Novel bioinformatic methods and machine learning approaches reveal candidate biomarkers of the intensity and timing of past exposure to Plasmodium falciparum.

Accurately quantifying the burden of malaria over time is an important goal of malaria surveillance efforts and can enable effective targeting and evaluation of interventions. Malaria surveillance methods capture active or recent infections which poses several challenges to achieving malaria surveillance goals. In high transmission settings, asymptomatic infections are common and therefore accurate measurement of malaria burden demands active surveillance; in low transmission regions where infections are rare accurate surveillance requires sampling large subsets of the population; and in any context monitoring malaria burden over time necessitates serial sampling. Antibody responses to Plasmodium falciparum parasites persist after infection and therefore measuring antibodies has the potential to overcome several of the current obstacles to accurate malaria surveillance. Identifying which antibody responses are markers of the timing and intensity of past exposure to P. falciparum remains challenging, particularly among adults who tend to be re-exposed multiple times over the course of their lifetime and therefore have similarly high antibody responses to many Plasmodium antigens. A previous analysis of 479 serum samples from individuals in three regions in southern Africa with different historical levels of P. falciparum malaria transmission (high, intermediate, and low) revealed regional differences in antibody responses to P. falciparum antigens among children under 5 years of age. Using a novel bioinformatic pipeline optimized for protein microarrays that minimizes between-sample technical variation, we used antibody responses to Plasmodium antigens as predictors in random forest models to classify samples from adults into these three regions of differing historical malaria transmission with high accuracy (AUC = 0.99). Many of the most important antigens for classification in these models do not overlap with previously published results and are therefore novel candidate markers for the timing and intensity of past exposure to P. falciparum. Measuring antibody responses to these antigens could lead to improved malaria surveillance.

A Bayesian hierarchical model for signal extraction from protein microarrays.

Protein microarrays are a promising technology that measure protein levels in serum or plasma samples. Due to their high technical variability and high variation in protein levels across serum samples in any population, directly answering biological questions of interest using protein microarray measurements is challenging. Analyzing preprocessed data and within-sample ranks of protein levels can mitigate the impact of between-sample variation. As for any analysis, ranks are sensitive to preprocessing, but loss function based ranks that accommodate major structural relations and components of uncertainty are very effective. Bayesian modeling with full posterior distributions for quantities of interest produce the most effective ranks. Such Bayesian models have been developed for other assays, for example, DNA microarrays, but modeling assumptions for these assays are not appropriate for protein microarrays. Consequently, we develop and evaluate a Bayesian model to extract the full posterior distribution of normalized protein levels and associated ranks for protein microarrays, and show that it fits well to data from two studies that use protein microarrays produced by different manufacturing processes. We validate the model via simulation and demonstrate the downstream impact of using estimates from this model to obtain optimal ranks.

Whole Blood Transfusion for Severe Malarial Anemia in a High Plasmodium falciparum Transmission Setting.

BACKGROUND: Severe malaria resulting from Plasmodium falciparum infection is the leading parasitic cause of death in children worldwide, and severe malarial anemia (SMA) is the most common clinical presentation. The evidence in support of current blood transfusion guidelines for patients with SMA is limited. METHODS: We conducted a retrospective cohort study of 911 hospitalized children with SMA in a holoendemic region of Zambia to examine the association of whole blood transfusion with in-hospital survival. Data were analyzed in adjusted logistic regression models using multiple imputation for missing data. RESULTS: The median age of patients was 24 months (interquartile range, 16-30) and overall case fatality was 16%. Blood transfusion was associated with 35% reduced odds of death in children with SMA (odds ratio, 0.65; 95% confidence interval, .52-.81; P = .0002) corresponding to a number-needed-to-treat (NNT) of 14 patients. Children with SMA complicated by thrombocytopenia were more likely to benefit from transfusion than those without thrombocytopenia (NNT = 5). Longer storage time of whole blood was negatively associated with survival and with the posttransfusion rise in the platelet count but was not associated with the posttransfusion change in hemoglobin concentration. CONCLUSIONS: Whole blood given to pediatric patients with SMA was associated with improved survival, mainly among those with thrombocytopenia who received whole blood stored for <4 weeks. These findings point to a potential use for incorporating thrombocytopenia into clinical decision making and management of severe malaria, which can be further assessed in prospective studies, and underline the importance of maintaining reliable blood donation networks in areas of high malaria transmission.

Accurate assignment of disease liability to genetic variants using only population data.

