Registration of Amyotrophic Lateral Sclerosis: Validity in the Danish National Patient Registry.

Purpose: Health care databases are a valuable source for epidemiological research on amyotrophic lateral sclerosis (ALS) if diagnosis codes are valid. We evaluated the validity of the diagnostic codes for ALS in the Danish National Patient Registry (DNPR). Patients and Methods: We obtained data from the DNPR for all adult (>17 years) patients registered with ALS in Denmark between 1987 and 2022 (median population of 4.2 million during the study period). We randomly selected adult patients living in the North Denmark Region and Central Denmark Region (median population 1.4 million), with a primary discharge diagnosis code of ALS, diagnosed at three departments of neurology. We retrieved and reviewed medical records and estimated the positive predictive value (PPV) of the ALS diagnosis. Results: Over 36 years, we identified 5679 patients. From the validation cohort of 300 patients, we were able to retrieve 240 (80%) medical records, and 215 ALS diagnoses were confirmed. The overall positive predictive value was 89.6% (95% confidence interval (CI): 85.1-92.8). The highest PPV was achieved for diagnoses registered for patients aged ≥70 years (93.8; 95% CI: 86.2-97.3) compared to patients <60 years (83.4; 95% CI: 73.3-90.7). Conclusion: We found a high PPV of primary diagnostic codes for ALS from Danish departments of neurology, demonstrating high validity. Thus, the DNPR is a well-suited data source for large-scale epidemiological research on ALS.

Using natural language processing for automated classification of disease and to identify misclassified ICD codes in cardiac disease.

Aims: ICD codes are used for classification of hospitalizations. The codes are used for administrative, financial, and research purposes. It is known, however, that errors occur. Natural language processing (NLP) offers promising solutions for optimizing the process. To investigate methods for automatic classification of disease in unstructured medical records using NLP and to compare these to conventional ICD coding. Methods and results: Two datasets were used: the open-source Medical Information Mart for Intensive Care (MIMIC)-III dataset (n = 55.177) and a dataset from a hospital in Belgium (n = 12.706). Automated searches using NLP algorithms were performed for the diagnoses 'atrial fibrillation (AF)' and 'heart failure (HF)'. Four methods were used: rule-based search, logistic regression, term frequency-inverse document frequency (TF-IDF), Extreme Gradient Boosting (XGBoost), and Bio-Bidirectional Encoder Representations from Transformers (BioBERT). All algorithms were developed on the MIMIC-III dataset. The best performing algorithm was then deployed on the Belgian dataset. After preprocessing a total of 1438 reports was retained in the Belgian dataset. XGBoost on TF-IDF matrix resulted in an accuracy of 0.94 and 0.92 for AF and HF, respectively. There were 211 mismatches between algorithm and ICD codes. One hundred and three were due to a difference in data availability or differing definitions. In the remaining 108 mismatches, 70% were due to incorrect labelling by the algorithm and 30% were due to erroneous ICD coding (2% of total hospitalizations). Conclusion: A newly developed NLP algorithm attained a high accuracy for classifying disease in medical records. XGBoost outperformed the deep learning technique BioBERT. NLP algorithms could be used to identify ICD-coding errors and optimize and support the ICD-coding process.

Maternal Mortality and COVID-19 Pandemic: Looking Beyond SARS CoV-2 Infection.

Objective: To study the impact of COVID-19 pandemic on maternal mortality ratio, aetiological and modifiable factors for maternal mortality and key interventions performed. Method: Retrospective exploratory study evaluating maternal mortality between April to November 2020 (study group) and 2019 (control group). Results: Demographic variations existed in the two groups. Increased maternal age and illiteracy were significantly more in the study group. Maternal mortality ratio (MMR) was significantly high in the study group (792 vs. 296 p value = 0.0). Hemorrhage accounted for 20% and COVID-19-related maternal deaths accounted for 15% deaths in the study group. Level 3 delay (delay in receiving care/inadequate care) was observed in 35% in the study group and 28% in control group (p value = 0.349). 17.5% of mothers in the study group as compared to 8% of control group were dead on arrival to hospital though not statistically significant (p value = 0.28). Significantly more women in study group died within 24 h of admission (45% vs. 20%, p value 0.04). Among the key interventions, the use of supplemental oxygen was significantly high in study group (p value = 0.02). Conclusion: Maternal mortality ratio was high in the pandemic year because of a significant decline in hospital delivery rate. The lesson learnt from this pandemic needs to be documented to guide better planning in the future to face similar situations.

