which are evaluated using an epidemiological model within the Estimation and Projection Package (EPP). The relatively simple four-parameter model of HIV incidence used in EPP through the previous round of estimates has encountered challenges when attempting to fit certain data series on prevalence over time, particularly in settings with long running epidemics where prevalence has increased recently. To address this, the most recent version of the modelling package (EPP 2011) includes a more flexible epidemiological model that allows HIV infection risk to vary over time. This paper describes the technical details of this flexible approach to modelling HIV transmission dynamics within EPP 2011. Methodology For the flexible modelling approach, the force of infection parameter, r, is allowed to vary over time through a random walk formulation, and an informative prior distribution is used to improve short-term projections beyond the last year of data. Model parameters are estimated using a Bayesian estimation approach in which models are fit to HIV seroprevalence data from surveillance sites. Results: This flexible model can yield better estimates of HIV prevalence over time in situations where the classic EPP model has difficulties, such as in Uganda, where prevalence is no longer falling. Based on formal out-of-sample projection tests, the flexible modelling approach also improves predictions and CIs for extrapolations beyond the last observed data point. Conclusions: We recommend use of a flexible modelling approach where data are sufficient (eg, where at least 5 years of observations are available), and particularly where an epidemic is beyond its peak.

Objective: We previously developed a flexible specification of the UNAIDS Estimation and Projection Package (EPP) that relied on splines to generate time-varying values for the force of infection parameter. Here, we test the feasibility of this approach for concentrated HIV/AIDS epidemics with very sparse data and compare two methods for making short-term future projections with the spline-based model. Methods: Penalised B-splines are used to model the average infection risk over time within the EPP 2011 modelling framework, which includes antiretroviral treatment effects and CD4 cell count progression, and is fit to sentinel surveillance prevalence data with a Bayesian algorithm. We compare two approaches for future projections: (1) an informative prior related to equilibrium prevalence and (2) a random walk formulation. Results: The spline-based model produced plausible fits across a range of epidemics, which included 87 subpopulations from 14 countries with concentrated epidemics and 75 subpopulations from 33 countries with generalised epidemics. The equilibrium prior and random walk approaches to future projections yielded similar prevalence estimates, and both performed well in tests of out-of-sample predictive validity for prevalence. In contrast, in some cases the two approaches varied substantially in estimates of incidence, with the random walk formulation avoiding extreme changes in incidence. Conclusions: A spline-based approach to allowing the force of infection parameter to vary over time within EPP 2011 is robust across a diverse array of epidemics, including concentrated ones with limited surveillance data. Future work on the EPP model should consider the impact that different modelling approaches have on estimates of HIV incidence.

Objectives: Population-based HIV testing surveys have become central to deriving estimates of national HIV prevalence in sub-Saharan Africa. However, limited participation in these surveys can lead to selection bias. We control for selection bias in national HIV prevalence estimates using a novel approach, which unlike conventional imputation can account for selection on unobserved factors. Methods: For 12 Demographic and Health Surveys conducted from 2001 to 2009 (N=138 300), we predict HIV status among those missing a valid HIV test with Heckman-type selection models, which allow for correlation between infection status and participation in survey HIV testing. We compare these estimates with conventional ones and introduce a simulation procedure that incorporates regression model parameter uncertainty into confidence intervals. Results: Selection model point estimates of national HIV prevalence were greater than unadjusted estimates for 10 of 12 surveys for men and 11 of 12 surveys for women, and were also greater than the majority of estimates obtained from conventional imputation, with significantly higher HIV prevalence estimates for men in Cote d'Ivoire 2005, Mali 2006 and Zambia 2007. Accounting for selective non-participation yielded 95% confidence intervals around HIV prevalence estimates that are wider than those obtained with conventional imputation by an average factor of 4.5. Conclusions: Our analysis indicates that national HIV prevalence estimates for many countries in sub-Saharan African are more uncertain than previously thought, and may be underestimated in several cases, underscoring the need for increasing participation in HIV surveys. Heckman-type selection models should be included in the set of tools used for routine estimation of HIV prevalence.

