Stroke Risk After Emergency Department Treat-and-Release Visit for a Fall.

BACKGROUND: Previous cohort studies of hospitalized patients with a delayed diagnosis of ischemic stroke found that these patients often had an initial emergency department (ED) diagnosis of a fall. We sought to evaluate whether ED visits for a fall resulting in discharge to home (ie, treat-and-release visits) were associated with increased short-term ischemic stroke risk. METHODS: A case-crossover design was used to compare ED visits for falls during case periods (0-15, 16-30, 31-90, and 91-180 days before stroke) and control periods (equivalent time periods exactly 1 year before stroke) using administrative data from the Healthcare Cost and Utilization Project on all hospital admissions and ED visits across 10 states from 2016 to 2020. To identify ED treat-and-release visits for a fall and patients hospitalized for acute ischemic stroke, we used previously validated International Classification of Diseases, Tenth Revision, Clinical Modification codes. Odds ratios and 95% CIs were calculated using conditional logistic regression. RESULTS: Among 90 592 hospitalized patients with ischemic stroke, 5230 (5.8%) had an ED treat-and-release visit for a fall within 180 days before their stroke. Patients with an ED treat-and-release visit for a fall were older (mean age, 74.7 [SD, 14.6] versus 70.8 [SD, 15.1] years), more often female (61.9% versus 53.4%), and had higher rates of vascular comorbidities than other patients with stroke. ED treat-and-release visits for a fall were significantly more common in the 15 days before stroke compared with the 15-day control period 1 year earlier (odds ratio, 2.7 [95% CI, 2.4-3.1]). The association between stroke and a preceding ED treat-and-release visit for a fall decreased in magnitude with increasing temporal distance from stroke. CONCLUSIONS: ED treat-and-release visits for a fall are associated with significantly increased short-term ischemic stroke risk. These visits may be opportunities to improve stroke diagnostic accuracy and treatment in the ED.

Power Has Pitfalls: (In)accuracy of Administrative Data for Nonoperative Management of Appendicitis.

INTRODUCTION: Nonoperative management (NOM) of uncomplicated appendicitis is increasingly common. Effectiveness of NOM has been studied by identifying patients via International Classification of Diseases (ICD) 9/ICD-10 codes for uncomplicated appendicitis and no code for appendectomy. We sought to assess the accuracy of such administrative definitions. METHODS: We retrospectively identified patients with ICD-9/ICD-10 codes for appendicitis at five sites across the United States. Initial management plan and clinical severity were recorded by trained abstractors. We identified a gold standard cohort of patients with surgeon-diagnosed uncomplicated appendicitis and planned NOM. We defined two administrative cohorts with ICD-9/ICD-10 codes for uncomplicated appendicitis and either no surgery during initial admission (definition #1) or no surgery on day 0-1 of admission (definition #2). We compared each definition to the gold standard. RESULTS: Among 1224 patients with uncomplicated appendicitis, 72 (5.9%) underwent planned NOM. NOM patients were older (median [Q1-Q3] of 37 [27-56] versus 32 [25-44] y) and less frequently male (51.4% versus 54.9%), White (54.1% versus 67.6%), and privately insured (38.9% versus 50.2%) than patients managed operatively. Definition #1 had sensitivity of 0.81 and positive predictive value of 0.87 for NOM of uncomplicated appendicitis. Definition #2 had sensitivity of 0.83 and positive predictive value of 0.72. The gold standard cohort had a true failure/recurrence rate of 23.6%, compared with apparent rates of 25.4% and 39.8%, respectively. CONCLUSIONS: Administrative definitions are prone to misclassification in identifying planned NOM of uncomplicated appendicitis. This likely impacts outcomes in studies using administrative databases. Investigators should disclose how misclassification may affect results and select an administrative definition that optimally balances sensitivity and specificity for their research question.

Prevalence and correlates of ICD-11 prolonged grief disorder among adults living in Ukraine during the war with Russia.

BACKGROUND: Although high rates of bereavement are evident in war-affected populations, no study has investigated the prevalence and correlates of probable ICD-11 prolonged grief disorder (PGD) under these circumstances. METHODS: Participants were 2050 adults who participated in a nationwide survey exploring the effects of the Ukraine-Russia war on the daily lives and mental health of Ukrainian people. RESULTS: Of the total sample, 87.7% (n = 1797) of people indicated a lifetime bereavement. In the full sample, 11.4% met the diagnostic requirements for probable ICD-11 PGD, and amongst those with a lifetime bereavement, the conditional rate of probable ICD-11 PGD was 13.0%. Significant risk factors of ICD-11 PGD included the recent loss of a loved one (6 months to a year ago), being most affected by a partner or spouse's death, loved one dying in the war, no recent contact with the deceased prior to their death, and meeting depression and anxiety diagnostic requirements. CONCLUSION: The study reveals that a significant percentage of Ukrainian bereaved individuals have probable ICD-11 PGD, and identifying risk factors, particularly war-related losses, will aid in the development of intervention and prevention programs for bereaved adults.

