RESUMEN
BACKGROUND: Women in Africa disproportionately acquire HIV-1. Understanding which women are most likely to acquire HIV-1 can guide focused prevention with pre-exposure prophylaxis (PrEP). Our objective is to identify women at highest risk of HIV-1 and estimate PrEP efficiency at different sensitivity levels. METHODS: Nationally representative data were collected from 2015-2019 from 15 population-based household surveys. This analysis included women aged 15-49 who tested HIV-1 sero-negative or had recent HIV-1. Least absolute shrinkage and selection operator regression models were fit with 28 variables to predict recent HIV-1. Models were trained on the full population and internally cross-validated. Performance was evaluated using area under the receiver-operating-characteristic curve (AUC), sensitivity, and number needed to treat (NNT) with PrEP to avert one infection. RESULTS: Among 209,012 participants 248 had recent HIV-1 infection, representing 118 million women and 402,000 (95% CI: 309,000-495,000) new annual infections. Two variables were retained in the model: living in a subnational area with high HIV-1 viremia and having a sexual partner living outside the home. Full-population AUC was 0.80 (95% CI: 0.76-0.84); cross-validated AUC was 0.79 (95% CI: 0.75-0.84). At a sensitivity of 33%, up to 130,000 cases could be averted if 7.9 million women were perfectly adherent to PrEP; NNT would be 61. At a sensitivity of 67%, up to 260,000 cases could be averted if 25.1 million women were perfectly adherent to PrEP; the NNT would be 96. CONCLUSIONS: This risk assessment tool was generalizable, predictive, and parsimonious with tradeoffs between reach and efficiency.
RESUMEN
Etiologic heterogeneity occurs when distinct sets of events or exposures give rise to different subtypes of disease. Inference about subtype-specific exposure effects from two-phase outcome-dependent sampling data requires adjustment for both confounding and the sampling design. Common approaches to inference for these effects do not necessarily appropriately adjust for these sources of bias, or allow for formal comparisons of effects across different subtypes. Herein, using inverse probability weighting (IPW) to fit a multinomial model is shown to yield valid inference with this sampling design for subtype-specific exposure effects and contrasts thereof. The IPW approach is compared to common regression-based methods for assessing exposure effect heterogeneity using simulations. The methods are applied to estimate subtype-specific effects of various exposures on breast cancer risk in the Carolina Breast Cancer Study.
RESUMEN
People living with HIV on antiretroviral therapy often have undetectable virus levels by standard assays, but "latent" HIV still persists in viral reservoirs. Eliminating these reservoirs is the goal of HIV cure research. The quantitative viral outgrowth assay (QVOA) is commonly used to estimate the reservoir size, that is, the infectious units per million (IUPM) of HIV-persistent resting CD4+ T cells. A new variation of the QVOA, the ultra deep sequencing assay of the outgrowth virus (UDSA), was recently developed that further quantifies the number of viral lineages within a subset of infected wells. Performing the UDSA on a subset of wells provides additional information that can improve IUPM estimation. This paper considers statistical inference about the IUPM from combined dilution assay (QVOA) and deep viral sequencing (UDSA) data, even when some deep sequencing data are missing. Methods are proposed to accommodate assays with wells sequenced at multiple dilution levels and with imperfect sensitivity and specificity, and a novel bias-corrected estimator is included for small samples. The proposed methods are evaluated in a simulation study, applied to data from the University of North Carolina HIV Cure Center, and implemented in the open-source R package SLDeepAssay.
