The Dartmouth Institute for Health Policy and Clinical Practice (TDI) is an organization within Dartmouth College "dedicated to improving health care through education, research, policy reform, leadership improvement, and communication with patients and the public." It was founded in 1988 by John Wennberg as the Center for the Evaluative Clinical Sciences (CECS); a reorganization in 2007 led to TDI's current structure.The institute provides a graduate-level education program involving elements of both Dartmouth's Graduate Arts and Sciences Programs and the Geisel School of Medicine. It grants Masters in Public Health degrees as well as Master of Science and Doctor of Philosophy in Health Policy and Clinical Science degrees. The institute is located at One Medical Center Drive, WTRB, Level 5 on the Dartmouth Hitchcock Hospital campus, Lebanon, NH. The institute's largest policy product is the Dartmouth Atlas of Health Care, which documents unwarranted variation in the American health care system.Dr. Anna Tosteson has served as interim director since October 2018, when former director Elliott S. Fisher and chief of strategy Adam Keller were placed on paid administrative leave following a complaint about conduct in the workplace. The investigation into misconduct concluded in April 2019 and resulted in Fisher's demotion and Keller's resignation. This followed the resignation of Professor H. Gilbert Welch in 2018 after Dartmouth College concluded he committed plagiarism. As a condition of his return directly set by the Geisel School of Medicine, Fisher has been banned physically from the 5th floor of the Williamson Building, where most of TDI is housed, for a period of two years..
Causal variable selection in time-varying treatment settings is challenging due to evolving confounding effects. Existing methods mainly focus on time-fixed exposures and are not directly applicable to time-varying scenarios. We propose a novel two-step procedure for variable selection when modeling the treatment probability at each time point. We first introduce a novel approach to longitudinal confounder selection using a Longitudinal Outcome Adaptive LASSO (LOAL) that will data-adaptively select covariates with theoretical justification of variance reduction of the estimator of the causal effect. We then propose an Adaptive Fused LASSO that can collapse treatment model parameters over time points with the goal of simplifying the models in order to improve the efficiency of the estimator while minimizing model misspecification bias compared with naive pooled logistic regression models. Our simulation studies highlight the need for and usefulness of the proposed approach in practice. We implemented our method on data from the Nicotine Dependence in Teens study to estimate the effect of the timing of alcohol initiation during adolescence on depressive symptoms in early adulthood.
Racial disparities in end-of-life care have been well documented, yet little is known about how both patient and provider races, as well as their concordance, influence the likelihood of advance care planning (ACP) discussions during hospitalization. To evaluate how patient race, provider race, and patient–provider racial concordance are associated with the occurrence of inpatient ACP conversations. Retrospective observational cohort study using hierarchical logistic regression. Seriously ill Medicare beneficiaries hospitalized between 2016 and 2019, managed by a national physician staffing organization (PSO) across 220 hospitals in 35 US states. The final sample included 390,392 hospitalizations and 2808 providers. The primary outcome was the occurrence of an ACP conversation, identified using CPT codes 99497 and 99498, assessed from admission through day 10 or discharge. Patient and provider races were categorized as White, Black, Hispanic, or Asian. Models included fixed effects for patient demographics, clinical risk, and hospital characteristics, and random effects for hospital clustering. Asian providers were more likely and Hispanic providers less likely to engage in ACP discussions. Patient–provider racial concordance modestly increased the likelihood of ACP for Black, White, and Hispanic patients, and several cross-race pairings also showed higher engagement. These effects were modest, varied across racial dyads, and occurred in the context of higher than national average inpatient ACP rates under the PSO’s quality improvement initiative. Provider race and patient–provider concordance each influenced the likelihood of inpatient ACP, though effects were modest and context-dependent. Concordance and certain racial pairings were associated with higher engagement, but disparities persisted across groups, highlighting that broader structural and communication barriers continue to shape inequities in end-of-life care.
