Realizing the Potential of Social Determinants Data: A Scoping Review of Approaches for Screening, Linkage, Extraction, Analysis and Interventions.

Background: Social determinants of health (SDoH) like socioeconomics and neighborhoods strongly influence outcomes, yet standardized SDoH data is lacking in electronic health records (EHR), limiting research and care quality. Methods: We searched PubMed using keywords "SDOH" and "EHR", underwent title/abstract and full-text screening. Included records were analyzed under five domains: 1) SDoH screening and assessment approaches, 2) SDoH data collection and documentation, 3) Use of natural language processing (NLP) for extracting SDoH, 4) SDoH data and health outcomes, and 5) SDoH-driven interventions. Results: We identified 685 articles, of which 324 underwent full review. Key findings include tailored screening instruments implemented across settings, census and claims data linkage providing contextual SDoH profiles, rule-based and neural network systems extracting SDoH from notes using NLP, connections found between SDoH data and healthcare utilization/chronic disease control, and integrated care management programs executed. However, considerable variability persists across data sources, tools, and outcomes. Discussion: Despite progress identifying patient social needs, further development of standards, predictive models, and coordinated interventions is critical to fulfill the potential of SDoH-EHR integration. Additional database searches could strengthen this scoping review. Ultimately widespread capture, analysis, and translation of multidimensional SDoH data into clinical care is essential for promoting health equity.

Mining for equitable health: Assessing the impact of missing data in electronic health records.

Electronic health records (EHR) are collected as a routine part of healthcare delivery, and have great potential to be utilized to improve patient health outcomes. They contain multiple years of health information to be leveraged for risk prediction, disease detection, and treatment evaluation. However, they do not have a consistent, standardized format across institutions, particularly in the United States, and can present significant analytical challenges- they contain multi-scale data from heterogeneous domains and include both structured and unstructured data. Data for individual patients are collected at irregular time intervals and with varying frequencies. In addition to the analytical challenges, EHR can reflect inequity- patients belonging to different groups will have differing amounts of data in their health records. Many of these issues can contribute to biased data collection. The consequence is that the data for under-served groups may be less informative partly due to more fragmented care, which can be viewed as a type of missing data problem. For EHR data in this complex form, there is currently no framework for introducing realistic missing values. There has also been little to no work in assessing the impact of missing data in EHR. In this work, we first introduce a terminology to define three levels of EHR data and then propose a novel framework for simulating realistic missing data scenarios in EHR to adequately assess their impact on predictive modeling. We incorporate the use of a medical knowledge graph to capture dependencies between medical events to create a more realistic missing data framework. In an intensive care unit setting, we found that missing data have greater negative impact on the performance of disease prediction models in groups that tend to have less access to healthcare, or seek less healthcare. We also found that the impact of missing data on disease prediction models is stronger when using the knowledge graph framework to introduce realistic missing values as opposed to random event removal.

A comparative analysis of body composition assessment by BIA and DXA in children with type II and III spinal muscular atrophy.

Background: Body composition analysis is a valuable tool for assessing and monitoring the nutritional status of children with spinal muscular atrophy (SMA). This study was designed to compare the consistency of bioelectrical impedance analysis (BIA) and dual-energy X-ray absorptiometry (DXA), as the gold standard method for assessing body composition in clinical practice when treating children with type II and III SMA. Methods: From 2019 to 2021, we performed a retrospective analysis of body composition by DXA and BIA measurement methods in patients with type II and III SMA treated at a Chinese tertiary children's hospital. Fat mass (FM), muscle mass (MM), bone mineral content (BMC), and visceral fat area (VFA) were compared using paired sample t-tests. We calculated Lin's concordance correlation coefficient (CCC) and Spearman correlation coefficient to verify the correlation between DXA and BIA measurements. Bland-Altman analysis was used to assess the consistency of the two methods. Results: Fifty-seven children with type II and III SMA were recruited. Compared with body composition measured by DXA, the average FM measured by BIA is significantly lower (P <0.001), whereas the average MM, BMC, and VFA measured by BIA are significantly higher (P < 0.001) in children with SMA. Overall, the difference between MM (Delta [BIA-DAX] = 1.6 kg) and FM (Delta [BIA-DAX] = -1.6 kg) measured by DXA and BIA was minor, whereas the difference of VFA (Delta [BIA-DAX] = -43.5 cm) was significantly large. Correlation analysis indicated a substantial correlation of MM (CCC = 0.96 [95% confidence interval (CI) = 0.93-0.98], r = 0.967 [P < 0.0001]) and FM (CCC = 0.95 [95% CI = 0.92-0.97], r = 0.953 [P < 0.0001]), and poor correlation of BMC (CCC = 0.61 [95% CI = 0.42-0.75], r = 0.612 [P < 0.0001]) and VFA (CCC = 0.54 [95% CI = 0.33-0.70], r = 0.689 [P < 0.0001]) measurements between the two methods. The Bland-Altman analysis suggests that the majority of participants were within LOA. In addition, differences in MM and VFA measurements between BIA and DAX increased according to patients' increasing height, whereas differences in FM and BMC did not differ with height. Conclusion: BIA overestimates MM and underestimates the FM, BMC, and VFA in children with SMA compared with DXA measurements. Overall, the non-invasive, easy-to-use, and repeatable BIA measurements were found to be in good agreement with DXA measurements, especially for FM and MM, which are essential parameters for the nutritional evaluation of children with SMA.

