Comparisons of High Versus Low Emergency Department Utilizers in Sickle Cell Disease

Introduction 

Approximately 77,000 Americans have sickle cell disease, an inherited hemolytic anemia affecting Americans primarily of African descent, but also of Mediterranean and other descents.1 Pain in sickle cell disease is ubiquitous, frequent, chronic, and debilitating and is the hallmark of the disease.2 Sickle cell disease pain may often be unpredictable, increased in patients with depression and anxiety,3 and can have a profoundly negative effect on a person's quality of life.4

Severe paroxysmal painful “crises” are episodes, often with no clear trigger, that are hypothesized to result from vaso-occlusion of vascular beds. Crises are responsible for the majority of sickle cell disease medical care contacts.5 However, most pain, even that which patients label as a crisis, is treated at home.6

Sickle cell disease patients who are heavy utilizers of the emergency department (ED) for pain relief have been stigmatized7 because they are perceived as abusing or misusing opioids. This perception is related to particular behavior patterns, including obtaining opioids from more than 1 provider on different occasions, requiring higher doses of opioids than other patients, long-term opioid use, or for being a “clock watcher” as relates to timing of opioid dosing.8 Although such behaviors can indicate opioid abuse or misuse, they can also represent inadequate pain control. Other studies in the ED document poor assessment and treatment of pain in general.9, 10, 11, 12, 13

Opioid analgesics are a mainstay of therapy for acute pain crises in sickle cell disease. Therefore, by adulthood, most patients have had more than 15 years of intermittent exposure to opioid analgesics. This chronic exposure may result in tolerance to opioids, requiring higher-than-average doses for pain relief. This tolerance may be perceived by health care providers as opioid abuse or misuse by sickle cell disease patients. This perception may be especially about sickle cell disease patients who frequent the ED. Physicians and nurses may greatly overestimate the prevalence of addiction in patients with chronic pain, especially those with sickle cell disease. In one study by Shapiro, 53% of emergency physicians and 23% of hematologists believed that greater than 20% of adult patients with sickle cell disease were addicted to opioids.14 The same study showed that 22% of emergency physicians and 9% of hematologists believed that more than 50% of the patients were addicted.14 There are no epidemiologic studies that have assessed the severity of illness or pain of sickle cell patients who utilize the ED and in particular those who are frequent utilizers of the ED. We therefore used existing prospective data from the Pain in Sickle Cell Epidemiology Project (PiSCES) study to quantify ED use in patients with sickle cell disease and compare clinical and demographic characteristics of sickle cell disease patents who were frequent utilizers versus those who seldom or never come to the ED. Using this cohort of patients from a large epidemiologic study, we sought to better characterize sickle cell disease patients who frequently utilize the ED.

Materials and Methods 

Primary Data Analysis 

χ2 Tests were used to compare categorical variables, t tests were used for continuous variables, and Wilcoxon rank sum test was used with ordinal variables to compare demographic, laboratory, clinical, and quality of life variables between high and low ED utilizers at baseline. Analysis of covariance was used to compare percentage of home opioid used among high and low ED utilizers, controlling for pain intensity. Mixed-model ANOVA was used to compare daily pain and distress (as measured by the diary) between high and low utilizers. For binary variables representing pain days, crisis days, generalized estimating equation methods were used. The mixed-model and generalized estimating equation methods allowed the diary day to be used as the unit of analysis, controlling for the correlation between days within subject (because of subjects providing diaries on multiple days). Differences between low and high utilizers were calculated, along with confidence intervals (CIs). When zero was not included within the CI, differences could be considered significant. Because many hypothesis tests were performed, the likelihood of spurious significance tests was increased. To control for this, for each set of analyses (Table 1, Table 2, Table 3) the Bonferroni method was used to adjust P values or CIs, so that the overall significance level was .05 for each table. Thus, an individual significance level of .05/18=.00278 was used for Table 1, .05/6=.0083 for Table 2 and .05/2 for Table 3 analyses. All analyses were performed with SAS version 9.1 statistical software (SAS Institute, Inc., Cary, NC).

