Mark Beaumont MD

January 12, 2022



Executive summary: 

Medical “hotspots” can be defined from a genetics perspective as an area of a gene that has a high rate of mutations or from an infectious disease epidemiological perspective, a region that has an elevated prevalence, higher transmission or higher probability of disease emergence. The phenomenon of high frequency users of emergency department services fits the definition of a medical hotspot. The emergency department (ED) is a viable and necessary option for patient care services when immediate care is a necessity. On the other hand, the open accessibility of the ER often creates excessive use of its services in non-urgent situations which can lead to elevated health care costs and poor health outcomes. This challenge to the US healthcare system indicates that visits by frequent users cannot simply be dismissed as inappropriate. Strategies to reduce ED visits must account for the overall health status of frequent users and the fact that multiple EDs are often involved. By using a multidisciplinary team approach, a group of providers led by ED physicians, can develop and implement a quality improvement program to reduce ED visits, subsequent hospitalizations and associated costs with the goal to improve the overall quality of medical care for frequent ED users.

Background and clinical significance: 

Frequent users of the emergency department are defined as patients who repeatedly show up to the emergency room for medical care. They choose to ride in an ambulance or walk into the hospital, over and over again, rather than schedule an appointment to see their own primary care physician. Frequent users of the emergency room are also known as high-utilizers of the healthcare system and are even at times referred to as “frequent flyers” highlighting the recurrent nature of their visits often times for non-urgent reasons.6, 8, 9

Frequent users of the emergency department often have complex medical, social, psychological and behavioral health needs. They are more likely to have a diagnosis of a chronic disease such as depression, diabetes, heart disease or substance abuse. They are also more likely to report a lower socioeconomic status and utilize healthcare services at higher rates, despite often having health insurance and identifying a usual source of care.2-4, 6 As ED visits continue to increase every year and health care payment models shift from the traditional volume-based fee-for-service towards value-based alternative payment models, it is critical that ED providers and the healthcare team become engaged in care coordination, particularly because the ED is the primary source of health care for many frequent users.7,8

The emergency department is an important and frequently used setting to receive health services. In 2014, more than 137 million ED visits took place in the United States.6 A small percentage of patients return to the ED frequently and this accounts for a disproportionately large share of overall visits and associated costs. Published studies estimate that in the U.S, on average, between 5 and 10 percent of patients revisit the ED four or more times per year and account for 21 percent to 28 percent of all ED visits.2-6 Many of these visits are non-urgent leading to overcrowding of the hospital, unnecessary delays in care, dissatisfaction with the healthcare system and avoidable patient harm.7-9 Since the ED is also an expensive place to receive care, ED visits can be a significant contributor to elevated and avoidable healthcare costs. It has been estimated that ED care costs two to five times as much as the same treatment delivered by a primary care physician. Eliminating revisits and inappropriate ED use can reduce health care spending by as much as $32 to $35 billion dollars each year.3-5, 9,10

As a basis of comparison, to provide context, it is important to consider the statistics of frequent ED users in the state of Massachusetts. In a 2003 study it was reported that only 24% of MA residents made 1 ED visit.1 An even smaller fraction of residents, only 1%, made 5 or more ED visits which represented 17.6% of all ED visits in MA.1 A large variation also existed in the number of distinct facilities visited by an individual, with patients visiting anywhere from 1 to 43 hospital EDs in the 12 month period.For the most frequent users 42% visited 1 ED, 34% visited 2, 15% visited 3, 5% visited 4 and 4% visited 5 or more.Female patients had a slightly higher rate of frequent ED utilization (1.0% vs. 0.9% of residents).The age of the patient had a much larger impact on the likelihood of being a frequent ED user. Among MA residents, only 0.3% of children younger than 15 years old visited an ED 5 or more times in FY 2003 compared to 1.5% of patients older than 65.  Frequent users were more likely to be non-white or Hispanic, with lower levels of education and lower income.Uninsured residents of MA were more likely to make any emergency room visits; 44% visited the ED between 1 to 4 times. However, 2.1% of uninsured users visited an ED on 5 or more occasions, comparable to the 2.1% of Medicaid enrollees and 2.0% of Medicare enrollees who were also frequent users. 