Asunto(s)
Infecciones por VIH , VIH-1 , Humanos , Latencia del Virus , VIH-1/genética , Linfocitos T CD4-Positivos , Simulación por Computador , Carga ViralRESUMEN
Assessing population-level effects of vaccines and other infectious disease prevention measures is important to the field of public health. In infectious disease studies, one person's treatment may affect another individual's outcome, that is, there may be interference between units. For example, the use of bed nets to prevent malaria by one individual may have an indirect effect on other individuals living in close proximity. In some settings, individuals may form groups or clusters where interference only occurs within groups, that is, there is partial interference. Inverse probability weighted estimators have previously been developed for observational studies with partial interference. Unfortunately, these estimators are not well suited for studies with large clusters. Therefore, in this paper, the parametric g-formula is extended to allow for partial interference. G-formula estimators are proposed for overall effects, effects when treated, and effects when untreated. The proposed estimators can accommodate large clusters and do not suffer from the g-null paradox that may occur in the absence of interference. The large sample properties of the proposed estimators are derived assuming no unmeasured confounders and that the partial interference takes a particular form (referred to as 'weak stratified interference'). Simulation studies are presented demonstrating the finite-sample performance of the proposed estimators. The Demographic and Health Survey from the Democratic Republic of the Congo is then analyzed using the proposed g-formula estimators to assess the effects of bed net use on malaria.
Asunto(s)
Malaria , Estudios Observacionales como Asunto , Humanos , Malaria/prevención & control , Mosquiteros Tratados con Insecticida/estadística & datos numéricos , Modelos Estadísticos , Simulación por Computador , República Democrática del Congo/epidemiologíaRESUMEN
While randomized controlled trials (RCTs) are critical for establishing the efficacy of new therapies, there are limitations regarding what comparisons can be made directly from trial data. RCTs are limited to a small number of comparator arms and often compare a new therapeutic to a standard of care which has already proven efficacious. It is sometimes of interest to estimate the efficacy of the new therapy relative to a treatment that was not evaluated in the same trial, such as a placebo or an alternative therapy that was evaluated in a different trial. Such dual-study comparisons are challenging because of potential differences between trial populations that can affect the outcome. In this article, two bridging estimators are considered that allow for comparisons of treatments evaluated in different trials, accounting for measured differences in trial populations. A "multi-span" estimator leverages a shared arm between two trials, while a "single-span" estimator does not require a shared arm. A diagnostic statistic that compares the outcome in the standardized shared arms is provided. The two estimators are compared in simulations, where both estimators demonstrate minimal empirical bias and nominal confidence interval coverage when the identification assumptions are met. The estimators are applied to data from the AIDS Clinical Trials Group 320 and 388 to compare the efficacy of two-drug vs four-drug antiretroviral therapy on CD4 cell counts among persons with advanced HIV. The single-span approach requires weaker identification assumptions and was more efficient in simulations and the application.
Asunto(s)
Antirretrovirales , Humanos , SesgoRESUMEN
Higher-order evidence is evidence about evidence. Epidemiologic examples of higher-order evidence include the settings where the study data constitute first-order evidence and estimates of misclassification comprise the second-order evidence (e.g., sensitivity, specificity) of a binary exposure or outcome collected in the main study. While sampling variability in higher-order evidence is typically acknowledged, higher-order evidence is often assumed to be free of measurement error (e.g., gold standard measures). Here we provide two examples, each with multiple scenarios where second-order evidence is imperfectly measured, and this measurement error can either amplify or attenuate standard corrections to first-order evidence. We propose a way to account for such imperfections that requires third-order evidence. Further illustrations and exploration of how higher-order evidence impacts results of epidemiologic studies is warranted.
Asunto(s)
Sesgo , Humanos , Sensibilidad y EspecificidadRESUMEN
We sought to examine cervical cancer screening barriers by sexual orientation among low-income women in North Carolina. The MyBodyMyTest-3 Trial recruited low-income women (< 250% of federal poverty level) aged 25-64 years who were 1+ year overdue for cervical cancer screening. We compared perceptions of cervical cancer screening among those who self-identified as lesbian, gay, bisexual, or queer (LGBQ; n = 70) to straight/heterosexual women (n = 683). For both LGBQ and straight respondents, the greatest barriers to screening were lack of health insurance (63% and 66%) and cost (49% and 50%). LGBQ respondents were more likely than straight respondents to report forgetting to screen (16% vs. 8%, p = .05), transportation barriers (10% vs. 2%, p = .001), and competing mental or physical health problems (39% vs. 27%, p = .10). Addressing access remains important for improving cervical cancer screening among those under-screened. For LGBQ women, additional attention may be needed for reminders, co-occurring health needs, and transportation barriers.
