- Email: [email protected]
Insurance Status, but Not Race, Predicts Perforation in Adult Patients with Acute Appendicitis
Insurance Status, but Not Race, Predicts Perforation in Adult Patients with Acute Appendicitis Fredric M Pieracci, MD, MPH, Soumitra R Eachempati, MD, FACS, Philip S Barie, MD, MBA, FACS, Mark A Callahan, MD Delay in treatment is a strong risk factor for perforation during acute appendicitis. In addition, lower socioeconomic status has been linked to impaired access to surgical care. We sought to examine the relationships among race, insurance status, and perforation in a recent, adult population with acute appendicitis. STUDY DESIGN: Data on adult patients with acute appendicitis were abstracted from the New York State Statewide Planning and Cooperative Systems Database for the years 2003 and 2004. A multiple logistic regression model, which adjusted for patient, community, and hospital factors, was used to examine the independent effects of both race and insurance status on likelihood of perforation. RESULTS: A total of 29,637 patients had acute appendicitis; 7,969 (26.9%) of these were perforated. Although Caucasian patients were more likely to perforate compared with minority patients, by univariate analysis, adjustment for age alone eliminated this disparity. In addition, by multivariable analysis, no difference existed in odds of perforation for Caucasian patients compared with African-American (odds ratio [OR] ⫽ 1.03, 95% CI [0.93, 1.15], p ⫽ 0.52), Hispanic (OR ⫽ 0.99, 95% CI [0.90, 1.08], p ⫽ 0.82), or Asian patients (OR ⫽ 0.85, 95% CI [0.73, 1.00], p ⫽ 0.05). But compared with privately insured patients, uninsured patients (OR 1.18, 95% CI [1.07 to 1.30], p ⫽ 0.0005), Medicaid patients (OR ⫽ 1.22, 95% CI [1.12 to 1.33], p ⬍ 0.0001), and Medicare patients (OR ⫽ 1.14, 95% CI [1.03, 1.25], p ⫽ 0.01) were significantly more likely to have perforation. CONCLUSIONS: Race does not appear to be an important variable in predicting perforation in adult patients with acute appendicitis, but the likelihood of perforation varies significantly according to insurance status. Future research is necessary to both understand and have an impact on this socioeconomic disparity. (J Am Coll Surg 2007;205:445–452. © 2007 by the American College of Surgeons) BACKGROUND:
The model of acute appendicitis has been studied widely as a gauge of effective and equitable health care. Use of this model affords several methodologic advantages because biologic, behavioral, and socioeconomic variability are unlikely to influence the probability that appendicitis will develop.1-3 Nearly all cases of appendicitis eventually result in hospitalization,3 at which point, if surgery is indicated, two distinct options exist (laparoscopic versus open appendectomy). Both racial and insurance-related disparities have been well described in the management of appendicitis,4-6 and
current focus has shifted from documentation of these disparities to analysis of their causes. Using the New York State administrative database, we reported recently that differential hospital use is an important component of both racial and insurance-related disparities in the surgical management of appendicitis.7 Specifically, both Caucasian and privately insured patients with acute appendicitis are more likely to present to hospitals that perform laparoscopic appendectomy with regularity. This differential hospital use represents one mechanism by which access to surgical care is impaired for certain patient populations. Appendiceal perforation signifies a second access marker of interest in the management of appendicitis. Compared with patients without appendiceal perforation, those with a perforated appendix incur increased morbidity, length of stay, and health-care costs.8 Although certain biologic factors, such as increasing age
Competing Interests Declared: None. Received October 18, 2006; Revised March 28, 2007; Accepted April 9, 2007. From the Departments of Surgery (Pieracci, Eachempati, Barie, Callahan) and Public Health (Pieracci, Eachempati, Barie), Weill Medical College of Cornell University, New York, NY. Correspondence address: Fredric M Pieracci, MD, 411 East 69th St, #KB220, New York, NY 10021.
© 2007 by the American College of Surgeons Published by Elsevier Inc.
445
ISSN 1072-7515/07/$32.00 doi:10.1016/j.jamcollsurg.2007.04.010
446
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
and comorbidities, influence the likelihood of perforation,9 time to operation is the most important predictor of perforation once symptoms occur.10-12 Delays in surgery are accounted for by the patient’s decision to seek timely medical attention, access to both primary and hospital care without cost concerns, and the quality of care once presentation has occurred. Minority populations, the uninsured, and those with public insurance are less likely to have a usual source of primary care,13 and hospitals that treat these populations disproportionately demonstrate worse surgical outcomes.14-16 Compared with patients with insurance, the uninsured are more likely to delay seeking medical care for fear of financial repercussions.17,18 So it is plausible that these patient populations are at increased risk for perforation once appendicitis has occurred. Recent studies addressing this issue have focused on the pediatric population, reporting both racial19-21 and insurance-related disparities20,21 in the likelihood of perforation in children with appendicitis. Although Bravemen and associates22 documented similar disparities in adult patients with appendicitis, this study was conducted nearly 2 decades ago, and the exclusion of elderly patients (age ⬎ 65 years) precluded their examination. Indeed, significant racial disparities in the management of other surgical diseases have been well documented in the elderly Medicare patient population.14,23,24 The objectives of this study were to examine the relationships among race, insurance status, and likelihood of perforation in adult patients with acute appendicitis. METHODS Data were extracted from the New York State Statewide Planning and Research Cooperative System (SPARCS) database for the years 2003 and 2004 using SAS version 9.1 (SAS Institute). The SPARCS database is a legislatively mandated patient data system that contains information on all patients discharged from an acute-care, nonfederal hospital in New York State. A total of 15 diagnostic fields and 15 procedure fields (principal and other 1 to 14), which are based on the International Classification for Disease, 9th revision (ICD-9) classification system, are available for each patient discharge. For this analysis, we created a data set that included all adult (age ⱖ 18 years) patients with the diagnosis of acute appendicitis (540.0, 540.1, or 540.9) in any diagnosis field. The primary outcomes variable was appendiceal perforation, defined as the presence of a diagnos-
J Am Coll Surg
