Specificity of International Classification of Diseases codes for bronchopulmonary dysplasia: an investigation using electronic health record data and a large insurance database.
Ontology highlight
ABSTRACT: Objective
International Classification of Diseases (ICD) codes in electronic health records (EHRs) are increasingly used for health services research, in spite of unknown diagnostic accuracy. The accuracy of ICD codes to identify bronchopulmonary dysplasia (BPD) is unknown.Study design
Retrospective cohort study in a single-center NICU (n = 166) to evaluate sensitivity and specificity of ICD-10 codes for the diagnosis of BPD. Analysis of large insurance claims database (n = 7887) to determine date of assignment of the code.Results
The sensitivity of any BPD-related ICD codes ranged from 0.82 to 0.95, while the specificity ranged from 0.25 to 0.36. In a large national insurance database, the most common date of ICD-9 or ICD-10 code assignment was the day of birth, which is inconsistent with the clinical definition.Conclusions
ICD codes registered for BPD are unlikely to accurately reflect the current clinical definition and should be interpreted with caution.
SUBMITTER: Beam KS
PROVIDER: S-EPMC7917960 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Obesity is a potentially modifiable risk factor for many diseases, and a better understanding of its impact on health care utilization, costs, and medical outcomes is needed. The ability to accurately evaluate obesity outcomes depends on a correct identification of the population with obesity. The primary objective of this study was to determine the prevalence and accuracy of International Classification of Diseases, Ninth Revision (ICD-9) coding for overweight and obesity within a US primary care electronic health record (EHR) database compared against actual body mass index (BMI) values from recorded clinical patient data; characteristics of patients with obesity who did or did not receive ICD-9 codes for overweight/obesity also were evaluated. The study sample included 5,512,285 patients in the database with any BMI value recorded between January 1, 2014, and June 30, 2014. Based on BMI, 74.6% of patients were categorized as being overweight or obese, but only 15.1% of patients had relevant ICD-9 codes. ICD-9 coding prevalence increased with increasing BMI category. Among patients with obesity (BMI ?30?kg/m2), those coded for obesity were younger, more often female, and had a greater comorbidity burden than those not coded; hypertension, dyslipidemia, type 2 diabetes mellitus, and gastroesophageal reflux disease were the most common comorbidities. KEY FINDINGS:US outpatients with overweight or obesity are not being reliably coded, making ICD-9 codes undependable sources for determining obesity prevalence and outcomes. BMI data available within EHR databases offer a more accurate and objective means of classifying overweight/obese status.
| S-EPMC5984561 | biostudies-literature
Project description:BACKGROUND:Accurate anesthesiology procedure code data are essential to quality improvement, research, and reimbursement tasks within anesthesiology practices. Advanced data science techniques, including machine learning and natural language processing, offer opportunities to develop classification tools for Current Procedural Terminology codes across anesthesia procedures. METHODS:Models were created using a Train/Test dataset including 1,164,343 procedures from 16 academic and private hospitals. Five supervised machine learning models were created to classify anesthesiology Current Procedural Terminology codes, with accuracy defined as first choice classification matching the institutional-assigned code existing in the perioperative database. The two best performing models were further refined and tested on a Holdout dataset from a single institution distinct from Train/Test. A tunable confidence parameter was created to identify cases for which models were highly accurate, with the goal of at least 95% accuracy, above the reported 2018 Centers for Medicare and Medicaid Services (Baltimore, Maryland) fee-for-service accuracy. Actual submitted claim data from billing specialists were used as a reference standard. RESULTS:Support vector machine and neural network label-embedding attentive models were the best performing models, respectively, demonstrating overall accuracies of 87.9% and 84.2% (single best code), and 96.8% and 94.0% (within top three). Classification accuracy was 96.4% in 47.0% of cases using support vector machine and 94.4% in 62.2% of cases using label-embedding attentive model within the Train/Test dataset. In the Holdout dataset, respective classification accuracies were 93.1% in 58.0% of cases and 95.0% among 62.0%. The most important feature in model training was procedure text. CONCLUSIONS:Through application of machine learning and natural language processing techniques, highly accurate real-time models were created for anesthesiology Current Procedural Terminology code classification. The increased processing speed and a priori targeted accuracy of this classification approach may provide performance optimization and cost reduction for quality improvement, research, and reimbursement tasks reliant on anesthesiology procedure codes.