Validity of ICD-10 diagnosis codes for placenta accreta spectrum disorders.

BACKGROUND: The 10th revision of the International Classification of Diseases, Clinical Modification (ICD-10) includes diagnosis codes for placenta accreta spectrum for the first time. These codes could enable valuable research and surveillance of placenta accreta spectrum, a life-threatening pregnancy complication that is increasing in incidence. OBJECTIVE: We sought to evaluate the validity of placenta accreta spectrum diagnosis codes that were introduced in ICD-10 and assess contributing factors to incorrect code assignments. METHODS: We calculated sensitivity, specificity, positive predictive value and negative predictive value of the ICD-10 placenta accreta spectrum code assignments after reviewing medical records from October 2015 to March 2020 at a quaternary obstetric centre. Histopathologic diagnosis was considered the gold standard. RESULTS: Among 22,345 patients, 104 (0.46%) had an ICD-10 code for placenta accreta spectrum and 51 (0.23%) had a histopathologic diagnosis. ICD-10 codes had a sensitivity of 0.71 (95% CI 0.56, 0.83), specificity of 0.98 (95% CI 0.93, 1.00), positive predictive value of 0.61 (95% CI 0.48, 0.72) and negative predictive value of 1.00 (95% CI 0.96, 1.00). The sensitivities of the ICD-10 codes for placenta accreta spectrum subtypes- accreta, increta and percreta-were 0.55 (95% CI 0.31, 0.78), 0.33 (95% CI 0.12, 0.62) and 0.56 (95% CI 0.31, 0.78), respectively. Cases with incorrect code assignment were less morbid than cases with correct code assignment, with a lower incidence of hysterectomy at delivery (17% vs 100%), blood transfusion (26% vs 75%) and admission to the intensive care unit (0% vs 53%). Primary reasons for code misassignment included code assigned to cases of occult placenta accreta (35%) or to cases with clinical evidence of placental adherence without histopatholic diagnostic (35%) features. CONCLUSION: These findings from a quaternary obstetric centre suggest that ICD-10 codes may be useful for research and surveillance of placenta accreta spectrum, but researchers should be aware of likely substantial false positive cases.

Model correction of diagnostic coding-based RSV incidence for children 0-4 years in the US.

BACKGROUND: Although administrative claims data have a high degree of completeness, not all medically attended Respiratory Syncytial Virus-associated lower respiratory tract infections (RSV-LRTIs) are tested or coded for their causative agent. We sought to determine the attribution of RSV to LRTI in claims data via modeling of temporal changes in LRTI rates against surveillance data. METHODS: We estimated the weekly incidence of LRTI (inpatient, outpatient, and total) for children 0-4 years using 2011-2019 commercial insurance claims, stratified by HHS region, matched to the corresponding weekly NREVSS RSV and influenza positivity data for each region, and modelled against RSV, influenza positivity rates, and harmonic functions of time assuming negative binomial distribution. LRTI events attributable to RSV were estimated as predicted events from the full model minus predicted events with RSV positivity rate set to 0. RESULTS: Approximately 42% of predicted RSV cases were coded in claims data. Across all regions, the percentage of LRTI attributable to RSV were 15-43%, 10-31%, and 10-31% of inpatient, outpatient, and combined settings, respectively. However, when compared to coded inpatient RSV-LRTI, 9 of 10 regions had improbable corrected inpatient LRTI estimates (predicted RSV/coded RSV ratio < 1). Sensitivity analysis based on separate models for PCR and antigen-based positivity showed similar results. CONCLUSIONS: Underestimation based on coding in claims data may be addressed by NREVSS-based adjustment of claims-based RSV incidence. However, where setting-specific positivity rates is unavailable, we recommend modeling across settings to mirror NREVSS's positivity rates which are similarly aggregated, to avoid inaccurate adjustments.

Inverse association between Paleolithic Diet Fraction and mortality and incidence of cardiometabolic disease in the prospective Malmö Diet and Cancer Study.