This study investigates the impact of different types of personal protective equipment (PPE), specifically Powered Air-Purifying Respirators (PAPRs) and traditional N95 masks with face shields, on the physiological stress responses of healthcare workers (HWs) during the COVID-19 pandemic. Utilizing an interventional randomized crossover trial design, the research encompasses a simulation phase with ten participants followed by field testing involving thirty frontline healthcare professionals in a tertiary-care hospital setting. Heart rate (HR) and movement data were collected through smartwatches, while trained observers recorded the duration and nature of various activities undertaken during simulations. Data analysis employed statistical techniques, including Principal Component Analysis (PCA) and t-Distributed Stochastic Neighbor Embedding (t-SNE), to explore potential correlations between PPE type, HR, and movement. Clustering validation measures such as the Calinski-Harabasz, Davies-Bouldin, and Silhouette scores were applied to evaluate the difference between each type of PPE. The results indicated no significant differentiation in HR responses between the two PPE types. However, because HR may lack the sensitivity to fully capture variations in cognitive load or stress, these findings should be interpreted as an exploratory baseline. Additionally, no clear distinctions were observed regarding individual user responses or the activities performed, even when considering movement data. Although the findings imply non-inferiority of the examined PPE, future research including heart rate variability as a more comprehensive indicator of stress would be informative. This research contributes valuable insights into PPE selection and its implications for healthcare worker performance and well-being in high-stress environments, ultimately aiming to inform guidelines and training programs to enhance healthcare delivery during infectious disease outbreaks.
CONTEXT:Whether differences in program composition and provider staffing translate into meaningful variation in end-of-life (EOL) outcomes remains poorly understood, underscoring the need to examine how palliative care (PC) is delivered within hospitals. OBJECTIVES:This study applies a data-driven approach to understand longitudinal PC delivery of service across different types of providers ("provider sequence") and explore whether these patterns of PC delivery correlate with EOL quality measures for patients with poor-prognosis advanced cancer. METHODS:We conducted a retrospective cohort study using 2018-2019 Medicare fee-for-service claims. For each patient, we defined a provider sequence as a pair of consecutive PC encounters, classifying each encounter into a provider type: team specialist (TS), independent specialist (IS), and primary (P). For a given hospital, we aggregated the provider sequences of patients assigned to that hospital. We defined nine possible provider sequence patterns and quantified the proportion of provider sequence patterns within hospitals. We tested the association between the provider sequence patterns and three EOL quality measures: < one emergency department visit in the last 30 days, hospice use, and hospice enrollment ≥ three days. RESULTS:Across 276 hospitals, a higher proportion of the IS→TS provider sequence pattern was associated with greater hospice use (β = 0.48; 95% confidence interval: 0.03, 0.93). For hospice enrollment ≥ three days, the TS→P provider sequence pattern had a negative association (β = -0.18; 95% confidence interval: -0.33, -0.04). CONCLUSION:A provider sequence quality signal may provide deeper insights into how palliative care is delivered within a hospital and may help explain EOL outcomes.
PURPOSE Patient, tumor, and treatment factors can help predict the chance that a woman with a history of breast cancer diagnosis will be diagnosed with a second breast cancer within a year of a negative mammogram. This qualitative study elucidates breast cancer survivor and multispecialty physician perspectives on barriers/facilitators to clinical implementation of a risk-prediction tool to support surveillance decisions. MATERIALS AND METHODS We enrolled women who completed primary breast cancer treatment and physicians from November 2023 to April 2024. Participants were recruited through Breast Cancer Surveillance Consortium's registries; patients participated in one of four focus groups and physicians participated in individual semistructured interviews. Participants were presented with information about an interval cancer risk prediction tool and were prompted to share perspectives on facilitators and barriers to using such a tool. To identify salient themes, thematic analysis was undertaken by three research team members. RESULTS Participants included 40 physicians and 23 patients. Three themes emerged: (1) evidence needed for tool acceptance, (2) tool features to facilitate usage, and (3) barriers to tool adoption. Both cancer survivor and physician groups were accepting of risk prediction tool use for surveillance imaging when tool development information was available; they perceived the tool would fit within workflows, and data integrity could be verified. Both groups anticipated structural (time) and technological barriers (magnetic resonance imaging availability) could impede adoption. CONCLUSION Qualitative findings from focus groups and interviews analyzed thematically suggest implementing a risk prediction tool for surveillance imaging requires evidence transparency, health record integration, data integrity protection, and system supports to promote ease of use in clinical settings while mitigating unintended consequences. All are important to consider during tool development and implementation planning.
