a b s t r a c t

Purpose: The aim of this study was to describe population-based patterns of chronic obstructive pulmonary disease (COPD)-related emergency department (ED) visits.

Methods: We analyzed all COPD-related ED visits made by North Carolina residents 45 years or older in 2008 to 2009 using statewide surveillance system data. Return visits were identified when patients returned to the same ED within 3 or 14 days of a prior COPD-related visit. We quantify the prevalence of hospitalization and return visits by age, sex, and payment method and describe ED disposition patterns.

Results: Nearly half (46.3%) of the 97 511 COPD-related ED visits resulted in hospital admission. The percent of visits preceded by another COPD-related visit within 3 and 14 days was 1.6% and 6.2%, respectively. Emergency department-related hospitalizations increased with age; there were no differences by sex. Hospitalizations were less likely for uninsured, Medicare, and Medicaid visits than for privately insured visits. In contrast, 3- and 14-day return visits were more likely to be uninsured, Medicare, and Medicaid visits than privately insured visits. Fourteen-day returns were more likely to be made by men. Return visits initially increased with age compared with the 45- to 49-year age group, then decreased steadily after age 65 years. When return visits were made, discharge at both visits was the most common disposition pattern. However, 33.7% of 3-day returns and 22.7% of 14-day returns were discharged at the First visit and hospitalized upon returning to the ED.

Conclusions: Chronic obstructive pulmonary disease-related hospital admissions and short-term return ED visits were common and varied by age and insurance status. Chronic obstructive pulmonary disease management remains a critical area for intervention and quality improvement.

(C) 2013

Introduction

Chronic obstructive pulmonary disease (COPD) is prevalent among older adults [1] and contributes to substantial Health care use, Medical costs [2], quality-of-life decrements [3], and mortality [4]. acute COPD exacerbations are often treated in the emergency department (ED)

? Presentation at a scientific meeting: This study was presented in an oral presentation at the International Society for Disease Surveillance, December 2012, San Diego, CA. The conference abstract was published in a special supplement in the Online Journal of Public Health Informatics (available at: http://pear.accc.uic.edu/htbin/cgiwrap/bin/ojs/index.php/ ojphi/article/view/4408/3457).

* Corresponding author. Department of Epidemiology, McGavran-Greenberg Hall, Gillings School of Global Public Health, University of North Carolina, Chapel Hill, NC 27599-7435. Tel.: +1 919 843 2361; fax: +1 919 843 1226.

E-mail address: [email protected] (S.J. Lippmann).

and result in immediate hospital admission [5,6]. Many patients treated in the ED return to the ED within days to weeks [7], including some who were admitted to the hospital at the initial visit. These return visits may be the result of new exacerbations, inadequate treatment or symptom management, infections, or reexposure to environmental triggers such as tobacco smoke. Exacerbation frequen- cy and health care use are important areas of COPD research [8] and have also attracted attention in the health care quality and performance field in light of readmissions measures proposed by the Centers for Medicare and Medicaid Services [9]. In a previous study, we examined COPD-related ED use over a 365-day follow-up period to learn about longer-term ED return visit frequency [10]. In this complementary study, we used statewide population-based surveillance data to investigate Short-term outcomes of COPD-related ED visits, including hospital admissions and short-term (3- and 14-day) return ED visits.

0735-6757/$ - see front matter (C) 2013 http://dx.doi.org/10.1016/j.ajem.2013.06.010

1394 S.J. Lippmann et al. / American Journal of Emergency Medicine 31 (2013) 1393-1396

Methods

We used public health surveillance data to perform a population- based study of ED visits for COPD. Visits are the unit of analysis. The North Carolina Disease Event Tracking and Epidemiologic Collection Tool (NC DETECT) collects ED visit data from all 24/7 acute care hospital-affiliated civilian EDs throughout the state [11]. The analysis data set included all ED visits made by North Carolina residents 45 years or older between January 1, 2008, and December 31, 2009, inclusive. Institutional review board (IRB) approval (UNC IRB No. 10- 1423) and a data use agreement were obtained from the University of North Carolina IRB and the North Carolina (NC) Division of Public Health, respectively.

