Case Definition for Diagnosed Alzheimer Disease and Related Dementias in Medicare.

Importance: Lack of a US dementia surveillance system hinders efforts to support and address disparities among persons living with Alzheimer disease and related dementias (ADRD). Objective: To review diagnosis and prescription drug code ADRD identification algorithms to develop and implement case definitions for national surveillance. Design, Setting, and Participants: In this cross-sectional study, a systematic literature review was conducted to identify unique International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) and prescription drug codes used by researchers to identify ADRD in administrative records. Code frequency of use, characteristics of beneficiaries identified by codes, and expert and author consensus around code definitions informed code placement into categories indicating highly likely, likely, and possible ADRD. These definitions were applied cross-sectionally to 2017 to 2019 Medicare fee-for-service (FFS) claims and Medicare Advantage (MA) encounter data to classify January 2019 Medicare enrollees. Data analysis was conducted from September 2022 to March 2024. Exposures: ICD-10-CM and national drug codes in FFS claims or MA encounters. Main Outcomes and Measures: The primary outcome was counts and rates of beneficiaries meeting each case definition. Category-specific age, sex, race and ethnicity, MA enrollment, dual-eligibility, long-term care utilization, mortality, and rural residence distributions, as well as frailty scores and FFS monthly expenditures were also analyzed. Beneficiary characteristics were compared across categories, and age-standardized to minimize confounding by age. Results: Of the 60 000 869 beneficiaries included (50 853 806 aged 65 years or older [84.8%]; 32 567 891 female [54.3%]; 5 555 571 Hispanic [9.3%]; 6 318 194 non-Hispanic Black [10.5%]; 44 384 980 non-Hispanic White [74.0%]), there were 4 312 496 (7.2%) with highly likely ADRD, 1 124 080 (1.9%) with likely ADRD, and 2 572 176 (4.3%) with possible ADRD, totaling more than 8.0 million with diagnostic evidence of at least possible ADRD. These beneficiaries were older, more frail, more likely to be female, more likely to be dual-eligible, more likely to use long-term care, and more likely to die in 2019 compared with beneficiaries with no evidence of ADRD. These differences became larger when moving from the possible ADRD group to the highly likely ADRD group. Mean (SD) FFS monthly spending was $2966 ($4921) among beneficiaries with highly likely ADRD compared with $936 ($2952) for beneficiaries with no evidence of ADRD. Differences persisted after age standardization. Conclusions and Relevance: This cross-sectional study of 2019 Medicare beneficiaries identified more than 5.4 million Medicare beneficiaries with evidence of at least likely ADRD in 2019 using the diagnostic case definition. Pending validation against clinical and other methods of ascertainment, this approach can be adopted provisionally for national surveillance.

[The analysis of age dynamics of causes of mortality of population of elderly age groups as reserve of maintaining population of older able-bodied age].

The article analyzes age dynamics of initial causes of death according to records in medical death certificates of population of older age groups. Materials and methods. The records of causes of death of 34.914 persons aged 60 years and older were used as primary source of information. The initial cause of death was determined according to the ICD-10 rules. The rate of registration by reason of death was calculated as intensive value per 100 deaths in concrete age and sex group. Each cause was coded according to the ICD-10 rules (revision 2014-2016). The belonging to group was determined by first character (letter) in four-digit code that corresponded to the Class. On the basis of analysis of structure of causes of death, the group A of causes that included five Classes of ICD-10, determined 81.4% of all deaths in population aged 60 years and older. Two Classes: "Diseases of the circulatory system" (Class IX) and "Neoplasms" (Class II) determine in all studied age groups more than a half of all deaths (from 55% to 71% of males and from 59% to 67% in females) and namely they determine mortality rate in older age groups. There are no gender differences in age characteristics of registration rate in these groups (p > 0.05), however age dynamics differ. In case of diseases of circulatory system initial cause of death is increase rate of registration at increasing of age. In case of neoplasms at increasing of age decrease of registration rate as initial cause of death is established. At that, rate of decline is higher than rate of increase that determines certain decrease of structural significance of combined contribution of these two groups of causes at increasing of age. The Group B of causes, including three Classes of ICD-10 "Respiratory diseases" (Class X), "Diseases of the digestive system" (Class XI) and "Diseases of the nervous system" (Class VI), determined in overall 11.9% of all deaths in population aged 60 years and older. The age dynamics of causes of death of population of older age groups exists for certain groups of causes and it should be considered in organizing medical care of population of older age groups.