PURPOSE: Paleolithic Diet Fraction (PDF) estimates how large a portion of the absolute dietary intake stems from food groups included in the Paleolithic diet. In randomized controlled trials higher PDFs have been associated with healthier levels of cardiometabolic risk markers. Our aim was to build upon these findings by examining associations between PDF and mortality and incidence of cardiometabolic disease in the prospective Malmö Diet and Cancer Study. METHODS: PDF was calculated from an interview-based, modified diet history method, and associations were estimated by using multivariable Cox proportional hazards regression. The examined cohort consisted of 24,104 individuals (44-74 years, 63% women) without previous coronary events, diabetes, or stroke at baseline (1992-1996). A total of 10,092 individuals died during a median follow-up of 18 years. RESULTS: Median PDF was 40% (0-90%). The adjusted hazard ratios (HR) for PDF as a continuous variable (from 0 to 100%) were for risk of death from all causes 0.55 [95% CI 0.45, 0.66], tumor 0.68 [95% CI 0.49, 0.93], cardiovascular 0.55 [95% CI 0.39, 0.78], respiratory 0.44 [95% CI 0.21, 0.90], neurological 0.26 [95% CI 0.11, 0.60], digestive, 0.10 [95% CI 0.03, 0.30], and other diseases 0.64 [95% CI 0.41, 1.00]. The corresponding HR for risk of coronary event was 0.61 [95% 0.43, 0.86], for ischemic stroke it was 0.73 [95% 0.48, 1.09] and for type 2 diabetes it was 0.82 [95% 0.61, 1.10]. CONCLUSION: Observational data suggest an inverse association between PDF and all-cause as well as cause-specific mortality and incidence of cardiometabolic disease.

Creating a computer assisted ICD coding system: Performance metric choice and use of the ICD hierarchy.

OBJECTIVE: Machine learning methods hold the promise of leveraging available data and generating higher-quality data while alleviating the data collection burden on healthcare professionals. International Classification of Diseases (ICD) diagnoses data, collected globally for billing and epidemiological purposes, represents a valuable source of structured information. However, ICD coding is a challenging task. While numerous previous studies reported promising results in automatic ICD classification, they often describe input data specific model architectures, that are heterogeneously evaluated with different performance metrics and ICD code subsets. This study aims to explore the evaluation and construction of more effective Computer Assisted Coding (CAC) systems using generic approaches, focusing on the use of ICD hierarchy, medication data and a feed forward neural network architecture. METHODS: We conduct comprehensive experiments using the MIMIC-III clinical database, mapped to the OMOP data model. Our evaluations encompass various performance metrics, alongside investigations into multitask, hierarchical, and imbalanced learning for neural networks. RESULTS: We introduce a novel metric, , tailored to the ICD coding task, which offers interpretable insights for healthcare informatics practitioners, aiding them in assessing the quality of assisted coding systems. Our findings highlight that selectively cherry-picking ICD codes diminish retrieval performance without performance improvement over the selected subset. We show that optimizing for metrics such as NDCG and AUPRC outperforms traditional F1-based metrics in ranking performance. We observe that Neural Network training on different ICD levels simultaneously offers minor benefits for ranking and significant runtime gains. However, our models do not derive benefits from hierarchical or class imbalance correction techniques for ICD code retrieval. CONCLUSION: This study offers valuable insights for researchers and healthcare practitioners interested in developing and evaluating CAC systems. Using a straightforward sequential neural network model, we confirm that medical prescriptions are a rich data source for CAC systems, providing competitive retrieval capabilities for a fraction of the computational load compared to text-based models. Our study underscores the importance of metric selection and challenges existing practices related to ICD code sub-setting for model training and evaluation.

Positive Predictive Value of Cirrhosis and Chronic Pancreatitis Diagnoses in Individuals with Alcohol Use Disorder: A Single-Center Study.

