The National Healthcare Safety Network's digital quality measures: CDC's automated measures for surveillance of patient safety.

OBJECTIVE: This article presents the National Healthcare Safety Network (NHSN)'s approach to automation for public health surveillance using digital quality measures (dQMs) via an open-source tool (NHSNLink) and piloting of this approach using real-world data in a newly established collaborative program (NHSNCoLab). The approach leverages Health Level Seven Fast Healthcare Interoperability Resources (FHIR) application programming interfaces to improve data collection and reporting for public health and patient safety beginning with common, clinically significant, and preventable patient harms, such as medication-related hypoglycemia, healthcare facility-onset Clostridioides difficile infection, and healthcare-associated venous thromboembolism. CONCLUSIONS: The NHSN's FHIR dQMs hold the promise of minimizing the burden of reporting, improving accuracy, quality, and validity of data collected by NHSN, and increasing speed and efficiency of public health surveillance.

Association between Gout and Gastric Cancer: A Nested Case-Control Study Using a National Health Sample Cohort.

Given the global significance of gout and gastric cancer (GC) as major health problems with interrelated impacts, we examined the development of GC in Korean patients with gout. We conducted a nested case-control study using data from 10,174 GC patients and 40,696 control patients from the Korean National Health Insurance Service-National Sample Cohort database. Propensity score matching (1:4) with propensity score overlap-weighted adjustment was used to reduce selection bias and estimate the odds ratio (OR) and 95% confidence intervals (CIs) for the association between gout and GC. An adjusted OR for GC was not significantly higher in patients with gout than in control patients (1.02; 95% CI, 0.93-1.12; p = 0.652). Additionally, no association between gout and GC was observed in subgroup analyses such as sex, age, level of income, region of residence, or Charlson Comorbidity Index score. In conclusion, these results suggest that gout is not a significant independent risk factor for GC among the Korean population. Additional investigation is required to establish a causal association between gout and GC, and to generalize these results to general populations.

Epidemiological Analysis of Traumatic Compartment Syndromes in Germany.

Background: Traumatic compartment syndrome is a critical condition that can lead to severe, lifelong disability. Methods: This retrospective study analyzed hospital billing data from 2015 to 2022, provided by the Federal Statistical Office of Germany, to examine the demographics and trends of traumatic compartment syndrome in Germany. The analysis included cases coded with ICD-10 codes T79.60 to T79.69 and any therapeutic OPS code starting with 5-79, focusing on diagnosis year, gender, ICD-10 code, and patient age. Results: The results showed that out of 13,305 cases, the majority were in the lower leg (44.4%), with males having a significantly higher incidence than females (2.3:1 ratio). A bimodal age distribution was observed, with peaks at 22-23 and 55 years. A notable annual decline of 43.87 cases in compartment syndrome was observed, with significant decreases across different genders and age groups, particularly in males under 40 (23.68 cases per year) and in the "foot" and "lower leg" categories (16.67 and 32.87 cases per year, respectively). Conclusions: The study highlights a declining trend in traumatic CS cases in Germany, with distinct demographic patterns. Through these findings, hospitals can adjust their therapeutic regimens, and it could increase awareness among healthcare professionals about this disease.

Barriers to healthcare data quality and recommendations in public health facilities in Dire Dawa city administration, eastern Ethiopia: a qualitative study.

Background: Maintaining good quality of healthcare data at various levels is a critical challenge in developing countries. The barriers to healthcare data quality remain largely unexplored in eastern Ethiopia. Objective: This study aimed to assess the barriers to quality of healthcare data in urban public health facilities in the Dire Dawa city administration from 7 April to 7 May 2019. Methods: An institutional-based qualitative exploratory approach was used among 17 purposefully selected key informants. In-depth interviews were inductively coded using the ATLAS.ti 7.5.4 version software. Inductive analysis was used by semantically analyzing the explicit content of the data to determine our themes. Results: Several key themes and subthemes with different barriers, some of which are mutually non-exclusive, were identified. These include: Organizational Barriers: Lack of an adequate health management information system and data clerk staff, poor management commitment, lack of post-training follow-up, work overload, frequent duty rotation, lack of incentives for good performers, lack of targeted feedback, and poor culture of information use. Behavioral/Individual Barriers: Gaps in the skill of managers and health professionals, lack of adequate awareness of each indicator and its definitions, inadequate educational competence, lack of feeling of ownership, poor commitment, lack of daily tallying, and lack of value for data. Technical Barriers: Lack of a standard form, diverse and too many data entry formats, manual data collection, shortage of supplies, failure to repair system break down in a timely manner, interruption in electricity and network, delay in digitizing health information systems, lack of post-training follow-up, and inadequate supervision. External Barriers: Poor collaboration between stakeholders, dependence on the software program of non-governmental organizations, and very hot weather conditions. Conclusion: Diverse and complex barriers to maintenance of data quality were identified. Developing standardized health management information system implementation plans, providing advanced supervisory-level training, supportive supervision, and site-level mentorship may be very effective in identifying and resolving bottleneck data quality issues. Healthcare managers should understand the imperative of data quality and accept responsibility for its improvement and maintenance. Interventions targeted only at supplies will not fully overcome limitations to data quality. Motivation of staff and recognition of best performance can motivate others and can create cooperation among staff.