Deaths caused by osteoporotic fractures in Japan: An epidemiological study.

BACKGROUND: Osteoporosis is a global issue with a worldwide prevalence of 18.3%, and the presence of coexisting fragility fractures can reduce the survival rate by approximately 20%. In Japan, the prevalence of osteoporosis is estimated to be 12.8 million, and the annual occurrence of hip fractures is approximately 193,400. Remarkably, coexisting hip or spinal fragility fractures caused by slight external force meet the Japanese diagnostic criterion for osteoporosis regardless of bone mineral density. However, only 191 deaths due to osteoporosis were published in 2021 in Japan. With the concern that some cases of hip and spinal fragility fractures were assigned an underlying cause of death of traumatic fracture instead of osteoporosis, this study aimed to elucidate the actual number of deaths due to osteoporosis in Japan. METHODS: We used the data from Japan in 2018. First, the number of deaths due to osteoporosis and hip or spinal fractures was reviewed using published vital statistics. Second, we calculated the number of elderly deaths (age ≥80 years) resulting from hip or spinal fractures caused by falls on the same level using data from approximately 1.4 million annual individual death certificates. Combining the above data, the actual number of deaths due to osteoporosis was estimated. RESULTS: Only 190 deaths due to osteoporosis were reported in the published data. The individual certificate data revealed 3437 elderly deaths due to hip or spinal fractures caused by falls on the same level, which could meet the criteria of osteoporotic fragility fractures. Accordingly, the estimated number of deaths caused by osteoporosis was calculated as 3,627, approximately 19 times the published value. CONCLUSIONS: After researching the individual death certificate data focusing on the coexisting hip or spinal fragility fracture, it was implied that osteoporosis may have a higher mortality rate in Japan than what is published.

An Artificial Intelligence Model for Predicting Trauma Mortality Among Emergency Department Patients in South Korea: Retrospective Cohort Study.

BACKGROUND: Within the trauma system, the emergency department (ED) is the hospital's first contact and is vital for allocating medical resources. However, there is generally limited information about patients that die in the ED. OBJECTIVE: The aim of this study was to develop an artificial intelligence (AI) model to predict trauma mortality and analyze pertinent mortality factors for all patients visiting the ED. METHODS: We used the Korean National Emergency Department Information System (NEDIS) data set (N=6,536,306), incorporating over 400 hospitals between 2016 and 2019. We included the International Classification of Disease 10th Revision (ICD-10) codes and chose the following input features to predict ED patient mortality: age, sex, intentionality, injury, emergent symptom, Alert/Verbal/Painful/Unresponsive (AVPU) scale, Korean Triage and Acuity Scale (KTAS), and vital signs. We compared three different feature set performances for AI input: all features (n=921), ICD-10 features (n=878), and features excluding ICD-10 codes (n=43). We devised various machine learning models with an ensemble approach via 5-fold cross-validation and compared the performance of each model with that of traditional prediction models. Lastly, we investigated explainable AI feature effects and deployed our final AI model on a public website, providing access to our mortality prediction results among patients visiting the ED. RESULTS: Our proposed AI model with the all-feature set achieved the highest area under the receiver operating characteristic curve (AUROC) of 0.9974 (adaptive boosting [AdaBoost], AdaBoost + light gradient boosting machine [LightGBM]: Ensemble), outperforming other state-of-the-art machine learning and traditional prediction models, including extreme gradient boosting (AUROC=0.9972), LightGBM (AUROC=0.9973), ICD-based injury severity scores (AUC=0.9328 for the inclusive model and AUROC=0.9567 for the exclusive model), and KTAS (AUROC=0.9405). In addition, our proposed AI model outperformed a cutting-edge AI model designed for in-hospital mortality prediction (AUROC=0.7675) for all ED visitors. From the AI model, we also discovered that age and unresponsiveness (coma) were the top two mortality predictors among patients visiting the ED, followed by oxygen saturation, multiple rib fractures (ICD-10 code S224), painful response (stupor, semicoma), and lumbar vertebra fracture (ICD-10 code S320). CONCLUSIONS: Our proposed AI model exhibits remarkable accuracy in predicting ED mortality. Including the necessity for external validation, a large nationwide data set would provide a more accurate model and minimize overfitting. We anticipate that our AI-based risk calculator tool will substantially aid health care providers, particularly regarding triage and early diagnosis for trauma patients.

