Development and implementation of a clinical decision support system tool for the evaluation of suspected monkeypox infection

Abstract

Monkeypox virus was historically rare outside of West and Central Africa until the current 2022 global outbreak, which has required clinicians to be alert to identify individuals with possible monkeypox, institute isolation, and take appropriate next steps in evaluation and management. Clinical decision support systems (CDSS), which have been shown to improve adherence to clinical guidelines, can support frontline clinicians in applying the most current evaluation and management guidance in the setting of an emerging infectious disease outbreak when those guidelines are evolving over time. Here, we describe the rapid development and implementation of a CDSS tool embedded in the electronic health record to guide frontline clinicians in the diagnostic evaluation of monkeypox infection and triage patients with potential monkeypox infection to individualized infectious disease physician review. We also present data on the initial performance of this tool in a large integrated healthcare system.

INTRODUCTION

In early May 2022, a cluster of monkeypox cases was initially identified in Great Britain, with the first identified case in the United States later that month.¹ The extent of the global monkeypox outbreak has continued to spread since, affecting tens of thousands of individuals in more than 80 nonendemic countries by mid-August 2022.² Prompt recognition of monkeypox infection is important not only to expedite potential initiation of therapy but also to institute transmission-based precautions to minimize risks of monkeypox transmission within healthcare facilities. As knowledge about the epidemiological risk factors and clinical presentation of people with monkeypox infection has grown during this outbreak, the definition of a person under investigation (PUI), a key element in driving diagnostic evaluation, has evolved, as have testing and infection control protocols. This rapidly evolving knowledge necessitates prompt dissemination of information to frontline clinicians as well as iterative adaptation of clinical resources to remain up to date with dynamic local and national guidelines.

Clinical decision support systems (CDSS), especially those computerized products that provide automated decision support as part of clinician workflow, have been shown to improve adherence to guidelines.³ In our large integrated health system, 2 years after the initial development of a COVID-19 CDSS, isolation, testing, and deisolation for patients with suspected and confirmed COVID-19 are largely automated, with only a small subset of the most complex cases triaged to clinical or infection control review.⁴ Building on this experience, we sought to create a CDSS to assist frontline clinicians in the evaluation of patients presenting with symptoms concerning for monkeypox. Here, we describe the development of the Monkeypox (MPX) Clinical Decision Support Tool, a dynamic CDSS embedded in the electronic health record (EHR) and capable of adapting to the outbreak as it develops. We detail major innovations and provide initial outcomes of monkeypox PUIs evaluated during early deployment.

MATERIALS AND METHODS

Massachusetts General Hospital (MGH) is a large academic medical center in Boston, MA, that was the site of the first case of confirmed monkeypox infection in the United States on May 18, 2022.¹ Immediately following that first diagnosis, frontline clinicians were instructed to contact a team of infectious disease physicians (ie, the MGH Biothreats Team) who have expertise in the evaluation of high-consequence infectious diseases when patients presented for care with a new rash and epidemiological risk factors for monkeypox infection. Biothreats Team physicians then individually reviewed patient details with the frontline clinicians and facilitated additional diagnostic workup for monkeypox at the Massachusetts Department of Public Health, when indicated.

Informed by these early evaluations of potential monkeypox PUIs, and based on prior experience in developing and implementing CDSS for COVID-19, Biothreats Team physicians collaborated with MGH infection control leadership and members of the Mass General Brigham (MGB) Digital Health Team to develop and implement the MPX Clinical Decision Support Tool into the healthcare system’s electronic health record (Epic Systems Inc., Verona, WI, USA). MGB is a large integrated healthcare system in Massachusetts and New Hampshire, which includes MGH, as well as 8 acute care hospitals, 2 specialty hospitals, and a broad network of postacute care, community health, and primary care centers. Across the system, the MPX tool was launched as an on-demand template that could be accessed by clinicians by typing “.MONKEYPOX” in any inpatient or outpatient note field.

The MPX CDSS tool consisted of a structured set of screening questions adapted from US Centers for Disease Control and Prevention (CDC) criteria for establishing PUI status for monkeypox infection.⁵ Based on a clinician’s answers to questions presented, if a patient met both epidemiological and clinical criteria for MPX PUI status, a secondary set of questions developed by Biothreats Team members was automatically added to the note to prompt the clinician to capture more data on individualized epidemiological risk for each patient. These questions included risk related both to the current monkeypox outbreak and to the classical modes of monkeypox transmission. A detailed description of the cascading questions and prompts is provided (Supplementary Figures S1–S11).