MA frequent users as a whole appear to be sicker than infrequent users, and thus one might expect them to need more ED visits than others. They have higher rates of mortality at the last ED visit (2.6% versus 1.1%), are substantially more likely to receive more resource-intensive visits (which is reflected in higher mean charges for outpatient visits, $812 versus $782), and are more likely to be admitted to the hospital (18.6% versus 12.1%). 1

Research plan and specific aims:

Historically programs to reduce ED usage and utilization among frequent ED users have employed several key factors which include educating the patient of the natural history of their medical illness and teaching them how to navigate the healthcare system. It also includes the development of a comprehensive disease treatment plan, led by the PCP, that is well coordinated among all providers ensuring the timely sharing of pertinent health information.9-12 For example, a gentleman with moderate persistent asthma who uses a daily controller medication and a rescue inhaler as a needed is at risk for an asthma exacerbation during the spring because pollen is a known trigger for him. His PCP is vigilant in reviewing the medical chart with the patient and prescribing refills of his inhalers including his daily allergy medication prior to the allergy season. The NP prints out a copy of his asthma action plan then reviews it with the patient and they also review the correct technique of using the inhalers with the MDI spacer. The social worker and case manager on the team have also scheduled a home visit to assess other risk factors for asthma attacks other than dust and pollen and suggest that the patient improve air quality in the living space by removing carpets and proving a voucher to purchase an air purifier.

Research has shown that intervention programs that support patients can reduce the frequency of ED visits.6-8 Community health worker programs are one of these effective interventions.12-15 They differ from traditional case management programs by utilizing community members, instead of licensed case managers or social workers, to assist with patient navigation of the healthcare system, patient education and facilitating communication among team members. CHWs are public health workers who are “trusted members of and/or have a close understanding of the community served” and have the potential to work as a catalyst to provide higher quality, more culturally competent care.18-20 CHW’s employed in the ambulatory care setting have been shown to reduce ED visits and healthcare utilization among patients with chronic illness and recent hospitalizations however there are a few peer-reviewed randomized controlled studies of the effects of CHW programs on ED visits among frequent ED users.12-15


The purpose of my grant proposal would be to conduct a randomized controlled trial of a pilot ED-based care coordination plan which includes using a CHW program targeting frequent ED users. This goal of the intervention would be to reduce ED visits, hospitalizations and associated costs among frequent ED users at Boston Medical Center, a large urban academic medical center in Boston, MA.

Study Design: 

Randomized controlled trial 

Research design and methods: 

Prior research studies have identified risk factors that drive patients’ frequent ED use. Researchers have worked to develop and establish tools to proactively identify frequent ED users. Six specific, modifiable risk factors have been selected to be used in this intervention because of the ability to address them through hospital-based interventions.1,3 

The risk factors for frequent ED use include:

  • Lack of health insurance
  • Lack of a primary care physician
  • Having a current or past medical history of psychiatric illness 
  • Having a current or past medical history of substance abuse
  • Cognitive or physical impairment
  • Difficulty understanding discharge instructions

Between the period of July 1, 2019, and December 30, 2019, the research team would identify patients from the Boston Medical Center emergency room who meet at least one of the criteria for being a frequent ED user. These criteria were:

  • Five or more prior ED visits in the previous 12 months or
  • Four or more prior ED visits in the previous 3 months or
  • Two or more prior ED visit in the previous 72 hours

Boston Medical Center is a private, not-for-profit, 487 bed, academic medical center located in Boston, MA. BMC emphasizes community-based care, with a mission to provide consistently accessible health services to all. BMC is the largest and busiest provider of trauma and emergency services in New England; the Emergency Department had 132,148 visits in 2016.

Study protocol:

The first phase of the program will include screening the identified “frequent ED users” by a research assistant using a questionnaire to assess what risk factors each patient has for frequent ED use. The research team would extract information from the patients’ medical records, including past diagnoses of any of chronic medical conditions. Records will be notated to assess what risk factors each of the high frequency ED users possesses.

