Faculty Bibliography

Groundbreaking insights into the origins of the human mind have been garnered through the study of eye movements in preverbal subjects who are unable to explain their thought processes. Developmental research has largely relied on in-lab testing with trained experimenters. This constraint provides a narrow window into infant cognition and impedes large-scale data collection in families from diverse socioeconomic, geographic, and cultural backgrounds. Here we introduce a new open-source methodology for automatically analyzing infant eye-tracking data collected on personal devices in the home. Using algorithms from computer vision, machine learning, and ecological psychology, we develop an online webcam-linked eye tracker (OWLET) that provides robust estimation of infants' point of gaze from smartphone and webcam recordings of infant assessments in the home. We validate OWLET in a large sample of 7-month-old infants (N = 127) tested remotely, using an established visual attention task. We show that this new method reliably estimates infants' point-of-gaze across a variety of contexts, including testing on both computers and mobile devices, and exhibits excellent external validity with parental-report measures of attention. Our platform fills a significant gap in current tools available for rapid online data collection and large-scale assessments of cognitive processes in infants. Remote assessment addresses the need for greater diversity and accessibility in human studies and may support the ecological validity of behavioral experiments. This constitutes a critical and timely advance in a core domain of developmental research and in psychological science more broadly.

Wildfire smoke is a rapidly growing threat to global cardiovascular health. We review the literature linking wildfire smoke exposures to cardiovascular effects. We find substantial evidence that short-term exposures are associated with key cardiovascular outcomes, including mortality, hospitalization, and acute coronary syndrome. Wildfire smoke exposures will continue to increase over the majority of Earth's surface. For example, the United States alone has experienced a 5-fold increase in annual area burned since 1972, with 82 million individuals estimated to be exposed to wildfire smoke by midcentury. The associated rise in excess morbidity and mortality constitutes a growing global public health crisis. Fortunately, the effect of wildfire smoke on cardiovascular health is modifiable at the individual and population levels through specific interventions. Health systems therefore have an opportunity to help safeguard patients from smoke exposures. We provide a roadmap of evidence-based interventions to reduce risk and protect cardiovascular health. Key interventions include preparing health systems for smoke events; identifying and educating vulnerable patients; reducing outdoor activities; creating cleaner air environments; using air filtration devices and personal respirators; and aggressive management of chronic diseases and traditional risk factors. Further research is needed to test the efficacy of interventions on reducing cardiovascular outcomes.

BACKGROUND:Surgery is the most common treatment for localized small kidney masses (SKMs) up to 4 cm, despite a lack of evidence for improved overall survival. Nonsurgical management options are gaining recognition, as evidence supports the indolence of most SKMs. Decision aids (DAs) have been shown to improve patient comprehension of the trade-offs of treatment options and overall decision quality, and may improve consideration of all major options according to individual health priorities and preferences. OBJECTIVE:This pilot randomized controlled trial (RCT) primarily aims to evaluate the impact of a new web-based DA on treatment decisions for patients with SKM; that is, selection of surgical versus nonsurgical treatment options. Secondary objectives include an assessment of decision-making outcomes: decisional conflict, decision satisfaction, and an understanding of individual preferences for treatment that incorporate the trade-offs associated with surgical versus nonsurgical interventions. METHODS:Three phases comprise the construction and evaluation of a new web-based DA on SKM treatment. In phase 1, this DA was developed in print format through a multidisciplinary design committee incorporating patient focus groups. Phase 2 was an observational study on patient knowledge and decision-making measures after randomization to receive the printed DA or institutional educational materials, which identified further educational needs applied to a web-based DA. Phase 3 will preliminarily evaluate the web-based DA: in a pilot RCT, 50 adults diagnosed with SKMs will receive the web-based DA or an existing web-based institutional website at urology clinics at a large academic medical center. The web-based DA applies risk communication and information about diagnosis and treatment options, elicits preferences regarding treatment options, and provides a set of options to consider with their doctor based on a decision-analytic model of benefits/harm analysis that accounts for comorbidity, age group, and tumor features. Questionnaires and treatment decision data will be gathered before and after viewing the educational material. RESULTS:This phase will consist of a pilot RCT from August 2022 to January 2023 to establish feasibility and preliminarily evaluate decision outcomes. Previous study phases from 2018 to 2020 supported the feasibility of providing the printed DA in urology clinics before clinical consultation and demonstrated increased patient knowledge about the diagnosis and treatment options and greater likelihood of favoring nonsurgical treatment just before consultation. This study was funded by the National Cancer Institute. Recruitment will begin in August 2022. CONCLUSIONS:A web-based DA has been designed to address educational needs for patients making treatment decisions for SKM, accounting for comorbidities and treatment-related benefits and risks. Outcomes from the pilot trial will evaluate the potential of a web-based DA in personalizing treatment decisions and in helping patients weigh attributes of surgical versus nonsurgical treatment options for their SKMs. TRIAL REGISTRATION/BACKGROUND:ClinicalTrials.gov NCT05387863; https://clinicaltrials.gov/ct2/show/NCT05387863. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)/UNASSIGNED:PRR1-10.2196/41451.

