Comparing patients who leave the ED prematurely, before vs after medical evaluation: a National Hospital Ambulatory Medical Care Survey analysis

American Journal of Emergency Medicine xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

American Journal of Emergency Medicine journal homepage: www.elsevier.com/locate/ajem

Original Contribution

Comparing patients who leave the ED prematurely, before vs after medical evaluation: a National Hospital Ambulatory Medical Care Survey analysis☆ Jessica Moe, MD a,⁎, Justin Brett Belsky, MD b a b

Department of Emergency Medicine, University of Alberta, Edmonton, Alberta, Canada Department of Emergency Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA

a r t i c l e

i n f o

Article history: Received 21 October 2015 Received in revised form 15 January 2016 Accepted 19 January 2016 Available online xxxx

a b s t r a c t Background: Many patients leave the Emergency Department (ED) before beginning or completing medical evaluation. Some of these patients may be at higher medical risk depending on their timing of leaving the ED. Objectives: To compare patient, hospital, and visit characteristics of patients who leave before completing medical care to patients who leave before ED evaluation. Methods: Retrospective cross-sectional analysis of ED visits using the 2009-2011 National Hospital Ambulatory Medical Care Survey. Results: A total of 100 962 ED visits were documented in the 2009-2011 National Hospital Ambulatory Medical Care Survey, representing a weighted count of 402 211 907 total ED visits. 2646 (2.62%) resulted in a disposition of left without completing medical care. Of these visits, 1792 (67.7%) left before being seen by a medical provider versus 854 (32.3%) who left after medical provider evaluation but before a final disposition. Patients who left after being assessed by a medical provider were older, had higher acuity visits, were more likely to have visited an ED without nursing triage, arrived more often by ambulance, and were more likely to have private insurance than to be self-paying or to have other payment arrangements (e.g. worker's compensation or charity/no charge). Conclusions: When comparing all patients who left the ED before completion of care, those who left after versus before medical provider evaluation differed in their patient, hospital, and visit characteristics and may represent a high risk patient group. © 2016 Elsevier Inc. All rights reserved.

1. Introduction Approximately 1.7% of all patients presenting to a United States Emergency Department (ED) leave after triage or before completing medical care (Pham 2009 [1]). While several studies suggest that patients who leave before completing ED care may be sicker, more likely to return and to require urgent admission (Baker 1991 [2], Rowe 2006 [3]), not all studies agree on this increased risk (Clarey 2012 [4], Kennedy 2008 [5], Tropea 2012 [6]). For instance, patients who leave without being seen tend to have lower triage levels (Clarey 2012 [4]), and may indicate long wait times (Rowe 2006 [3], Clarey 2012 [4], Weiss 2005 [7], Polevoi 2005 [8]). Rates of emergent 30-day admissions were higher among patients who left against medical advice after medical assessment (4.4%) versus those who left before being seen by a medical provider ☆ Supports (financial or in-kind): None. Meetings and Awards: None. Conflicts of interest: The authors report no conflict of interests for this manuscript. ⁎ Corresponding author at: RCPS Emergency Medicine Residency Program, University of Alberta, 750 University Terrace, 8303 - 112 Street, Edmonton, Alberta T6G 2T4. Tel.: +1 780 934 1896; fax: +1 780 492 4341. E-mail address: [email protected] (J. Moe).

(2.6%) (Ding 2007 [9]). For patients who leave before completing care, clinical risk may be related to their timing of leaving the ED relative to medical assessment. To our knowledge, patients who leave without completing their medical care before and after assessment by a medical provider have not been compared on a national level. The objective of this study is to compare patient, visit and hospital characteristics associated with patients who leave before completing their ED care either before or after being evaluated by a medical provider using the 2009-2011 National Hospital Ambulatory Medical Care Survey (NHAMCS). This information will provide a sense of whether these patients differ in their baseline demographic and clinical characteristics, which is the first step in determining potential interventions, if indicated. 2. Methods 2.1. Study design This study is a retrospective cross-sectional secondary analysis of ED visits using the 2009-2011 NHAMCS database. Institutional review board approval for the secondary analysis of these data was submitted

http://dx.doi.org/10.1016/j.ajem.2016.01.015 0735-6757/© 2016 Elsevier Inc. All rights reserved.