PURPOSE: The growing size of public variant repositories prompted us to test the accuracy of pathogenicity prediction of DNA variants using population data alone. METHODS: Under the a priori assumption that the ratio of the prevalence of variants in healthy population vs that in affected populations form 2 distinct distributions (pathogenic and benign), we used a Bayesian method to assign probability to a variant belonging to either distribution. RESULTS: The approach, termed Bayesian prevalence ratio (BayPR), accurately parsed 300 of 313 expertly curated CFTR variants: 284 of 296 pathogenic/likely pathogenic variants in 1 distribution and 16 of 17 benign/likely benign variants in another. BayPR produced an area under the receiver operating characteristic curve of 0.99 for 103 functionally confirmed missense CFTR variants, which is equal to or exceeds 10 commonly used algorithms (area under the receiver operating characteristic curve range = 0.54-0.99). Application of BayPR to expertly curated variants in 8 genes associated with 7 Mendelian conditions led to the assignment of a disease-causing probability of ≥80% to 1350 of 1374 (98.3%) pathogenic/likely pathogenic variants and of ≤20% to 22 of 23 (95.7%) benign/likely benign variants. CONCLUSION: Irrespective of the variant type or functional effect, the BayPR approach provides probabilities of pathogenicity for DNA variants responsible for Mendelian disorders using only the variant counts in affected and unaffected population samples.

A pre-processing pipeline to quantify, visualize, and reduce technical variation in protein microarray studies.

Technical variation, or variation from non-biological sources, is present in most laboratory assays. Correcting for this variation enables analysts to extract a biological signal that informs questions of interest. However, each assay has different sources and levels of technical variation, and the choice of correction methods can impact downstream analyses. Compared to similar assays such as DNA microarrays, relatively few methods have been developed and evaluated for protein microarrays, a versatile tool for measuring levels of various proteins in serum samples. Here, we propose a pre-processing pipeline to correct for some common sources of technical variation in protein microarrays. The pipeline builds upon an existing normalization method by using controls to reduce technical variation. We evaluate our method using data from two protein microarray studies and by simulation. We demonstrate that pre-processing choices impact the fluorescent-intensity based ranks of proteins, which in turn, impact downstream analysis.

Assessing and communicating heterogeneity of treatment effects for patient subpopulations: Panel discussion on considerations in design and analysis.

Clinical trials are primarily conducted to understand the average effects treatments have on patients. However, patients are heterogeneous in the severity of the condition and in ways that affect what treatment effect they can expect. It is therefore important to understand and characterize how treatment effects vary. The design and analysis of clinical studies play critical roles in evaluating and characterizing heterogeneous treatment effects. This panel discussed considerations in design and analysis under the recognition that there are heterogeneous treatment effects across subgroups of patients. Panel members discussed many questions including: What is a good estimate of the treatment effect in me, a 65-year-old, bald, Caucasian-American, male patient? What magnitude of heterogeneity of treatment effects (HTE) is sufficiently large to merit attention? What role can prior evidence about HTE play in confirmatory trial design and analysis? Is there anything described in the 21st Century Cures Act that would benefit from greater attention to HTE? An example of a Bayesian approach addressing multiplicity when testing for treatment effects in subgroups will be provided. We can do more or better at understanding heterogeneous treatment effects and providing the best information on heterogeneous treatment effects.

Individualized treatment effects with censored data via fully nonparametric Bayesian accelerated failure time models.

Individuals often respond differently to identical treatments, and characterizing such variability in treatment response is an important aim in the practice of personalized medicine. In this article, we describe a nonparametric accelerated failure time model that can be used to analyze heterogeneous treatment effects (HTE) when patient outcomes are time-to-event. By utilizing Bayesian additive regression trees and a mean-constrained Dirichlet process mixture model, our approach offers a flexible model for the regression function while placing few restrictions on the baseline hazard. Our nonparametric method leads to natural estimates of individual treatment effect and has the flexibility to address many major goals of HTE assessment. Moreover, our method requires little user input in terms of model specification for treatment covariate interactions or for tuning parameter selection. Our procedure shows strong predictive performance while also exhibiting good frequentist properties in terms of parameter coverage and mitigation of spurious findings of HTE. We illustrate the merits of our proposed approach with a detailed analysis of two large clinical trials (N = 6769) for the prevention and treatment of congestive heart failure using an angiotensin-converting enzyme inhibitor. The analysis revealed considerable evidence for the presence of HTE in both trials as demonstrated by substantial estimated variation in treatment effect and by high proportions of patients exhibiting strong evidence of having treatment effects which differ from the overall treatment effect.