tic code for either perforated appendicitis without abscess (540.0), or perforated or nonperforated appendicitis with peritoneal abscess (540.1) in any diagnostic field. All other cases of appendicitis were considered nonperforated. Patients with diagnosis codes for both perforated and nonperforated appendicitis were included in the perforation group. The main predictor variables were race and insurance status. Race was grouped according to SPARCS database categories for race and ethnicity, which are based on current federal standards defined by the Office of Management and Budget.25 For this analysis, the SPARCS race and ethnicity fields were combined to form five race groups: Caucasian, Hispanic, African American, Asian, and other. Patients lacking race information were excluded from analysis. Insurance status was based on the SPARCS database field for primary reimbursement source. Five groups were created: private insurance, Medicare, Medicaid, uninsured, and other. Additional biologic characteristics abstracted were age (grouped as 18 to 40, 41 to 64, and ⱖ 65 years), gender, and comorbid conditions. Comorbidities were quantified using the Charlson comorbidity index26 adapted for use with administrative data by Deyo and associates.27 All diagnostic fields were queried for the presence of relevant comorbid conditions. Because the median comorbidity score was 0 (range 0 to 5), and a small proportion of subjects had at least 1 Charlson comorbid condition coded (n ⫽ 1,577, 5.3%), this variable was dichotomized as either zero or ⱖ 1. The following patient community factors, obtained from US census data for the year 2000 and based on subject ZIP code, were abstracted: the percentage of families living below the federal poverty line (⬍ 10%, 10% to 24.9%, ⬎ 25%); the percentage of adults who had achieved at least a high school education (dichotomized around the median value of 43.0%); and the percentage of rural residences (⬍ 10%, 10% to 50%, ⬎ 50%). Hospital characteristics known to influence the likelihood of perforation and so were abstracted were admission source (emergency department versus other), hospital teaching status, and volume of patients with acute appendicitis treated during the study period (divided into quartiles). All statistical analyses were computed using SAS version 9.1 (SAS Institute). Assessment of differences in nominal categorical variables was performed using the chi-square test. All p values were two-sided, with sta-
Vol. 205, No. 3, September 2007
Pieracci et al
tistical significance evaluated at the 0.05 ␣ level, with the following exception. Because multiple comparisons were conducted when comparing individual race (Caucasian versus African American, Hispanic, Asian, and other) and insurance (private insurance versus Medicare, Medicaid, uninsured, and other) groups, a Bonferroni-corrected ␣ levelwasused.TheBonferronicorrected ␣ level was calculated by dividing the standard ␣ error level of 0.05 by the number of tests performed (0.05 / 4 ⫽ 0.0125 in the case of both race and insurance status).28 Multiple logistic regression, which included predictor variables significant at the 0.25 level by univariate analysis,29 was used to obtain adjusted odds of perforation. All variables were entered into the model as outlined previously using k – 1 indicator variables, where k is equal to the number of levels within the variable. To avoid overcontrolling, which may result from the addition of multiple, correlated socioeconomic factors to the regression model,30 a model that controlled for age only (the strongest predictor of perforation by univariate analysis) was tested first. Next, a model that contained each of the previously mentioned patient, community, and hospital factors was tested to obtain fully adjusted odds of perforation. Variables were added to the model using a forward selection method according to descending magnitude of unadjusted odds ratios (ORs). RESULTS A total of 31,245 patients had appendicitis, and race information was available for 29,637 (94.9%) of them. Of these 29,637 patients, 7,969 (26.9%) had perforation. Additional sample characteristics, and the results of univariate analysis, are shown in Table 1. The likelihood of perforation varied significantly by all variables tested. The prevalence of perforation increased markedly with age; more than one-half of patients aged 65 years or older had perforation. In addition, subjects with a comorbidity score ⱖ 1 were significantly more likely to have perforated as compared with those with a comorbidity score of zero (44.6% versus 25.9%, respectively, p ⬍ 0.0001). Caucasian patients demonstrated the highest prevalence of perforation by univariate analysis (28.6%), followed by African-American (27.0%), Hispanic (23.8%), and Asian patients (23.6%). Analysis of insurance status revealed that Medicare patients had the highest prevalence of perforation (52.0%), followed by privately insured (24.3%) and Medicaid
Insurance Status Predicts Appendiceal Perforation
447
(23.9%) patients. Patients without insurance had the lowest prevalence of perforation (22.5%). Finally, patients who resided in communities with the lowest levels of poverty were most likely to suffer perforation (28.0%). The higher observed prevalence of perforation in patients who were Caucasian, privately insured, and resided in wealthy communities was suspected to result from confounding by age, so it was examined further by multivariable analysis. Finally, male gender, lower community levels of education, an increased proportion of rural residences, an admission source other than the emergency department, and treatment at both a nonteaching and lower-volume hospital were associated with an increased likelihood of perforation. Confounding by age explained all of the variability in likelihood of perforation by race. As shown in Table 2, there were no longer significant differences in the likelihood of perforation when comparing Caucasian with African-American, Hispanic, and Asian patients after adjusting for age alone. In addition, these associations remained nonsignificant after adjustment for the remaining covariates. By contrast, significant differences in likelihood of perforation between insurance groups remained after adjusting for age and other covariates used in the regression model. Specifically, as compared with patients with private insurance, uninsured (OR ⫽ 1.18, 95% CI [1.07, 1.30], p ⫽ 0.0005), Medicaid (OR ⫽ 1.22, 95% CI [1.12, 1.33], p ⬍ 0.0001), and Medicare patients (OR 1.14, 95% CI [1.03, 1.25], p ⫽ 0.01) were all more likely to have perforation. We also performed a stratified analysis of Medicare patients aged 65 years and older (n ⫽ 2,432) to eliminate possible confounding of the relationship between race and perforation by insurance status (Table 3). By univariate analysis, minority patients were not more likely to have perforation than Caucasian patients. Controlling for the previously mentioned patient, community, and hospital factors did not affect this relationship. DISCUSSION In this study of a statewide administrative database, we observed that race was not associated with likelihood of perforation in adult patients with appendicitis, in contrast to previous reports. But insurance status was independently associated with perforation; uninsured, Medicaid, and Medicare patients were all significantly more likely to have perforation compared with patients with private insurance. These relationships persisted after
448
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
J Am Coll Surg
Table 1. Sample Demographics and Univariate Analysis Perforated†
Total* Variable
n
n
%
29,637
n
%
p Value‡
7,969
26.9
—
28.6
Main predictor variables ⬍ 0.0001
Race Caucasian
18,598
62.8
5,314
Hispanic
4,465
15.1
1,063
23.8
African American
2,684
9.1
725
27.0
Asian
1,193
4.0
281
23.6
Other
2,697
9.1
586
21.7
17,149
57.9
4,166
24.3
Medicare
3,128
10.6
1,626
52.0
Medicaid
5,126
17.3
1,224
23.9
Uninsured
3,691
12.5
831
22.5
543
1.8
122
22.5
18–40
16,348
55.2
2,831
17.3
41–64
10,430
35.2
3,567
34.2
ⱖ 65
2,859
9.7
1,571
54.9
Male
16,125
54.4
4,451
27.6
Female
13,512
45.6
3,518
26.0
Zero
28,060
94.7
7,266
25.9
ⱖ1
1,577
5.3
703
44.6
⬍ 10
14,095
47.6
3,956
28.1
10–25
10,999
37.1
2,854
26.0
⬎ 25
4,543
15.3
1,159
25.5
⬍ 10
23,234
78.4
6,058
26.1
10–50
3,679
12.4
1,072
29.1
⬎ 50
2,724
9.2
839
30.8
Below median, ⬍ 43
14,275
48.2
3,938
27.6
Above median, ⱖ 43
15,362
51.8
4,031
26.2
26,763
90.3
7,012
26.2
2,874
9.7
957
33.3
21,447
72.4
5,650
2.3
8,190
27.6
2,319
28.3
937
3.2
291
31.1
Low
4,143
14.0
1,229
29.7
High
6,506
22.0
1,711
26.3
18,051
60.9
4,738
26.2
⬍ 0.0001
Insurance Private
Other Biologic factors
⬍ 0.0001
Age, y
Gender
0.002
⬍ 0.0001
Comorbidity score
Community factors ⬍ 0.0001
Below federal poverty line, %
⬍ 0.0001
Rural residences, %
High school educated, %
0.01
Hospital factors ⬍ 0.0001
Admission source Emergency department Other Teaching status Teaching Nonteaching
0.001
Hospital volume Lowest
Highest
0.004
*Refers to the number of patients in that group divided by the total sample size. † Refers to the number of patients with a perforation over the total number of patients in that group. ‡ Refers to results from chi-square analysis for significant differences in likelihood of perforation between groups.
