Prevalence and Mortality of Life-Threatening and Life-Shortening Diseases in Children and Adolescents in Germany.

This study provides prevalence and mortality data for 0- to 19-year-old children and adolescents with medically documented life-threatening and life-shortening diagnoses in Germany. A secondary data analysis of more than 12 million insured persons documented by the statutory health insurance funds in Germany from 2014 to 2019 was conducted in collaboration with the German Association of Statutory Health Insurance Funds (GKV-SV) and the Institute for Applied Health Research Berlin (InGef), whose data sets vary in collection methods. Diagnosis prevalence and mortality were calculated based on selected International Classification of Diseases, 10th Revision (ICD-10) codes reported in inpatient and outpatient care settings. In Germany, the diagnosis prevalence of life-threatening and life-shortening diseases in children and adolescents ranges between 319 948 (InGef-adapted Fraser list) and 402 058 (GKV-SV). These diagnoses can be differentiated into different disease groups (Together-for-Short-Lives [TfSL] 1-4). The TfSL-1 group in which curative treatment can be feasible represents the largest one, with 190 865 persons. In 2019, approximately 1458 children and adolescents with life-threatening and life-shortening diseases died. The current diagnostic and mortality data of affected children and adolescents in Germany serve as the essential foundation for further research into the health care of the target group.

Performance of International Classification of Disease-10 codes in detecting emergency department patients with opioid misuse.

BACKGROUND AND AIMS: Accurate case discovery is critical for disease surveillance, resource allocation and research. International Classification of Disease (ICD) diagnosis codes are commonly used for this purpose. We aimed to determine the sensitivity, specificity and positive predictive value (PPV) of ICD-10 codes for opioid misuse case discovery in the emergency department (ED) setting. DESIGN AND SETTING: Retrospective cohort study of ED encounters from January 2018 to December 2020 at an urban academic hospital in the United States. A sample of ED encounters enriched for opioid misuse was developed by oversampling ED encounters with positive urine opiate screens or pre-existing opioid-related diagnosis codes in addition to other opioid misuse risk factors. CASES: A total of 1200 randomly selected encounters were annotated by research staff for the presence of opioid misuse within health record documentation using a 5-point scale for likelihood of opioid misuse and dichotomized into cohorts of opioid misuse and no opioid misuse. MEASUREMENTS: Using manual annotation as ground truth, the sensitivity and specificity of ICD-10 codes entered during the encounter were determined with PPV adjusted for oversampled data. Metrics were also determined by disposition subgroup: discharged home or admitted. FINDINGS: There were 541 encounters annotated as opioid misuse and 617 with no opioid misuse. The majority were males (54.4%), average age was 47 years and 68.5% were discharged directly from the ED. The sensitivity of ICD-10 codes was 0.56 (95% confidence interval [CI], 0.51-0.60), specificity 0.99 (95% CI, 0.97-0.99) and adjusted PPV 0.78 (95% CI, 0.65-0.92). The sensitivity was higher for patients discharged from the ED (0.65; 95% CI, 0.60-0.69) than those admitted (0.31; 95% CI, 0.24-0.39). CONCLUSIONS: International Classification of Disease-10 codes appear to have low sensitivity but high specificity and positive predictive value in detecting opioid misuse among emergency department patients in the United States.

Applying Deep Learning Model to Predict Diagnosis Code of Medical Records.

The International Classification of Diseases (ICD) code is a diagnostic classification standard that is frequently used as a referencing system in healthcare and insurance. However, it takes time and effort to find and use the right diagnosis code based on a patient's medical records. In response, deep learning (DL) methods have been developed to assist physicians in the ICD coding process. Our findings propose a deep learning model that utilized clinical notes from medical records to predict ICD-10 codes. Our research used text-based medical data from the outpatient department (OPD) of a university hospital from January to December 2016. The dataset used clinical notes from five departments, and a total of 21,953 medical records were collected. Clinical notes consisted of a subjective component, objective component, assessment, plan (SOAP) notes, diagnosis code, and a drug list. The dataset was divided into two groups: 90% for training and 10% for test cases. We applied natural language processing (NLP) technique (word embedding, Word2Vector) to process the data. A deep learning-based convolutional neural network (CNN) model was created based on the information presented above. Three metrics (precision, recall, and F-score) were used to calculate the achievement of the deep learning CNN model. Clinically acceptable results were achieved through the deep learning model for five departments (precision: 0.53-0.96; recall: 0.85-0.99; and F-score: 0.65-0.98). With a precision of 0.95, a recall of 0.99, and an F-score of 0.98, the deep learning model performed the best in the department of cardiology. Our proposed CNN model significantly improved the prediction performance for an automated ICD-10 code prediction system based on prior clinical information. This CNN model could reduce the laborious task of manual coding and could assist physicians in making a better diagnosis.