Model selection and inference for censored lifetime medical expenditures.

Identifying factors associated with increased medical cost is important for many micro- and macro-institutions, including the national economy and public health, insurers and the insured. However, assembling comprehensive national databases that include both the cost and individual-level predictors can prove challenging. Alternatively, one can use data from smaller studies with the understanding that conclusions drawn from such analyses may be limited to the participant population. At the same time, smaller clinical studies have limited follow-up and lifetime medical cost may not be fully observed for all study participants. In this context, we develop new model selection methods and inference procedures for secondary analyses of clinical trial data when lifetime medical cost is subject to induced censoring. Our model selection methods extend a theory of penalized estimating function to a calibration regression estimator tailored for this data type. Next, we develop a novel inference procedure for the unpenalized regression estimator using perturbation and resampling theory. Then, we extend this resampling plan to accommodate regularized coefficient estimation of censored lifetime medical cost and develop postselection inference procedures for the final model. Our methods are motivated by data from Southwest Oncology Group Protocol 9509, a clinical trial of patients with advanced nonsmall cell lung cancer, and our models of lifetime medical cost are specific to this population. But the methods presented in this article are built on rather general techniques and could be applied to larger databases as those data become available.

Multiple imputation in the presence of high-dimensional data.

Missing data are frequently encountered in biomedical, epidemiologic and social research. It is well known that a naive analysis without adequate handling of missing data may lead to bias and/or loss of efficiency. Partly due to its ease of use, multiple imputation has become increasingly popular in practice for handling missing data. However, it is unclear what is the best strategy to conduct multiple imputation in the presence of high-dimensional data. To answer this question, we investigate several approaches of using regularized regression and Bayesian lasso regression to impute missing values in the presence of high-dimensional data. We compare the performance of these methods through numerical studies, in which we also evaluate the impact of the dimension of the data, the size of the true active set for imputation, and the strength of correlation. Our numerical studies show that in the presence of high-dimensional data the standard multiple imputation approach performs poorly and the imputation approach using Bayesian lasso regression achieves, in most cases, better performance than the other imputation methods including the standard imputation approach using the correctly specified imputation model. Our results suggest that Bayesian lasso regression and its extensions are better suited for multiple imputation in the presence of high-dimensional data than the other regression methods.

Variable selection in the presence of missing data: resampling and imputation.

In the presence of missing data, variable selection methods need to be tailored to missing data mechanisms and statistical approaches used for handling missing data. We focus on the mechanism of missing at random and variable selection methods that can be combined with imputation. We investigate a general resampling approach (BI-SS) that combines bootstrap imputation and stability selection, the latter of which was developed for fully observed data. The proposed approach is general and can be applied to a wide range of settings. Our extensive simulation studies demonstrate that the performance of BI-SS is the best or close to the best and is relatively insensitive to tuning parameter values in terms of variable selection, compared with several existing methods for both low-dimensional and high-dimensional problems. The proposed approach is further illustrated using two applications, one for a low-dimensional problem and the other for a high-dimensional problem.

Increased cardiovascular disease, resource use, and costs before the clinical diagnosis of diabetes in veterans in the southeastern U.S.

