Assessing electronic health record phenotypes against gold-standard diagnostic criteria for diabetes mellitus

Spratt, Susan E.; Pereira, Katherine; Granger, Bradi B.; Batch, Bryan C.; Phelan, Matthew; Pencina, Michael J.; Miranda, Marie Lynn; Boulware, L. Ebony; Lucas, Joseph E.; Nelson, Charlotte L.; Neely, Benjamin; Goldstein, Benjamin A.; Barth, Pamela; Richesson, Rachel L.; Riley, Isaretta; Corsino, Leonor; Hinz, Eugenia R. McPeek; Rusincovitch, Shelley A.; Green, Jennifer; Barton, Anna Beth; Kelley, Carly E.; Hyland, Kristen A.; Tang, Mónica; Elliott, Amanda; Ruel, Ewa; Clark, Alexander M.; Mabrey, Melanie; Morrissey, Kay Lyn; Rao, Jyothi; Hong, Beatrice; Pierre-Louis, Marjorie; Kelly, Katherine Patterson; Jelesoff, Nicole

doi:10.1093/jamia/ocw123

Cited by 53 publications

(53 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This need is particularly relevant to the success of studies such as the PMI All of Us Research Program, 10 other national research initiatives, 27,28 the future of CVD epidemiology, 3–5 and the development of learning healthcare systems. 7 Although several studies have evaluated the different dimensions of EHR data quality and developed EHR phenotyping algorithms, 16,29–32 they frequently are from a single healthcare institution and use data from clinical care as a gold standard reference, including paper charts, patient and physician interviews, standardized patient encounters, registry data, and claims data. 23 …”

Section: Discussionmentioning

confidence: 99%

“…In recent years, national research consortiums, such as eMERGE (electronic medical records and genomics) Network, the NIH Collaboratory, and other research groups, have developed, tested, and validated EHR detection algorithms for various cardiovascular disease risk factors and events, including diabetes, HTN, heart failure, and coronary heart disease. 16,31,32 Many of these algorithms are developed at a single center, tested at other centers, and adapted if needed. 30 Consistent with prior studies, our results underscore the additive value of clinical data to ICD codes in case detection, especially for obesity, which has previously been reported as under-detected in administrative databases.…”

Section: Discussionmentioning

confidence: 99%

“…We defined obesity in HealthLNK by ICD-9 codes (278.0, 278.00, 278.01, 278.03) or a BMI ≥ 30 kg/m 2 , whereas in MESA, investigators defined obesity as BMI ≥ 30 kg/m 2 measured during in-person examination. In HealthLNK, we defined diabetes with ICD-9 codes (249.XX, 250.XX, 357.2, 362.0X), 14,16 hemoglobin A1C ≥ 6.5%, or diabetes medication use. In MESA, diabetes was defined as either use of diabetes medications or a fasting glucose ≥ 126 mg/dL.…”

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Validity of Cardiovascular Data From Electronic Sources

