Explore the vast range of titles available.

Although prior research has identified multiple risk factors for diabetic ketoacidosis (DKA), clinicians continue to lack clinic-ready models to predict dangerous and costly episodes of DKA. We asked whether we could apply deep learning, specifically the use of a long short-term memory (LSTM) model, to accurately predict the 180-day risk of DKA-related hospitalization for youth with type 1 diabetes (T1D). We aimed to describe the development of an LSTM model to predict the 180-day risk of DKA-related hospitalization for youth with T1D. We used 17 consecutive calendar quarters of clinical data (January 10, 2016, to March 18, 2020) for 1745 youths aged 8 to 18 years with T1D from a pediatric diabetes clinic network in the Midwestern United States. The input data included demographics, discrete clinical observations (laboratory results, vital signs, anthropometric measures, diagnosis, and procedure codes), medications, visit counts by type of encounter, number of historic DKA episodes, number of days since last DKA admission, patient-reported outcomes (answers to clinic intake questions), and data features derived from diabetes- and nondiabetes-related clinical notes via natural language processing. We trained the model using input data from quarters 1 to 7 (n=1377), validated it using input from quarters 3 to 9 in a partial out-of-sample (OOS-P; n=1505) cohort, and further validated it in a full out-of-sample (OOS-F; n=354) cohort with input from quarters 10 to 15. DKA admissions occurred at a rate of 5% per 180-days in both out-of-sample cohorts. In the OOS-P and OOS-F cohorts, the median age was 13.7 (IQR 11.3-15.8) years and 13.1 (IQR 10.7-15.5) years; median glycated hemoglobin levels at enrollment were 8.6% (IQR 7.6%-9.8%) and 8.1% (IQR 6.9%-9.5%); recall was 33% (26/80) and 50% (9/18) for the top-ranked 5% of youth with T1D; and 14.15% (213/1505) and 12.7% (45/354) had prior DKA admissions (after the T1D diagnosis), respectively. For lists rank ordered by the probability of hospitalization, precision increased from 33% to 56% to 100% for positions 1 to 80, 1 to 25, and 1 to 10 in the OOS-P cohort and from 50% to 60% to 80% for positions 1 to 18, 1 to 10, and 1 to 5 in the OOS-F cohort, respectively. The proposed LSTM model for predicting 180-day DKA-related hospitalization was valid in this sample. Future research should evaluate model validity in multiple populations and settings to account for health inequities that may be present in different segments of the population (eg, racially or socioeconomically diverse cohorts). Rank ordering youth by probability of DKA-related hospitalization will allow clinics to identify the most at-risk youth. The clinical implication of this is that clinics may then create and evaluate novel preventive interventions based on available resources.

