Predicting maintenance lithium response for bipolar disorder from electronic health records—a retrospective study

Background Optimising maintenance drug treatment selection for people with bipolar disorder is challenging. There is some evidence that clinical and demographic features may predict response to lithium. However, attempts to personalise treatment choice have been limited. Method We aimed to determine...

Full description

Saved in:
Bibliographic Details
Main Authors: Joseph F. Hayes, Fehmi Ben Abdesslem, Sandra Eloranta, David P. J. Osborn, Magnus Boman
Format: Article
Language:English
Published: PeerJ Inc. 2024-10-01
Series:PeerJ
Subjects:
Online Access:https://peerj.com/articles/17841.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850183984986718208
author Joseph F. Hayes
Fehmi Ben Abdesslem
Sandra Eloranta
David P. J. Osborn
Magnus Boman
author_facet Joseph F. Hayes
Fehmi Ben Abdesslem
Sandra Eloranta
David P. J. Osborn
Magnus Boman
author_sort Joseph F. Hayes
collection DOAJ
description Background Optimising maintenance drug treatment selection for people with bipolar disorder is challenging. There is some evidence that clinical and demographic features may predict response to lithium. However, attempts to personalise treatment choice have been limited. Method We aimed to determine if machine learning methods applied to electronic health records could predict differential response to lithium or olanzapine. From electronic United Kingdom primary care records, we extracted a cohort of individuals prescribed either lithium (19,106 individuals) or olanzapine (12,412) monotherapy. Machine learning models were used to predict successful monotherapy maintenance treatment, using 113 clinical and demographic variables, 8,017 (41.96%) lithium responders and 3,831 (30.87%) olanzapine responders. Results We found a quantitative structural difference in that lithium maintenance responders were weakly predictable in our holdout sample, consisting of the 5% of patients with the most recent exposure. Age at first diagnosis, age at first treatment and the time between these were the most important variables in all models. Discussion Even if we failed to predict successful monotherapy olanzapine treatment, and so to definitively separate lithium vs. olanzapine responders, the characterization of the two groups may be used for classification by proxy. This can, in turn, be useful for establishing maintenance therapy. The further exploration of machine learning methods on EHR data for drug treatment selection could in the future play a role for clinical decision support. Signals in the data encourage further experiments with larger datasets to definitively separate lithium vs. olanzapine responders.
format Article
id doaj-art-43cbb58a40ac472bb46d522a4501decb
institution OA Journals
issn 2167-8359
language English
publishDate 2024-10-01
publisher PeerJ Inc.
record_format Article
series PeerJ
spelling doaj-art-43cbb58a40ac472bb46d522a4501decb2025-08-20T02:17:10ZengPeerJ Inc.PeerJ2167-83592024-10-0112e1784110.7717/peerj.17841Predicting maintenance lithium response for bipolar disorder from electronic health records—a retrospective studyJoseph F. Hayes0Fehmi Ben Abdesslem1Sandra Eloranta2David P. J. Osborn3Magnus Boman4Department of Psychiatry, University College London, University of London, London, United KingdomDepartment of Psychiatry, University College London, University of London, London, United KingdomDivision of Clinical Epidemiology, Department of Medicine Solna, Karolinska Institutet, Stockholm, SwedenDepartment of Psychiatry, University College London, University of London, London, United KingdomDepartment of Psychiatry, University College London, University of London, London, United KingdomBackground Optimising maintenance drug treatment selection for people with bipolar disorder is challenging. There is some evidence that clinical and demographic features may predict response to lithium. However, attempts to personalise treatment choice have been limited. Method We aimed to determine if machine learning methods applied to electronic health records could predict differential response to lithium or olanzapine. From electronic United Kingdom primary care records, we extracted a cohort of individuals prescribed either lithium (19,106 individuals) or olanzapine (12,412) monotherapy. Machine learning models were used to predict successful monotherapy maintenance treatment, using 113 clinical and demographic variables, 8,017 (41.96%) lithium responders and 3,831 (30.87%) olanzapine responders. Results We found a quantitative structural difference in that lithium maintenance responders were weakly predictable in our holdout sample, consisting of the 5% of patients with the most recent exposure. Age at first diagnosis, age at first treatment and the time between these were the most important variables in all models. Discussion Even if we failed to predict successful monotherapy olanzapine treatment, and so to definitively separate lithium vs. olanzapine responders, the characterization of the two groups may be used for classification by proxy. This can, in turn, be useful for establishing maintenance therapy. The further exploration of machine learning methods on EHR data for drug treatment selection could in the future play a role for clinical decision support. Signals in the data encourage further experiments with larger datasets to definitively separate lithium vs. olanzapine responders.https://peerj.com/articles/17841.pdfBipolar disorderLithiumMaintenance response predictionMachine learningRetrospective study
spellingShingle Joseph F. Hayes
Fehmi Ben Abdesslem
Sandra Eloranta
David P. J. Osborn
Magnus Boman
Predicting maintenance lithium response for bipolar disorder from electronic health records—a retrospective study
PeerJ
Bipolar disorder
Lithium
Maintenance response prediction
Machine learning
Retrospective study
title Predicting maintenance lithium response for bipolar disorder from electronic health records—a retrospective study
title_full Predicting maintenance lithium response for bipolar disorder from electronic health records—a retrospective study
title_fullStr Predicting maintenance lithium response for bipolar disorder from electronic health records—a retrospective study
title_full_unstemmed Predicting maintenance lithium response for bipolar disorder from electronic health records—a retrospective study
title_short Predicting maintenance lithium response for bipolar disorder from electronic health records—a retrospective study
title_sort predicting maintenance lithium response for bipolar disorder from electronic health records a retrospective study
topic Bipolar disorder
Lithium
Maintenance response prediction
Machine learning
Retrospective study
url https://peerj.com/articles/17841.pdf
work_keys_str_mv AT josephfhayes predictingmaintenancelithiumresponseforbipolardisorderfromelectronichealthrecordsaretrospectivestudy
AT fehmibenabdesslem predictingmaintenancelithiumresponseforbipolardisorderfromelectronichealthrecordsaretrospectivestudy
AT sandraeloranta predictingmaintenancelithiumresponseforbipolardisorderfromelectronichealthrecordsaretrospectivestudy
AT davidpjosborn predictingmaintenancelithiumresponseforbipolardisorderfromelectronichealthrecordsaretrospectivestudy
AT magnusboman predictingmaintenancelithiumresponseforbipolardisorderfromelectronichealthrecordsaretrospectivestudy