Linking census data with electronic medical records for clinical research: A systematic review

Colorectal Cancer Screening Disparities Among Race: A Zip Code Level Analysis

BACKGROUND

Colorectal cancer (CRC) screening can prevent disease by early identification. Existing disparities in CRC screening have been associated with factors including race, socioeconomic status, insurance, and even geography. Our study takes a deeper look into how social determinants related to zip code tabulation areas affect CRC screenings.

MATERIALS AND METHODS

We conducted a retrospective cross-sectional study of CRC screenings by race at a zip code level, evaluating for impactful social determinant factors such as the social deprivation index (SDI). We used publicly available data from CDC 500 Cities Project (2016-2019), PLACES Project (2020), and the American Community Survey (2019). We conducted multivariate and confirmatory factor analyses among race, income, health insurance, check-up visits, and SDI.

RESULTS

Increasing the tertile of SDI was associated with a higher likelihood of being Black or Hispanic, as well as decreased median household income (P < .01). Lower rates of regular checkup visits were found in the third tertile of SDI (P < .01). The multivariate analysis showed that being Black, Hispanic, lower income, being uninsured, lack of regular check-ups, and increased SDI were related to decreased CRC screening. In the confirmatory factor analysis, we found that SDI and access to insurance were the variables most related to decreased CRC screening.

CONCLUSION

Our results reveal the top 2 factors that impact a locality's CRC screening rates are the social deprivation index and access to health care. This data may help implement interventions targeting social barriers to further promote CRC screenings within disadvantaged communities and decrease overall mortality via early screening.

Systematic review of approaches to use of neighborhood-level risk factors with clinical data to predict clinical risk and recommend interventions

BACKGROUND

Despite a large body of literature investigating how the environment influences health outcomes, most published work to date includes only a limited subset of the rich clinical and environmental data that is available and does not address how these data might best be used to predict clinical risk or expected impact of clinical interventions.

OBJECTIVE

Identify existing approaches to inclusion of a broad set of neighborhood-level risk factors with clinical data to predict clinical risk and recommend interventions.

METHODS

A systematic review of scientific literature published and indexed in PubMed, Web of Science, Association of Computing Machinery (ACM) and SCOPUS from 2010 through October 2020 was performed. To be included, articles had to include search terms related to Electronic Health Record (EHR) data Neighborhood-Level Risk Factors (NLRFs), and Machine Learning (ML) Methods. Citations of relevant articles were also reviewed for additional articles for inclusion. Articles were reviewed and coded by two independent reviewers to capture key information including data sources, linkage of EHR to NRLFs, methods, and results. Articles were assessed for quality using a modified Quality Assessment Tool for Systematic Reviews of Observational Studies (QATSO).

RESULTS

A total of 334 articles were identified for abstract review. 36 articles were identified for full review with 19 articles included in the final analysis. All but two of the articles included socio-demographic data derived from the U.S. Census and we found great variability in sources of NLRFs beyond the Census. The majority or the articles (14 of 19) included broader clinical (e.g. medications, labs and co-morbidities) and demographic information about the individual from the EHR in addition to the clinical outcome variable. Half of the articles (10) had a stated goal to predict the outcome(s) of interest. While results of the studies reinforced the correlative association of NLRFs to clinical outcomes, only one article found that adding NLRFs into a model with other data added predictive power with the remainder concluding either that NLRFs were of mixed value depending on the model and outcome or that NLRFs added no predictive power over other data in the model. Only one article scored high on the quality assessment with 13 scoring moderate and 4 scoring low.

CONCLUSIONS

In spite of growing interest in combining NLRFs with EHR data for clinical prediction, we found limited evidence that NLRFs improve predictive power in clinical risk models. We found these data and methods are being used in four ways. First, early approaches to include broad NLRFs to predict clinical risk primarily focused on dimension reduction for feature selection or as a data preparation step to input into regression analysis. Second, more recent work incorporates NLRFs into more advanced predictive models, such as Neural Networks, Random Forest, and Penalized Lasso to predict clinical outcomes or predict value of interventions. Third, studies that test how inclusion of NLRFs predict clinical risk have shown mixed results regarding the value of these data over EHR or claims data alone and this review surfaced evidence of potential quality challenges and biases inherent to this approach. Finally, NLRFs were used with unsupervised learning to identify underlying patterns in patient populations to recommend targeted interventions. Further access to computable, high quality data is needed along with careful study design, including sub-group analysis, to better determine how these data and methods can be used to support decision making in a clinical setting.

