- Email: [email protected]
Race and socioeconomic disparities in national stoma reversal rates
The American Journal of Surgery (2016) 211, 710-715
Society of Black Academic Surgeons
Race and socioeconomic disparities in national stoma reversal rates Syed Nabeel Zafar, M.D., M.P.H.a,*, Navin R. Changoor, M.D.a,b, Kibileri Williams, M.D.a, Rafael D. Acosta, M.D.a, Wendy R. Greene, M.D.a, Terrence M. Fullum, M.D.a, Adil H. Haider, M.D., M.P.H.b, Edward E. Cornwell, III, M.D.a, Daniel D. Tran, M.D.a a
Department of Surgery, Howard University Hospital, 2041 Georgia Ave NW, Washington, DC, 20060, USA; bCenter for Surgery and Public Health, Brigham and Women’s Hospital, Boston, MA, USA
KEYWORDS: Ostomy; Stoma reversal; Health care disparities; Access to health care; Surgery
Abstract BACKGROUND: Many temporary stomas are never reversed leading to significantly worse quality of life. Recent evidence suggests a lower rate of reversal among minority patients. Our study aimed to elucidate disparities in national stoma closure rates by race, medical insurance status, and household income. METHODS: Five years of data from the Nationwide Inpatient Sample (2008 to 2012) was used to identify the annual rates of stoma formation and annual rates of stoma closure. Stomas labeled as ‘‘permanent’’ or those created secondary to colorectal cancers were excluded. Temporary stoma closure rates were calculated, and differences were tested with the chi-square test. Separate analyses were performed by race/ethnicity, insurance status, and household income. Nationally representative estimates were calculated using discharge-level weights. RESULTS: The 5-year average annual rate of temporary stoma creation was 76,551 per year (46% colostomies and 54% ileostomies). The annual rate of stoma reversal was 50,155 per year that equated to an annual reversal rate of 65.5%. Reversal rates were higher among white patients compared with black patients (67% vs 56%, P , .001) and among privately insured patients compared with uninsured patients (88% vs 63%, P , .001). Reversal rates increased as the household income increased from 61% in the lowest income quartile to 72% in the highest quartile (P , .001). CONCLUSIONS: Stark disparities exist in national rates of stoma closure. Stoma closure is associated with race, insurance, and income status. This study highlights the lack of access to surgical health care among patients of minority race and low-income status. Ó 2016 Elsevier Inc. All rights reserved.
Oral presentation at the 25th Annual Scientific Assembly of the Society of Black Academic Surgeons, April 9–11, 2015, Chapel Hill, NC. * Corresponding author. Tel.: 11-202-865-1446; fax: 11-202-8656728. E-mail address: [email protected] Manuscript received June 2, 2015; revised manuscript October 1, 2015 0002-9610/$ - see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2015.11.020
The fashioning of an ileostomy or colostomy is a relatively common surgical procedure performed for a myriad of indications spanning trauma, diverticulitis, malignancy, inflammatory bowel disease, and ischemia. These stomas can be permanent or temporary, with the latter being reversed within a very variable time frame.
S.N. Zafar et al.
Disparities in stoma reversal
An estimated 450,000 patients in the United States have an ostomy.1 An unknown proportion of these await reversal. From clinical experience and previous reports, not all stomas are reversed.2,3 There is a paucity of data regarding the national rates of stoma formation and reversal. In a retrospective analysis of 3,899 trauma patients in the state of California who underwent stoma creation, 28% were not reversed within 5 years.2 In a similar study of 1,176 patients in California with stomas created for diverticulitis, an estimated 35% of patients did not have their stomas reversed.4 Living with an ostomy has been associated with a lower quality of life, worse illness perceptions, and higher health care consumption compared with a similar patient population without an ostomy.5,6 Recent studies have shown insurance status, race, age, and income level to be associated with stoma reversal.2,3,7,8 African Americans have been found to be 4 times less likely to undergo reversal than Caucasians.3 Lack of adequate health insurance has been identified as a predictor of greater disease severity, suboptimal surgical treatment, and mortality.9 Older patients and those of lower socioeconomic status were also less likely to have their stomas reversed in a timely fashion.7 Most of these studies are, however, single-institution retrospective reviews. It is, therefore, difficult to extrapolate these results at a national level. The aim of this study was to assess race, insurance, and income-related disparities in the rates of stoma reversal on a national scale in the United States.