Asunto(s)
Detección Precoz del Cáncer , Accesibilidad a los Servicios de Salud , Pobreza , Neoplasias del Cuello Uterino , Humanos , Femenino , Neoplasias del Cuello Uterino/diagnóstico , North Carolina , Persona de Mediana Edad , Adulto , Detección Precoz del Cáncer/estadística & datos numéricos , Accesibilidad a los Servicios de Salud/estadística & datos numéricos , Conducta Sexual , Minorías Sexuales y de Género/estadística & datos numéricos , Minorías Sexuales y de Género/psicología , Aceptación de la Atención de Salud/estadística & datos numéricos , Aceptación de la Atención de Salud/psicología , Tamizaje Masivo/estadística & datos numéricosRESUMEN
"Fusion" study designs combine data from different sources to answer questions that could not be answered (as well) by subsets of the data. Studies that augment main study data with validation data, as in measurement-error correction studies or generalizability studies, are examples of fusion designs. Fusion estimators, here solutions to stacked estimating functions, produce consistent answers to identified research questions using data from fusion designs. In this paper, we describe a pair of examples of fusion designs and estimators, one where we generalize a proportion to a target population and one where we correct measurement error in a proportion. For each case, we present an example motivated by human immunodeficiency virus research and summarize results from simulation studies. Simulations demonstrate that the fusion estimators provide approximately unbiased results with appropriate 95% confidence interval coverage. Fusion estimators can be used to appropriately combine data in answering important questions that benefit from multiple sources of information.
Asunto(s)
Proyectos de Investigación , Humanos , Simulación por ComputadorRESUMEN
In first-line antiretroviral therapy (ART) for human immunodeficiency virus (HIV) treatment, some subgroups of patients may respond better to an efavirenz-based regimen than an integrase strand transfer inhibitor (InSTI)-based regimen, or vice versa, due to patient characteristics modifying treatment effects. Using data based on nearly 16,000 patients from the North American AIDS Cohort Collaboration on Research and Design from 2009-2016, statistical methods for precision medicine were employed to estimate an optimal treatment rule that minimizes the 5-year risk of the composite outcome of acquired immune deficiency syndrome (AIDS)-defining illnesses, serious non-AIDS events, and all-cause mortality. The treatment rules considered were functions that recommend either an efavirenz- or InSTI-based regimen conditional on baseline patient characteristics such as demographic information, laboratory results, and health history. The estimated 5-year risk under the estimated optimal treatment rule was 10.0% (95% confidence interval (CI): 8.6, 11.3), corresponding to an absolute risk reduction of 2.3% (95% CI: 0.9, 3.8) when compared with recommending an efavirenz-based regimen for all patients and 2.6% (95% CI: 1.0, 4.2) when compared with recommending an InSTI-based regimen for all. Tailoring ART to individual patient characteristics may reduce 5-year risk of the composite outcome compared with assigning all patients the same drug regimen.