Emergency department visits were classified as being COPD related if the first- or second-listed discharge diagnoses contained 1 of the following International Classification of Diseases, Ninth Revision, Clinical Modification codes: 491.* (chronic bronchitis), 492.* (emphy- sema), 493.2* (chronic obstructive asthma), 494.* (bronchiectasis), or 496.* (chronic airway obstruction, not elsewhere classified) [10]. Hospital admissions included those patients admitted to any inpatient hospital unit from the ED, identified in ED disposition codes. Three- day and 14-day return visits were ascertained by examining whether the patient had made another COPD-related ED visit to the same facility within the 3 or 14 days before the current visit, respectively. Patient identifiers in this data system are facility specific; because visits to other facilities cannot be linked, only return visits to the same facility are captured, and the metrics reported here represent minimum return visit prevalences. For ED visits in January 2008, we looked for prior visits in December 2007.

We compared the prevalence of hospital admission and return visits for each time interval (3 and 14 days) by age, sex, and payment method. When a return visit was identified, we also examined permutations of the first and second ED Visit dispositions (eg, to determine how often a return visit was made after a prior visit that required hospitalization or when both/neither visits led to hospital-

ization). Adjusted prevalence ratios (AdjPRs) and 95% confidence intervals (CIs) were estimated using log-binomial regression. All analyses were performed using SAS version 9.2 (Cary, NC).

Results

There were 97 511 COPD-related ED visits made by 64 568 eligible adults in 2008 to 2009 [10]. Of the 64 568 individual patients, 75.1%

(48 508) had only 1 visit during the 2-year period, 14.6% (9401) had 2

visits, and 10.3% (6659) had 3 or more visits.

Hospital admissions

Nearly half (46.3%) of all COPD-related ED visits resulted in hospital admission [10]. The prevalence of hospital admission increased approximately monotonically by 5-year age increments, with patients 90 years or older almost 50% more likely to be hospitalized (AdjPR, 1.49; 95% CI, 1.42-1.57) than their 45- to 49-year-old counterparts (Table 1). There were no differences in likelihood of hospital admission by sex. After adjusting for age and sex, ED visits that were uninsured (self-pay) were far less likely (AdjPR, 0.64; 95% CI, 0.61-0.66) to lead to hospitalization compared with those with private insurance. Medi- care- and Medicaid-paid visits were 5% (AdjPR, 0.95; 95% CI, 0.93-0.97) and 10% (AdjPR, 0.90; 95% CI, 0.88-0.93) less likely to be admitted than private insurance visits, respectively. The crude differences in hospitalization prevalence by payment method were driven, in part, by differences in age among these groups; adjustment for age reduced the size of the prevalence ratios for self-pay and Medicaid.

Return visits

Of the COPD-related ED visits, 1.6% (1607) and 6.2% (6018) were categorized as return visits within 3 and 14 days, respectively. There were no differences by sex for 3-day return visit prevalence (Table 1). Return visits within 14 days were 10% more likely (AdjPR, 1.10; 95%

Table 1

Number, prevalence, and prevalence ratios for hospital admission and return visits resulting from COPD-related ED visits by adults 45 years or older, North Carolina, 2008 to 2009

Population characteristic Total COPD Hospital admissions Return visits^a (3 d) Return visits^a (14 d)