Importance:Given higher reimbursement rates, hospitals primarily serving privately insured patients may invest more in intensive coding than hospitals serving publicly insured patients. This may lead these hospitals to code more diagnoses for all patients. Objective:To estimate whether, for the same Medicaid enrollee with multiple hospitalizations, a hospital's share of privately insured patients is associated with the number of diagnoses on claims. Design, Setting, and Participants:This cross-sectional study used patient-level fixed effects regression models on inpatient Medicaid claims from Medicaid enrollees with at least 2 admissions in at least 2 different hospitals in New York State between 2010 and 2017. Analyses were conducted from 2019 to 2021. Exposures:The annual share of privately insured patients at the admitting hospital. Main Outcomes and Measures:Number of diagnostic codes per admission. Probability of diagnoses being from a list of conditions shown to be intensely coded in response to payment incentives. Results:This analysis included 1 614 630 hospitalizations for Medicaid-insured patients (mean [SD] age, 48.2 [20.1] years; 829 684 [51.4%] women and 784 946 [48.6%] men). Overall, 74 998 were Asian (4.6%), 462 259 Black (28.6%), 375 591 Hispanic (23.3%), 486 313 White (30.1%), 128 896 unknown (8.0%), and 86 573 other (5.4%). When the same patient was seen in a hospital with a higher share of privately insured patients, more diagnoses were recorded (0.03 diagnoses per percentage point [pp] increase in share of privately insured; 95% CI, 0.02-0.05; P < .001). Patients discharged from hospitals in the bottom quartile of privately insured patient share received 1.37 more diagnoses when they were subsequently discharged from hospitals in the top quartile, relative to patients whose admissions were both in the bottom quartile (95% CI, 1.21-1.53; P < .001). Those going from hospitals in the top quartile to the bottom had 1.67 fewer diagnoses (95% CI, -1.84 to -1.50; P < .001). Diagnoses in hospitals with a higher private payer share were more likely to be for conditions sensitive to payment incentives (0.08 pp increase for each pp increase in private share; 95% CI, 0.06-0.10; P < .001). These findings were replicated in 2016 to 2017 data. Conclusions and Relevance:In this cross-sectional study of Medicaid enrollees, admission to a hospital with a higher private payer share was associated with more diagnoses on Medicaid claims. This suggests payment policy may drive differential investments in infrastructure to document diagnoses. This may create a feedback loop that exacerbates resource inequity.

Violations of the positivity assumption (also called the common support condition) challenge health policy research and can result in significant bias, large variance, and invalid inference. We define positivity in the single- and multiple-timepoint (i.e., longitudinal) health policy evaluation setting, and discuss real-world threats to positivity. We show empirical evidence of the practical positivity violations that can result when attempting to estimate the effects of health policies (in this case, Naloxone Access Laws). In such scenarios, an alternative is to estimate the effect of a shift in law enactment (e.g., the effect if enactment had been delayed by some number of years). Such an effect corresponds to what is called a modified treatment policy, and dramatically weakens the required positivity assumption, thereby offering a means to estimate policy effects even in scenarios with serious positivity problems. We apply the approach to define and estimate the longitudinal effects of Naloxone Access Laws on opioid overdose rates.

