Donna Spiegelman’s research while affiliated with Yale University and other places

Hybrid studies allow investigators to simultaneously study an intervention effectiveness outcome and an implementation research outcome. In particular, type 2 hybrid studies support research that places equal importance on both outcomes rather than focusing on one and secondarily on the other (i.e., type 1 and type 3 studies). Hybrid 2 studies introduce the statistical issue of multiple testing, complicated by the fact that they are typically also cluster randomized trials. Standard statistical methods do not apply in this scenario. Here, we describe the design methodologies available for validly powering hybrid type 2 studies and producing reliable sample size calculations in a cluster‐randomized design with a focus on binary outcomes. Through a literature search, 18 publications were identified that included methods relevant to the design of hybrid 2 studies. Five methods were identified, two of which did not account for clustering but are extended in this article to do so, namely the combined outcomes approach and the single 1‐degree of freedom combined test. Procedures for powering hybrid 2 studies using these five methods are described and illustrated using input parameters inspired by a study from the Community Intervention to Reduce CardiovascuLar Disease in Chicago (CIRCL‐Chicago) Implementation Research Center. In this illustrative example, the intervention effectiveness outcome was controlled blood pressure, and the implementation outcome was reach. The conjunctive test resulted in higher power than the popular p value adjustment methods, and the newly extended combined outcomes and single 1‐DF test were found to be the most powerful among all of the tests.

Behavioral health interventions, such as trainings or incentives, are implemented in settings where individuals are interconnected, and the intervention assigned to some individuals may also affect others within their network. Evaluating such interventions requires assessing both the effect of the intervention on those who receive it and the spillover effect on those connected to the treated individuals. With behavioral interventions, spillover effects can be heterogeneous in that certain individuals, due to their social connectedness and individual characteristics, are more likely to respond to the intervention and influence their peers' behaviors. Targeting these individuals can enhance the effectiveness of interventions in the population. In this paper, we focus on an Egocentric Network-based Randomized Trial (ENRT) design, wherein a set of index participants is recruited from the population and randomly assigned to the treatment group, while concurrently collecting outcome data on their nominated network members, who remina untreated. In such design, spillover effects on network members may vary depending on the characteristics of the index participant. Here, we develop a testing method, the Multiple Comparison with Best (MCB), to identify subgroups of index participants whose treatment exhibits the largest spillover effect on their network members. Power and sample size calculations are then provided to design ENRTs that can detect key influencers. The proposed methods are demonstrated in a study on network-based peer HIV prevention education program, providing insights into strategies for selecting peer educators in peer education interventions.

BACKGROUND The HIV epidemic is a national focus within the United States (U.S.), Ending the HIV Epidemic (EHE) Initiative aiming to reduce new HIV infections by 90% by 2023 through scaling of prevention and treatment strategies. Data are crucial to understanding HIV-related needs, barriers to care, and the effectiveness of interventions. While several publicly available datasets exist, few integrate multiple topic domains such as HIV outcomes, social determinants of health (SDOH), and community-level factors. The lack of unified data and difficulty linking datasets across these domains hampers efforts to tailor HIV management and treatment strategies. Existing datasets are often siloed and difficult to analyze together, limiting potential for comprehensive research. Combining and integrating data across geographic and thematic strata offers significant potential to better understand the factors influencing HIV outcomes and to optimize intervention strategies. OBJECTIVE This study explores the feasibility of combining public HIV and community data to identify factors influencing HIV outcomes and interventions. We described the process of sourcing, extracting, and linking variables into a unified database to highlight key community factors and demonstrate the potential of this integrated approach for overcoming data access barriers in HIV research. METHODS A team of two researchers was integrated within a larger network to undergo this task. Approximately 350 hours were spent developing a single unified database. The build was conducted across three phases: 1) initial sourcing of datasets, 2) extracting and linking data, and 3) quality control/quality assurance. Experts were consulted to identify relevant data sources, and data was extracted and linked across multiple geographics levels. Quality control included manual spot checks and web scraping with Web Scraper, a third-party tool to verify accuracy of the extracted data. Data were uploaded to SAS for further analysis. RESULTS The resulting database comprises 242 variables sources from eight publicly available datasets. These variables were linked to 104 Ryan White-associated clinics across five U.S. states, covering geographic levels from state to individual clinic. All data were sourced and scraped between August 2022 and January 2023. Following data preparation, descriptive tables were generated, grouped into three domains: environmental variables, HIV-related variables, and implementation outcome variables. The final database and data dictionary are publicly accessible on the Yale Center for Implementation and Prevention Science (CMIPS) website for future research. CONCLUSIONS This integrated HIV and SDOH database overcomes key challenges in accessing fragmented data, offering valuable insights for addressing HIV outcomes. Despite barriers like technology learning curves and data availability, the project successfully created a resource that can inform EHE goals and guide future interventions. Expanding the database will enable deeper analysis and support ongoing effort to end the HIV epidemic.

Introduction Cancer is the primary cause of death globally, and despite the significant advancements in treatment and survival rates, it is still stigmatized in many parts of the world. However, there is limited public health research on cancer stigma among the general female population in Nepal. Therefore, this study aims to determine the prevalence of cancer stigma and its associated factors in this group. Methods We conducted a cross-sectional study among 426 healthy women aged 30 to 60 years who were residents of Dhulikhel and Banepa in central Nepal. We measured cancer stigma using the Cancer Stigma Scale (CASS). CASS measures cancer stigma in six domains (awkwardness, avoidance, severity, personal responsibility, policy opposition, financial discrimination) on a 6-point Likert scale (strongly disagree to agree strongly) with higher mean stigma scores correlating with higher levels of stigma. We utilized Generalized Estimating Equations (GEE) with multivariable linear regression to identify the socio-demographic factors associated with the CASS score. Results Overall, the level of cancer stigma was low, with a mean stigma score of 2.6 (0.6), but it was still present among participants. Stigma related to personal responsibility had the highest levels, with a mean score of 3.9 (1.3), followed by severity with a mean score of 3.2 (1.3), and financial discrimination with a mean score of 2.9 (1.6). There was a significant association between the mean CASS score and older age (mean difference in stigma score: 0.11 points; 95% CI: 0.02–0.20) as well as lower education (difference: -0.02 points; 95% CI: -0.03 to -0.003), after adjusting for age, ethnicity, education, marital status, religion, occupation, and parity. Conclusion While overall cancer stigma was low, some domains of stigma were higher among women in a suburban area in central Nepal; thus, indicating that cancer stigma persists in this region despite its low overall prevalence.