Ahmad¹,

Chan²,

Rosenman³

et al. 2017

Circulation

View full text Add to dashboard Cite

Background Understanding the validity of data from electronic data research networks is critical to national research initiatives and learning healthcare systems for cardiovascular care. Our goal was to evaluate the degree of agreement of electronic data research networks compared with data collected by standardized research approaches in a cohort study. Methods We linked individual-level data from The Multi-Ethnic Study of Atherosclerosis (MESA), a community-based cohort, with HealthLNK, a 2006–2012 database of electronic health records (EHRs) from six, Chicago health systems. To evaluate the correlation and agreement of blood pressure (BP) in HealthLNK as compared with in-person MESA examinations, and BMI in HealthLNK compared with MESA, we used Pearson Correlation Coefficients and Bland-Altman plots. Using diagnoses in MESA as the criterion standard, we calculated the performance of HealthLNK for hypertension (HTN), obesity, and diabetes diagnosis using ICD-9 codes and clinical data. We also identified potential myocardial infarctions (MIs), strokes, and heart failure events in HealthLNK and compared them with adjudicated events in MESA. Results Of the 1,164 MESA participants enrolled at the Chicago Field Center, 802 (68.9%) participants had data in HealthLNK. The correlation was low for systolic BP (0.39; P<0.0001). Compared with MESA, HealthLNK overestimated systolic BP by 6.5 mmHg (95%CI: 4.2, 7.8). There was a high correlation between BMI in MESA and HealthLNK (0.94; P<0.0001). HealthLNK underestimated BMI by 0.3 kg/m2 (95%CI: −0.4, −0.1). Using ICD-9 codes and clinical data, the sensitivity and specificity of HealthLNK queries for HTN were 82.4% and 59.4%, for obesity were 73.0% and 89.8%, and for diabetes were 79.8% and 93.3%. Compared with adjudicated CVD events in MESA, the concordance rates for MI, stroke, and heart failure were, respectively, 41.7% (5/12), 61.5% (8/13), and 62.5% (10/16). Conclusions These findings illustrate the limitations and strengths of electronic data repositories compared with information collected by traditional standardized epidemiologic approaches for the ascertainment of CVD risk factors and events.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Validity of Cardiovascular Data From Electronic Sources

Ahmad¹,

Chan²,

Rosenman³

et al. 2017

Circulation

View full text Add to dashboard Cite

show abstract

“…Large well‐designed observational studies can help identify the benefits and harms of specific treatments among patients with T2DM, particularly among those who may be underrepresented in clinical trials (ie, the elderly and those with certain comorbidities such as heart and renal disease) . The identification of patients with T2DM from EHR and administrative databases has been conducted extensively in the prior literature …”

Section: Introductionmentioning

confidence: 99%

“…Most previous studies have identified and studied patients with T2DM using EHR and administrative databases within well‐defined but single data systems. However, less is known about the implementation of those strategies across diverse systems and at a large scale . Therefore, T2DM represented an optimal condition for us to construct, characterize and validate a cohort of patients with T2DM across four diverse PCORnet sites to determine the utility of using these integrated data to conduct epidemiological and comparative effectiveness research.…”

Section: Introductionmentioning

confidence: 99%

Performance of a computable phenotype for identification of patients with diabetes within PCORnet: The Patient‐Centered Clinical Research Network

Wiese

Roumie

Buse

et al. 2019

Pharmacoepidemiology and Drug

View full text Add to dashboard Cite

Purpose PCORnet, the National Patient‐Centered Clinical Research Network, represents an innovative system for the conduct of observational and pragmatic studies. We describe the identification and validation of a retrospective cohort of patients with type 2 diabetes (T2DM) from four PCORnet sites. Methods We adapted existing computable phenotypes (CP) for the identification of patients with T2DM and evaluated their performance across four PCORnet sites (2012‐2016). Patients entered the cohort on the earliest date they met one of three CP categories: (CP1) coded T2DM diagnosis (ICD‐9/ICD‐10) and an antidiabetic prescription, (CP2) diagnosis and glycosylated hemoglobin (HbA1c) ≥6.5%, or (CP3) an antidiabetic prescription and HbA1c ≥6.5%. We required evidence of health care utilization in each of the 2 prior years for each patient, as we also developed an incident T2DM CP to identify the subset of patients without documentation of T2DM in the 365 days before t0. Among a systematic sample of patients, we calculated the positive predictive value (PPV) for the T2DM CP and incident‐T2DM CP using electronic health record (EHR) review as reference. Results The CP identified 50 657 patients with T2DM. The PPV of patients randomly selected for validation was 96.2% (n = 1572; CI:95.1‐97.0) and was consistently high across sites. The PPV for the incident‐T2DM CP was 5.8% (CI:4.5‐7.5). Conclusions The T2DM CP accurately and efficiently identified patients with T2DM across multiple sites that participate in PCORnet, although the incident T2DM CP requires further study. PCORnet is a valuable data source for future epidemiological and comparative effectiveness research among patients with T2DM.