Methods We used 5 years of data from the Nationwide Inpatient Sample (NIS; admission years 2008 to 2012). The NIS is the largest all-payer inpatient database available.10 It is a 20% probability sample of all community hospitals and is derived from state inpatient databases. It represents 95% of the US population and includes demographic, clinical, and outcomes information. Discharge-level weights provided for each patient allow for the calculation of nationally representative estimates. The NIS does not have longitudinal data on patients beyond hospital discharge; therefore, follow-up information, such as whether an individual patient’s stoma was reversed after a few months, is unavailable. However, because the NIS is a representative sample, it does provide us with national estimates. In an ideal situation where all temporary stomas are reversed, the national 5-year average number of stomas created should equal the national 5-year average number of stomas reversed. Using discharge-level weights and survey estimation commands, we are able to determine these numbers. Therefore, we are able to calculate national stoma reversal rates by dividing the national 5-year average number of stomas reversed by the national 5-year average number of stomas created. From the NIS 2008 to 2012, we created 2 separate data sets: one with all patients who had a temporary stoma
711 created and the second with all patients who had a stoma reversal. For both data sets, we calculate 5-year national average rates. For the first data set, we used International Classification of Diseases, Ninth Revision (ICD 9), procedure codes and the NIS-provided Clinical Classification Software to select all patients who had a procedure code for ‘‘colostomy’’ (ICD 9 codes 46.10, 46.11, 46.12, 46.13, and 46.14) and ‘‘ileostomy’’ (ICD 9 codes 46.20, 46.21, 46.22, 46.23, 46.24, 46.31, 46.32, and 46.39) creation. We then excluded patients if they satisfied either one of the following exclusion criteria: (a) procedure labeled as ‘‘permanent’’ stoma, (b) stoma created because of a diagnosis of malignancy of the colon, rectum, or anus, (c) the patient died during the same hospitalization, or (d) the stoma was reversed during the same admission. In particular, malignancy of colon, rectum, and anus were excluded because of the lack of staging information that made it impossible to determine which patients had permanent ostomies because of advanced-stage cancers. We, therefore, arrived at a cohort of patients with ‘‘temporary’’ stomas. From the database, it is impossible to determine with 100% accuracy whether a stoma created was intended to be temporary or permanent. The earlier mentioned exclusion criteria offer us the closest and most reasonable estimate of ‘‘temporary’’ stomas; however, it is not an entirely accurate estimate. For example, many stomas for cancers of the colon, rectum, and anus may have been intended to be temporary, which we exclude, and conversely many trauma stomas, eg, in paraplegic patients, may have been intended to be permanent, which we include. These inaccuracies in identification are, however, thought to be unrelated to demographic and socioeconomic categories. Because the objective of our analysis is to study disparities in stoma reversal rates, it is reasonable to accept this marginal inaccuracy in identification of ‘‘temporary’’ stomas as it is unlikely to affect the disparity. Other demographic and clinical variables included in the data set were age (in years), gender, race, number of chronic conditions, admission type (elective vs emergent), insurance (Medicare, Medicaid, private insurance, uninsured or self pay, and other insurance), income quartile of patients, Zip code, primary diagnosis, length of hospital stay, and mortality. Using NIS-provided discharge-level weights, we were able to estimate the 5-year average annual rate of ‘‘temporary’’ stoma creation. We calculated this rate for different demographic categories. For the second data set, we used ICD 9 procedure codes to identify patients who underwent an ‘‘ileostomy closure’’ (ICD 9 code 46.51), ‘‘colostomy closure’’ (46.52), or ‘‘unspecified stoma closure’’ (46.50). A similar set of demographic and clinical variables were abstracted and tabulated. Using discharge-level weights, we calculated 5-year average annual rates of stoma reversal for different demographic categories as mentioned earlier. We then calculated stoma reversal rates by dividing the annual average number of stomas reversed by the annual average number of stomas created. Ideally, these 2 numbers should
712
The American Journal of Surgery, Vol 211, No 4, April 2016
be equal to provide a reversal rate of 100%. Stoma reversal rates were then calculated for different demographic and socioeconomic categories, and differences were tested for statistical significance using the chi-square test. All analyses were performed on STATA, version 12 (StataCorp, College station, TX).
Results Over the 5-year period from 2008 to 2012, a total of 566,625 stomas were created. Of these, 59,571 (10.5%) were labeled as permanent stomas, 83,298 (14.7%) were created because of malignancy of the colon, rectum, or anus, 18,949 (3.3%) had their ostomy reversed during the same admission, and 54,815 (9.7%) died during the same hospitalization. After these patients were excluded, 382,754 ‘‘temporary’’ stomas remained, of which 175,477 (45.8%)
were colostomies and 211,085 (55.1%) were ileostomies. The 5-year average annual rate of stoma creation was 76,551 stomas per year. The demographic and clinical characteristics of these patients are presented in Table 1. The average age was 59 years, 66% of patients were of white race, whereas 9.7% were black. Gender distribution was roughly equal with 48% of patients being women. Only 3.9% of patients were uninsured. On average, each patient had 5 comorbidities, 70% were admitted nonelectively, and the primary diagnosis was categorized as inflammation or infection in 45% of patients. During the same 5-year period, 250,776 stomas were reversed. Of these, 44.9% (112,483) were ileostomy closures, whereas 55.4% (138,873) were colostomy closures. The 5-year annual average rate of stoma reversal was 50,155 per year. The characteristics of patients undergoing stoma reversal are presented in Table 2. The mean age was
Table 1 Characteristics of patients receiving temporary stomas: NIS 2008 to 2012 (weighted to provide national estimates; n 5 382,754) Variable
Categories
Frequency/mean
%/SD
Age (y) Race
Mean White Black Hispanic Others Missing Female Mean 2008 2009 2010 2011 2012 Medicare Medicaid Private Uninsured Other Missing Yes Lowest 2 3 Highest Missing Inflammation or infection Trauma Complications of surgery or medical device Other cancer Obstruction Others Ileostomy Colostomy Mean
58.7 y 252,265 37,262 27,914 17,783 47,529 186,524 5.2 70,565 74,932 78,652 84,090 74,515 179,597 44,905 129,212 15,042 11,425 832 268,168 96,971 98,532 94,688 84,581 7,983 173,156 7,604 39,554 58,151 32,344 43,004 211,085 175,477 17.8 d
620.0 y 65.9% 9.7% 7.3% 4.6% 12.4% 48.7% 63.1 18.4% 19.6% 20.6% 22.0% 19.5% 47.0% 11.8% 33.8% 3.9% 3.0% .2% 70.1% 25.3% 25.7% 24.7% 22.1 2.1% 45.2% 2.0% 10.3% 15.2% 8.5% 15.3% 55.1% 45.8% 619.1 d
Gender Number of chronic conditions Year
Insurance
Nonelective admission Income quartile of patient ZIP
Primary diagnosis
Type of stoma Length of hospital stay SD 5 standard deviation.
S.N. Zafar et al.