Self-Reported ICD-11 Personality Disorder Severity in Peruvian Adolescents: Structure, Validity, and Tentative Cutoffs.

Personality disorder (PD) is particularly common in adolescents, which underscores the significance of early screening, diagnosis, and intervention. To date, the definition of PD in the new ICD-11 has not yet been investigated in adolescents. This study therefore aimed to investigate the unidimensionality and criterion validity of self-reported ICD-11 PD features in Peruvian adolescents using the Personality Disorder Severity ICD-11 (PDS-ICD-11) scale. A total of 1,073 students (63% female; age range 12-16 years) were administered the PDS-ICD-11 scale along with criterion measures of personality pathology and symptom distress. The PDS-ICD-11 score showed adequate unidimensionality and conceptually meaningful associations with external criterion variables. The findings indicate that ICD-11 PD features, as measured with the PDS-ICD-11 scale, are structurally and conceptually sound when employed with adolescents. Norm-based cutoffs derived from the present study may be used for clinical interpretation. The PDS-ICD-11 may be employed as an efficient screening tool for personality dysfunction in adolescents.

[Clinical systematics of chronic delusional psychoses in schizophrenia and schizophrenia-spectrum pathology]. / Klinicheskaya sistematika khronicheskikh bredovykh psikhozov pri shizofrenii i patologii shizofrenicheskogo spektra.

OBJECTIVE: To develop a systematics of chronic delusional psychoses in schizophrenia and pathology of the schizophrenic spectrum that takes into account psychopathological structure (the mechanism of delusional formation) and the trajectories of the disease course (the ratio of negative/positive dimensions). MATERIAL AND METHODS: The study sample was recruited from the large Moscow psychiatric hospital in 2019-2024 and included 126 patients (94 male, 32 female, mean age 36.5±12.1 years) hospitalized with ICD-10 diagnosis of paranoid schizophrenia (F20.00) or delusional disorder (F22). A clinical-psychopathological method was used. RESULTS: The study identifies three variants of paranoid domains, the typology of which takes into account both the psychopathological structure and the trajectories of their development, which determine the dominance of predominantly negative or positive symptoms throughout the disease: 1) psychosis with delusional interpretations and predominant orientation of symptoms towards the negative pole of the schizophrenia psychopathology (n=37, 29.4%); 2) psychosis with hallucinatory delusions and predominant direction of symptoms towards the positive pole of the schizophrenia psychopathology (n=50, 39.7%); and 3) «combined¼ hallucinatory-delusional psychosis with a simultaneous orientation of symptoms to the negative/positive poles of the schizophrenia psychopathology (n=39, 30.9%). Detailed phenomenological characteristics of each of the three presented variants are given. CONCLUSION: The developed three-component taxonomy confirms, on the model of chronic delusional psychoses appearing in schizophrenia and the pathology of the schizophrenia spectrum, the concept of simultaneous representation of two relatively independent domains determined by neurobiological processes in the schizophrenia psychopathology positive and negative disorders.

[Schizophrenia spectrum disorders (aspects of psychopathology and clinical taxonomy)]. / Rasstroistva shizofrenicheskogo spektra (aspekty psikhopatologii i klinicheskaya sistematika).

