ABSTRACT
The conditional survival function of a time-to-event outcome subject to censoring and truncation is a common target of estimation in survival analysis. This parameter may be of scientific interest and also often appears as a nuisance in nonparametric and semiparametric problems. In addition to classical parametric and semiparametric methods (e.g., based on the Cox proportional hazards model), flexible machine learning approaches have been developed to estimate the conditional survival function. However, many of these methods are either implicitly or explicitly targeted toward risk stratification rather than overall survival function estimation. Others apply only to discrete-time settings or require inverse probability of censoring weights, which can be as difficult to estimate as the outcome survival function itself. Here, we employ a decomposition of the conditional survival function in terms of observable regression models in which censoring and truncation play no role. This allows application of an array of flexible regression and classification methods rather than only approaches that explicitly handle the complexities inherent to survival data. We outline estimation procedures based on this decomposition, empirically assess their performance, and demonstrate their use on data from an HIV vaccine trial. Supplementary materials for this article are available online.
ABSTRACT
In the absence of data from a randomized trial, researchers may aim to use observational data to draw causal inference about the effect of a treatment on a time-to-event outcome. In this context, interest often focuses on the treatment-specific survival curves, that is, the survival curves were the population under study to be assigned to receive the treatment or not. Under certain conditions, including that all confounders of the treatment-outcome relationship are observed, the treatment-specific survival curve can be identified with a covariate-adjusted survival curve. In this article, we propose a novel cross-fitted doubly-robust estimator that incorporates data-adaptive (e.g. machine learning) estimators of the conditional survival functions. We establish conditions on the nuisance estimators under which our estimator is consistent and asymptotically linear, both pointwise and uniformly in time. We also propose a novel ensemble learner for combining multiple candidate estimators of the conditional survival estimators. Notably, our methods and results accommodate events occurring in discrete or continuous time, or an arbitrary mix of the two. We investigate the practical performance of our methods using numerical studies and an application to the effect of a surgical treatment to prevent metastases of parotid carcinoma on mortality.
ABSTRACT
OBJECTIVE: We aimed to evaluate the feasibility and utility of an unsupervised testing mechanism, in which participants pick up a swab kit, self-test (unsupervised) and return the kit to an on-campus drop box, as compared with supervised self-testing at staffed locations. DESIGN: University SARS-CoV-2 testing cohort. SETTING: Husky Coronavirus Testing provided voluntary SARS-CoV-2 testing at a university in Seattle, USA. OUTCOME MEASURES: We computed descriptive statistics to describe the characteristics of the study sample. Adjusted logistic regression implemented via generalised estimating equations was used to estimate the odds of a self-swab being conducted through unsupervised versus supervised testing mechanisms by participant characteristics, including year of study enrolment, pre-Omicron versus post-Omicron time period, age, sex, race, ethnicity, affiliation and symptom status. RESULTS: From September 2021 to July 2022, we received 92 499 supervised and 26 800 unsupervised self-swabs. Among swabs received by the laboratory, the overall error rate for supervised versus unsupervised swabs was 0.3% vs 4%, although this declined to 2% for unsupervised swabs by the spring of the academic year. Results were returned for 92 407 supervised (5% positive) and 25 836 unsupervised (4%) swabs from 26 359 participants. The majority were students (79%), 61% were female and most identified as white (49%) or Asian (34%). The use of unsupervised testing increased during the Omicron wave when testing demand was high and stayed constant in spring 2022 even when testing demand fell. We estimated the odds of using unsupervised versus supervised testing to be significantly greater among those <25 years of age (p<0.001), for Hispanic versus non-Hispanic individuals (OR 1.2, 95% CI 1.0 to 1.3, p=0.01) and lower among individuals symptomatic versus asymptomatic or presymptomatic (0.9, 95% CI 0.8 to 0.9, p<0.001). CONCLUSIONS: Unsupervised swab collection permitted increased testing when demand was high, allowed for access to a broader proportion of the university community and was not associated with a substantial increase in testing errors.
