Faculty Bibliography

BACKGROUND:Missing data is a pervasive problem in longitudinal data analysis. Several single-imputation (SI) and multiple-imputation (MI) approaches have been proposed to address this issue. In this study, for the first time, the function of the longitudinal regression tree algorithm as a non-parametric method after imputing missing data using SI and MI was investigated using simulated and real data. METHOD/METHODS:Using different simulation scenarios derived from a real data set, we compared the performance of cross, trajectory mean, interpolation, copy-mean, and MI methods (27 approaches) to impute missing longitudinal data using parametric and non-parametric longitudinal models and the performance of the methods was assessed in real data. The real data included 3,645 participants older than 18 years within six waves obtained from the longitudinal Tehran cardiometabolic genetic study (TCGS). The data modeling was conducted using systolic and diastolic blood pressure (SBP/DBP) as the outcome variables and included predictor variables such as age, gender, and BMI. The efficiency of imputation approaches was compared using mean squared error (MSE), root-mean-squared error (RMSE), median absolute deviation (MAD), deviance, and Akaike information criteria (AIC). RESULTS:The longitudinal regression tree algorithm outperformed based on the criteria such as MSE, RMSE, and MAD than the linear mixed-effects model (LMM) for analyzing the TCGS and simulated data using the missing at random (MAR) mechanism. Overall, based on fitting the non-parametric model, the performance of the 27 imputation approaches was nearly similar. However, the SI traj-mean method improved performance compared with other imputation approaches. CONCLUSION/CONCLUSIONS:Both SI and MI approaches performed better using the longitudinal regression tree algorithm compared with the parametric longitudinal models. Based on the results from both the real and simulated data, we recommend that researchers use the traj-mean method for imputing missing values of longitudinal data. Choosing the imputation method with the best performance is widely dependent on the models of interest and the data structure.

Persons living with advanced cancer have intensive symptoms and psychosocial needs that often result in visits to the Emergency Department (ED). We report on program engagement, advance care planning (ACP), and hospice use for a 6-month longitudinal nurse-led, telephonic palliative care intervention for patients with advanced cancer as part of a larger randomized trial. Patients 50 years and older with metastatic solid tumors were recruited from 18 EDs and randomized to receive nursing calls focused on ACP, symptom management, and care coordination or specialty outpatient palliative care (ClinicialTrials.gov: NCT03325985). One hundred and five (50%) graduated from the 6-month program, 54 (26%) died or enrolled in hospice, 40 (19%) were lost to follow-up, and 19 (9%) withdrew prior to program completion. In a Cox proportional hazard regression, withdrawn subjects were more likely to be white and have a low symptom burden compared to those who did not withdraw. Two hundred eighteen persons living with advanced cancer were enrolled in the nursing arm, and 182 of those (83%) completed some ACP. Of the subjects who died, 43/54 (80%) enrolled in hospice. Our program demonstrated high rates of engagement, ACP, and hospice enrollment. Enrolling subjects with a high symptom burden may result in even greater program engagement.

BACKGROUND AND AIMS/OBJECTIVE:Opioid overdose is a leading cause of death during the immediate time after release from jail or prison. Most jails in the United States do not provide methadone and buprenorphine treatment for opioid use disorder (MOUD), and research in estimating its impact in jail settings is limited. We aimed to test the hypothesis that in-jail MOUD is associated with lower overdose mortality risk post-release. DESIGN, SETTING AND PARTICIPANTS/METHODS:Retrospective, observational cohort study of 15 797 adults with opioid use disorder who were released from New York City jails to the community in 2011-17. They experienced 31 382 incarcerations and were followed up to 1 year. MEASUREMENTS/METHODS:The primary outcomes were death caused by accidental drug poisoning and all-cause death. The exposure was receipt of MOUD (17 119 events) versus out-of-treatment (14 263 events) during the last 3 days before community reentry. Covariates included demographic, clinical, behavioral, housing, healthcare utilization, and legal characteristics variables. We performed multivariable, mixed-effect Cox regression analysis to test association between in-jail MOUD and deaths. FINDINGS/RESULTS:A majority were male (82%) and their average age was 42 years. Receiving MOUD was associated with misdemeanor charges, being female, injection drug use, and homelessness. During 1 year post-release, 111 overdose deaths occurred, and crude death rates were 0.49 and 0.83 per 100 person-years for in-jail MOUD and out-of-treatment groups, respectively. Accounting for confounding and random effects, in-jail MOUD was associated with lower overdose mortality risk (adjusted hazard ratio = 0.20, 95% CI = 0.08-0.46), and all-cause mortality risk (adjusted hazard ratio = 0.22, 95% CI = 0.11-0.42) for the first month post-release. CONCLUSIONS:Methadone and buprenorphine treatment for opioid use disorder during incarceration was associated with an 80% reduction in overdose mortality risk for the first month post-release.