IMPORTANCE: Screening for diabetes might be more widespread if adverse associations with cardiovascular disease (CVD), resource use, and costs were known to occur earlier than conventional clinical diagnosis. OBJECTIVE: The purpose of this study was to determine whether adverse effects associated with diabetes begin prior to clinical diagnosis. DESIGN: Veterans with diabetes were matched 1:2 with controls by follow-up, age, race/ethnicity, gender, and VA facility. CVD was obtained from ICD-9 codes, and resource use and costs from VA datasets. SETTING: VA facilities in SC, GA, and AL. PARTICIPANTS: Patients with and without diagnosed diabetes. MAIN OUTCOME MEASURES: Diagnosed CVD, resource use, and costs. RESULTS: In this study, the 2,062 diabetic patients and 4,124 controls were 63 years old on average, 99 % male, and 29 % black; BMI was 30.8 in diabetic patients vs. 27.8 in controls (p<0.001). CVD prevalence was higher and there were more outpatient visits in Year -4 before diagnosis through Year +4 after diagnosis among diabetic vs. control patients (all p<0.01); in Year -2, CVD prevalence was 31 % vs. 24 %, and outpatient visits were 22 vs. 19 per year, respectively. Total VA costs/year/veteran were higher in diabetic than control patients from Year -4 ($4,083 vs. $2,754) through Year +5 ($8,347 vs. $5,700) (p<0.003) for each, reflecting underlying increases in outpatient, inpatient, and pharmacy costs (p<0.05 for each). Regression analysis showed that diabetes contributed an average of $1,748/year to costs, independent of CVD (p<0.001). CONCLUSIONS AND RELEVANCE: VA costs per veteran are higher--over $1,000/year before and $2,000/year after diagnosis of diabetes--due to underlying increases in outpatient, inpatient, and pharmacy costs, greater number of outpatient visits, and increased CVD. Moreover, adverse associations with veterans' health and the VA healthcare system occur early in the natural history of the disease, several years before diabetes is diagnosed. Since adverse associations begin before diabetes is recognized, greater consideration should be given to systematic screening in order to permit earlier detection and initiation of preventive management. Keeping frequency of CVD and marginal costs in line with those of patients before diabetes is currently diagnosed has the potential to save up to $2 billion a year.

A nonparametric multiple imputation approach for data with missing covariate values with application to colorectal adenoma data.

A nearest neighbor-based multiple imputation approach is proposed to recover missing covariate information using the predictive covariates while estimating the association between the outcome and the covariates. To conduct the imputation, two working models are fitted to define an imputing set. This approach is expected to be robust to the underlying distribution of the data. We show in simulation and demonstrate on a colorectal data set that the proposed approach can improve efficiency and reduce bias in a situation with missing at random compared to the complete case analysis and the modified inverse probability weighted method.

Epidemiological profile of Clonorchis sinensis infection in one community, Guangdong, People's Republic of China.

BACKGROUND: Clonorchiasis caused by ingesting improperly prepared fish ranks among the most important but still neglected food-borne parasitic diseases, especially in the People's Republic of China (P.R. China). To promote the implementation of interventions efficiently, the demonstration of an epidemiological profile of Clonorchis sinensis infection is essential in hyper-epidemic areas. METHODS: In one community with higher levels of economic development in Guangdong province, P.R. China, villagers were motivated to provide stool samples for examining helminth eggs. Then, those infected with C. sinensis completed the structured questionnaire including demographical characteristics, knowledge and behavior. RESULTS: A total of 293 villagers infected with C. sinensis participated in questionnaire investigation. Among them, 94.54% were adult and 93.17% were indigenous. The geometric mean of C. sinensis eggs per gram of feces in the children, adult females and adult males was 58, 291 and 443, respectively. The divergence between knowledge and behavior in the adults, especially the adult males, was shown. Out of 228 persons eating raw fish, 160 did it more frequently at restaurants, the proportion of which varied in different populations, showing 25.00%, 54.88% and 80.28% in the children, adult females and adult males, respectively. CONCLUSIONS: Different interventions need to be adopted in different populations. Chemotherapy should be prioritized in the adults, especially the adult males. In addition, health education targeting the children, is essential and may play a crucial role in controlling clonorchiasis in the long term. In order to successfully control clonorchiasis, intervention in the restaurant should not be overlooked in some endemic areas.

Screening for diabetes and prediabetes should be cost-saving in patients at high risk.

