Faculty Bibliography

Extensive empirical health research leverages variation in the timing and location of policy changes as quasi-experiments. Multiple social policies may be adopted simultaneously in the same locations, creating co-occurrence which must be addressed analytically for valid inferences. The pervasiveness and consequences of co-occurring policies have received limited attention. We analyzed a systematic sample of 13 social policy databases covering diverse domains including poverty, paid family leave, and tobacco. We quantified policy co-occurrence in each database as the fraction of variation in each policy measure across different jurisdictions and times that could be explained by co-variation with other policies (R2). We used simulations to estimate the ratio of the variance of effect estimates under the observed policy co-occurrence to variance if policies were independent. Policy co-occurrence ranged from very high for state-level cannabis policies to low for country-level sexual minority rights policies. For 65% of policies, greater than 90% of the place-time variation was explained by other policies. Policy co-occurrence increased the variance of effect estimates by a median of 57-fold. Co-occurring policies are common and pose a major methodological challenge to rigorously evaluating health effects of individual social policies. When uncontrolled, co-occurring policies confound one another, and when controlled, resulting positivity violations may substantially inflate the variance of estimated effects. Tools to enhance validity and precision for evaluating co-occurring policies are needed.

Men diagnosed with low-risk prostate cancer (PC) are increasingly electing active surveillance (AS) as their initial management strategy. While this may reduce the side effects of treatment for prostate cancer, many men on AS eventually convert to active treatment. PC is one of the most heritable cancers, and genetic factors that predispose to aggressive tumors may help distinguish men who are more likely to discontinue AS. To investigate this, we undertook a multi-institutional genome-wide association study (GWAS) of 5,222 PC patients and 1,139 other patients from replication cohorts, all of whom initially elected AS and were followed over time for the potential outcome of conversion from AS to active treatment. In the GWAS we detected 18 variants associated with conversion, 15 of which were not previously associated with PC risk. With a transcriptome-wide association study (TWAS), we found two genes associated with conversion (MAST3, p = 6.9×10^-7 and GAB2, p = 2.0×10^-6). Moreover, increasing values of a previously validated 269-variant genetic risk score (GRS) for PC was positively associated with conversion (e.g., comparing the highest to the two middle deciles gave a hazard ratio [HR] = 1.13; 95% Confidence Interval [CI]= 0.94-1.36); whereas, decreasing values of a 36-variant GRS for prostate-specific antigen (PSA) levels were positively associated with conversion (e.g., comparing the lowest to the two middle deciles gave a HR = 1.25; 95% CI, 1.04-1.50). These results suggest that germline genetics may help inform and individualize the decision of AS-or the intensity of monitoring on AS-versus treatment for the initial management of patients with low-risk PC.

OBJECTIVE:To evaluate the healthcare resource utilisation for chronic kidney disease (CKD) and other major non-communicable chronic diseases (NCDs) in China. DESIGN:A cross-sectional study. SETTING:A national inpatient database of tertiary hospitals in China. PARTICIPANTS:The study included a total of 19.5 million hospitalisations of adult patients from July 2013 to June 2014. Information on CKD and other major NCDs, including coronary heart disease (CHD), stroke, hypertension, diabetes, chronic obstructive pulmonary disease (COPD) and cancer, was extracted from the unified discharge summary form. OUTCOME MEASURES:Cost, length of hospital stay and in-hospital mortality. RESULTS:The percentages of hospitalisations with CKD, CHD, stroke, hypertension, diabetes, COPD and cancer were 4.5%, 9.2%, 8.2%, 18.8%, 7.9%, 2.3% and 19.4%, respectively. For each major NCD, the presence of CKD was independently associated with longer hospital stay, with increased percentages ranging from 7.69% (95% CI 7.11% to 8.28%) for stroke to 21.60% (95% CI 21.09% to 22.10%) for CHD. Hospital mortality for other NCDs was also higher in the presence of CKD, with fully adjusted relative risk ranging from 1.91 (95% CI 1.82 to 1.99) for stroke to 2.65 (95% CI 2.55 to 2.75) for cancer. Compared with other NCDs, CKD was associated with the longest hospital stay (22.1% increase) and resulted in the second highest in-hospital mortality, only lower than that of cancer (relative risk, 2.23 vs 2.87, respectively). CONCLUSIONS:The presence of diagnosed CKD alongside each major NCD was associated with an additional burden on the healthcare system. Healthcare resource utilisation and prognosis of CKD were comparable with those of other major NCDs, which highlights the importance of CKD as a major public health burden.