Getting our ducks in a row: The need for data utility comparisons of healthcare systems data for clinical trials.

BACKGROUND: Better use of healthcare systems data, collected as part of interactions between patients and the healthcare system, could transform planning and conduct of randomised controlled trials. Multiple challenges to widespread use include whether healthcare systems data captures sufficiently well the data traditionally captured on case report forms. "Data Utility Comparison Studies" (DUCkS) assess the utility of healthcare systems data for RCTs by comparison to data collected by the trial. Despite their importance, there are few published UK examples of DUCkS. METHODS-AND-RESULTS: Building from ongoing and selected recent examples of UK-led DUCkS in the literature, we set out experience-based considerations for the conduct of future DUCkS. Developed through informal iterative discussions in many forums, considerations are offered for planning, protocol development, data, analysis and reporting, with comparisons at "patient-level" or "trial-level", depending on the item of interest and trial status. DISCUSSION: DUCkS could be a valuable tool in assessing where healthcare systems data can be used for trials and in which trial teams can play a leading role. There is a pressing need for trials to be more efficient in their delivery and research waste must be reduced. Trials have been making inconsistent use of healthcare systems data, not least because of an absence of evidence of utility. DUCkS can also help to identify challenges in using healthcare systems data, such as linkage (access and timing) and data quality. We encourage trial teams to incorporate and report DUCkS in trials and funders and data providers to support them.

Comparative Analysis of Outpatient Antibiotic Prescribing in Early Life: A Population-Based Study Across Birth Cohorts in Denmark and Germany.

INTRODUCTION: Comparing antibiotic prescribing between countries can provide important insights into potential needs of improving antibiotic stewardship programs. We aimed to compare outpatient antibiotic prescribing in early life between children born in Denmark and Germany. METHODS: Using the Danish nationwide healthcare registries and a German claims database (GePaRD, ~ 20% population coverage), we included children born between 2004 and 2016, and followed them regarding outpatient antibiotic prescriptions until end of enrollment or the end of 2018. We then determined the median time to first antibiotic prescription. Based on all prescriptions in the first 2 years of life, we calculated the rate of antibiotic treatment episodes and for the children's first prescriptions in this period, we determined established quality indicators. All analyses were stratified by birth year and country. RESULTS: In the 2016 birth cohorts, the median time to first antibiotic prescription was ~ 21 months in Denmark and ~ 28 in Germany; the rate of antibiotic treatment episodes per 1000 person-years was 537 in Denmark and 433 in Germany; the percentage of prescribed antibiotics with higher concerns regarding side effects and/or resistance potential was 6.2% in Denmark and 44.2% in Germany. In the 2016 birth cohorts, the age at first antibiotic prescription was 50-59% higher compared to the 2004 birth cohorts; the rate of antibiotic treatment episodes was 43-44% lower. CONCLUSIONS: Infants in Denmark received antibiotics markedly earlier and more frequently than in Germany, while quality indicators of antibiotic prescribing were more favorable in Denmark. Although both countries experienced positive changes towards more rational antibiotic prescribing in early life, our findings suggest potential for further improvement. This particularly applies to prescribing antibiotics with a lower potential for side effects and/or resistance in Germany.

Structured Data Retrieval and Analysis of HL7 v2 Messages with Elasticsearch.