Subject(s)
COVID-19 Testing , COVID-19 , SARS-CoV-2 , Specimen Handling , Humans , COVID-19/diagnosis , COVID-19/epidemiology , Female , Male , Adult , Universities , COVID-19 Testing/methods , COVID-19 Testing/statistics & numerical data , Middle Aged , Young Adult , Specimen Handling/methods , Cohort Studies , Washington/epidemiology , Self-Testing , Adolescent , Aged , Pandemics , Feasibility StudiesABSTRACT
Individualizing treatment assignment can improve outcomes for diseases with patient-to-patient variability in comparative treatment effects. When a clinical trial demonstrates that some patients improve on treatment while others do not, it is tempting to assume that treatment effect heterogeneity exists. However, if outcome variability is mainly driven by factors other than variability in the treatment effect, investigating the extent to which covariate data can predict differential treatment response is a potential waste of resources. Motivated by recent meta-analyses assessing the potential of individualizing treatment for major depressive disorder using only summary statistics, we provide a method that uses summary statistics widely available in published clinical trial results to bound the benefit of optimally assigning treatment to each patient. We also offer alternate bounds for settings in which trial results are stratified by another covariate. Our upper bounds can be especially informative when they are small, as there is then little benefit to collecting additional covariate data. We demonstrate our approach using summary statistics from a depression treatment trial. Our methods are implemented in the rct2otrbounds R package.
Subject(s)
Depressive Disorder, Major , Precision Medicine , Humans , Depressive Disorder, Major/drug therapy , Depressive Disorder, Major/therapy , Precision Medicine/methods , Treatment Outcome , Data Interpretation, Statistical , Clinical Trials as Topic , Randomized Controlled Trials as Topic , Models, Statistical , Antidepressive Agents/therapeutic useABSTRACT
A central goal of vaccine research is to characterize and validate immune correlates of protection (CoPs). In addition to helping elucidate immunological mechanisms, a CoP can serve as a valid surrogate endpoint for an infectious disease clinical outcome and thus qualifies as a primary endpoint for vaccine authorization or approval without requiring resource-intensive randomized, controlled phase 3 trials. Yet, it is challenging to persuasively validate a CoP, because a prognostic immune marker can fail as a reliable basis for predicting/inferring the level of vaccine efficacy against a clinical outcome, and because the statistical analysis of phase 3 trials only has limited capacity to disentangle association from cause. Moreover, the multitude of statistical methods garnered for CoP evaluation in phase 3 trials renders the comparison, interpretation, and synthesis of CoP results challenging. Toward promoting broader harmonization and standardization of CoP evaluation, this article summarizes four complementary statistical frameworks for evaluating CoPs in a phase 3 trial, focusing on the frameworks' distinct scientific objectives as measured and communicated by distinct causal vaccine efficacy parameters. Advantages and disadvantages of the frameworks are considered, dependent on phase 3 trial context, and perspectives are offered on how the frameworks can be applied and their results synthesized.
Subject(s)
Vaccine Efficacy , Vaccines , Research Design , Biomarkers/analysis , Causality , Randomized Controlled Trials as TopicABSTRACT
BACKGROUND: SARS-CoV-2 antigen-detection rapid diagnostic tests (Ag-RDTs) have become widely utilized but longitudinal characterization of their community-based performance remains incompletely understood. METHODS: This prospective longitudinal study at a large public university in Seattle, WA utilized remote enrollment, online surveys, and self-collected nasal swab specimens to evaluate Ag-RDT performance against real-time reverse transcription polymerase chain reaction (rRT-PCR) in the context of SARS-CoV-2 Omicron. Ag-RDT sensitivity and specificity within 1 day of rRT-PCR were evaluated by symptom status throughout the illness episode and Orf1b cycle threshold (Ct). RESULTS: From February to December 2022, 5757 participants reported 17 572 Ag-RDT results and completed 12 674 rRT-PCR tests, of which 995 (7.9%) were rRT-PCR positive. Overall sensitivity and specificity were 53.0% (95% confidence interval [CI], 49.6%-56.4%) and 98.8% (95% CI, 98.5%-99.0%), respectively. Sensitivity was comparatively higher for Ag-RDTs used 1 day after rRT-PCR (69.0%), 4-7 days after symptom onset (70.1%), and Orf1b Ct ≤20 (82.7%). Serial Ag-RDT sensitivity increased with repeat testing ≥2 (68.5%) and ≥4 (75.8%) days after an initial Ag-RDT-negative result. CONCLUSIONS: Ag-RDT performance varied by clinical characteristics and temporal testing patterns. Our findings support recommendations for serial testing following an initial Ag-RDT-negative result, especially among recently symptomatic persons or those at high risk for SARS-CoV-2 infection.