Utilizing International Classification of Diseases Codes to Identify Shoulder Dystocia and Neonatal Brachial Plexus Injury.

BACKGROUND: The utilization of International Classification of Diseases, Ninth or Tenth Revision, (ICD-9/10) coding to identify the incidence of disease is frequently performed in medical research. This study attempts to assess the validity of using ICD-9/10 codes to identify patients with shoulder dystocia (SD) with concurrent neonatal brachial plexus palsy (NBPP). METHODS: This retrospective cohort study examined patients evaluated at the University of Michigan Brachial Plexus and Peripheral Nerve Program (UM-BP/PN) from 2004 to 2018. We reported the percentage of patients with reported NBPP ICD-9/10 and SD ICD-9/10 discharged at birth who were later diagnosed with NBPP by a specialty clinic by interdisciplinary faculty and staff utilizing physical evaluations and ancillary testing such as such as electrodiagnostics and imaging. The relationship of reported NBPP ICD-9/10, SD ICD-9/10, extent of NBPP nerve involvement, and NBPP persistence at age two years were examined via chi-square or Fischer exact test. RESULTS: Of the 51 mother-infant dyads with complete birth discharge records evaluated at the UM-BP/PN, 26 (51%) were discharged without an ICD-9/10 code documenting NBPP; of these 26 patients, only four had ICD-9/10 documentation of SD at discharge, which left 22 patients with no ICD-9/10 code documentation of either SD or NBPP (43%). Patients with pan-plexopathy were more likely to be discharged with an NBBP ICD-9/10 code than those infants with upper nerve involvement (77% vs 39%, P < 0.02). CONCLUSION: Use of ICD-9/10 codes for the identification of NBPP appears to undercount the true incidence. This underestimation is more pronounced for milder forms of NBPP.

Risk of Duplicate ICD Codes for Orthopedic and Injury Related Research.

The World Health Organization's International Classification of Diseases (ICD) has become the international standard diagnostic classification for reporting morbidity and mortality. In 2015, the United States transitioned from the 9th to 10th Revision. The update was necessary due to major structural limitations of the ICD-9 system. Concerns of the transition mainly centered around clinical usage and cost; however, there were concerns for overlapping codes with the same classification but different meanings between the two versions. Duplicate codes could pose an issue for big data retrospective studies that overlap between the two systems. Therefore, the goals of this study are to further explore and identify duplicate ICD codes between the systems. ICD-9-CM and ICD-10-CM code files were obtained from the Centers for Medicare & Medicaid Services. There were 14,567 ICD-9-CM codes and 91,737 unique ICD-10-CM codes tabulated. Duplicated items between the files were isolated. Four hundred sixty-nine duplicate codes were identified, consisting of 39 E Codes and 430 V Codes. These twin codes contain classifications for external causes of injury and factors influencing health status and contact with health services. Therefore, special attention should be drawn to retrospective research involving methods of injury spanning ICD-9 and ICD-10 systems.

Utilidad de los diagnósticos alertantes CIE-10 para identificar acontecimientos adversos por los medicamentos en los servicios de urgencias / Usefulness of ICD-10 diagnosis triggers to identify adverse drug events in emergency care

Objetivos: Evaluar la utilidad de una herramienta basada en los códigos diagnósticos CIE-10 para identificar a los pacientes que consultan a un servicio de urgencias por acontecimientos adversos por medicamentos (AAM). Métodos: Estudio observacional prospectivo, en el cual se incluyeron los pacientes que acudieron a un servicio de urgencias durante el periodo de mayo-agosto de 2022 con un diagnóstico codificado con alguno de los 27 diagnósticos CIE-10 establecidos como alertantes para el estudio. La confirmación de la presencia de AAM a partir de dichos diagnósticos se realizó analizando los fármacos prescritos previamente al ingreso, a través de un debate entre expertos y mediante una entrevista telefónica con los pacientes. Resultados: Se evaluaron 1.143 pacientes con diagnósticos alertantes, de los cuales 310 (27,1%) correspondieron a pacientes cuya consulta se atribuyó a un AAM. El 58,4% de los AAM se detectaron mediante 3 códigos diagnósticos: K59.0-Estreñimiento (n = 87; 28,1%), I16.9-Crisis hipertensiva (n = 72; 23,2%) e I95.1-Hipotensión ortostática (n = 22; 7,1%). Los códigos diagnósticos con mayor grado de asociación con AAM fueron: E16.2-Hipoglucemia no especificada (73,7%) y E11.65-Diabetes mellitus tipo 2 con hiperglucemia (71,4%), mientras que los diagnósticos D62-Anemia poshemorrágica aguda e I74.3-Embolia y trombosis de arterias de los miembros inferiores no identificaron ningún AAM. Conclusiones: Los códigos CIE-10 asociados a diagnósticos alertantes son una herramienta de utilidad para identificar a los pacientes que consultan los servicios de urgencias por AAM y podrían ser utilizados para abordar las intervenciones de prevención secundaria dirigidas a evitar nuevas consultas al sistema sanitario. (AU)

