Emergency Department Use: A Reflection of Poor Primary Care Access?

March 30, 2015
Daniel Weisz, MD, MPA

,
Michael K. Gusmano, PhD

,
Grace Wong, MBA, MPH

,
John Trombley II, MPP

Volume 21, Issue 2

An original emergency department patient survey, insurance claims data, and administrative records are used to examine the characteristics of nonurgent users.

ABSTRACT

Objectives

To determine whether the use of the emergency department (ED) for nonurgent care reflects poor access to community-based primary care providers (PCPs).

Study Design

Using a survey of ED patients, insurance claims data, and administrative records identifying demographic factors, we analyzed the use of the ED in an impoverished area of Brooklyn, New York.

Methods

We examined original survey data to investigate the extent to which residents of northern and central Brooklyn use EDs for nonemergencies and whether these patients have access to PCPs. We used data from health insurers operating in northern and central Brooklyn, and New York state hospital ED visit data to investigate the factors influencing ED visits for ambulatory care—sensitive conditions (ACSCs). Logistic regression was used to identify characteristics that predict ED visits not resulting in admission for ACSCs.

Results

Of 11,546 patients that completed our survey, the presenting complaint was self-described as emergent by 57%, 30% had no PCP, and 19% reported no health insurance coverage. Using health insurance plan encounter data, only 15 % of patients had seen any provider within 1 week of the ED visit. Insurance type, age, gender, race/ethnicity, and socioeconomic status of area of residence influence the likelihood of these ED visits.

Conclusions

Correlating data from 3 sources, we suggest that the expansion of insurance under the Affordable Care Act may not be sufficient to reduce ED use for nonurgent conditions.

Am J Manag Care. 2015;21(2):e152-e160

We analyzed whether the nonurgent use of the emergency department (ED) in an impoverished area of Brooklyn, New York, is associated with poor access to primary care and other modifiable factors. We found:

  • Insurance type, age, gender, race/ethnicity, and socioeconomic status of area of residence influence the likelihood of these ED visits.
  • Medicaid clients, rather than the uninsured, account for the largest share of patients using EDs for nonurgent conditions.
  • Nonurgent ED visits are rarely preceded by a visit to a primary care provider, and this circumstance was unrelated to the type of health insurance.

The implementation of the Patient Protection and Affordable Care Act, with its expansion of public and private health insurance, raises questions about future use of the emergency department (ED), especially in deprived areas. There is evidence that, despite healthcare reform and the expansion of insurance coverage, there will be inadequate access to primary care for the newly insured. This, combined with a reduction in financial barriers to those using the ED, could increase demand for ED services.1-3

The ED is not an optimal site for nonurgent care. Nonemergent care delivered in the ED worsens overcrowding, reduces patient satisfaction, and is more costly than care offered in other settings.4 Treatment is often delivered without the availability of the complete medical history and without the capacity to ensure follow-up, resulting in episodic, fragmented, low-value care.

Annual ED visit rates in the United States have increased more than would be expected from population growth.5 An increasing number of uninsured and Medicaid patients contribute to rising ED visit rates.6,7 Factors known to influence an individual’s decision to visit the ED for nonurgent care have been reviewed.8 The recent healthcare reform debates and the pressure to redesign Medicaid to reduce state budget deficits have prompted implementation of measures designed to reduce ED use, despite evidence that diverting low-acuity cases away from the ED results in smaller savings than strategies to reduce inpatient admissions.9

The areas of northern and central Brooklyn are deprived sections of New York City with a large percentage of minorities, high unemployment, poverty and crime rates, and a significant percentage of households on public assistance and receiving food stamps. There are known poor health outcomes, including high rates of inpatient stays for ambulatory care—sensitive conditions (ACSCs) and premature mortality.10

In this paper we investigate the extent to which ED visits in northern and central Brooklyn occur for diagnoses that could have been treated in another setting, and we explore the reasons patients use EDs for the treatment of nonurgent conditions, including limitations in access to primary care. Preventable encounters with the healthcare system in the ED for ACSCs, a measure of limited access to quality primary care and poor care coordination,11-13 should be responsive to system changes that direct patients to non-ED care settings.