OBJECTIVE: To create a new taxonomy of schizophrenia spectrum disorders (SSD) based on the comparability of the design of SSD and borderline states. MATERIAL AND METHODS: The total sample consists of 205 patients with an established diagnosis of SSD (F21; F25; F22 according to ICD-10) collected from studies of the department of borderline mental pathology and psychosomatic disorders of the Federal State Budgetary Institution Mental Health Research Center and the Department of Psychiatry and Psychosomatics of Moscow State Medical University in the period 2014 to 2024. Clinical, psychometric, statistical methods were used. RESULTS: A new two-level model of schizotypal personality disorder (STPD) has been developed: the first level is psychopathic-like disorders of the «Ferschroben¼ type; the second level are psychopathological disorders (positive, negative, etc.), appearing under their «mask¼, constituting a «tracing paper¼ of manifestations of schizophrenia «in miniature¼. The two-level psychopathological model of STPD is a complex clinical phenotype, including independent but overlapping phenotypic formations: psychopathic-like - the «Ferschroben¼ type; and basic - schizophreniform disorders. CONCLUSION: The clinical classification of schizophrenia spectrum disorders has been developed; pseudoneuroses and stress-induced disorders of the endogenous circle are considered in the aspect of the dynamics of STPD.

Accuracy of ICD-10 Diagnostic Codes for Identifying Housing Instability.

This cohort study assesses the performance of International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) Z59 codes for identifying housing instability during health care encounters.

Sensitivity of classification criteria from time of diagnosis in an incident systemic lupus erythematosus cohort: a population-based study from Norway.

OBJECTIVES: To compare the sensitivity of 2019 European Alliance of Associations for Rheumatology/American College of Rheumatology (EULAR/ACR) classification criteria against 1997 ACR criteria for systemic lupus erythematosus (SLE), for incident SLE cases in the presumably complete population-based Nor-SLE cohort from Southeast Norway (2.9 million inhabitants). METHODS: All cases International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) coded as SLE during 2000-2017 were individually reviewed. Those with a confirmed SLE diagnosis by expert clinical assessment were included in the Nor-SLE cohort. Core clinical data were recorded, and the cases were classified according to 2019 EULAR/ACR and 1997 ACR criteria. Juvenile SLE was defined as <16 years at diagnosis and adult SLE was defined as ≥16 years at diagnosis. RESULTS: We included 737 incident SLE cases (701 adults, 36 juveniles). At diagnosis, 2019 EULAR/ACR criteria were more sensitive than 1997 ACR criteria for adults (91.6% vs 77.3%; p<0.001), but not for juveniles (97.2% vs 88.9%). The 2019 EULAR/ACR counts at diagnosis differed by age group and ethnicity, being higher in young cases and those originating from Asia. From time of diagnosis to study end the fulfilment rate of 2019 EULAR/ACR criteria for the adult cohort increased from 92.5% and 86.5% to 94.6% and 91.0%, respectively, for females and males (mean disease duration of 7.5 years). CONCLUSION: Showing 92% criteria fulfilment already at time of SLE diagnosis by 2019 EULAR/ACR criteria versus 77% by 1997 ACR criteria, the results from this population-based study suggest that the 2019 EULAR/ACR criteria will achieve its goal of capturing more early-SLE cases for clinical trials.

What Kind of Transformer Models to Use for the ICD-10 Codes Classification Task.

Coding according to the International Classification of Diseases (ICD)-10 and its clinical modifications (CM) is inherently complex and expensive. Natural Language Processing (NLP) assists by simplifying the analysis of unstructured data from electronic health records, thereby facilitating diagnosis coding. This study investigates the suitability of transformer models for ICD-10 classification, considering both encoder and encoder-decoder architectures. The analysis is performed on clinical discharge summaries from the Medical Information Mart for Intensive Care (MIMIC)-IV dataset, which contains an extensive collection of electronic health records. Pre-trained models such as BioBERT, ClinicalBERT, ClinicalLongformer, and ClinicalBigBird are adapted for the coding task, incorporating specific preprocessing techniques to enhance performance. The findings indicate that increasing context length improves accuracy, and that the difference in accuracy between encoder and encoder-decoder models is negligible.

Prevalence and correlates of ICD-11-based prolonged grief disorder in a representative Slovakian sample of recently bereaved adults.