Subject(s)
COVID-19 Serological Testing , COVID-19 , SARS-CoV-2 , Sensitivity and Specificity , Humans , COVID-19/diagnosis , SARS-CoV-2/immunology , SARS-CoV-2/isolation & purification , SARS-CoV-2/genetics , Prospective Studies , Longitudinal Studies , Male , Female , Middle Aged , Adult , COVID-19 Serological Testing/methods , Antigens, Viral/analysis , COVID-19 Nucleic Acid Testing/methods , Aged , Washington , Young Adult , AdolescentABSTRACT
Vaccine effectiveness (VE) studies utilizing the test-negative design are typically conducted in clinical settings, rather than community populations, leading to bias in VE estimates against mild disease and limited information on VE in healthy young adults. In a community-based university population, we utilized data from a large SARS-CoV-2 testing program to estimate relative VE of COVID-19 mRNA vaccine primary series and monovalent booster dose versus primary series only against symptomatic SARS-CoV-2 infection from September 2021 to July 2022. We used the test-negative design and logistic regression implemented via generalized estimating equations adjusted for age, calendar time, prior SARS-CoV-2 infection, and testing frequency (proxy for test-seeking behavior) to estimate relative VE. Analyses included 2,218 test-positive cases (59 % received monovalent booster dose) and 9,615 test-negative controls (62 %) from 9,066 individuals, with median age of 21 years, mostly students (71 %), White (56 %) or Asian (28 %), and with few comorbidities (3 %). More cases (23 %) than controls (6 %) had COVID-19-like illness. Estimated adjusted relative VE of primary series and monovalent booster dose versus primary series only against symptomatic SARS-CoV-2 infection was 40 % (95 % CI: 33-47 %) during the overall analysis period and 46 % (39-52 %) during the period of Omicron circulation. Relative VE was greater for those without versus those with prior SARS-CoV-2 infection (41 %, 34-48 % versus 33 %, 9 %-52 %, P < 0.001). Relative VE was also greater in the six months after receiving a booster dose (41 %, 33-47 %) compared to more than six months (27 %, 8-42 %), but this difference was not statistically significant (P = 0.06). In this relatively young and healthy adult population, an mRNA monovalent booster dose provided increased protection against symptomatic SARS-CoV-2 infection, overall and with the Omicron variant. University testing programs may be utilized for estimating VE in healthy young adults, a population that is not well-represented by routine VE studies.
Subject(s)
COVID-19 Vaccines , COVID-19 , Young Adult , Humans , Adult , COVID-19/epidemiology , COVID-19/prevention & control , COVID-19 Testing , Universities , SARS-CoV-2 , RNA, MessengerABSTRACT
BACKGROUND: While epidemiologic evidence links higher levels of exposure to fine particulate matter (PM2.5) to decreased cognitive function, fewer studies have investigated links with traffic-related air pollution (TRAP), and none have examined ultrafine particles (UFP, ≤100 nm) and late-life dementia incidence. OBJECTIVE: To evaluate associations between TRAP exposures (UFP, black carbon [BC], and nitrogen dioxide [NO2]) and late-life dementia incidence. METHODS: We ascertained dementia incidence in the Seattle-based Adult Changes in Thought (ACT) prospective cohort study (beginning in 1994) and assessed ten-year average TRAP exposures for each participant based on prediction models derived from an extensive mobile monitoring campaign. We applied Cox proportional hazards models to investigate TRAP exposure and dementia incidence using age as the time axis and further adjusting for sex, self-reported race, calendar year, education, socioeconomic status, PM2.5, and APOE genotype. We ran sensitivity analyses where we did not adjust for PM2.5 and other sensitivity and secondary analyses where we adjusted for multiple pollutants, applied alternative exposure models (including total and size-specific UFP), modified the adjustment covariates, used calendar year as the time axis, assessed different exposure periods, dementia subtypes, and others. RESULTS: We identified 1,041 incident all-cause dementia cases in 4,283 participants over 37,102 person-years of follow-up. We did not find evidence of a greater hazard of late-life dementia incidence with elevated levels of long-term TRAP exposures. The estimated hazard ratio of all-cause dementia was 0.98 (95 % CI: 0.92-1.05) for every 2000 pt/cm3 increment in UFP, 0.95 (0.89-1.01) for every 100 ng/m3 increment in BC, and 0.96 (0.91-1.02) for every 2 ppb increment in NO2. These findings were consistent across sensitivity and secondary analyses. DISCUSSION: We did not find evidence of a greater hazard of late-life dementia risk with elevated long-term TRAP exposures in this population-based prospective cohort study.