Disparities in stoma reversal
713
Table 2 Characteristics of patients undergoing stoma reversal, NIS 2008 to 2012 (weighted to provide national estimates; n 5 250,776)
Table 3 Annual stoma reversal rates by patient characteristics (5-year average annual rates)
Variable
Categories Frequency/mean %/SD
Stoma
Created
Reversed
Reversal rate %
Age (y) Race
Mean White Black Hispanic Others Missing Female Mean
Age 651 Age18–65 Age ,18 Female Male White Black Medicare Medicaid Private Uninsured Income lowest Income level 2 Income level 3 Income highest Ileostomy Colostomy
33,449 39,499 3,603 37,306 39,219 50,453 7,452 36,992 8,981 25,843 3,008 19,394 19,706 18,938 16,916 42,217 35,095
16,153 31,262 2,741 23,625 26,411 33,865 4,190 17,128 6,016 22,807 1,887 11,866 12,640 12,466 12,143 22,497 27,775
48.3 79.1 76.1 63.3 67.3 67.1 56.2 46.3 67.0 88.3 62.7 61.2 64.1 65.8 71.8 53.1 79.1
54.1 y 169,323 20,949 19,564 11,163 29,777 118,125 3.9
619.4 y 67.5% 8.4% 7.8% 4.5% 11.9% 47.1% 62.5
2008 50,614 2009 50,195 2010 49,706 2011 51,531 2012 48,730 Insurance Medicare 85,639 Medicaid 30,082 Private 114,035 Uninsured 9,436 Other 9,536 Missing 483 Nonelective admission Yes 31,089 Income quartile of Lowest 59,330 patient ZIP 2 63,201 3 62,329 Highest 60,715 Missing 5,201 Type of stoma Ileostomy 112,483 Colostomy 138,873 Length of hospital stay Mean 8.0 Mortality Yes 2,002
20.2% 20.0% 19.8% 20.6% 19.4% 34.2% 12.0% 45.5% 3.8% 3.8% .2% 12.4% 23.7%
Gender Number of chronic conditions Year
25.2% 24.9% 24.2% 2.1% 44.8% 55.3% 612.2 .8%
SD 5 standard deviation.
54 years and 68% of patients were of white race. Women made up 47% of patients and 3.8% of patients were uninsured. The average number of comorbid conditions per patient was 4, and 12.4% were admitted ‘‘nonemergently’’ with an average length of hospital stay of 8 days and a mortality rate of .8%. The overall annual reversal rate was 65.5% (50,155 of 76,551). Table 3 displays the reversal rates for each demographic category. Reversal rates were significantly lower in patients aged 65 years or older. Lower reversal rates were also seen in black compared with white patients (67% vs 56%, P , .001) and uninsured patients compared with those with private insurance (63% vs 88%, P , .001). As income quartile increased, so did the stoma reversal ratedfrom 61% in the lowest income group to 72% among the highest income group (P , .001). Colostomies were also more likely to be reversed than ileostomies (79% vs 53%, P , .001). When combining various demographic categories, the following results were obtained. Middle-aged patients belonging to the highest income group with private
P value ,.001 ,.001 ,.001 ,.001
,.001
,.001
insurance had a 100% reversal rate regardless of race. When race was kept constant, insurance was still significantly associated with reversal ratesdblack middle-aged uninsured patients had a stoma reversal rate of 72% compared with a 82% reversal rate among black middleaged insured patients (P , .001). Similarly for white patientsdwhite middle-aged uninsured patients had a 59% reversal rate compared with a 96% reversal rate for white middle-aged privately insured patients (P , .001). When insurance was kept constant, racial disparities were also observeddinsured middle-aged black patient had a stoma reversal rate of 82% compared with 96% among insured middle-aged white patients (P , .001).
Comments This study demonstrates higher rates of stoma reversal for white patients, patients with private insurance, and those with higher household income. This raises serious concerns regarding disparities in access to surgical care among black patients, those uninsured, and those of lower socioeconomic status. Daluvoy et al3 found that black patients were 4 times less likely to undergo stoma reversal than white patients (odds ratio .24). However, theirs was a singleinstitution study that may not have been representative of the national population. Godat et al,2 while studying stoma reversal for trauma patients in the state of California, also had similar findings where black race was associated with a decreased rate of stoma reversal after discharge (hazard ratio .74, P , .001). Our study uses a national database, inclusive of both traumatic and nontraumatic causes, which is nationally representative. The results of our study do contrast those of Maggard et al4 who found that neither
714 race nor socioeconomic status affect stoma reversal rates. This study, however, was limited to the state of California and patients with ostomies for diverticulitis. The factors contributing to lower reversal rates among black patients cannot be determined from this study. Associations might include severity of disease process, insurance status, compliance with follow-up, or even nutritional status or BMI, but these data were not used in this study, and therefore, such inferences cannot be made. Further studies using longitudinal data sets should be performed to further elucidate the reasons behind these apparent disparities. Our study nonetheless highlights some very important issues. The lower reversal rates in black patients, and those of lower socioeconomic status, emphasize the lack of access to surgical care for these patient populations. It is very likely that lack of education about ostomy care and follow-up may be contributory to decreased stoma reversal rates and should be addressed at the patient level. Additionally, our study suggests that clinicians may need a higher threshold for creating stomas in these populations or a greater push for reversal at the same admission or better education and support with regard to follow-up. We understand that stoma reversal is not without risk of morbidity11,12 and may not improve quality of life in some instances.13 However, delays in reversal contribute significantly to postoperative complications and thereby further increase the health care burden.14 There is also a low priority given to the procedure in some institutions.12,15–18 Our study has shown that in addition to race, insurance status and income level are also significantly associated with stoma reversal rates with the lowest rates among the uninsured and those of lower income levels. These 3 factors may be inter-related and together represent a major disparity in access to surgical care for these populations. Alkire et al19 found that globally at least 5.8 billion people do not have access to surgical care. Although this is more prevalent in low- to middle-income countries, these disparities do exist in the United States. Hayanga et al20 showed that in the most segregated counties in the United States, each percentage point increase in the black or Hispanic population was associated with a statistically significant decrease in outpatient surgery volume, ambulatory surgical facilities, and number of general surgeons. These disparities account for unnecessary costs, and as of 2009, it was estimated that elimination of disparities can reduce medical expenditure by $229.4 billion over 3 years.21 There are limitations to our study that must be discussed. This is not a longitudinal data set; therefore, it is not a direct comparison. The same patients are not followed over time to see if their stomas were reversed or not. Such data on a national level do not exist at present. However, as explained earlier, for the objectives of our study, this indirect comparison is valid. Second, our definition of ‘‘temporary’’ stomas may not be accurate. For example, as discussed in the ‘‘Methods’’ section, some trauma stomas may be intended to be permanent, whereas some oncologic stomas may have been intended to be temporary. It is for
The American Journal of Surgery, Vol 211, No 4, April 2016 this reason that we do not conclude on a national ‘‘reversal rate,’’ as this figure would be not be very accurate. However, for the purposes of studying disparities, these limitations hold less value. It is thought that such inaccuracies in categorization would be roughly equal between all population subgroups and, therefore, should not affect the measurement of differences between them. Other limitations are those inherent to the use of large administrative databases for the purposes of secondary research. The accuracy is limited to the accuracy of data collection.