	ED visits, n (column %)	n (%)	AdjPRb (95% CI)		n (%)	AdjPRb (95% CI)		n (%)	AdjPRb (95% CI)
Age (y)
45-49	8650 (8.9)	2835 (32.8)	Reference		132 (1.5)	Reference		521 (6.0)	Reference
50-54	11 153 (11.4)	4079 (36.6)	1.10 (1.05-1.14)		224 (2.0)	1.31 (1.06-1.62)		779 (7.0)	1.16 (1.04-1.29)
55-59	11 776 (12.1)	4901 (41.6)	1.21 (1.17-1.26)		202 (1.7)	1.11 (0.89-1.39)		791 (6.7)	1.11 (1.00-1.24)
60-64	13 413 (13.8)	6128 (45.7)	1.31 (1.26-1.36)		252 (1.9)	1.21 (0.98-1.49)		918 (6.8)	1.13 (1.02-1.26)
65-69	14 001 (14.4)	6910 (49.4)	1.38 (1.33-1.43)		286 (2.0)	1.24 (1.00-1.54)		962 (6.9)	1.12 (1.00-1.25)
70-74	12 488 (12.8)	6416 (51.4)	1.43 (1.38-1.49)		182 (1.5)	0.88 (0.69-1.11)		708 (5.7)	0.92 (0.82-1.03)
75-79	11 107 (11.4)	5795 (52.2)	1.45 (1.40-1.51)		164 (1.5)	0.88 (0.70-1.12)		662 (6.0)	0.96 (0.85 1.08)
80-84	8284 (8.5)	4509 (54.4)	1.52 (1.46-1.58)		109 (1.3)	0.79 (0.61-1.03)		424 (5.1)	0.82 (0.72-0.94)
85-89	4727 (4.9)	2548 (53.9)	1.50 (1.44-1.56)		41 (0.9)	0.52 (0.36-0.74)		193 (4.1)	0.66 (0.56-0.78)
90 +	1912 (2.0)	1025 (53.6)	1.49 (1.42-1.57)		15 (0.8)	0.47 (0.28-0.81)		60 (3.1)	0.51 (0.39-0.66)
Sex
Female	54 053 (55.4)	25 128 (46.5)	Reference		854 (1.6)	Reference		3184 (5.9)	Reference
Male	43 447 (44.6)	20 012 (46.1)	0.99 (0.97-1.00)		753 (1.7)	1.09 (0.99-1.20)		2833 (6.5)	1.10 (1.05-1.16)
Unknown	11 (0.0)	n b 10c			0 (0.0)			n b 10c
Payment method
Noninsured
Self-pay (no insurance)	8465 (8.7)	2375 (28.1)	0.64 (0.61-0.66)		144 (1.7)	1.28 (1.03-1.60)		505 (6.0)	1.27 (1.13-1.42)
Insured
Private insurance	13 810 (14.2)	6673 (48.3)	Reference	177 (1.3)		Reference	635 (4.6)		Reference
Medicare	57 396 (58.9)	28 541 (49.7)	0.95 (0.93-0.97)	959 (1.7)		1.35 (1.10-1.66)	3546 (6.2)		1.46 (1.34-1.59)
Medicaid	10 499 (10.8)	4251 (40.5)	0.90 (0.88-0.93)	190 (1.8)		1.46 (1.24-1.72)	834 (7.9)		1.68 (1.51-1.85)
Other government payments/	1039 (1.1)	463 (44.6)	0.94 (0.87-1.00)	14 (1.3)		1.01 (0.59-1.74)	57 (5.5)		1.15 (0.89-1.50)
workers compensation
Other/unknown/missing	6302 (6.5)	2843 (45.1)	0.90 (0.87-0.93)	123 (2.0)		1.60 (1.27-2.01)	441 (7.0)		1.58 (1.40-1.77)
Total	97 511	45 146 (46.3)		1607 (1.6)			6018 (6.2)

^a Return visits are ED visits in which the patient had a previous COPD-related ED visit within the previous 3 or 14 days.

^b Adjusted for the other variables in the table.

^c Small cell sizes (0 b n b 10) are suppressed to comply with data use agreements.

S.J. Lippmann et al. / American Journal of Emergency Medicine 31 (2013) 1393-1396 1395

CI, 1.05-1.16) to be made by men compared with women. Prevalence of return visits for both time intervals initially increased with age but decreased steadily after the 65- to 69-year-old group. Visits that were uninsured (self-pay), paid by Medicare or Medicaid, or with “other, unknown, or missing” payment method were more likely to be 3- 14- day return visits than those paid by private insurance.