Please cite this article as: Moe J, Belsky JB, Comparing patients who leave the ED prematurely, before vs after medical evaluation: a National Hospital Ambulatory Medical Care S..., Am J Emerg Med (2016), http://dx.doi.org/10.1016/j.ajem.2016.01.015

2

J. Moe, J.B. Belsky / American Journal of Emergency Medicine xxx (2016) xxx–xxx

and considered exempt by the Massachusetts General Hospital Institutional Review Board. 2.2. Study setting and population The NHAMCS is a probability sample survey that generates a nationally representative sample of ED visits. The survey methodology has previously been well described (CDC/National Center for Statistics, 2010[10]). The NHAMCS uses a multi-stage probability design (McCaig 1994 [11]) and rigorous data processing and quality assurance measures (McCaig 2012 [12]). The survey design incorporates clustering of hospitals within geographic primary sampling units and of visits within hospitals (McCaig 2012 [12]). Patients, hospitals, and providers may be represented more than once in a year and may not be repeated in subsequent years as surveys are not linked to each other. 2.3. Measurements Our population of interest consisted of those who left the ED before completing their medical evaluation and treatment. In order to identify this population, we combined all visits with coded dispositions of “left before triage,” “left after triage” (for 2010 and 2011), “left before medical screening exam,” “left after medical screening exam” (for 2009) or “left against medical advice” (all years). We then defined two groups based on whether or not they had been evaluated by a medical provider. We defined a medical provider as either an attending physician, resident/intern, consult physician, physician assistant, nurse practitioner, or mental health provider. Patient, visit and hospital characteristics were assessed for each unique visit entry and included: age (b18, 18-39, 40-64, and ≥ 65), sex, race/ethnicity (Caucasian/non-Hispanic, African American/nonHispanic, Hispanic/other), insurance status (private, Medicaid/CHIP, Medicare, and self-pay/other), metropolitan status area (defined by the U.S. Office of Management and Budget and designated by the U.S. Census Bureau), geographic region (Northeast, Midwest, South, and West), hospital ownership (proprietary, voluntary non-profit, and government [non-federal]), EDs that use separate fast track unit for non-urgent care, triage status (emergent/immediate, urgent, semiurgent/non-urgent or visit occurred in an ED that does not conduct triage), ambulance arrival to ED, chief complaint (based on coded “reason for visit for ambulatory care”), pain scale out 10 (0, 1-4, 5-10), day of the week (weekday versus weekend), number of times seen in the same ED in last 12 months (0, 1-4, ≥ 5), and whether the patient had been seen in the same ED in last 72 h. Age in years was the only continuous variable analyzed. All variable cell counts had a minimum of 30 unweighted records. 2.4. Data analysis The 2009, 2010, and 2011 NHAMCS public-use data files were obtained and imported into SAS 9.4 (SAS Institute Inc., Cary, NC). Due to the complex sample design, sampling errors were determined with SUDAAN 11.0.1 (RTI, Research Triangle Park, NC) which takes into account the clustered nature of the sample. Descriptive statistics (means with standard errors [SE], proportions) were used to describe patient, visit, and hospital characteristics. Differences in categorical variables between patients who left before and after evaluation by a medical provider were assessed using chisquared tests. Bivariate logistic regressions were conducted to generate unadjusted odds ratios for visits ending before and after evaluation by a medical provider. A multivariate logistic regression model was created to generate adjusted odds ratios. Categorical variables that were statistically significant on unadjusted bivariate analysis were included in the model and variables with ≥ 15% missing entries were excluded. A 2-sided P b .05 was considered statistically significant for all comparisons. No correction was made to adjust for multiple comparisons.