Social Determinants of Health Mediate COVID-19 Disparities in South Florida

BACKGROUND

There are several reports of health disparities related to COVID-19. Understanding social determinants of health (SDoH) could help develop mitigation strategies to prevent further COVID-19 spread. Our aim is to evaluate self-reported and census-based SDoH as a mediator of health disparities in COVID-19.

METHODS

We conducted a cross-sectional ecological study and included all COVID-19 cases report by the COVID-19 Florida dashboard as the dependent variable. The independent variables were census-based median household income, population and household size, and self-reported SDoH using a validated survey. We calculated the incidence rate ratio (IRR) of COVID-19 by zip code using Poisson regression and structured equation modelling to evaluate the mediation effect of income and SDoH on COVID-19 cases.

RESULTS

We included 97,594 COVID-19 positive cases across 79 Miami-Dade ZIP codes with a median age of 43 years; females represented 50.7% of the cases. The highest IRR (4.44) were for ZIP code 33125 (income $21,106, 6% Black, 93% Hispanic), while the lowest IRR (0.86) was for ZIP code 33146 (median household incomes $96,609, 3% Black and 53% Hispanic). In structured equation models, the indirect coefficient of income in the relationship between race/ethnicity and COVID-19 were only significant for Blacks and not Hispanics.

CONCLUSIONS

This ecological analysis using ZIP code and aggregate individual-level SDoH shows that in Miami-Dade county, COVID infection is associated with economic disadvantage in a particular geographical area and not with racial/ethnic distribution.

Prior cardiovascular risk and screening echocardiograms predict hospitalization and severity of coronavirus infection among elderly medicare patients

Background

The COVID-19 pandemic has disproportionally impacted the elderly. In the United States and Europe the mortality rate of elderly patients with COVID-19 is greater than 30%. Our aim is to determine predictors of COVID-19 related hospitalization and severity of disease among elderly Medicare patients in the United States.

Methods

We conducted a retrospective cohort study including elderly Medicare COVID-19 patients across eight states. We collected data from the inpatient and outpatient electronic health record, demographic, clinical and echocardiographic predictors. Our primary outcomes were hospitalization and adult respiratory distress syndrome (ARDS). Our secondary outcome was mortality.

Results

We identified 400 COVID-19 positive patients (incidence 5.2; (95% CI 4.7-5.7) per 1000 patients). The mean age of our patients was 72 ​± ​8, 60% were female, 82% were minorities and had a mean Charlson score of 2.9 ​± ​1.4. Two-hundred and forty-four patients were hospitalized due to COVID-19 (63%) and the mortality rate was 18%; 95% CI 14-22 with 1 patient still in the hospital. Age, socioeconomic status, Charlson score, systolic blood pressure, body mass index, grade 2 or 3 diastolic dysfunction, moderate or severe left ventricular hypertrophy were significant predictors of hospitalization and ARDS (p ​< ​0.05).

Conclusions

Our study reports a lower incidence on a COVID-19 cohort than previously reported. Predictors of poor outcomes included socio-economic, cardiovascular risk and echocardiographic measures. High touch care with early cardiovascular risk factor modification could explain the low risk of events in our population.

Using Security Questions to Link Participants in Longitudinal Data Collection

Anonymous data collection systems are often necessary when assessing sensitive behaviors but can pose challenges to researchers seeking to link participants over time. To assist researchers in anonymously linking participants, we outlined and tested a novel security question linking (security question linking; SEEK) method. The SEEK method includes four steps: (1) data management and standardization, (2) many-to-many matching, (3) fuzzy matching, and (4) rematching and verification. The method is demonstrated in SAS with two samples from a longitudinal study of adolescent dating violence. After an initial assessment during a laboratory visit, participants were asked to complete an online assessment either (a) once, 3 months later (Sample 1, n = 60), or (b) three times at 1-month intervals (Sample 2, n = 140). Demographics, eye color, and responses to nine security questions were used as key variables to link responses from the laboratory and online follow-up assessments. The rates of matched cases were 100% in Sample 1 and from 94.3 to 98.3% in Sample 2. To quantify the confidence in the data quality of successfully matched pairs, we reported the means and standard deviations of the number of matched security questions. In addition, we reported the rank order and counts of the mismatched components in key variables. Results indicate that the SEEK method provides a feasible and reliable solution to link responses in longitudinal studies with sensitive questions.