Extracting social determinants of health from electronic health records using natural language processing: a systematic review

ObjectiveSocial determinants of health (SDoH) are nonclinical dispositions that impact patient health risks and clinical outcomes. Leveraging SDoH in clinical decision-making can potentially improve diagnosis, treatment planning, and patient outcomes. Despite increased interest in capturing SDoH in electronic health records (EHRs), such information is typically locked in unstructured clinical notes. Natural language processing (NLP) is the key technology to extract SDoH information from clinical text and expand its utility in patient care and research. This article presents a systematic review of the state-of-the-art NLP approaches and tools that focus on identifying and extracting SDoH data from unstructured clinical text in EHRs.Materials and MethodsA broad literature search was conducted in February 2021 using 3 scholarly databases (ACL Anthology, PubMed, and Scopus) following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A total of 6402 publications were initially identified, and after applying the study inclusion criteria, 82 publications were selected for the final review.ResultsSmoking status (n = 27), substance use (n = 21), homelessness (n = 20), and alcohol use (n = 15) are the most frequently studied SDoH categories. Homelessness (n = 7) and other less-studied SDoH (eg, education, financial problems, social isolation and support, family problems) are mostly identified using rule-based approaches. In contrast, machine learning approaches are popular for identifying smoking status (n = 13), substance use (n = 9), and alcohol use (n = 9).ConclusionNLP offers significant potential to extract SDoH data from narrative clinical notes, which in turn can aid in the development of screening tools, risk prediction models, and clinical decision support systems.     

Demonstrating an approach for evaluating synthetic geospatial and temporal epidemiologic data utility: results from analyzing >1.8 million SARS-CoV-2 tests in the United States National COVID Cohort Collaborative (N3C)

ObjectiveThis study sought to evaluate whether synthetic data derived from a national coronavirus disease 2019 (COVID-19) dataset could be used for geospatial and temporal epidemic analyses.Materials and MethodsUsing an original dataset (n = 1 854 968 severe acute respiratory syndrome coronavirus 2 tests) and its synthetic derivative, we compared key indicators of COVID-19 community spread through analysis of aggregate and zip code-level epidemic curves, patient characteristics and outcomes, distribution of tests by zip code, and indicator counts stratified by month and zip code. Similarity between the data was statistically and qualitatively evaluated.ResultsIn general, synthetic data closely matched original data for epidemic curves, patient characteristics, and outcomes. Synthetic data suppressed labels of zip codes with few total tests (mean = 2.9 ± 2.4; max = 16 tests; 66% reduction of unique zip codes). Epidemic curves and monthly indicator counts were similar between synthetic and original data in a random sample of the most tested (top 1%; n = 171) and for all unsuppressed zip codes (n = 5819), respectively. In small sample sizes, synthetic data utility was notably decreased.DiscussionAnalyses on the population-level and of densely tested zip codes (which contained most of the data) were similar between original and synthetically derived datasets. Analyses of sparsely tested populations were less similar and had more data suppression.ConclusionIn general, synthetic data were successfully used to analyze geospatial and temporal trends. Analyses using small sample sizes or populations were limited, in part due to purposeful data label suppression—an attribute disclosure countermeasure. Users should consider data fitness for use in these cases.     

Bias and fairness assessment of a natural language processing opioid misuse classifier: detection and mitigation of electronic health record data disadvantages across racial subgroups

ObjectivesTo assess fairness and bias of a previously validated machine learning opioid misuse classifier.Materials & MethodsTwo experiments were conducted with the classifier’s original (n = 1000) and external validation (n = 53 974) datasets from 2 health systems. Bias was assessed via testing for differences in type II error rates across racial/ethnic subgroups (Black, Hispanic/Latinx, White, Other) using bootstrapped 95% confidence intervals. A local surrogate model was estimated to interpret the classifier’s predictions by race and averaged globally from the datasets. Subgroup analyses and post-hoc recalibrations were conducted to attempt to mitigate biased metrics.ResultsWe identified bias in the false negative rate (FNR = 0.32) of the Black subgroup compared to the FNR (0.17) of the White subgroup. Top features included “heroin” and “substance abuse” across subgroups. Post-hoc recalibrations eliminated bias in FNR with minimal changes in other subgroup error metrics. The Black FNR subgroup had higher risk scores for readmission and mortality than the White FNR subgroup, and a higher mortality risk score than the Black true positive subgroup (P < .05).DiscussionThe Black FNR subgroup had the greatest severity of disease and risk for poor outcomes. Similar features were present between subgroups for predicting opioid misuse, but inequities were present. Post-hoc mitigation techniques mitigated bias in type II error rate without creating substantial type I error rates. From model design through deployment, bias and data disadvantages should be systematically addressed.ConclusionStandardized, transparent bias assessments are needed to improve trustworthiness in clinical machine learning models.     

