Emergency Department Use by Primary Care Patients at a Safety-Net Hospital | Emergency Medicine | JAMA Internal Medicine | JAMA Network
[Skip to Navigation]
Table. Demographic, Clinical, and Utilization Characteristics of Primary Care Patients According to ED Use, 2009-2010
Table. Demographic, Clinical, and Utilization Characteristics of Primary Care Patients According to ED Use, 2009-2010
Research Letter
Feb 13, 2012

Emergency Department Use by Primary Care Patients at a Safety-Net Hospital

Author Affiliations

Author Affiliations: Section of General Internal Medicine, Department of Medicine, Boston Medical Center and Boston University School of Medicine (Drs Lasser, Kronman, and Samet), and Departments of Community Health Sciences (Drs Lasser and Samet), Biostatistics (Drs Cabral), and Epidemiology (Dr Samet), Boston University School of Public Health, Boston, Massachusetts.

Arch Intern Med. 2012;172(3):278-280. doi:10.1001/archinternmed.2011.709

In fee-for-service payment models, there are strong financial incentives for hospitals to tolerate high levels of emergency department (ED) use, including use by established primary care patients. Yet, as health care reform introduces global payment models, high levels of ED use will no longer be financially tenable. Understanding the magnitude of the problem of ED use by established primary care patients is crucial to redesigning primary care delivery and reimbursement in the United States. We had 2 objectives: (1) to characterize ED use at an urban safety-net hospital after the implementation of Massachusetts health reform, focusing on patients who had primary care providers (PCPs) and (2) to identify patterns of ED use that might inform the hospital-based primary care practices' transformation to a medical home, and eventually, to an accountable care organization.


Boston Medical Center (BMC) is an urban safety-net hospital with 8 primary care practices staffed by 105 PCPs. The practices predominantly serve a minority and low-income population. We identified patients who had 1 or more primary care visits from July 1, 2009, to July 1, 2010, and examined their ED use over this period. We defined frequent ED users as patients with 4 or more ED visits in the past year and occasional ED users as those with 1 to 3 ED visits in the past year. We used an algorithm developed by Billings et al1 to categorize each visit's principal International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis and to determine the probability that a visit required ED care. Using a validation of this algorithm,2 we defined visits as high severity if the probability that ED care was needed was 0.75 or higher for the visit's principal ICD-9-CM diagnosis. Similarly, we defined visits as low severity if the probability that ED care was needed was 0.25 or lower. Visits of indeterminate severity were defined as those with a probability higher than 0.25 and lower than 0.75. Visits classified as high severity have been found to have a strong association with future hospitalization or death.3 We performed χ2 tests and t tests to compare differences in demographics between persons with and without any ED use and with occasional vs frequent ED use. The Boston University Medical Campus institutional review board has approved this study as exempt.


During the study period (2009-2010), 39 603 patients had 1 or more primary care visits. Most patients (65.4%) did not make any ED visits, while 11 787 (29.8%) were occasional ED users and 1928 were frequent ED users (4.9%) (Table). The 11 787 patients with occasional ED use made 17 759 visits over the study period, while the 1928 patients with frequent ED use made 12 289 visits. Approximately half (49.8%) of all ED visits occurred on weekdays, while BMC primary care practices were open. Most ED visits were for low-severity conditions.


Emergency department use by primary care patients at an urban safety-net hospital was high, though most visits were of low severity. One possible reason for this is lack of access to primary care,4 with few available appointments to see a PCP. While data on time to third next available appointment, a standard measure of primary care access,5 are not available for the primary care practices during the study period, other practice metrics suggest that access may have been a problem. For example, missed primary care appointment rates were high, averaging 24.5%. High missed appointment rates are often correlated with long wait times to schedule appointments.6 In addition, monthly telephone call statistics show that only between 72.4% and 88.1% of patient telephone calls were answered by the primary care call center over the study period. It is possible that patients called the practices with an urgent problem, did not have their telephone call answered promptly, and decided to seek care in the ED instead. Indeed, 13% of telephone calls were abandoned by patients over the study period (patients called and subsequently hung up while they were kept on hold). The fact that nearly half of all ED visits took place during the hours of primary care clinic operation further suggests that appointment availability may have been an issue. In addition, a sizable minority, roughly one-fifth, of primary care is provided by residents,7 who have limited availability when they are not in clinic. It is also possible that Massachusetts health reform has affected access to primary care. As newly insured patients have entered primary care in large numbers, it is possible that access to primary care has worsened for other patients.

