Phenotyping Antidepressant Treatment Response with Deep Learning in Electronic Health Records

Abstract

Efficient, accurate phenotyping for antidepressant treatment response in electronic health records (EHRs) could facilitate precision psychiatry applications but remains a challenge. Increasingly, artificial intelligence methods using "deep learning" applied to clinical data have shown promise in complex classification problems. Here, we systematically evaluate the performance of eight deep-learning-based natural language processing models in classifying response to antidepressants in a large real-world healthcare setting. We obtained data spanning 1990-2018 for adults with depression and a co-occurring antidepressant prescription from the EHR data warehouse of the Mass General Brigham healthcare system (n=111,572). Clinical notes were collected for the following time windows after antidepressant initiation: (1) 2 days to 4 weeks, (2) 4-12 weeks, and (3) 12-26 weeks. A stratified random sample of these note sets (total 4,299 across time periods) were manually reviewed to classify response status as "improved" or "no evidence of improvement" in depression symptoms. All models performed well, with areas under the receiver operator curve (AUROC) of at least 0.80. Positive predictive values (PPVs) ranged from 0.72 - 0.91. In general, models incorporating more information-dense and longer text sequences performed better than others. The best performing model (Longformer-large with sliding window) had an AUROC = 0.88 and PPV = 0.84 at a specificity of 0.88. Our results indicate that deep learning methods applied to EHR data can accurately classify antidepressant response in a real-world healthcare setting. Automated treatment response classification may facilitate a range of research and clinical decision support applications.

Investigating genetic overlap between antidepressant and lithium response and treatment resistance in major depressive disorder

“…Data from clinical trials are useful for this purpose; however, there is a paucity of clinical trials with su cient size and adequate length of follow-up. Recently, emerging efforts utilizing Electronic Health Records provide an exciting opportunity to study treatment response and resistance in large-scale, real-world healthcare settings 16,42 . Using comprehensive clinical records from the Swedish healthcare registers, our work extends these efforts and further demonstrates that integrating the information of ECT use can provide a useful TRD de nition.…”

Real-world observation on response to cholinesterase inhibitors or selective serotonin reuptake inhibitors prescribed to outpatients with dementia using electronic medical records

“…The manual review presented in this study can be an important first step that can be utilized to develop a reference standard used to build Natural Language Processing (NLP) model for structuring clinical data of dementia patients for further automated, large-scale analyses of EMRs. 25 26 Second, our analysis was performed in the EMRs with American English. Clinical text may naturally and structurally differ between healthcare systems, so the English words found and categorized in our study may not be fully generalizable to EMRs of other healthcare systems; adjustments may be necessary if similar analyses were to be conducted at other sites.…”

Translating clinical notes into quantitative measures—a real-world observation on the response to cholinesterase inhibitors or selective serotonin reuptake inhibitors prescribed to outpatients with dementia using electronic medical records

“…We expect these limitations can be addressed by incorporating artificial intelligence. The manual review presented in this study can be an important first step that can be utilized to develop a reference standard used to build the natural language processing (NLP) model for structuring clinical data of dementia patients for further automated, large-scale analyses of EMRs ( Vaci et al, 2020 ; Sheu et al, 2021 ; Vaci et al, 2021 ). In the same light, we chose to focus on CEIs and SSRIs in this study because these two classes of medications are the most commonly used in our study cohort.…”