Building and validating 5-feature models to predict preeclampsia onset time from electronic health record data

Ballard, Hailey; Yang, Xiaotong; Mahadevan, Aditya; Garmire, Lana; Lemas, Dominick J.

doi:10.1101/2023.03.23.23287655

Cited by 3 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Future studies incorporating causal inference or randomized controlled trials may offer more insights into the causal pathways between PE and subsequent health outcomes. Given the increasing recognized subtypes of preeclampsia, which exhibit different pathological processes, diagnosis time, symptoms, time to deliveries, as well as outcome, it may be necessary to stratify PE by subtype if the patient size is su ciently large in future studies [42][43][44]. Lastly, our ndings are purely based on EHR data.…”

Section: Discussionmentioning

confidence: 96%

Discover overlooked complications after preeclampsia using electronic health records

Garmire,

Zhu,

Yang

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Background Preeclampsia (PE) is a severe pregnancy complication characterized by hypertension and end-organ damage such as proteinuria. PE poses a significant threat to women's long-term health, including an increased risk of cardiovascular and renal diseases. Most previous studies have been hypothesis-based, potentially overlooking certain significant complications. This study conducts a comprehensive, non-hypothesis-based analysis of PE-complicated diagnoses after pregnancies using multiple large-scale electronic health records (EHR) datasets. Method From the University of Michigan (UM) Healthcare System, we collected 4,348 PE patients for the cases and 27,377 patients with pregnancies not complicated by PE or related conditions for the controls. We first conducted a non-hypothesis-based analysis to identify any long-term adverse health conditions associated with PE using logistic regression with adjustments to demographics, social history, and medical history. We confirmed the identified complications with UK Biobank data which contain 443 PE cases and 14,870 non-PE controls. We then conducted a survival analysis on complications that exhibited significance in more than 5 consecutive years post-PE. We further examined the potential racial disparities of identified complications between Caucasian and African American patients. Findings Uncomplicated hypertension, complicated diabetes, congestive heart failure, renal failure, and obesity exhibited significantly increased risks whereas hypothyroidism showed decreased risks, in 5 consecutive years after PE in the UM discovery data. UK Biobank data confirmed the increased risks of uncomplicated hypertension, complicated diabetes, congestive heart failure, renal failure, and obesity. Further survival analysis using UM data indicated significantly increased risks in uncomplicated hypertension, complicated diabetes, congestive heart failure, renal failure, and obesity, and significantly decreased risks in hypothyroidism. There exist racial differences in the risks of developing hypertension and hypothyroidism after PE. PE protects against hypothyroidism in African American postpartum women but not Cacausians; it also increases the risks of uncomplicated hypertension but less severely in African American postpartum women as compared to Cacausians. Interpretation This study addresses the lack of a comprehensive examination of PE's long-term effects utilizing large-scale EHR and advanced statistical methods. Our findings underscore the need for long-term monitoring and interventions for women with a history of PE, emphasizing the importance of personalized postpartum care. Notably, the racial disparities observed in the impact of PE on hypertension and hypothyroidism highlight the necessity of tailored aftercare based on race.

show abstract

Section: Discussionmentioning

confidence: 96%

Discover overlooked complications after preeclampsia using electronic health records

Garmire,

Zhu,

Yang

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Future studies incorporating causal inference or randomized controlled trials may offer more insights into the causal pathways between PE and subsequent health outcomes. Given the increasing recognized subtypes of preeclampsia, which exhibit different pathological processes, diagnosis time, symptoms, time to deliveries, as well as outcome, it may be necessary to stratify PE by subtype if the patient size is sufficiently large in future studies [42–44]. Lastly, our findings are purely based on EHR data.…”

Section: Discussionmentioning

confidence: 99%

Discover overlooked complications after preeclampsia from three real-world medical record datasets of over 100,000 pregnancies