Background: Prolonged Grief Disorder (PGD) has recently been included in both the ICD-11 and DSM-5-TR diagnostic manuals. Studying its prevalence and correlates across cultures is vital for more effective identification, treatment, and prevention.Objective: This study aimed to examine prevalence rates of ICD-11-based PGD, in a representative Slovakian sample in response to deaths of loved ones occurring during the previous year. Further aims were to examine the factor structure of PGD symptoms and correlates of summed PGD item scores and PGD 'caseness'.Method: Self-reported data on PGD, depression, anxiety, alcohol use, and descriptive characteristics were gathered from a representative sample of the Slovak population (N = 319).Results: Data were gathered from N = 1853 people; 319 participants (17.2%) reported a loss in the past year. The prevalence of probable PGD among these bereaved participants was 1.99% for recent losses (<6 months, n = 151) and 7.75% for more distant losses (6-12 months, n = 130). The most frequently endorsed symptoms included longing/yearning for the deceased, sadness, denial/unrealness, and difficulty accepting the death. PGD symptoms had a unitary factor structure which was consistent for subsamples bereaved 1-5 and 6-12 months. The severity of PGD varied with kinship. Depression and anxiety, but not alcohol misuse, were associated with PGD severity and PGD caseness.Conclusions: These findings underscore that a significant group of people develop PGD between 6-12 months following a loss. This emphasises the need for targeted psychological interventions.

Prolonged Grief Disorder (PGD) is newly included in ICD-11 and knowledge about its prevalence and correlates in the general population is urgently needed.In a representative Slovakian sample (N = 1853), 319 people (17.2%) reported a loss during the past year; 7.75% of people, bereaved 6­12 months earlier, met criteria for ICD-11-based PGD.PGD severity and caseness were associated with kinship (but less strongly with other sociodemographic and loss characteristics) and with depression and anxiety (but less strongly with problematic alcohol use).At 6­12 months following loss, PGD seems fairly common in the general population and timely identification and mitigation of PGD is an important public health issue.

Reformatted Crosswalk-Tables Between Annual ICD-10-Versions for Cross Version Data Analysis.

The International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) is internationally used for coding diagnoses, with the ICD-10 German Modification (GM) being prescribed for morbidity coding in Germany. ICD-10-GM is subject to annual revisions. This can lead to backward compatibility issues leading to undesirable consequences for cross-version data analysis. A study of annual crosswalk-tables concerning 21 ICD-10-GM versions showed that the ratio of difficult transitions from an older to a newer version (0.89 %) and vice versa (0.48 %) is not particularly significant but should nevertheless not be neglected. In this paper we present two solutions (Neo4J database and FHIR ConceptMaps) for the automated handling of different ICD-10-GM versions.

Linking Health Terminologies: A Unified Approach to the WHO Family of International Classifications.

This paper presents an effort by the World Health Organization (WHO) to integrate the reference classifications of the Family of International Classifications (ICD, ICF, and ICHI) into a unified digital framework. The integration was accomplished via an expanded Content Model and a single Foundation that hosts all entities from these classifications, allowing the traditional use cases of individual classifications to be retained while enhancing their combined use. The harmonized WHO-FIC Content Model and the unified Foundation has streamlined the content management, enhanced the web-based tool functionalities, and provided opportunities for linkage with external terminologies and ontologies. This integration promises reduced maintenance cost, seamless joint application, complete representation of health-related concepts while enabling better interoperability with other informatics infrastructures.

Challenges of Integrating ICD 11 into Automatic Alerting Systems.

Automatic alerting systems (AASs) can identify adverse health events but emergency communication relies on human operators and natural languages. For complete automation, we need to code the diversity of adverse events in a granularity that supports optimal dispatches. Hence, AAs shall integrate with the International Classification of Diseases (ICD). The ICD-11 coding system includes chapters for external causes of injury. However, ICD-11 supports coding injury incidents in electronic health records (EHRs) after they have occurred, while disregarding integrating real-time injury reporting within its framework. We explore the potential challenges associated with integrating ICD-11 into AAS by analyzing external causes of morbidity or mortality and the dimensions of external causes as potential areas of integration. We recognize the themes: (i) incident of injury, (ii) mode of transport, (iii) indoor location, (iv) outdoor location, and (v) type of building, and identify four challenges: (i) conceptual differences between the two systems, (ii) injury identification, (iii) presence of entities below the shoreline in ICD-11, and (iv) lack of specificity in certain ICD-11 codes related to AASs. For easy integration of ICD-11 into AASs, we recommend an AAS data dictionary and propose ICD-11 updates related to external causes of injury.