Objectives: To assess the usefulness of a tool based on ICD-10 diagnostic codes to identify patients who consult an emergency department for adverse drug events (ADE). Methods: Prospective observational study, in which patients discharged from an emergency department during May to August 2022 with a diagnosis coded with one of the 27 ICD-10 diagnoses considered as triggers were included. ADE confirmation was carried out by analyzing drugs prescribed prior to admission, and through a discussion among experts and a phone interview with patients after hospital discharge. Results: 1,143 patients with trigger diagnoses were evaluated, of which 310 (27.1%) corresponded to patients whose emergency visit was attributed to an ADE. A 58.4% of ADE consultations were found with three diagnostic codes: K59.0-Constipation (n = 87; 28.1%), I16.9-Hypertensive Crisis (n = 72; 23.2%) and I95.1-Orthostatic hypotension (n = 22; 7.1%). The diagnoses with the highest degree of association with consultations attributed to ADE were E16.2-Hypoglycemia, unspecified (73.7%) and E11.65-Type 2 diabetes mellitus with hyperglycemia (71.4%), while diagnoses D62-Acute posthemorrhagic anemia and I74.3-Embolism and thrombosis of arteries of the lower limbs were not attributed to any case of ADE. Conclusions: The ICD-10 codes associated with trigger diagnoses are a useful tool to identify patients who consult the emergency services with ADE and could be used to apply secondary prevention programs to avoid new consultations to the health care system. (AU)

Neuropsychiatric disorders in adults undergoing chimeric antigen receptor T cells therapy for aggressive lymphomas and acute lymphoblastic leukemia.

Objective: To evaluate risk factors for neuropsychiatric disorders (NPD) in recipients of CART therapy. Methods: Patients ≥ 18 years with acute lymphoblastic leukemia (ALL), and aggressive B-cell lymphomas who received CART in 2018 were evaluated. Patients with and without NPD were compared. Results: NPD was diagnosed in 31.2% of patients. Compared to patients without NPD, patients with NPD were likely to be females (P = 0.035) and have ALL (P = 0.039). NPD was significantly associated with female gender (OR = 2.03) and diagnosis of ALL (OR = 2.76). No association between NPD and outcomes. Conclusions: Female gender and ALL were risk factors for NPD.

[Translated article] Usefulness of ICD-10 diagnostic triggers to identify adverse drug events in emergency care.

OBJECTIVES: To assess the usefulness of a tool based on ICD-10 diagnostic codes to identify patients who consult an emergency department for adverse drug events (ADE). METHODS: Prospective observational study, in which patients discharged from an emergency department during May to August 2022 with a diagnosis coded with one of the 27 ICD-10 diagnoses considered as triggers were included. ADE confirmation was carried out by analyzing drugs prescribed prior to admission, and through a discussion among experts and a phone interview with patients after hospital discharge. RESULTS: 1143 patients with trigger diagnoses were evaluated, of which 310 (27.1%) corresponded to patients whose emergency visit was attributed to an ADE. A 58.4% of ADE consultations were found with three diagnostic codes: K59.0-Constipation (n = 87; 28.1%), I16.9-Hypertensive Crisis (n = 72; 23.2%) and I95.1-Orthostatic hypotension (n = 22; 7.1%). The diagnoses with the highest degree of association with consultations attributed to ADE were E16.2-Hypoglycemia, unspecified (73.7%) and E11.65-Type 2 diabetes mellitus with hyperglycemia (71.4%), while diagnoses D62-Acute posthemorrhagic anemia and I74.3-Embolism and thrombosis of arteries of the lower limbs were not attributed to any case of ADE. CONCLUSIONS: The ICD-10 codes associated with trigger diagnoses are a useful tool to identify patients who consult the emergency services with ADE and could be used to apply secondary prevention programs to avoid new consultations to the health care system.