Table 1. Demographic and clinical sample description measured at baseline, by ED utilization category.

	Demographics	Low ED Utilizer,⁎	High ED Utilizer,⁎
		N=150,	N=82,
		No. (%)	No. (%)
	Sex
	Female	91(60.7)	52(63.4)
	Age, y
	16–24	32(21.3)	19(23.2)
	25–34	38(25.3)	31(37.8)
	35–44	43(28.7)	23(28.1)
	45–54	29(19.3)	6(7.3)
	54–64	8(5.3)	3(3.7)
	Education
	<High school	18(12.0)	10(12.2)
	High school	56(37.3)	32(39.0)
	>High school	76(50.7)	40(48.8)
	Income, $
	<10,000	49(33.3)	39(48.7)
	10,000-19,999	33(22.5)	19(23.7)
	20,000-29,999	23(15.7)	11(13.7)
	>30,000	42(28.6)	11(13.7)
	Clinical variables
	Genotype
	Sβ+ thalassemia	5(3.3)	1(1.2)
	β0 Thalassemia	4(2.7)	1(1.2)
	Hemoglobin SC	42(28.0)	14(17.3)
	Hemoglobin SS	99(66.0)	65(80.2)
	Anxiety	5(3.4)	10(12.2)
	Depression	38(25.3)	26(31.7)
	Hydroxyurea use	30(20.0)	26(31.7)
	Specialty center	75(52.4)	35(43.2)
	Transfusion in past 3 mo†	25(16.9)	36(44.4)
	Comorbidities
	TIA or stroke	17(11.4)	13(15.8)
	Kidney failure	8(5.4)	2(2.4)
	Avascular necrosis	24(16.1)	24(29.3)
	Laboratory values, mean (SD)
	HCT†	28.4(5.9)	25.1(4.8)
	WBC	10.4(3.9)	12.1(5.0)
	% F	3.7(7.3)	4.2(7.4)
	Quality of life
	SF-36 mental function	48.1(10.3)	45.8(12.5)
	SF-36 physical function†	36.0(10.6)	32.7(8.7)

TIA; Transient ischemic attack; HCT, hematocrit; WBC, white blood cell, %F, percent fetal hemoglobin. SF-36, Medical Outcome Study 36 Item Short Form.

⁎ Low ED utilizer: <3 ED visits/year; high ED utilizer: ≥3 ED visits/year. † p < 0.00278.

Table 2. Pain and distress levels reported on diary days stratified by utilization class.

	Outcome measures	Low ED Utilizer,⁎	High ED Utilizer,⁎	High Versus Low Utilizers,
		Mean (SD)	Mean (SD)	Difference (99.1% CI)
	Pain intensity on pain days	3.97(0.13)	4.87(0.16)	0.90(0.34to1.46)
	Pain intensity on noncrisis pain days	3.77(0.13)	4.44(0.17)	0.63(0.09to1.24)
	Pain intensity on crisis days	4.99(0.16)	6.05(0.18)	1.05(0.42to1.69)
	Distress intensity on pain days	2.63(0.17)	3.52(0.22)	0.89(0.15to1.63)
	Distress intensity on noncrisis pain days	2.45(0.16)	3.03(0.20)	0.58(–0.10to1.26)
	Distress intensity on crisis days	3.63(0.24)	4.78(0.26)	1.15(0.21to2.10)

⁎ Low ED utilizer: <3 ED visits per year; high ED utilizer: ≥3 ED visits per year.

Table 3. Probability of pain or crisis reported on diary days, stratified by utilization class.

	Outcome measures	Low ED Utilizer,*	High ED Utilizer,*	High Versus Low Utilizers,
		Mean (SD)	Mean (SD)	Difference (97.5% CI)
	Pain days	0.47(0.03)	0.73(0.04)	0.26(0.14–0.37)
	Crisis days	0.12(0.02)	0.22(0.03)	0.10(0.02–0.19)

†Low ED utilizer: <3 ED visits per year; high ED utilizer: ≥3 ED visits per year.

