key: cord-0763220-gwp4e78r authors: Perelman, Sharon C; Erde, Steven; Torre, Lynda; Ansari, Tunaidi title: Rapid Deployment of an Algorithm to Triage Dental Emergencies During COVID-19 Pandemic date: 2021-03-13 journal: J Am Med Inform Assoc DOI: 10.1093/jamia/ocab045 sha: 54a5686968f0a340c8a0dddd0422230aa5bd55b2 doc_id: 763220 cord_uid: gwp4e78r COVID-19 quickly immobilized healthcare systems in the United States during the early stages of the outbreak. While much of the ensuing response focused on supporting the medical infrastructure, Columbia University College of Dental Medicine pursued a solution to triage and safely treat patients with dental emergencies amidst the pandemic. Considering rapidly changing guidelines from governing bodies, dental infection control protocols and our clinical faculty's expertise, we modeled, built, and implemented a screening algorithm, which provides decision support as well as insight into COVID-19 status and clinical comorbidities, within a newly integrated Electronic Health Record (EHR). Once operationalized, we analyzed the data and outcomes of its utilization and found that it had effectively guided providers in triaging patient needs in a standardized methodology. This article describes the algorithm's rapid development to assist faculty providers in identifying patients with the most urgent needs, thus prioritizing treatment of dental emergencies during the pandemic. At a media briefing on March 11, 2020 , the Director-General of the World Health Organization (WHO) characterized the novel coronavirus (COVID-19) as being responsible for a worldwide pandemic [1] . This virus, transmitted mainly by respiratory droplets, is known as severe acute respiratory distress syndrome (SARS-CoV-2) [ 1 2] . It leads to infection, which may vary in patients from being asymptomatic to pneumonia, and sometimes results in death [1] . As of January 31, 2021, there have been worldwide over 102 million documented cases resulting in 2.2 million deaths, of which an estimated 26 million cases and 440,094 deaths have occurred in the United States [3] . On March 20, 2020 the mayor of New York City declared the city as the epicenter of the pandemic with 5151 cases and 29 deaths at the time [4] . The College of Dental Medicine (CDM) at Columbia University, located in New York City, has approximately 120,000 patient visits annually. It serves as the primary dental provider to a significant portion of the surrounding metropolitan area, a large population of which is underserved. 170 students, 83 postgraduate students (PG), and 40 hospital dental residents (HD) (all under the supervision of dental faculty) provide direct patient care in the dental clinics. A CDM COVID-19 task force suspended all elective procedures after considering a multitude of factors, including the city's high population density, its reliance on public transportation, and the virulence of COVID-19 infection. The CDM emergency dental clinics remained open, continuing to serve the community and reducing the need for patients to seek urgent dental care at already overburdened emergency rooms. Initially, PG students were allowed to provide https://mc.manuscriptcentral.com/jamia treatment, but shortly after, 50 faculty from specialties including General, Oral Surgery, Pediatrics, Endodontics and Orthodontics, and 9 HD residents treated emergency patients. The providers screened patients for COVID-19 symptoms; if the patient was symptomatic, febrile or had recent exposure to someone with COVID-19, non-urgent care was deferred. Learning that many of the walk-in emergency patients indeed fit the criteria to postpone treatment or manage medically, CDM rapidly deployed a telehealth strategy, piloting a Clinical Decision Support (CDS) algorithm for four weeks, ultimately deploying at full scale. The use of telehealth visits, conducted by dental faculty, quickly became essential in allowing patients to be triaged to appropriate care while at the same time limiting potential viral exposure for patients, providers, and ancillary staff. The fact that there were no documented incidents of viral transmission among faculty, staff, or residents is evidence of the strategy's effectiveness. Furthermore, following recommendations from The Center for Medicare and Medicaid Services (CMS), these actions preserved personal protective equipment (PPE), ensuring its availability for patients with emergent needs [5] . It became increasingly difficult for providers to keep up with rapidly developing guidelines for treating patients with suspected or confirmed COVID-19 and to define which dental emergencies to manage medically via telehealth. The decision to treat a patient in the dental clinic has an added layer of complexity due to the generation of droplets and aerosols during