A theory-based meta-regression of factors influencing clinical decision support adoption and implementation

ObjectiveThe purpose of the study was to explore the theoretical underpinnings of effective clinical decision support (CDS) factors using the comparative effectiveness results.Materials and MethodsWe leveraged search results from a previous systematic literature review and updated the search to screen articles published from January 2017 to January 2020. We included randomized controlled trials and cluster randomized controlled trials that compared a CDS intervention with and without specific factors. We used random effects meta-regression procedures to analyze clinician behavior for the aggregate effects. The theoretical model was the Unified Theory of Acceptance and Use of Technology (UTAUT) model with motivational control.ResultsThirty-four studies were included. The meta-regression models identified the importance of effort expectancy (estimated coefficient = −0.162; P = .0003); facilitating conditions (estimated coefficient = 0.094; P = .013); and performance expectancy with motivational control (estimated coefficient = 1.029; P = .022). Each of these factors created a significant impact on clinician behavior. The meta-regression model with the multivariate analysis explained a large amount of the heterogeneity across studies (R2 =88.32%).DiscussionThree positive factors were identified: low effort to use, low controllability, and providing more infrastructure and implementation strategies to support the CDS. The multivariate analysis suggests that passive CDS could be effective if users believe the CDS is useful and/or social expectations to use the CDS intervention exist.ConclusionsOverall, a modified UTAUT model that includes motivational control is an appropriate model to understand psychological factors associated with CDS effectiveness and to guide CDS design, implementation, and optimization.     

Can decision support combat incompleteness and bias in routine primary care data?

ObjectiveRoutine primary care data may be used for the derivation of clinical prediction rules and risk scores. We sought to measure the impact of a decision support system (DSS) on data completeness and freedom from bias.Materials and MethodsWe used the clinical documentation of 34 UK general practitioners who took part in a previous study evaluating the DSS. They consulted with 12 standardized patients. In addition to suggesting diagnoses, the DSS facilitates data coding. We compared the documentation from consultations with the electronic health record (EHR) (baseline consultations) vs consultations with the EHR-integrated DSS (supported consultations). We measured the proportion of EHR data items related to the physician’s final diagnosis. We expected that in baseline consultations, physicians would document only or predominantly observations related to their diagnosis, while in supported consultations, they would also document other observations as a result of exploring more diagnoses and/or ease of coding.ResultsSupported documentation contained significantly more codes (incidence rate ratio [IRR] = 5.76 [4.31, 7.70] P < .001) and less free text (IRR = 0.32 [0.27, 0.40] P < .001) than baseline documentation. As expected, the proportion of diagnosis-related data was significantly lower (b = −0.08 [−0.11, −0.05] P < .001) in the supported consultations, and this was the case for both codes and free text.ConclusionsWe provide evidence that data entry in the EHR is incomplete and reflects physicians’ cognitive biases. This has serious implications for epidemiological research that uses routine data. A DSS that facilitates and motivates data entry during the consultation can improve routine documentation.     

Is rs759853 polymorphism in promoter of aldose reductase gene a risk factor for diabetic nephropathy? A meta-analysis

Background

So far, a number of case-control or cohort studies have been carried out to investigate the relationship between rs759853 polymorphism in the promoter of aldose reductase (AR) gene and the risk of diabetic nephropathy (DN). However, the results have generated considerable controversy. We performed this study to clarify the linkage between this gene mutation and the risk of DN.

Methods

A comprehensive literature search of electronic databases and a well-organized meta-analysis were conducted.

