Predicting opioid overdose risk of patients with opioid prescriptions using electronic health records based on temporal deep learning

Dong, Xinyu; Deng, Jianyuan; Hou, Wei; Rashidian, Sina; Rosenthal, Richard N.; Saltz, Mary; Saltz, Joel; Wang, Fusheng

doi:10.1016/j.jbi.2021.103725

Cited by 27 publications

(30 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We obtained similar results with medications such as naloxone, acetaminophen-IV and epinephrine (Table 3), and methadone and heroin (Table 4) being predictive features for OUD. This is also similar to the findings in [37] where the authors mention that the medication class of opioids are among the most important features for OUD prediction. Similar to [36], we found methadone (Table 4) and folic acid (Table 3) to be present in EHRs of OUD patients.…”

Section: Prediction Of Opioid Use Disorder (Oud)supporting

confidence: 89%

“…They also found opioid dependent patients to be commonly malnourished and suffer from psychiatric disorders. Researchers trained an LSTM [38] model with attention to predict opioid overdose risk for patients prescribed opioids from EHRs [37]. They found the most important features for identifying patients at risk of overdose to be the mention of opioids, pain scale score, antipropulsives, general anesthetics and alcohol use among others.…”

Section: Use Of Machine Learning Models To Predict Opioid Dependence or Use Disordermentioning

confidence: 99%

“…In other research [36], a machine learning model was built on EHRs to predict opioid dependence, which obtained a mean area under the receiver operating characteristic curve (AUC-ROC) of 92%. An LSTM model with attention built in [37] to predict overdose risk achieved an F1 score of 0.78 and an AUC-ROC of 0.8449. The OUD model built in this paper achieved an F1 score of 0.82 ± 0.05, and an AUC-ROC score of 0.94 ± 0.008 (Table 2).…”

Section: Prediction Of Opioid Use Disorder (Oud)mentioning

confidence: 99%

“…Our models overcome some limitations that researchers studying opioid dependence in EHR data have faced in the past. A lot of previous research relies on domain knowledge from clinical experts for feature engineering, as mentioned in [37]. This typically requires a lot of effort and can be quite cumbersome and expensive.…”

Section: Our Deep Learning Models Overcome Known Limitations In the Fieldmentioning

confidence: 99%

See 3 more Smart Citations

A Deep Learning Method to Detect Opioid Prescription and Opioid Use Disorder from Electronic Health Records

Kashyap

Callison-Burch

Boland

2021

Preprint

View full text Add to dashboard Cite

Objective: As the opioid epidemic continues across the United States, methods are needed to accurately and quickly identify patients at risk for opioid use disorder (OUD). The purpose of this study is to develop two predictive algorithms: one to predict opioid prescription and one to predict OUD. Materials and Methods: We developed an informatics algorithm that trains two deep learning models over patient EHRs using the MIMIC-III database. We utilize both the structured and unstructured parts of the EHR and show that it is possible to predict both of these challenging outcomes. Results: Our deep learning models incorporate both structured and unstructured data elements from the EHRs to predict opioid prescription with an F1-score of 0.88 +/- 0.003 and an AUC-ROC of 0.93 +/- 0.002. We also constructed a model to predict OUD diagnosis achieving an F1-score of 0.82 +/- 0.05 and AUC-ROC of 0.94 +/- 0.008. Discussion: Our model for OUD prediction outperformed prior algorithms for specificity, F1 score and AUC-ROC while achieving equivalent sensitivity. This demonstrates the importance of a.) deep learning approaches in predicting OUD and b.) incorporating both structured and unstructured data for this prediction task. No prediction models for opioid prescription as an outcome were found in the literature and therefore this represents an important contribution of our work as opioid prescriptions are more common than OUDs. Conclusion: Algorithms such as those described in this paper will become increasingly important to understand the drivers underlying this national epidemic.

