A Novel Approach to Delivering Evidence-based, High-quality Care in Psychiatry Through an Electronic Integrated Care Pathway (eICP) Pilot

Abstract

Integrated care pathways (ICPs) are evidence-based decision support tools intended to reduce variation and improve quality of care. Historically, adoption of ICPs has been difficult to measure, as the pathways were outside of the electronic health record (EHR), where care delivery documentation and orders were completed. This Technology Column describes the innovative development and implementation of a diagnosis specific electronic ICP that directly embeds pathway steps into an EHR to facilitate order sets, clinical decision-making, and usage tracking. The pathway was implemented at a seven-hospital academic medical center, and details the technology, team structure, early adoption results, and future directions. As such, the importance of investing and organizing resources to create an eICP (e.g., time, technology, and specialized teams) to provide a user-friendly experience to support early adoption is underscored. Preliminary findings show that the eICP had consistent use in the first year of implementation. This manuscript is intended to serve as a practical guide to build eICPs within behavioral health service areas across institutions.

Variations in care occur within and across psychiatric facilities, which can negatively impact quality of care [1], worsen patient satisfaction, and increase healthcare costs [2]. Integrated clinical care pathways (ICPs) have emerged as a model to reduce variation while facilitating timely, accurate clinical decision-making. An ICP provides clinicians with concrete, evidence-based clinical guidance when treating patients who present with an identified clinical diagnosis, and aims to incorporate research knowledge while reducing use of non-evidence-based decisions [3]. ICPs can be leveraged by health systems to ensure high quality and appropriate clinical standards are being utilized across settings. Use of ICPs in clinical practice can help minimize delayed discharges, optimize best treatment practices, and improve patient experience [4]. Additionally, ICPs have the potential to reduce spending by discouraging ordering of unnecessary tests and shortening length of stays [4, 5].

Despite their benefits, ICPs have had difficulties with implementation due to lack of sustained use and difficulty identifying a unified best practice [6, 7]. Historically, ICPs have been separate from the clinical workflow, with paper guidelines that are not interactive, preventing providers from incorporating ICPs consistently and easily into clinical practice [7]. Similarly, identifying a unified “best practice” can be difficult, since individuals and groups may have differing views on treatment protocols in specialty versus primary care practices. Both these challenges are important to address while developing ICPs.

To address existing implementation challenges with ICPs, a new direction for ICP development is the use of interactive, electronic ICPs (eICPs). Informatics and technology have rapidly advanced since ICPs were first introduced, and ICPs can now be directly integrated into clinical workflows through the electronic health record (EHR). In fact, eICPs allow for real-time, interactive decision-making and clinical decision support (i.e., orders, labs, order sets, calculators, pre-populated medication dosing), and are more effective and easier to access than paper-based ICPs. In addition, eICPs have increased providers’ use of care pathways and facilitate more precise documentation of patients’ outcomes [8, 9].

Although eICPs have been described in medical settings, few articles have described its application in a real-life setting for psychiatric care delivery [6, 10]. In this column, we describe the development and application of an eICP pilot for the diagnosis and treatment of Acute Mania in Bipolar Disorder. We focus on feasibility, proof-of-concept, and adoption, and provide early findings on use patterns for the eICP pilot.

Development of an eICP Pilot for Acute Mania Management

An eICP pilot was developed and used at a large seven-hospital academic medical center for application in the emergency room and inpatient psychiatric settings. Acute mania was selected as the first pathway, since the clinical management for acute mania in bipolar disorder is well-defined (e.g., use of lithium as a first-line agent). However, the team still identified practice variability in the starting doses of mood stabilizers and management of insomnia in bipolar disorder. More nuanced variation was found with guidance on referral to electroconvulsive therapy and treatment of pregnant patients. 

Building the Team and Infrastructure for eICPs

As part of the broader goal of the academic medical center, a distinct, administrative framework was created to advance a “Care Signature” across hospitals within the same health system. The framework included a dedicated Care Signature team, which was led by a senior physician with significant experience building clinical pathways and included team members across disciplines and subspecialties. The team included clinical leads (physicians, advanced practice providers, and pharmacists) with protected time to work on eICPs, including weekly trainings on building, iterating, and executing eICPs. A clinical lead from the main Care Signature team was designated to each subspecialty service line to help organize and build eICPs.

In the Psychiatry service line, both a larger leadership group Care Signature Council (CSC) as well as a smaller, expert-led group Clinical Consensus Group (CCG) were formed. The CSC oversaw the prioritization, facilitation, and governance of all pathway-related work within Psychiatry and convened monthly. CSC membership consisted of the department chair, a senior hospital leader, section chiefs, system pharmacists, and leaders from across the health system. The CCG was formed to focus on one specific pathway and consisted of approximately 15 clinicians with relevant clinical experience.

Our Acute Mania Clinical Consensus Group (CCG) team consisted of inpatient and emergency psychiatrists, a pharmacist, resident physician, registered nurse, and clinical experts in women’s mental health and electroconvulsive therapy, who collaborated during weekly meetings to develop the pathway prototype. Initially, the team mapped clinical steps in care and identified steps for which variation existed due to lack of consensus. The CCG then divided into smaller teams to perform literature reviews and develop consensus statements for these steps, which included recommendations for efficacious, safe, and cost-effective interventions.