Asunto(s)
Síndrome de Inmunodeficiencia Adquirida , Infecciones por VIH , Humanos , VIH , Infecciones por VIH/tratamiento farmacológico , Inhibidores de la Transcriptasa Inversa/uso terapéutico , Medicina de Precisión , Síndrome de Inmunodeficiencia Adquirida/tratamiento farmacológicoRESUMEN
INTRODUCTION: Endoscopic healing has been associated with improved long-term clinical outcomes in inflammatory bowel disease (IBD) and is a recommended target for treatment. Evidence is limited regarding real-world uptake and patterns of treat-to-target monitoring to assess endoscopic healing after treatment initiation. We aimed to estimate the proportion of patients in the Study of a Prospective Adult Research Cohort with IBD (SPARC IBD) who received colonoscopy in the 3-15 months after starting a new IBD treatment. METHODS: We identified SPARC IBD patients who initiated a new biologic (infliximab, adalimumab, certolizumab pegol, golimumab, vedolizumab, and ustekinumab) or tofacitinib. We estimated the proportion of patients who received colonoscopies in the 3-15 months after IBD treatment initiation and described use by patient subgroups. RESULTS: Among 1,708 eligible initiations from 2017 to 2022, the most common medications were ustekinumab (32%), infliximab (22%), vedolizumab (20%), and adalimumab (16%). The median patient age was 38 years, with 66% Crohn's disease; 55% were female, and 12% were non-White. In the 3-15 months after medication initiation, 49.3% (95% confidence interval 46.2%-52.5%) of initiations were followed by a colonoscopy. Colonoscopy use was similar between ulcerative colitis and Crohn's disease, but was higher among male patients, those older than 40 years, and those who received colonoscopy within 3 months of initiation. Colonoscopy use varied between study sites, from 26.6% (15.0%-38.3%) to 63.2% (54.5%-72.0%). DISCUSSION: Approximately half of SPARC IBD patients received colonoscopy in the 3-15 months after initiation to a new IBD treatment, suggesting a low uptake of treat-to-target colonoscopy for the assessment of mucosal healing in real-world clinical practice. The variation in colonoscopy use between study sites suggests a lack of consensus and a need for more robust evidence around whether or not the practice of routine monitoring colonoscopy is associated with improved patient outcomes.
Asunto(s)
Colitis Ulcerosa , Enfermedad de Crohn , Enfermedades Inflamatorias del Intestino , Humanos , Adulto , Masculino , Femenino , Enfermedad de Crohn/tratamiento farmacológico , Infliximab/uso terapéutico , Adalimumab/uso terapéutico , Ustekinumab/uso terapéutico , Estudios Prospectivos , Enfermedades Inflamatorias del Intestino/tratamiento farmacológico , Enfermedades Inflamatorias del Intestino/inducido químicamente , Colitis Ulcerosa/tratamiento farmacológico , ColonoscopíaRESUMEN
We describe an approach to sensitivity analysis introduced by Robins et al (1999), for the setting where the outcome is missing for some observations. This flexible approach focuses on the relationship between the outcomes and missingness, where data can be missing completely at random, missing at random given observed data, or missing not at random. We provide examples from HIV that include the sensitivity of the estimation of a mean and proportion under different missingness mechanisms. The approach illustrated provides a method for examining how the results of epidemiologic studies might shift as a function of bias due to missing data.
Asunto(s)
Modelos Estadísticos , Humanos , Sesgo , Estudios EpidemiológicosRESUMEN
Studies of social networks provide unique opportunities to assess the causal effects of interventions that may impact more of the population than just those intervened on directly. Such effects are sometimes called peer or spillover effects, and may exist in the presence of interference, that is, when one individual's treatment affects another individual's outcome. Randomization-based inference (RI) methods provide a theoretical basis for causal inference in randomized studies, even in the presence of interference. In this article, we consider RI of the intervention effect in the eX-FLU trial, a randomized study designed to assess the effect of a social distancing intervention on influenza-like-illness transmission in a connected network of college students. The approach considered enables inference about the effect of the social distancing intervention on the per-contact probability of influenza-like-illness transmission in the observed network. The methods allow for interference between connected individuals and for heterogeneous treatment effects. The proposed methods are evaluated empirically via simulation studies, and then applied to data from the eX-FLU trial.