Massachusetts has been a bellwether for the implementation of health reform and will be a bellwether for the transformation of primary care, with the move away from fee-for-service payments and the introduction of global payments for health care. Overall ED volume has continued to increase in Massachusetts after health reform.8 It is unclear if changes in primary care practice and payment will be sufficient to reduce high levels of ED use among patients at an urban safety-net hospital.

Back to top
Article Information

Correspondence: Dr Lasser, Health/Care Disparities Research Program, 801 Massachusetts Ave, Boston Medical Center, Section of General Internal Medicine, Room #2091, Boston, MA 02118 (karen.lasser@bmc.org).

Author Contributions: Dr Lasser, the primary author, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Lasser, Kronman, and Samet. Acquisition of data: Lasser. Analysis and interpretation of data: Lasser, Kronman, Cabral, and Samet. Drafting of the manuscript: Lasser, Cabral, and Samet. Critical revision of the manuscript for important intellectual content: Lasser, Kronman, Cabral, and Samet. Statistical analysis: Lasser, Kronman, and Cabral. Obtained funding: Lasser and Samet. Administrative, technical, and material support: Lasser. Study supervision: Lasser and Samet.

Financial Disclosure: Dr Lasser was a consultant to Rise Health (June-November 2011).

Funding/Support: This study was supported by the Department of Medicine at Boston Medical Center/Boston University School of Medicine.

Previous Presentation: This work was presented at the Society of General Internal Medicine Annual Meeting; May 7, 2011; Phoenix, Arizona.

Additional Contributions: Meredith Manze D’Amore, PhD (Section of General Internal Medicine, Boston Medical Center/Boston University School of Medicine), assisted in manuscript preparation; Linda C. Rosen, BSEE (Boston University Office of Clinical Research), performed data extraction; and Maxim D Shrayer, PhD (Department of Slavic and Eastern Languages and Literatures, Boston College), and William Fernandez, MD (Department of Emergency Medicine, Boston Medical Center/Boston University School of Medicine), provided comments on an earlier draft of the manuscript.

Billings J, Parikh N, Mijanovich T. Emergency department use: the New York Story.  Issue Brief (Commonw Fund). 2000;(434):1-1211665699PubMedGoogle Scholar
Ballard DW, Price M, Fung V,  et al.  Validation of an algorithm for categorizing the severity of hospital emergency department visits.  Med Care. 2010;48(1):58-6319952803PubMedGoogle ScholarCrossref
Wharam JF, Landon BE, Galbraith AA, Kleinman KP, Soumerai SB, Ross-Degnan D. Emergency department use and subsequent hospitalizations among members of a high-deductible health plan.  JAMA. 2007;297(10):1093-110217356030PubMedGoogle ScholarCrossref
Smulowitz P, Landon BE, Burke L, Baugh C, Gunn H, Lipton R. Emergency department use by the uninsured after health care reform in Massachusetts.  Intern Emerg Med. 2009;4(6):501-50619777165PubMedGoogle ScholarCrossref
Murray M, Berwick DM. Advanced access: reducing waiting and delays in primary care.  JAMA. 2003;289(8):1035-104012597760PubMedGoogle ScholarCrossref
Lasser KE, Mintzer IL, Lambert A, Cabral H, Bor DH. Missed appointment rates in primary care: the importance of site of care.  J Health Care Poor Underserved. 2005;16(3):475-48616118837PubMedGoogle ScholarCrossref
Zallman L, Ma J, Xiao L, Lasser KE. Quality of US primary care delivered by resident and staff physicians.  J Gen Intern Med. 2010;25(11):1193-119720645018PubMedGoogle ScholarCrossref
Smulowitz PB, Lipton R, Wharam JF,  et al.  Emergency department utilization after the implementation of Massachusetts health reform.  Ann Emerg Med. 2011;58(3):225-234.e121570157PubMedGoogle ScholarCrossref