The era of the electronic health record (EHR) requires lots of semantic interoperability for data sharing and reusability. We select HL7 v2 messages as the most common structured data type in hospital information systems, to investigate the plausibility of using Elasticsearch (ES) as a healthcare search engine and data analytics tool. Due to the facts, Elasticsearch can be integrated as a powerful searchable database for practical healthcare applications, to analyze structured healthcare data from various locations. It allows easy and efficient searching for complex query tasks.

An Approach for Generating Realistic Australian Synthetic Healthcare Data.

Healthcare data is a scarce resource and access is often cumbersome. While medical software development would benefit from real datasets, the privacy of the patients is held at a higher priority. Realistic synthetic healthcare data can fill this gap by providing a dataset for quality control while at the same time preserving the patient's anonymity and privacy. Existing methods focus on American or European patient healthcare data but none is exclusively focused on the Australian population. Australia is a highly diverse country that has a unique healthcare system. To overcome this problem, we used a popular publicly available tool, Synthea, to generate disease progressions based on the Australian population. With this approach, we were able to generate 100,000 patients following Queensland (Australia) demographics.

Pneumococcal Vaccination Lowers the Risk of Alzheimer's Disease: A Study Utilizing Data from the IBM® MarketScan® Database.

Previous studies demonstrated an association between influenza vaccination and the likelihood of developing Alzheimer's disease. This study was aimed at assessing whether pneumococcal vaccinations are associated with a lower risk of Alzheimer's disease based on analysis of data from the IBM® MarketScan® Database. Vaccinated and unvaccinated matched cohorts were generated using propensity-score matching with the greedy nearest-neighbor matching algorithm. The conditional logistic regression method was used to estimate the relationship between pneumococcal vaccination and the onset of Alzheimer's disease. There were 142,874 subjects who received the pneumococcal vaccine and 14,392 subjects who did not. The conditional logistic regression indicated that the people who received the pneumococcal vaccine had a significantly lower risk of developing Alzheimer's disease as compared to the people who did not receive any pneumococcal vaccine (OR=0.37; 95%CI: 0.33-0.42; P-value < .0001). Our findings demonstrated that the pneumococcal vaccine was associated with a 63% reduction in the risk of Alzheimer's disease among US adults aged 65 and older.

Systematic reviews of machine learning in healthcare: a literature review.

INTRODUCTION: The increasing availability of data and computing power has made machine learning (ML) a viable approach to faster, more efficient healthcare delivery. METHODS: A systematic literature review (SLR) of published SLRs evaluating ML applications in healthcare settings published between1 January 2010 and 27 March 2023 was conducted. RESULTS: In total 220 SLRs covering 10,462 ML algorithms were reviewed. The main application of AI in medicine related to the clinical prediction and disease prognosis in oncology and neurology with the use of imaging data. Accuracy, specificity, and sensitivity were provided in 56%, 28%, and 25% SLRs respectively. Internal and external validation was reported in 53% and less than 1% of the cases respectively. The most common modeling approach was neural networks (2,454 ML algorithms), followed by support vector machine and random forest/decision trees (1,578 and 1,522 ML algorithms, respectively). EXPERT OPINION: The review indicated considerable reporting gaps in terms of the ML's performance, both internal and external validation. Greater accessibility to healthcare data for developers can ensure the faster adoption of ML algorithms into clinical practice.

CODA: an open-source platform for federated analysis and machine learning on distributed healthcare data.

OBJECTIVES: Distributed computations facilitate multi-institutional data analysis while avoiding the costs and complexity of data pooling. Existing approaches lack crucial features, such as built-in medical standards and terminologies, no-code data visualizations, explicit disclosure control mechanisms, and support for basic statistical computations, in addition to gradient-based optimization capabilities. MATERIALS AND METHODS: We describe the development of the Collaborative Data Analysis (CODA) platform, and the design choices undertaken to address the key needs identified during our survey of stakeholders. We use a public dataset (MIMIC-IV) to demonstrate end-to-end multi-modal FL using CODA. We assessed the technical feasibility of deploying the CODA platform at 9 hospitals in Canada, describe implementation challenges, and evaluate its scalability on large patient populations. RESULTS: The CODA platform was designed, developed, and deployed between January 2020 and January 2023. Software code, documentation, and technical documents were released under an open-source license. Multi-modal federated averaging is illustrated using the MIMIC-IV and MIMIC-CXR datasets. To date, 8 out of the 9 participating sites have successfully deployed the platform, with a total enrolment of >1M patients. Mapping data from legacy systems to FHIR was the biggest barrier to implementation. DISCUSSION AND CONCLUSION: The CODA platform was developed and successfully deployed in a public healthcare setting in Canada, with heterogeneous information technology systems and capabilities. Ongoing efforts will use the platform to develop and prospectively validate models for risk assessment, proactive monitoring, and resource usage. Further work will also make tools available to facilitate migration from legacy formats to FHIR and DICOM.