Asunto(s)
Gripe Humana , Distanciamiento Físico , Ensayos Clínicos Controlados Aleatorios como Asunto , Red Social , Causalidad , Ensayos Clínicos Controlados Aleatorios como Asunto/estadística & datos numéricos , Gripe Humana/prevención & control , Gripe Humana/transmisión , HumanosRESUMEN
Many research questions in public health and medicine concern sustained interventions in populations defined by substantive priorities. Existing methods to answer such questions typically require a measured covariate set sufficient to control confounding, which can be questionable in observational studies. Differences-in-differences rely instead on the parallel trends assumption, allowing for some types of time-invariant unmeasured confounding. However, most existing difference-in-differences implementations are limited to point treatments in restricted subpopulations. We derive identification results for population effects of sustained treatments under parallel trends assumptions. In particular, in settings where all individuals begin follow-up with exposure status consistent with the treatment plan of interest but may deviate at later times, a version of Robins' g-formula identifies the intervention-specific mean under stable unit treatment value assumption, positivity, and parallel trends. We develop consistent asymptotically normal estimators based on inverse-probability weighting, outcome regression, and a double robust estimator based on targeted maximum likelihood. Simulation studies confirm theoretical results and support the use of the proposed estimators at realistic sample sizes. As an example, the methods are used to estimate the effect of a hypothetical federal stay-at-home order on all-cause mortality during the COVID-19 pandemic in spring 2020 in the United States.
Asunto(s)
Modelos Estadísticos , Pandemias , Humanos , Simulación por Computador , Probabilidad , Tamaño de la MuestraRESUMEN
We consider nonparametrically estimating the joint distribution of a survival time and mark variable, where the survival time is subject to right censoring and the mark variable is only observed when the survival time is not censored. The possibility of dependent censoring is allowed for using inverse probability of censoring weights. The proposed estimator is shown to be consistent and asymptotically normal. Finite sample behavior of the proposed methods are investigated via simulation study. Finally, we illustrate the nonparametric estimator from a recent HIV vaccine efficacy trial.
Asunto(s)
Análisis de Supervivencia , Humanos , Probabilidad , Simulación por ComputadorRESUMEN
BACKGROUND: Although severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infectious virus isolation in outpatients with coronavirus disease 2019 (COVID-19) has been associated with viral RNA levels and symptom duration, little is known about the host, disease, and viral determinants of infectious virus detection. METHODS: COVID-19 adult outpatients were enrolled within 7 days of symptom onset. Clinical symptoms were recorded via patient diary. Nasopharyngeal swabs were collected to quantitate SARS-CoV-2 RNA by reverse transcriptase polymerase chain reaction and for infectious virus isolation in Vero E6-cells. SARS-CoV-2 antibodies were measured in serum using a validated ELISA assay. RESULTS: Among 204 participants with mild-to-moderate symptomatic COVID-19, the median nasopharyngeal viral RNA was 6.5 (interquartile range [IQR] 4.7-7.6 log10 copies/mL), and 26% had detectable SARS-CoV-2 antibodies (immunoglobulin (Ig)A, IgM, IgG, and/or total Ig) at baseline. Infectious virus was recovered in 7% of participants with SARS-CoV-2 antibodies compared to 58% of participants without antibodies (prevalence ratio [PR] = 0.12, 95% confidence interval [CI]: .04, .36; P = .00016). Infectious virus isolation was also associated with higher levels of viral RNA (mean RNA difference +2.6 log10, 95% CI: 2.2, 3.0; P < .0001) and fewer days since symptom onset (PR = 0.79, 95% CI: .71, .88 per day; P < .0001). CONCLUSIONS: The presence of SARS-CoV-2 antibodies is strongly associated with clearance of infectious virus. Seropositivity and viral RNA levels are likely more reliable markers of infectious virus clearance than subjective measure of COVID-19 symptom duration. Virus-targeted treatment and prevention strategies should be administered as early as possible and ideally before seroconversion. CLINICAL TRIALS REGISTRATION: NCT04405570.