Because there is no single database that allows us to explore these questions, we combine data from 3 different sources to help us develop and test hypotheses. To assess the extent to which patients are treated in EDs for nonurgent conditions, we employ the rates of ED visits for ACSCs not resulting in admission. Administrative data, from the New York Statewide Planning and Research Cooperative System (SPARCS), document the extent of this potentially preventable ED use, and we analyze associated factors including insurance status and demographic and neighborhood factors. To complement the administrative data, we report findings from an original survey describing Brooklyn patients’ use of the ED. The survey allows us to explore whether patients believe their visit to the ED is an emergency and to identify other factors that may influence the use of EDs for nonurgent conditions. We use insurance data from plans operating in Brooklyn to investigate whether patients who have visited the ED for ACSCs have received outpatient care in the weeks before or after the ED visit. This information, with survey data, can help us understand whether patients who present in the ED for nonurgent conditions are receiving primary care.

METHODS

eAppendix

The Brooklyn Health Care Improvement Project (BHIP) (see [available at www.ajmc.com] for additional information),14 directed by researchers at the State University of New York—Downstate, includes a coalition of northern and central Brooklyn federally qualified health centers (FQHCs), hospitals, insurance plans, and community-based organizations. BHIP was funded by a 2009 New York State Health Care Efficiency and Affordability Law grant to develop a comprehensive health planning strategy for a designated area within northern and central Brooklyn covering more than 1 million lives. BHIP collected data to develop and evaluate strategies for reducing unnecessary ED visits.

We triangulated our analysis of ED healthcare-seeking behavior for primary care-treatable conditions (ie, ACSCs) by our study population. We examined original survey data from a large sample of ED patients to investigate the extent to which EDs are used for self-reported nonemergencies and whether these patients report having access to primary care providers. The survey, of a total of 11,546 patients or their representatives, was completed in 2 rounds (two 2-week spans, each week including its full 168 hours) and represents approximately one-third of individuals visiting the ED at any 1 of the 6 BHIP participating hospitals. The survey was completed during the last 2 weeks in January 2011 and the first two weeks of August 2011. (The survey instrument is available in the eAppendix.)

Encounter data from health insurers operating in northern and central Brooklyn is used to investigate the relation between ED visits and visits to other outpatient care sites by patients with Medicaid and commercial health coverage. We examined de-identified insurance claims data, institutional and professional files (which are the parts of the insurance claims data) from 2007 to 2009, from 8 insurers serving the study area. These data were pooled and classified as commercial (approximately 500,000 covered lives) or Medicaid (389,000 covered lives), and divided into children (aged <18 years of age and representing 47% of members with claims) and adults (aged 18 years and over) for analysis.

Finally, ED visits for ACSCs (Billings’ definition11) that do not result in admission to the hospital are analyzed using SPARCS hospital ED visit data. This population-based administrative data set is used to conduct a logistic regression analysis (SPSS version 19, IBM, Amonk, New York). The ED visit data for years 2007 through 2009 included randomly generated personal identifiers and residence geo-coded to 2010 US Census Tract (CT) boundaries, to link the data to 2010 US CT-level variables. The dependent variable is an ED visit for an ACSC diagnosis that did not result in an inpatient stay. Individual-level independent variables include age, gender, race/ethnicity, and primary payer. The model controls for CT community-level variables (derived from US Census 2010, file SF-1) include median household income, education achieved, self-assessed English language competence, and area housing vacancy rates. As a proxy for primary care provider (PCP) availability, the model includes the CT age-adjusted rate of inpatient discharges for ACSCs, a better indicator of PCP availability than the number of physicians in the neighborhood.15

RESULTS

Table 1

Survey respondents were 57% female and 64% native-born, and 86% had lived in the same residence for more than 1 year (). Self-identified race and ethnicity reflect the population in the study area: 79% identified themselves as black, 3% as white, and 0.5% as Asian. Hispanics accounted for 20.5%. Of all respondents, 57% indicated that their visit was a medical emergency, 30% stated that they had no PCP, and 19% reported no health insurance coverage. Rates of coverage exceeded 90% in those younger than 18 years and older than 65 years, and averaged 76% for the 18- to 64-year-old age group. Medicare accounted for 10% of all payers, while 20% were covered by commercial insurance. Medicaid including Child Health Plus or Family Health Plus accounted for 51%.

