Trajectories: a framework for detecting temporal clinical event sequences from health data standardized to the Observational Medical Outcomes Partnership (OMOP) Common Data Model.

Objective: To develop a framework for identifying temporal clinical event trajectories from Observational Medical Outcomes Partnership-formatted observational healthcare data. Materials and Methods: A 4-step framework based on significant temporal event pair detection is described and implemented as an open-source R package. It is used on a population-based Estonian dataset to first replicate a large Danish population-based study and second, to conduct a disease trajectory detection study for type 2 diabetes patients in the Estonian and Dutch databases as an example. Results: As a proof of concept, we apply the methods in the Estonian database and provide a detailed breakdown of our findings. All Estonian population-based event pairs are shown. We compare the event pairs identified from Estonia to Danish and Dutch data and discuss the causes of the differences. The overlap in the results was only 2.4%, which highlights the need for running similar studies in different populations. Conclusions: For the first time, there is a complete software package for detecting disease trajectories in health data.

Use of unstructured text in prognostic clinical prediction models: a systematic review.

OBJECTIVE: This systematic review aims to assess how information from unstructured text is used to develop and validate clinical prognostic prediction models. We summarize the prediction problems and methodological landscape and determine whether using text data in addition to more commonly used structured data improves the prediction performance. MATERIALS AND METHODS: We searched Embase, MEDLINE, Web of Science, and Google Scholar to identify studies that developed prognostic prediction models using information extracted from unstructured text in a data-driven manner, published in the period from January 2005 to March 2021. Data items were extracted, analyzed, and a meta-analysis of the model performance was carried out to assess the added value of text to structured-data models. RESULTS: We identified 126 studies that described 145 clinical prediction problems. Combining text and structured data improved model performance, compared with using only text or only structured data. In these studies, a wide variety of dense and sparse numeric text representations were combined with both deep learning and more traditional machine learning methods. External validation, public availability, and attention for the explainability of the developed models were limited. CONCLUSION: The use of unstructured text in the development of prognostic prediction models has been found beneficial in addition to structured data in most studies. The text data are source of valuable information for prediction model development and should not be neglected. We suggest a future focus on explainability and external validation of the developed models, promoting robust and trustworthy prediction models in clinical practice.

Trends in the conduct and reporting of clinical prediction model development and validation: a systematic review.

OBJECTIVES: This systematic review aims to provide further insights into the conduct and reporting of clinical prediction model development and validation over time. We focus on assessing the reporting of information necessary to enable external validation by other investigators. MATERIALS AND METHODS: We searched Embase, Medline, Web-of-Science, Cochrane Library, and Google Scholar to identify studies that developed 1 or more multivariable prognostic prediction models using electronic health record (EHR) data published in the period 2009-2019. RESULTS: We identified 422 studies that developed a total of 579 clinical prediction models using EHR data. We observed a steep increase over the years in the number of developed models. The percentage of models externally validated in the same paper remained at around 10%. Throughout 2009-2019, for both the target population and the outcome definitions, code lists were provided for less than 20% of the models. For about half of the models that were developed using regression analysis, the final model was not completely presented. DISCUSSION: Overall, we observed limited improvement over time in the conduct and reporting of clinical prediction model development and validation. In particular, the prediction problem definition was often not clearly reported, and the final model was often not completely presented. CONCLUSION: Improvement in the reporting of information necessary to enable external validation by other investigators is still urgently needed to increase clinical adoption of developed models.

A Higher Incidence of Isolated Biliary Atresia in Rural Areas: Results From an Epidemiological Study in The Netherlands.

OBJECTIVES: Environmental factors may be involved in the pathogenesis of biliary atresia (BA). This epidemiological study aimed to analyze the relationships between the incidence of BA, the incidence of confirmed viral or bacterial infections and population density, and geographical and temporal clustering of BA in the Netherlands. STUDY DESIGN: Correlations between the monthly incidence of BA and the number of confirmed infections were assessed. BA incidence per province was calculated and compared to the province with highest population density. Birthplaces were classified as rural or urban. Temporal clustering of month of birth and month of conception were analyzed. We performed analyses for isolated BA (IBA) and syndromic BA (SBA) separately. Chi2, logistic regression, and Walter and Elwood test were used. RESULTS: A total of 262 IBA and 49 SBA patients, born between 1987 and 2018, were included. IBA incidence correlated to the number of confirmed infections of, for example, Chlamydia trachomatis (R = 0.14; P = 0.02) and adenovirus (R = 0.22; P = 0.005). We observed a higher incidence of IBA (0.75/10,000; odds ratio [OR] = 1.86; P = 0.04) and SBA (0.27/10,000; OR = 6.91; P = 0.001) in Groningen and a higher incidence of SBA in Gelderland (0.13/10,000; OR = 3.35; P = 0.03). IBA incidence was 68% higher in rural (0.67/10,000) versus urban areas (0.40/10,000) (P = 0.02). The estimated month of conception of patients with SBA clustered in November (85% increase compared to average SBA incidence [0.09/10,000; P = 0.04]). CONCLUSIONS: IBA incidence correlated weakly with national confirmed infections. IBA and SBA incidence varied geographically in the Netherlands. IBA incidence was higher in rural than in urban areas, which may be explained decreased exposure to pathogens. Our results provide support for a role of environmental factors in the pathogenesis of IBA.