show abstract

An exploratory study of plasma ceramides in comorbidities in Down syndrome

Worley

Byeon

Smith

et al. 2023

American J of Med Genetics Pt A

View full text Add to dashboard Cite

Plasma ceramide levels (henceforth, “ceramides”) are biomarkers of some diseases that are comorbidities of Down syndrome (DS). We sought to determine if comorbidities in DS were associated with ceramides, studying a convenience cohort of 35 study participants, all ≥12 months old. To identify comorbidities, we reviewed the problem lists in electronic health records that were concurrent with sample collection. We placed clinically related comorbidities into one of five categories of comorbidities, henceforth, categories: obesity/overweight; autoimmune disease; congenital heart disease; bacterial infection; and central nervous system (CNS) condition. We measured the eight ceramides most frequently associated with disease using liquid chromatography–tandem mass spectrometry. We calculated a ceramide composite outcome score (CCOS) for each participant by normalizing each ceramide level to the mean for that level in the study population and then summing the normalized levels, to be proxy variable for all eight ceramides in aggregate. We used multivariable linear regression models adjusted for age and sex to test associations of categories with ceramides and with CCOSs. Post hoc, we realized that co‐occurring comorbidities might interfere with establishing associations between predictor categories and ceramides and that stratified analyses might eliminate their influence on associations. We posited that CCOSs could be used to screen for associations of categories with multiple ceramides, since most diseases have been associated with more than one ceramide. We chose to omit in the stratified analyses the two categories that were the most different from one another in their associations with their CCOSs, having the most divergent regression coefficients (the highest positive and lowest negative coefficients). We first omitted one of these two divergent categories in a stratified analysis and tested in the remaining participants (those without a comorbidity in the interfering category) for associations of the other four categories with their CCOSs and then did the same for the other divergent category. In each of these two screening stratified analyses, we found one category was significantly associated with its CCOS. In the two identified categories, we then tested for associations with each of the eight ceramides, using the appropriate stratified analysis. Next, we sought to determine if the associations of the two categories with ceramides we found by omitting participants in the interfering categories held in our small sample for participants in the omitted categories as well. For each of the two categories, we therefore omitted participants without the interfering category and determined associations between the predictor category and individual ceramides in the remaining participants (those with a comorbidity in the interfering category). In the a priori analyses, autoimmune disease was inversely associated with C16 and CNS condition was inversely associated with C23. Obesity/overweight and CNS condition were the two categories with the most divergent regression coefficients (0.037 vs. −0.048). In post hoc stratified analyses, after omitting participants with obesity/overweight, thereby leaving participants without obesity/overweight, bacterial infection was associated with its CCOS and then with C14, C20, and C22. However, in the companion stratified analyses, omitting participants without obesity/overweight, thereby leaving participants with obesity/overweight, bacterial infection was not associated with any of the eight ceramides. Similarly, in post hoc stratified analyses after omitting participants with a CNS condition, thereby leaving participants without a CNS condition, obesity/overweight was associated with its CCOS and then with C14, C23, and C24. In the companion analyses, omitting participants without a CNS condition, thereby leaving participants with a CNS condition, obesity/overweight was inversely associated with C24.1. In conclusion, CNS and autoimmune disease were inversely associated with one ceramide each in a priori analyses. In post hoc analyses, we serendipitously omitted categories that interfered with associations of other categories with ceramides in stratified analyses. We found that bacterial infection was associated with three ceramides in participants without obesity/overweight and that obesity/overweight was associated with three ceramides in participants without a CNS condition. We therefore identified obesity/overweight and CNS conditions as potential confounders or effect modifiers for these associations. This is the first report of ceramides in DS and in human bacterial infection. Further study of ceramides in comorbidities of DS is justified.

show abstract

Assessing electronic health record phenotypes against gold-standard diagnostic criteria for diabetes mellitus

Cited by 53 publications

References 20 publications

Validity of Cardiovascular Data From Electronic Sources

Validity of Cardiovascular Data From Electronic Sources

Performance of a computable phenotype for identification of patients with diabetes within PCORnet: The Patient‐Centered Clinical Research Network

An exploratory study of plasma ceramides in comorbidities in Down syndrome

Contact Info

Product

Resources

About