Cost of behavioral weight loss programs implemented in clinical practice: The POWER trial at Johns Hopkins.

Obesity presents an important public health problem that affects more than a third of the U.S. adult population and that is associated with increased morbidity, mortality, and costs. Previously, we documented that two primary care-based weight loss interventions were clinically effective. To encourage the implementation of and reimbursement for these interventions, we evaluated their relative cost-effectiveness. We performed a cost analysis of the Practice-based Opportunities for Weight Reduction (POWER) trial, a three-arm trial that enrolled 415 patients with obesity from six primary care practices. Trial participants were randomized to a control arm, an in-person support intervention, or a remote support intervention; in the two intervention arms, behavioral interventions were delivered over 24 months, in two phases. Weight loss was measured at 6, 12, and 24 months. Using timesheets and empirical data, we evaluated the cost of the in-person and remote support interventions from the perspective of a health care system delivering the interventions. A univariate sensitivity analysis was conducted to evaluate uncertainty around model assumptions. All comparisons were tested using independent t-tests. Cost of the in-person intervention was higher at 6 months ($113 per participant per month and $117 per kg lost) than the remote support intervention ($101 per participant per month and $99 per kg lost; p < .001). Costs were also higher for the in-person support intervention at 24 months ($73 per participant per month and $342 per kg lost) than for the remote support intervention ($53 per participant per month and $275 per kg lost; p < .001). In the sensitivity analyses, cost ranged from $274/kg lost to $456/kg lost for the in-person support intervention and from $218/kg to $367/kg lost for the remote support intervention. A primary care weight loss intervention administered remotely was relatively more cost-effective than an in-person intervention. Expanding the scope of reimbursable programs to include other cost-effective interventions could help ensure that a broader range of patients receive the type of support needed.

Comparative Study of Confidence Intervals for Proportions in Complex Sample Surveys.

The most widespread method of computing confidence intervals (CIs) in complex surveys is to add and subtract the margin of error (MOE) from the point estimate, where the MOE is the estimated standard error multiplied by the suitable Gaussian quantile. This Wald-type interval is used by the American Community Survey (ACS), the largest US household sample survey. For inferences on small proportions with moderate sample sizes, this method often results in marked under-coverage and lower CI endpoint less than 0. We assess via simulation the coverage and width, in complex sample surveys, of seven alternatives to the Wald interval for a binomial proportion with sample size replaced by the 'effective sample size,' that is, the sample size divided by the design effect. Building on previous work by the present authors, our simulations address the impact of clustering, stratification, different stratum sampling fractions, and stratum-specific proportions. We show that all intervals undercover when there is clustering and design effects are computed from a simple design-based estimator of sampling variance. Coverage can be better calibrated for the alternatives to Wald by improving estimation of the effective sample size through superpopulation modeling. This approach is more effective in our simulations than previously proposed modifications of effective sample size. We recommend intervals of the Wilson or Bayes uniform prior form, with the Jeffreys prior interval not far behind.

Bayesian Estimation of MSM Population Size in Côte d'Ivoire.

Côte d'Ivoire has among the most generalized HIV epidemics in West Africa with an estimated half million people living with HIV. Across West Africa, key populations, including gay men and other men who have sex with men (MSM), are often disproportionately burdened with HIV due to specific acquisition and transmission risks. Quantifying population sizes of MSM at the subnational level is critical to ensuring evidence-based decisions regarding the scale and content of HIV prevention interventions. While survey-based direct estimates of MSM numbers are available in a few urban centers across Côte d'Ivoire, no data on MSM population size exists in other areas without any community group infrastructure to facilitate sufficient access to communities of MSM. The data are used in a Bayesian regression setup to produce estimates of the numbers of MSM in areas of Côte d'Ivoire prioritized in the HIV response. Our hierarchical model imputes missing covariates using geo-spatial information and allows for proper uncertainty quantification leading to confidence bounds for predicted MSM population size estimates. This process provided population size estimates where there are no empirical data, to guide the prioritization of further collection of empirical data on MSM and inform evidence-based scaling of HIV prevention and treatment programs for MSM across Côte d'Ivoire.

The Role of Expert Judgment in Statistical Inference and Evidence-Based Decision-Making.

This article resulted from our participation in the session on the "role of expert opinion and judgment in statistical inference" at the October 2017 ASA Symposium on Statistical Inference. We present a strong, unified statement on roles of expert judgment in statistics with processes for obtaining input, whether from a Bayesian or frequentist perspective. Topics include the role of subjectivity in the cycle of scientific inference and decisions, followed by a clinical trial and a greenhouse gas emissions case study that illustrate the role of judgments and the importance of basing them on objective information and a comprehensive uncertainty assessment. We close with a call for increased proactivity and involvement of statisticians in study conceptualization, design, conduct, analysis, and communication.