Vol. 205, No. 3, September 2007
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
449
Table 2. Multivariable Analysis of Predictors of Appendiceal Perforation Variable
Race Caucasian African American Hispanic Asian Other Insurance Private insurance Uninsured Medicaid Medicare Other
Unadjusted odds ratio (95% CI)
Age-adjusted odds ratio (95% CI)
Fully adjusted† odds ratio (95% CI)
1.00 0.93 (0.85, 1.01) 0.78 (0.72, 0.84)* 0.77 (0.67, 0.88)* 0.69 (0.63, 0.77)*
1.00 1.08 (0.98, 1.19) 1.06 (0.98, 1.15) 0.86 (0.75, 0.99) 0.91 (0.82, 1.07)
1.00 1.03 (0.93, 1.15) 0.99 (0.90, 1.08) 0.85 (0.73, 1.00) 0.84 (0.75, 0.94)*
1.00 0.91 (0.83, 0.99) 0.98 (0.91, 1.05) 3.37 (3.12, 3.64)* 0.90 (0.74, 1.11)
1.00 1.16 (1.06, 1.27)* 1.18 (1.10, 1.28)* 1.16 (1.05, 1.27)* 1.05 (0.85, 1.29)
1.00 1.18 (1.07, 1.30)* 1.22 (1.12, 1.33)* 1.14 (1.03, 1.25)* 1.01 (0.82, 1.24)
*p ⬍ 0.0125 (Bonferroni-corrected ␣ error level ⫽ 0.05 / 4). † Controlling for age, gender, comorbidity score, insurance, race, community poverty level, community education level, community level of urbanization, hospital teaching status, and hospital volume.
controlling for patient, community, and hospital factors, each of which was significantly associated with likelihood of perforation by univariate analysis. To our knowledge, only one other study has addressed these relationships in an adult population. Braveman and colleagues22 used State of California data from 1984 to 1989 to report that, compared with patients with private insurance, patients with Medicaid (OR 1.49, 95% CI [1.49, 1.59]) and without insurance (OR 1.46, 95% CI [1.39, 1.54]) were significantly more likely to have perforation. Age, gender, race, comorbidities, community poverty level, hospital teaching status, and hospital volume were controlled for, and patients older than 65 years were excluded secondary to near-universal hospital reimbursement through Medicare (in our sample, by contrast, nearly 15% [n ⫽ 427] of elderly patients used private insurance as their principal method of hospital payment). Our findings are consistent with those reported by Braveman and coauthors, although the effect size of our study is somewhat attenuated. These data, in addition to those in similar reports in the pedi-
atric literature,20,21 provide evidence for an independent effect of insurance status on likelihood of perforation. This effect may be mediated through several mechanisms. Because both uninsured and publicly insured patients are less likely to have a usual source of care,31,32 difficulty in consulting a provider may delay hospitalization. Longer waiting times once care has been sought may also contribute to delayed treatment. Second, uninsured patients may delay seeking care for fear of financial repercussions. In one large multicenter study, uninsured patients were 9.5 times as likely to report delaying care because of cost concerns as compared with privately insured patients.17 Delays in seeking care were 60% more frequent for emergent versus elective admissions. Several alternative explanations for our findings must also be considered. Socioeconomic variables, such as race, income, and community resources, frequently confound the relationship between insurance status and adverse health outcomes. But previous reports have not found an association between individual or community income and the likelihood of perforation in patients
Table 3. Subgroup Analysis of Likelihood of Appendiceal Perforation among Medicare Patients Aged 65 Years or Older Race
Caucasian African American Hispanic Asian Other
Unadjusted odds ratio (95% CI)
p Value*
Fully adjusted† odds ratio (95% CI)
p Value*
1.00 1.12 (0.73, 1.71) 0.83 (0.52, 1.33) 0.75 (0.39, 1.43) 1.10 (0.64, 1.91)
— 0.52 0.31 0.26 0.66
1.00 1.13 (0.78, 1.63) 0.85 (0.58, 1.26) 0.81 (0.48, 1.37) 1.15 (0.73, 1.80)
— 0.52 0.43 0.43 0.54
*p ⬍ 0.0125 (Bonferroni-corrected ␣ error level ⫽ 0.05 / 4). † Controlling for age, gender, comorbidity score, community poverty level, community education level, community level of urbanization, hospital teaching status, and hospital volume.
450
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
with appendicitis.19,20 In addition, both race and community poverty level were controlled for in this analysis. Additional factors that may delay treatment33-35 were also addressed in our analysis. Specifically, we controlled for community level of education, degree of community urbanization, admission source, hospital teaching status, and hospital volume. But although all of these variables were significantly associated with likelihood of perforation by univariate analysis, their addition to the regression model did not affect the significance of the relationship between insurance status and perforation. The lack of an association between race and perforation was an unexpected finding. Minority patients have been documented to use fewer health-care services,36-38 have altered preferences about the utility of surgical interventions,39,40 receive recommendation for necessary procedures less often,41,42 and receive care at hospitals that incur worse outcomes.15,16 Each of the previously mentioned disparities would suggest that this patient group is at increased risk for appendiceal perforation. In addition, reports on both adult22 and pediatric19-21 populations have observed that minority patients are significantly more likely to have perforation compared with Caucasian patients. By contrast, our results did not find a relationship between race and likelihood of perforation in both our overall sample analysis and in a subgroup analysis of elderly Medicare patients. Because differences in odds of perforation among Caucasian, African-American, and Hispanic patients were eliminated after controlling for age only, it is unlikely that confounding by additional socioeconomic factors explained the original disparities. One possible explanation is that an improved awareness of racial and ethnic disparities within the medical community has led to a reduction in these disparities. Abatement of racial disparities over time was noted recently in Medicare patients undergoing coronary artery bypass grafting.43 In addition, a recent single-institution study of pediatric patients with acute appendicitis observed neither racial nor insurance-related disparities in the likelihood of perforation.44 Our decision to include elderly patients was based on the following two reasons: First, the predominant use of Medicare for hospital reimbursement facilitated subgroup analysis of the effect of race on perforation in this population. Second, we were able to distinguish between hospital reimbursement through Medicare and the nearly 15% of cases in which private insurance was
J Am Coll Surg
the principal method of hospital payment. Elderly patients with private insurance may also be less inclined to delay treatment because of potential financial repercussions. In this case, concerns may involve the possibility of extended hospitals stays, costly outpatient prescription drugs, and longterm care. This claim is supported by the 14% adjusted increased odds of perforation in patients with Medicare as compared with patients with private insurance. Health policy implications based on our results must still be considered with caution. Although insurance status appears to be associated with perforation in patients with acute appendicitis, it is not clear that elimination of insurance variability will reduce or eliminate delays in treatment. Indeed, substantial disparities in access to care have been observed in countries with compulsory health insurance.45-47 So it is possible that patients with and without private insurance differ systematically with respect to unmeasured variables. But our results indicate that insurance status has an independent effect on the likelihood of appendiceal perforation among the variables tested. Patient characteristics analyzed in this study were limited to the fields available within the SPARCS database. Accordingly, covariates such as level of education and individual income were approximated using community-level data based on patient ZIP codes. Because race information was available in 94.9% of patients with appendicitis, exclusion bias was minimal. But nearly 10% of our sample was composed of patients whose race was coded as “other.” Interestingly, by multivariable analysis, this patient group remained significantly less likely to perforate compared with Caucasian patients. But an absence of detailed race information for this group precluded further analysis of this finding. Although adjustment for the degree of comorbidity was performed using a validated metric, undercoding of these comorbidities within the SPARCS database may have biased our results. Additional unmeasured differences in underlying health status may have existed among both race and insurance groups. Finally, tremendous interstate variation exists in rates of both uninsurance and public insurance. Our report on New York State represents only one constituent within the larger context of national health insurance-related disparities. In conclusion, in this analysis of adult patients with acute appendicitis, the likelihood of perforation did not vary by race. Uninsured, Medicaid, and Medicare pa-
Vol. 205, No. 3, September 2007
Pieracci et al
tients, by contrast, were all significantly more likely to have perforation, as compared with privately insured patients. The lack of association between race and perforation is an encouraging finding that contradicts a large body of literature documenting impaired access to health care among minority populations. By contrast, the increased likelihood of perforation observed in patients without private insurance provides evidence for continued insurance-related disparities in access to surgical care. Reduction of these disparities through insurance reform may provide one mechanism by which both patient outcomes and hospital costs are improved. Author Contributions Study conception and design: Pieracci, Eachempati Acquisition of data: Pieracci Analysis and interpretation of data: Pieracci, Eachempati, Barie, Callahan Drafting of manuscript: Pieracci Critical revision: Eachempati, Barie, Callahan Acknowledgment: We thank Heather Taffet Gold, PhD, and Huong T Do, MA, for their assistance with acquisition and analysis of US census data.