Asunto(s)
COVID-19 , Enfermedades Transmisibles , Adulto , Anticuerpos Antivirales , Prueba de COVID-19 , Humanos , Inmunoglobulina A , Pacientes Ambulatorios , ARN Viral , SARS-CoV-2RESUMEN
In the analysis of observational studies, inverse probability weighting (IPW) is commonly used to consistently estimate the average treatment effect (ATE) or the average treatment effect in the treated (ATT). The variance of the IPW ATE estimator is often estimated by assuming that the weights are known and then using the so-called "robust" (Huber-White) sandwich estimator, which results in conservative standard errors (SEs). Here we show that using such an approach when estimating the variance of the IPW ATT estimator does not necessarily result in conservative SE estimates. That is, assuming the weights are known, the robust sandwich estimator may be either conservative or anticonservative. Thus, confidence intervals for the ATT using the robust SE estimate will not be valid, in general. Instead, stacked estimating equations which account for the weight estimation can be used to compute a consistent, closed-form variance estimator for the IPW ATT estimator. The 2 variance estimators are compared via simulation studies and in a data analysis of the association between smoking and gene expression.
Asunto(s)
Modelos Estadísticos , Simulación por Computador , Humanos , ProbabilidadRESUMEN
BACKGROUND: Of new sexually transmitted infections (STIs) in the United States, 50% occur among youth aged 15 to 24 years. Previous studies among youth with HIV (YHIV) do not distinguish STI trends among individuals with perinatally (YPHIV) and nonperinatally (YNPHIV) acquired HIV. METHODS: Among 3 Adolescent Medicine Trials Network for HIV/AIDS Interventions (ATN) studies conducted between 2009 and 2015, we estimated incident diagnoses of trichomonal, bacterial, viral, and overall STIs stratified by sex assigned at birth, mode of HIV acquisition (perinatal [YPHIV] and nonperinatal [YNPHIV]), age (13-17 and 18-24 years), CD4 count (<200, 200-499, and ≥500/µL), and HIV viral load (VL) (<400 and ≥400 copies/mL). RESULTS: Among 3131 YHIV, across the 3 studies, mean (SD) age was 20.6 (2.6) years, 888 (28%) were female, 2498 (80%) had nonperinatal HIV acquisition recorded, and 2298 (73%) were African American/Black. Mean follow-up was 0.9 (0.3) years. Compared with YPHIV, YNPHIV spent less person-time with VL <400 copies/mL (47% vs. 53%) and more time off antiretroviral therapy (49% vs. 15%), and had higher overall STI rates (males, 65.9 vs. 8.5/100 person-years [PY]; females, 54.7 vs. 17.2/100 PY). Among YPHIV, bacterial STIs were higher during person-time spent with VL ≥400 vs. <400 copies/mL (male YPHIV, 10.9 vs. 0.6/100 PY; female YPHIV, 11.2 vs. 2.9/100 PY); no difference was observed among YNPHIV, which may be due to concurrent acquisition of HIV and other STIs and limited follow-up. CONCLUSIONS: Compared with YPHIV, YNPHIV spent less time on antiretroviral therapy and virologically suppressed; YNPHIV also had higher STI diagnosis rates. Very high STI diagnosis rates among YHIV, including among those without virologic suppression, highlight the importance of youth-focused efforts to support durable virologic suppression and identify and treat STIs.