Risk factors and drugs that trigger the onset of Stevens-Johnson syndrome and toxic epidermal necrolysis: A population-based cohort study using the Shizuoka Kokuho database.

Background: Evidence of factors associated with Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) from population-based studies is scarce. Objective: We aimed to identify the incidence, risk factors, and drugs that trigger the development of SJS/TEN in the general population. Methods: A regional, population-based, longitudinal cohort with 2,398,393 Japanese individuals was analyzed using the Shizuoka Kokuho Database from 2012 to 2020. Results: Among 1,909,570 individuals, 223 (0.01%, 2.3 cases/100,000 person-years) patients were diagnosed with SJS/TEN during the observational period of a maximum of 7.5 years. In a multivariable analysis, the risks of SJS/TEN were an older age, and the presence of type 2 diabetes, peripheral vascular disease, and systemic autoimmune diseases. The administration of drugs, such as immune checkpoint inhibitors, insulin, and type 2 diabetes agents, triggered the onset of SJS/TEN. Limitations: The results may apply only to the Japanese population. Conclusion: In this cohort population from a database representing the general population, the risks of developing SJS/TEN were old age and a history of type 2 diabetes, peripheral vascular disease, and systemic autoimmune disease. Furthermore, in addition to previously reported drugs, the administration of immune checkpoint inhibitors, insulin, and type 2 diabetes agents, may trigger the development of SJS/TEN.

Dearth of ICD Codes for Complications of Immune Checkpoint Inhibitors Impedes Clinical Care and Research.

Immune checkpoint inhibitors (ICIs) are a rapidly expanding class of targeted therapies effective in the treatment of various cancers. However, while efficacious, ICIs have been associated with treatment complications, namely immune-related adverse events (irAEs). IrAEs of the endocrine system are among the most commonly reported irAEs, but despite their high incidence, standardized disease definitions and endocrine IrAE-specific International Classification of Diseases (ICD) codes remain lacking. This dearth of standardized nomenclature and ICD codes has in many ways impeded both the clinical care of patients and the progress of endocrine irAE-related research. ICD codes are used internationally and are essential for medical claims reporting in the health care setting, and they provide a universal language system for recording, reporting, and monitoring diseases. These codes are also a well-accepted form of electronic health record data capture that facilitates the collection, storage, and sharing of data. Therefore, the lack of standardized disease definitions and ICD codes has been associated with misclassification and suboptimal management of individuals with endocrine irAEs and has also been associated with reduced data availability, comparability, and quality. Harmonized and clinically relevant disease definitions along with the subsequent development of endocrine-irAE-specific ICD codes will provide a systematic approach to understanding the spectrum and burden of endocrine irAE diseases, and will have a positive effect across clinical, public health, and research settings.

Usefulness of ICD-10 diagnosis triggers to identify adverse drug events in emergency care. / Utilidad de los diagnósticos alertantes CIE-10 para identificar acontecimientos adversos por los medicamentos en los servicios de urgencias.

OBJECTIVES: To assess the usefulness of a tool based on ICD-10 diagnostic codes to identify patients who consult an emergency department for adverse drug events (ADE). METHODS: Prospective observational study, in which patients discharged from an emergency department during May to August 2022 with a diagnosis coded with one of the 27 ICD-10 diagnoses considered as triggers were included. ADE confirmation was carried out by analyzing drugs prescribed prior to admission, and through a discussion among experts and a phone interview with patients after hospital discharge. RESULTS: 1,143 patients with trigger diagnoses were evaluated, of which 310 (27.1%) corresponded to patients whose emergency visit was attributed to an ADE. A 58.4% of ADE consultations were found with three diagnostic codes: K59.0-Constipation (n = 87; 28.1%), I16.9-Hypertensive Crisis (n = 72; 23.2%) and I95.1-Orthostatic hypotension (n = 22; 7.1%). The diagnoses with the highest degree of association with consultations attributed to ADE were E16.2-Hypoglycemia, unspecified (73.7%) and E11.65-Type 2 diabetes mellitus with hyperglycemia (71.4%), while diagnoses D62-Acute posthemorrhagic anemia and I74.3-Embolism and thrombosis of arteries of the lower limbs were not attributed to any case of ADE. CONCLUSIONS: The ICD-10 codes associated with trigger diagnoses are a useful tool to identify patients who consult the emergency services with ADE and could be used to apply secondary prevention programs to avoid new consultations to the health care system.