dental procedures. Aerosol-generating procedures (AGPs) increase the risk of transmission of the virus between patient and provider [2 6 7] . In response to these challenges, CDM developed a decision support algorithm which was embedded into a newly deployed, enterprise-wide, joint medical and dental (EHR) system, EPIC. This algorithm provides evidence-based best practices based on Center for Disease Control and Prevention (CDC), American Dental Association (ADA), and CDM infection control policies for safely treating patients with dental emergencies during this pandemic [8 9] . It analyzes variables that one should consider when screening patients before treatment in the dental clinic, including the nature of the patient's dental problem, COVID-19 status, and comorbidities. Decision support generates customized recommendations ranging from delaying treatment, to referral to the cough and fever clinic, to treatment with full PPE, including N95 masks and negative pressure rooms. This pilot's goal was to maximize visits to the dental clinic for patients requiring emergent care. CDM sought to do this by developing a CDS tool that guided faculty screeners to evaluate patients with dental emergencies objectively and consistently. The design of the algorithm is to stratify risk in order to identify patients who are likely to have a poor outcome if care is delayed, as well as reduce patients with non-emergent needs seeking care in the emergency rooms, preventing overuse of dental emergency services. CDM developed Acuity Levels (Table1) with corresponding recommendations as the output of the algorithm based on a decision tree (Figure 1 ). The Acuity Levels are based on a patient history of pain, swelling, bleeding, trauma, and comorbidities and are critical to delineating treatment [8] . To best establish accurate Acuity Level as related to pain, the need for more descriptive terminology describing the pain experience and impact on quality of life rather than a numeric value was identified and incorporated in the algorithm [10] . A variety of factors, including age, sex, ethnic background, sleep status, and previous experiences may contribute to individual assessments of pain [10] [11] [12] . What may be a pain level of nine for an individual may be a four for another person or for that same person in different settings. Since this is a subjective measurement, we determined that an objective and reproducible measure should be derived to establish the patient's pain assessment score. As such, the faculty screener asks the patient a series of objective questions (Figure 2) , some leading to cascading items, from which the algorithm calculates a pain assessment score of None, Moderate or Severe. This pain assessment score is a crucial factor in determining a final Acuity Level and providing guidance for treatment recommendations. Of the patients that reported pain, 80% indicated that it impacted their daily activities. with conditions such as swelling or pain (without relief by medication) [13] . Although comorbidities are self-reported at this time, the EHR's patient problem list is readily viewable, which may prompt further discussion. The comorbidities we considered included diabetes, patients undergoing chemotherapy or radiation, transplant patients, artificial cardiac valve, and immunocompromised patients [14 15 ]. Depending on the results of the dental screening algorithmic calculation, a new series of question for COVID-19 symptoms will be presented. A final recommendation considering both the dental screening and COVID-19 status guides the provider to the appropriate recommendations ( Figure 3 ). Although this paper's focus is on telehealth pre-screening, dental faculty perform screenings in the emergency dental clinic, whether the patients are walk-ins or referred based on telehealth encounters. The goal of the in-person screenings is to quickly determine the patient's COVID-19 status using the algorithm questions, allowing the faculty screener to immediately move a patient with potential COVID-19 to an isolation room where the dental screening is completed ( Figure 3A ). Evaluating the faculty screener's assessment and treatment recommendations, the algorithm calculates additional levels of precautions based on the need for treatment that may include AGPs ( Figure 3B ). In this pilot, the clinician used the algorithm; however, it was designed to enable non-clinicians to run the algorithm to determine acuity level, thus freeing up provider resources. In order to implement the algorithm, we utilized calculation functions within clinical documentation flowsheets and developed a decision tree to guide the providers through the complex process of determining the appropriate treatment pathway while adhering to the CDC, ADA, and CDM