Conclusions Stark disparities exist in the rates of stoma closure in the United States. National stoma reversal rates are associated with the patient’s race, insurance status, and income level. There is an urgent need to address issues of access to surgical care in these populations.
References 1. Turnbull GB. The ostomy files: ostomy statistics: the $64000 question. Ostomy Wound Manage 2003;49:22–3. 2. Godat L, Kobavashi I, Chang DC, et al. Do trauma stomas ever get reversed? J Am Coll Surg 2014;219:70–7. 3. Daluvoy S, Gonzalez F, Vaziri K, et al. Factors associated with ostomy reversal. Surg Endosc 2008;22:2168–70. 4. Maggard MA, Zingmond D, O’Connell JB, et al. What proportion of patients with an ostomy (for diverticulitis) get reversed? Am Surg 2004;70:928–31. 5. Mols F, Lemmens V, Bosscha K, et al. Living with the physical and mental consequences of an ostomy: a study among 1-10 year rectal cancer survivors from the population-based PROFILES registry. Psychooncology 2014;23:998–1004. 6. Fucini C, Gattai R, Urena C, et al. Quality of life among five-year survivors after treatment for very low rectal cancer with or without a permanent abdominal stoma. Ann Surg Oncol 2008;15:1099–106. 7. Dodgion CM, Neville BA, Lipsitz SR, et al. Do older Americans undergo stoma reversal following low anterior resection for rectal cancer? J Surg Res 2013;183:238–45. 8. Tokode OM, Akingboye A, Coker O. Factors affecting reversal following Hartmann’s procedure: experience from two district general hospitals in the UK. Surg Today 2011;41:79–83. 9. Lidor AO, Gearhart SL, Wu AW, et al. Effect of race and insurance status on presentation, treatment, and mortality in patients undergoing surgery for diverticulitis. Arch Surg 2008;143:1160–5. 10. HCUP Databases. Healthcare Cost and Utilization Project (HCUP). Rockville, MD: Agency for Healthcare Research and Quality; 2015 Available at: www.hcup-us.ahrq.gov/nisoverview.jsp; 2015. Accessed January 15, 2016. 11. Chow A, Tilney HS, Paraskeva P, et al. The morbidity surrounding reversal of defunctioning ileostomies: a systematic review of 48 studies including 6,107 cases. Int J Colorectal Dis 2009;24:711–23. 12. Gessler B, Haglind E, Angenete E. Loop ileostomies in colorectal cancer patientsdmorbidity and risk factors for nonreversal. J Surg Res 2012;178:708–14. 13. Siassi M, Hohenberger W, Losel F, et al. Quality of life and patient’s expectations after closure of a temporary stoma. Int J Colorectal Dis 2008;23:1207–12. 14. Rubio-Perez I, Leon M, Pastor D, et al. Increased postoperative complications after protective ileostomy closure delay: an institutional study. World J Gastrointest Surg 2014;6:169–74.
S.N. Zafar et al.
Disparities in stoma reversal
15. Floodeen H, Lindgren R, Matthiessen P. When are defunctioning stomas in rectal cancer surgery really reversed? Results from a population-based single center experience. Scand J Surg 2013;102:246–50. 16. Alizai PH, Schulze-Hagen M, Klink CD, et al. Primary anastomosis with a defunctioning stoma versus Hartmann’s procedure for perforated diverticulitisda comparison of stoma reversal rates. Int J Colorectal Dis 2013;28:1681–8. 17. Lim SW, Kim HJ, Kim CH, et al. Risk factors for permanent stoma after low anterior resection for rectal cancer. Langenbecks Arch Surg 2013;398:259–64.
715 18. Williams NS. Stoma reversal: limitations and pitfalls. Lancet Oncol 2007;8:278–9. 19. Alkire BC, Raykar NP, Shrime MG, et al. Global access to surgical care: a modeling study. Lancet Glob Health 2015;3:e316–23. 20. Hayanga AJ, Kaiser HE, Sinha R, et al. Residential segregation and access to surgical care by minority populations in us counties. J Am Coll Surg 2009;208:1017–22. 21. LaVeist TA, Gaskin DJ, Richard P. The Economic Burden of Health Inequalities in the United States. Washington, DC: Joint Center for Political and Economic Studies; 2009.