Of those who stated that their visit was nonemergent, one-third used the ED for convenience or because it is their usual source of care; 29% could not schedule a PCP appointment, believed the wait for an appointment was too long, or were advised by their usual source of care to present to the ED. Lack of insurance or lower out-of-pocket cost accounted for 6% of visits. Of those survey respondents claiming to not have a PCP, 20% are covered by a Medicaid managed care plan.

Table 2

Encounter data from the insurance plans reveal that most patients visiting an ED in northern and central Brooklyn for ACSCs do not make regular visits to community or institution based non-ED providers. Timing of visits made before and after an ED visit for ACSCs are presented in . Fifteen percent of the patients presenting to the ED with an ACSC had seen any covered provider within 1 week prior to their visit and 38% had seen a provider within 4 weeks. The percentage of patients seeing a provider prior to an ED visit for an ACSC varied little with type of coverage or age group. Follow-up visits with a non-ED provider after an ED visit for an ACSC diagnosis occurred in about 45% of patients for whom we have encounter data, despite the generic instructions to seek follow-up appointments with their PCP. These findings are consistent with self-reported information from the survey about access to primary care providers.

Table 3

Table 4

A summary of the characteristics of the patients residing in our study area visiting EDs as compared with the residents of all of Brooklyn is presented in . The odds of an ED visit for an ACSC diagnosis not resulting in admission is lower among younger people, women, and among those classified as Asian (). The odds of visits by Medicare beneficiaries and Medicaid clients are significantly higher than those for patients with commercial insurance. The odds of an uninsured patient visiting an ED for an ACSC are no different than the odds of a privately insured patient making such a visit. Residing in a census tract with the lowest quartile of household median income, with highest rates of those without at least a high school education, with the highest vacant housing rates, and with the highest rates of those who do not speak English well are all associated with statistically significant higher odds ratios of ED visits for ACSCs. There is almost no relationship between ED visits for ACSCs and access to local primary care as measured by area hospital discharge rates for ACSCs.

DISCUSSION

Patients who receive ambulatory care in safety net hospital EDs often have a high prevalence of chronic medical conditions and substance abuse and rarely attend a primary care clinic, preferring to return to the ED for subsequent care.16,17 The analysis of ED visits for ACSCs not resulting in a hospital admission is helpful in understanding 1 use of EDs that could be avoided through availability and appropriate use of community-based primary care.18 The Agency for Healthcare Research and Quality has verified reliability of ED visits for ACSCs in terms of precision, minimum bias, and construct validity.12 Other factors outside the direct control of the healthcare system, such as poor environmental conditions or lack of patient compliance, can result in ED visits for ACSCs. However, our analysis, the first using 3 data sources, suggests that unsatisfactory access to health services in the community is an important factor. Rates of ED visits for ACSCs not resulting in admission can be used to provide a measure of unmet community healthcare needs, to monitor how well complications from a number of common conditions are being avoided, and to compare the performance of healthcare systems across communities.

EDs have long served as the safety net for medically underserved patients, particularly adults with Medicaid and patients without any health insurance.5 Self-reported health status and rates of diagnosis of such chronic illnesses as hypertension or diabetes19 suggest that our study population is sicker than the borough and citywide averages, but this does not fully explain the high rate of ED use in this community. In contrast to a previous report from another state,20 the uninsured were not responsible for the majority of primary care-treatable ED visits. Rather, the Medicaid population makes the largest percentage of these visits (40.7%), despite estimates that Medicaid clients represent only 15.2% of the study area population.19 Black patients are at higher risk for ED visits for ACSCs, but this is in proportion to racial demographics of the study area. Our findings of high ED utilization for chronic ACSCs by black persons and Medicaid patients does not differ significantly from data indicating that nationally, barriers to primary care contribute to higher ED and hospital utilization rates seen in these groups.21