The team focused on actionable steps that would directly influence clinical care—ordering laboratory tests, recommending starting and loading doses of medications and titration schedules, identifying side effects that impact treatment selection, providing indications for referral to specialists in pregnant patients, and escalating care to interventional psychiatry service when warranted. The team also incorporated recommendations for clinical measurement during hospitalization and selected the Young-Mania Rating Scale to embed into the pathway for assessing clinical symptom change during hospitalization. See Appendix 1, eICP Bipolar Disorder (Acute Mania).

Leveraging the Electronic Health Record (EHR) to Support Care Delivery

Once the CCG team reviewed the literature and devised all relevant consensus statements, the eICP was built to reflect the recommended action steps. To support the work, a third-party software company (AgileMD, San Francisco, CA) was utilized to integrate the eICP directly into the EHR. The external software is a graphical tool that enables clinicians to create and deploy eICPs in an interactive flowchart format. Functionality included text or image pop-up windows, links to other portions of the EHR (e.g., documentation or medication administration), and the ability to link orders directly from the pathway. AgileMD was easy to use and facilitated the creation of rapid drafts of pathway steps. Further, clinical experts and team members could quickly provide feedback. The eICP was then linked directly to the EHR and took advantage of EHR capabilities to include orders, default order settings, and knowledge synthesis to improve clinician efficiency, decrease cognitive burden, and enhance patient care. Notably, prior optimization requests directly in the EHR could take months to implement, and AgileMD allowed for more rapid development.

One example is ordering valproic acid, which involves a loading dose followed by a maintenance dose and a serum level test ordered four days in the future. The eICP linked to a custom-built order set that contained all three orders with default settings that required six clicks and no additional typing. In contrast, previously for the same orders, a clinician needed to use at least 14 clicks and manual typing in multiple fields. The additional fields can lead to added time spent on creating orders, additional risk for error, and inefficiencies due to team members downstream needing clarification (i.e., pharmacist calling the psychiatrist to ask about the order). A second example is the default of a higher initial lithium dose: 450 mg twice a day (BID) instead of 300 mg BID, to achieve faster acute mania symptom resolution and attempt to shorten length of stay. Thirdly, a table was created of antipsychotic medications with evidence for managing acute mania, recommendations for usage (including titration schedules and lab monitoring suggestions), and contraindications. The aggregation and concise presentation of medication information obviated the need for clinicians to rely on memory or perform multiple discrete searches, which may be incomplete or disruptive to clinical workflow.

Iterative Development and Launch of the eICP

After the CCG team agreed on the draft prototype for the clinical pathway, the eICP was then reviewed with the Psychiatry CSC, which allowed leaders and a broad stakeholder group to provide input. In addition, the eICP was reviewed with medical staff and resident physicians to solicit feedback and promote utilization, before being linked and launched within the EHR. Each eICP incorporated feedback from the CCG team on appropriate quality metrics that can highlight the impact of the clinical pathway. For all pathways, an electronic clinical dashboard was created to monitor eICP utilization as a process metric and can identify individualized use by role and name. Once available in the EHR, the eICP was available for clinicians so that the eICP was accessible throughout the health system, thus driving consistency in practice across care teams and sites. The smaller CCG group met at least weekly to develop consensus statements, and the CSC met monthly. The total time for development and pathway build took approximately 5 months, given that it was the first care pathway built within the psychiatric setting. Concurrently, as the pathway was launched, the CCG and CSC provided education and communication about the eICP to multiple clinical staff members, residents, and faculty (e.g., resident specific training meetings, Grand Rounds, medical staff meetings, departmental and hospital newsletters).

Initial Results and Clinician Feedback

The eICP for Acute Mania management went live on May 27, 2021. Preliminary results from the first year indicate that the Acute Mania eICP had been used in 374 sessions, on average, once per day. Similarly, the average number of encounters and providers who have used it per day is one. The greatest number of sessions completed in a day (13 total) using the eICP was reported on June 4, 2021. Our initial pilot data shows that the eICP has been utilized by all but one hospital offering inpatient level of care within the same health system. We found that providers in the emergency psychiatry setting utilized the eICP more often than those on inpatient units.

We obtained direct feedback from members in both CCGs and CSCs suggesting that team members were very engaged and interested in pathway building to improve clinical care and reduce variation across locations. Despite external factors such as COVID-19 occurring at the same time, the CCGs and CSCs met regularly to work on the eICP.