Results

Twelve comparisons and 4,735 individuals from nine published case-control or cohort studies were included finally. From none to large heterogeneity was observed, therefore, both fixed and random models were used. Significant differences were found between AR rs759853 polymorphism and susceptibility of DN from both type 1 and type 2 diabetes in all genetic models (allele contrast, OR = 1.37, CI (1.18, 1.59), P < 0.0001; additive model, OR = 1.78, CI (1.25, 2.53), P = 0.01; recessive model OR = 1.33 CI (1.08, 1.63), P = 0.008; dominant model, OR = 1.52, CI (1.26, 1.84), P < 0.0001; codominance model OR = 1.30 (1.15, 1.47), P < 0.0001). In stratified meta-analyses for type 2 diabetes by ethnicity, the significant relationship was found in allele contrast and dominant model in Caucasians, and in allele contrast and codominance model in Asians. However, data do not support the linkage between this gene mutation and the progression of DN. There was no significant publication bias.

Conclusions

The evidence currently available shows that the AR rs759853 polymorphism may correlate with the susceptibility of DN. However, data do not support the association between this DNA variation and the progression of DN.     

Effects of technology-supported exercise programs on the knee pain,physical function,and quality of life of individuals with knee osteoarthritis and/or chronic knee pain: A systematic review and meta-analysis of randomized controlled trials

ObjectiveThe study sought to examine the effects of technology-supported exercise programs on the knee pain, physical function, and quality of life of individuals with knee osteoarthritis and/or chronic knee pain by a systematic review and meta-analysis of randomized controlled trials.Materials and MethodsWe searched MEDLINE, EMBASE, CINAHL Plus, and the Cochrane Library from database inception to August 2020. A meta-analysis and subgroup analyses, stratified by technology type and program feature, were conducted.ResultsTwelve randomized controlled trials were reviewed, all of which implemented the programs for 4 weeks to 6 months. Telephone, Web, mobile app, computer, and virtual reality were used to deliver the programs. The meta-analysis showed that these programs were associated with significant improvements in knee pain (standardized mean difference [SMD] = −0.29; 95% confidence interval [CI], −0.48 to −0.10; P = .003) and quality of life (SMD = 0.25; 95% CI, 0.04 to 0.46; P = .02) but not with significant improvement in physical function (SMD = 0.22; 95% CI, 0 to 0.43; P = .053). Subgroup analyses showed that some technology types and program features were suggestive of potential benefits.ConclusionsUsing technology to deliver the exercise programs appears to offer benefits. The technology types and program features that were associated with health values have been identified, based on which suggestions are discussed for the further research and development of such programs.     

Real-world insights from launching remote peer-to-peer mentoring in a safety net healthcare delivery setting

Peer mentors have been proven to improve diabetes outcomes, especially among diverse patients. Delivering peer mentoring via remote strategies (phone, text, mobile applications) is critical, especially in light of the recent pandemic. We conducted a real-world evaluation of a remote diabetes intervention in a safety-net delivery system in New York. We summarized the uptake, content, and pre-post clinical effectiveness for English- and Spanish-speaking participants. Of patients who could be reached, 71% (n = 690/974) were enrolled, and 90% of those (n = 618/690) participated in coaching. Patients and mentors had a mean of 32 check-ins, and each patient set an average of 10 goals. 29% of the participants accessed the program via the smartphone application. Among participants with complete hemoglobin A1c data (n = 179), there was an absolute 1.71% reduction (P < .01). There are multiple lessons for successful implementation of remote peer coaching into settings serving diverse patients, including meaningful patient-mentor matching and addressing social determinants.     

Characterizing physician EHR use with vendor derived data: a feasibility study and cross-sectional analysis