Term Candidate Generation to Enrich Clinical Terminologies with Large Language Models.

Annotated language resources derived from clinical routine documentation form an intriguing asset for secondary use case scenarios. In this investigation, we report on how such a resource can be leveraged to identify additional term candidates for a chosen set of ICD-10 codes. We conducted a log-likelihood analysis, considering the co-occurrence of approximately 1.9 million de-identified ICD-10 codes alongside corresponding brief textual entries from problem lists in German. This analysis aimed to identify potential candidates with statistical significance set at p < 0.01, which were used as seed terms to harvest additional candidates by interfacing to a large language model in a second step. The proposed approach can identify additional term candidates at suitable performance values: hypernyms MAP@5=0.801, synonyms MAP@5 = 0.723 and hyponyms MAP@5 = 0.507. The re-use of existing annotated clinical datasets, in combination with large language models, presents an interesting strategy to bridge the lexical gap in standardized clinical terminologies and real-world jargon.

Incorrect and Sex-Inconsistent Mapping of Disorders: Identification of Sex Biases in the ICD-10, ICD-11 and SNOMED CT and How to Work Around Them.

In the international classifications ICD-10-WHO and ICD-11-WHO, many sex-specific diseases have incomplete coding. It is possible to further enhance semantic interoperability using SNOMED CT additionally to ICD. Part of the analysis of semantic interoperability of diagnoses in the ICD are Sexual Dysfunctions, Postpartum Depression, Sexual Assault, Premenstrual Tension Syndrome and Premenstrual Dysphoric Disorder, Female Genital Mutilation and Cutting, Gender Incongruence and Disorders of Breast. Labeling biases have been identified in all diagnoses, either in SNOMED CT or ICD. For mental disorders associated with pregnancy, gender incongruence and sexual violence the use of the GPS of SNOMED CT can help enhance semantic interoperability additionally to ICD.

[Cannabis use and cannabis use disorders]. / Cannabiskonsum und Cannabiskonsumstörungen.

Cannabis use and cannabis use disorders have taken on a new social significance as a result of partial legalization. In 2021 a total of 4.5 million adults (8.8%) in Germany used the drug. The number of users as well as problematic use have risen in the last decade. Cannabis products with a high delta-9-tetrahydrocannabinol (THC) content and their regular use lead to changes in cannabinoid receptor distribution in the brain and to modifications in the structure and functionality of relevant neuronal networks. The consequences of cannabinoid use are particularly in the psychological functioning and can include intoxication, harmful use, dependence with withdrawal symptoms and cannabis-induced mental disorders. Changes in the diagnostics between ICD-10 and ICD-11 are presented. Interdisciplinary S3 guidelines on cannabis-related disorders are currently being developed and will be finalized shortly.

Demographics, medication use, and admission characteristics of patients hospitalized with diabetes in Ontario, Canada: A retrospective cohort study.