Only about 57% of the patients we surveyed maintained that they were in the ED for a “medical emergency,” and 30% claimed they had no primary care provider. Our findings from insurance claims data, showing that about 38% of patients presenting at the ED with ACSCs had been seen by a primary care provider during the preceding month, are consistent with the self-reported perceptions about lack of access to primary care captured by our survey. The fact that nearly two-thirds of the patients who required ambulatory care medical services appeared at the ED without seeking care from a primary care physician is striking, since many of these patients have Medicaid coverage and have been auto-assigned to a PCP even if they did not select one; this is true because of their mandatory participation in a Medicaid managed care organization.

These data suggest that plans operating in northern and central Brooklyn need to improve their primary care networks, work with the providers in their networks to improve their accessibility to patients, and/or do a better job of communicating with their clients about the primary care services to which they are entitled. We have no explanation for why admission rates for ACSCs measured at the census tract level do not have a significant influence on the odds of ACSC ED visits. Since the total number of PCPs generally appears to be adequate (based on a review of the lists of health providers provided by area insurers, hospitals, and clinics) and many of the patients using the ED for conditions that could be treated by a community-based PCP are covered by Medicaid, other factors must contribute to the gap in access. Inadequate PCP availability after usual working hours, the low rates of providers speaking languages other than English, a lack of cultural competence, and substandard customer service, on the one hand, combined with 24/7 ED availability, the perception of high quality of care in the ED, and the convenience of “one-stop shopping,” on the other, may all contribute to patients choosing the ED as a source of primary care.22

Our survey findings, in which 29% reported having difficulty making an appointment with a PCP, are consistent with these explanations. Previous studies suggest that the use of an ED for primary care reflects the extent to which patients have a relationship with a PCP and the perception of the urgency of need. A program that provided an enhanced referral system to family medicine homes from the ED was associated with decreased subsequent ED utilization by uninsured patients.23

Although the number of primary care providers may not be a problem in northern and central Brooklyn, patient perception of access to these providers is. Survey responses show that many patients in Brooklyn, including a large number who have Medicaid coverage, are using EDs for conditions that can be treated in a primary care setting. Claims data from health insurers demonstrate that patients have often not contacted the provider designated to serve as their usual source of care prior to an ED visit. In our interviews, patients cite difficulties in getting appointments and a lack of “convenience” as barriers to the use of community-based primary care. Medicaid managed care plans are designed to provide early interventions and preventive care and to facilitate access to primary care in the attempt to provide a less expensive alternative to emergency care.24 In northern and central Brooklyn, the Medicaid plans are not accomplishing this aim.

A range of factors contribute to ED visit rates for ACSCs; our evidence, however, suggests that a key determinant is not simply the number of PCPs in the area but the real or perceived convenience, quality, and effectiveness of local primary care. Universal strategies for reduction of ED overutilization by increasing access to, and timeliness and quality of, primary care for all patients likely to experience ACSCs are important.25 Outpatient safety-net providers can help to reduce ED visits for ACSCs in the Medicaid population by making care more expedient.26 If primary care providers in these areas of Brooklyn make it easier for patients to schedule appointments, offer convenient hours of operation, and provide a broad range of services in a dignified atmosphere, patients will be more likely to use this system for ACSCs. Furthermore, if the system can place such resources in proximity to an ED, patients would have the peace of mind that an ED is close at hand if there is the need for true emergency care. To decrease the reliance of this Brooklyn population on the ED as source of primary care, it is important to provide information regarding the accessibility of appropriate sources of care.

A community-based primary care program for uninsured low-income adults at Virginia Commonwealth University Medical Center in Richmond resulted in a decline in ED visits and inpatient admissions, while primary care visits increased.27 When researchers followed uninsured Baltimore patients with no regular healthcare provider, they learned that improving access to primary care services by referral to a community health center was not sufficient to reduce visits to the ED.28 This demonstrated that the characteristics of the source of care are important.