[PraeKids: diagnosis prevalence of life-threatening and life-limiting diseases in children and adolescents in Germany]. / PraeKids: Diagnoseprävalenz lebensbedrohlicher und lebensverkürzender Erkrankungen bei Kindern und Jugendlichen in Deutschland.

INTRODUCTION: In Germany, a prevalence of approximately 50,000 children and adolescents living with life-threatening and life-limiting diseases is currently assumed. This number, which is communicated in the supply landscape, is based on a simple transfer of empirical data from England. METHODS: In cooperation with the German National Association of Health Insurance Funds (GKV-SV) and the Institute for Applied Health Research Berlin GmbH (InGef), the billing data of the specific treatment diagnoses documented by statutory health insurance funds for the years 2014-2019 were analyzed and - for the first time - prevalence data of affected 0­ to 19-year-olds were collected. In addition, the data from InGef was used to calculate the prevalence by diagnosis grouping, the Together for Short Lives (TfSL) groups 1-4, and based on the (updated) coding lists from the English prevalence studies. RESULTS: The data analysis determined a prevalence range of 319,948 (InGef - adapted Fraser list) to 402,058 (GKV-SV), with consideration of the TfSL groups. The TfSL­1 group represents the largest group with 190,865 patients. DISCUSSION AND CONCLUSION: This study is the first to provide the prevalence of 0­ to 19-year-olds living with life-threatening or life-limiting diseases in Germany. Since the case definitions and the included care settings (outpatient/inpatient) differ in the research design, the prevalence values collected from the GKV-SV and InGef are also different. Due to the very heterogeneous course of the diseases, chances of survival, and mortality rates, no direct conclusions can be drawn for palliative and hospice care structures.

Identifying Disease of Interest With Deep Learning Using Diagnosis Code.

BACKGROUND: Autoencoder (AE) is one of the deep learning techniques that uses an artificial neural network to reconstruct its input data in the output layer. We constructed a novel supervised AE model and tested its performance in the prediction of a co-existence of the disease of interest only using diagnostic codes. METHODS: Diagnostic codes of one million randomly sampled patients listed in the Korean National Health Information Database in 2019 were used to train, validate, and test the prediction model. The first used AE solely for a feature engineering tool for an input of a classifier. Supervised Multi-Layer Perceptron (sMLP) was added to train a classifier to predict a binary level with latent representation as an input (AE + sMLP). The second model simultaneously updated the parameters in the AE and the connected MLP classifier during the learning process (End-to-End Supervised AE [EEsAE]). We tested the performances of these two models against baseline models, eXtreme Gradient Boosting (XGB) and naïve Bayes, in the prediction of co-existing gastric cancer diagnosis. RESULTS: The proposed EEsAE model yielded the highest F1-score and highest area under the curve (0.86). The EEsAE and AE + sMLP gave the highest recalls. XGB yielded the highest precision. Ablation study revealed that iron deficiency anemia, gastroesophageal reflux disease, essential hypertension, gastric ulcers, benign prostate hyperplasia, and shoulder lesion were the top 6 most influential diagnoses on performance. CONCLUSION: A novel EEsAE model showed promising performance in the prediction of a disease of interest.

Identifying monoclonal gammopathy of undetermined significance from electronic health records.