The second phase of the program would include employing a team of providers led by ED physicians and nurse care coordinators would then develop and implement the pilot program. In order to identify chronic frequent users, rather than those with an isolated health event requiring multiple visits, patients would be identified with the most ED visits during both the 30-day period and the 12-month period preceding the introduction of the program. 

The goal would be to enroll at least 40-50 patients in the study. A sample size that is too small increases the likelihood of a Type II error skewing the results, which decreases the power of the study. The sample size is directly proportional to the Z-score and inversely proportional to the margin of error. Consequently, reducing the sample size reduces the confidence level of the study, which is related to the Z-score and also increases the margin of error. Half of the patients would randomly be assigned to the intervention group and the other half to traditional care, control group, which includes the patient receiving discharge instructions and a recommendation to follow up with the primary care physician. Utilization of ED services and associated costs would be studied during the first 6 months and the 12 months of the pilot program.

Research design and methods:

The intervention would consist of several key elements, one being the development of an interdisciplinary care treatment plan which will include discharge planning to guide ED care and the assignment of an ED-based CHW who will assist with care coordination, team communication and address social issues contributing to unmet health needs of frequent users. First, trained ED providers will perform a detailed chart review for all patients randomized to the intervention to identify medical and social issues driving frequent ED visits, specifically highlighting the presence or absence of the six risk factors for frequent ED that have already been identified. Then, an acute care plan will be developed with the assistance of social workers, case managers and community health workers to improve the quality, efficiency, and coordination of ED care. The plan will be developed in conjunction with the patient’s longitudinal providers including the primary care provider and specialists, to reduce variation in acute care services. The plan will then be uploaded to the electronic health record (EHR) and electronically “flagged” in a location visible to clinicians during ED encounters.

The CHW’s goals are to better engage patients with their long term providers and to help address unmet social and behavioral health needs that contribute to ED utilization. At the start of the intervention, the CHW would review each patient’s chart and conduct a standardized intake assessment to determine unmet needs that are related to physical, psychological and social aspects of care. The CHW would then communicate with patients by phone or in person, including during scheduled home visits, to address the identified needs. During working hours,  when intervention patients register on arrival to the ED, CHWs will be notified to meet with patients and provide follow-up care and community-based resources.

During ED visits, the CHW would continue to work with the frequent ED use patients to advance their acute care plan if they had previously been enrolled or enroll new patients who haven’t been enrolled previously. The CHW would assist with specific tasks tailored to each patient’s needs, such as coordinating transportation to clinic visits, providing information on local food banks, and establishing linkage to a primary care provider for patients without one. An interdisciplinary team consisting of the CHW, a physician, and a nurse care coordinator would meet weekly to discuss the needs of enrolled patients, assess progress of enrolled patients, and assign tasks for future encounters. The evaluation of this quality improvement intervention would need to be approved by the institutional review board.11-15

Data Analysis:

The primary aim of the analysis will be to assess the program’s impact on the number of ED visits, subsequent hospitalizations and overall health-care costs.  After the trial period ends, the frequency and nature of ED visits will be analyzed along with frequency of hospitalizations and associated costs for the patients in the intervention group compared to the control group. This will be accomplished by comparing the total number and frequency of visits to the ED prior to and after implementation of the program for high frequency users and comparing the data to the control group, high frequency users who were not treated. It would also be helpful to calculate the mean number of patient visits and the standard deviation for high frequency users in the treatment group and control groups. A secondary aim would be to assess the financial impact of the program on the direct and indirect costs of care per patient. This information can be used to calculate the amount of money saved by decreased ED visits. Lastly, a goal would be to determine the accuracy, validity and usefulness of using risk factors to predict frequency of ED use by making correlations, performing forecasting and running regression analysis to help practitioners proactively identify patients at a high risk for recidivism.