| S-EPMC7665375 | biostudies-literature
Project description:Objective:Standards such as the Logical Observation Identifiers Names and Codes (LOINC®) are critical for interoperability and integrating data into common data models, but are inconsistently used. Without consistent mapping to standards, clinical data cannot be harmonized, shared, or interpreted in a meaningful context. We sought to develop an automated machine learning pipeline that leverages noisy labels to map laboratory data to LOINC codes. Materials and Methods:Across 130 sites in the Department of Veterans Affairs Corporate Data Warehouse, we selected the 150 most commonly used laboratory tests with numeric results per site from 2000 through 2016. Using source data text and numeric fields, we developed a machine learning model and manually validated random samples from both labeled and unlabeled datasets. Results:The raw laboratory data consisted of >6.5 billion test results, with 2215 distinct LOINC codes. The model predicted the correct LOINC code in 85% of the unlabeled data and 96% of the labeled data by test frequency. In the subset of labeled data where the original and model-predicted LOINC codes disagreed, the model-predicted LOINC code was correct in 83% of the data by test frequency. Conclusion:Using a completely automated process, we are able to assign LOINC codes to unlabeled data with high accuracy. When the model-predicted LOINC code differed from the original LOINC code, the model prediction was correct in the vast majority of cases. This scalable, automated algorithm may improve data quality and interoperability, while substantially reducing the manual effort currently needed to accurately map laboratory data.
| S-EPMC7646911 | biostudies-literature
Project description: Not available
| S-EPMC6857496 | biostudies-literature
Project description:To compare three groupings of Electronic Health Record (EHR) billing codes for their ability to represent clinically meaningful phenotypes and to replicate known genetic associations. The three tested coding systems were the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, the Agency for Healthcare Research and Quality Clinical Classification Software for ICD-9-CM (CCS), and manually curated "phecodes" designed to facilitate phenome-wide association studies (PheWAS) in EHRs.We selected 100 disease phenotypes and compared the ability of each coding system to accurately represent them without performing additional groupings. The 100 phenotypes included 25 randomly-chosen clinical phenotypes pursued in prior genome-wide association studies (GWAS) and another 75 common disease phenotypes mentioned across free-text problem lists from 189,289 individuals. We then evaluated the performance of each coding system to replicate known associations for 440 SNP-phenotype pairs.Out of the 100 tested clinical phenotypes, phecodes exactly matched 83, compared to 53 for ICD-9-CM and 32 for CCS. ICD-9-CM codes were typically too detailed (requiring custom groupings) while CCS codes were often not granular enough. Among 440 tested known SNP-phenotype associations, use of phecodes replicated 153 SNP-phenotype pairs compared to 143 for ICD-9-CM and 139 for CCS. Phecodes also generally produced stronger odds ratios and lower p-values for known associations than ICD-9-CM and CCS. Finally, evaluation of several SNPs via PheWAS identified novel potential signals, some seen in only using the phecode approach. Among them, rs7318369 in PEPD was associated with gastrointestinal hemorrhage.Our results suggest that the phecode groupings better align with clinical diseases mentioned in clinical practice or for genomic studies. ICD-9-CM, CCS, and phecode groupings all worked for PheWAS-type studies, though the phecode groupings produced superior results.
| S-EPMC5501393 | biostudies-literature
Project description:BackgroundThe effectiveness of weight loss therapies is commonly measured using body mass index and other obesity-related variables. Although these data are often stored in electronic health records (EHRs) and potentially very accessible, few studies on obesity and weight loss have used data derived from EHRs. We developed processes for obtaining data from the EHR in order to construct a database on patients undergoing Roux-en-Y gastric bypass (RYGB) surgery.MethodsClinical data obtained as part of standard of care in a bariatric surgery program at an integrated health delivery system were extracted from the EHR and deposited into a data warehouse. Data files were extracted, cleaned, and stored in research datasets. To illustrate the utility of the data, Kaplan-Meier analysis was used to estimate length of post-operative follow-up.ResultsDemographic, laboratory, medication, co-morbidity, and survey data were obtained from 2028 patients who had undergone RYGB at the same institution since 2004. Pre-and post-operative diagnostic and prescribing information were available on all patients, while survey laboratory data were available on a majority of patients. The number of patients with post-operative laboratory test results varied by test. Based on Kaplan-Meier estimates, over 74% of patients had post-operative weight data available at 4 years.ConclusionA variety of EHR-derived data related to obesity can be efficiently obtained and used to study important outcomes following RYGB.
| S-EPMC3508953 | biostudies-literature
Project description:ObjectiveTo quantify and compare levels of potential biomarkers in neonates with (i) Bronchopulmonary dysplasia (BPD); (ii) BPD-associated pulmonary hypertension (BPD-PH); (iii) PH without BPD; and (iv) neonates without lung disease at ~36 weeks postmenstrual age.Study designMultiple potential biomarkers were measured in plasma samples of 90 patients using a multi-spot enzyme-linked immunosorbent assay. Statistical tests done included one-way ANOVA to compare levels of biomarkers between different groups.ResultsHigher levels of ICAM-1 were present in infants with BPD and correlated with its severity. Infants with BPD have significantly higher levels of ANG-2 and lower levels of ANG-1. Infants with PH have higher levels of: IL-6, IL-8, IL-10, and TNF-α. Infants with BPD-PH have significantly lower levels of MCP-1 and higher levels of IL-1β than infants with PH without BPD.ConclusionICAM-1 may be used as a specific biomarker for diagnosis of BPD and its severity.