BACKGROUND: Although accuracy of diagnosis codes for cirrhosis and chronic pancreatitis (CP) has been evaluated in multiple studies, none have focused on patients with alcohol use disorders (AUD). We evaluated the positive predictive value (PPV) for a verified diagnosis of cirrhosis and CP in AUD patients treated at a tertiary care center. METHODS: We performed a detailed review of electronic health records for AUD patients assigned ICD-9 or 10 codes for alcoholic cirrhosis (ALC) (n = 199), CP (n = 200), or both (n = 200). We calculated PPV for a verified diagnosis of cirrhosis and CP and performed multivariable regression analysis to assess the impact of relevant factors on PPV for a verified diagnosis. RESULTS: PPV of cirrhosis was 81.2% (95% CI 77.0 to 84.9%) which increased to 87.5% (95% CI 83.8 to 90.6%) if the definition was relaxed to include alcohol-related hepatitis. PPV of CP was 54.5% (95% CI 49.5 to 59.5%) which increased to 78% (95% CI 73.6 to 82.0%) when recurrent acute pancreatitis was included in the definition. In multivariable analyses, the odds of a verified diagnosis were significantly higher in individuals aged 65+ years for both cirrhosis (OR 12.23, 95% CI 2.19 to 68.42) and CP (OR 8.84, 95% CI 2.7 to 28.93) and in ever smokers for CP (OR 1.95, 95% CI 1.05 to 3.65). CONCLUSION: PPV for diagnosis codes in AUD patients is high for a verified diagnosis of cirrhosis but only modest for CP. While administrative datasets can provide reliable information for cirrhosis, future studies should focus on ways to boost the diagnostic validity of administrative datasets for CP.

The application of the ICD-10 for antepartum stillbirth patients in a referral centre of Eastern China: a retrospective study from 2015 to 2022.

BACKGROUND: The causes of some stillbirths are unclear, and additional work must be done to investigate the risk factors for stillbirths. OBJECTIVE: To apply the International Classification of Disease-10 (ICD-10) for antepartum stillbirth at a referral center in eastern China. METHODS: Antepartum stillbirths were grouped according to the cause of death according to the International Classification of Disease-10 (ICD-10) criteria. The main maternal condition at the time of antepartum stillbirth was assigned to each patient. RESULTS: Antepartum stillbirths were mostly classified as fetal deaths of unspecified cause, antepartum hypoxia. Although more than half of the mothers were without an identified condition at the time of the antepartum stillbirth, where there was a maternal condition associated with perinatal death, maternal medical and surgical conditions and maternal complications during pregnancy were most common. Of all the stillbirths, 51.2% occurred between 28 and 37 weeks of gestation, the main causes of stillbirth at different gestational ages also differed. Autopsy and chromosomal microarray analysis (CMA) were recommended in all stillbirths, but only 3.6% received autopsy and 10.5% underwent chromosomal microarray analysis. CONCLUSIONS: The ICD-10 is helpful in classifying the causes of stillbirths, but more than half of the stillbirths in our study were unexplained; therefore, additional work must be done. And the ICD-10 score may need to be improved, such as by classifying stillbirths according to gestational age. Autopsy and CMA could help determine the cause of stillbirth, but the acceptance of these methods is currently low.

Linking physiology and demographics, non-ocular pathology and pharmaceutical drug use to standard OCT measures of the inner retina: The PPP project.

PURPOSE: To assess the associations between physiology and demographics, non-ocular pathology and pharmaceutical drug use against peri-papillary retinal nerve fibre layer thickness (pRNFL T) and other optical coherence tomography (OCT) inner retinal measures in normal, healthy eyes. METHODS: A retrospective, cross-sectional study of 705 consecutive participants with bilateral normal, healthy optic nerves and maculae. PRNFL Ts, vertical cup/disc ratio (CDR), cup volume and macular ganglion cell layer-inner plexiform layer (GCL-IPL) Ts were extracted from Cirrus OCT scans, then regressed against predictor variables of participants' physiology and demographics (eye laterality, refraction, intraocular pressure [IOP], age, sex, race/ethnicity, etc.) and non-ocular pathology and pharmaceutical drug use according to the World Health Organisation classifications. Associations were assessed for statistical significance (p < 0.05) and clinical significance (|ß| > 95% limits of agreement for repeated measures). RESULTS: A multitude of non-ocular pathology and pharmaceutical drug use were statistically and clinically significantly associated with deviations in standard OCT inner retinal measures, exceeding the magnitude of other factors such as age, IOP and race/ethnicity. Thinner inner retina and larger optic nerve cup measures were linked to use of systemic corticosteroids, sex hormones/modulators, presence of vasomotor/allergic rhinitis and other diseases and drugs (up to -29.3 [-49.88, -8.72] µm pRNFL T, 0.31 [0.07, 0.54] vertical CDR, 0.29 [0.03, 0.54] mm3 cup volume and -10.18 [-16.62, -3.74] µm macular GCL-IPL T; all p < 0.05). Thicker inner retina and smaller optic nerve cup measures were diffusely associated with use of antineoplastic agents, presence of liver or urinary diseases and other diseases and drugs (up to 67.12 [64.92, 69.31] µm pRNFL T, -0.31 [-0.53, -0.09] vertical CDR, -0.06 [-0.11, 0] mm3 cup volume and 28.84 [14.51, 43.17] µm macular GCL-IPL T; all p < 0.05). CONCLUSION: There are a multitude of systemic diseases and drugs associated with altered OCT inner retinal measures, with magnitudes far exceeding those of other factors such as age, IOP and race/ethnicity. These systemic factors should at least be considered during OCT assessments to ensure precise interpretation of normal versus pathological inner retinal health.