BACKGROUND: Monoclonal gammopathy of undetermined significance (MGUS) precedes multiple myeloma (MM). Use of electronic health records may facilitate large-scale epidemiologic research to elucidate risk factors for the progression of MGUS to MM or other lymphoid malignancies. AIMS: We evaluated the accuracy of an electronic health records-based approach for identifying clinically diagnosed MGUS cases for inclusion in studies of patient outcomes/ progression risk. METHODS AND RESULTS: Data were retrieved from Kaiser Permanente Southern California's comprehensive electronic health records, which contain documentation of all outpatient and inpatient visits, laboratory tests, diagnosis codes and a cancer registry. We ascertained potential MGUS cases diagnosed between 2008 and 2014 using the presence of an MGUS ICD-9 diagnosis code (273.1). We initially excluded those diagnosed with MM within 6 months after MGUS diagnosis, then subsequently those with any lymphoid malignancy diagnosis from 2007 to 2014. We reviewed medical charts for 100 randomly selected potential cases for evidence of a physician diagnosis of MGUS, which served as our gold standard for case confirmation. To assess sensitivity, we also investigated the presence of the ICD-9 code in the records of 40 randomly selected and chart review-confirmed MGUS cases among patients with a laboratory report of elevated circulating monoclonal (M-) protein (a key test for MGUS diagnosis) and no subsequent lymphoid malignancy (as described above). The positive predictive value (PPV) for the ICD-9 code was 98%. All MGUS cases confirmed by chart review also had confirmatory laboratory test results. Of the confirmed cases first identified via M-protein test results, 88% also had the ICD-9 diagnosis code. CONCLUSION: The diagnosis code-based approach has excellent PPV and likely high sensitivity for detecting clinically diagnosed MGUS. The generalizability of this approach outside an integrated healthcare system warrants further evaluation.

Applying Convolutional Neural Networks to Predict the ICD-9 Codes of Medical Records.

The International Statistical Classification of Disease and Related Health Problems (ICD) is an international standard system for categorizing and reporting diseases, injuries, disorders, and health conditions. Most previously-proposed disease predicting systems need clinical information collected by the medical staff from the patients in hospitals. In this paper, we propose a deep learning algorithm to classify disease types and identify diagnostic codes by using only the subjective component of progress notes in medical records. In this study, we have a dataset, consisting of about one hundred and sixty-eight thousand medical records, from a medical center, collected during 2003 and 2017. First, we apply standard text processing procedures to parse the sentences and word embedding techniques for vector representations. Next, we build a convolution neural network model on the medical records to predict the ICD-9 code by using a subjective component of the progress note. The prediction performance is evaluated by ten-fold cross-validation and yields an accuracy of 0.409, recall of 0.409 and precision of 0.436. If we only consider the "chapter match" of ICD-9 code, our model achieves an accuracy of 0.580, recall of 0.580, and precision of 0.582. Since our diagnostic code prediction model is solely based on subjective components (mainly, patients' self-report descriptions), the proposed approach could serve as a remote and self-diagnosis assistance tool, prior to seeking medical advice or going to the hospital. In addition, our work may be used as a primary evaluation tool for discomfort in the rural area where medical resources are restricted.

Validity of ICD-10 diagnoses of overweight and obesity in Danish hospitals.

PURPOSE: Health care databases may be a valuable source for epidemiological research in obesity, if diagnoses are valid. We examined the validity and completeness of International Classification of Diseases, 10th revision [ICD-10] diagnosis coding for overweight/obesity in Danish hospitals. PATIENTS AND METHODS: We linked data from the Danish National Patient Registry on patients with a hospital diagnosis code of overweight/obesity (ICD-10 code E66) with computerized height and weight measurements made during hospital contacts in the Central Denmark Region Clinical Information System. We computed the positive predictive value (PPV) of the IDC-10 diagnosis of overweight/obesity, using a documented body mass index (BMI) ≥25 kg/m2 as gold standard. We also examined the completeness of obesity/overweight diagnosis coding among all patients recorded with BMI ≥25 kg/m2. RESULTS: Of all 19,672 patients registered with a first diagnosis code of overweight/obesity in the National Patient Registry, 17,351 patients (88.2%) had any BMI measurement recorded in the Central Denmark Region Clinical Information System, and 17,240 patients (87.6%) had a BMI ≥25 kg/m2, yielding a PPV of 87.6% (95% CI: 87.2-88.1). The PPV was slightly higher for primary diagnosis codes of overweight/obesity: 94.1% (95% CI: 93.3-94.8) than for secondary diagnosis codes: 86.1% (95% CI: 85.6-86.6). The PPV increased with higher patient age: from 75.3% (95% CI: 73.8-76.9) in those aged 18-29 years to 94.7% (95% CI: 92.6-96.9) in patients aged 80 years and above. Completeness of obesity/overweight diagnosis coding among patients recorded with BMI ≥25 kg/m2 was only 10.9% (95% CI: 10.8-11.0). CONCLUSION: Our findings indicate a high validity of the ICD-10 code E66 for overweight/obesity when recorded; however, completeness of coding was low. Nonetheless, ICD-10 discharge codes may be a suitable source of data on overweight/obesity for epidemiological research.