Results 

A total of 308 patients enrolled in PiSCES and completed baseline surveys. Twenty-three subjects did not fill out any diaries, whereas a further 43 completed fewer than 30 of a possible 188 diaries that could be completed in 6 months. The study sample consists of the 232 subjects who completed at least 30 days of daily diaries. This cohort of 232 patients submitted a total of 31,294 diaries. Among included patients, 16% of patients completed fewer than 60 diaries, 17% completed 60 to 120 diaries, and 67% completed more than 120 diaries. Subjects in this analysis cohort were similar on baseline demographics to those excluded except for age, where excluded subjects were younger [30.2 versus 34.3; difference is 4.1 (95% CI 1.3, 7.0)].

In our cohort, the estimated number of ED visits per year ranged from 0 to 125 visits (Figure). Eighty-two subjects (35%) had an estimated 3 or more ED visits per year and constitute our high utilizer group. Of the 150 subjects in the low utilizer group, 123 reported no ED visits. The number of diaries completed by subjects in the low utilizer group did not differ statistically from the number completed by the high utilizer group.

View Large Image Download to PowerPoint

Figure. Estimated distribution of self-reported ED visits per year.

Demographics of the cohort are presented in Table 1, comparing high and low utilizers of the ED. More than half of studied patients were women, approximately 50% had greater than a high school education, and 80% were younger than 45 years. Genotype distribution was consistent with that of other studies in that most patients had homozygous sickle cell anemia (Hb SS). High ED utilizers had lower hematocrit levels and more transfusion therapy than subjects who used the ED infrequently. Although the data were not statistically significant, high ED utilizers trended toward a higher incidence of anxiety, higher incidence of avascular necrosis, and higher WBC count. Although subjects did not differ on the mental summary scale of Medical Outcome Study 36 Item Short Form quality of life, there were significant differences in the physical functioning summary scale. The 2 groups did not differ significantly on depression.

Table 2 shows that frequent ED utilizers consistently reported higher mean pain and distress levels than other sickle cell disease patients, regardless of whether only pain days were considered, only noncrisis pain days were considered, or only crisis days were considered (with the exception of distress on noncrisis days). We also found an almost 2-fold increase in the probability of being in pain or reporting a crisis on the diary (Table 3). Subjects in the high ED utilizer group took opioids on pain days at home significantly more often than those in the low utilizer ED group (62.9% of days versus 35.9% of days). But when we controlled for mean pain levels, the difference between groups was no longer significant (47.4% versus 44.1).

Limitations 

This was a longitudinal study based on patients' self-report of pain and utilization. The frequency of ED visits is therefore self-reported. However, we believe the patient's account is more likely to represent underreporting than overreporting of their ED utilization. In addition, the annual utilization rates were extrapolated from the number of utilization days during the study period. We acknowledge that the lack of a full year of data is a limitation that may overestimate or underestimate utilization; however, analysis of patients with different levels of diary day contributions showed no differences in utilization between the groups. We believe the number of days studied is sufficient to make predictions about pain and utilization patterns. Comorbid conditions, quality of life data, and biological data were assessed at baseline and were not reassessed post–study completion. The data are representative of 1 state, and therefore generalizations may be limited.

Discussion 

Our results are among the first to demonstrate the proportion of sickle cell disease patients who are frequent ED utilizers and to distinguish them from other sickle cell disease patients according to their medical characteristics and their pain and distress levels.

We found that although a third of the subjects in the study utilized the ED 3 or more times per year, according to daily diaries, these patients were more symptomatic with their disease, having more pain and more crises and requiring more opioid use. However, after controlling for number of pain days and the magnitude of pain, frequent ED users did not use opioids any more often than other sickle cell disease patients. Frequent ED utilization was associated with higher clinical severity of the disease, as reflected in hematocrit values and need for transfusion treatment. Frequent ED utilization was also associated with decrements in physical functioning according to Medical Outcome Study 36 Item Short Form quality-of-life measures.