The long and the short of pediatric emergency department antibiotic prescribing: A retrospective observational study.

BACKGROUND: Most antibiotics prescribed to children are provided in the outpatient and emergency department (ED) settings, yet these prescribers are seldom engaged by antibiotic stewardship programs. We reviewed ED antibiotic prescriptions for three common infections to describe current prescribing practices. METHODS: Prescription data between 2018 and 2021 were extracted from the electronic records of children discharged from the Children's Hospital of Eastern Ontario ED with urinary tract infection (UTI), community acquired pneumonia (CAP), and acute otitis media ≥2 years of age (AOM). Antibiotic choice, duration, as well as the provider's time in practice and training background were collected. Antibiotic durations were compared with Canadian guideline recommendations to assess concordance. Provider-level prescribing practices were analyzed using k-means cluster analysis. RESULTS: 10,609 prescriptions were included: 2868 for UTI, 2958 for CAP, and 4783 for AOM. Guideline-concordant durations prescribed was generally high (UTI 84.9%, CAP 94.0%, AOM 52.8%), a large proportion of antibiotic-days prescribed were in excess of the minimally recommended duration for each infection (UTI 16.8%, 19.3%, AOM 25.5%). Cluster analysis yielded two clusters of prescribers, with those in one cluster more commonly prescribing durations at the lower end of recommended interval, and the others more commonly prescribing longer durations for all three infections reviewed. No statistically significant differences were found between clusters by career stage or training background. CONCLUSIONS: While guideline-concordant antibiotic prescribing was generally high, auditing antibiotic prescriptions identified shifting prescribing towards the minimally recommended duration as a potential opportunity to reduce antibiotic use among children for these infections.

Gout and Colorectal Cancer Likelihood: Insights from a Nested Case-Control Study of the Korean Population Utilizing the Korean National Health Insurance Service-National Sample Cohort.

Considering the global importance of both gout and colorectal cancer (CRC) as significant health issues with mutual relevance, we aimed to examine the risk of colorectal cancer in Korean patients with gout. In this nested case-control study, we used data from 9920 CRC patients and 39,680 controls the Korean National Health Insurance Service-National Sample Cohort database. Propensity score overlap-weighted multivariate logistic regression analyses, adjusted for confounders, were used to assess the odds ratio (OR) and 95% confidence interval (CI) of the association between gout and CRC. Adjusted OR for CRC were similar between patients with gout and the control group (0.95; 95% CI, 0.86-1.04; p = 0.282). However, after adjustment, subgroup analysis revealed an 18% reduction in the probability of CRC among patients younger than 65 years with gout (95% CI, 0.70-0.95; p = 0.009). Conversely, absence of an association between gout and subsequent CRC persisted regardless of sex, income, residence, and Charlson Comorbidity Index score, even among individuals aged 65 years or older. These results imply that gout may not be a significant independent risk factor for CRC among the general population. However, in patients younger than 65 years with gout, a slightly reduced likelihood of CRC was observed. Further research is necessary to establish a causal relationship between gout and CRC and to generalize these findings to other populations.

Proton Pump Inhibitors and Likelihood of Colorectal Cancer in the Korean Population: Insights from a Nested Case-Control Study Using National Health Insurance Data.

The potential connection between proton pump inhibitors (PPIs) and colorectal cancer (CRC) risk remains unclear, with specific ethnic genetic backgrounds playing a role in PPI-induced adverse effects. In this nested case-control study, we investigated the risk of CRC in relation to preceding PPI use and the duration of use using data from the Korean National Health Insurance Service-National Sample Cohort database, including 9374 incident CRC patients and 37,496 controls. To assess the impact of preceding PPI exposure (past vs. current) and use duration (days: <30, 30-90, and ≥90) on incident CRC, we conducted propensity score overlap-weighted multivariate logistic regression analyses, adjusted for confounding factors. Our findings revealed that past and current PPI users had an increased likelihood of developing CRC. Regardless of duration, individuals who used PPIs also had higher odds of developing CRC. Subgroup analyses revealed that CRC occurrence increased independent of history or duration of prior PPI use, consistent across various factors such as age, sex, income level, and residential area. These findings suggest that PPI use, regardless of past or present use and duration of use, may be related to an increased risk of developing CRC in the Korean population.