Hybrid studies allow investigators to simultaneously study an intervention effectiveness outcome and an implementation research outcome. In particular, type 2 hybrid studies support research that places equal importance on both outcomes rather than focusing on one and secondarily on the other (i.e., type 1 and type 3 studies). Hybrid 2 studies introduce the statistical issue of multiple testing, complicated by the fact that they are typically also cluster randomized trials. Standard statistical methods do not apply in this scenario. Here, we describe the design methodologies available for validly powering hybrid type 2 studies and producing reliable sample size calculations in a cluster-randomized design with a focus on binary outcomes. Through a literature search, 18 publications were identified that included methods relevant to the design of hybrid 2 studies. Five methods were identified, two of which did not account for clustering but are extended in this article to do so, namely the combined outcomes approach and the single 1-degree of freedom combined test. Procedures for powering hybrid 2 studies using these five methods are described and illustrated using input parameters inspired by a study from the Community Intervention to Reduce CardiovascuLar Disease in Chicago (CIRCL-Chicago) Implementation Research Center. In this illustrative example, the intervention effectiveness outcome was controlled blood pressure, and the implementation outcome was reach. The conjunctive test resulted in higher power than the popular p-value adjustment methods, and the newly extended combined outcomes and single 1-DF test were found to be the most powerful among all of the tests.

Background The Nepalese government endorsed and implemented the Package of Essential Non-Communicable Disease Interventions (PEN) by the World Health Organization (WHO) to prevent and manage four major non-communicable diseases (NCDs): cardiovascular disease (CVD), diabetes, cancers, and chronic respiratory diseases. This study explored barriers and facilitators to patient utilization of NCD services at primary healthcare facilities in Nepal. Methodology: We conducted a qualitative study with a 35 purposive sample of patients living with one or more NCDs (hypertension, diabetes, chronic obstructive pulmonary disease (COPD/ asthma) who sought healthcare at primary healthcare facilities in 14 randomly selected districts in seven provinces in Nepal that implemented PEN. Trained qualitative researchers conducted in-depth interviews in-person in a private setting using a semi-structured interview guide developed based on the Health Belief Model in the local language. The interviews were audio-recorded, transcribed verbatim, coded inductively and deductively, and analyzed by a framework approach using Dedoose software. Results From the perspectives of patients, key facilitators of service utilization encompassed free medicines, low-cost services, geographical and financial accessibility, less waiting time, positive interactions with health service providers, experiencing improvements in their health conditions, and support from family and peers. Barriers to utilizing services included inadequate health services (e.g., lack of medications and equipment), inaccessibility and affordability, inadequate health-related information from health service providers, low knowledge of NCD care, and lack of reminders or follow ups. Conclusion Enhancing NCD service utilization is potentially attainable through interventions that address patients’ knowledge, self-motivation, and misconceptions. Furthermore, strengthening the availability and accessibility of crucial services such as laboratory investigations, medications, equipment, and the patient-provider relationship is crucial for sustainable implementation of PEN.

Environmental exposures are often defined using buffer zones around geocoded home addresses, but these static boundaries can miss dynamic daily activity patterns, leading to biased results. This paper presents a novel measurement error correction method for spatially defined environmental exposures within a survival analysis framework using the Cox proportional hazards model. The method corrects high-dimensional surrogate exposures from geocoded residential data at multiple buffer radii by applying principal component analysis for dimension reduction and leveraging external GPS-tracked validation datasets containing true exposure measurements. It also derives the asymptotic properties and variances of the proposed estimators. Extensive simulations are conducted to evaluate the performance of the proposed estimators, demonstrating its ability to improve accuracy in estimated exposure effects. An illustrative application assesses the impact of greenness exposure on depression incidence in the Nurses' Health Study (NHS). The results demonstrate that correcting for measurement error significantly enhances the accuracy of exposure estimates. This method offers a critical advancement for accurately assessing the health impacts of environmental exposures, outperforming traditional static buffer approaches.

One way to quantify exposure to air pollution and its constituents in epidemiologic studies is to use an individual's nearest monitor. This strategy results in potential inaccuracy in the actual personal exposure, introducing bias in estimating the health effects of air pollution and its constituents, especially when evaluating the causal effects of correlated multi-pollutant constituents measured with correlated error. This paper addresses estimation and inference for the causal effect of one constituent in the presence of other PM2.5 constituents, accounting for measurement error and correlations. We used a linear regression calibration model, fitted with generalized estimating equations in an external validation study, and extended a double/debiased machine learning (DML) approach to correct for measurement error and estimate the effect of interest in the main study. We demonstrated that the DML estimator with regression calibration is consistent and derived its asymptotic variance. Simulations showed that the proposed estimator reduced bias and attained nominal coverage probability across most simulation settings. We applied this method to assess the causal effects of PM2.5 constituents on cognitive function in the Nurses' Health Study and identified two PM2.5 constituents, Br and Mn, that showed a negative causal effect on cognitive function after measurement error correction.