Although many patients were discharged to home at both ED visits (an identified visit and the one found to precede it within either 3 or 14 days), 45.4% (730/1607) of the 3-day and 54.2% (3260/6018) of the 14-day return visits included a hospital admission at one or both visits (Table 2). In 8.0% (128/1607) of the 3-day return visits, the patients were hospitalized at the prior ED visit but returned to the ED within 3 days; for the 14-day visits, 28.9% (1741/6018) returned to the ED despite the patient being admitted at the prior visit. Overall, the most common patterns for 3-day return visits (“prior/return visit disposi- tion”) were as follows: “discharged/discharged” (43.6%; 700), “dis- charged/admitted” (33.7%; 542), “admitted/admitted” (4.4%; 71), and “admitted/discharged” (3.0%; 49). For 14-day returns, the most common patterns were as follows: “discharged/discharged” (35.8%; 2155), “discharged/admitted” (22.7%; 1367), “admitted/admitted” (17.8%; 1072), and “admitted/discharged” (9.7%; 582). See Table 2 for all permutation frequencies.

Discussion

Chronic obstructive pulmonary disease-related ED visits resulted in a high prevalence of hospitalization and 3- and 14-day return ED

visits. The 46.3% prevalence of hospital admissions in our study was similar to a Canadian ED-based cohort (49.3%) [12]. In contrast, the reported hospitalization rates were lower (32.5%) in a population- based administrative database study in Alberta, Canada [5] and higher (61%) in a COPD cohort from the Multicenter Airway Research Collaboration (MARC) study network in the United States and Canada [6]. Unlike the MARC-based study, which found large sex differences in admissions [6], in our study and the Canadian cohort [12], there were no sex differences in hospital admissions [6]. All 3 studies found a positive association between age and admission [6,12]. Uninsured patients and those with Public insurance (Medicare/Medicaid) were less likely to be admitted to the hospital than the privately insured, as suggested by the MARC study [6].

Our findings confirm previous findings demonstrating high return/ relapse rates for COPD-related ED visits. We found that approximately 1 in 16 COPD-related ED visits was made by patients who had made another visit in the previous 14 days and that approximately 1 in 60 visits followed another visit within the previous 3 days. In a Canadian administrative database study of a cohort of patients discharged after a COPD-related ED visit, 5.7% of patients had a subsequent ED visit within 7 days [5]. We include return visits regardless of the disposition at the prior visit; we found that in 8% of the 3-day and in 29% of the 14-day return visits, patients returned to the ED despite being recently admitted to the hospital. By restricting their cohort to only discharged patients, Rowe et al [5] and other studies that include only discharged patients may underestimate the number of return visits. Similar to Rowe et al [5], we found that short-term return ED

Table 2

Disposition patterns for return visit pairs of COPD-related ED visits by adults 45 years or older, North Carolina, 2008 to 2009

First visit (prior visit) disposition Second visit (current/return visit) disposition 3-d return visit pairs 14-d return visit pairs

		n	%		n	%
Admitted to a hospital department	Admitted to a hospital department	71	4.42		1072	17.81

Died 0 0 ^{a a}

Discharged to home or self-care 49 3.05 582 9.67

Left without treatment or against medical advice ^{a a} 12 0.2

Other 0 0 13 0.22

Transferred to another location ^{a a} 30 0.5

Unknown/missing ^{a a} 30 0.5

Discharged to home or self-care Admitted to a hospital department 542 33.73 1367 22.72

Died a a a a

Discharged to home or self-care	700	43.56	2155	35.81
Left without treatment or against medical advice	18	1.12	48	0.8

Other ^{a a} 22 0.37

	Transferred to another location	26	1.62	72	1.2
	Unknown/missing	39	2.43	110	1.83
Left without treatment or	Admitted to a hospital department	21	1.31	33	0.55
against medical advice	Discharged to home or self-care	25	1.56	45	0.75

Left without treatment or against medical advice Other

a	a	a	a
a	a	a	a

Transferred to another location 0 0 ^{a a}

Unknown/missing ^{a a a a}

Other Admitted to a hospital department ^{a a} 12 0.20

Discharged to home or self-care ^{a a} 15 0.25

Left without treatment or against medical advice ^{a a a}

Other 0 0 11 0.18

Unknown/missing ^{a a a a}

Transferred to another location Admitted to a hospital department 12 0.75 37 0.61

Discharged to home or self-care ^{a a} 17 0.28

Transferred to another location ^{a a} 20 0.33

Unknown/missing ^{a a a a}

Unknown/missing	Admitted to a hospital department	26	1.62	70	1.16
	Discharged to home or self-care	11	0.68	69	1.15

Left without treatment or against medical advice 0 0 ^{a a}

Other ^{a a a a}

Transferred to another location 0 0 ^{a a}

	Unknown/missing	21	1.31	124	2.06
Total		1607	100	6018	100

Return visit pairs are defined as the current visit and the most recent prior visit made by the same patient, if the prior visit was within the respective time interval (prior 3 or 14 days).