Prevalence of and Associated Factors for Eyelid Cancer in the American Academy of Ophthalmology Intelligent Research in Sight Registry.

Purpose: To estimate the prevalence of eyelid cancers in the American Academy of Ophthalmology Intelligent Research in Sight (IRIS) Registry and evaluate the associated factors. Design: Retrospective IRIS Registry database study. Participants: All patients in the IRIS Registry between December 1, 2010, and December 1, 2018, with International Classification of Disease, ninth and 10th revisions, codes for eyelid cancers (basal cell carcinoma [BCC], squamous cell carcinoma [SCC], malignant melanoma [MM], sebaceous carcinoma/other specified malignant neoplasm [SBC], melanoma in situ [MIS], and unspecified malignant neoplasm [UMN]). Methods: The prevalence of each eyelid cancer type was estimated overall and by age group, sex, race, ethnicity, and smoking status. The associations between any eyelid cancer (AEC) or each cancer type and possible risk factors were examined using univariate and multivariate logistic regression models. Main Outcome Measures: Prevalence of and associated factors for each eyelid cancer type. Results: There were 82 136 patients with eyelid cancer identified. The prevalence of AEC was 145.1 per 100 000 population. The cancer-specific prevalence ranged from 87.9 (BCC) to 25.6 (UMN), 11.1 (SCC), 5.0 (SBC), 4.1 (MM), and 0.4 (MIS) per 100 000 population. The prevalence of AEC and each cancer type increased with increasing age (all P < 0.0001), and the prevalence of AEC, BCC, SCC, and MM was higher in males (all P < 0.0001), MIS (P = 0.02). The prevalence of BCC, SCC, MM, SBC, and AEC was highest in Whites versus that in patients of any other race (all P < 0.0001). In the multivariate logistic regression model with associated risk factors (age, sex, race, ethnicity, and smoking status), AEC was associated with older age groups ([< 20 years reference {ref.}]; odds ratio [95% confidence interval]: 20-39 years: 3.35 [1.96-5.72]; 40-65 years: 24.21 [14.80-39.59]; and > 65 years: 42.78 [26.18-69.90]), male sex (female [ref.]; 1.40 [1.33-1.48]), White race (inverse associations with African Americans [0.12 {0.09-0.16}], Asians [0.19 {0.13-0.26}], others [0.59 {0.40-0.89}]), and ethnicity (non-Hispanic [ref.]; Hispanic: 0.38 [0.33-0.45]; unknown: 0.81 [0.75-0.88]). Active smoking (never smoker [ref.]) was associated with AEC (1.11 [1.01-1.21]), BCC (1.27 [1.23-1.31]), SCC (1.59 [1.46-1.73]), and MM (1.26 [1.08-1.46]). Conclusions: This study reports the overall and cancer-specific prevalence of eyelid cancers using a large national clinical eye disease database. Smoking was found to be associated with AEC, BCC, SCC, and MM, which is a new observation. This epidemiologic profile of on-eyelid cancers is valuable for identifying patients at a higher risk of malignancy, allocating medical resources, and improving cancer care.

Predictors of Long-term Ophthalmic Complications after Closed Globe Injuries Using the Intelligent Research in Sight (IRIS®) Registry.