Society of Black Academic Surgeons
Race and socioeconomic disparities in national stoma reversal rates Syed Nabeel Zafar, M.D., M.P.H.a,*, Navin R. Changoor, M.D.a,b, Kibileri Williams, M.D.a, Rafael D. Acosta, M.D.a, Wendy R. Greene, M.D.a, Terrence M. Fullum, M.D.a, Adil H. Haider, M.D., M.P.H.b, Edward E. Cornwell, III, M.D.a, Daniel D. Tran, M.D.a a
Department of Surgery, Howard University Hospital, 2041 Georgia Ave NW, Washington, DC, 20060, USA; bCenter for Surgery and Public Health, Brigham and Women’s Hospital, Boston, MA, USA
KEYWORDS: Ostomy; Stoma reversal; Health care disparities; Access to health care; Surgery
Abstract BACKGROUND: Many temporary stomas are never reversed leading to significantly worse quality of life. Recent evidence suggests a lower rate of reversal among minority patients. Our study aimed to elucidate disparities in national stoma closure rates by race, medical insurance status, and household income. METHODS: Five years of data from the Nationwide Inpatient Sample (2008 to 2012) was used to identify the annual rates of stoma formation and annual rates of stoma closure. Stomas labeled as ‘‘permanent’’ or those created secondary to colorectal cancers were excluded. Temporary stoma closure rates were calculated, and differences were tested with the chi-square test. Separate analyses were performed by race/ethnicity, insurance status, and household income. Nationally representative estimates were calculated using discharge-level weights. RESULTS: The 5-year average annual rate of temporary stoma creation was 76,551 per year (46% colostomies and 54% ileostomies). The annual rate of stoma reversal was 50,155 per year that equated to an annual reversal rate of 65.5%. Reversal rates were higher among white patients compared with black patients (67% vs 56%, P , .001) and among privately insured patients compared with uninsured patients (88% vs 63%, P , .001). Reversal rates increased as the household income increased from 61% in the lowest income quartile to 72% in the highest quartile (P , .001). CONCLUSIONS: Stark disparities exist in national rates of stoma closure. Stoma closure is associated with race, insurance, and income status. This study highlights the lack of access to surgical health care among patients of minority race and low-income status. Ó 2016 Elsevier Inc. All rights reserved.
Oral presentation at the 25th Annual Scientific Assembly of the Society of Black Academic Surgeons, April 9–11, 2015, Chapel Hill, NC. * Corresponding author. Tel.: 11-202-865-1446; fax: 11-202-8656728. E-mail address: [email protected] Manuscript received June 2, 2015; revised manuscript October 1, 2015 0002-9610/$ - see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2015.11.020
The fashioning of an ileostomy or colostomy is a relatively common surgical procedure performed for a myriad of indications spanning trauma, diverticulitis, malignancy, inflammatory bowel disease, and ischemia. These stomas can be permanent or temporary, with the latter being reversed within a very variable time frame.
S.N. Zafar et al.
Disparities in stoma reversal
An estimated 450,000 patients in the United States have an ostomy.1 An unknown proportion of these await reversal. From clinical experience and previous reports, not all stomas are reversed.2,3 There is a paucity of data regarding the national rates of stoma formation and reversal. In a retrospective analysis of 3,899 trauma patients in the state of California who underwent stoma creation, 28% were not reversed within 5 years.2 In a similar study of 1,176 patients in California with stomas created for diverticulitis, an estimated 35% of patients did not have their stomas reversed.4 Living with an ostomy has been associated with a lower quality of life, worse illness perceptions, and higher health care consumption compared with a similar patient population without an ostomy.5,6 Recent studies have shown insurance status, race, age, and income level to be associated with stoma reversal.2,3,7,8 African Americans have been found to be 4 times less likely to undergo reversal than Caucasians.3 Lack of adequate health insurance has been identified as a predictor of greater disease severity, suboptimal surgical treatment, and mortality.9 Older patients and those of lower socioeconomic status were also less likely to have their stomas reversed in a timely fashion.7 Most of these studies are, however, single-institution retrospective reviews. It is, therefore, difficult to extrapolate these results at a national level. The aim of this study was to assess race, insurance, and income-related disparities in the rates of stoma reversal on a national scale in the United States.