Clinicians were additionally prompted in the MPX CDSS tool to upload images of the patient’s rash into the EHR to aid in virtual review by infectious disease and public health officials. The tool also included directions regarding immediate infection control protocols. Finally, to ensure the outcome of the PUI assessment was communicated across clinicians caring for the patient, the tool automatically added a label, called an infection status, to the patient record to indicate they were a PUI for monkeypox (“MPX-Risk”) upon completion of the note. The MPX-Risk infection status, similar to other infection statuses used to communicate the need for specific transmission-based precautions, can be seen by all users of the EHR, drives other automated systems (eg, isolation orders), and feeds into data and analytics across the EHR.

On June 8, 2022, the MPX CDSS tool was launched across MGB with the ultimate goal of standardizing evaluation and management of monkeypox PUIs throughout the system. At each site, once the clinician completed the note, facility-specific instructions on the next point of contact (ie, infection control, infectious diseases, etc.) were provided based on facility-specific workflows. We evaluated outcomes between June 8, 2022, and July 20, 2022, including overall utilization, determination of PUI status, testing frequency, and test positivity among patients evaluated with the MPX CDSS tool. Records for patients with visits documented in the MGH Sexual Health Clinic, subject to 42CFR protection, were excluded from this analysis. Over the course of the study period, several updates were made to capture the evolving PUI definition, collect information to facilitate decisions regarding home isolation, and to modify the CDSS tool for end-user clarity.

This study was conducted under MGB IRB protocol 2012P002359.

RESULTS

MPX tool utilization

Between June 8, 2022, and July 20, 2022, the MPX CDSS tool was used for 55 distinct patients. Utilization within the MGB system occurred predominantly within the MGH with 39 encounters. Other MGB system evaluations included 5 at Brigham & Women’s Hospital, 5 at Newton-Wellesley Hospital, 1 at Cooley-Dickinson Hospital, 1 at Salem Hospital, and 4 across several urgent care settings. More than half of these completed encounters took place in an Emergency Medicine context (37), with most of the remainder occurring either in primary care settings (8) or MGH infectious diseases clinic (6). The predominant use of the tool at MGH reflects overall volume of distribution of PUIs across the MGB healthcare system, which was focused at MGH (Figure 1).

Figure 1.

Care settings in which the MPX CDSS tool was used. The distribution of care settings in which the tool was utilized during its first 6 weeks of implementation is shown. CDSS: clinical decision support system; ID: infectious diseases; MPX: Monkeypox.

Descriptive characteristics of PUIs

The CDC PUI definition used to identify PUIs is provided (Supplementary Table S1). Among 55 patients presenting with an unexplained rash for which the tool was used, 31 (56%) were identified by the tool to be possible PUIs. High rates of PUI identification were seen in the MGH Infectious Diseases Clinic (5 PUIs identified among 6 uses of the MPX tool, 83%). The PUI identification rate in an Emergency Medicine setting was 16/37 (43%), while 3/8 (38%) patients in primary care settings were determined to be possible PUIs for monkeypox. The average age among possible PUIs was 34.7 [standard deviation (SD) 8.6] compared to 41.0 (SD 15.1) for patients who were not identified as possible PUIs. Twenty-six possible PUIs identified as men, 1 as a transgender man, and 4 as women.

All patients identified as possible PUI, by definition, met the clinical PUI criterion of having an otherwise unexplained rash. Among the 31 patients who also met epidemiological criteria and thus were identified by the MPX CDSS tool as possible PUIs for monkeypox, 24 reported close contact with members of a social network experiencing monkeypox, 7 reported contact with a person with a similar rash or who was thought to have monkeypox, and 2 reported travel to an area of active monkeypox transmission. No patients reported contact with animals or animal products as described with monkeypox transmission in endemic settings.