OBJECTIVE: Although screening for diabetes and prediabetes is recommended, it is not clear how best or whom to screen. We therefore compared the economics of screening according to baseline risk. RESEARCH DESIGN AND METHODS: Five screening tests were performed in 1,573 adults without known diabetes--random plasma/capillary glucose, plasma/capillary glucose 1 h after 50-g oral glucose (any time, without previous fast, plasma glucose 1 h after a 50-g oral glucose challenge [GCTpl]/capillary glucose 1 h after a 50-g oral glucose challenge [GCTcap]), and A1C--and a definitive 75-g oral glucose tolerance test. Costs of screening included the following: costs of testing (screen plus oral glucose tolerance test, if screen is positive); costs for false-negative results; and costs of treatment of true-positive results with metformin, all over the course of 3 years. We compared costs for no screening, screening everyone for diabetes or high-risk prediabetes, and screening those with risk factors based on age, BMI, blood pressure, waist circumference, lipids, or family history of diabetes. RESULTS: Compared with no screening, cost-savings would be obtained largely from screening those at higher risk, including those with BMI >35 kg/m(2), systolic blood pressure ≥130 mmHg, or age >55 years, with differences of up to -46% of health system costs for screening for diabetes and -21% for screening for dysglycemia110, respectively (all P < 0.01). GCTpl would be the least expensive screening test for most high-risk groups for this population over the course of 3 years. CONCLUSIONS: From a health economics perspective, screening for diabetes and high-risk prediabetes should target patients at higher risk, particularly those with BMI >35 kg/m(2), systolic blood pressure ≥130 mmHg, or age >55 years, for whom screening can be most cost-saving. GCTpl is generally the least expensive test in high-risk groups and should be considered for routine use as an opportunistic screen in these groups.

Disability weight of Clonorchis sinensis infection: captured from community study and model simulation.

BACKGROUND: Clonorchiasis is among the most neglected tropical diseases. It is caused by ingesting raw or undercooked fish or shrimp containing the larval of Clonorchis sinensis and mainly endemic in Southeast Asia including China, Korea and Vietnam. The global estimations for population at risk and infected are 601 million and 35 million, respectively. However, it is still not listed among the Global Burden of Disease (GBD) and no disability weight is available for it. Disability weight reflects the average degree of loss of life value due to certain chronic disease condition and ranges between 0 (complete health) and 1 (death). It is crucial parameter for calculating the morbidity part of any disease burden in terms of disability-adjusted life years (DALYs). METHODOLOGY/PRINCIPAL FINDINGS: According to the probability and disability weight of single sequelae caused by C. sinensis infection, the overall disability weight could be captured through Monte Carlo simulation. The probability of single sequelae was gained from one community investigation, while the corresponding disability weight was searched from the literatures in evidence-based approach. The overall disability weights of the male and female were 0.101 and 0.050, respectively. The overall disability weights of the age group of 5-14, 15-29, 30-44, 45-59 and 60+ were 0.022, 0.052, 0.072, 0.094 and 0.118, respectively. There was some evidence showing that the disability weight and geometric mean of eggs per gram of feces (GMEPG) fitted a logarithmic equation. CONCLUSION/SIGNIFICANCE: The overall disability weights of C. sinensis infection are differential in different sex and age groups. The disability weight captured here may be referred for estimating the disease burden of C. sinensis infection.

Estimation of recurrence of colorectal adenomas with dependent censoring using weighted logistic regression.

In colorectal polyp prevention trials, estimation of the rate of recurrence of adenomas at the end of the trial may be complicated by dependent censoring, that is, time to follow-up colonoscopy and dropout may be dependent on time to recurrence. Assuming that the auxiliary variables capture the dependence between recurrence and censoring times, we propose to fit two working models with the auxiliary variables as covariates to define risk groups and then extend an existing weighted logistic regression method for independent censoring to each risk group to accommodate potential dependent censoring. In a simulation study, we show that the proposed method results in both a gain in efficiency and reduction in bias for estimating the recurrence rate. We illustrate the methodology by analyzing a recurrent adenoma dataset from a colorectal polyp prevention trial.

[Cost-effectiveness analysis of integrated control strategy of parasitic diseases in demonstration plots].

OBJECTIVE: To study the cost-effectiveness of different deworming schemes in demonstration plots of integrated control of parasitic diseases. METHODS: The cost-effectiveness was analyzed between mass drug administration and drug administration to focal population according to the different infection rates of parasites. RESULTS: In the demonstration plots of soil-born nematodes control, the costs for reducing one infected case and the cost for reducing 1% infection rate per ten thousand people in mass drug administration groups (Group 1 and Group 2) and drug administration to focal population were 20.73, 14.42 Yuan and 14.33 Yuan, and 1 700.49, 1 503.19 Yuan and 145.41 Yuan, respectively. In the demonstration plots of control of clonorchiasis sinensis, the costs for reducing one infected case and the cost for reducing 1% infection rate per ten thousand people in mass drug administration group and drug administration to focal population were 31.03 Yuan and 37.01 Yuan, and 3 115.10 Yuan and 3841.38 Yuan, respectively. The multiple effectiveness indexes for control of soil-transmitted nematodes in the mass drug administration groups (Group 1 and Group 2) and drug administration to focal population were 76.72, 80.27 and 97.64, respectively. The multiple effectiveness indexes for control of clonorchiasis sinensis in the mass drug administration group and the drug administration to focal population group were 112.93 and 65.49, respectively. CONCLUSIONS: We should choose the deworming schemes not only to get a great reduction of human parasite infection rate and a rapid effective reduction of the source of infection, but also to make a full use of the limited funds on target population.