Methods We used 5 years of data from the Nationwide Inpatient Sample (NIS; admission years 2008 to 2012). The NIS is the largest all-payer inpatient database available.10 It is a 20% probability sample of all community hospitals and is derived from state inpatient databases. It represents 95% of the US population and includes demographic, clinical, and outcomes information. Discharge-level weights provided for each patient allow for the calculation of nationally representative estimates. The NIS does not have longitudinal data on patients beyond hospital discharge; therefore, follow-up information, such as whether an individual patient’s stoma was reversed after a few months, is unavailable. However, because the NIS is a representative sample, it does provide us with national estimates. In an ideal situation where all temporary stomas are reversed, the national 5-year average number of stomas created should equal the national 5-year average number of stomas reversed. Using discharge-level weights and survey estimation commands, we are able to determine these numbers. Therefore, we are able to calculate national stoma reversal rates by dividing the national 5-year average number of stomas reversed by the national 5-year average number of stomas created. From the NIS 2008 to 2012, we created 2 separate data sets: one with all patients who had a temporary stoma
711 created and the second with all patients who had a stoma reversal. For both data sets, we calculate 5-year national average rates. For the first data set, we used International Classification of Diseases, Ninth Revision (ICD 9), procedure codes and the NIS-provided Clinical Classification Software to select all patients who had a procedure code for ‘‘colostomy’’ (ICD 9 codes 46.10, 46.11, 46.12, 46.13, and 46.14) and ‘‘ileostomy’’ (ICD 9 codes 46.20, 46.21, 46.22, 46.23, 46.24, 46.31, 46.32, and 46.39) creation. We then excluded patients if they satisfied either one of the following exclusion criteria: (a) procedure labeled as ‘‘permanent’’ stoma, (b) stoma created because of a diagnosis of malignancy of the colon, rectum, or anus, (c) the patient died during the same hospitalization, or (d) the stoma was reversed during the same admission. In particular, malignancy of colon, rectum, and anus were excluded because of the lack of staging information that made it impossible to determine which patients had permanent ostomies because of advanced-stage cancers. We, therefore, arrived at a cohort of patients with ‘‘temporary’’ stomas. From the database, it is impossible to determine with 100% accuracy whether a stoma created was intended to be temporary or permanent. The earlier mentioned exclusion criteria offer us the closest and most reasonable estimate of ‘‘temporary’’ stomas; however, it is not an entirely accurate estimate. For example, many stomas for cancers of the colon, rectum, and anus may have been intended to be temporary, which we exclude, and conversely many trauma stomas, eg, in paraplegic patients, may have been intended to be permanent, which we include. These inaccuracies in identification are, however, thought to be unrelated to demographic and socioeconomic categories. Because the objective of our analysis is to study disparities in stoma reversal rates, it is reasonable to accept this marginal inaccuracy in identification of ‘‘temporary’’ stomas as it is unlikely to affect the disparity. Other demographic and clinical variables included in the data set were age (in years), gender, race, number of chronic conditions, admission type (elective vs emergent), insurance (Medicare, Medicaid, private insurance, uninsured or self pay, and other insurance), income quartile of patients, Zip code, primary diagnosis, length of hospital stay, and mortality. Using NIS-provided discharge-level weights, we were able to estimate the 5-year average annual rate of ‘‘temporary’’ stoma creation. We calculated this rate for different demographic categories. For the second data set, we used ICD 9 procedure codes to identify patients who underwent an ‘‘ileostomy closure’’ (ICD 9 code 46.51), ‘‘colostomy closure’’ (46.52), or ‘‘unspecified stoma closure’’ (46.50). A similar set of demographic and clinical variables were abstracted and tabulated. Using discharge-level weights, we calculated 5-year average annual rates of stoma reversal for different demographic categories as mentioned earlier. We then calculated stoma reversal rates by dividing the annual average number of stomas reversed by the annual average number of stomas created. Ideally, these 2 numbers should
712
The American Journal of Surgery, Vol 211, No 4, April 2016
be equal to provide a reversal rate of 100%. Stoma reversal rates were then calculated for different demographic and socioeconomic categories, and differences were tested for statistical significance using the chi-square test. All analyses were performed on STATA, version 12 (StataCorp, College station, TX).
Results Over the 5-year period from 2008 to 2012, a total of 566,625 stomas were created. Of these, 59,571 (10.5%) were labeled as permanent stomas, 83,298 (14.7%) were created because of malignancy of the colon, rectum, or anus, 18,949 (3.3%) had their ostomy reversed during the same admission, and 54,815 (9.7%) died during the same hospitalization. After these patients were excluded, 382,754 ‘‘temporary’’ stomas remained, of which 175,477 (45.8%)
were colostomies and 211,085 (55.1%) were ileostomies. The 5-year average annual rate of stoma creation was 76,551 stomas per year. The demographic and clinical characteristics of these patients are presented in Table 1. The average age was 59 years, 66% of patients were of white race, whereas 9.7% were black. Gender distribution was roughly equal with 48% of patients being women. Only 3.9% of patients were uninsured. On average, each patient had 5 comorbidities, 70% were admitted nonelectively, and the primary diagnosis was categorized as inflammation or infection in 45% of patients. During the same 5-year period, 250,776 stomas were reversed. Of these, 44.9% (112,483) were ileostomy closures, whereas 55.4% (138,873) were colostomy closures. The 5-year annual average rate of stoma reversal was 50,155 per year. The characteristics of patients undergoing stoma reversal are presented in Table 2. The mean age was
Table 1 Characteristics of patients receiving temporary stomas: NIS 2008 to 2012 (weighted to provide national estimates; n 5 382,754) Variable
Categories
Frequency/mean
%/SD
Age (y) Race
Mean White Black Hispanic Others Missing Female Mean 2008 2009 2010 2011 2012 Medicare Medicaid Private Uninsured Other Missing Yes Lowest 2 3 Highest Missing Inflammation or infection Trauma Complications of surgery or medical device Other cancer Obstruction Others Ileostomy Colostomy Mean
58.7 y 252,265 37,262 27,914 17,783 47,529 186,524 5.2 70,565 74,932 78,652 84,090 74,515 179,597 44,905 129,212 15,042 11,425 832 268,168 96,971 98,532 94,688 84,581 7,983 173,156 7,604 39,554 58,151 32,344 43,004 211,085 175,477 17.8 d
620.0 y 65.9% 9.7% 7.3% 4.6% 12.4% 48.7% 63.1 18.4% 19.6% 20.6% 22.0% 19.5% 47.0% 11.8% 33.8% 3.9% 3.0% .2% 70.1% 25.3% 25.7% 24.7% 22.1 2.1% 45.2% 2.0% 10.3% 15.2% 8.5% 15.3% 55.1% 45.8% 619.1 d
Gender Number of chronic conditions Year
Insurance
Nonelective admission Income quartile of patient ZIP
Primary diagnosis
Type of stoma Length of hospital stay SD 5 standard deviation.
S.N. Zafar et al.