Subnational estimation of modern contraceptive prevalence in five sub-Saharan African countries: a Bayesian hierarchical approach.

BACKGROUND: Global monitoring efforts have relied on national estimates of modern contraceptive prevalence rate (mCPR) for many low-income countries. However, most contraceptive delivery programs are implemented by health departments at lower administrative levels, reflecting a persisting gap between the availability of and need for subnational mCPR estimates. METHODS: Using woman-level data from multiple semi-annual national survey rounds conducted between 2013 and 2016 in five sub-Saharan African countries (Burkina Faso, Ethiopia, Ghana, Kenya, and Uganda) by the Performance, Monitoring and Accountability 2020 project, we propose a Bayesian Hierarchical Model with a standard set of covariates and temporally correlated random effects to estimate the level and trend of mCPR for first level administrative divisions in each country. RESULTS: There is considerable narrowing of the uncertainty interval (UI) around the model-based estimates, compared to the estimates directly based on the survey data. We find substantial variations in the estimated subnational mCPRs. Uganda, for example, shows a gain in mCPR of 6.4% (95% UI: 4.5-8.3) based on model estimates of 20.9% (19.6-22.2) in mid-2014 and 27.3% (26.0-28.8) in mid-2016, with change across 10 regions ranging from - 0.6 points in Karamoja to 9.4 points in Central 2 region. The lower bound of the UIs of the change over four rounds was above 0 in 6 regions. Similar upward trends are observed for most regions in the other four countries, and there is noticeable within-country geographic variation. CONCLUSIONS: Reliable subnational estimates of mCPR empower health departments in evidence-based policy making. Despite nationally increasing mCPRs, regional disparities exist within countries suggesting uneven contraceptive access. Raising investments in disadvantaged areas may be warranted to increase equity in access to modern contraceptive methods.

Mathematical Modeling of "Chronic" Infectious Diseases: Unpacking the Black Box.

BACKGROUND: Mathematical models are increasingly used to understand the dynamics of infectious diseases, including "chronic" infections with long generation times. Such models include features that are obscure to most clinicians and decision-makers. METHODS: Using a model of a hypothetical active case-finding intervention for tuberculosis in India as an example, we illustrate the effects on model results of different choices for model structure, input parameters, and calibration process. RESULTS: Using the same underlying data, different transmission models produced different estimates of the projected intervention impact on tuberculosis incidence by 2030 with different corresponding uncertainty ranges. We illustrate the reasons for these differences and present a simple guide for clinicians and decision-makers to evaluate models of infectious diseases. CONCLUSIONS: Mathematical models of chronic infectious diseases must be understood to properly inform policy decisions. Improved communication between modelers and consumers is critical if model results are to improve the health of populations.

Perfluoroalkyl Chemicals, Menstrual Cycle Length, and Fecundity: Findings from a Prospective Pregnancy Study.

BACKGROUND: Perfluoroalkyl substances have been associated with changes in menstrual cycle characteristics and fecundity, when modeled separately. However, these outcomes are biologically related, and we evaluate their joint association with exposure to perfluoroalkyl substances. METHODS: We recruited 501 couples from Michigan and Texas in 2005-2009 upon their discontinuing contraception and followed them until pregnancy or 12 months of trying. Female partners provided a serum sample on enrollment and completed daily journals on menstruation, intercourse, and pregnancy test results. We measured seven perfluoroalkyl substances in serum using liquid chromatography-tandem mass spectrometry. We assessed the association between perfluoroalkyl substances and menstrual cycle length using accelerated failure time models and between perfluoroalkyl substances and fecundity using a Bayesian joint modeling approach to incorporate cycle length. RESULTS: Menstrual cycles were 3% longer comparing women in the second versus first tertile of perfluorodecanoate (PFDeA; acceleration factor [AF] = 1.03, 95% credible interval [CrI] = [1.00, 1.05]), but 2% shorter for women in the highest versus lowest tertile of perfluorooctanoic acid (PFOA; AF = 0.98, 95% CrI = [0.96, 1.00]). When accounting for cycle length, relevant covariates, and remaining perfluoroalkyl substances, the probability of pregnancy was lower for women in second versus first tertile of perfluorononanoate (PFNA; odds ratio [OR] = 0.6, 95% CrI = [0.4, 1.0]) although not when comparing the highest versus lowest (OR = 0.7, 95% CrI = [0.3, 1.1]) tertile. CONCLUSIONS: In this prospective cohort study, we observed associations between two perfluoroalkyl substances and menstrual cycle length changes, and between select perfluoroalkyl substances and diminished fecundity at some (but not all) concentrations. See video abstract at, http://links.lww.com/EDE/B136.