Identification of a novel CHEK2 variant and assessment of its contribution to the risk of breast cancer in French Canadian women.

BACKGROUND: BRCA1 and BRCA2 account for the majority of the known familial breast cancer risk, however, the impact of other cancer susceptibility genes largely remains to be elucidated. Checkpoint Kinase 2 (CHEK2) is an important signal transducer of cellular responses to DNA damage, whose defects have been associated with an increase in breast cancer risk. Previous studies have identified low penetrance CHEK2 alleles such as 1100delC and I157T, as well as variants such as S428F in the Ashkenazi Jewish population and IVS2 + 1G>A in the Polish population. No founder allele has been specifically identified in the French Canadian population. METHODS: The 14 coding exons of CHEK2 were fully sequenced for variant alleles in a panel of 25 affected French Canadian women and 25 healthy controls. Two variants were identified of which one novel variant was further screened for in an additional panel of 667 breast cancer patients and 6548 healthy controls. Additional genotyping was conducted using allele specific PCR and a restriction digest assay. Significance of amino acid substitutions were deduced by employing comparative analysis techniques. RESULTS: Two variants were identified: the previously reported silent substitution 252A>G (E84E) and the novel missense variant, 1217G>A (R406H). No significant difference in allele distribution between French Canadian women with breast cancer and healthy controls was observed (3/692, 0.43% vs. 22/6573, 0.33%, respectively, P = 0.73). CONCLUSION: The novel CHEK2 missense variant identified in this study, R406H, is unlikely to contribute to breast cancer risk in French Canadian women.

[Study on the mixed testing of serum samples in seroepidemiological survey of parasitic diseases].

OBJECTIVE: In order to accumulate experiences for improving the efficiency in serological tests, the present study on mixed testing of serum samples was performed by taking the serological test of trichinellosis and toxoplasmosis as the examples, and had proved the effects on cost-effectiveness of seroepidemiological survey of parasitic disease with method of mixed-samples test. METHODS: According to the binomial distribution principle, to develop an approach to the feasibility of mixed testing of serum samples, and to work on a cost-effectiveness analysis of one-by-one testing and mixed testing using hygienic economic analysis method was performed. For serological test of trichinellosis and toxoplasmosis, 3 kinds of mixed testing methods, namely 3 serum sample mixture, 5 serum sample mixture and 10 serum sample mixture, were performed. RESULTS: The results showed that all the 3 kinds of mixed tests of trichinellosis and toxoplasmosis showing positive result if only 1 weak positive serum sample were mixed with. When the serum samples being mixed were all negative ones, then among the 24 groups tested with each kind of negative serum sample mixture of trichinellosis (3 serum samples, 5 serum samples and 10 serum samples), they all showed negative. However, among the 12 groups tested with 2 kinds of negative serum mixture of toxoplasmosis (3 serum samples and 5 serum samples), all showed negative while among the 18 groups tested with the 10 serum sample mixture, 16 groups showed negative and 2 were positive. The mixed testing of trichinellosis and toxoplasmosis showed that the efficiency of mixed testing was related to the serological positive rate of the parasitic diseases to be examined. When serological positive rate was 10%, the efficiency of mixed testing was higher in 4 serum sample group. When serological positive rate was 1%, the efficiency of mixed testing was higher in 10 serum sample group and when serological positive rate was 0.1%, the in crease of the size of mixed serum samples could decrease the number of testing, but the prerequisite was that there must be one positive sample, so that the positivity for all the mixed tests could be detected. If mixed testing were performed on all negative samples, no positivity could be detected. CONCLUSION: The result of cost-effectiveness analysis demonstrated that for seroepidemiological survey of parasitic diseases, the cost for mixed testing was low, especially when the serological positive rate was expected low (< or = 1%, thus the mixed testing could save a large amount of the cost.