REFERENCES 1. Temple CL, Huchcroft SA, Temple WJ. The natural history of appendicitis in adults. A prospective study. Ann Surg 1995;221: 278–281. 2. Korner H, Sondenaa K, Soreide JA, et al. Incidence of acute nonperforated and perforated appendicitis: age-specific and sexspecific analysis. World J Surg 1997;21:313–317. 3. Addiss DG, Shaffer N, Fowler BS, Tauxe RV. The epidemiology of appendicitis and appendectomy in the United States. Am J Epidemiol 1990;132:910–925. 4. Guller U, Jain N, Curtis L, et al. Insurance status and race represent independent predictors of undergoing laparoscopic surgery for appendicitis: Secondary data analysis of 145,546 patients. J Am Coll Surg 2004;199:567–577. 5. Guller U, Jain N, Peterson ED, et al. Laparoscopic appendectomy in the elderly. Surgery 2004;135:479–488. 6. Nguyen NT, Zainabdi K, Mavandadi S, et al. Trends in utilization and outcomes of laparoscopic versus open appendectomy. Am J Surg 2004;188:813–820. 7. Pieracci FM, Eachempati SR, Christos PJ, et al. Explaining racial and insurance-related disparities in the surgical management of acute appendicitis. Am J Surg 2007;194:57–62. 8. Velanovich V, Satava R. Balancing the normal appendectomy rate with the perforated appendicitis rate. Am Surg 1992;58: 264–269.
Insurance Status Predicts Appendiceal Perforation
451
9. Luckmann R. Incidence and case fatality rates of acute appendicitis in California: A population-based study of the effects of age. Am J Epidemiol 1989;129:905–918. 10. Buckman TG, Zuidema GD. Reasons for delay of the diagnosis of acute appendicitis. Surg Gynecol Obstet 1984;158: 260–266. 11. Pittman-Waller VA, Myers JG, Stewart RM, et al. Appendicitis: why so complicated? Analysis of 5,755 consencutive appendectomies. Am Surg 2000;66:548–554. 12. Klein SR, Layden L, Wright JF, White RA. Appendicitis in the elderly: a diagnositic challenge. Postgrad Med 1988;83:247– 254. 13. Lieu TA, Newacheck PW, McManus MA. Race, ethnicity, and access to ambulatory care among US adolescents. Am J Publ Health 1993;83:960. 14. Groeneveld PW, Laufer SB, Garber AM. Technology diffusion, hospital variation, and racial disparities among elderly medicare beneficiaries. Med Care 2005;43:320–329. 15. Fiscella K, Franks P, Meldrum S, Barnett S. Racial disparity in surgical complications in New York State. Ann Surg 2005;242: 151–155. 16. Skinner J, Chandra A, Staiger D, et al. Mortality after acute myocardial infarction in hospitals that disproportionately treat Black patients. Circulation 2005;112:2634–2641. 17. Weissman JS, Stern R, Fielding SL, Epstein AM. Delayed access to health care: Risk factors, reasons, and consequences. Ann Intern Med 1991;114:325–331. 18. Hadley J. Sicker and poorer—the consequences of being uninsured: A review of the research on the relationship between health insurance, medical care use, work, income and education. Med Care Res Rev 2003;60:3S. 19. Guagliardo MF, Teach SJ, Huang ZJ, et al. Racial and ethnic disparities in pediatric appendicitis rupture rate. Acad Emerg Med 2003;10:1218–1227. 20. Smink DS, Fishman SJ, Kleinman K, Finklestein JA. Effects of race, insurance status, and hospital volume on perforated appendicitis in children. Pediatrics 2005;115:920–925. 21. Ponsky TA, Huang ZJ, Kittle K, et al. Hospital and patient-level characteristics and the risk of appendiceal rupture and negative appendectomy in children. JAMA 2004;292:1977–1982. 22. Braveman P, Schaaf VM, Egerter S, et al. Insurance-related differences in the risk of ruptured appendix. N Engl J Med 1994; 331:444–449. 23. Friedman E. Money isn’t everything: Nonfinancial barriers to access. JAMA 1994;271:1535. 24. Gornick ME, Eggers PW, Reilly TW, et al. Effects of race and income on mortality and use of services among Medicare beneficiaries. N Engl J Med 1996;335:791. 25. US Government. Revisions to the Standards for the Classification of Federal Data on Race and Ethnicity. Available at: http:// www.whitehouse.gov/omb/fedreg/ombdir15.html. Accessed: May 22, 2007. 26. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chron Dis 1987;40: 373–383. 27. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–619. 28. Kleinbaum DG, Kupper LL, Muller KE, Nizam A. Applied regression analysis and other multivariable methods. Pacific Grove, CA: Brooks/Cole Publishing; 1998:29.
452
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
29. Hosmer DW, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons; 2000:95. 30. Kleinbaum DG, Kupper LL, Muller KE, Nizam A. Applied regression analysis and other multivariable methods. Pacific Grove, CA: Brooks/Cole Publishing; 1998:237–248. 31. Millmann M, ed. Access to health care in America. Washington, DC: National Academy Press, Institute of Medicine; 1993. 32. Billings J, Teicholz N. Uninsured patients in the District of Columbia. Health Aff 1990;15:8. 33. Hannan EL, O’Donnell JF, Kilburn H Jr, et al. Investigation of the relationship between volume and mortality for surgical procedures performed in New York State hospitals. JAMA 1989; 262:503–510. 34. Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002; 346:1128–1137. 35. Smink DS, Finkelstein JA, Kleinman K, Fishman SJ. The effect of hospital volume of pediatric appendectomies on the misdiagnosis of appendicitis in children. Pediatrics 2004;113:18–23. 36. Yelin EH, Kramer JS, Epstein W. Is health care use equivalent across social groups? A diagnosis-based study. Am J Publ Health 1983;73:563–571. 37. Richards RJ, Reker DM. Racial differences in use of colonoscopy, sigmoidoscopy, and barium enema in Medicare beneficiaries. Dig Dis Sci 2002;47:2715–2719. 38. Bach PB, Cramer LD, Warren JL, Begg CB. Racial differences in the treatment of early-stage lung cancer. N Engl J Med 1999; 341:1198–1205.