There were no differences between high utilizers and low utilizers in sex, age, education, or income, nor was frequent ED utilization associated with whether the source of the patient's ambulatory care was a sickle cell specialist versus another physician.

Our results may correct some unfortunate stereotypes that may negatively affect the health care provided to these patients. Sickle cell disease patients who are frequent ED utilizers should be expected to utilize more services because they are more severely ill, have higher anxiety, and experience poorer functional status from their disease, not because they are opioid abusers. In fact, our results mirror that of many studies of other diseases that show that frequent users of EDs, hospitals, and ambulatory primary care are those who have significant acute and chronic medical problems.21, 22, 23, 24 Frequent users of EDs more often self-report fair to poor health status compared with non-ED users or those who use the ED only once or twice in a given year.23 Frequent ED users also more often perceive their pain or other symptoms as a threat to life or personal autonomy.25

Previous studies26, 27, 28 suggest that heavy users of EDs are likely to have serious social problems, including alcohol abuse and psychiatric dysfunction. We did find more anxiety among sickle cell disease patients who were frequent ED users. However, paradoxically, in another report from this study, we found alcohol abusers with sickle cell disease had fewer unscheduled clinic visits, ED visits, hospital days, and any health care utilization for sickle cell disease. This was only statistically significant for ED visits.29

Our results represent analyses that address potential misperceptions by health care providers, especially ED physicians, nurses, and case managers in urban cities who treat a substantial amount of sickle cell disease patients. Emergency physicians are challenged with determining how to manage a given ED visit for sickle cell disease. They must dispense sometimes large amounts of opioids, both in the ED and as a prescription after ED care, to a patient they are unsure they can trust really requires the opioid. They also must provide a means of follow-up care for this patient, who often may not have a regular primary care physician, let alone a sickle cell specialist who can provide continuity of expert care. A recent study suggests this dilemma arises often in ED care.30 Thus, sickle cell disease patients who are “frequent flyers” in urban EDs may provoke uncertainty and mistrust in emergency physicians who repeatedly care for them.

Although not addressed in our study, race and socioeconomic barriers may affect the quality of care received. Further, because the majority of sickle cell patients are of African descent, cultural and racial barriers may arise, preventing prompt and proper treatment during a painful crisis.31 In a qualitative study, patients reported being neglected, being stigmatized, and experiencing mistrust from health care providers.31 Particularly, pain management of a given sickle cell disease visit can be problematic.32 Disparities in pain management persist, with less analgesic use in EDs for blacks and Hispanics compared with whites.33 Emergency physicians may suspect drug-seeking behavior among sickle cell disease patients according to ethnicity/race, requests for specific drugs and doses, or administration of large doses of opioid without relief. This in turn may reflect the general lack of knowledge about pain management among health care professionals.32

We believe our results demonstrate clearly here that some of the concern associated with managing sickle cell disease in EDs may be unfounded. Sickle cell disease patients who utilize the ED frequently are truly sicker patients. It is likely that these patients' requests for opioids are in line with their pain at home and their overall disease severity when pain magnitude and frequency of pain days are taken into consideration. Emergency physicians treating sickle cell disease pain in the ED and planning follow-up of an ED visit for sickle cell disease should factor into their treatment plans the patient's background and severity of disease. Clinicians should be assured that frequent ED visitors with sickle cell disease are more severely ill and needy, and they should be more sensitive to the natural progression of sickle cell disease and the stigmas these patients endure.

Further, we urge EDs to establish a better comprehensive management program for sickle cell disease ED care, with the goal of more timely and compassionate pain management and improved patient satisfaction. In summary, we believe our data suggest that emergency physicians see a biased sample of sickle cell patients, those who have a good deal more pain and distress than the average sickle cell patient and therefore may have more ED utilization.