BACKGROUND:Blood pressure (BP) control is suboptimal in minority communities, including Asian populations. We evaluate the feasibility, adoption, and effectiveness of an integrated CHW-led health coaching and practice-level intervention to improve hypertension control among South Asian patients in New York City, Project IMPACT (Integrating Million Hearts for Provider and Community Transformation). The primary outcome was BP control, and secondary outcomes were systolic BP and diastolic BP at 6-month follow-up. METHODS:A randomized-controlled trial took place within community-based primary care practices that primarily serve South Asian patients in New York City between 2017 and 2019. A total of 303 South Asian patients aged 18-85 with diagnosed hypertension and uncontrolled BP (systolic BP ≥140 mm Hg or diastolic BP ≥90 mm Hg) within the previous 6 months at 14 clinic sites consented to participate. After completing 1 education session, individuals were randomized into treatment (n=159) or control (n=144) groups. Treatment participants received 4 additional group education sessions and individualized health coaching over a 6-month period. A mixed effect generalized linear model with a logit link function was used to assess intervention effectiveness for controlled hypertension (Yes/No), adjusting for practice level random effect, age, sex, baseline systolic BP, and days between BP measurements. RESULTS:<0.001). In final adjusted analysis, treatment group participants had 3.7 [95% CI, 2.1-6.5] times the odds of achieving BP control at follow-up compared with the control group. CONCLUSIONS:A CHW-led health coaching intervention was effective in achieving BP control among South Asian Americans in New York City primary care practices. Findings can guide translation and dissemination of this model across other communities experiencing hypertension disparities. REGISTRATION/BACKGROUND:URL: https://www. CLINICALTRIALS/RESULTS:gov; Unique identifier: NCT03159533.

BACKGROUND:Numerous clinical trials have been initiated to find effective treatments for COVID-19. These trials have often been initiated in regions where the pandemic has already peaked. Consequently, achieving full enrollment in a single trial might require additional COVID-19 surges in the same location over several years. This has inspired us to pool individual patient data (IPD) from ongoing, paused, prematurely-terminated, or completed randomized controlled trials (RCTs) in real-time, to find an effective treatment as quickly as possible in light of the pandemic crisis. However, pooling across trials introduces enormous uncertainties in study design (e.g., the number of RCTs and sample sizes might be unknown in advance). We sought to develop a versatile treatment efficacy assessment model that accounts for these uncertainties while allowing for continuous monitoring throughout the study using Bayesian monitoring techniques. METHODS:We provide a detailed look at the challenges and solutions for model development, describing the process that used extensive simulations to enable us to finalize the analysis plan. This includes establishing prior distribution assumptions, assessing and improving model convergence under different study composition scenarios, and assessing whether we can extend the model to accommodate multi-site RCTs and evaluate heterogeneous treatment effects. In addition, we recognized that we would need to assess our model for goodness-of-fit, so we explored an approach that used posterior predictive checking. Lastly, given the urgency of the research in the context of evolving pandemic, we were committed to frequent monitoring of the data to assess efficacy, and we set Bayesian monitoring rules calibrated for type 1 error rate and power. RESULTS:The primary outcome is an 11-point ordinal scale. We present the operating characteristics of the proposed cumulative proportional odds model for estimating treatment effectiveness. The model can estimate the treatment's effect under enormous uncertainties in study design. We investigate to what degree the proportional odds assumption has to be violated to render the model inaccurate. We demonstrate the flexibility of a Bayesian monitoring approach by performing frequent interim analyses without increasing the probability of erroneous conclusions. CONCLUSION:This paper describes a translatable framework using simulation to support the design of prospective IPD meta-analyses.

BACKGROUND:The emergency department (ED) is a critical juncture in the care of persons living with dementia (PLwD), as they have a high rate of hospital admission, ED revisits, and subsequent inpatient stays. We examine ED disposition of PLwD compared with older adults with non-dementia chronic disease as well as healthcare utilization and survival. METHODS:Medicare claims data were used to identify community-dwelling older adults 66+ years old from 34 hospitals with either Alzheimer's disease/Alzheimer's disease related dementias (AD/ADRD) or a non-AD/ADRD chronic condition between January 1, 2014, and December 31, 2018. We compared ED disposition at the index visit, as well as healthcare utilization and mortality in the 12 months following an index ED visit, and adjusted for age, gender, and risk of mortality. RESULTS:There were 29,626 patients in the AD/ADRD sample, and 317,046 in the comparison sample. The AD/ADRD sample was older (82.4 years old [SD: 8.2] vs. 76.0 years old [SD: 7.7]) and had more female patients (59.9% vs. 54.7%). The AD/ADRD sample was more likely to experience ED disposition to acute care (OR 1.039, p < 0.001, 95% CI 1.029-1.050), to have an ED revisit (OR 1.077, p < 0.001, 95% CI 1.066-1.087), and an inpatient stay in the subsequent 12 months (OR 1.085, p < 0.001, 95% CI 1.075-1.095). ED disposition to hospice was low in both samples (0.2%). AD/ADRD patients had a higher risk of mortality (OR 1.099, p < 0.001, 95% CI 1.091-1.107) and high short-term mortality (31.9% within 12 months) than those without AD/ADRD (15.3% within 12 months). CONCLUSIONS:PLwD who visit the ED have high short-term mortality. Despite this, disposition to acute care, ED revisits, and inpatient stays, rather than hospice, remain the predominant mode of care delivery. Transition directly from the ED to hospice for PLwD is rare.