Specifically we would obtain and calculate, retrospectively, a statistical analysis of all ED visits, hospitalizations and average costs per patient for the intervention group and control groups by ED discharge diagnosis, age, race, sex and type of insurance (private, Medicare, Medicaid or uninsured), if the patient was a U.S citizen, education level, work status, family income and history of selected chronic medical problems such as substance abuse, congestive heart failure, low back pain, disorders of the teeth, diabetes and psychological disorders for all high frequency ED users. Statistical comparisons of demographic characteristics, ED utilization, hospitalization frequency and average costs would be based on hypothesis or t tests for continuous variables and χ2 tests for categorical variables. Utilization, demographic, and patient-level financial data would be obtained from the healthcare system’s clinical data warehouse, which includes the EHR and cost accounting data. Program costs would be identified through the hospital accounting system and cost center reports. 

 A simple regression analysis will be performed to study how the six patient characteristics predicted the frequency for ED use and if they played a statistically significant role in driving ED visits compared to what was expected. For example, I could use a dummy variable where the variable 1= patients in the treatment group and 0=patients in the control group where the equation frequency of ED use=b0  + b1 (dummy variable) + b2 * (history of mental illness) + b3 * (history of substance abuse) etc. in order to try and control for all characteristics. Also, as a dummy variable, if I knew the prior frequency for each characteristic, I could include that information as a variable as the following, frequency of ED use= b0  + b1 (dummy variable) + b2 * (history of mental illness) + b3 * (history of substance abuse) + b4 * (prior frequency) and lastly I could also create the above + b5  * (the prior frequency) * (dummy variable).  Just having the dummy variable itself captures a change in the line. The (frequency) * (dummy with shift) can give a picture with a change in the intercept and a change in the slope. The goal of the analysis would be to create the most homogenous groups as possible for the study to try and remove as much variability as possible.  By plotting out the characteristics of all of the people in the groups, I could see if someone was an outlier for one characteristic or another.

Statistical analysis will explore the relative strength of correlations and relationships between the frequency of ED use and any particular combination of characteristics or risk factors. The analysis could be used to yield an equation which shows how each risk factor, with all others equal, relatively increases the risk of ED use. Regression analysis will be helpful to study the relationships between the dependent and explanatory variables including a regression line to designate whether the relationships are strong or weak also reviewing the R2 , the coefficient of determination, as the proportion of the variance in the frequency of ED that is predictable from the six independent variable(s). The adjusted R2, which compares the descriptive power of regression models that include a diverse number of independent variables. Every predictor or independent variable, risk factor for frequent ED use, added to the model increases the R2 value.

 I would also evaluate the p-values, F statistics and t statistics for each coefficient to test for statistical significance. The F statistic is the value you get when you run a regression analysis to find out if the means or variance between the two populations are significantly different. It’s similar to the T statistic which will tell us if a single variable is statistically significant and an F test will tell you if a group of variables are jointly significant.

 The analysis would eventually be used to help with the ability to predict and forecast what patients are at the highest risk to be frequent ED users given their set of characteristics. Once the data is obtained a fair amount of time would be devoted to data visualization, the graphical representation of information and data which includes using charts, graphs, maps and other data visualization tools provide an accessible way to see and understand trends, outliers, and patterns in data.

The detailed project would approximately cost $700,000 over 3 years. As a result of decreased service utilization, it is estimated that the average direct costs of the high frequency ED users in the intervention group would be lower than the costs of the patients in the control group resulting in a substantial total annualized cost savings to the hospital. Implementation costs would be related to hiring and training of new study personnel in addition to the administrative costs to perform the statistical analysis. Accounting for additional revenue from increased capacity as a result of fewer ED visits and hospitalizations would result in a net higher return on investment.

It is in my opinion that effective solutions to the problem of frequent revisits to the ED will have to emerge from a well-coordinated, systems-level approach that includes input from administrators, providers and auxiliary healthcare team members. These might range from limited efforts to create more accessible care options such as non-ED-based after-hours care to broad-based efforts aimed at providing complex patients with resources and education to help them navigate the healthcare system and more actively engage in managing their conditions.


Selected graphs related to Boston Medical Center readmission and discharge rate, 2016-2017

Center for Health Information and Analysis;


Fuda, K, Immmekus, R. Frequent Users of MA Emergency Departments: A Statewide Analysis. Health Policy and Clinical Practice. 2006; Volume 48, No. 1.