Congenital microcephaly hospitalizations in California infants: 1999-2013.

INTRODUCTION: Population-level changes in microcephaly incidence risk (IR) could signal circulation of neurotropic pathogens or potential emerging teratogen exposure. METHODS: In this retrospective population cohort study, we estimated the IR of hospitalizations with a microcephaly ICD-9-CM discharge diagnosis code among infants ≤1 year over a 15-year period (1999-2013) using the Electronic Health Record (EHR) database from all hospital discharges in California from the Office of Statewide Hospital Planning and Development (OSHPD) database. We calculated the overall and yearly IRs per 10,000 live births (LBs) and per 10,000 hospitalizations in infants ≤1 year, and explored the impact in the IR estimates when children with microcephaly associated comorbidities were excluded or not. RESULTS: Among 8,860,153 hospital discharges of infants ≤1 year in the OSHPD database over this 15 year period, we identified 6,004 hospitalizations with a microcephaly discharge diagnosis code; 3,526 of those were in neonates ≤30 days. The IR of microcephaly hospitalizations for infants ≤1 year was 7.70/10,000 LB (for neonates it was 4.52/10,000 LB) and 6.78 per 10,000 hospitalizations ≤1 year. There was large heterogeneity in the yearly microcephaly IRs (I2 = 66.6%). DISCUSSION: EHR collected data could be used as a complementary approach to track epidemiologic changes in microcephaly IRs. However, standardization in the use of microcephaly discharge diagnosis code and harmonization in the types of additional comorbidities to be excluded across analyses is mandatory to allow for prompt identification of true changes in microcephaly rates over time.

Validation of diagnosis codes to identify side of colon in an electronic health record registry.

BACKGROUND: The use of real-world data to generate evidence requires careful assessment and validation of critical variables before drawing clinical conclusions. Prospective clinical trial data suggest that anatomic origin of colon cancer impacts prognosis and treatment effectiveness. As an initial step in validating this observation in routine clinical settings, we explored the feasibility and accuracy of obtaining information on tumor sidedness from electronic health records (EHR) billing codes. METHODS: Nine thousand four hundred three patients with metastatic colorectal cancer (mCRC) were selected from the Flatiron Health database, which is derived from de-identified EHR data. This study included a random sample of 200 mCRC patients. Tumor site data derived from International Classification of Diseases (ICD) codes were compared with data abstracted from unstructured documents in the EHR (e.g. surgical and pathology notes). Concordance was determined via observed agreement and Cohen's kappa coefficient (κ). Accuracy of ICD codes for each tumor site (left, right, transverse) was determined by calculating the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), and corresponding 95% confidence intervals, using abstracted data as the gold standard. RESULTS: Study patients had similar characteristics and side of colon distribution compared with the full mCRC dataset. The observed agreement between the ICD codes and abstracted data for tumor site for all sampled patients was 0.58 (κ = 0.41). When restricting to the 62% of patients with a side-specific ICD code, the observed agreement was 0.84 (κ = 0.79). The specificity (92-98%) of structured data for tumor location was high, with lower sensitivity (49-63%), PPV (64-92%) and NPV (72-97%). Demographic and clinical characteristics were similar between patients with specific and non-specific side of colon ICD codes. CONCLUSIONS: ICD codes are a highly reliable indicator of tumor location when the specific location code is entered in the EHR. However, non-specific side of colon ICD codes are present for a sizable minority of patients, and structured data alone may not be adequate to support testing of some research hypotheses. Careful assessment of key variables is required before determining the need for clinical abstraction to supplement structured data in generating real-world evidence from EHRs.

Using natural language processing for identification of herpes zoster ophthalmicus cases to support population-based study.