Inferring actual treatment pathways from patient records.

OBJECTIVE: Treatment pathways are step-by-step plans outlining the recommended medical care for specific diseases; they get revised when different treatments are found to improve patient outcomes. Examining health records is an important part of this revision process, but inferring patients' actual treatments from health data is challenging due to complex event-coding schemes and the absence of pathway-related annotations. The objective of this study is to develop a method for inferring actual treatment steps for a particular patient group from administrative health records - a common form of tabular healthcare data - and address several technique- and methodology-based gaps in treatment pathway-inference research. METHODS: We introduce Defrag, a method for examining health records to infer the real-world treatment steps for a particular patient group. Defrag learns the semantic and temporal meaning of healthcare event sequences, allowing it to reliably infer treatment steps from complex healthcare data. To our knowledge, Defrag is the first pathway-inference method to utilise a neural network (NN), an approach made possible by a novel, self-supervised learning objective. We also developed a testing and validation framework for pathway inference, which we use to characterise and evaluate Defrag's pathway inference ability, establish benchmarks, and compare against baselines. RESULTS: We demonstrate Defrag's effectiveness by identifying best-practice pathway fragments for breast cancer, lung cancer, and melanoma in public healthcare records. Additionally, we use synthetic data experiments to demonstrate the characteristics of the Defrag inference method, and to compare Defrag to several baselines, where it significantly outperforms non-NN-based methods. CONCLUSIONS: Defrag offers an innovative and effective approach for inferring treatment pathways from complex health data. Defrag significantly outperforms several existing pathway-inference methods, but computationally-derived treatment pathways are still difficult to compare against clinical guidelines. Furthermore, the open-source code for Defrag and the testing framework are provided to encourage further research in this area.

Machine Learning Algorithm to Perform the American Society of Anesthesiologists Physical Status Classification.

OBJECTIVE: The American Society of Anesthesiologists (ASA) Physical Status (PS) Classification System defines perioperative patient scores ranging from 1 to 6 (healthy to brain dead, respectively). The scoring is performed and used by physician anesthesiologists and providers to classify surgical patients based on co-morbidities and various clinical characteristics. There is potentially a variability in scoring stemming from individual biases. The biases impact the prediction of operating times, length of stay in the hospital, anesthetic management, and billing. This study's purpose was to develop an automated system to achieve reproducible scoring. METHODS: A machine learning (ML) model was trained on already assigned ASA PS scores of 12,064 patients. The ML algorithm was automatically selected by Wolfram Mathematica (Wolfram Research, Champaign, IL) and tested with retrospective records not used in training. Manual scoring was performed by the anesthesiologist as part of the standard preoperative evaluation. Intraclass correlation coefficient (ICC) in R (version 4.2.2; R Development Core Team, Vienna, Austria) was calculated to assess the consistency of scoring. RESULTS: An ML model was trained on the data corresponding to 12,064 patients. Logistic regression was chosen automatically, with an accuracy of 70.3±1.0% against the training dataset. The accuracy against 1,999 patients (the test dataset) was 69.6±1.0%. The ICC for the comparison between ML and the anesthesiologists' ASA PS scores was greater than 0.4 ("fair to good"). CONCLUSIONS: We have shown the feasibility of applying ML to assess the ASA PS score within an oncology patient population. Though our accuracy was not very good, we feel that, as more data are mined, a valid foundation for refinement to ML will emerge.

Socioeconomic differences in the utilization of diagnostic imaging and non-pharmaceutical conservative therapies for spinal diseases.