Asunto(s)
Síndrome de Inmunodeficiencia Adquirida , Infecciones por VIH , Enfermedades de Transmisión Sexual , Adolescente , Adulto , Negro o Afroamericano , Recuento de Linfocito CD4 , Femenino , Infecciones por VIH/diagnóstico , Infecciones por VIH/epidemiología , Humanos , Recién Nacido , Masculino , Enfermedades de Transmisión Sexual/diagnóstico , Enfermedades de Transmisión Sexual/epidemiología , Estados Unidos/epidemiología , Adulto JovenRESUMEN
BACKGROUND: Low-income and uninsured people with a cervix (PWC) are at the highest risk of being underscreened for cervical cancer. We evaluated the prevalence of high-risk human papillomavirus (hrHPV) on home self-collected samples, as well as rates of in-clinic follow-up and risk factors associated with hrHPV positivity in this at-risk population. METHODS: My Body My Test 3 was conducted between 2016 and 2019 in North Carolina among individuals aged 25 to 64 years, overdue for cervical cancer screening, and with incomes of <250% of the US Federal Poverty Level. Our analytic sample included participants randomized to the self-collection arm who returned self-collected cervicovaginal brush samples for HPV testing (n = 329). Samples were tested for 14 hrHPV types by an HPV RNA assay and further genotyped for HPV-16 and HPV-18/45. We examined behavioral risk factors for hrHPV positivity using logistic regression and between-subject t tests. RESULTS: High-risk HPV RNA prevalence was 16% (n = 52/329) in self-collected samples. Of the hrHPV-positive participants, 24 (46%) presented for in-clinic cervical cancer screening, compared with 56 (20%) of hrHPV-negative participants. Those with ≥2 sexual partners in the past year were twice as likely to be hrHPV positive in adjusted analyses (adjusted odds ratio, 2.00 [95% confidence interval, 1.03-3.88]). High-risk HPV-positive and HPV-negative participants had similar attitudes toward screening, with the exception of hrHPV-positive participants who reported a lower perceived risk of cervical cancer than those who were hrHPV negative (P < 0.05). CONCLUSION: The hrHPV RNA prevalence was similar to findings in other underscreened PWC in the United States. Efforts to reach underscreened PWC are critical for cervical cancer prevention. Future studies aimed at home self-collection should address methods of increasing clinic attendance and completion of treatment among those with HPV-positive results.
Asunto(s)
Alphapapillomavirus , Infecciones por Papillomavirus , Neoplasias del Cuello Uterino , Adulto , Alphapapillomavirus/genética , Detección Precoz del Cáncer/métodos , Femenino , Humanos , Persona de Mediana Edad , North Carolina/epidemiología , Papillomaviridae/genética , Infecciones por Papillomavirus/diagnóstico , Infecciones por Papillomavirus/epidemiología , Aceptación de la Atención de Salud , Prevalencia , ARN , Neoplasias del Cuello Uterino/diagnóstico , Neoplasias del Cuello Uterino/epidemiología , Neoplasias del Cuello Uterino/prevención & controlRESUMEN
Inverse probability of treatment weights (IPTWs) are commonly used to control for confounding when estimating causal effects of point exposures from observational data. When planning a study that will be analyzed with IPTWs, determining the required sample size for a given level of statistical power is challenging because of the effect of weighting on the variance of the estimated causal means. This paper considers the utility of the design effect to quantify the effect of weighting on the precision of causal estimates. The design effect is defined as the ratio of the variance of the causal mean estimator divided by the variance of a naïve estimator if, counter to fact, no confounding had been present and weights were not needed. A simple, closed-form approximation of the design effect is derived that is outcome invariant and can be estimated during the study design phase. Once the design effect is approximated for each treatment group, sample size calculations are conducted as for a randomized trial, but with variances inflated by the design effects to account for weighting. Simulations demonstrate the accuracy of the design effect approximation, and practical considerations are discussed.
Asunto(s)
Modelos Estadísticos , Proyectos de Investigación , Causalidad , Probabilidad , Tamaño de la MuestraRESUMEN
Estimating population-level effects of a vaccine is challenging because there may be interference, that is, the outcome of one individual may depend on the vaccination status of another individual. Partial interference occurs when individuals can be partitioned into groups such that interference occurs only within groups. In the absence of interference, inverse probability weighted (IPW) estimators are commonly used to draw inference about causal effects of an exposure or treatment. Tchetgen Tchetgen and VanderWeele proposed a modified IPW estimator for causal effects in the presence of partial interference. Motivated by a cholera vaccine study in Bangladesh, this paper considers an extension of the Tchetgen Tchetgen and VanderWeele IPW estimator to the setting where the outcome is subject to right censoring using inverse probability of censoring weights (IPCW). Censoring weights are estimated using proportional hazards frailty models. The large sample properties of the IPCW estimators are derived, and simulation studies are presented demonstrating the estimators' performance in finite samples. The methods are then used to analyze data from the cholera vaccine study.