BACKGROUND: In Canada, one in seven adults has diabetes (i.e., 2.3 million) and the lifetime risk of developing diabetes is approximately 30% by age 65. Although 30% of patients admitted to the hospital have diabetes, data from inpatient hospitalizations for patients with diabetes are lacking, both in Canada and globally. OBJECTIVE: To validate International Classification of Diseases 10th edition Canadian version (ICD-10-CA) codes for the identification of patients with diabetes, to create a multicenter database of patients with diabetes hospitalized under internal medicine in Ontario, and to determine their baseline characteristics, medication use, and admission characteristics. STUDY DESIGN: We created a database of people who had diabetes and were hospitalized between 2010 and 2020 at 8 hospitals in Ontario that were part of the General Medicine Inpatient Initiative (GEMINI) hospital data-sharing network. Patients who had diabetes were identified using chart review, based upon either (i) a previous physician diagnosis of diabetes, (ii) a recorded hemoglobin A1c ≥ 6.5% or (iii) outpatient prescription of a diabetes medication preceding the hospitalization. The test characteristics of ICD-10-CA codes for diabetes were evaluated. We compared baseline demographics, medication use and hospitalization details among patients with and without diabetes. For hospitalization details, we collected information on the admission diagnosis, comorbidity index, length of stay, receipt of ICU-level care, and inpatient mortality. RESULTS: There were 384,588 admissions within the total study cohort, of which 118,987 (30.9%) had an ICD-10-CA diagnosis code of diabetes (E10.x, E11.x, E13.x, E14.x). The sensitivity and specificity of ICD-10-CA diagnostic codes was 95.9% and 98.8%, respectively. Most patients with an ICD-10-CA code for diabetes had a code for type 2 diabetes (93.9%) and a code for type 1 diabetes was rare (6.1%). The mean age was 66.4 years for patients without diabetes and 71.3 years for those with an ICD-10-CA diagnosis code for diabetes. Patients with diabetes had a higher prevalence of hypertension (64% vs. 37.9%), coronary artery disease (28.7% vs. 15.3%), heart failure (24.5% vs. 12.1%) and renal failure (33.8% vs. 17.3%) in comparison to those without diabetes. The most prevalent diabetes medications received in hospital were metformin (43%), DPP4 inhibitors (22.7%) and sulfonylureas (18.8%). The most common reason for admission among patients with diabetes was heart failure (9.0%), and among patients without diabetes was pneumonia (7.8%). Median length of stay was longer for patients with diabetes (5.5 vs. 4.5 days) and in-hospital mortality was similar between groups (6.8% with diabetes vs. 6.5% without diabetes). IMPORTANCE: Diabetes is one of the most prevalent chronic medical conditions, affecting roughly one third of all patients hospitalized on an internal medicine ward and is associated with other comorbidities and longer hospital stays. ICD-10-CA codes were highly accurate in identifying patients with diabetes. The development of an inpatient cohort will allow for further study of in-hospital practices and outcomes among patients with diabetes.

Standardization of coding definitions for sickle cell disease complications: A systematic literature review.

PURPOSE: Sickle cell disease (SCD) affects all organ systems and is characterized by numerous acute and chronic complications and comorbidities. Standardized codes are needed for complications/comorbidities used in real-world evidence (RWE) studies that rely on administrative and medical coding. This systematic literature review was conducted to produce a comprehensive list of complications/comorbidities associated with SCD, along with their diagnosis codes used in RWE studies. METHODS: A search in MEDLINE and Embase identified studies published from 2016 to 2023. Studies were included if they were conducted in US SCD populations and reported complications/comorbidities and respective International Classification of Diseases, Clinical Modification (ICD-CM) codes. All identified complications/comorbidities and codes were reviewed by a certified medical coding expert and hematologist. RESULTS: Of 1851 identified studies, 39 studies were included. The most reported complications/comorbidities were stroke, acute chest syndrome, pulmonary embolism, venous thromboembolism, and vaso-occlusive crisis. Most of the studies used ICD-9-CM codes (n = 21), while some studies used ICD-10-CM codes (n = 3) or both (n = 15), depending on the study period. Most codes reported in literature were heterogeneous across complications/comorbidities. The medical coding expert and hematologist recommended modifications for several conditions. CONCLUSION: While many studies we identified did not report their codes and were excluded from this review, the studies with codes exhibited diverse coding definitions. By providing a standardized set of diagnosis codes that were reported by studies and reviewed by a coding expert and hematologist, our review can serve as a foundation for accurately identifying complications/comorbidities in future research, and may reduce heterogeneity, enhance transparency, and improve reproducibility. Future efforts focused on validating these code lists are needed.

ICDXML: enhancing ICD coding with probabilistic label trees and dynamic semantic representations.