Background: Increasing evidence supporting the value of screening women for cervical cancer once in their lifetime, coupled with mounting interest in scaling up successful screening demonstration projects, present challenges to public health decision makers seeking to take full advantage of the single-visit opportunity to provide additional services. We present an analytic framework for packaging multiple interventions during a single point of contact, explicitly taking into account a budget and scarce human resources, constraints acknowledged as significant obstacles for provision of health services in poor countries. Methods and Findings: We developed a binary integer programming (IP) model capable of identifying an optimal package of health services to be provided during a single visit for a particular target population. Inputs to the IP model are derived using state-transition models, which compute lifetime costs and health benefits associated with each intervention. In a simplified example of a single lifetime cervical cancer screening visit, we identified packages of interventions among six diseases that maximized disability-adjusted life years (DALYs) averted subject to budget and human resource constraints in four resource-poor regions. Data were obtained from regional reports and surveys from the World Health Organization, international databases, the published literature, and expert opinion. With only a budget constraint, interventions for depression and iron deficiency anemia were packaged with cervical cancer screening, while the more costly breast cancer and cardiovascular disease interventions were not. Including personnel constraints resulted in shifting of interventions included in the package, not only across diseases but also between low- and high-intensity intervention options within diseases. Conclusions: The results of our example suggest several key themes: Packaging other interventions during a one-time visit has the potential to increase health gains; the shortage of personnel represents a real-world constraint that can impact the optimal package of services; and the shortage of different types of personnel may influence the contents of the package of services. Our methods provide a general framework to enhance a decision maker's ability to simultaneously consider costs, benefits, and important nonmonetary constraints. We encourage analysts working on real-world problems to shift from considering costs and benefits of interventions for a single disease to exploring what synergies might be achievable by thinking across disease burdens.

Background: Many mathematical models have investigated the impact of expanding access to antiretroviral therapy (ART) on new HIV infections. Comparing results and conclusions across models is challenging because models have addressed slightly different questions and have reported different outcome metrics. This study compares the predictions of several mathematical models simulating the same ART intervention programmes to determine the extent to which models agree about the epidemiological impact of expanded ART. Methods and Findings: Twelve independent mathematical models evaluated a set of standardised ART intervention scenarios in South Africa and reported a common set of outputs. Intervention scenarios systematically varied the CD4 count threshold for treatment eligibility, access to treatment, and programme retention. For a scenario in which 80% of HIV-infected individuals start treatment on average 1 y after their CD4 count drops below 350 cells/µl and 85% remain on treatment after 3 y, the models projected that HIV incidence would be 35% to 54% lower 8 y after the introduction of ART, compared to a counterfactual scenario in which there is no ART. More variation existed in the estimated long-term (38 y) reductions in incidence. The impact of optimistic interventions including immediate ART initiation varied widely across models, maintaining substantial uncertainty about the theoretical prospect for elimination of HIV from the population using ART alone over the next four decades. The number of person-years of ART per infection averted over 8 y ranged between 5.8 and 18.7. Considering the actual scale-up of ART in South Africa, seven models estimated that current HIV incidence is 17% to 32% lower than it would have been in the absence of ART. Differences between model assumptions about CD4 decline and HIV transmissibility over the course of infection explained only a modest amount of the variation in model results. Conclusions: Mathematical models evaluating the impact of ART vary substantially in structure, complexity, and parameter choices, but all suggest that ART, at high levels of access and with high adherence, has the potential to substantially reduce new HIV infections. There was broad agreement regarding the short-term epidemiologic impact of ambitious treatment scale-up, but more variation in longer term projections and in the efficiency with which treatment can reduce new infections. Differences between model predictions could not be explained by differences in model structure or parameterization that were hypothesized to affect intervention impact.