Accuracy assessment of patient safety incident (PSI) codes and present-on-admission (POA) indicators: a cross-sectional analysis using the Patient Safety Incidents Inquiry (PSII) in Korea.

BACKGROUND: Among the various methods used, administrative data collected for claims and billing purposes, such as diagnosis codes and present-on-admission (POA) indicators, can easily be employed to assess patient safety status. However, it is crucial that administrative data be accurate to generate valid estimates of adverse event (AE) occurrence. Thus, we aimed to evaluate the accuracy of diagnosis codes and POA indicators in patients with confirmed AEs in the hospital admission setting. METHODS: We analysed the diagnosis codes of 1,032 confirmed AE cases and 6,754 non-AE cases from the 2019 Patient Safety Incidents Inquiry, which was designed as a cross-sectional study, to determine their alignment with the Korean Patient Safety Incidents (PSIs) Code Classification System. The unit of analysis was the individual case rather than the patient, because two or more AEs may occur in one patient. We examined whether the primary and secondary diagnostic codes had PSIs codes matching the AE type and checked each PSI code for whether the POA indicator had an 'N' tag. We reviewed the presence of PSI codes in patients without identified AEs and calculated the correlation between the AE incidence rate and PSI code and POA indicator accuracy across 15 hospitals. RESULTS: Ninety (8.7%) of the AE cases had PSI codes with an 'N' tag on the POA indicator compared to 294 (4.4%) of the non-AE cases. Infection- (20.4%) and surgery/procedure-related AEs (13.6%) had relatively higher instances of correctly tagged PSI codes. We did not identify any PSI codes for diagnosis-related incidents. While we noted significant differences in AE incidence rates, PSI code accuracy, and POA indicator accuracy among the hospitals, the correlations between these variables were not statistically significant. CONCLUSION: Currently, PSI codes and POA indicators in South Korea appear to have low validity. To use administrative data in medical quality improvement activities such as monitoring patient safety levels, improving the accuracy of administrative data should be a priority. Possible strategies include targeted education on PSI codes and POA indicators and introduction of new evaluation indicators regarding the accuracy of administrative data.

Evaluating the coding accuracy of type 2 diabetes mellitus among patients with non-alcoholic fatty liver disease.

BACKGROUND: Non-alcoholic fatty liver disease (NAFLD) describes a spectrum of chronic fattening of liver that can lead to fibrosis and cirrhosis. Diabetes has been identified as a major comorbidity that contributes to NAFLD progression. Health systems around the world make use of administrative data to conduct population-based prevalence studies. To that end, we sought to assess the accuracy of diabetes International Classification of Diseases (ICD) coding in administrative databases among a cohort of confirmed NAFLD patients in Calgary, Alberta, Canada. METHODS: The Calgary NAFLD Pathway Database was linked to the following databases: Physician Claims, Discharge Abstract Database, National Ambulatory Care Reporting System, Pharmaceutical Information Network database, Laboratory, and Electronic Medical Records. Hemoglobin A1c and diabetes medication details were used to classify diabetes groups into absent, prediabetes, meeting glycemic targets, and not meeting glycemic targets. The performance of ICD codes among these groups was compared to this standard. Within each group, the total numbers of true positives, false positives, false negatives, and true negatives were calculated. Descriptive statistics and bivariate analysis were conducted on identified covariates, including demographics and types of interacted physicians. RESULTS: A total of 12,012 NAFLD patients were registered through the Calgary NAFLD Pathway Database and 100% were successfully linked to the administrative databases. Overall, diabetes coding showed a sensitivity of 0.81 and a positive predictive value of 0.87. False negative rates in the absent and not meeting glycemic control groups were 4.5% and 6.4%, respectively, whereas the meeting glycemic control group had a 42.2% coding error. Visits to primary and outpatient services were associated with most encounters. CONCLUSION: Diabetes ICD coding in administrative databases can accurately detect true diabetic cases. However, patients with diabetes who meets glycemic control targets are less likely to be coded in administrative databases. A detailed understanding of the clinical context will require additional data linkage from primary care settings.

Using a severity threshold to improve occupational injury surveillance: Assessment of a severe traumatic injury-based occupational health indicator across the International Classification of Diseases lexicon transition.