Background: Integrative medicine (IM) is a growing subspecialty among the American healthcare system, but little is known about geographical and sociodemographic variability in access to services. Objective: To better understand access to IM healthcare services, we aim to: 1.) document the number of IM providers within the hospital service area (HSA) of various hospitals across the United States (US) and, 2.) explore the relationship between age, income, and race as predictors of the number of IM providers. Methods: We conducted an environmental scan to document the number of IM providers including naturopathic, acupuncture, chiropractic, and massage therapy providers within the HSA of 16 US hospitals using state and national search databases. We examined predictors of the number of providers per HSA using population and demographic data from the U.S. Census Bureau. Search database quality was evaluated using the Center for Disease Control and Prevention Clear Communication Index. Results: The number of IM providers varied from 11.6 "“ 67.4 providers/100,000 persons. Massage therapists were the most prevalent (n = 13.8/100,000), followed by chiropractors (n = 5.2/100,000), acupuncturists (n = 4.6/100,000), and naturopathic physicians (n =.5/100,000). Higher average household income and population >65 years old were associated with more IM providers within an HSA (Rate Ratio (RR) 4.22, 95% CI 1.49-12.01; and 1.14, 1.05 - 1.24, respectively). In addition, the quality of publicly available search databases varied widely among US states (4.84 - 8.00/10), but less so among IM provider types (6.21 - 7.57/10). Conclusions: The high variability in number of IM providers and search database quality among various HSAs across the US warrants further investigation into factors influencing access to services. Our findings regarding income and older adult population raise concern for inequitable access to care, but are also promising when considering the increasing demand for healthcare services among the older adult population.

BACKGROUND:In 2021, Delta became the predominant SARS-CoV-2 variant worldwide. While vaccines have effectively prevented COVID-19 hospitalization and death, vaccine breakthrough infections increasingly occurred. The precise role of clinical and genomic determinants in Delta infections is not known, and whether they contributed to increased rates of breakthrough infections compared to unvaccinated controls. METHODS:We studied SARS-CoV-2 variant distribution, dynamics, and adaptive selection over time in relation to vaccine status, phylogenetic relatedness of viruses, full genome mutation profiles, and associated clinical and demographic parameters. FINDINGS/RESULTS:We show a steep and near-complete replacement of circulating variants with Delta between May and August 2021 in metropolitan New York. We observed an increase of the Delta sublineage AY.25 (14% in vaccinated, 7% in unvaccinated), its spike mutation S112L, and AY.44 (8% in vaccinated, 2% in unvaccinated) with its nsp12 mutation F192V in breakthroughs. Delta infections were associated with younger age and lower hospitalization rates than Alpha. Delta breakthrough infections increased significantly with time since vaccination, and, after adjusting for confounders, they rose at similar rates as in unvaccinated individuals. INTERPRETATION/CONCLUSIONS:We observed a modest adaptation of Delta genomes in breakthrough infections in New York, suggesting an improved genomic framework to support Delta's epidemic growth in times of waning vaccine protection despite limited impact on vaccine escape. FUNDING/BACKGROUND:The study was supported by NYU institutional funds. The NYULH Genome Technology Center is partially supported by the Cancer Center Support Grant P30CA016087 at the Laura and Isaac Perlmutter Cancer Center.

INTRODUCTION/BACKGROUND:Effective communication is essential to good health care, and hearing loss disrupts patient-provider communication. For the more than 2 million veterans with severe hearing loss, communication is particularly challenging in noisy health care environments such as emergency departments. The purpose of this qualitative study was to describe patient and provider perspectives of feasibility and potential benefit of providing a hearing assistance device, a personal amplifier, during visits to an emergency department in an urban setting affiliated with the Department of Veterans Affairs. METHODS:This qualitative descriptive study was conducted in parallel with a randomized controlled study. We completed a semistructured interview with 11 veterans and 10 health care providers to elicit their previous experiences with patient-provider communication in the ED setting and their perspectives on hearing screening and using the personal amplifier in the emergency department. Interview data were analyzed using content analysis and Atlas.ti V8.4 software (Scientific Software Development GmbH, Berlin, Germany). RESULTS:The veteran sample (n = 11) had a mean age of 80.3 years (SD = 10.2). The provider sample included 7 nurses and 3 physicians. In the ED setting, hearing loss disrupts patient-provider communication. Screening for hearing loss in the emergency department was feasible except in urgent/emergent cases. The use of the personal amplifier made communication more effective and less effortful for both veterans and providers. DISCUSSION/CONCLUSIONS:Providing the personal amplifier improved the ED experience for veterans and offers a promising intervention that could improve health care quality and safety for ED patient populations.