| S-EPMC7664991 | biostudies-literature
Project description:ObjectivesElectronic Health Records (EHRs) contain scanned documents from a variety of sources such as identification cards, radiology reports, clinical correspondence, and many other document types. We describe the distribution of scanned documents at one health institution and describe the design and evaluation of a system to categorize documents into clinically relevant and non-clinically relevant categories as well as further sub-classifications. Our objective is to demonstrate that text classification systems can accurately classify scanned documents.MethodsWe extracted text using Optical Character Recognition (OCR). We then created and evaluated multiple text classification machine learning models, including both "bag of words" and deep learning approaches. We evaluated the system on three different levels of classification using both the entire document as input, as well as the individual pages of the document. Finally, we compared the effects of different text processing methods.ResultsA deep learning model using ClinicalBERT performed best. This model distinguished between clinically-relevant documents and not clinically-relevant documents with an accuracy of 0.973; between intermediate sub-classifications with an accuracy of 0.949; and between individual classes with an accuracy of 0.913.DiscussionWithin the EHR, some document categories such as "external medical records" may contain hundreds of scanned pages without clear document boundaries. Without further sub-classification, clinicians must view every page or risk missing clinically-relevant information. Machine learning can automatically classify these scanned documents to reduce clinician burden.ConclusionUsing machine learning applied to OCR-extracted text has the potential to accurately identify clinically-relevant scanned content within EHRs.
| S-EPMC7731898 | biostudies-literature
Project description:BackgroundPulmonary hypertension (PH) in patients with bronchopulmonary dysplasia (BPD) results from vasoconstriction and/or vascular remodeling, which can be regulated by mitogen-activated protein kinases (MAPKs). MAPKs are deactivated by dual-specificity phosphatases (DUSPs). We hypothesized that single-nucleotide polymorphisms (SNPs) in DUSP genes could be used to predict PH in BPD.MethodsPreterm infants diagnosed with BPD (n = 188) were studied. PH was defined by echocardiographic criteria. Genomic DNA isolated from patient blood samples was analyzed for 31 SNPs in DUSP genes. Clinical characteristics and minor allele frequencies were compared between BPD-PH (cases) and BPD-without PH (control) groups. Biomarker models to predict PH in BPD using clinical and SNP data were tested by calculations of area under the ROC curve.ResultsIn our BPD cohort, 32% (n = 61) had PH. Of the DUSP SNPs evaluated, DUSP1 SNP rs322351 was less common, and DUSP5 SNPs rs1042606 and rs3793892 were more common in cases than in controls. The best fit biomarker model combines clinical and DUSP genetic data with an area under the ROC curve of 0.76.ConclusionWe identified three DUSP SNPs as potential BPD-PH biomarkers. Combining clinical and DUSP genetic data yields the most robust predictor for PH in BPD.
| S-EPMC6962530 | biostudies-literature
Project description:Evaluation of large-scale data sets is needed to better understand the epidemiology, cost, and burden of atopic dermatitis (AD). We sought to validate the use of ICD-9-CM codes for identifying AD.Patients from a large metropolitan quaternary care medical center with a diagnostic code of either 691.8 (AD) or 692.9 (eczema and contact dermatitis) were queried. Medical records were reviewed for demographics, Hanifin & Rajka (H&R) and United Kingdom Working Party (UKWP) criteria. Sensitivity, specificity, and positive predictive values (PPV) of the codes were calculated.Of 43 278 patients identified with associated ICD-9 codes of 691.8 or 692.9, 519 and 253 with 691.8 and 692.9 were randomly selected for chart review. There was extensive overlap: 34.3% had ?1 occurrences of 691.8 and 692.9 and 25.6% had multiple occurrences of both codes. Among patients with ?1 occurrence of 691.8, 29.9% and 30.8% met the H&R and UKWP criteria, respectively. Similarly, among patients with ?1 occurrence of 692.9, 33.7% and 32.2% met the H&R and UKWP criteria. Increased PPV was associated with concomitant diagnoses of asthma, hay fever, and food allergy and increased disease severity.In the outpatient setting, the ICD-9-CM codes 691.8 and 692.9 alone have poor PPV. Incorporation of diagnoses of asthma, hay fever, and food allergy improves PPV and specificity. In the inpatient setting, a primary discharge diagnosis of 691.8 had excellent PPV. Although ICD-10 has been adopted in Europe and more recently in the USA, the same systematic errors would likely occur unless providers standardize their coding.
| S-EPMC5461190 | biostudies-literature