IMPORTANCE: Diagnosis codes are inadequate for accurately identifying herpes zoster (HZ) ophthalmicus (HZO). There is significant lack of population-based studies on HZO due to the high expense of manual review of medical records. BACKGROUND: To assess whether HZO can be identified from the clinical notes using natural language processing (NLP). To investigate the epidemiology of HZO among HZ population based on the developed approach. DESIGN: A retrospective cohort analysis. PARTICIPANTS: A total of 49 914 southern California residents aged over 18 years, who had a new diagnosis of HZ. METHODS: An NLP-based algorithm was developed and validated with the manually curated validation data set (n = 461). The algorithm was applied on over 1 million clinical notes associated with the study population. HZO versus non-HZO cases were compared by age, sex, race and co-morbidities. MAIN OUTCOME MEASURES: We measured the accuracy of NLP algorithm. RESULTS: NLP algorithm achieved 95.6% sensitivity and 99.3% specificity. Compared to the diagnosis codes, NLP identified significant more HZO cases among HZ population (13.9% vs. 1.7%). Compared to the non-HZO group, the HZO group was older, had more males, had more Whites and had more outpatient visits. CONCLUSIONS AND RELEVANCE: We developed and validated an automatic method to identify HZO cases with high accuracy. As one of the largest studies on HZO, our finding emphasizes the importance of preventing HZ in the elderly population. This method can be a valuable tool to support population-based studies and clinical care of HZO in the era of big data.

Exploration of association rule mining for coding consistency and completeness assessment in inpatient administrative health data.

OBJECTIVE: Data quality assessment is a challenging facet for research using coded administrative health data. Current assessment approaches are time and resource intensive. We explored whether association rule mining (ARM) can be used to develop rules for assessing data quality. MATERIALS AND METHODS: We extracted 2013 and 2014 records from the hospital discharge abstract database (DAD) for patients between the ages of 55 and 65 from five acute care hospitals in Alberta, Canada. The ARM was conducted using the 2013 DAD to extract rules with support ≥0.0019 and confidence ≥0.5 using the bootstrap technique, and tested in the 2014 DAD. The rules were compared against the method of coding frequency and assessed for their ability to detect error introduced by two kinds of data manipulation: random permutation and random deletion. RESULTS: The association rules generally had clear clinical meanings. Comparing 2014 data to 2013 data (both original), there were 3 rules with a confidence difference >0.1, while coding frequency difference of codes in the right hand of rules was less than 0.004. After random permutation of 50% of codes in the 2014 data, average rule confidence dropped from 0.72 to 0.27 while coding frequency remained unchanged. Rule confidence decreased with the increase of coding deletion, as expected. Rule confidence was more sensitive to code deletion compared to coding frequency, with slope of change ranging from 1.7 to 184.9 with a median of 9.1. CONCLUSION: The ARM is a promising technique to assess data quality. It offers a systematic way to derive coding association rules hidden in data, and potentially provides a sensitive and efficient method of assessing data quality compared to standard methods.

Limited Accuracy of Administrative Data for the Identification and Classification of Adult Congenital Heart Disease.

BACKGROUND: Administrative data sets utilize billing codes for research and quality assessment. Previous data suggest that such codes can accurately identify adults with congenital heart disease (CHD) in the cardiology clinic, but their use has yet to be validated in a larger population. METHODS AND RESULTS: All administrative codes from an entire health system were queried for a single year. Adults with a CHD diagnosis code (International Classification of Diseases, Ninth Revision, (ICD-9) codes 745-747) defined the cohort. A previously validated hierarchical algorithm was used to identify diagnoses and classify patients. All charts were reviewed to determine a gold standard diagnosis, and comparisons were made to determine accuracy. Of 2399 individuals identified, 206 had no CHD by the algorithm or were deemed to have an uncertain diagnosis after provider review. Of the remaining 2193, only 1069 had a confirmed CHD diagnosis, yielding overall accuracy of 48.7% (95% confidence interval, 47-51%). When limited to those with moderate or complex disease (n=484), accuracy was 77% (95% confidence interval, 74-81%). Among those with CHD, misclassification occurred in 23%. The discriminative ability of the hierarchical algorithm (C statistic: 0.79; 95% confidence interval, 0.77-0.80) improved further with the addition of age, encounter type, and provider (C statistic: 0.89; 95% confidence interval, 0.88-0.90). CONCLUSIONS: ICD codes from an entire healthcare system were frequently erroneous in detecting and classifying CHD patients. Accuracy was higher for those with moderate or complex disease or when coupled with other data. These findings should be taken into account in future studies utilizing administrative data sets in CHD.

Methods of defining hypertension in electronic medical records: validation against national survey data.