J Am Coll Surg
39. Seldis SP, Fisher VJ, Tice D, et al. Racial differences in performance of invasive cardiac procedures in a Department of Veterans Affairs Medical Center. J Clin Epidemiol 1997;50:899–901. 40. Ayanian JZ, Cleary PD, Weissman JS, Epstein AM. The effect of patients’ preferences on racial differences in access to renal transplantation. N Engl J Med 1999;341:1661–1669. 41. Epstein AM, Weissman JS, Schnieder EC, et al. Race and gender disparities in rates of cardiac revascularization. Med Care 2003; 41:1240–1255. 42. Hannan EL, van Ryn M, Burke J, et al. Access to coronary artery bypass surgery by race/ethnicity and gender among patients who are approporaite for surgery. Med Care 1999;37:68–77. 43. Nixdorff N, Rimm A. Reduction in racial disparities between 1986 and 2001 for coronary artery bypass graft surgery in the Medicare population. Abstract presented at the 2005 meeting of the American College of Preventive Medicine, Reno, NV. 44. Nwomeh BC, Chisolm DJ, Caniano DA, Kelleher KJ. Racial and socioeconomic disparity in perforated appendicitis among children: where is the problem? Pediatrics 2006;117:870–875. 45. Siemiatycki J, Richardson L, Pless IB. Equity in medical care under national health insurance in Montreal. N Engl J Med 1980;303:10–15. 46. McKinlay JB. Some approaches and problems in the study of the use of services: an overview. J Health Soc Behav 1972;13:115– 152. 47. Rudge CJ, Fuggle SV. Geographic disparities in access to organ transplantation in the United Kingdom. Transplantation 2003; 15:1395–1398.
The model of acute appendicitis has been studied widely as a gauge of effective and equitable health care. Use of this model affords several methodologic advantages because biologic, behavioral, and socioeconomic variability are unlikely to influence the probability that appendicitis will develop.1-3 Nearly all cases of appendicitis eventually result in hospitalization,3 at which point, if surgery is indicated, two distinct options exist (laparoscopic versus open appendectomy). Both racial and insurance-related disparities have been well described in the management of appendicitis,4-6 and
current focus has shifted from documentation of these disparities to analysis of their causes. Using the New York State administrative database, we reported recently that differential hospital use is an important component of both racial and insurance-related disparities in the surgical management of appendicitis.7 Specifically, both Caucasian and privately insured patients with acute appendicitis are more likely to present to hospitals that perform laparoscopic appendectomy with regularity. This differential hospital use represents one mechanism by which access to surgical care is impaired for certain patient populations. Appendiceal perforation signifies a second access marker of interest in the management of appendicitis. Compared with patients without appendiceal perforation, those with a perforated appendix incur increased morbidity, length of stay, and health-care costs.8 Although certain biologic factors, such as increasing age
Competing Interests Declared: None. Received October 18, 2006; Revised March 28, 2007; Accepted April 9, 2007. From the Departments of Surgery (Pieracci, Eachempati, Barie, Callahan) and Public Health (Pieracci, Eachempati, Barie), Weill Medical College of Cornell University, New York, NY. Correspondence address: Fredric M Pieracci, MD, 411 East 69th St, #KB220, New York, NY 10021.
© 2007 by the American College of Surgeons Published by Elsevier Inc.
445
ISSN 1072-7515/07/$32.00 doi:10.1016/j.jamcollsurg.2007.04.010
446
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
and comorbidities, influence the likelihood of perforation,9 time to operation is the most important predictor of perforation once symptoms occur.10-12 Delays in surgery are accounted for by the patient’s decision to seek timely medical attention, access to both primary and hospital care without cost concerns, and the quality of care once presentation has occurred. Minority populations, the uninsured, and those with public insurance are less likely to have a usual source of primary care,13 and hospitals that treat these populations disproportionately demonstrate worse surgical outcomes.14-16 Compared with patients with insurance, the uninsured are more likely to delay seeking medical care for fear of financial repercussions.17,18 So it is plausible that these patient populations are at increased risk for perforation once appendicitis has occurred. Recent studies addressing this issue have focused on the pediatric population, reporting both racial19-21 and insurance-related disparities20,21 in the likelihood of perforation in children with appendicitis. Although Bravemen and associates22 documented similar disparities in adult patients with appendicitis, this study was conducted nearly 2 decades ago, and the exclusion of elderly patients (age ⬎ 65 years) precluded their examination. Indeed, significant racial disparities in the management of other surgical diseases have been well documented in the elderly Medicare patient population.14,23,24 The objectives of this study were to examine the relationships among race, insurance status, and likelihood of perforation in adult patients with acute appendicitis. METHODS Data were extracted from the New York State Statewide Planning and Research Cooperative System (SPARCS) database for the years 2003 and 2004 using SAS version 9.1 (SAS Institute). The SPARCS database is a legislatively mandated patient data system that contains information on all patients discharged from an acute-care, nonfederal hospital in New York State. A total of 15 diagnostic fields and 15 procedure fields (principal and other 1 to 14), which are based on the International Classification for Disease, 9th revision (ICD-9) classification system, are available for each patient discharge. For this analysis, we created a data set that included all adult (age ⱖ 18 years) patients with the diagnosis of acute appendicitis (540.0, 540.1, or 540.9) in any diagnosis field. The primary outcomes variable was appendiceal perforation, defined as the presence of a diagnos-
J Am Coll Surg
tic code for either perforated appendicitis without abscess (540.0), or perforated or nonperforated appendicitis with peritoneal abscess (540.1) in any diagnostic field. All other cases of appendicitis were considered nonperforated. Patients with diagnosis codes for both perforated and nonperforated appendicitis were included in the perforation group. The main predictor variables were race and insurance status. Race was grouped according to SPARCS database categories for race and ethnicity, which are based on current federal standards defined by the Office of Management and Budget.25 For this analysis, the SPARCS race and ethnicity fields were combined to form five race groups: Caucasian, Hispanic, African American, Asian, and other. Patients lacking race information were excluded from analysis. Insurance status was based on the SPARCS database field for primary reimbursement source. Five groups were created: private insurance, Medicare, Medicaid, uninsured, and other. Additional biologic characteristics abstracted were age (grouped as 18 to 40, 41 to 64, and ⱖ 65 years), gender, and comorbid conditions. Comorbidities were quantified using the Charlson comorbidity index26 adapted for use with administrative data by Deyo and associates.27 All diagnostic fields were queried for the presence of relevant comorbid conditions. Because the median comorbidity score was 0 (range 0 to 5), and a small proportion of subjects had at least 1 Charlson comorbid condition coded (n ⫽ 1,577, 5.3%), this variable was dichotomized as either zero or ⱖ 1. The following patient community factors, obtained from US census data for the year 2000 and based on subject ZIP code, were abstracted: the percentage of families living below the federal poverty line (⬍ 10%, 10% to 24.9%, ⬎ 25%); the percentage of adults who had achieved at least a high school education (dichotomized around the median value of 43.0%); and the percentage of rural residences (⬍ 10%, 10% to 50%, ⬎ 50%). Hospital characteristics known to influence the likelihood of perforation and so were abstracted were admission source (emergency department versus other), hospital teaching status, and volume of patients with acute appendicitis treated during the study period (divided into quartiles). All statistical analyses were computed using SAS version 9.1 (SAS Institute). Assessment of differences in nominal categorical variables was performed using the chi-square test. All p values were two-sided, with sta-