CONCLUSIONS

Most Medicaid clients in New York state are enrolled in managed care and have been assigned a PCP. In Brooklyn, Medicaid clients account for the largest share of patients using EDs for conditions treatable by community-based PCPs. State data suggest that adequate numbers of primary care physicians and clinics serve northern and central Brooklyn, but nonetheless, many Medicaid patients use the ED as their usual source of care and do not establish a relationship with a PCP. Educational efforts could improve the use of community-based PCPs among this patient population, but our findings suggest that it is a mistake to assume that people use EDs for non-emergency conditions simply out of ignorance. The primary care system in Brooklyn needs to do a better job of meeting patients’ needs. Survey and insurance encounter data suggest that many Medicaid patients chose the ED because the healthcare system has failed to provide easily accessible, culturally competent, timely, quality primary care. Ideally, this would include urgent care appointments with PCPs during daytime hours, the availability of same-day appointments, access to after-hours care, a means for urgent communication with a PCP, and convenient access to laboratory and x-ray testing. The many cultures represented in our study area will require a variety of solutions to meet these demands.

New interventions should be data-driven in order to detect effective strategies and to understand where to dedicate scarce resources. By identifying the neighborhoods with the highest rates of ED visits not resulting in admission, we have identified several areas in Brooklyn where the evaluation of the success of pilot projects designed to encourage the use of sources of care other than the ED should be feasible.Author Affiliations: State University of New York Downstate Medical Center (DW, GW, JT), Brooklyn, NY; International Longevity Center, Mailman School of Public Health, Columbia University (DW), New York, NY; The Hastings Center (MKG), Garrison, NY.

Funding Source: 2009 New York State Health Care Efficiency and Affordability Law grant (HEAL-9).

Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (DW, MKG, GW, JT); acquisition of data (DW, GW, JT); analysis and interpretation of data (DW, MKG, GW, JT); drafting of the manuscript (DW and MKG); critical revision of the manuscript for important intellectual content (DW, MKG, GW, JT); statistical analysis (DW, GW); ; obtaining funding (GW); administrative, technical, or logistic support (GW, JT); and supervision (GW).

Address correspondence to: Daniel Weisz, MD, MPA, 722 W 168th St, Room 1403, Mailman School of Public Health, Columbia University, New York, NY 10032. E-mail: dw2493@columbia.edu.REFERENCES

1. Chen C, Scheffler G, Chandra A. Massachusetts’ health care reform and emergency department utilization. N Engl J Med. 2011;365(12):e25.

2. Pitts SR, Carrier ER, Rich EC, Kellermann AL. Where Americans get acute care: increasingly, it’s not at their doctors’ office. Health Aff (Millwood). 2010;29(9):1620-1629.

3. Smulowitz PB, O’Malley J, Yang X, Landon BE. Increased use of the emergency department after health care reform in Massachusetts [published online March 20, 2014]. Ann Emerg Med. doi: 10.1016/j.annemergmed.2014.02.011.

4. McCaig LF, Burt CW. National Hospital Ambulatory Medical Care Survey: 2003 Emergency Department Summary. Hyattsville, MD: National Center for Health Statistics; 2005.

5. Tang N, Stein J, Hsia RY, Maselli JH, Gonzales R. Trends and characteristics of US emergency department visits, 1997-2007. JAMA. 2010;304(6):664-670.

6. Garcia TC, Bernstein AB, Bush MA. Emergency department visitors and visits: who used the emergency room in 2007? NCHS Data Brief. 2010;(38):1-8.

7. McConville S, Lee H. Emergency department care in California: who uses it and why? California Counts: Population Trends and Profiles. 2008;10(1):1-23. Public Policy Institute of California website. http://www.ppic.org/content/pubs/cacounts/CC_808SMCC.pdf. Accessed January 26, 2015.

8. Uscher-Pines L, Pines J, Kellerman A, Gillen E, Mehrotra A. Emergency department visits for nonurgent conditions: systematic literature review. Am J Manag Care. 2013;19(1):47-59.

9. Smulowitz PB, Honigman L, Landon BE. A novel approach to identifying targets for cost reduction in the emergency department. Ann Emerg Med. 2013;61(3):293-300.