Accurately assigning standardized diagnosis and procedure codes from clinical text is crucial for healthcare applications. However, this remains challenging due to the complexity of medical language. This paper proposes a novel model that incorporates extreme multi-label classification tasks to enhance International Classification of Diseases (ICD) coding. The model utilizes deformable convolutional neural networks to fuse representations from hidden layer outputs of pre-trained language models and external medical knowledge embeddings fused using a multimodal approach to provide rich semantic encodings for each code. A probabilistic label tree is constructed based on the hierarchical structure existing in ICD labels to incorporate ontological relationships between ICD codes and enable structured output prediction. Experiments on medical code prediction on the MIMIC-III database demonstrate competitive performance, highlighting the benefits of this technique for robust clinical code assignment.

Improving posttraumatic stress disorder assessment in young children: comparing measures and identifying clinically-relevant symptoms in children ages six and under.

Young children exposed to trauma are vulnerable to developing posttraumatic stress disorder (PTSD). Although experts agree on the importance of tailoring PTSD assessments to be developmentally-appropriate for young children, there is little research on which assessment methods best identify clinically significant symptomatology in this difficult-to-assess population. Two competing models for assessing PTSD have been proposed by the DSM-5 and ICD-11. This study compared a DSM-5 measure to an ICD-11 measure in young children (ages 1-6) exposed to a natural disaster. The measures identified similar rates of PTSD in children; however, diagnostic agreement between the measures was low (31-36%). Both PTSD measures were associated with actual and perceived life threat, functional impairment, and comorbid psychopathology. PTSD symptom and cluster endorsement rates were also binned by age and compared to functional impairment to identify commonly-reported and highly impairing symptoms in trauma-exposed young children, as potential candidates for inclusion in future revisions of diagnostic criteria. Across age ranges, increased clinginess post-trauma was the most commonly reported symptom and was associated with functional impairment. Arousal symptoms (startle, hypervigilance) and Re-experiencing symptoms (nightmares) also emerged as relevant for young children. Findings may contribute to efforts to improve assessment for PTSD in young children.

Developing and testing a framework for coding general practitioners' free-text diagnoses in electronic medical records - a reliability study for generating training data in natural language processing.

BACKGROUND: Diagnoses entered by general practitioners into electronic medical records have great potential for research and practice, but unfortunately, diagnoses are often in uncoded format, making them of little use. Natural language processing (NLP) could assist in coding free-text diagnoses, but NLP models require local training data to unlock their potential. The aim of this study was to develop a framework of research-relevant diagnostic codes, to test the framework using free-text diagnoses from a Swiss primary care database and to generate training data for NLP modelling. METHODS: The framework of diagnostic codes was developed based on input from local stakeholders and consideration of epidemiological data. After pre-testing, the framework contained 105 diagnostic codes, which were then applied by two raters who independently coded randomly drawn lines of free text (LoFT) from diagnosis lists extracted from the electronic medical records of 3000 patients of 27 general practitioners. Coding frequency and mean occurrence rates (n and %) and inter-rater reliability (IRR) of coding were calculated using Cohen's kappa (Κ). RESULTS: The sample consisted of 26,980 LoFT and in 56.3% no code could be assigned because it was not a specific diagnosis. The most common diagnostic codes were, 'dorsopathies' (3.9%, a code covering all types of back problems, including non-specific lower back pain, scoliosis, and others) and 'other diseases of the circulatory system' (3.1%). Raters were in almost perfect agreement (Κ ≥ 0.81) for 69 of the 105 diagnostic codes, and 28 codes showed a substantial agreement (K between 0.61 and 0.80). Both high coding frequency and almost perfect agreement were found in 37 codes, including codes that are particularly difficult to identify from components of the electronic medical record, such as musculoskeletal conditions, cancer or tobacco use. CONCLUSION: The coding framework was characterised by a subset of very frequent and highly reliable diagnostic codes, which will be the most valuable targets for training NLP models for automated disease classification based on free-text diagnoses from Swiss general practice.