Lin, et al. ED-Based Care Coordination Reduces Costs for Frequent ED Users. The American Journal of Managed Care. December 2017 – Published on: December 14, 2017.

Pham et al. Characteristics of Frequent Users of Three Hospital Emergency Departments. Agency for Healthcare Research and Quality, Rockville, MD. July 2017.

CDC/National Center for Health Statistics

Center for Health Information and Analysis;

Hunt KA, Weber EJ, Showstack JA, Colby DC, Callaham ML. Characteristics of frequent users of emergency departments. Ann Emerg Med. 2006;48(1):1-8. doi: 10.1016/j.annemergmed.2005.12.030.

Emergency Department Visits in Massachusetts: Who Uses Emergency Care and Why? Massachusetts Health Reform Survey Policy Brief. September 2009.

LaCalle E, Rabin E. Frequent users of emergency departments: the myths, the data, and the policy implications. Ann Emerg Med. 2010;56(1):42-48. doi: 10.1016/j.annemergmed.2010.01.032.

Sun BC, Burstin HR, Brennan TA. Predictors and outcomes of frequent emergency department users. Acad Emerg Med. 2003;10(4):320-328. doi: 10.1111/j.1553-2712.2003.tb01344.x.

Hansagi H, Olsson M, Sjöberg S, Tomson Y, Göransson S. Frequent use of the hospital emergency department is indicative of high use of other health care services. Ann Emerg Med. 2001;37(6):561-567. doi: 10.1067/mem.2001.111762.

 Colligan EM, Pines JM, Colantuoni E, Wolff JL. Factors associated with frequent emergency department use in the Medicare population. Med Care Res Rev. 2017;74(3):311-327. doi: 10.1177/1077558716641826.

Colligan EM, Pines JM, Colantuoni E, Howell B, Wolff JL. Risk factors for persistent frequent emergency department use in Medicare beneficiaries. Ann Emerg Med. 2016;67(6):721-729. doi: 10.1016/j.annemergmed.2016.01.033

Soril LJ, Leggett LE, Lorenzetti DL, Noseworthy TW, Clement FM. Reducing frequent visits to the emergency department: a systematic review of interventions. PLoS One. 2015;10(4):e0123660. doi: 10.1371/journal.pone.0123660.

Kumar GS, Klein R. Effectiveness of case management strategies in reducing emergency department visits in frequent user patient populations: a systematic review. J Emerg Med. 2012;44(3):717-729. doi: 10.1016/j.jemermed.2012.08.035.

Althaus F, Paroz S, Hugli O, et al. Effectiveness of interventions targeting frequent users of emergency departments: a systematic review. Ann Emerg Med. 2011;58(1):41-52.e42. doi: 10.1016/j.annemergmed.2011.03.007.

Okin RL, Boccellari A, Azocar F, et al. The effects of clinical case management on hospital service use among ED frequent users. Am J Emerg Med. 2000;18(5):603-608. doi: 10.1053/ajem.2000.9292.

Stergiopoulos V, Gozdzik A, Tan de Bibiana J, et al. Brief case management versus usual care for frequent users of emergency departments: the Coordinated Access to Care from Hospital Emergency Departments (CATCH-ED) randomized controlled trial. BMC Health Serv Res. 2016;16(1):432. doi: 10.1186/s12913-016-1666-1.

Shumway M, Boccellari A, O’Brien K, Okin RL. Cost-effectiveness of clinical case management for ED frequent users: results of a randomized trial. Am J Emerg Med. 2008;26(2):155-164. doi: 10.1016/j.ajem.2007.04.021.

Horwitz SM, Busch SH, Balestracci KM, Ellingson KD, Rawlings J. Intensive intervention improves primary care follow-up for uninsured emergency department patients. Acad Emerg Med. 2005;12(7):647-652. doi: 10.1197/j.aem.2005.02.015.

Baren JM, Boudreaux ED, Brenner BE, et al. Randomized controlled trial of emergency department interventions to improve primary care follow-up for patients with acute asthma. Chest. 2006;129(2):257-265. doi: 10.1378/chest.129.2.257.