10. NYC Community Health Profiles. New York City Department of Health and Mental Hygiene website. http://www.nyc.gov/html/doh/html/data/nyc-health-profiles.shtml. Published 2006. Accessed January 26, 2015.

11. Billings J, Zeitel L, Lukomnik J, Carey TS, Blank AE, Newman L. Impact of socioeconomic status on hospital use in New York City. Health Aff (Millwood). 1993;12(1):162-173.

12. Agency for Healthcare Research and Quality (AHRQ). Refinement of the HCUP Quality Indicators. Rockville, MD: AHRQ; 2001. (AHRQ Report No. 01-0035).

13. Weinick RM, Billings J, eds. Tools for monitoring the local safety net. Agency for Health Care Research and Quality website. Available at http://archive.ahrq.gov/data/safetynet. Published July 2003. Accessed January 26, 2015.

14. The Brooklyn Healthcare Improvement Project. SUNY Downstate Medical Center website. www.downstate.edu/bhip/. Accessed January 26, 2015.

15. Gusmano MK, Rodwin VG, Weisz D. Health Care in World Cities: New York, Paris, and London. Baltimore, MD: Johns Hopkins University Press, 2010.

16. Blackburn J, Becker DJ, Sen B, Morrisey MA, Caldwell C, Menachemi N. Characteristics of low-severity emergency department use among CHIP enrollees. Am J Manag Care. 2013;19(12):e391-e399.

17. Rask KJ, Williams MV, McNagny SE, Parker RM, Baker DW. Ambulatory health care use by patients in a public hospital emergency department. J Gen Intern Med. 1998;13(9):614-620.

18. Johnson PJ, Ghildayal N, Ward AC, Westgard BC, Boland LL, Hokanson JS. Disparities in potentially avoidable emergency department (ED) care: ED visits for ambulatory care sensitive conditions. Med Care. 2012;50(12):1020-1028.

19. Community Health Survey. New York City Department of Health and Mental Hygiene website. https://a816-healthpsi.nyc.gov/SASStoredProcess/guest?_PROGRAM=%2FEpiQuery%2FCHS%2Fchsindex& year=2009. Published 2009. Accessed January 26, 2015.

20. Begley CE, Vojvodic RW, Seo M, Burau K. Emergency room use and access to primary care: evidence from Houston, Texas. J Health Care Poor Underserved. 2006;17(3):610-624.

21. Oster A, Bindman AB. Emergency department visits for ambulatory care sensitive conditions: insights into preventable hospitalizations. Med Care. 2003;41(2):198-207.

22. McCusker J, Roberge D, Lévesque JF, et al. Emergency department visits and primary care among adults with chronic conditions. Med Care. 2010;48(11):972-980.

23. Murnik M, Randal F, Guevara M, Skipper B, Kaufman A. Web-based primary care referral program associated with reduced emergency department utilization. Fam Med. 2006;38(3):185-189.

24. Long JA, Chang VW, Ibrahim SA, Asch DA. Update on the health disparities literature. Ann Intern Med. 2004;141(10):805-812.

25. Chukmaitov AS, Tang A, Carretta HJ, Menachemi N, Brooks RG. Characteristics of all, occasional, and frequent emergency department visits due to ambulatory care-sensitive conditions in Florida. J Ambul Care Manage. 2012;35(2):149-158.

26. Falik M, Needleman J, Wells B, Korb J. Ambulatory care sensitive hospitalizations and emergency visits: experiences of Medicaid patients using federally qualified health centers. Med Care. 2001;39(6):551-561.

27. Bradley CJ, Gandhi SO, Neumark D, Garland S, Retchin SM. Lessons for coverage expansion: a Virginia primary care program for the uninsured reduced utilization and cut costs. Health Aff (Millwood). 2012;31(2):350-359.

28. McCarthy ML, Hirshon JM, Ruggles RL, Docimo AB, Welinsky M, Bessman ES. Referral of medically uninsured emergency department patients to primary care. Acad Emerg Med. 2002;9(6):639-642.