Vol. 205, No. 3, September 2007
Pieracci et al
tistical significance evaluated at the 0.05 ␣ level, with the following exception. Because multiple comparisons were conducted when comparing individual race (Caucasian versus African American, Hispanic, Asian, and other) and insurance (private insurance versus Medicare, Medicaid, uninsured, and other) groups, a Bonferroni-corrected ␣ levelwasused.TheBonferronicorrected ␣ level was calculated by dividing the standard ␣ error level of 0.05 by the number of tests performed (0.05 / 4 ⫽ 0.0125 in the case of both race and insurance status).28 Multiple logistic regression, which included predictor variables significant at the 0.25 level by univariate analysis,29 was used to obtain adjusted odds of perforation. All variables were entered into the model as outlined previously using k – 1 indicator variables, where k is equal to the number of levels within the variable. To avoid overcontrolling, which may result from the addition of multiple, correlated socioeconomic factors to the regression model,30 a model that controlled for age only (the strongest predictor of perforation by univariate analysis) was tested first. Next, a model that contained each of the previously mentioned patient, community, and hospital factors was tested to obtain fully adjusted odds of perforation. Variables were added to the model using a forward selection method according to descending magnitude of unadjusted odds ratios (ORs). RESULTS A total of 31,245 patients had appendicitis, and race information was available for 29,637 (94.9%) of them. Of these 29,637 patients, 7,969 (26.9%) had perforation. Additional sample characteristics, and the results of univariate analysis, are shown in Table 1. The likelihood of perforation varied significantly by all variables tested. The prevalence of perforation increased markedly with age; more than one-half of patients aged 65 years or older had perforation. In addition, subjects with a comorbidity score ⱖ 1 were significantly more likely to have perforated as compared with those with a comorbidity score of zero (44.6% versus 25.9%, respectively, p ⬍ 0.0001). Caucasian patients demonstrated the highest prevalence of perforation by univariate analysis (28.6%), followed by African-American (27.0%), Hispanic (23.8%), and Asian patients (23.6%). Analysis of insurance status revealed that Medicare patients had the highest prevalence of perforation (52.0%), followed by privately insured (24.3%) and Medicaid
Insurance Status Predicts Appendiceal Perforation
447
(23.9%) patients. Patients without insurance had the lowest prevalence of perforation (22.5%). Finally, patients who resided in communities with the lowest levels of poverty were most likely to suffer perforation (28.0%). The higher observed prevalence of perforation in patients who were Caucasian, privately insured, and resided in wealthy communities was suspected to result from confounding by age, so it was examined further by multivariable analysis. Finally, male gender, lower community levels of education, an increased proportion of rural residences, an admission source other than the emergency department, and treatment at both a nonteaching and lower-volume hospital were associated with an increased likelihood of perforation. Confounding by age explained all of the variability in likelihood of perforation by race. As shown in Table 2, there were no longer significant differences in the likelihood of perforation when comparing Caucasian with African-American, Hispanic, and Asian patients after adjusting for age alone. In addition, these associations remained nonsignificant after adjustment for the remaining covariates. By contrast, significant differences in likelihood of perforation between insurance groups remained after adjusting for age and other covariates used in the regression model. Specifically, as compared with patients with private insurance, uninsured (OR ⫽ 1.18, 95% CI [1.07, 1.30], p ⫽ 0.0005), Medicaid (OR ⫽ 1.22, 95% CI [1.12, 1.33], p ⬍ 0.0001), and Medicare patients (OR 1.14, 95% CI [1.03, 1.25], p ⫽ 0.01) were all more likely to have perforation. We also performed a stratified analysis of Medicare patients aged 65 years and older (n ⫽ 2,432) to eliminate possible confounding of the relationship between race and perforation by insurance status (Table 3). By univariate analysis, minority patients were not more likely to have perforation than Caucasian patients. Controlling for the previously mentioned patient, community, and hospital factors did not affect this relationship. DISCUSSION In this study of a statewide administrative database, we observed that race was not associated with likelihood of perforation in adult patients with appendicitis, in contrast to previous reports. But insurance status was independently associated with perforation; uninsured, Medicaid, and Medicare patients were all significantly more likely to have perforation compared with patients with private insurance. These relationships persisted after
448
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
J Am Coll Surg
Table 1. Sample Demographics and Univariate Analysis Perforated†
Total* Variable
n
n
%
29,637
n
%
p Value‡
7,969
26.9
—
28.6
Main predictor variables ⬍ 0.0001
Race Caucasian
18,598
62.8
5,314
Hispanic
4,465
15.1
1,063
23.8
African American
2,684
9.1
725
27.0
Asian
1,193
4.0
281
23.6
Other
2,697
9.1
586
21.7
17,149
57.9
4,166
24.3
Medicare
3,128
10.6
1,626
52.0
Medicaid
5,126
17.3
1,224
23.9
Uninsured
3,691
12.5
831
22.5
543
1.8
122
22.5
18–40
16,348
55.2
2,831
17.3
41–64
10,430
35.2
3,567
34.2
ⱖ 65
2,859
9.7
1,571
54.9
Male
16,125
54.4
4,451
27.6
Female
13,512
45.6
3,518
26.0
Zero
28,060
94.7
7,266
25.9
ⱖ1
1,577
5.3
703
44.6
⬍ 10
14,095
47.6
3,956
28.1
10–25
10,999
37.1
2,854
26.0
⬎ 25
4,543
15.3
1,159
25.5
⬍ 10
23,234
78.4
6,058
26.1
10–50
3,679
12.4
1,072
29.1
⬎ 50
2,724
9.2
839
30.8
Below median, ⬍ 43
14,275
48.2
3,938
27.6
Above median, ⱖ 43
15,362
51.8
4,031
26.2
26,763
90.3
7,012
26.2
2,874
9.7
957
33.3
21,447
72.4
5,650
2.3
8,190
27.6
2,319
28.3
937
3.2
291
31.1
Low
4,143
14.0
1,229
29.7
High
6,506
22.0
1,711
26.3
18,051
60.9
4,738
26.2
⬍ 0.0001
Insurance Private
Other Biologic factors
⬍ 0.0001
Age, y
Gender
0.002
⬍ 0.0001
Comorbidity score
Community factors ⬍ 0.0001
Below federal poverty line, %
⬍ 0.0001
Rural residences, %
High school educated, %
0.01
Hospital factors ⬍ 0.0001
Admission source Emergency department Other Teaching status Teaching Nonteaching
0.001
Hospital volume Lowest
Highest
0.004
*Refers to the number of patients in that group divided by the total sample size. † Refers to the number of patients with a perforation over the total number of patients in that group. ‡ Refers to results from chi-square analysis for significant differences in likelihood of perforation between groups.
Vol. 205, No. 3, September 2007
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
449
Table 2. Multivariable Analysis of Predictors of Appendiceal Perforation Variable
Race Caucasian African American Hispanic Asian Other Insurance Private insurance Uninsured Medicaid Medicare Other
Unadjusted odds ratio (95% CI)
Age-adjusted odds ratio (95% CI)
Fully adjusted† odds ratio (95% CI)
1.00 0.93 (0.85, 1.01) 0.78 (0.72, 0.84)* 0.77 (0.67, 0.88)* 0.69 (0.63, 0.77)*
1.00 1.08 (0.98, 1.19) 1.06 (0.98, 1.15) 0.86 (0.75, 0.99) 0.91 (0.82, 1.07)
1.00 1.03 (0.93, 1.15) 0.99 (0.90, 1.08) 0.85 (0.73, 1.00) 0.84 (0.75, 0.94)*
1.00 0.91 (0.83, 0.99) 0.98 (0.91, 1.05) 3.37 (3.12, 3.64)* 0.90 (0.74, 1.11)
1.00 1.16 (1.06, 1.27)* 1.18 (1.10, 1.28)* 1.16 (1.05, 1.27)* 1.05 (0.85, 1.29)
1.00 1.18 (1.07, 1.30)* 1.22 (1.12, 1.33)* 1.14 (1.03, 1.25)* 1.01 (0.82, 1.24)
*p ⬍ 0.0125 (Bonferroni-corrected ␣ error level ⫽ 0.05 / 4). † Controlling for age, gender, comorbidity score, insurance, race, community poverty level, community education level, community level of urbanization, hospital teaching status, and hospital volume.