Subject(s)
Air Pollutants , Air Pollution , Dementia , Adult , Humans , Air Pollutants/analysis , Air Pollution/analysis , Environmental Exposure/analysis , Prospective Studies , Nitrogen Dioxide/analysis , Incidence , Particulate Matter/analysis , Dementia/epidemiologyABSTRACT
In many applications, it is of interest to assess the relative contribution of features (or subsets of features) toward the goal of predicting a response - in other words, to gauge the variable importance of features. Most recent work on variable importance assessment has focused on describing the importance of features within the confines of a given prediction algorithm. However, such assessment does not necessarily characterize the prediction potential of features, and may provide a misleading reflection of the intrinsic value of these features. To address this limitation, we propose a general framework for nonparametric inference on interpretable algorithm-agnostic variable importance. We define variable importance as a population-level contrast between the oracle predictiveness of all available features versus all features except those under consideration. We propose a nonparametric efficient estimation procedure that allows the construction of valid confidence intervals, even when machine learning techniques are used. We also outline a valid strategy for testing the null importance hypothesis. Through simulations, we show that our proposal has good operating characteristics, and we illustrate its use with data from a study of an antibody against HIV-1 infection.
ABSTRACT
The COVE trial randomized participants to receive two doses of mRNA-1273 vaccine or placebo on Days 1 and 29 (D1, D29). Anti-SARS-CoV-2 Spike IgG binding antibodies (bAbs), anti-receptor binding domain IgG bAbs, 50% inhibitory dilution neutralizing antibody (nAb) titers, and 80% inhibitory dilution nAb titers were measured at D29 and D57. We assessed these markers as correlates of protection (CoPs) against COVID-19 using stochastic interventional vaccine efficacy (SVE) analysis and principal surrogate (PS) analysis, frameworks not used in our previous COVE immune correlates analyses. By SVE analysis, hypothetical shifts of the D57 Spike IgG distribution from a geometric mean concentration (GMC) of 2737 binding antibody units (BAU)/mL (estimated vaccine efficacy (VE): 92.9% (95% CI: 91.7%, 93.9%)) to 274 BAU/mL or to 27,368 BAU/mL resulted in an overall estimated VE of 84.2% (79.0%, 88.1%) and 97.6% (97.4%, 97.7%), respectively. By binary marker PS analysis of Low and High subgroups (cut-point: 2094 BAU/mL), the ignorance interval (IGI) and estimated uncertainty interval (EUI) for VE were [85%, 90%] and (78%, 93%) for Low compared to [95%, 96%] and (92%, 97%) for High. By continuous marker PS analysis, the IGI and 95% EUI for VE at the 2.5th percentile (519.4 BAU/mL) vs. at the 97.5th percentile (9262.9 BAU/mL) of D57 Spike IgG concentration were [92.6%, 93.4%] and (89.2%, 95.7%) vs. [94.3%, 94.6%] and (89.7%, 97.0%). Results were similar for other D29 and D57 markers. Thus, the SVE and PS analyses additionally support all four markers at both time points as CoPs.