BACKGROUND: Traumatic injury is a leading cause of death and disability among US workers. Severe injuries are less subject to systematic ascertainment bias related to factors such as reporting barriers, inpatient admission criteria, and workers' compensation coverage. A state-based occupational health indicator (OHI #22) was initiated in 2012 to track work-related severe traumatic injury hospitalizations. After 2015, OHI #22 was reformulated to account for the transition from the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) to ICD-10-CM. This study describes rates and trends in OHI #22, alongside corresponding metrics for all work-related hospitalizations. METHODS: Seventeen states used hospital discharge data to calculate estimates for calendar years 2012-2019. State-panel fixed-effects regression was used to model linear trends in annual work-related hospitalization rates, OHI #22 rates, and the proportion of work-related hospitalizations resulting from severe injuries. Models included calendar year and pre- to post-ICD-10-CM transition. RESULTS: Work-related hospitalization rates showed a decreasing monotonic trend, with no significant change associated with the ICD-10-CM transition. In contrast, OHI #22 rates showed a monotonic increasing trend from 2012 to 2014, then a significant 50% drop, returning to a near-monotonic increasing trend from 2016 to 2019. On average, OHI #22 accounted for 12.9% of work-related hospitalizations before the ICD-10-CM transition, versus 9.1% post-transition. CONCLUSIONS: Although hospital discharge data suggest decreasing work-related hospitalizations over time, work-related severe traumatic injury hospitalizations are apparently increasing. OHI #22 contributes meaningfully to state occupational health surveillance efforts by reducing the impact of factors that differentially obscure minor injuries; however, OHI #22 trend estimates must account for the ICD-10-CM transition-associated structural break in 2015.

Evaluating a Natural Language Processing-Driven, AI-Assisted International Classification of Diseases, 10th Revision, Clinical Modification, Coding System for Diagnosis Related Groups in a Real Hospital Environment: Algorithm Development and Validation Study.

BACKGROUND: International Classification of Diseases codes are widely used to describe diagnosis information, but manual coding relies heavily on human interpretation, which can be expensive, time consuming, and prone to errors. With the transition from the International Classification of Diseases, Ninth Revision, to the International Classification of Diseases, Tenth Revision (ICD-10), the coding process has become more complex, highlighting the need for automated approaches to enhance coding efficiency and accuracy. Inaccurate coding can result in substantial financial losses for hospitals, and a precise assessment of outcomes generated by a natural language processing (NLP)-driven autocoding system thus assumes a critical role in safeguarding the accuracy of the Taiwan diagnosis related groups (Tw-DRGs). OBJECTIVE: This study aims to evaluate the feasibility of applying an International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM), autocoding system that can automatically determine diagnoses and codes based on free-text discharge summaries to facilitate the assessment of Tw-DRGs, specifically principal diagnosis and major diagnostic categories (MDCs). METHODS: By using the patient discharge summaries from Kaohsiung Medical University Chung-Ho Memorial Hospital (KMUCHH) from April 2019 to December 2020 as a reference data set we developed artificial intelligence (AI)-assisted ICD-10-CM coding systems based on deep learning models. We constructed a web-based user interface for the AI-assisted coding system and deployed the system to the workflow of the certified coding specialists (CCSs) of KMUCHH. The data used for the assessment of Tw-DRGs were manually curated by a CCS with the principal diagnosis and MDC was determined from discharge summaries collected at KMUCHH from February 2023 to April 2023. RESULTS: Both the reference data set and real hospital data were used to assess performance in determining ICD-10-CM coding, principal diagnosis, and MDC for Tw-DRGs. Among all methods, the GPT-2 (OpenAI)-based model achieved the highest F1-score, 0.667 (F1-score 0.851 for the top 50 codes), on the KMUCHH test set and a slightly lower F1-score, 0.621, in real hospital data. Cohen κ evaluation for the agreement of MDC between the models and the CCS revealed that the overall average κ value for GPT-2 (κ=0.714) was approximately 12.2 percentage points higher than that of the hierarchy attention network (κ=0.592). GPT-2 demonstrated superior agreement with the CCS across 6 categories of MDC, with an average κ value of approximately 0.869 (SD 0.033), underscoring the effectiveness of the developed AI-assisted coding system in supporting the work of CCSs. CONCLUSIONS: An NLP-driven AI-assisted coding system can assist CCSs in ICD-10-CM coding by offering coding references via a user interface, demonstrating the potential to reduce the manual workload and expedite Tw-DRG assessment. Consistency in performance affirmed the effectiveness of the system in supporting CCSs in ICD-10-CM coding and the judgment of Tw-DRGs.