controlling for patient, community, and hospital factors, each of which was significantly associated with likelihood of perforation by univariate analysis. To our knowledge, only one other study has addressed these relationships in an adult population. Braveman and colleagues22 used State of California data from 1984 to 1989 to report that, compared with patients with private insurance, patients with Medicaid (OR 1.49, 95% CI [1.49, 1.59]) and without insurance (OR 1.46, 95% CI [1.39, 1.54]) were significantly more likely to have perforation. Age, gender, race, comorbidities, community poverty level, hospital teaching status, and hospital volume were controlled for, and patients older than 65 years were excluded secondary to near-universal hospital reimbursement through Medicare (in our sample, by contrast, nearly 15% [n ⫽ 427] of elderly patients used private insurance as their principal method of hospital payment). Our findings are consistent with those reported by Braveman and coauthors, although the effect size of our study is somewhat attenuated. These data, in addition to those in similar reports in the pedi-
atric literature,20,21 provide evidence for an independent effect of insurance status on likelihood of perforation. This effect may be mediated through several mechanisms. Because both uninsured and publicly insured patients are less likely to have a usual source of care,31,32 difficulty in consulting a provider may delay hospitalization. Longer waiting times once care has been sought may also contribute to delayed treatment. Second, uninsured patients may delay seeking care for fear of financial repercussions. In one large multicenter study, uninsured patients were 9.5 times as likely to report delaying care because of cost concerns as compared with privately insured patients.17 Delays in seeking care were 60% more frequent for emergent versus elective admissions. Several alternative explanations for our findings must also be considered. Socioeconomic variables, such as race, income, and community resources, frequently confound the relationship between insurance status and adverse health outcomes. But previous reports have not found an association between individual or community income and the likelihood of perforation in patients
Table 3. Subgroup Analysis of Likelihood of Appendiceal Perforation among Medicare Patients Aged 65 Years or Older Race
Caucasian African American Hispanic Asian Other
Unadjusted odds ratio (95% CI)
p Value*
Fully adjusted† odds ratio (95% CI)
p Value*
1.00 1.12 (0.73, 1.71) 0.83 (0.52, 1.33) 0.75 (0.39, 1.43) 1.10 (0.64, 1.91)
— 0.52 0.31 0.26 0.66
1.00 1.13 (0.78, 1.63) 0.85 (0.58, 1.26) 0.81 (0.48, 1.37) 1.15 (0.73, 1.80)
— 0.52 0.43 0.43 0.54
*p ⬍ 0.0125 (Bonferroni-corrected ␣ error level ⫽ 0.05 / 4). † Controlling for age, gender, comorbidity score, community poverty level, community education level, community level of urbanization, hospital teaching status, and hospital volume.
450
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
with appendicitis.19,20 In addition, both race and community poverty level were controlled for in this analysis. Additional factors that may delay treatment33-35 were also addressed in our analysis. Specifically, we controlled for community level of education, degree of community urbanization, admission source, hospital teaching status, and hospital volume. But although all of these variables were significantly associated with likelihood of perforation by univariate analysis, their addition to the regression model did not affect the significance of the relationship between insurance status and perforation. The lack of an association between race and perforation was an unexpected finding. Minority patients have been documented to use fewer health-care services,36-38 have altered preferences about the utility of surgical interventions,39,40 receive recommendation for necessary procedures less often,41,42 and receive care at hospitals that incur worse outcomes.15,16 Each of the previously mentioned disparities would suggest that this patient group is at increased risk for appendiceal perforation. In addition, reports on both adult22 and pediatric19-21 populations have observed that minority patients are significantly more likely to have perforation compared with Caucasian patients. By contrast, our results did not find a relationship between race and likelihood of perforation in both our overall sample analysis and in a subgroup analysis of elderly Medicare patients. Because differences in odds of perforation among Caucasian, African-American, and Hispanic patients were eliminated after controlling for age only, it is unlikely that confounding by additional socioeconomic factors explained the original disparities. One possible explanation is that an improved awareness of racial and ethnic disparities within the medical community has led to a reduction in these disparities. Abatement of racial disparities over time was noted recently in Medicare patients undergoing coronary artery bypass grafting.43 In addition, a recent single-institution study of pediatric patients with acute appendicitis observed neither racial nor insurance-related disparities in the likelihood of perforation.44 Our decision to include elderly patients was based on the following two reasons: First, the predominant use of Medicare for hospital reimbursement facilitated subgroup analysis of the effect of race on perforation in this population. Second, we were able to distinguish between hospital reimbursement through Medicare and the nearly 15% of cases in which private insurance was
J Am Coll Surg
the principal method of hospital payment. Elderly patients with private insurance may also be less inclined to delay treatment because of potential financial repercussions. In this case, concerns may involve the possibility of extended hospitals stays, costly outpatient prescription drugs, and longterm care. This claim is supported by the 14% adjusted increased odds of perforation in patients with Medicare as compared with patients with private insurance. Health policy implications based on our results must still be considered with caution. Although insurance status appears to be associated with perforation in patients with acute appendicitis, it is not clear that elimination of insurance variability will reduce or eliminate delays in treatment. Indeed, substantial disparities in access to care have been observed in countries with compulsory health insurance.45-47 So it is possible that patients with and without private insurance differ systematically with respect to unmeasured variables. But our results indicate that insurance status has an independent effect on the likelihood of appendiceal perforation among the variables tested. Patient characteristics analyzed in this study were limited to the fields available within the SPARCS database. Accordingly, covariates such as level of education and individual income were approximated using community-level data based on patient ZIP codes. Because race information was available in 94.9% of patients with appendicitis, exclusion bias was minimal. But nearly 10% of our sample was composed of patients whose race was coded as “other.” Interestingly, by multivariable analysis, this patient group remained significantly less likely to perforate compared with Caucasian patients. But an absence of detailed race information for this group precluded further analysis of this finding. Although adjustment for the degree of comorbidity was performed using a validated metric, undercoding of these comorbidities within the SPARCS database may have biased our results. Additional unmeasured differences in underlying health status may have existed among both race and insurance groups. Finally, tremendous interstate variation exists in rates of both uninsurance and public insurance. Our report on New York State represents only one constituent within the larger context of national health insurance-related disparities. In conclusion, in this analysis of adult patients with acute appendicitis, the likelihood of perforation did not vary by race. Uninsured, Medicaid, and Medicare pa-
Vol. 205, No. 3, September 2007
Pieracci et al
tients, by contrast, were all significantly more likely to have perforation, as compared with privately insured patients. The lack of association between race and perforation is an encouraging finding that contradicts a large body of literature documenting impaired access to health care among minority populations. By contrast, the increased likelihood of perforation observed in patients without private insurance provides evidence for continued insurance-related disparities in access to surgical care. Reduction of these disparities through insurance reform may provide one mechanism by which both patient outcomes and hospital costs are improved. Author Contributions Study conception and design: Pieracci, Eachempati Acquisition of data: Pieracci Analysis and interpretation of data: Pieracci, Eachempati, Barie, Callahan Drafting of manuscript: Pieracci Critical revision: Eachempati, Barie, Callahan Acknowledgment: We thank Heather Taffet Gold, PhD, and Huong T Do, MA, for their assistance with acquisition and analysis of US census data.