Subject(s)
2019-nCoV Vaccine mRNA-1273 , COVID-19 , Humans , Antibodies, Neutralizing , Antibodies, Viral , COVID-19/prevention & control , Immunoglobulin G , Vaccine EfficacyABSTRACT
An immune correlate of risk (CoR) is an immunologic biomarker in vaccine recipients associated with an infectious disease clinical endpoint. An immune correlate of protection (CoP) is a CoR that can be used to reliably predict vaccine efficacy (VE) against the clinical endpoint and hence is accepted as a surrogate endpoint that can be used for accelerated approval or guide use of vaccines. In randomized, placebo-controlled trials, CoR analysis is limited by not assessing a causal vaccine effect. To address this limitation, we construct the controlled risk curve of a biomarker, which provides the causal risk of an endpoint if all participants are assigned vaccine and the biomarker is set to different levels. Furthermore, we propose a causal CoP analysis based on controlled effects, where for the important special case that the biomarker is constant in the placebo arm, we study the controlled vaccine efficacy curve that contrasts the controlled risk curve with placebo arm risk. We provide identification conditions and formulae that account for right censoring of the clinical endpoint and two-phase sampling of the biomarker, and consider G-computation estimation and inference under a semiparametric model such as the Cox model. We add modular approaches to sensitivity analysis that quantify robustness of CoP evidence to unmeasured confounding. We provide an application to two phase 3 trials of a dengue vaccine indicating that controlled risk of dengue strongly varies with 50$\%$ neutralizing antibody titer. Our work introduces controlled effects causal mediation analysis to immune CoP evaluation.
Subject(s)
Vaccines , Humans , Vaccines/therapeutic use , Biomarkers/analysisABSTRACT
We propose causal isotonic calibration, a novel nonparametric method for calibrating predictors of heterogeneous treatment effects. In addition, we introduce a novel data-efficient variant of calibration that avoids the need for hold-out calibration sets, which we refer to as cross-calibration. Causal isotonic cross-calibration takes cross-fitted predictors and outputs a single calibrated predictor obtained using all available data. We establish under weak conditions that causal isotonic calibration and cross-calibration both achieve fast doubly-robust calibration rates so long as either the propensity score or outcome regression is estimated well in an appropriate sense. The proposed causal isotonic calibrator can be wrapped around any black-box learning algorithm to provide strong distribution-free calibration guarantees while preserving predictive performance.
ABSTRACT
INTRODUCTION: Although SARS-CoV-2 vaccines were first approved under Emergency Use Authorization by the Food and Drug Administration in late 2020 for adults, authorisation for young children 6 months to <5 years of age did not occur until 2022. These authorisations were based on clinical trials, understanding real-world vaccine effectiveness (VE) in the setting of emerging variants is critical. The primary goal of this study is to evaluate SARS-CoV-2 VE against infection among children aged >6 months and adults aged <50 years. METHODS: CASCADIA is a 4-year community-based prospective study of SARS-CoV-2 VE among 3500 adults and paediatric populations aged 6 months to 49 years in Oregon and Washington, USA. At enrolment and regular intervals, participants complete a sociodemographic questionnaire. Individuals provide a blood sample at enrolment and annually thereafter, with optional blood draws every 6 months and after infection and vaccination. Participants complete weekly self-collection of anterior nasal swabs and symptom questionnaires. Swabs are tested for SARS-CoV-2 and other respiratory pathogens by reverse transcription-PCR, with results of selected pathogens returned to participants; nasal swabs with SARS-CoV-2 detected will undergo whole genome sequencing. Participants who test positive for SARS-CoV-2 undergo serial swab collection every 3 days for 21 days. Serum samples are tested for SARS-CoV-2 antibody by binding and neutralisation assays. ANALYSIS: The primary outcome is SARS-CoV-2 infection. Cox regression models will be used to estimate the incidence rate ratio associated with SARS-CoV-2 vaccination among the paediatric and adult population, controlling for demographic factors and other potential confounders. ETHICS AND DISSEMINATION: All study materials including the protocol, consent forms, data collection instruments, participant communication and recruitment materials, were approved by the Kaiser Permanente Interregional Institutional Review Board, the IRB of record for the study. Results will be disseminated through peer-reviewed publications, presentations, participant newsletters and appropriate general news media.