Background:Clinicians currently lack an effective means for identifying youth with type 1 diabetes (T1D) who are at risk for experiencing glycemic deterioration between diabetes clinic visits. As a result, their ability to identify youth who may optimally benefit from targeted interventions designed to address rising glycemic levels is limited. Although electronic health records (EHR)–based risk predictions have been used to forecast health outcomes in T1D, no study has investigated the potential for using EHR data to identify youth with T1D who will experience a clinically significant rise in glycated hemoglobin (HbA1c) ≥0.3% (approximately 3 mmol/mol) between diabetes clinic visits.Objective:We aimed to evaluate the feasibility of using routinely collected EHR data to develop a machine learning model to predict 90-day unit-change in HbA1c (in % units) in youth (aged 9‐18 y) with T1D. We assessed our model’s ability to augment clinical decision-making by identifying a percent change cut point that optimized identification of youth who would experience a clinically significant rise in HbA1c.Methods:From a cohort of 2757 youth with T1D who received care from a network of pediatric diabetes clinics in the Midwestern United States (January 2012-August 2017), we identified 1743 youth with 9643 HbA1c observation windows (ie, 2 HbA1c measurements separated by 70‐110 d, approximating the 90-day time interval between routine diabetes clinic visits). We used up to 5 years of youths’ longitudinal EHR data to transform 17,466 features (demographics, laboratory results, vital signs, anthropometric measures, medications, diagnosis codes, procedure codes, and free-text data) for model training. We performed 3-fold cross-validation to train random forest regression models to predict 90-day unit-change in HbA1c(%).Results:Across all 3 folds of our cross-validation model, the average root-mean-square error was 0.88 (95% CI 0.85‐0.90). Predicted HbA1c(%) strongly correlated with true HbA1c(%) (r=0.79; 95% CI 0.78‐0.80). The top 10 features impacting model predictions included postal code, various metrics related to HbA1c, and the frequency of a diagnosis code indicating difficulty with treatment engagement. At a clinically significant percent rise threshold of ≥0.3% (approximately 3 mmol/mol), our model’s positive predictive value was 60.3%, indicating a 1.5-fold enrichment (relative to the observed frequency that youth experienced this outcome [3928/9643, 40.7%]). Model sensitivity and positive predictive value improved when thresholds for clinical significance included smaller changes in HbA1c, whereas specificity and negative predictive value improved when thresholds required larger changes in HbA1c.Conclusions:Routinely collected EHR data can be used to create an ML model for predicting unit-change in HbA1c between diabetes clinic visits among youth with T1D. Future work will focus on optimizing model performance and validating the model in additional cohorts and in other diabetes clinics.

Clinicians currently lack an effective means for identifying youth with type 1 diabetes (T1D) who are at risk for experiencing glycemic deterioration between diabetes clinic visits. As a result, their ability to identify youth who may optimally benefit from targeted interventions designed to address rising glycemic levels is limited. Although electronic health records (EHR)-based risk predictions have been used to forecast health outcomes in T1D, no study has investigated the potential for using EHR data to identify youth with T1D who will experience a clinically significant rise in glycated hemoglobin (HbA1c) ≥0.3% (approximately 3 mmol/mol) between diabetes clinic visits. We aimed to evaluate the feasibility of using routinely collected EHR data to develop a machine learning model to predict 90-day unit-change in HbA1c (in % units) in youth (aged 9-18 y) with T1D. We assessed our model's ability to augment clinical decision-making by identifying a percent change cut point that optimized identification of youth who would experience a clinically significant rise in HbA1c. From a cohort of 2757 youth with T1D who received care from a network of pediatric diabetes clinics in the Midwestern United States (January 2012-August 2017), we identified 1743 youth with 9643 HbA1c observation windows (ie, 2 HbA1c measurements separated by 70-110 d, approximating the 90-day time interval between routine diabetes clinic visits). We used up to 5 years of youths' longitudinal EHR data to transform 17,466 features (demographics, laboratory results, vital signs, anthropometric measures, medications, diagnosis codes, procedure codes, and free-text data) for model training. We performed 3-fold cross-validation to train random forest regression models to predict 90-day unit-change in HbA1c(%). Across all 3 folds of our cross-validation model, the average root-mean-square error was 0.88 (95% CI 0.85-0.90). Predicted HbA1c(%) strongly correlated with true HbA1c(%) (r=0.79; 95% CI 0.78-0.80). The top 10 features impacting model predictions included postal code, various metrics related to HbA1c, and the frequency of a diagnosis code indicating difficulty with treatment engagement. At a clinically significant percent rise threshold of ≥0.3% (approximately 3 mmol/mol), our model's positive predictive value was 60.3%, indicating a 1.5-fold enrichment (relative to the observed frequency that youth experienced this outcome [3928/9643, 40.7%]). Model sensitivity and positive predictive value improved when thresholds for clinical significance included smaller changes in HbA1c, whereas specificity and negative predictive value improved when thresholds required larger changes in HbA1c. Routinely collected EHR data can be used to create an ML model for predicting unit-change in HbA1c between diabetes clinic visits among youth with T1D. Future work will focus on optimizing model performance and validating the model in additional cohorts and in other diabetes clinics.