REFERENCES 1. Temple CL, Huchcroft SA, Temple WJ. The natural history of appendicitis in adults. A prospective study. Ann Surg 1995;221: 278–281. 2. Korner H, Sondenaa K, Soreide JA, et al. Incidence of acute nonperforated and perforated appendicitis: age-specific and sexspecific analysis. World J Surg 1997;21:313–317. 3. Addiss DG, Shaffer N, Fowler BS, Tauxe RV. The epidemiology of appendicitis and appendectomy in the United States. Am J Epidemiol 1990;132:910–925. 4. Guller U, Jain N, Curtis L, et al. Insurance status and race represent independent predictors of undergoing laparoscopic surgery for appendicitis: Secondary data analysis of 145,546 patients. J Am Coll Surg 2004;199:567–577. 5. Guller U, Jain N, Peterson ED, et al. Laparoscopic appendectomy in the elderly. Surgery 2004;135:479–488. 6. Nguyen NT, Zainabdi K, Mavandadi S, et al. Trends in utilization and outcomes of laparoscopic versus open appendectomy. Am J Surg 2004;188:813–820. 7. Pieracci FM, Eachempati SR, Christos PJ, et al. Explaining racial and insurance-related disparities in the surgical management of acute appendicitis. Am J Surg 2007;194:57–62. 8. Velanovich V, Satava R. Balancing the normal appendectomy rate with the perforated appendicitis rate. Am Surg 1992;58: 264–269.
Insurance Status Predicts Appendiceal Perforation
451
9. Luckmann R. Incidence and case fatality rates of acute appendicitis in California: A population-based study of the effects of age. Am J Epidemiol 1989;129:905–918. 10. Buckman TG, Zuidema GD. Reasons for delay of the diagnosis of acute appendicitis. Surg Gynecol Obstet 1984;158: 260–266. 11. Pittman-Waller VA, Myers JG, Stewart RM, et al. Appendicitis: why so complicated? Analysis of 5,755 consencutive appendectomies. Am Surg 2000;66:548–554. 12. Klein SR, Layden L, Wright JF, White RA. Appendicitis in the elderly: a diagnositic challenge. Postgrad Med 1988;83:247– 254. 13. Lieu TA, Newacheck PW, McManus MA. Race, ethnicity, and access to ambulatory care among US adolescents. Am J Publ Health 1993;83:960. 14. Groeneveld PW, Laufer SB, Garber AM. Technology diffusion, hospital variation, and racial disparities among elderly medicare beneficiaries. Med Care 2005;43:320–329. 15. Fiscella K, Franks P, Meldrum S, Barnett S. Racial disparity in surgical complications in New York State. Ann Surg 2005;242: 151–155. 16. Skinner J, Chandra A, Staiger D, et al. Mortality after acute myocardial infarction in hospitals that disproportionately treat Black patients. Circulation 2005;112:2634–2641. 17. Weissman JS, Stern R, Fielding SL, Epstein AM. Delayed access to health care: Risk factors, reasons, and consequences. Ann Intern Med 1991;114:325–331. 18. Hadley J. Sicker and poorer—the consequences of being uninsured: A review of the research on the relationship between health insurance, medical care use, work, income and education. Med Care Res Rev 2003;60:3S. 19. Guagliardo MF, Teach SJ, Huang ZJ, et al. Racial and ethnic disparities in pediatric appendicitis rupture rate. Acad Emerg Med 2003;10:1218–1227. 20. Smink DS, Fishman SJ, Kleinman K, Finklestein JA. Effects of race, insurance status, and hospital volume on perforated appendicitis in children. Pediatrics 2005;115:920–925. 21. Ponsky TA, Huang ZJ, Kittle K, et al. Hospital and patient-level characteristics and the risk of appendiceal rupture and negative appendectomy in children. JAMA 2004;292:1977–1982. 22. Braveman P, Schaaf VM, Egerter S, et al. Insurance-related differences in the risk of ruptured appendix. N Engl J Med 1994; 331:444–449. 23. Friedman E. Money isn’t everything: Nonfinancial barriers to access. JAMA 1994;271:1535. 24. Gornick ME, Eggers PW, Reilly TW, et al. Effects of race and income on mortality and use of services among Medicare beneficiaries. N Engl J Med 1996;335:791. 25. US Government. Revisions to the Standards for the Classification of Federal Data on Race and Ethnicity. Available at: http:// www.whitehouse.gov/omb/fedreg/ombdir15.html. Accessed: May 22, 2007. 26. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chron Dis 1987;40: 373–383. 27. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–619. 28. Kleinbaum DG, Kupper LL, Muller KE, Nizam A. Applied regression analysis and other multivariable methods. Pacific Grove, CA: Brooks/Cole Publishing; 1998:29.
452
Pieracci et al
Insurance Status Predicts Appendiceal Perforation
29. Hosmer DW, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons; 2000:95. 30. Kleinbaum DG, Kupper LL, Muller KE, Nizam A. Applied regression analysis and other multivariable methods. Pacific Grove, CA: Brooks/Cole Publishing; 1998:237–248. 31. Millmann M, ed. Access to health care in America. Washington, DC: National Academy Press, Institute of Medicine; 1993. 32. Billings J, Teicholz N. Uninsured patients in the District of Columbia. Health Aff 1990;15:8. 33. Hannan EL, O’Donnell JF, Kilburn H Jr, et al. Investigation of the relationship between volume and mortality for surgical procedures performed in New York State hospitals. JAMA 1989; 262:503–510. 34. Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002; 346:1128–1137. 35. Smink DS, Finkelstein JA, Kleinman K, Fishman SJ. The effect of hospital volume of pediatric appendectomies on the misdiagnosis of appendicitis in children. Pediatrics 2004;113:18–23. 36. Yelin EH, Kramer JS, Epstein W. Is health care use equivalent across social groups? A diagnosis-based study. Am J Publ Health 1983;73:563–571. 37. Richards RJ, Reker DM. Racial differences in use of colonoscopy, sigmoidoscopy, and barium enema in Medicare beneficiaries. Dig Dis Sci 2002;47:2715–2719. 38. Bach PB, Cramer LD, Warren JL, Begg CB. Racial differences in the treatment of early-stage lung cancer. N Engl J Med 1999; 341:1198–1205.
J Am Coll Surg
39. Seldis SP, Fisher VJ, Tice D, et al. Racial differences in performance of invasive cardiac procedures in a Department of Veterans Affairs Medical Center. J Clin Epidemiol 1997;50:899–901. 40. Ayanian JZ, Cleary PD, Weissman JS, Epstein AM. The effect of patients’ preferences on racial differences in access to renal transplantation. N Engl J Med 1999;341:1661–1669. 41. Epstein AM, Weissman JS, Schnieder EC, et al. Race and gender disparities in rates of cardiac revascularization. Med Care 2003; 41:1240–1255. 42. Hannan EL, van Ryn M, Burke J, et al. Access to coronary artery bypass surgery by race/ethnicity and gender among patients who are approporaite for surgery. Med Care 1999;37:68–77. 43. Nixdorff N, Rimm A. Reduction in racial disparities between 1986 and 2001 for coronary artery bypass graft surgery in the Medicare population. Abstract presented at the 2005 meeting of the American College of Preventive Medicine, Reno, NV. 44. Nwomeh BC, Chisolm DJ, Caniano DA, Kelleher KJ. Racial and socioeconomic disparity in perforated appendicitis among children: where is the problem? Pediatrics 2006;117:870–875. 45. Siemiatycki J, Richardson L, Pless IB. Equity in medical care under national health insurance in Montreal. N Engl J Med 1980;303:10–15. 46. McKinlay JB. Some approaches and problems in the study of the use of services: an overview. J Health Soc Behav 1972;13:115– 152. 47. Rudge CJ, Fuggle SV. Geographic disparities in access to organ transplantation in the United Kingdom. Transplantation 2003; 15:1395–1398.