^a Small cell sizes (0 b n b 10) are suppressed to comply with data use agreements. The total number of suppressed pairs was 46 for 3-day return visit pairs and 52 for 14-day return visit pairs.

1396 S.J. Lippmann et al. / American Journal of Emergency Medicine 31 (2013) 1393-1396

visits were more common among men. However, unlike Rowe et al [5], we found a decrease in return visit prevalence with increased age greater than 70 years; this may be caused by our retrospective approach to defining return visits and not restricting our analysis to discharged patients. Kim et al [7] reported a 21% relapse rate within 14 days after ED treatment of COPD in a follow-up cohort of 140 patients discharged to home; however, that study did not distinguish between follow-up visits made at the ED vs other clinics. Our finding that return visits are more commonly made by men corroborates the suggestive but not statistically significant finding of Kim et al [7] regarding a greater likelihood of relapse among men.

We are not aware of any other studies that have reported disposition patterns for return and prior visit “pairs,” in part because previous studies have focused on follow-up of discharged patients only or have studied Hospital readmissions rather than ED return visits. Given the high prevalence of hospital admissions overall among patients with COPD, it is not surprising that there was a high proportion of return visits that included hospital admission at one or both visits. It is concerning, however, that many hospitalizations occurred either after the patient was discharged at a recent Previous ED visit or even after the patient had been hospitalized after an ED visit in the prior 3 or 14 days. These patterns not only demonstrate the severe illness that many patients with COPD experience but also suggest possible opportunities for interventions to improve care [13]. This is an area that warrants further exploration, especially in light of health care Policy changes that penalize hospitals with high read- mission rates [9].

Limitations

We can only ascertain return visits to the same facility because NC DETECT patient identifiers cannot link patients across facilities [11]; thus, we may be underestimating return ED visits with the underestimate proportional to the frequency with which patients go to different EDs in a short period for COPD-related reasons. The limited clinical data in NC DETECT preclude our ascertaining whether return visits result from new exacerbations or failure to stabilize the original exacerbation [8], as well as our ability to examine clinical factors, such as GOLD assessment ratings [14] and Treatment practices, that likely inform admissions decisions and may predict return visits. Finally, our findings may not be generalizable to the nation as a whole; although the percentage of population 65 years or older in NC is similar to that in other states [15], the COPD prevalence [1] and adult smoking prevalence [16] are slightly higher.

Conclusions

This population-based study describes the short-term outcomes of a large number of COPD-related ED visits using a unique statewide surveillance system. We found a high prevalence of hospital admissions and return ED visits, including some repeat hospitaliza- tions. We also provide new information about how COPD-related hospital admissions and return ED visits vary by age, sex, and insurance status in the population. This information could be used to develop, target, or evaluate clinical, pharmacologic, system, and behavioral interventions to reduce short-term return ED visits among patients with COPD.

Acknowledgments

The authors thank Winston Liao for valuable contributions and insights. The statistical analysis for this research was supported by a University of North Carolina, Gillings School of Public Health Gillings Innovation Laboratory grant. Effort by K.HL was partially supported by Award Number KL2RR025746 from the National Center for Research Resources. M.W. is supported by a VA Research Career Scientist Award from the HSR&D Service. Data were obtained from the NC DHHS/DPH NC DETECT system under a data use agreement. The NC DETECT Data Oversight Committee does not take responsibility for the scientific validity or accuracy of methodology, results, statistical analyses, or conclusions presented.