Subject(s)
COVID-19 , United States , Adult , Humans , Child , Child, Preschool , Infant , COVID-19/diagnosis , COVID-19/epidemiology , COVID-19/prevention & control , SARS-CoV-2 , COVID-19 Vaccines , Prospective Studies , Vaccine Efficacy , InternetABSTRACT
In Bayesian data analysis, it is often important to evaluate quantiles of the posterior distribution of a parameter of interest (e.g., to form posterior intervals). In multi-dimensional problems, when non-conjugate priors are used, this is often difficult generally requiring either an analytic or sampling-based approximation, such as Markov chain Monte-Carlo (MCMC), Approximate Bayesian computation (ABC) or variational inference. We discuss a general approach that reframes this as a multi-task learning problem and uses recurrent deep neural networks (RNNs) to approximately evaluate posterior quantiles. As RNNs carry information along a sequence, this application is particularly useful in time-series. An advantage of this risk-minimization approach is that we do not need to sample from the posterior or calculate the likelihood. We illustrate the proposed approach in several examples.
ABSTRACT
BACKGROUND: People experiencing homelessness (PEH) are at increased risk for acquiring SARS-CoV-2, but the burden of long COVID in this population is unknown. METHODS: We conducted a matched prospective cohort study to assess the prevalence, characteristics, and impact of long COVID among sheltered PEH in Seattle, WA between September 2020-April 2022. Adults ≥ 18 years, residing across nine homeless shelters with active respiratory virus surveillance, were eligible to complete in-person baseline surveys and interval follow-up phone surveys. We included a subset of 22 COVID-19-positive cases who tested positive or inconclusive for SARS-CoV-2 and 44 COVID-19-negative controls who tested negative for SARS-CoV-2, frequency matched on age and sex. Among controls, 22 were positive and 22 were negative for one of 27 other respiratory virus pathogens. To assess the impact of COVID-19 on the risk of symptom presence at follow-up (day 30-225 post-enrollment test), we performed log-linear regression with robust standard errors, adjusting for confounding by shelter site and demographic variables determined a priori. RESULTS: Of 53 eligible COVID-19 cases, 22 (42%) completed ≥ 1 follow-up survey. While five (23%) cases reported ≥ 1 symptom at baseline, this increased to 77% (10/13) between day 30-59 and 33% (4/12) day 90 + . The most commonly reported symptoms day 30 + were fatigue (27%) and rhinorrhea (27%), with 8 (36%) reporting symptoms that interfered with or prevented daily activities. Four (33%) symptomatic cases reported receiving medical care outside of a medical provider at an isolation facility. Of 44 controls, 12 (27%) reported any symptoms day 90 + . Risk of any symptoms at follow-up was 5.4 times higher among COVID-19 cases compared to controls (95% CI: 2.7-10.5). CONCLUSIONS: Shelter residents reported a high prevalence of symptoms 30 + days after their SARS-CoV-2 detection, though few accessed medical care for persistent illness. The impact of COVID-19 extends beyond acute illness and may exacerbate existing challenges that marginalized populations face in maintaining their health and wellbeing.
Subject(s)
COVID-19 , Ill-Housed Persons , Humans , Adult , COVID-19/epidemiology , SARS-CoV-2 , Post-Acute COVID-19 Syndrome , Longitudinal Studies , Prospective StudiesABSTRACT
The best assay or marker to define mRNA-1273 vaccine-induced antibodies as a correlate of protection (CoP) is unclear. In the COVE trial, participants received two doses of the mRNA-1273 COVID-19 vaccine or placebo. We previously assessed IgG binding antibodies to the spike protein (spike IgG) or receptor binding domain (RBD IgG) and pseudovirus neutralizing antibody 50 or 80% inhibitory dilution titer measured on day 29 or day 57, as correlates of risk (CoRs) and CoPs against symptomatic COVID-19 over 4 months after dose. Here, we assessed a new marker, live virus 50% microneutralization titer (LV-MN50), and compared and combined markers in multivariable analyses. LV-MN50 was an inverse CoR, with a hazard ratio of 0.39 (95% confidence interval, 0.19 to 0.83) at day 29 and 0.51 (95% confidence interval, 0.25 to 1.04) at day 57 per 10-fold increase. In multivariable analyses, pseudovirus neutralization titers and anti-spike binding antibodies performed best as CoRs; combining antibody markers did not improve correlates. Pseudovirus neutralization titer was the strongest independent correlate in a multivariable model. Overall, these results supported pseudovirus neutralizing and binding antibody assays as CoRs and CoPs, with the live virus assay as a weaker correlate in this sample set. Day 29 markers performed as well as day 57 markers as CoPs, which could accelerate immunogenicity and immunobridging studies.