Zhu,

Yang,

Tao

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

BackgroundPreeclampsia (PE) is a severe pregnancy complication characterized by hypertension and proteinuria. PE poses a significant threat to women’s health, including an increased risk of cardiovascular and renal diseases. Most previous studies have been hypothesis-based, potentially overlooking significant comorbidities. This study conducts a comprehensive, non-hypothesis-based analysis of PE comorbidities using large-scale electronic health records (EHR) data.MethodFrom the University of Michigan (UM) Healthcare System, we collected 4,348 PE patients for the case cohort and 27,377 patients with pregnancies not complicated by PE or related conditions for the control cohort. We first conducted a non-hypothesis-based exploration to identify any comorbidities associated with PE using logistic regression with adjustments to demographics, lifestyles, and medical history. We validated the results by performing logistic regression on the post-PE risks of the identified comorbidities from UKBiobank data. The validation data contain 443 PE cases and 14,954 non-PE controls from UKBiobank. We then conducted survival analysis on comorbidities that exhibited significance in more than 5 consecutive years post-PE. We further examined the racial disparities of identified comorbidities between Caucasian and African American patients.FindingsUncomplicated hypertension, complicated diabetes, congestive heart failure, renal failure, obesity, and hypothyroidism exhibited significantly increased risks in 5 consecutive years after PE in the UM discovery data. Caucasians were more sensitive to PE’s impact on hypertension and hypothyroidism compared to African Americans. Validation using UKBiobank data confirmed increased risks in uncomplicated hypertension, complicated diabetes, congestive heart failure, renal failure, and obesity.InterpretationThis study addresses the lack of a comprehensive examination of PE’s long-term effects utilizing large-scale EHR and modern statistical methods. Our findings underscored the need for long-term monitoring and interventions for women with a history of PE, emphasizing the importance of personalized postpartum care. Notably, the racial disparities observed in the impact of PE on hypertension and hypothyroidism highlight the necessity of tailored aftercare based on ethnicity.

show abstract

Deep learning-based prognosis models accurately predict the time to delivery among preeclampsia patients using health records at the time of diagnosis

Yang

Ballard

Mahadevan

et al. 2022

Preprint

View full text Add to dashboard Cite

BackgroundPreeclampsia (PE) is one of the leading factors in maternal and perinatal mortality and morbidity worldwide. The only cure for PE to date is to deliver the placenta and stop gestation. However, the timing of delivery among PE patients is essential to minimize the risk of severe maternal morbidities, and at the same time ensure the survival of the baby.MethodsIn this study, we constructed a series of deep learning-based models to predict the prognosis, or the time to delivery, since the initial diagnosis of PE using electronic health record (EHR) data. We extracted and processed 1578 pregnancies in Michigan Medicine at the University of Michigan in 2015-2021 as the discovery cohort. Using the Cox-nnet v2 algorithm, we built the baseline model with EHR information prior to diagnosis, as well as the full model including baseline information and lab testing results and vital signs at the time of diagnosis. We evaluated the models using the C-index and log-rank p-values in KM survival curves, using both 20% testing data of the Michigan cohort, as well as 1177 PE pregnancy EHR data from the Medical Center of the University of Florida.ResultsThe baseline prognosis model for time to delivery since PE diagnosis achieved C-index values of 0.75 and 0.72 on the training and testing set respectively. While the full model reached C-indices of 0.77 and 0.74 in the same training and testing sets. Both models performed better than their Cox-PH model counterparts. The seven most important features in the baseline model in descending order were diagnosis gestational age, severe PE, past PE, age, parity, gravidity, and uncomplicated diabetes. Meanwhile, 14 most important features were selected and interpreted in the full model, including diagnosis gestational age, parity, severe PE, past PE, features in lab tests (white blood cell, platelet, and red blood cell counts, AST value), min respiratory rate, and features measuring blood pressure (minimum, mean and standard deviation of systolic blood pressure, and maximum and standard deviation of diastolic blood pressure).ConclusionThe time to delivery predicting models provide clinicians valuable tools and options to quantify the delivery risks and make better decisions on the optimal delivery time of PE patients at the time of diagnosis. Implementation of these actionable models into PE clinical care practice is expected to significantly improve the management of PE patients.

show abstract

Building and validating 5-feature models to predict preeclampsia onset time from electronic health record data

Cited by 3 publications

References 17 publications

Discover overlooked complications after preeclampsia using electronic health records

Discover overlooked complications after preeclampsia using electronic health records

Discover overlooked complications after preeclampsia from three real-world medical record datasets of over 100,000 pregnancies

Deep learning-based prognosis models accurately predict the time to delivery among preeclampsia patients using health records at the time of diagnosis

Contact Info

Product

Resources

About