Automated Decision Support For Community Mental Health Services Using National Electronic Health Records: Qualitative Implementation Case Study

Yasmin van Kasteren; Jörg Strobel; Tarun Bastiampillai; Ecushla Linedale; Niranjan Bidargaddi

doi:10.2196/35403

Automated Decision Support For Community Mental Health Services Using National Electronic Health Records: Qualitative Implementation Case Study

JMIR Hum Factors. 2022 Jul 5;9(3):e35403. doi: 10.2196/35403.

Authors

Yasmin van Kasteren^{1

2}, Jörg Strobel³, Tarun Bastiampillai¹, Ecushla Linedale⁴, Niranjan Bidargaddi^{1

2}

Affiliations

¹ Flinders Health and Medical Research Institute, College of Medicine and Public Health, Flinders University, Adelaide, Australia.
² Flinders Digital Health Research Centre, Flinders University, Adelaide, Australia.
³ Barossa Fleurieu Adelaide Hills Local Health Network, South Australia, Australia.
⁴ Health Translation SA, South Australian Health and Medical Research Institute, Adelaide, Australia.

PMID: 35788103
PMCID: PMC9297136
DOI: 10.2196/35403

Abstract

Background: A high proportion of patients with severe mental illness relapse due to nonadherence to psychotropic medication. In this paper, we use the normalization process theory (NPT) to describe the implementation of a web-based clinical decision support system (CDSS) for Community Mental Health Services (CMHS) called Actionable Intime Insights or AI². AI² has two distinct functions: (1) it provides an overview of medication and treatment history to assist in reviewing patient adherence and (2) gives alerts indicating nonadherence to support early intervention.

Objective: Our objective is to evaluate the pilot implementation of the AI² application to better understand the challenges of implementing a web-based CDSS to support medication adherence and early intervention in CMHS.

Methods: The NPT and participatory action framework were used to both explore and support implementation. Qualitative data were collected over the course of the 14-month implementation, in which researchers were active participants. Data were analyzed and coded using the NPT framework. Qualitative data included discussions, meetings, and work products, including emails and documents.

Results: This study explores the barriers and enablers of implementing a CDSS to support early intervention within CMHS using Medicare data from Australia's national electronic record system, My Health Record (MyHR). The implementation was a series of ongoing negotiations, which resulted in a staged implementation with compromises on both sides. Clinicians were initially hesitant about using a CDSS based on MyHR data and expressed concerns about the changes to their work practice required to support early intervention. Substantial workarounds were required to move the implementation forward. This pilot implementation allowed us to better understand the challenges of implementation and the resources and support required to implement and sustain a model of care based on automated alerts to support early intervention.

Conclusions: The use of decision support based on electronic health records is growing, and while implementation is challenging, the potential benefits of early intervention to prevent relapse and hospitalization and ensure increased efficiency of the health care system are worth pursuing.

Keywords: EHR; adherence; algorithm; algorithms; automated alert; automated alerts; computerised clinical decision system; decision support; decision system; eHealth; electronic health records; health record; implementation; medication; medication adherence; mental health; mental illness; monitoring; nonadherence; normalization process theory; participatory action framework; psychotropic; psychotropic medication; web-based.

©Yasmin van Kasteren, Jörg Strobel, Tarun Bastiampillai, Ecushla Linedale, Niranjan Bidargaddi. Originally published in JMIR Human Factors (https://humanfactors.jmir.org), 05.07.2022.