BACKGROUND: Electronic medical records (EMR) can be a cost-effective source for hypertension surveillance. However, diagnosis of hypertension in EMR is commonly under-coded and warrants the needs to review blood pressure and antihypertensive drugs for hypertension case identification. METHODS: We included all the patients actively registered in The Health Improvement Network (THIN) database, UK, on 31 December 2011. Three case definitions using diagnosis code, antihypertensive drug prescriptions and abnormal blood pressure, respectively, were used to identify hypertension patients. We compared the prevalence and treatment rate of hypertension in THIN with results from Health Survey for England (HSE) in 2011. RESULTS: Compared with prevalence reported by HSE (29.7%), the use of diagnosis code alone (14.0%) underestimated hypertension prevalence. The use of any of the definitions (38.4%) or combination of antihypertensive drug prescriptions and abnormal blood pressure (38.4%) had higher prevalence than HSE. The use of diagnosis code or two abnormal blood pressure records with a 2-year period (31.1%) had similar prevalence and treatment rate of hypertension with HSE. CONCLUSIONS: Different definitions should be used for different study purposes. The definition of 'diagnosis code or two abnormal blood pressure records with a 2-year period' could be used for hypertension surveillance in THIN.

An empirical evaluation of supervised learning approaches in assigning diagnosis codes to electronic medical records.

BACKGROUND: Diagnosis codes are assigned to medical records in healthcare facilities by trained coders by reviewing all physician authored documents associated with a patient's visit. This is a necessary and complex task involving coders adhering to coding guidelines and coding all assignable codes. With the popularity of electronic medical records (EMRs), computational approaches to code assignment have been proposed in the recent years. However, most efforts have focused on single and often short clinical narratives, while realistic scenarios warrant full EMR level analysis for code assignment. OBJECTIVE: We evaluate supervised learning approaches to automatically assign international classification of diseases (ninth revision) - clinical modification (ICD-9-CM) codes to EMRs by experimenting with a large realistic EMR dataset. The overall goal is to identify methods that offer superior performance in this task when considering such datasets. METHODS: We use a dataset of 71,463 EMRs corresponding to in-patient visits with discharge date falling in a two year period (2011-2012) from the University of Kentucky (UKY) Medical Center. We curate a smaller subset of this dataset and also use a third gold standard dataset of radiology reports. We conduct experiments using different problem transformation approaches with feature and data selection components and employing suitable label calibration and ranking methods with novel features involving code co-occurrence frequencies and latent code associations. RESULTS: Over all codes with at least 50 training examples we obtain a micro F-score of 0.48. On the set of codes that occur at least in 1% of the two year dataset, we achieve a micro F-score of 0.54. For the smaller radiology report dataset, the classifier chaining approach yields best results. For the smaller subset of the UKY dataset, feature selection, data selection, and label calibration offer best performance. CONCLUSIONS: We show that datasets at different scale (size of the EMRs, number of distinct codes) and with different characteristics warrant different learning approaches. For shorter narratives pertaining to a particular medical subdomain (e.g., radiology, pathology), classifier chaining is ideal given the codes are highly related with each other. For realistic in-patient full EMRs, feature and data selection methods offer high performance for smaller datasets. However, for large EMR datasets, we observe that the binary relevance approach with learning-to-rank based code reranking offers the best performance. Regardless of the training dataset size, for general EMRs, label calibration to select the optimal number of labels is an indispensable final step.

Refining Measurement of Substance Use Disorders Among Women of Child-Bearing Age Using Hospital Records: The Development of the Explicit-Mention Substance Abuse Need for Treatment in Women (EMSANT-W) Algorithm.

Substance use disorder (SUD) in women of reproductive age is associated with adverse health consequences for both women and their offspring. US states need a feasible population-based, case-identification tool to generate better approximations of SUD prevalence, treatment use, and treatment outcomes among women. This article presents the development of the Explicit Mention Substance Abuse Need for Treatment in Women (EMSANT-W), a gender-tailored tool based upon existing International Classification of Diseases, 9th Edition, Clinical Modification diagnostic code-based groupers that can be applied to hospital administrative data. Gender-tailoring entailed the addition of codes related to infants, pregnancy, and prescription drug abuse, as well as the creation of inclusion/exclusion rules based on other conditions present in the diagnostic record. Among 1,728,027 women and associated infants who accessed hospital care from January 1, 2002 to December 31, 2008 in Massachusetts, EMSANT-W identified 103,059 women with probable SUD. EMSANT-W identified 4,116 women who were not identified by the widely used Clinical Classifications Software for Mental Health and Substance Abuse (CCS-MHSA) and did not capture 853 women identified by CCS-MHSA. Content and approach innovations in EMSANT-W address potential limitations of the Clinical Classifications Software, and create a methodologically sound, gender-tailored and feasible population-based tool for identifying women of reproductive age in need of further evaluation for SUD treatment. Rapid changes in health care service infrastructure, delivery systems and policies require tools such as the EMSANT-W that provide more precise identification methods for sub-populations and can serve as the foundation for analyses of treatment use and outcomes.