3. Results A total of 100,962 patient record forms were obtained from the 20092011 NHAMCS database equating to a weighted count of 402,211,907 total visits. 2646 (2.62%) visits resulted in a patient leaving before completing medical care, representing a weighted count of 10,875,283 visits. Of the 2646 patients who left before completion of medical care, 854 (32.3%) left after being evaluated by a medical professional versus 1792 patients (67.7%) who left before being assessed by a medical professional. Patient, visit and hospital characteristics of ED visits ending before completion of care are presented in Tables 1, 2 and 3, respectively. On bivariate analysis, age, race/ethnicity, insurance status, triage urgency, arrival by ambulance, reason for visit, and frequency of ED visits in the past 12 months were statistically significant. The previous variables were included in our multivariate model except for frequency of visits in the past 12 months given a missing value rate of 36.3%. On multivariate analysis, patients who left after compared to those who left before being assessed by a medical provider were older (higher proportion in the 4065 range than b18 or 18-39), and more likely to have private insurance than to be self-paying or to have other payment arrangements (e.g. worker's compensation or charity/no charge) (Table 1). Visits associated with patients who left after medical provider assessment were of higher acuity (more likely triaged as emergent/immediate or urgent than semiurgent or non-urgent), more likely to have occurred in an ED that did not conduct nursing triage, and more likely to have arrived to the ED by ambulance than visits made by patients who left before medical provider assessment (Table 2). The following variables were significantly associated with leaving after compared to before medical provider assessment on unadjusted analysis but were no longer significant in the adjusted, multivariate model: a decreased likelihood of being Hispanic, greater likelihood of Medicare, and psychiatric reason for visit. 4. Discussion In this secondary analysis of the 2009-2011 NHAMCS database, we found that patients who left before completing care were older, more likely to arrive by ambulance, and had higher acuity visits than patients who left without being seen. Our findings are consistent with previous research demonstrating that patients who leave without being seen by a medical provider are younger, have lower acuity visits and tend not to arrive by ambulance (Goodacre 2005 [13], Crilly 2013 [14]). Increased rates of patients who leave before medical assessment are associated with high ED volumes and crowding (Weiss 2005 [7], Hobbs 2000 [15], McMullan 2004 [16]). Patients consistently identify wait times as a major reason for leaving before medical assessment (Rowe 2006 [3], McNamara 1995 [17], Johnson 2009 [18], Mohsin 2007 [19], Monzon 2005 [20], Arendt 2003 [21], Baker 1991 [2]). Conversely, high rates of patients leaving after provider assessment have not been shown to correlate with ED overcrowding (Ding 2007 [9]), and reasons for leaving prematurely have not been clearly identified in this group. Previous research has concluded that many of these patients come from vulnerable populations with poor overall access to care (Ding 2006 [22], Sun. 2007 [23]), so minimizing attrition is important. Our study has several important limitations. Although the NHAMCS reports high survey response rates and performs regular quality checks for accurate coding, there are inherent limitations in ensuring completeness and accuracy of information when performing large database analysis. Missing data could affect the accuracy of our analysis; our a priori decision to exclude variables with ≥15% missing values from the multivariate model limits this problem. Furthermore, we recognize the potential information bias that could arise from using coding to identify variables. Our use of provider coding to identify patients who did and did not see a medical provider may not be accurate in all cases. It is possible that a medical provider “signed up” for a patient without actually seeing them. It is also possible that more complete or accurate information was available for those patients who were seen by a medical provider versus those who were not.

Please cite this article as: Moe J, Belsky JB, Comparing patients who leave the ED prematurely, before vs after medical evaluation: a National Hospital Ambulatory Medical Care S..., Am J Emerg Med (2016), http://dx.doi.org/10.1016/j.ajem.2016.01.015

J. Moe, J.B. Belsky / American Journal of Emergency Medicine xxx (2016) xxx–xxx

3

Table 1 Patient characteristics of ED visits ending before completion of medical treatment Eval by MP

No eval by MP

Unadjusted OR of

Adjusted OR of

Patient characteristics

(n = 854)

(n = 1792)

Seen (95% CI)

Seen (95% CI)

Age, mean ± SE (y)⁎ b18 18-39 40-65 N65 Sex Male Female Race/ethnicity⁎

41.47 ± 0.92 9.54 37.61 40.55 12.30

29.52 ± 0.69⁎ 26.18 45.18 23.11 5.53

0.21 (0.15-0.29)⁎ 0.47 (0.38-0.60)⁎ Reference 1.27 (0.86-1.86)

0.27 (0.17-0.43)⁎ 0.61 (0.47-0.79)⁎ Reference 0.95 (0.57-1.58)

48.78 51.22

44.28 55.72

Reference 0.83 (0.66-1.05)

Reference

white/non-Hispanic African American/non-Hispanic Hispanic/Other Payment⁎ Private Medicaid/CHIP Medicare Self-pay/Other

57.44 28.91 13.66

48.25 31.42 20.34

Reference 0.77 (0.59-1.01) 0.56 (0.38-0.83)⁎

Reference 0.80 (0.55-1.15) 0.68 (0.43-1.07)

22.56 30.60 19.54 27.30

20.17 33.62 9.87 36.35

Reference 0.81 (0.60-1.10) 1.77 (1.29-2.42)⁎ 0.67 (0.49-0.92)⁎

Reference 1.03 (0.70-1.54) 0.98 (0.63-1.52) 0.69 (0.49-0.99)⁎

Model adjusts for age, race/ethnicity, payment, triage urgency, arrival by ambulance, and reason for visit. Eval, evaluation; MP, medical provider; OR, odds ratio; CI, confidence interval; CHIP, Children's Health Insurance Program. ⁎ P b .05

In conclusion, among patients who leave the ED before completing their care, those who leave before versus after medical assessment have different demographic and clinical characteristics. Future research should focus on determining reasons for which patients leave after being assessed by a medical provider, to identify risk factors that could allow intervention.