Infant exposure to per/polyfluoroalkyl compounds is associated with immune disruption. We examined associations between neonatal concentrations of perflurooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) and immunoglobulin (Ig) isotype profiles in a prospective cohort of infants. We measured Ig isotypes, including IgA, IgE, IgM and the IgG subclasses IgG₁, IgG₂, IgG₃, and IgG_4, and PFOA and PFOS in newborn dried bloodspots from N = 3175 infants in the Upstate KIDS Study (2008-2010). We examined the association between newborn Ig isotype levels and individual PFOS and PFOA concentrations using mixed effects regression models with a random intercept to account for twins among study participants. We assessed the joint effect PFOA and PFOS with quantile-based g-computation on all singletons and one randomly selected twin (N = 2901), with Ig categorized as above or below median value. Models were adjusted for infant sex, and maternal pre-pregnancy body mass index, race, parity, age and infertility treatment. In adjusted models, PFOA was inversely associated with IgE (coefficient = -0.12 per unit increase in PFOA, 95% CI: -0.065, -0.17), whereas IgG₂, IgM, and IgA were positively associated with PFOA (coefficient for IgG₂ = 0.22, 95% CI: 0.15, 0.27; coefficient for IgM = 0.11, 95% CI: 0.08, 0.15; and coefficient for IgA = 0.15, 95% CI: 0.07, 0.18). There was no relation between PFOS and Ig isotypes. Analysis of the joint effect of PFOA and PFOS showed an OR of 1.2 (95% CI: 1.04, 1.36) for IgA and OR of 1.12 (95% CI: 1.00, 1.24) for IgG₂ levels above the median for every quartile increase. PFOA levels were significantly associated with elevated IgA, IgM, IgG₂, and reduced levels of IgE in single-pollutant models. A small but significant joint effect of PFOA and PFOS was observed. Our results suggest that early exposure to PFOA and PFOS may disrupt neonatal immunoglobulin levels.

OBJECTIVE:To predict adverse kidney outcomes for use in optimizing medical management and clinical trial design. RESEARCH DESIGN AND METHODS/METHODS:In this meta-analysis of individual participant data, 43 cohorts (N = 1,621,817) from research studies, electronic medical records, and clinical trials with global representation were separated into development and validation cohorts. Models were developed and validated within strata of diabetes mellitus (presence or absence) and estimated glomerular filtration rate (eGFR; ≥60 or <60 mL/min/1.73 m2) to predict a composite of ≥40% decline in eGFR or kidney failure (i.e., receipt of kidney replacement therapy) over 2-3 years. RESULTS:There were 17,399 and 24,591 events in development and validation cohorts, respectively. Models predicting ≥40% eGFR decline or kidney failure incorporated age, sex, eGFR, albuminuria, systolic blood pressure, antihypertensive medication use, history of heart failure, coronary heart disease, atrial fibrillation, smoking status, and BMI, and, in those with diabetes, hemoglobin A1c, insulin use, and oral diabetes medication use. The median C-statistic was 0.774 (interquartile range [IQR] = 0.753, 0.782) in the diabetes and higher-eGFR validation cohorts; 0.769 (IQR = 0.758, 0.808) in the diabetes and lower-eGFR validation cohorts; 0.740 (IQR = 0.717, 0.763) in the no diabetes and higher-eGFR validation cohorts; and 0.750 (IQR = 0.731, 0.785) in the no diabetes and lower-eGFR validation cohorts. Incorporating the previous 2-year eGFR slope minimally improved model performance, and then only in the higher-eGFR cohorts. CONCLUSIONS:Novel prediction equations for a decline of ≥40% in eGFR can be applied successfully for use in the general population in persons with and without diabetes with higher or lower eGFR.