Subject(s)
2019-nCoV Vaccine mRNA-1273 , COVID-19 , Humans , Vaccine Efficacy , COVID-19/prevention & control , Antibodies, Neutralizing , Immunoglobulin G , Antibodies, ViralABSTRACT
INTRODUCTION: Overall, resilient health systems build upon sufficient, qualified, well-distributed, and motivated health workers; however, this precious resource is limited in numbers to meet people's demands, particularly in LMICs. Understanding the subnational distribution of health workers from different lens is critical to ensure quality healthcare and improving health outcomes. METHODS: Using data from Health Personnel Information System, facility-level Service Availability and Readiness Assessment, and other sources, we performed a district-level longitudinal analysis to assess health workforce density and the ratio of male to female health workers between January 2016 and June 2020 across all districts in Mozambique. RESULTS: 22 011 health workers were sampled, of whom 10 405 (47.3%) were male. The average age was 35 years (SD: 9.4). Physicians (1025, 4.7%), maternal and child health nurses (4808, 21.8%), and nurses (6402, 29.1%) represented about 55% of the sample. In January 2016, the average district-level workforce density was 75.8 per 100 000 population (95% CI 65.9, 87.1), and was increasing at an annual rate of 8.0% (95% CI 6.00, 9.00) through January 2018. The annual growth rate declined to 3.0% (95% CI 2.00, 4.00) after January 2018. Two provinces, Maputo City and Maputo Province, with 268.3 (95% CI 186.10, 387.00) and 104.6 (95% CI 84.20, 130.00) health workers per 100 000 population, respectively, had the highest workforce density at baseline (2016). There were 3122 community health workers (CHW), of whom 72.8% were male, in January 2016. The average number of CHWs per 10 000 population was 1.33 (95% CI 1.11, 1.59) in 2016 and increased by 18% annually between January 2016 and January 2018. This trend reduced to 11% (95% CI 0.00, 13.00) after January 2018. The sex ratio was twice as high for all provinces in the central and northern regions relative to Maputo Province. Maputo City (OR: 0.34; 95% CI 0.32, 0.34) and Maputo Province (OR: 0.56; 95% CI 0.49, 0.65) reported the lowest sex ratio at the baseline. Encouragingly, important sex ratio improvements were observed after January 2018, particularly in the northern and central regions. CONCLUSION: Mozambique made substantial progress in health workers' availability during the study period; however, with a critical slowdown after 2018. Despite the progress, meaningful shortages and distribution disparities persist.
Subject(s)
Health Personnel , Quality of Health Care , Child , Humans , Male , Female , Adult , Longitudinal Studies , Mozambique/epidemiology , WorkforceABSTRACT
Introduction: Overall, resilient health systems build upon sufficient, qualified, well-distributed, and motivated health workers; however, this precious resource is limited in numbers to meet people's demands, particularly in LMICs. Understanding the subnational distribution of health workers from different lens is critical to ensure quality healthcare and improving health outcomes. Methods: Using data from Health Personnel Information System, facility-level Service Availability and Readiness Assessment, and other sources, we performed a district-level longitudinal analysis to assess health workforce density and the ratio of male to female health workers between January 2016 and June 2020 across all districts in Mozambique. Results: 22 011 health workers were sampled, of whom 10 405 (47.3%) were male. The average age was 35 years (SD: 9.4). Physicians (1025, 4.7%), maternal and child health nurses (4808, 21.8%), and nurses (6402, 29.1%) represented about 55% of the sample. In January 2016, the average district-level workforce density was 75.8 per 100 000 population (95% CI 65.9, 87.1), and was increasing at an annual rate of 8.0% (95% CI 6.00, 9.00) through January 2018. The annual growth rate declined to 3.0% (95% CI 2.00, 4.00) after January 2018. Two provinces, Maputo City and Maputo Province, with 268.3 (95% CI 186.10, 387.00) and 104.6 (95% CI 84.20, 130.00) health workers per 100 000 population, respectively, had the highest workforce density at baseline (2016). There were 3122 community health workers (CHW), of whom 72.8% were male, in January 2016. The average number of CHWs per 10 000 population was 1.33 (95% CI 1.11, 1.59) in 2016 and increased by 18% annually between January 2016 and January 2018. This trend reduced to 11% (95% CI 0.00, 13.00) after January 2018. The sex ratio was twice as high for all provinces in the central and northern regions relative to Maputo Province. Maputo City (OR: 0.34; 95% CI 0.32, 0.34) and Maputo Province (OR: 0.56; 95% CI 0.49, 0.65) reported the lowest sex ratio at the baseline. Encouragingly, important sex ratio improvements were observed after January 2018, particularly in the northern and central regions. Conclusion: Mozambique made substantial progress in health workers' availability during the study period; however, with a critical slowdown after 2018. Despite the progress, meaningful shortages and distribution disparities persist..