Purpose: To identify clinical factors associated with the need for future surgical intervention following closed globe ocular trauma. Design: Retrospective cohort study. Subjects Participants and/or Controls: Patients in the American Academy of Ophthalmology Intelligent Research in Sight (IRIS®) Registry with a diagnosis of closed globe ocular trauma occurring between 2013 and 2019, identified using International Classification of Disease, 10th Revision and Systematized Nomenclature of Medicine codes. Methods: Diagnosis codes were used to identify multiple concomitant diagnoses present on the date of closed globe ocular trauma. Survival analyses were performed for each outcome of interest, and linear regression was used to identify clinical factors associated with the risk of surgical intervention. Main Outcome Measures: Outcomes included retinal break treatment, retinal detachment (RD) repair, retinal break treatment or RD repair, glaucoma surgery, and cataract surgery. Results: Of the 206 807 patients with closed globe ocular trauma, 9648 underwent surgical intervention during the follow-up period (mean, 444 days): 1697 (0.8%) had RD repair, 1658 (0.8%) had retinal break treatment, 600 (0.3%) had glaucoma surgery, and 5693 (2.8%) had cataract surgery. Traumatic cataract was the strongest risk factor for cataract surgery (hazard ratio, 13.0; 95% confidence interval, 10.8-15.6), traumatic hyphema showed highest risk for glaucoma surgery (7.24; 4.60-11.4), and vitreous hemorrhage was the strongest risk factor for retinal break treatment and detachment repair (11.01; 9.18-13.2 and 14.2; 11.5-17.6, respectively) during the first 60 days after trauma date. Vitreous hemorrhage was a risk factor for cataract surgery at > 60 days after trauma date only. Iris-angle injury was the strongest risk factor for glaucoma surgery > 60 days after trauma, while vitreous hemorrhage remained the strongest factor for retinal break treatment and detachment repair at > 60 days. Traumatic hyphema was a risk factor for all surgical outcomes during all follow-up intervals. Conclusions: Diagnosis of concomitant traumatic cataract, vitreous hemorrhage, traumatic hyphema, and other risk factors may increase the likelihood of requiring surgical intervention after closed globe ocular trauma.

Development and validation of an electronic health records-based opioid use disorder algorithm by expert clinical adjudication among patients with prescribed opioids.

BACKGROUND: In the US, over 200 lives are lost from opioid overdoses each day. Accurate and prompt diagnosis of opioid use disorders (OUD) may help prevent overdose deaths. However, international classification of disease (ICD) codes for OUD are known to underestimate prevalence, and their specificity and sensitivity are unknown. We developed and validated algorithms to identify OUD in electronic health records (EHR) and examined the validity of OUD ICD codes. METHODS: Through four iterations, we developed EHR-based OUD identification algorithms among patients who were prescribed opioids from 2014 to 2017. The algorithms and OUD ICD codes were validated against 169 independent "gold standard" EHR chart reviews conducted by an expert adjudication panel across four healthcare systems. After using 2014-2020 EHR for validating iteration 1, the experts were advised to use 2014-2017 EHR thereafter. RESULTS: Of the 169 EHR charts, 81 (48%) were reviewed by more than one expert and exhibited 85% expert agreement. The experts identified 54 OUD cases. The experts endorsed all 11 OUD criteria from the Diagnostic and Statistical Manual of Mental Disorders-5, including craving (72%), tolerance (65%), withdrawal (56%), and recurrent use in physically hazardous conditions (50%). The OUD ICD codes had 10% sensitivity and 99% specificity, underscoring large underestimation. In comparison our algorithm identified OUD with 23% sensitivity and 98% specificity. CONCLUSIONS AND RELEVANCE: This is the first study to estimate the validity of OUD ICD codes and develop validated EHR-based OUD identification algorithms. This work will inform future research on early intervention and prevention of OUD.

Model for Predicting In-Hospital Mortality of Physical Trauma Patients Using Artificial Intelligence Techniques: Nationwide Population-Based Study in Korea.

BACKGROUND: Physical trauma-related mortality places a heavy burden on society. Estimating the mortality risk in physical trauma patients is crucial to enhance treatment efficiency and reduce this burden. The most popular and accurate model is the Injury Severity Score (ISS), which is based on the Abbreviated Injury Scale (AIS), an anatomical injury severity scoring system. However, the AIS requires specialists to code the injury scale by reviewing a patient's medical record; therefore, applying the model to every hospital is impossible. OBJECTIVE: We aimed to develop an artificial intelligence (AI) model to predict in-hospital mortality in physical trauma patients using the International Classification of Disease 10th Revision (ICD-10), triage scale, procedure codes, and other clinical features. METHODS: We used the Korean National Emergency Department Information System (NEDIS) data set (N=778,111) compiled from over 400 hospitals between 2016 and 2019. To predict in-hospital mortality, we used the following as input features: ICD-10, patient age, gender, intentionality, injury mechanism, and emergent symptom, Alert/Verbal/Painful/Unresponsive (AVPU) scale, Korean Triage and Acuity Scale (KTAS), and procedure codes. We proposed the ensemble of deep neural networks (EDNN) via 5-fold cross-validation and compared them with other state-of-the-art machine learning models, including traditional prediction models. We further investigated the effect of the features. RESULTS: Our proposed EDNN with all features provided the highest area under the receiver operating characteristic (AUROC) curve of 0.9507, outperforming other state-of-the-art models, including the following traditional prediction models: Adaptive Boosting (AdaBoost; AUROC of 0.9433), Extreme Gradient Boosting (XGBoost; AUROC of 0.9331), ICD-based ISS (AUROC of 0.8699 for an inclusive model and AUROC of 0.8224 for an exclusive model), and KTAS (AUROC of 0.1841). In addition, using all features yielded a higher AUROC than any other partial features, namely, EDNN with the features of ICD-10 only (AUROC of 0.8964) and EDNN with the features excluding ICD-10 (AUROC of 0.9383). CONCLUSIONS: Our proposed EDNN with all features outperforms other state-of-the-art models, including the traditional diagnostic code-based prediction model and triage scale.