Coding rules for uncertain and "ruled out" diagnoses in ICD-10 and ICD-11.

The International Classification of Diseases, 11th Revision (ICD-11) has significantly improved the ability to navigate coding challenges beyond prior iterations of the ICD. Commonly encountered sources of complexity in clinical documentation include coding of uncertain and "ruled out" diagnoses. Assessing official international guidelines and rules, this paper documents extensive variation across countries in existing practices for coding and reporting unconfirmed and "ruled out" clinical concepts in ICD-10 (and modifications thereof). The design of ICD-11 is intended to mitigate these coding challenges by introducing postcoordination, expanding the range of codable clinical concepts, and offering clearer guidance in the ICD-11 Reference Guide. ICD-11 offers substantial progress towards more precise capture of uncertain and "ruled out" diagnoses, including international consensus on coding rules for these historically challenging clinical concepts. However, we identify the need for further clarification of the concepts of "provisional diagnosis" and "differential diagnosis."

A retrospective analysis using comorbidity detecting algorithmic software to determine the incidence of International Classification of Diseases (ICD) code omissions and appropriateness of Diagnosis-Related Group (DRG) code modifiers.

BACKGROUND: The mechanism for recording International Classification of Diseases (ICD) and diagnosis related groups (DRG) codes in a patient's chart is through a certified medical coder who manually reviews the medical record at the completion of an admission. High-acuity ICD codes justify DRG modifiers, indicating the need for escalated hospital resources. In this manuscript, we demonstrate that value of rules-based computer algorithms that audit for omission of administrative codes and quantifying the downstream effects with regard to financial impacts and demographic findings did not indicate significant disparities. METHODS: All study data were acquired via the UCLA Department of Anesthesiology and Perioperative Medicine's Perioperative Data Warehouse. The DataMart is a structured reporting schema that contains all the relevant clinical data entered into the EPIC (EPIC Systems, Verona, WI) electronic health record. Computer algorithms were created for eighteen disease states that met criteria for DRG modifiers. Each algorithm was run against all hospital admissions with completed billing from 2019. The algorithms scanned for the existence of disease, appropriate ICD coding, and DRG modifier appropriateness. Secondarily, the potential financial impact of ICD omissions was estimated by payor class and an analysis of ICD miscoding was done by ethnicity, sex, age, and financial class. RESULTS: Data from 34,104 hospital admissions were analyzed from January 1, 2019, to December 31, 2019. 11,520 (32.9%) hospital admissions were algorithm positive for a disease state with no corresponding ICD code. 1,990 (5.8%) admissions were potentially eligible for DRG modification/upgrade with an estimated lost revenue of $22,680,584.50. ICD code omission rates compared against reference groups (private payors, Caucasians, middle-aged patients) demonstrated significant p-values < 0.05; similarly significant p-value where demonstrated when comparing patients of opposite sexes. CONCLUSIONS: We successfully used rules-based algorithms and raw structured EHR data to identify omitted ICD codes from inpatient medical record claims. These missing ICD codes often had downstream effects such as inaccurate DRG modifiers and missed reimbursement. Embedding augmented intelligence into this problematic workflow has the potential for improvements in administrative data, but more importantly, improvements in administrative data accuracy and financial outcomes.

Evaluating MedDRA-to-ICD terminology mappings.

BACKGROUND: In this era of big data, data harmonization is an important step to ensure reproducible, scalable, and collaborative research. Thus, terminology mapping is a necessary step to harmonize heterogeneous data. Take the Medical Dictionary for Regulatory Activities (MedDRA) and International Classification of Diseases (ICD) for example, the mapping between them is essential for drug safety and pharmacovigilance research. Our main objective is to provide a quantitative and qualitative analysis of the mapping status between MedDRA and ICD. We focus on evaluating the current mapping status between MedDRA and ICD through the Unified Medical Language System (UMLS) and Observational Medical Outcomes Partnership Common Data Model (OMOP CDM). We summarized the current mapping statistics and evaluated the quality of the current MedDRA-ICD mapping; for unmapped terms, we used our self-developed algorithm to rank the best possible mapping candidates for additional mapping coverage. RESULTS: The identified MedDRA-ICD mapped pairs cover 27.23% of the overall MedDRA preferred terms (PT). The systematic quality analysis demonstrated that, among the mapped pairs provided by UMLS, only 51.44% are considered an exact match. For the 2400 sampled unmapped terms, 56 of the 2400 MedDRA Preferred Terms (PT) could have exact match terms from ICD. CONCLUSION: Some of the mapped pairs between MedDRA and ICD are not exact matches due to differences in granularity and focus. For 72% of the unmapped PT terms, the identified exact match pairs illustrate the possibility of identifying additional mapped pairs. Referring to its own mapping standard, some of the unmapped terms should qualify for the expansion of MedDRA to ICD mapping in UMLS.