Acknowledgements The authors would like to thank Dr. Ellen McCarthy and Shimon Shaykevich for their guidance with the NHAMCS database and statistical considerations.

Table 2 ED visit characteristics ED Characteristics Triage status⁎ Emergent/immediate Urgent Semi-urgent/non-urgent Does not conduct nursing triage Arrival by ambulance Yes No Primary reason for visit⁎ Neurologic Psychiatry CVS/respiratory MSK/skin GI/GU Injury/poisoning Other Pain severity (0-10) 0 1-4 5-10 Day of week Weekday Weekend Visits to same ED in last 12 months⁎ 0 1-4 ≥5 Previous visit to same ED in past 3 days Yes No

Eval by MP (n = 854)

No Eval by MP (n = 1,792)

Unadjusted OR of Seen (95% CI)

Adjusted OR of Seen (95% CI)

15.64 50.32 30.30 3.75

7.86 37.91 52.71 1.53

3.46 (2.39-5.01)⁎ 2.31 (1.79-2.98)⁎ Reference 4.27 (2.08-8.73)⁎

2.81 (1.77-4.47)⁎ 1.91 (1.43-2.56)⁎ Reference 5.21 (1.65-16.39)⁎

22.87 77.13

8.02 91.98

Reference 0.29 (0.22-0.40)⁎

Reference 0.42 (0.29-0.60)⁎

8.94 4.11 10.03 14.88 20.72 11.08 30.24

7.02 2.43 12.38 19.82 21.97 10.52 25.85

1.57 (0.99-2.48) 2.09 (1.13-3.89)⁎ Reference 0.93 (0.64-1.35) 1.16 (0.79-1.71) 1.30 (0.79-2.15) 1.44 (0.99-2.10)

1.57 (0.92-2.68) 1.52 (0.73-3.18) Reference 1.00 (0.64-1.56) 1.16 (0.75-1.80) 1.57 (0.87-2.85) 1.51 (1.01-2.27)⁎

23.85 16.13 60.02

27.87 14.44 57.69

Reference 1.30 (0.86-1.98) 1.22 (0.85-1.73)

71.51 28.49

74.88 25.12

Reference 1.19 (0.94-1.51)

40.44 44.31 15.25

42.11 48.40 9.49

0.60 (0.41-0.88)⁎ 0.57 (0.35-0.92)⁎ Reference

5.50 94.50

7.06 92.94

Reference 1.31 (0.87-1.95)

CVS, Cardiovascular; MSK, Musculoskeletal; GI, Gastrointestinal; GU, Genitourinary. ⁎ P b .05

Please cite this article as: Moe J, Belsky JB, Comparing patients who leave the ED prematurely, before vs after medical evaluation: a National Hospital Ambulatory Medical Care S..., Am J Emerg Med (2016), http://dx.doi.org/10.1016/j.ajem.2016.01.015

4

J. Moe, J.B. Belsky / American Journal of Emergency Medicine xxx (2016) xxx–xxx

Table 3 Hospital characteristics of ED visits ending before completion of medical treatment Hospital characteristics

Metropolitan status MSA Non MSA Geographic region Northeast Midwest South West Ownership Nonprofit Government, non-Federal 16.37 Proprietary Presence of ED fast track unit for non-urgent care Yes No

Eval by MP (n = 854)

No Eval by MP (n = 1,792)

Unadjusted OR of Seen (95% CI)

87.45 12.55

87.50 12.50

Reference 1.0 (0.57–1.77)

15.64 23.66 40.37 20.32

13.82 19.09 49.70 17.39

0.91 (0.60-1.38) Reference 0.66 (0.45-0.95)⁎ 0.94 (0.58-1.53)

71.45 16.37

74.35 16.75

0.70 (0.42-1.17) 0.71 (0.39-1.31)

12.18

8.89

75.73 24.27

73.69 26.31

Reference

Reference 0.90 (0.62-1.29)

MSA, metropolitan statistical area. ⁎ P b .05

References [1] Pham JC, Ho GK, Hill PM, McCarthy ML, Pronovost PJ. National study of patient, visit, and hospital characteristics associated with leaving an emergency department without being seen: predicting LWBS. Acad Emerg Med 2009;16(10):949–55. [2] Baker DW, Stevens CD, Brook RH. Patients Who Leave a Public Emergency Department Without Being Seen by a Physician: Causes and Consequences. JAMA 1991; 266(8):1085–90. [3] Rowe BH, Channon P, Bullard M, Blitz S, Saunders D, Rosychuk RJ, et al. Characteristics of Patients Who Leave Emergency Departments without Being Seen. Acad Emerg Med 2006;13:848–52. [4] Clarey AJ, Cooke MW. Patients who leave emergency departments without being seen: literature review and english data analysis. Emerg Med J 2012;29(8):617–21. [5] Kennedy M, MacBean CE, Brand C, Sundararajan V, Taylor DM. Review article: Leaving the emergency department without being seen. Emerg Med Australas 2008;20: 306–13.