BACKGROUND: A different utilization of health care services due to socioeconomic status on the same health plan contradicts the principle of equal treatment. We investigated the presence and magnitude of socioeconomic differences in utilization of diagnostic imaging and non-pharmaceutical conservative therapies for patients with spinal diseases. METHODS: The cohort study based on routine healthcare data from Germany with 11.7 million patient-years between 2012 and 2016 for patients with physician-confirmed spinal diseases (ICD-10: M40-M54), occupation and age 20 to 64 years. A Poisson model estimated the effects of the socioeconomic status (school education, professional education and occupational position) for the risk ratio of receiving diagnostic imaging (radiography, computed tomography, magnetic resonance imaging) and non-pharmaceutical conservative therapies (physical therapy including exercise therapy, manual therapy and massage, spinal manipulative therapy, acupuncture). RESULTS: Patients received diagnostic imaging in 26%, physical therapy in 32%, spinal manipulative therapy in 25%, and acupuncture in 4% of all patient-years. Similar to previous survey-based studies higher rates of utilization were associated with higher socioeconomic status. These differences were most pronounced for manual therapy, exercise therapy, and magnetic resonance imaging. CONCLUSIONS: The observed differences in health care utilization were highly related to socioeconomic status. Socioeconomic differences were higher for more expensive health services. Further research is necessary to identify barriers to equitable access to health services and to take appropriate action to decrease existing social disparities.

Use of routine healthcare data in randomised implementation trials: a methodological mixed-methods systematic review.

BACKGROUND: Routine data are increasingly used in randomised controlled trials evaluating healthcare interventions. They can aid participant identification, outcome assessment, and intervention delivery. Randomised implementation trials evaluate the effect of implementation strategies on implementation outcomes. Implementation strategies, such as reminders, are used to increase the uptake of evidence-based interventions into practice, while implementation outcomes, such as adoption, are key measures of the implementation process. The use of routine data in effectiveness trials has been explored; however, there are no reviews on implementation trials. We therefore aimed to describe how routine data have been used in randomised implementation trials and the design characteristics of these trials. METHODS: We searched MEDLINE (Ovid) and Cochrane Central Register of Controlled Trials from Jan 2000 to Dec 2021 and manually searched protocols from trial registers. We included implementation trials and type II and type III hybrid effectiveness-implementation trials conducted using routine data. We extracted quantitative and qualitative data and narratively synthesised findings. RESULTS: From 4206 titles, we included 80 trials, of which 22.5% targeted implementation of evidence-based clinical guidelines. Multicomponent implementation strategies were more commonly evaluated (70.0%) than single strategies. Most trials assessed adoption as the primary outcome (65.0%). The majority of trials extracted data from electronic health records (EHRs) (62.5%), and 91.3% used routine data for outcome ascertainment. Reported reasons for using routine data were increasing efficiency, assessing outcomes, reducing research burden, improving quality of care, identifying study samples, confirming findings, and assessing representativeness. Data quality, the EHR system, research governance, and external factors such as government policy could act either as facilitators or barriers. CONCLUSIONS: Adherence to guidance on designing and reporting implementation studies, and specifically to harmonise the language used in describing implementation strategies and implementation outcomes, would aid identification of studies and data extraction. Routine healthcare data are widely used for participant identification, outcome assessment and intervention delivery. Researchers should familiarise themselves with the barriers and facilitators to using routine data, and efforts could be made to improve data quality to overcome some of the barriers. REGISTRATION: PROSPERO CRD42022292321.

Developing a prediction model of children asthma risk using population-based family history health records.

BACKGROUND: Identifying children at high risk of developing asthma can facilitate prevention and early management strategies. We developed a prediction model of children's asthma risk using objectively collected population-based children and parental histories of comorbidities. METHODS: We conducted a retrospective population-based cohort study using administrative data from Manitoba, Canada, and included children born from 1974 to 2000 with linkages to ≥1 parent. We identified asthma and prior comorbid condition diagnoses from hospital and outpatient records. We used two machine-learning models: least absolute shrinkage and selection operator (LASSO) logistic regression (LR) and random forest (RF) to identify important predictors. The predictors in the base model included children's demographics, allergic conditions, respiratory infections, and parental asthma. Subsequent models included additional multiple comorbidities for children and parents. RESULTS: The cohort included 195,666 children: 51.3% were males and 17.7% had asthma diagnosis. The base LR model achieved a low predictive performance with sensitivity of 0.47, 95% confidence interval (0.45-0.48), and specificity of 0.67 (0.66-0.67) using a predicted probability threshold of 0.20. Sensitivity significantly improved when children's comorbidities were included using LASSO LR: 0.71 (0.69-0.72). Predictive performance further improved by including parental comorbidities (sensitivity = 0.72 [0.70-0.73], specificity = 0.69 [0.69-0.70]). We observed similar results for the RF models. Children's menstrual disorders and mood and anxiety disorders, parental lipid metabolism disorders and asthma were among the most important variables that predicted asthma risk. CONCLUSION: Including children and parental comorbidities to children's asthma prediction models improves their accuracy.