ObjectiveTo derive 7 proposed core electronic health record (EHR) use metrics across 2 healthcare systems with different EHR vendor product installations and examine factors associated with EHR time.Materials and MethodsA cross-sectional analysis of ambulatory physicians EHR use across the Yale-New Haven and MedStar Health systems was performed for August 2019 using 7 proposed core EHR use metrics normalized to 8 hours of patient scheduled time.ResultsFive out of 7 proposed metrics could be measured in a population of nonteaching, exclusively ambulatory physicians. Among 573 physicians (Yale-New Haven N = 290, MedStar N = 283) in the analysis, median EHR-Time₈ was 5.23 hours. Gender, additional clinical hours scheduled, and certain medical specialties were associated with EHR-Time₈ after adjusting for age and health system on multivariable analysis. For every 8 hours of scheduled patient time, the model predicted these differences in EHR time (P < .001, unless otherwise indicated): female physicians +0.58 hours; each additional clinical hour scheduled per month −0.01 hours; practicing cardiology −1.30 hours; medical subspecialties −0.89 hours (except gastroenterology, P = .002); neurology/psychiatry −2.60 hours; obstetrics/gynecology −1.88 hours; pediatrics −1.05 hours (P = .001); sports/physical medicine and rehabilitation −3.25 hours; and surgical specialties −3.65 hours.ConclusionsFor every 8 hours of scheduled patient time, ambulatory physicians spend more than 5 hours on the EHR. Physician gender, specialty, and number of clinical hours practicing are associated with differences in EHR time. While audit logs remain a powerful tool for understanding physician EHR use, additional transparency, granularity, and standardization of vendor-derived EHR use data definitions are still necessary to standardize EHR use measurement.     

Risk factors associated with medication ordering errors

ObjectiveWe utilized a computerized order entry system–integrated function referred to as “void” to identify erroneous orders (ie, a “void” order). Using voided orders, we aimed to (1) identify the nature and characteristics of medication ordering errors, (2) investigate the risk factors associated with medication ordering errors, and (3) explore potential strategies to mitigate these risk factors.Materials and MethodsWe collected data on voided orders using clinician interviews and surveys within 24 hours of the voided order and using chart reviews. Interviews were informed by the human factors–based SEIPS (Systems Engineering Initiative for Patient Safety) model to characterize the work systems–based risk factors contributing to ordering errors; chart reviews were used to establish whether a voided order was a true medication ordering error and ascertain its impact on patient safety.ResultsDuring the 16-month study period (August 25, 2017, to December 31, 2018), 1074 medication orders were voided; 842 voided orders were true medication errors (positive predictive value = 78.3 ± 1.2%). A total of 22% (n = 190) of the medication ordering errors reached the patient, with at least a single administration, without causing patient harm. Interviews were conducted on 355 voided orders (33% response). Errors were not uniquely associated with a single risk factor, but the causal contributors of medication ordering errors were multifactorial, arising from a combination of technological-, cognitive-, environmental-, social-, and organizational-level factors.ConclusionsThe void function offers a practical, standardized method to create a rich database of medication ordering errors. We highlight implications for utilizing the void function for future research, practice and learning opportunities.     

Impact of a problem-oriented view on clinical data retrieval

ObjectiveThe electronic health record (EHR) data deluge makes data retrieval more difficult, escalating cognitive load and exacerbating clinician burnout. New auto-summarization techniques are needed. The study goal was to determine if problem-oriented view (POV) auto-summaries improve data retrieval workflows. We hypothesized that POV users would perform tasks faster, make fewer errors, be more satisfied with EHR use, and experience less cognitive load as compared with users of the standard view (SV).MethodsSimple data retrieval tasks were performed in an EHR simulation environment. A randomized block design was used. In the control group (SV), subjects retrieved lab results and medications by navigating to corresponding sections of the electronic record. In the intervention group (POV), subjects clicked on the name of the problem and immediately saw lab results and medications relevant to that problem.ResultsWith POV, mean completion time was faster (173 seconds for POV vs 205 seconds for SV; P < .0001), the error rate was lower (3.4% for POV vs 7.7% for SV; P = .0010), user satisfaction was greater (System Usability Scale score 58.5 for POV vs 41.3 for SV; P < .0001), and cognitive task load was less (NASA Task Load Index score 0.72 for POV vs 0.99 for SV; P < .0001).DiscussionThe study demonstrates that using a problem-based auto-summary has a positive impact on 4 aspects of EHR data retrieval, including cognitive load.ConclusionEHRs have brought on a data deluge, with increased cognitive load and physician burnout. To mitigate these increases, further development and implementation of auto-summarization functionality and the requisite knowledge base are needed.     