Test Strategies to Predict Inflammatory Bowel Disease Among Children With Nonbloody Diarrhea.

OBJECTIVES: We evaluated 4 diagnostic strategies to predict the presence of inflammatory bowel disease (IBD) in children who present with chronic nonbloody diarrhea and abdominal pain. METHODS: We conducted a prospective cohort study including 193 patients aged 6 to 18 years who underwent a standardized diagnostic workup in secondary or tertiary care hospitals. Each patient was assessed for symptoms, C-reactive protein (>10 mg/L), hemoglobin (<-2 SD for age and sex), and fecal calprotectin (≥250 µg/g). Patients with rectal bleeding or perianal disease were excluded because the presence of these findings prompted endoscopy regardless of their biomarkers. Primary outcome was IBD confirmed by endoscopy or IBD ruled out by endoscopy or uneventful clinical follow-up for 6 months. RESULTS: Twenty-two of 193 (11%) children had IBD. The basic prediction model was based on symptoms only. Adding blood or stool markers increased the AUC from 0.718 (95% confidence interval [CI]: 0.604-0.832) to 0.930 (95% CI: 0.884-0.977) and 0.967 (95% CI: 0.945-0.990). Combining symptoms with blood and stool markers outperformed all other strategies (AUC 0.997 [95% CI: 0.993-1.000]). Triaging with a strategy that involves symptoms, blood markers, and calprotectin will result in 14 of 100 patients being exposed to endoscopy. Three of them will not have IBD, and no IBD-affected child will be missed. CONCLUSIONS: Evaluating symptoms plus blood and stool markers in patients with nonbloody diarrhea is the optimal test strategy that allows pediatricians to reserve a diagnostic endoscopy for children at high risk for IBD.

Application of the urban exposome framework using drinking water and quality of life indicators: a proof-of-concept study in Limassol, Cyprus.

BACKGROUND: Cities face rapid changes leading to increasing inequalities and emerging public health issues that require cost-effective interventions. The urban exposome concept refers to the continuous monitoring of urban environmental and health indicators using the city and smaller intra-city areas as measurement units in an interdisciplinary approach that combines qualitative and quantitative methods from social sciences, to epidemiology and exposure assessment. METHODS: In this proof of concept study, drinking water and quality of life indicators were described as part of the development of the urban exposome of Limassol (Cyprus) and were combined with agnostic environment-wide association analysis. This study was conducted as a two-part project with a qualitative part assessing the perceptions of city stakeholders, and quantitative part using a cross-sectional study design (an urban population study). We mapped the water quality parameters and participants' opinions on city life (i.e., neighborhood life, health care, and green space access) using quarters (small administrative areas) as the reference unit of the city. In an exploratory, agnostic, environment-wide association study analysis, we used all variables (questionnaire responses and water quality metrics) to describe correlations between them. RESULTS: Overall, urban drinking-water quality using conventional indicators of chemical (disinfection byproducts-trihalomethanes (THM)) and microbial (coliforms, E. coli, and Enterococci) quality did not raise particular concerns. The general health and chronic health status of the urban participants were significantly (false discovery rate corrected p-value < 0.1) associated with different health conditions such as hypertension and asthma, as well as having financial issues in access to dental care. Additionally, correlations between THM exposures and participant behavioral characteristics (e.g., household cleaning, drinking water habits) were documented. CONCLUSION: This proof-of-concept study showed the potential of using integrative approaches to develop urban exposomic profiles and identifying within-city differences in environmental and health indicators. The characterization of the urban exposome of Limassol will be expanded via the inclusion of biomonitoring tools and untargeted metabolomics.

Biomarkers of end of shift exposure to disinfection byproducts in nurses.

Increased disinfectant use commonly takes place in hospitals and other health care settings. A cross-sectional study among active nurses in two Cypriot public hospitals (n=179) was conducted to examine the prevalence of exposure to disinfection byproducts (DBPs), such as trihalomethanes (THMs) using both self-reported information and biomarker measurements. The objectives of this study were to: i) quantify the magnitude and variability of occupational exposure to disinfectants/DBPs in nurses, ii) generate job exposure matrices (JEM) and job task exposure matrices (JTEM) for disinfectants, and iii) assess the major determinants of urinary THMs in nurses. End of shift urinary total THM values showed high variability among the nurses, but did not differ between hospitals. The disinfectant group of alcohols/phenols was used by >98% of nurses, followed by octenidine (82%), iodine and chlorine (39%, each), chlorhexidine (25%), formaldehyde (12%), hydrogen peroxide (11%), and peracetic acid/ammonia/quaternary ammonium compounds (QACs), all being <8% each. Chlorine use during the past 24hr was associated with significantly (p<0.05) lower brominated THMs (BrTHMs) after adjusting for age, gender and BMI, while a positive association was shown for TCM and the sum of all THMs (TTHMs), albeit not significant. Nurses were exposed to nearly double the levels of TTHMs and BrTHMs (median and IQR, 1027 [560, 2475] ng/g and 323 [212, 497] ng/g, respectively) when compared to those of the general population (552 [309,989] ng/g and 152 [87,261] ng/g, respectively). This was the first occupational health dataset reporting measurements of biomarkers of end of shift exposures to disinfectants/DBPs.