show abstract

Section: Prediction Of Opioid Use Disorder (Oud)supporting

confidence: 89%

Section: Use Of Machine Learning Models To Predict Opioid Dependence or Use Disordermentioning

confidence: 99%

Section: Prediction Of Opioid Use Disorder (Oud)mentioning

confidence: 99%

Section: Our Deep Learning Models Overcome Known Limitations In the Fieldmentioning

confidence: 99%

See 2 more Smart Citations

A Deep Learning Method to Detect Opioid Prescription and Opioid Use Disorder from Electronic Health Records

Kashyap

Callison-Burch

Boland

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…For each patient, we aggregated the diagnoses codes from all inpatient encounters within one calendar year. The International Classification of Diseases (ICD) codes for (1) poisoning and (2) adverse effect by opium, heroin, methadone, and other related narcotics, among others, were then used to identify patients with opioid overdose [ 35 , 36 ]. All relevant ICD Ninth Revision (ICD-9) and Tenth Revision (ICD-10) codes were summarized in Table S1 of the Electronic Supplementary Material (ESM) [ 37 ].…”

Section: Methodsmentioning

confidence: 99%

A Large-Scale Observational Study on the Temporal Trends and Risk Factors of Opioid Overdose: Real-World Evidence for Better Opioids

Deng

Hou

Dong

et al. 2021

Drugs - Real World Outcomes

Self Cite

View full text Add to dashboard Cite

Background The USA is in the midst of an opioid overdose epidemic. To address the epidemic, we conducted a large-scale population study on opioid overdose. Objectives The primary objective of this study was to evaluate the temporal trends and risk factors of inpatient opioid overdose. Based on its patterns, the secondary objective was to examine the innate properties of opioid analgesics underlying reduced overdose effects. Methods A retrospective cross-sectional study was conducted based on a large-scale inpatient electronic health records database, Cerner Health Facts ® , with (1) inclusion criteria for participants as patients admitted between 1 January, 2009 and 31 December, 2017 and (2) measurements as opioid overdose prevalence by year, demographics, and prescription opioid exposures.

show abstract

Predicting Chronic Opioid Use Among Patients With Osteoarthritis Using Electronic Health Record Data

Mohl

Stempniewicz

Cuddeback

et al. 2022

Arthritis Care & Research

View full text Add to dashboard Cite

Objective To estimate the risk of a patient with osteoarthritis (OA) developing chronic opioid use (COU) within 1 year of a new opioid prescription by using electronic health record (EHR) data and predictive models. Methods We used EHR data from 13 health care organizations to identify patients with OA with an opioid prescription between March 1, 2017 and February 28, 2019 and no record of opioid use in the prior 6 months. We evaluated 4 machine learning models to estimate patients’ risk of COU (≥3 prescriptions ≥84 days, maximum gap ≤60 days). We also estimated the transportability of models to organizations outside the training set. Results The cohort consisted of 33,894 patients with OA, of whom 2,925 (8.6%) developed COU within 1 year. All models demonstrated good discrimination, with the best‐performing model (random forest) achieving an area under the receiver operating characteristic curve (AUC) of 0.728 (95% CI 0.711–0.745), but the simplest regression model performed nearly as well (AUC 0.717 [95% CI 0.699–0.734]). Predicted risk deciles spanned a range of 2% risk for the 10th percentile to 18% risk for the 90th percentile for well‐calibrated models. Models showed highly variable discrimination across organizations (AUC 0.571–0.842). Conclusions We found that EHR‐based predictive models could estimate the risk of future COU among patients with OA to help inform care decisions. Black‐box methods did not have significant advantages over more interpretable models. Care should be taken when extending all models into organizations not included in model training because of a high variability in performance across held‐out organizations.

show abstract

Predicting opioid overdose risk of patients with opioid prescriptions using electronic health records based on temporal deep learning

Cited by 27 publications

References 21 publications

A Deep Learning Method to Detect Opioid Prescription and Opioid Use Disorder from Electronic Health Records

A Deep Learning Method to Detect Opioid Prescription and Opioid Use Disorder from Electronic Health Records

A Large-Scale Observational Study on the Temporal Trends and Risk Factors of Opioid Overdose: Real-World Evidence for Better Opioids

Predicting Chronic Opioid Use Among Patients With Osteoarthritis Using Electronic Health Record Data

Contact Info

Product

Resources

About