Evaluation of federated learning variations for COVID-19 diagnosis using chest radiographs from 42 US and European hospitals

ObjectiveFederated learning (FL) allows multiple distributed data holders to collaboratively learn a shared model without data sharing. However, individual health system data are heterogeneous. “Personalized” FL variations have been developed to counter data heterogeneity, but few have been evaluated using real-world healthcare data. The purpose of this study is to investigate the performance of a single-site versus a 3-client federated model using a previously described Coronavirus Disease 19 (COVID-19) diagnostic model. Additionally, to investigate the effect of system heterogeneity, we evaluate the performance of 4 FL variations.Materials and methodsWe leverage a FL healthcare collaborative including data from 5 international healthcare systems (US and Europe) encompassing 42 hospitals. We implemented a COVID-19 computer vision diagnosis system using the Federated Averaging (FedAvg) algorithm implemented on Clara Train SDK 4.0. To study the effect of data heterogeneity, training data was pooled from 3 systems locally and federation was simulated. We compared a centralized/pooled model, versus FedAvg, and 3 personalized FL variations (FedProx, FedBN, and FedAMP).ResultsWe observed comparable model performance with respect to internal validation (local model: AUROC 0.94 vs FedAvg: 0.95, P = .5) and improved model generalizability with the FedAvg model (P < .05). When investigating the effects of model heterogeneity, we observed poor performance with FedAvg on internal validation as compared to personalized FL algorithms. FedAvg did have improved generalizability compared to personalized FL algorithms. On average, FedBN had the best rank performance on internal and external validation.ConclusionFedAvg can significantly improve the generalization of the model compared to other personalization FL algorithms; however, at the cost of poor internal validity. Personalized FL may offer an opportunity to develop both internal and externally validated algorithms.     

Association of a variant in the regulatory region of NADPH oxidase 4 gene and metabolic syndrome in patients with chronic hepatitis C

Background

Given the important contribution of the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase system to the generation of reactive oxygen species induced by hepatitis C virus (HCV), we investigated two single nucleotide polymorphisms (SNPs) in the putative regulatory region of the genes encoding NADPH oxidase 4 catalytic subunit (NOX4) and its regulatory subunit p22phox (CYBA) and their relation with metabolic and histological variables in patients with HCV.

Methods

One hundred seventy eight naïve HCV patients (49.3% male; 65% HCV genotype 1) with positive HCV RNA were genotyped using specific primers and fluorescent-labeled probes for SNPs rs3017887 in NOX4 and −675 T → A in CYBA.

Results

No association was found between the genotype frequencies of NOX4 and CYBA SNPs and inflammation scores or fibrosis stages in the overall population. The presence of the CA + AA genotypes of the NOX4 SNP was nominally associated with a lower alanine aminotransferase (ALT) concentration in the male population (CA + AA = 72.23 ± 6.34 U/L versus CC = 100.22 ± 9.85; mean ± SEM; P = 0.05). The TT genotype of the CYBA SNP was also nominally associated with a lower ALT concentration in the male population (TT = 84.01 ± 6.77 U/L versus TA + AA = 109.67 ± 18.37 U/L; mean ± SEM; P = 0.047). The minor A-allele of the NOX4 SNP was inversely associated with the frequency of metabolic syndrome (MS) in the male population (odds ratio (OR): 0.15; 95% confidence interval (CI): 0.03 to 0.79; P = 0.025).

Conclusions

The results suggest that the evaluated NOX4 and CYBA SNPs are not direct genetic determinants of fibrosis in HCV patients, but nevertheless NOX4 rs3017887 SNP could indirectly influence fibrosis susceptibility due to its inverse association with MS in male patients.     

Development of a standardized laparoscopic caecum resection model to simulate laparoscopic appendectomy in rats

Background

Laparoscopic appendectomy (LA) has become one of the most common surgical procedures to date. To improve and standardize this technique further, cost-effective and reliable animal models are needed.

Methods

In a pilot study, 30 Wistar rats underwent laparoscopic caecum resection (as rats do not have an appendix vermiformis), to optimize the instrumental and surgical parameters. A subsequent test study was performed in another 30 rats to compare three different techniques for caecum resection and bowel closure.

Results

Bipolar coagulation led to an insufficiency of caecal stump closure in all operated rats (Group 1, n = 10). Endoloop ligation followed by bipolar coagulation and resection (Group 2, n = 10) or resection with a LigaSure™ device (Group 3, n = 10) resulted in sufficient caecal stump closure.

Conclusions

We developed a LA model enabling us to compare three different caecum resection techniques in rats. In conclusion, only endoloop closure followed by bipolar coagulation proved to be a secure and cost-effective surgical approach.     