Bayesian analysis of heterogeneous treatment effects for patient-centered outcomes research.

Evaluation of heterogeneity of treatment effect (HTE) is an essential aspect of personalized medicine and patient-centered outcomes research. Our goal in this article is to promote the use of Bayesian methods for subgroup analysis and to lower the barriers to their implementation by describing the ways in which the companion software beanz can facilitate these types of analyses. To advance this goal, we describe several key Bayesian models for investigating HTE and outline the ways in which they are well-suited to address many of the commonly cited challenges in the study of HTE. Topics highlighted include shrinkage estimation, model choice, sensitivity analysis, and posterior predictive checking. A case study is presented in which we demonstrate the use of the methods discussed.

Cataract Surgery Outcomes in Uveitis: The Multicenter Uveitis Steroid Treatment Trial.

PURPOSE: To assess the visual outcomes of cataract surgery in eyes that received fluocinolone acetonide implant or systemic therapy with oral corticosteroids and immunosuppression during the Multicenter Uveitis Steroid Treatment (MUST) Trial. DESIGN: Nested prospective cohort study of patients enrolled in a randomized clinical trial. PARTICIPANTS: Patients that underwent cataract surgery during the first 2 years of follow-up in the MUST Trial. METHODS: Visual outcomes of cataract surgery were evaluated 3, 6, and 9 months after surgery using logarithmic visual acuity charts. Change in visual acuity over time was assessed using a mixed-effects model. MAIN OUTCOME MEASURES: Best-corrected visual acuity. RESULTS: After excluding eyes that underwent cataract surgery simultaneously with implant surgery, among the 479 eyes in the MUST Trial, 117 eyes (28 eyes in the systemic, 89 in the implant group) in 82 patients underwent cataract surgery during the first 2 years of follow-up. Overall, visual acuity increased by 23 letters from the preoperative visit to the 3-month visit (95% confidence interval [CI], 17-29 letters; P < 0.001) and was stable through 9 months of follow-up. Eyes presumed to have a more severe cataract, as measured by inability to grade vitreous haze, gained an additional 42 letters (95% CI, 34-56 letters; P < 0.001) beyond the 13-letter gain in eyes that had gradable vitreous haze before surgery (95% CI, 9-18 letters; P < 0.001) 3 months after surgery, making up for an initial difference of -45 letters at the preoperative visit (95% CI, -56 to -34 letters; P < 0.001). Black race, longer time from uveitis onset, and hypotony were associated with worse preoperative visual acuity (P < 0.05), but did not affect postsurgical recovery (P > 0.05, test of interaction). After adjusting for other risk factors, there was no significant difference in the improvement in visual acuity between the 2 treatment groups (implant vs. systemic therapy, 2 letters; 95% CI, -10 to 15 letters; P = 0.70). CONCLUSIONS: Cataract surgery resulted in substantial, sustained, and similar visual acuity improvement in the eyes of patients with uveitis treated with the fluocinolone acetonide implant or standard systemic therapy.

A Bayesian joint model of menstrual cycle length and fecundity.

Menstrual cycle length (MCL) has been shown to play an important role in couple fecundity, which is the biologic capacity for reproduction irrespective of pregnancy intentions. However, a comprehensive assessment of its role requires a fecundity model that accounts for male and female attributes and the couple's intercourse pattern relative to the ovulation day. To this end, we employ a Bayesian joint model for MCL and pregnancy. MCLs follow a scale multiplied (accelerated) mixture model with Gaussian and Gumbel components; the pregnancy model includes MCL as a covariate and computes the cycle-specific probability of pregnancy in a menstrual cycle conditional on the pattern of intercourse and no previous fertilization. Day-specific fertilization probability is modeled using natural, cubic splines. We analyze data from the Longitudinal Investigation of Fertility and the Environment Study (the LIFE Study), a couple based prospective pregnancy study, and find a statistically significant quadratic relation between fecundity and menstrual cycle length, after adjustment for intercourse pattern and other attributes, including male semen quality, both partner's age, and active smoking status (determined by baseline cotinine level 100 ng/mL). We compare results to those produced by a more basic model and show the advantages of a more comprehensive approach.