Disparities in stoma reversal
713
Table 2 Characteristics of patients undergoing stoma reversal, NIS 2008 to 2012 (weighted to provide national estimates; n 5 250,776)
Table 3 Annual stoma reversal rates by patient characteristics (5-year average annual rates)
Variable
Categories Frequency/mean %/SD
Stoma
Created
Reversed
Reversal rate %
Age (y) Race
Mean White Black Hispanic Others Missing Female Mean
Age 651 Age18–65 Age ,18 Female Male White Black Medicare Medicaid Private Uninsured Income lowest Income level 2 Income level 3 Income highest Ileostomy Colostomy
33,449 39,499 3,603 37,306 39,219 50,453 7,452 36,992 8,981 25,843 3,008 19,394 19,706 18,938 16,916 42,217 35,095
16,153 31,262 2,741 23,625 26,411 33,865 4,190 17,128 6,016 22,807 1,887 11,866 12,640 12,466 12,143 22,497 27,775
48.3 79.1 76.1 63.3 67.3 67.1 56.2 46.3 67.0 88.3 62.7 61.2 64.1 65.8 71.8 53.1 79.1
54.1 y 169,323 20,949 19,564 11,163 29,777 118,125 3.9
619.4 y 67.5% 8.4% 7.8% 4.5% 11.9% 47.1% 62.5
2008 50,614 2009 50,195 2010 49,706 2011 51,531 2012 48,730 Insurance Medicare 85,639 Medicaid 30,082 Private 114,035 Uninsured 9,436 Other 9,536 Missing 483 Nonelective admission Yes 31,089 Income quartile of Lowest 59,330 patient ZIP 2 63,201 3 62,329 Highest 60,715 Missing 5,201 Type of stoma Ileostomy 112,483 Colostomy 138,873 Length of hospital stay Mean 8.0 Mortality Yes 2,002
20.2% 20.0% 19.8% 20.6% 19.4% 34.2% 12.0% 45.5% 3.8% 3.8% .2% 12.4% 23.7%
Gender Number of chronic conditions Year
25.2% 24.9% 24.2% 2.1% 44.8% 55.3% 612.2 .8%
SD 5 standard deviation.
54 years and 68% of patients were of white race. Women made up 47% of patients and 3.8% of patients were uninsured. The average number of comorbid conditions per patient was 4, and 12.4% were admitted ‘‘nonemergently’’ with an average length of hospital stay of 8 days and a mortality rate of .8%. The overall annual reversal rate was 65.5% (50,155 of 76,551). Table 3 displays the reversal rates for each demographic category. Reversal rates were significantly lower in patients aged 65 years or older. Lower reversal rates were also seen in black compared with white patients (67% vs 56%, P , .001) and uninsured patients compared with those with private insurance (63% vs 88%, P , .001). As income quartile increased, so did the stoma reversal ratedfrom 61% in the lowest income group to 72% among the highest income group (P , .001). Colostomies were also more likely to be reversed than ileostomies (79% vs 53%, P , .001). When combining various demographic categories, the following results were obtained. Middle-aged patients belonging to the highest income group with private
P value ,.001 ,.001 ,.001 ,.001
,.001
,.001
insurance had a 100% reversal rate regardless of race. When race was kept constant, insurance was still significantly associated with reversal ratesdblack middle-aged uninsured patients had a stoma reversal rate of 72% compared with a 82% reversal rate among black middleaged insured patients (P , .001). Similarly for white patientsdwhite middle-aged uninsured patients had a 59% reversal rate compared with a 96% reversal rate for white middle-aged privately insured patients (P , .001). When insurance was kept constant, racial disparities were also observeddinsured middle-aged black patient had a stoma reversal rate of 82% compared with 96% among insured middle-aged white patients (P , .001).
Comments This study demonstrates higher rates of stoma reversal for white patients, patients with private insurance, and those with higher household income. This raises serious concerns regarding disparities in access to surgical care among black patients, those uninsured, and those of lower socioeconomic status. Daluvoy et al3 found that black patients were 4 times less likely to undergo stoma reversal than white patients (odds ratio .24). However, theirs was a singleinstitution study that may not have been representative of the national population. Godat et al,2 while studying stoma reversal for trauma patients in the state of California, also had similar findings where black race was associated with a decreased rate of stoma reversal after discharge (hazard ratio .74, P , .001). Our study uses a national database, inclusive of both traumatic and nontraumatic causes, which is nationally representative. The results of our study do contrast those of Maggard et al4 who found that neither
714 race nor socioeconomic status affect stoma reversal rates. This study, however, was limited to the state of California and patients with ostomies for diverticulitis. The factors contributing to lower reversal rates among black patients cannot be determined from this study. Associations might include severity of disease process, insurance status, compliance with follow-up, or even nutritional status or BMI, but these data were not used in this study, and therefore, such inferences cannot be made. Further studies using longitudinal data sets should be performed to further elucidate the reasons behind these apparent disparities. Our study nonetheless highlights some very important issues. The lower reversal rates in black patients, and those of lower socioeconomic status, emphasize the lack of access to surgical care for these patient populations. It is very likely that lack of education about ostomy care and follow-up may be contributory to decreased stoma reversal rates and should be addressed at the patient level. Additionally, our study suggests that clinicians may need a higher threshold for creating stomas in these populations or a greater push for reversal at the same admission or better education and support with regard to follow-up. We understand that stoma reversal is not without risk of morbidity11,12 and may not improve quality of life in some instances.13 However, delays in reversal contribute significantly to postoperative complications and thereby further increase the health care burden.14 There is also a low priority given to the procedure in some institutions.12,15–18 Our study has shown that in addition to race, insurance status and income level are also significantly associated with stoma reversal rates with the lowest rates among the uninsured and those of lower income levels. These 3 factors may be inter-related and together represent a major disparity in access to surgical care for these populations. Alkire et al19 found that globally at least 5.8 billion people do not have access to surgical care. Although this is more prevalent in low- to middle-income countries, these disparities do exist in the United States. Hayanga et al20 showed that in the most segregated counties in the United States, each percentage point increase in the black or Hispanic population was associated with a statistically significant decrease in outpatient surgery volume, ambulatory surgical facilities, and number of general surgeons. These disparities account for unnecessary costs, and as of 2009, it was estimated that elimination of disparities can reduce medical expenditure by $229.4 billion over 3 years.21 There are limitations to our study that must be discussed. This is not a longitudinal data set; therefore, it is not a direct comparison. The same patients are not followed over time to see if their stomas were reversed or not. Such data on a national level do not exist at present. However, as explained earlier, for the objectives of our study, this indirect comparison is valid. Second, our definition of ‘‘temporary’’ stomas may not be accurate. For example, as discussed in the ‘‘Methods’’ section, some trauma stomas may be intended to be permanent, whereas some oncologic stomas may have been intended to be temporary. It is for
The American Journal of Surgery, Vol 211, No 4, April 2016 this reason that we do not conclude on a national ‘‘reversal rate,’’ as this figure would be not be very accurate. However, for the purposes of studying disparities, these limitations hold less value. It is thought that such inaccuracies in categorization would be roughly equal between all population subgroups and, therefore, should not affect the measurement of differences between them. Other limitations are those inherent to the use of large administrative databases for the purposes of secondary research. The accuracy is limited to the accuracy of data collection.