BACKGROUND:Accumulating evidence links ultra-processed foods to poor diet quality and chronic diseases. Understanding dietary trends is essential to inform priorities and policies to improve diet quality and prevent diet-related chronic diseases. Data are lacking, however, for trends in ultra-processed food intake. OBJECTIVES/OBJECTIVE:We examined US secular trends in food consumption according to processing level from 2001 to 2018. METHODS:We analyzed dietary data collected by 24-h recalls from adult participants (aged >19 y; N  = 40,937) in 9 cross-sectional waves of the NHANES (2001-2002 to 2017-2018). We calculated participants' intake of minimally processed foods, processed culinary ingredients, processed foods, and ultra-processed foods as the relative contribution to daily energy intake (%kcal) using the NOVA framework. Trends analyses were performed using linear regression, testing for linear trends by modeling the 9 surveys as an ordinal independent variable. Models were adjusted for age, sex, race/ethnicity, education level, and income. Consumption trends were reported for the full sample and stratified by sex, age groups, race/ethnicity, education level, and income level. RESULTS:Adjusting for changes in population characteristics, the consumption of ultra-processed foods increased among all US adults from 2001-2002 to 2017-2018 (from 53.5 to 57.0 %kcal; P-trend < 0.001). The trend was consistent among all sociodemographic subgroups, except Hispanics, in stratified analyses. In contrast, the consumption of minimally processed foods decreased significantly over the study period (from 32.7 to 27.4 %kcal; P-trend < 0.001) and across all sociodemographic strata. The consumption of processed culinary ingredients increased from 3.9 to 5.4 %kcal (P-trend < 0.001), whereas the intake of processed foods remained stable at ∼10 %kcal throughout the study period (P-trend = 0.052). CONCLUSIONS:The current findings highlight the high consumption of ultra-processed foods in all parts of the US population and demonstrate that intake has continuously increased in the majority of the population in the past 2 decades.

OBJECTIVE:Rapid automatized naming (RAN) tasks have been utilized for decades to evaluate neurological conditions. Time scores for the Mobile Universal Lexicon Evaluation System (MULES, rapid picture naming) and Staggered Uneven Number (SUN, rapid number naming) are prolonged (worse) with concussion, mild cognitive impairment, multiple sclerosis and Parkinson's disease. The purpose of this investigation was to compare paper/pencil versions of MULES and SUN with a new digitized format, the MICK app. METHODS:Participants (healthy office-based volunteers, professional women's hockey players), completed two trials of the MULES and SUN tests on both platforms (tablet, paper/pencil). The order of presentation of the testing platforms was randomized. Between-platform variability was calculated using the two-way random-effects intraclass correlation coefficient (ICC). RESULTS:Among 59 participants (median age 32, range 22-83), no significant differences were observed for comparisons of mean best scores for the paper/pencil versus MICK app platforms, counterbalanced for order of administration (P = 0.45 for MULES, P = 0.50 for SUN, linear regression). ICCs for agreement between the MICK and paper/pencil tests were 0.92 (95% CI 0.86, 0.95) for MULES and 0.94 (95% CI 0.89, 0.96) for SUN, representing excellent levels of agreement. Inter-platform differences did not vary systematically across the range of average best time score for either test. CONCLUSION/CONCLUSIONS:The MICK app for digital administration of MULES and SUN demonstrates excellent agreement of time scores with paper/pencil testing. The computerized app allows for greater accessibility and scalability in neurological diseases, inclusive of remote monitoring. Sideline testing for sports-related concussion may also benefit from this technology.