Subject(s)
Humans , Male , Female , Child , Adult , Quality of Health Care , Health Personnel , Longitudinal Studies , Workforce , Mozambique/epidemiologyABSTRACT
Measuring immune correlates of disease acquisition and protection in the context of a clinical trial is a prerequisite for improved vaccine design. We analysed binding and neutralizing antibody measurements 4 weeks post vaccination as correlates of risk of moderate to severe-critical COVID-19 through 83 d post vaccination in the phase 3, double-blind placebo-controlled phase of ENSEMBLE, an international randomized efficacy trial of a single dose of Ad26.COV2.S. We also evaluated correlates of protection in the trial cohort. Of the three antibody immune markers we measured, we found most support for 50% inhibitory dilution (ID50) neutralizing antibody titre as a correlate of risk and of protection. The outcome hazard ratio was 0.49 (95% confidence interval 0.29, 0.81; P = 0.006) per 10-fold increase in ID50; vaccine efficacy was 60% (43%, 72%) at non-quantifiable ID50 (<2.7 IU50 ml-1) and increased to 89% (78%, 96%) at ID50 = 96.3 IU50 ml-1. Comparison of the vaccine efficacy by ID50 titre curves for ENSEMBLE-US, the COVE trial of the mRNA-1273 vaccine and the COV002-UK trial of the AZD1222 vaccine supported the ID50 titre as a correlate of protection across trials and vaccine types.
Subject(s)
Ad26COVS1 , COVID-19 , Humans , COVID-19/prevention & control , ChAdOx1 nCoV-19 , 2019-nCoV Vaccine mRNA-1273 , Vaccine Efficacy , Antibodies, NeutralizingABSTRACT
An immune correlate of risk (CoR) is an immunologic biomarker in vaccine recipients associated with an infectious disease clinical endpoint. An immune correlate of protection (CoP) is a CoR that can be used to reliably predict vaccine efficacy (VE) against the clinical endpoint and hence is accepted as a surrogate endpoint that can be used for accelerated approval or guide use of vaccines. In randomized, placebo-controlled trials, CoR analysis is limited by not assessing a causal vaccine effect. To address this limitation, we construct the controlled risk curve of a biomarker, which provides the causal risk of an endpoint if all participants are assigned vaccine and the biomarker is set to different levels. Furthermore, we propose a causal CoP analysis based on controlled effects, where for the important special case that the biomarker is constant in the placebo arm, we study the controlled vaccine efficacy curve that contrasts the controlled risk curve with placebo arm risk. We provide identification conditions and formulae that account for right censoring of the clinical endpoint and two-phase sampling of the biomarker, and consider G-computation estimation and inference under a semiparametric model such as the Cox model. We add modular approaches to sensitivity analysis that quantify robustness of CoP evidence to unmeasured confounding. We provide an application to two phase 3 trials of a dengue vaccine indicating that controlled risk of dengue strongly varies with 50$\%$ neutralizing antibody titer. Our work introduces controlled effects causal mediation analysis to immune CoP evaluation.