Challenges and Lessons Learned

The development and implementation of an eICP for Acute Mania management demonstrates an early example of incorporating a clinical pathway directly into an EHR for timely, meaningful use of clinical decision support and more efficient care. The utilization data of Psychiatry’s first clinical pathway suggests that the eICP has spread through different settings with relatively consistent use per day. Notably, the eICP development process involved several challenges, such as significant investment of resources, including time, expert knowledge, review of literature, infrastructure development, dedicated staff, and licensing of an external software partnership. Although this first pathway required 5 months to build, a substantial portion of time was dedicated to developing the processes and governance infrastructure. More recently, the average care pathways are requiring 2–4 months to construct. Our work suggests that after initial investments are made in the processes and governance, pathways can be developed more quickly by a core team focused on pathway building, with specialty members invited as subject matter experts. When the Acute Mania eICP was initially launched, it could not be linked to the EHR storyboard through patient-specific suggestion criteria. Nevertheless, this feature is now available using a software link between AgileMD and the EHR.

There were several lessons learned. Team members participated actively throughout the eICP development process, suggesting that the work was meaningful and allowed for active engagement toward a collective common resource for care delivery. Individuals felt that they contributed to improving patient care directly, unlike traditional quality improvement projects, which often had regulatory and compliance components not always associated with direct clinical care implications. Additionally, the structured sponsorship and engagement by both health system and departmental leaders directly influenced the direction and completion of the project. The eICP process also had significant visibility in the department and hospitals highlighting the importance of the initiative toward improving care. Furthermore, the broad stakeholder engagement facilitated system adoption across multiple hospitals and use of the eICP by different disciplines.

The health system-wide Care Signature group identified and used several key principles for pathway building. Pathways were intentionally built to incorporate action steps at each decision-making point in the care delivery process. Although clinical literature is important in guiding caveats to practice and rare side effects of medications, creating actionable steps and decision points helped leverage both the literature’s evidence-based practice recommendations as well as local clinical expertise. In addition, eICPs require careful consideration and balance of content: too much content in one eICP can lead to key information being buried and not directly visible, and too little content can make the eICP less useful by omitting details that impact clinical decision-making. The value of eICPs included the ability to build order sets and incorporate clinical calculators, which can save time for clinicians, improve efficiency, and minimize variation in practice. Finally, the Care Signature group noted that scheduling of group members can impact project momentum, so the group developed templated kickoff meeting and utilized note taking and video recordings for shared understanding about project goals.

Conclusions

The eICP concept builds on existing ICPs by incorporating a user-friendly experience aimed at optimizing patient care. The Bipolar Disorder (Acute Mania) eICP provides an illustration of a pilot that has been implemented in a large academic medical center. Its development and implementation demonstrate that resources, personnel, and infrastructure were all needed to help drive the development of the eICP. Although initially intensive to build, eICPs are a helpful tool to improve clinical workflows and facilitate decision-making by consolidating information for real-time use. This pilot suggests that eICPs can be used as a novel approach to providing just-in-time evidence-based practices in psychiatric settings. The eICPs also act as an educational tool to help familiarize clinicians with a health system’s unified “Care Signature” approach to treating a particular clinical condition across multiple hospitals, thus codifying best practice with a goal of reducing unsafe or unnecessary clinical variation.

Future directions will need to measure clinical outcomes and changes in variation and go beyond initial feasibility and utilization. While the model and infrastructure to create eICPs exists today, the challenge will be to build a library of care pathways across many common psychiatric conditions while maintaining updated evidence-based practices.

Luming Li, M.D., M.H.S.

is currently the Chief Medical Officer for The Harris Center for Mental Health and IDD. As the Chief Medical Officer, Dr. Li is responsible for medical staff oversight and direct leadership of clinical quality and safety. Her primary goals professionally are to improve care for patients with severe psychiatric conditions. She actively practices in the outpatient and psychiatric emergency settings. Her research and educational interests focus on administrative psychiatry and leadership, quality and safety, and clinical service delivery. She has written more than 35 articles, book chapters, and peer-reviewed journal articles about delivering psychiatric services during the COVID-19 pandemic, quality improvement in health systems, psychiatric leadership, and caring for mentally ill individuals in complex systems of care. 

Dr. Li completed a 7-year B.A./M.D. program at Rutgers/Robert Wood Johnson Medical School, and residency training and a M.H.S. at Yale School of Medicine and Yale New Haven Hospital. She has also served on national committees within the American Psychiatric Association, including the Health Systems and Financing Committee (2017-2018), and Innovation Committee (2020-current). She was previous recipient of the APA’s Public Psychiatry Fellowship, the Health and Aging Policy Fellowship, and the American Political Science Association Congressional Fellowship. Prior to joining The Harris Center, Dr. Li served as the Medical Director of Quality Improvement of the Yale New Haven Psychiatric Hospital and the Medical Director of Clinical Operations at Yale New Haven Health System (2018-2021). She currently is the chair of the Texas Council Medical Directors Consortium, a member for the Technical Expert Panel for several MIPS Measures, Epic Behavioral Health Steering Board, and an executive board member for the American Association of Psychiatric Administrative Leaders. She maintains active faculty positions at Baylor School of Medicine, UTHealth Houston, and Yale School of Medicine.

Appendix

The Appendix shows the pathway that directly is visible in the electronic health record.

Declarations

Ethical Considerations

This study was reviewed by the institutional review board at Yale University and received an exemption determination.

Conflict of Interest

The authors declare that they have no conflict of interest.