Testing outcomes

Among the 31 patients identified as possible PUIs by the MPX CDSS tool, testing was not pursued for 7 PUIs assessed as low risk after specialist review at the hospital or in discussion with the state epidemiologist. Of the 24 patients for whom monkeypox testing was pursued, 5 (21%) were PCR positive, and 15 (63%) were PCR negative. Two (8%) patients had inconclusive tests results, generally indicating a lack of amplification from the assay positive control; neither of these was retested. Results from 1 patient (4%) were not recorded in the medical record, and 1 patient (4%) was lost to follow-up prior to testing. We are aware of an additional 3 patients who tested positive for monkeypox at MGH during this period who were not triaged through the MPX tool; all presented through the infectious diseases clinic. The average age of patients who tested positive for MPX was 33.9 (SD 8.5) years; all were male. A summary of the MPX CDSS tool utilization and downstream outcomes by week over the first 6 weeks of implementation is presented (Figure 2).

Figure 2.

Trends in use of the MPX CDSS tool. Bars depict the number of patients for whom the tool was used (total use), the number of possible PUIs identified by the tool (CDSS PUI), the number of possible PUIs in whom testing was pursued after expert review (Testing Pursued), and the number of patients who tested positive (MPX positive) during each of the first 6 weeks of implementation. CDSS: clinical decision support system; MPX: Monkeypox; PUI: person under investigation.

Versioning

During the study period, updates occurred as expected to respond to changing epidemiology, infection control protocols, as well as end-user feedback to improve utilization. While the tool can be added to individual clinicians’ notes or their personal library of note templates, it was constructed as a central system template in the EHR to maintain consistency across users. As updates were requested by the Biothreats Team to adapt to state or national guidance or changes in the epidemiology of the outbreak, these changes were automatically pushed to all users of the note simultaneously (Figure 3).

Figure 3.

Versioning of the MPX CDSS tool. The timeline showing major milestones prompting development of the MPX CDSS tool (eg, first diagnosed case at MGH on May 18, 2022), followed by request for development and implementation within 24 h of request on June 8, 2022. The top half of the timeline describes multiple changes to the PUI definition and IC protocols, while the bottom half describes versioning related to changes in the CDS to allow for more efficient collection of pertinent patient information, as well as ensuring access by clinicians to the most up to date public health guidance and local points of contact. CDC: Centers for Disease Control and Prevention; CDSS: clinical decision support system; IC: infection control; MPX: Monkeypox; PUI: person under investigation.

DISCUSSION

We have described the development, implementation, and early usage patterns of an embedded CDSS to facilitate the initial isolation and evaluation of patients in whom monkeypox infection is suspected. Once the decision was made to develop this CDSS, we were able to implement the MPX CDSS tool across the MGB system in a matter of days, and over the first 6 weeks, this tool was utilized 55 times for the evaluation of monkeypox infection in our healthcare system.

Importantly, this CDSS allows for rapid and frequent iterations with the input of subject matter experts based on evolving epidemiological risk factors, testing strategies, and to improve the efficiency of diagnostic evaluation. Improvements such as automated patient chart labeling to identify PUIs and thus alert clinicians of the monkeypox work up in progress, and the ability to update front-line clinicians of infection control protocols in real time, have provided the flexibility required in an emerging outbreak.

Several limitations are noted. The use of the MPX CDSS tool requires clinician buy-in as there is no forcing function within the EHR to compel use, and this can lead to missed opportunities for tool application. Despite this, with messaging to frontline clinicians and attention to integration into standard workflow, clinician acceptance of the tool can be achieved. In this study, we did not aim to evaluate the sensitivity or specificity of the MPX CDSS as a clinical decision tool; MPX CDSS tool questions were derived from CDC definitions, but more data will be needed to validate the test characteristics of the instrument.

This approach has facilitated up-to-date decision support to frontline providers by centralizing MPX tool updating within a team of subject matter experts, including feedback from end users to improve functionality. Though described here only for monkeypox, developing similar tools for the evaluation of other high-consequence infectious diseases such as Middle East respiratory syndrome, Ebola virus disease, and avian influenza may also help to facilitate efficient and adaptable evaluation of PUIs. In doing so, frontline providers will have valuable frameworks of support for identifying, isolating, and informing approaches^6–9 in evaluating patients amid rapidly emerging infectious disease outbreaks.

FUNDING

This work was supported by the US Assistant Secretary for Preparedness and Response (6 U3REP150548-05-08 to ESS).

AUTHOR CONTRIBUTION

All authors contributing to the manuscript approve of this submission.

SUPPLEMENTARY MATERIAL

Supplementary material is available at Journal of the American Medical Informatics Association online.

Data Availability

The data underlying this article cannot be shared publicly due to the privacy of individuals that participated in the study.