Infant and toddler MRI enables unprecedented insight into the developing brain. However, consensus about optimal data collection practices is lacking, which slows growth of the field and impedes replication efforts. The goal of this study was to collect systematic data across a large number of infant/toddler research laboratories to better understand preferred practices. Survey data addressed MRI acquisition strategies, scan success rates, visit preparations, scanning protocols, accommodations for families, study design, and policies regarding incidental findings. Respondents had on average 8 years' experience in early life neuroimaging and represented more than fifty research laboratories. Areas of consensus across labs included higher success rates among newborns compared to older infants or toddlers, high rates of data loss across age groups, endorsement of multiple layers of hearing protection, and age-specific scan preparation and participant accommodation. Researchers remain divided on decisions in longitudinal study design and practices regarding incidental findings. This study summarizes practices honed over years of work by a large collection of scientists, which may serve as an important resource for those new to the field. The ability to reference data about best practices facilitates future harmonization, data sharing, and reproducibility, all of which advance this important frontier in developmental science.

First-person accounts of COVID-19 illness and treatment can complement and enrich data derived from electronic medical or public health records. With patient-reported data, it is uniquely possible to ascertain in-depth contextual information as well as behavioral and emotional responses to illness. The Novel Coronavirus Illness Patient Report (NCIPR) dataset includes complete survey responses from 1,584 confirmed COVID-19 patients ages 18 to 98. NCIPR survey questions address symptoms, medical complications, home and hospital treatments, lasting effects, anxiety about illness, employment impacts, quarantine behaviors, vaccine-related behaviors and effects, and illness of other family/household members. Additional questions address financial security, perceived discrimination, pandemic impacts (relationship, social, stress, sleep), health history, and coping strategies. Detailed patient reports of illness, environment, and psychosocial impact, proximal to timing of infection and considerate of demographic variation, is meaningful for understanding pandemic-related public health from the perspective of those that contracted the disease.

BACKGROUND:Persistent cognitive symptoms have been reported following COVID-19 hospitalization. We investigated the relationship between demographics, social determinants of health (SDOH) and cognitive outcomes 6-months after hospitalization for COVID-19. METHODS:We analyzed 6-month follow-up data collected from a multi-center, prospective study of hospitalized COVID-19 patients. Demographic and SDOH variables (age, race/ethnicity, education, employment, health insurance status, median income, primary language, living arrangements, and pre-COVID disability) were compared between patients with normal versus abnormal telephone Montreal Cognitive Assessments (t-MOCA; scores<18/22). Multivariable logistic regression models were constructed to evaluate predictors of t-MoCA. RESULTS:Of 382 patients available for 6-month follow-up, 215 (56%) completed the t-MoCA (n = 109/215 [51%] had normal and n = 106/215 [49%] abnormal results). 14/215 (7%) patients had a prior history of dementia/cognitive impairment. Significant univariate predictors of abnormal t-MoCA included older age, ≤12 years of education, unemployment pre-COVID, Black race, and a pre-COVID history of cognitive impairment (all p < 0.05). In multivariable analyses, education ≤12 years (adjusted OR 5.21, 95%CI 2.25-12.09), Black race (aOR 5.54, 95%CI 2.25-13.66), and the interaction of baseline functional status and unemployment prior to hospitalization (aOR 3.98, 95%CI 1.23-12.92) were significantly associated with abnormal t-MoCA scores after adjusting for age, history of dementia, language, neurological complications, income and discharge disposition. CONCLUSIONS:Fewer years of education, Black race and unemployment with baseline disability were associated with abnormal t-MoCA scores 6-months post-hospitalization for COVID-19. These associations may be due to undiagnosed baseline cognitive dysfunction, implicit biases of the t-MoCA, other unmeasured SDOH or biological effects of SARS-CoV-2.

Many older adults with diabetes (DM) have co-occurring Alzheimer's Disease (AD) and AD-Related Dementias (ADRD). Complex treatment plans may impose treatment burden for caregivers responsible for day-to-day self-management. The purpose of this qualitative study was to describe caregiver perceptions of treatment burden for people with DM-AD/ADRD. Caregivers (n = 33) of patients with DM-AD/ADRD participated in semi-structured interviews about their caregiver role and perceptions of treatment burden of DM-AD/ADRD management. Qualitative data were analyzed using content analysis (ATLAS.ti). Caregivers reported high levels of burden related to complex treatment/self-management for patients with DM-AD/ADRD that varied day-to-day with the patient's cognitive status. Four themes were: (1) trajectory of treatment burden; (2) navigating multiple healthcare providers/systems of care; (3) caregiver role conflict; and (4) emotional burden. Interventions to reduce caregiver treatment burden should include activating supportive services, education, and care coordination especially, if patient treatment increases in complexity over time.