Risk Factors for Retinal Detachment Repair After Pediatric Cataract Surgery in the United States.

Purpose: To determine the cumulative incidence of retinal detachment (RD) repair following pediatric cataract surgery and identify the associated risk factors. Design: US population-based insurance claims retrospective cohort study. Participants: Patients ≤ 18 years old who underwent cataract surgery in 2 large databases: Optum Clinformatics (2003-2021) and IBM MarketScan (2007-2016). Methods: Individuals with ≥ 6 months of prior enrollment were included, and those with a history of RD, RD repair, traumatic cataract, spherophakia, or ectopia lentis were excluded. The primary outcome was time between initial cataract surgery and RD repair. The risk factors investigated included age, sex, persistent fetal vasculature (PFV), prematurity, intraocular lens (IOL) placement, and pars plana lensectomy approach. Main Outcome Measures: Kaplan-Meier estimated cumulative incidence of RD repair 5 years after cataract surgery and hazard ratios (HRs) with 95% confidence intervals (CIs) from multivariable Cox proportional hazards regression models. Results: Retinal detachment repair was performed on 47 of 3289 children included in this study. The cumulative incidence of RD repair within 5 years of cataract surgery was 2.0% (95% CI, 1.3%-2.6%). Children requiring RD repair were more likely to have a history of prematurity or PFV and less likely to have an IOL placed (all P < 0.001). Factors associated with RD repair in the multivariable analysis included a history of prematurity (HR, 6.89; 95% CI, 3.26-14.56; P < 0.001), PFV diagnosis (HR, 8.20; 95% CI, 4.11-16.37; P < 0.001), and IOL placement (HR, 0.44; 95% CI, 0.21-0.91; P = 0.03). Age at surgery, sex, and pars plana lensectomy approach were not significantly associated with RD repair after adjusting for all other covariates. Conclusions: Approximately 2% of patients will undergo RD repair within 5 years of pediatric cataract surgery. Children with a history of PFV and prematurity undergoing cataract surgery without IOL placement are at the greatest risk.

Administrative data in pediatric critical care research-Potential, challenges, and future directions.

Heterogenous patient populations with small case numbers constitute a relevant barrier to research in pediatric critical care. Prospective studies bring along logistic barriers and-if interventional-ethical concerns. Therefore, retrospective observational investigations, mainly multicenter studies or analyses of registry data, prevail in the field of pediatric critical care research. Administrative health care data represent a possible alternative to overcome small case numbers and logistic barriers. However, their current use is limited by a lack of knowledge among clinicians about the availability and characteristics of these data sets, along with required expertise in the handling of large data sets. Specifically in the field of critical care research, difficulties to assess the severity of the acute disease and estimate organ dysfunction and outcomes pose additional challenges. In contrast, trauma research has shown that classification of injury severity from administrative data can be achieved and chronic disease scores have been developed for pediatric patients, nurturing confidence that the remaining obstacles can be overcome. Despite the undoubted challenges, interdisciplinary collaboration between clinicians and methodologic experts have resulted in impactful publications from across the world. Efforts to enable the estimation of organ dysfunction and measure outcomes after critical illness are the most urgent tasks to promote the use of administrative data in critical care. Clever analysis and linking of different administrative health care data sets carry the potential to advance observational research in pediatric critical care and ultimately improve clinical care for critically ill children.