References

1. Chronic obstructive pulmonary disease among adults-United States. MMWR Morb Mortal Wkly Rep 2012;61(46):938-43.
2. Elixhauser A, Au DH, Podulka J. Readmissions for chronic obstructive pulmonary disease, 2008 [Internet]. Rockville, MD: Agency for Health Care Research and Quality (US); 2011 Sep. Report No.: #121. Available at: http://www.hcup-us.ahrq. gov/reports/statbriefs/sb121.pdf.
3. Brown DW, Pleasants R, Ohar JA, Kraft M, Donohue JF, Mannino DM, et al. Health- related quality of life and chronic obstructive pulmonary disease in North Carolina. N Am J Med Sci 2010;2(2):60-5.
4. Mannino DM, Braman S. The epidemiology and economics of chronic obstructive pulmonary disease. Proc Am Thorac Soc 2007;4(7):502-6.
5. Rowe BH, Voaklander DC, Marrie TJ, Senthilselvan A, Klassen TP, Rosychuk RJ. Outcomes following chronic obstructive pulmonary disease presentations to emergency departments in Alberta: a population-based study. Can Respir J 2010;17(6):295.
6. Tsai C-L, Clark S, Cydulka RK, Rowe BH, Camargo CA. Factors associated with hospital admission among emergency department patients with chronic obstruc- tive pulmonary disease exacerbation. Acad Emerg Med 2007;14(1):6-14.
7. Kim S, Emerman CL, Cydulka RK, Rowe BH, Clark S, Camargo CA. Prospective multicenter study of relapse following emergency department treatment of COPD exacerbation. Chest 2004;125(2):473-81.
8. Soler-Cataluna JJ, Rodriguez-Roisin R. Frequent chronic obstructive pulmonary disease exacerbators: how much real, how much fictitious? COPD: Journal of Chronic Obstructive Pulmonary Disease 2010;7(4):276-84.
9. Horwitz L, Partovian C, Lin Z, Herrin J, Grady J, Conover M, et al. Hospital-wide (all- condition) 30-day risk-standardized readmission measure [Internet]. Available from: https://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment- Instruments/MMS/downloads/MMSHospital-WideAll-ConditionReadmission Rate.pdf; 2011.
10. Yeatts K, Lippmann SJ, Waller A, Hassmiller Lich K, Travers D, Weinberger M, et al. Population-based burden of COPD-related visits in the emergency department (ED): return ED visits, hospital admissions, and comorbidity risks. CHEST [Internet] 2013 Accessed April 15, 2013; Available at: http://journal.publications. chestnet. org/article.aspx?articleID=1676810.
11. Carolina Center for Health Informatics, University of North Carolina at Chapel Hill. The UNC Department of Emergency Medicine Carolina Center for Health Informatics Report, NC DETECT Emergency Department Data: 2008 [Internet]. 2010. Available from: http://www.ncdetect.org/pubs.html.
12. Rowe BH, Villa-Roel C, Guttman A, Ross S, Mackey D, Sivilotti MLA, et al. Predictors of hospital admission for chronic obstructive pulmonary disease exacerbations in Canadian emergency departments. Acad Emerg Med 2009;16(4):316-24.
13. Yip NH, Yuen G, Lazar EJ, Regan BK, Brinson MD, Taylor B, et al. Analysis of hospitalizations for COPD exacerbation: opportunities for improving care. COPD: Journal of Chronic Obstructive Pulmonary Disease 2010 Apr;7(2):85-92.
14. Vestbo J, Hurd SS, Agusti AG, Jones PW, Vogelmeier C, Anzueto A, et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease GOLD executive summary. Am J Respir Crit Care Med 2013;187(4):347-65.
15. U.S. Census Bureau. North Carolina QuickFacts from the US Census Bureau [Internet]. Accessed March 13, 2013. Available at: http://quickfacts.census.gov/ qfd/states/37000.html.
16. Centers for Disease Control, Prevention. CDC-STATE System: Highlights Report

    -“rlinet1tbl15North Carolina [Internet]. Accessed March 13, 2013. Available at: http://apps.nccd.cdc.gov/statesystem/HighlightReport/HighlightReport.aspx? FromHomePage=Y& StateName=North%20Carolina&StateId=NC.