guidelines [5 8] . Each flowsheet row prompted the user to input choices, and the formulas in the background Taking these into consideration, Acuity Level IV would be satisfied with this criteria:  Then set Acuity Level to IV After implementing this new algorithm, we tracked utilization and assessed the results from the screening to determine if it was achieving its goals. Data was analyzed to determine whether the algorithm was successfully delineating patients that could postpone treatment, be managed medically, or needed to visit the clinic for an in-person evaluation. Over the course of our pilot, algorithm utilization increased during the first 4 weeks, averaging 83 per week. From 3/23/2020 to 6/4/2020 a total of 1274 telehealth screenings were performed on 1108 patients (Table 2 ). Due to organizational policies, the emergency clinic was operating at a reduced capacity with limited resources (providers, operatories, PPE, and negative pressure rooms). To ensure that we were able to treat the patients with the greatest need, CDM developed a standardized method to prioritize the needs of emergent patients over those that may be treated medically or remotely. The algorithm ultimately provided the solution, enabling consistent evaluation across all providers, standardizing acuity levels, and allowing schedulers and triage personnel to prioritized urgent needs. The overall rate of follow up visit to the dental emergency clinic post screening was 30% with or without algorithmic usage. However, when utilizing the algorithm, the rates of patients that did need emergency care based on Acuity Level was 63%. One of the primary goals of building this clinical decision tool was to optimize the usage of limited resources during the COVID-19 pandemic, in this case, access to emergency care dental services. By creating the screening tool and algorithm based on recommendations from the ADA, CDC and CDM infection control policies, we were able to shift the mix of patients that were seen for emergency care. This result shows that the use of this screening tool did meet its goals of optimizing which patients were seen in the emergency dental clinic as opposed to patients that did not require urgent care and could be managed medically or deferred until the clinics re-opened for routine care. As we enter the new normal that is a result of COVID-19, we must rethink the dental care delivery model. By virtue of the patient's age and comorbid risk factors, coronavirus exposure potentially places one at high risk for serious illness and possible death. Healthcare delivery, in general, will look very different based on the need to maintain social distancing, reduce crowding in reception areas and the physical spacing requirements to prevent transmission of the virus. In a short time, CDM was able to develop a screening algorithm that limited patients, providers and staff exposure to the virus, while at the same time identifying patients in need of emergent care and facilitating the delivery of that care. Utilization of this screening algorithm is an efficient and productive means to evaluate potential dental and oral health emergencies and may be beneficial to consider as a model for screening dental emergencies in the future. COVID-19 Virtual Press Conference The Severe Acute Respiratory Syndrome Coronavirus-2 (SARS CoV-2) in Dentistry. Management of Biological Risk in Dental Practice Coronavirus Resource Center New York mayor declares city as epicenter of pandemic. Secondary New York mayor declares city as epicenter of pandemic 2020 Center for Medicare & Medicaid Services C. CMS Releases Recommendations on Adult Elective Surgeries, Non-Essential Medical, Surgical, and Dental Procedures During COVID-19 Respinse Aerosols and splatter in dentistry: A brief review of the literature and infection control implications Transmission routes of 2019-nCoV and controls in dental practice ADA releases interim guidance on minimizing COVID-19 transmission risk when treating dental emergencies Guidance for Dental Settings Interim Infection Prevention and Control Guidance for Dental Settings During the COVID-19 Response Are Pain Beliefs, Cognitions, and Behaviors Influenced by Race, Ethnicity, and Culture in Patients with Chronic Musculoskeletal Pain: A Systematic Review The short-form McGill pain questionnaire Dental abscess death and morbidity A systematic review of dental disease in patients undergoing cancer therapy The bacteremia of dental origin and its implications in the appearance of bacterial endocarditis Data available on request. No grants or funding was received for this project AUTHOR'S CONTRIBUTIONS All authors made substantial contributions to the development, design and this project, as well as the writing of the manuscript and interpretation of the data.