Optimization of the allergen classification of the International Classification Of Diseases, 11th Revision (ICD-11).

BACKGROUND: Accurate diagnosis of triggers or causative allergens is essential for appropriate risk assessment, providing correct advice to patients with allergy and their caregivers and personalized treatment. However, allergens have never been represented in the World Health Organization International Classification of Diseases (ICD). OBJECTIVE: In this article, we present the process of selection of allergens to better fit the ICD, 11th Revision (ICD-11) structure and the outcomes of this process. METHODS: The Logical Observation Identifiers Names and Codes database, containing 1444 allergens, was used as the basis for the selection process. Two independent experts were responsible for the first selection of the allergens according to specific technical criteria. The second step of the selection process was based on real-life relevance of the allergens according to the frequency of requests regarding each of them. RESULTS: We selected 1109 allergens (76.8%) from all 1444 present in the Logical Observation Identifiers Names and Codes database, with considerable agreement between experts (Cohen κ = 8.6). After assessment of real-life data, 297 additional relevant allergens worldwide were selected and grouped as plants (36.4%), drugs (32.6%), animal proteins (21%), mold and other microorganisms (1.5%), occupational allergens (0.4%), and miscellaneous allergens (0.5%). CONCLUSION: The stepwise approach allowed us to select the most relevant allergens in practice, which is the first step to building a classification of allergens for the WHO ICD-11. Aligned with the achievement in the construction of the pioneer section addressed to the allergic and hypersensitivity conditions in the ICD-11, the introduction of a classification for allergens can be considered timely and much needed in clinical practice.

Association Between Hospital-Documented Atrial Fibrillation and Central Retinal Artery Occlusion.

BACKGROUND: Carotid stenosis is thought to be the primary risk factor for central retinal artery occlusion (CRAO); however, it is not known whether atrial fibrillation (AF)-a cardiac arrhythmia that underlies over 25% of cerebral ischemic strokes-predisposes patients to CRAO. METHODS: A retrospective, observational, cohort study was performed using data from the State Inpatient Databases and State Emergency Department Databases from New York (2006-2015), California (2003-2011), and Florida (2005-2015) to determine the association between AF and CRAO. The primary exposure was hospital-documented AF. The primary end point was hospital-documented CRAO, defined as having an International Classification of Diseases, Ninth Revision, Clinical Modification, code 362.31 in the primary diagnosis position. Cause-specific hazard models were used to model CRAO-free survival among patients according to hospital-documented AF status. RESULTS: Of 39 834 885 patients included in the study, 2 723 842 (median age, 72.7 years; 48.5% women) had AF documented during the exposure window. The median follow-up duration was 6 years and 1 month. Patients with AF were older, more likely to be of non-Hispanic White race/ethnicity, and had a higher burden of cardiovascular comorbidities compared with patients without AF. The cumulative incidence of CRAO determined prospectively after exclusions was 8.69 per 100 000 at risk in those with AF and 2.39 per 100 000 at risk in those without AF over the study period. Before adjustment, AF was associated with higher risk of CRAO (hazard ratio, 2.55 [95% CI, 2.15-3.03]). However, after adjustment for demographics, state, and cardiovascular comorbidities, there was an inverse association between AF and risk of CRAO (adjusted hazard ratio, 0.72 [95% CI, 0.60-0.87]). These findings were robust in our prespecified sensitivity analyses. By contrast, positive control outcomes of embolic and ischemic stroke showed an expected strong relationship between AF and risk of stroke. CONCLUSIONS: We found an inverse association between AF and CRAO in a large, representative study of hospitalized patients; however, this cohort did not ascertain AF or CRAO occurring outside of hospital or emergency department settings.

Increased Hospitalizations Involving Fungal Infections during COVID-19 Pandemic, United States, January 2020-December 2021.

Hospitalizations involving fungal infections increased 8.5% each year in the United States during 2019-2021. During 2020-2021, patients hospitalized with COVID-19-associated fungal infections had higher (48.5%) in-hospital mortality rates than those with non-COVID-19-associated fungal infections (12.3%). Improved fungal disease surveillance is needed, particularly during respiratory virus pandemics.