[6] Tropea J, Sundararajan V, Gorelik A, Kennedy M, Cameron P, Brand CA. Patients Who Leave Without Being Seen in Emergency Departments: An Analysis of Predictive Factors and Outcomes. Acad Emerg Med 2012;19:439–47. [7] Weiss SJ, Ernst AA, Derlet R, King R, Bair A, Nick TG. Relationship between the national ED overcrowding scale and the number of patients who leave without being seen in an academic ED. Am J Emerg Med 2005;23(3):288–94. [8] Polevoi SK, Quinn JV, Kramer NR. Factors Associated with Patients Who Leave without Being Seen. Acad Emerg Med 2005;12:232–6. [9] Ding R, Jung JJ, Kirsch TD, Levy F, McCarthy ML. Uncomplicated emergency department care: patients who leave against medical advice. Acad Emerg Med 2007;14: 870–6. [10] Scope and Sample Design: NHAMCS Scope and Sample Design [Internet]. Atlanta, GA: CDC/National Center for Health Statistics. [updated 2010 Jan 15; cited 2016 Jan 14; Available from] http://www.cdc.gov/nchs/ahcd/ahcd_scope.htm#nhamcs_ scope. [11] McCaig LF, McLemore T. Plan and operation of the national hospital ambulatory medical survey. Series 1: programs and collection procedures. Vital Health Stat 1994;34:1–78. [12] McCaig LF, Burt CW. Understanding and interpreting the national hospital ambulatory care survey: key questions and answers. Ann Emerg Med 2012;60(6):716–21. [13] Goodacre S, Webster A. Who waits longest in the emergency department and who leaves without being seen? Emerg Med J 2005;22:93–6. [14] Crilly J, Bost N, Thalib L, Timms J, Gleeson H. Patients who present to the emergency department and leave without being seen: prevalence, predictors and outcomes. Eur J Emerg Med 2013;20:248–55. [15] Hobbs D, Kunzman SC, Tandberg D, Sklar D. Hospital Factors Associated with Emergency Center Patients Leaving Without Being Seen. Am J Emerg Med 2000;18:767–72. [16] McMullan JT, Veser FH. Emergency Department Volume and Acuity as Factors in Patients Leaving Without Treatment. South Med J 2004;97(8):729–33. [17] McNamara KJ. Patients Leaving the ED Without Being Seen by a Physician: Is SameDay Follow-Up Indicated? Am J Emerg Med 1995;13:136–41. [18] Johnson M, Myers S, Wineholt J, Pollack M, Kusmiesz AL. Patients Who Leave the Emergency Department Without Being Seen. J Emerg Nurs 2009;35(2): 105–8. [19] Mohsin M, Forero R, Ieraci S, Bauman AE, Young L, Santiano N. A population followup study of patients who left an emergency department without being seen by a medical officer. Emerg Med J 2007;24:175–9. [20] Monzon J, Friedman SM, Clarke C, Arenovich T. Patients who leave the emergency department without being seen by a physician: a control-matched study. Can J Emerg Med 2005;7(2):107–13. [21] Arendt KW, Sadosty AT, Weaver AL, Brent CR, Boie ET. The Left-Without-Being-Seen Patients: What Would Keep Them From Leaving? Ann Emerg Med 2003;42:317–23. [22] Ding R, McCarthy ML, G Li, TD Kirsch, JJ Jung, GD Kelen. Patients Who Leave Without Being Seen: Their Characteristics and History of Emergency Department Use. Ann Emerg Med 2006;48:686–93. [23] Sun BC, Binstadt ES, Pelletier A, Camargo CA. Characteristics and Temporal Trends of “Left Before Being Seen” Visits in US Emergency Departments, 1995-2002. J Emerg Med 2007;32(2):211–5.

Please cite this article as: Moe J, Belsky JB, Comparing patients who leave the ED prematurely, before vs after medical evaluation: a National Hospital Ambulatory Medical Care S..., Am J Emerg Med (2016), http://dx.doi.org/10.1016/j.ajem.2016.01.015