Conclusions Stark disparities exist in the rates of stoma closure in the United States. National stoma reversal rates are associated with the patient’s race, insurance status, and income level. There is an urgent need to address issues of access to surgical care in these populations.
References 1. Turnbull GB. The ostomy files: ostomy statistics: the $64000 question. Ostomy Wound Manage 2003;49:22–3. 2. Godat L, Kobavashi I, Chang DC, et al. Do trauma stomas ever get reversed? J Am Coll Surg 2014;219:70–7. 3. Daluvoy S, Gonzalez F, Vaziri K, et al. Factors associated with ostomy reversal. Surg Endosc 2008;22:2168–70. 4. Maggard MA, Zingmond D, O’Connell JB, et al. What proportion of patients with an ostomy (for diverticulitis) get reversed? Am Surg 2004;70:928–31. 5. Mols F, Lemmens V, Bosscha K, et al. Living with the physical and mental consequences of an ostomy: a study among 1-10 year rectal cancer survivors from the population-based PROFILES registry. Psychooncology 2014;23:998–1004. 6. Fucini C, Gattai R, Urena C, et al. Quality of life among five-year survivors after treatment for very low rectal cancer with or without a permanent abdominal stoma. Ann Surg Oncol 2008;15:1099–106. 7. Dodgion CM, Neville BA, Lipsitz SR, et al. Do older Americans undergo stoma reversal following low anterior resection for rectal cancer? J Surg Res 2013;183:238–45. 8. Tokode OM, Akingboye A, Coker O. Factors affecting reversal following Hartmann’s procedure: experience from two district general hospitals in the UK. Surg Today 2011;41:79–83. 9. Lidor AO, Gearhart SL, Wu AW, et al. Effect of race and insurance status on presentation, treatment, and mortality in patients undergoing surgery for diverticulitis. Arch Surg 2008;143:1160–5. 10. HCUP Databases. Healthcare Cost and Utilization Project (HCUP). Rockville, MD: Agency for Healthcare Research and Quality; 2015 Available at: www.hcup-us.ahrq.gov/nisoverview.jsp; 2015. Accessed January 15, 2016. 11. Chow A, Tilney HS, Paraskeva P, et al. The morbidity surrounding reversal of defunctioning ileostomies: a systematic review of 48 studies including 6,107 cases. Int J Colorectal Dis 2009;24:711–23. 12. Gessler B, Haglind E, Angenete E. Loop ileostomies in colorectal cancer patientsdmorbidity and risk factors for nonreversal. J Surg Res 2012;178:708–14. 13. Siassi M, Hohenberger W, Losel F, et al. Quality of life and patient’s expectations after closure of a temporary stoma. Int J Colorectal Dis 2008;23:1207–12. 14. Rubio-Perez I, Leon M, Pastor D, et al. Increased postoperative complications after protective ileostomy closure delay: an institutional study. World J Gastrointest Surg 2014;6:169–74.
S.N. Zafar et al.
Disparities in stoma reversal
15. Floodeen H, Lindgren R, Matthiessen P. When are defunctioning stomas in rectal cancer surgery really reversed? Results from a population-based single center experience. Scand J Surg 2013;102:246–50. 16. Alizai PH, Schulze-Hagen M, Klink CD, et al. Primary anastomosis with a defunctioning stoma versus Hartmann’s procedure for perforated diverticulitisda comparison of stoma reversal rates. Int J Colorectal Dis 2013;28:1681–8. 17. Lim SW, Kim HJ, Kim CH, et al. Risk factors for permanent stoma after low anterior resection for rectal cancer. Langenbecks Arch Surg 2013;398:259–64.
715 18. Williams NS. Stoma reversal: limitations and pitfalls. Lancet Oncol 2007;8:278–9. 19. Alkire BC, Raykar NP, Shrime MG, et al. Global access to surgical care: a modeling study. Lancet Glob Health 2015;3:e316–23. 20. Hayanga AJ, Kaiser HE, Sinha R, et al. Residential segregation and access to surgical care by minority populations in us counties. J Am Coll Surg 2009;208:1017–22. 21. LaVeist TA, Gaskin DJ, Richard P. The Economic Burden of Health Inequalities in the United States. Washington, DC: Joint Center for Political and Economic Studies; 2009.