Improving discharge data fidelity for use in large administrative databases.

OBJECT: Large administrative databases have assumed a major role in population-based studies examining health care delivery. Lumbar fusion surgeries specifically have been scrutinized for rising rates coupled with ill-defined indications for fusion such as stenosis and spondylosis. Administrative databases classify cases with the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). The ICD-9-CM discharge codes are not designated by surgeons, but rather are assigned by trained hospital medical coders. It is unclear how accurately they capture the surgeon's indication for fusion. The authors first sought to compare the ICD-9-CM code(s) assigned by the medical coder according to the surgeon's indication based on a review of the medical chart, and then to elucidate barriers to data fidelity. METHODS: A retrospective review was undertaken of all lumbar fusions performed in the Department of Neurosurgery at the authors' institution between August 1, 2011, and August 31, 2013. Based on this review, the indication for fusion in each case was categorized as follows: spondylolisthesis, deformity, tumor, infection, nonpathological fracture, pseudarthrosis, adjacent-level degeneration, stenosis, degenerative disc disease, or disc herniation. These surgeon diagnoses were compared with the primary ICD-9-CM codes that were generated by the medical coders and submitted to administrative databases. A follow-up interview with the hospital's coders and coding manager was undertaken to review causes of error and suggestions for future improvement in data fidelity. RESULTS: There were 178 lumbar fusion operations performed in the course of 170 hospital admissions. There were 44 hospitalizations in which fusion was performed for tumor, infection, or nonpathological fracture. Of these, the primary diagnosis matched the surgical indication for fusion in 98% of cases. The remaining 126 hospitalizations were for degenerative diseases, and of these, the primary ICD-9-CM diagnosis matched the surgeon's diagnosis in only 61 (48%) of 126 cases of degenerative disease. When both the primary and all secondary ICD-9-CM diagnoses were considered, the indication for fusion was identified in 100 (79%) of 126 cases. Still, in 21% of hospitalizations, the coder did not identify the surgical diagnosis, which was in fact present in the chart. There are many different causes of coding inaccuracy and data corruption. They include factors related to the quality of documentation by the physicians, coder training and experience, and ICD code ambiguity. CONCLUSIONS: Researchers, policymakers, payers, and physicians should note these limitations when reviewing studies in which hospital claims data are used. Advanced domain-specific coder training, increased attention to detail and utilization of ICD-9-CM diagnoses by the surgeon, and improved direction from the surgeon to the coder may augment data fidelity and minimize coding errors. By understanding sources of error, users of these large databases can evaluate their limitations and make more useful decisions based on them.

Assessing the burden of diabetes mellitus in emergency departments in the United States: the National Hospital Ambulatory Medical Care Survey (NHAMCS).

OBJECTIVE: To evaluate the performance of three alternative methods to identify diabetes in patients visiting Emergency Departments (EDs), and to describe the characteristics of patients with diabetes who are not identified when the alternative methods are used. RESEARCH DESIGN AND METHODS: We used data from the National Hospital Ambulatory Medical Care Survey (NHAMCS) 2009 and 2010. We assessed the sensitivity and specificity of using providers' diagnoses and diabetes medications (both excluding and including biguanides) to identify diabetes compared to using the checkbox for diabetes as the gold standard. We examined the characteristics of patients whose diabetes was missed using multivariate Poisson regression models. RESULTS: The checkbox identified 5,567 ED visits by adult patients with diabetes. Compared to the checkbox, the sensitivity was 12.5% for providers' diagnoses alone, 20.5% for providers' diagnoses and diabetes medications excluding biguanides, and 21.5% for providers' diagnoses and diabetes medications including biguanides. The specificity of all three of the alternative methods was >99%. Older patients were more likely to have diabetes not identified. Patients with self-payment, those who had glucose measured or received IV fluids in the ED, and those with more diagnosis codes and medications, were more likely to have diabetes identified. CONCLUSIONS: NHAMCS's providers' diagnosis codes and medication lists do not identify the majority of patients with diabetes visiting EDs. The newly introduced checkbox is helpful in measuring ED resource utilization by patients with diabetes.