Association of CYP1A1 MspI polymorphism in the esophageal cancer risk: a meta-analysis in the Chinese population

Background

Although many epidemiologic studies have investigated the CYP1A1 MspI gene polymorphisms and their associations with esophageal cancer (EC), definite conclusions cannot be drawn. To clarify the effects of CYP1A1 MspI polymorphisms on the risk of EC, a meta-analysis was performed in Chinese population.

Methods

Related studies were identified from PubMed, Springer Link, Ovid, Chinese Wanfang Data Knowledge Service Platform, Chinese National Knowledge Infrastructure (CNKI), and Chinese Biology Medicine (CBM) till October 2014. Pooled ORs and 95% CIs were used to assess the strength of the associations.

Results

A total of 13 studies including 1,519 EC cases and 1,962 controls were involved in this meta-analysis. Overall, significant association was found between CYP1A1 MspI polymorphism and EC risk when all studies in the Chinese population pooled into this meta-analysis (C vs. T: OR = 1.25, 95% CI = 1.04 to 1.51; CC + CT vs. TT: OR = 1.35, 95% CI = 1.06 to 1.72; CC vs. TT + CT: OR = 1.35, 95% CI = 1.03 to 1.76). When we performed stratified analyses by geographical locations, histopathology type, and source of control, significantly increased risks were found in North China (C vs. T: OR = 1.38, 95% CI = 1.12 to 1.70; CC vs. TT: OR = 1.72, 95% CI = 1.16 to 2.56; CC + CT vs. TT: OR = 1.52, 95% CI = 1.14 to 2.02; CC vs. TT + CT: OR = 1.55, 95% CI = 1.17 to 2.06), in the population-based studies (C vs. T: OR = 1.22, 95% CI = 1.05 to 1.42; CC vs. TT: OR = 1.38, 95% CI = 1.02 to 1.88; CC + CT vs. TT: OR = 1.36, 95% CI = 1.10 to 1.69; CC vs. TT + CT: OR = 1.43, 95% CI = 1.13 to 1.81) and ESCC (C vs. T: OR = 1.17, 95% CI = 1.04 to 1.32; CC + CT vs. TT: OR = 1.28, 95% CI = 1.08 to 1.52).

Conclusions

This meta-analysis provides the evidence that CYP1A1 MspI polymorphism may contribute to the EC development in the Chinese population.     

Impact of event notification services on timely follow-up and rehospitalization among primary care patients at two Veterans Affairs Medical Centers

ObjectiveTo examine the effectiveness of event notification service (ENS) alerts on health care delivery processes and outcomes for older adults.Materials and methodsWe deployed ENS alerts in 2 Veterans Affairs (VA) medical centers using regional health information exchange (HIE) networks from March 2016 to December 2019. Alerts targeted VA-based primary care teams when older patients (aged 65+ years) were hospitalized or attended emergency departments (ED) outside the VA system. We employed a concurrent cohort study to compare postdischarge outcomes between patients whose providers received ENS alerts and those that did not (usual care). Outcome measures included: timely follow-up postdischarge (actual phone call within 7 days or an in-person primary care visit within 30 days) and all-cause inpatient or ED readmission within 30 days. Generalized linear mixed models, accounting for clustering by primary care team, were used to compare outcomes between groups.ResultsCompared to usual care, veterans whose primary care team received notification of non-VA acute care encounters were 4 times more likely to have phone contact within 7 days (AOR = 4.10, P < .001) and 2 times more likely to have an in-person visit within 30 days (AOR = 1.98, P = .007). There were no significant differences between groups in hospital or ED utilization within 30 days of index discharge (P = .057).DiscussionENS was associated with increased timely follow-up following non-VA acute care events, but there was no associated change in 30-day readmission rates. Optimization of ENS processes may be required to scale use and impact across health systems.ConclusionGiven the importance of ENS to the VA and other health systems, this study provides guidance for future research on ENS for improving care coordination and population outcomes.Trial RegistrationClinicalTrials.gov {"type":"clinical-trial","attrs":{"text":"NCT02689076","term_id":"NCT02689076"}}NCT02689076. “Regional Data Exchange to Improve Care for Veterans After Non-VA Hospitalization.” Registered February 23, 2016.