Socioeconomic disparities in ostomy reversal among older adults with diverticulitis are more substantial among non-Hispanic Black patients

Abstract

Background:

While ostomies for diverticulitis are often intended to be temporary, ostomy reversal rates can be as low as 46%. There are few comprehensive studies evaluating the effects of socioeconomic status as a disparity in ostomy reversal. We hypothesized that among the elderly Medicare population undergoing partial colectomy for diverticulitis, lower socioeconomic status would be associated with reduced reversal rates.

Methods:

Retrospective cohort study using a 20% representative sample of Medicare beneficiaries >65 years old with diverticulitis who received ostomies between January 1, 2010, to December 31, 2017. We evaluated the effect of neighborhood socioeconomic status, measured by the Social Deprivation Index, on ostomy reversal within 1 year. Secondary outcomes were complications and mortality.

Results:

Of 10,572 patients, ostomy reversals ranged from 21.2% (low socioeconomic status) to 29.8% (highest socioeconomic status), with a shorter time to reversal among higher socioeconomic status groups. Patients with low socioeconomic status were less likely to have their ostomies reversed, compared with the highest socioeconomic status group (hazard ratio 0.83, 95% confidence interval 0.74–0.93) and were more likely to die (hazard ratio 1.21, 95% confidence interval 1.10–1.33). When stratified by race/ethnicity and socioeconomic status, non-Hispanic White patients at every socioeconomic status had a higher reversal rate than non-Hispanic Black patients (White patients 32.0%–24.8% vs Black patients 19.6%–14.7%). Socioeconomic status appeared to have a higher relative impact among non-Hispanic Black patients.

Conclusion:

Among Medicare diverticulitis patients, ostomy reversal rates are low. Patients with lower socioeconomic status are less likely to undergo stoma reversal and are more likely to die; Black patients are least likely to have an ostomy reversal.

Introduction

Diverticular disease is a common pathology in the United States, and the incidence is increasing as our population ages and becomes more obese.¹ The rate of patients progressing to acute diverticulitis is approximately 4%, with the majority managed non-operatively.^2,3 Among patients who develop generalized peritonitis requiring emergency surgery, the operation generally involves a partial colectomy with a Hartmann’s pouch and end-colostomy, with some patients undergoing a partial colectomy with primary anastomosis and a diverting loop ileostomy to protect the new anastomosis.^2,4–6 These ostomies are often intended to be temporary, but research has shown that rates of ostomy reversal for all temporary ostomies range from 46% to 83%.^6–12 The decision to reverse an ostomy and the timing of reversal can be dependent on a number of factors including the primary operation, postoperative complications, and the patient’s health status and preferences, as well as their comorbidities.^13,14 Financial constraints may play a part as well and may increase loss to follow up. Due to a lack of guidelines, patient selection and timing are commonly based on surgeon preference.¹²

Ostomy reversal is an area where disparities exist.^10,15 A number of studies suggest that in patients with diverticulitis requiring surgery and ostomy creation, patients who are Black, Hispanic, uninsured, or have a lower household income are less likely to have their ostomies reversed or are more likely to undergo reversal in a delayed fashion.^{9,11,12,15,16} Additionally, it appears that patients are less likely to have an ostomy reversal if they are older, regardless of comorbidity status.⁹ These trends open the door to significant disparities in quality of life and complication risk owing to delayed ostomy closure.^13,14 Most of these studies only investigated patients who had a Hartmann’s procedure, did not evaluate socioeconomic status (SES), were single institution, or were limited in geographic distribution.^{9,11,12,15,16} These studies also included patients of all ages, despite the burden of diverticulitis being higher in older adults and age being a significant determinant of surgical outcomes. Thus, the purpose of this study is to expand upon these early data that suggest significant disparities in ostomy reversal but restrict to a universally insured, older adult population. We hypothesized that lower SES would be independently associated with lower rates of loop ileostomy and end colostomy reversal following sigmoid colectomy for diverticulitis. We also hypothesized that lower SES would be associated with increased complications and mortality.

Methods

Data source

This study was conducted using a 20% representative sample of all Medicare beneficiaries ≥65 years old with fee-for-service coverage. Beneficiaries were included if they had simultaneous coverage of Parts A, B, and D for at least 1 calendar month between 2007 and 2017, although all available fee-for-service claims and death data are included once those criteria are met. As of 2014, roughly 70% of Medicare beneficiaries had Part D coverage, and older adults (≥65 years old) make up over 80% of the covered population (over 40 million adults each year).¹⁷ We captured diagnoses and procedures using International Classification of Diseases Ninth and Tenth edition, clinical modification (ICD-9-CM and ICD-10-CM) diagnosis and procedure codes, as well as Current Procedural Terminology (CPT) codes.

Study population

All older adults with diverticulitis that received an ostomy between January 1, 2010, to December 31, 2017, were eligible for inclusion. We required patients to have continuous enrollment in Medicare Parts A and B for at least 365 days before their ostomy procedure; thus, patients <66 years old were excluded (n = 193) (Fig 1). Patients were censored when their coverage changed or was dropped. The number of censored patients was low such that a coverage issue was unlikely to contribute to reversal rates. We also excluded patients diagnosed with colorectal cancer (n = 3,517) or who had another ostomy procedure (placement or reversal) within the year before their included ostomy procedure (n = 55). The lookback period of 1 year was chosen to allow for calculation of frailty index and to ensure that stoma reversals were specific to the ostomy placement of interest. The ICD-9-CM, ICD-10-CM, and CPT codes used to identify diagnoses and procedures of interest are reported in Supplemental Table S1. We identified ICD-9-CM and CPT codes using prior studies and clinical expertise; we identified ICD-10-CM procedure codes using forward and backward mapping with Centers for Medicare and Medicaid Services Generalized Equivalence Mappings and then reviewed the codes.

Fig 1.

Study design and timeline to capture outcomes among patients with diverticulitis who underwent ostomy.

Socioeconomic status (SES)

The primary exposure of interest was neighborhood SES, which was measured using the Social Deprivation Index (SDI). The SDI is a composite measure of area-level deprivation (in this study, ZIP codes), based on 7 demographic characteristics captured in the American Community Survey: (1) percent living in poverty, (2) percent with <12 years of education, (3) percent of single parent households, (4) percent living in rented housing unit, (5) percent living in overcrowded housing unit, (6) percent of households without a car, and (7) percent of non-employed adults <65 years old. A higher score indicates a higher level of deprivation and lower SES. This measure was initially developed by Butler et al and has been updated using data from the 2008 to 2012 and 2011 to 2015 American Community Survey data.^18,19

Patient ZIP code was captured using the billing ZIP code listed on the hospitalization for the original ostomy surgery and then linked to a database of SDIs for individual ZIP codes. Because SDI is only measured for ZIP codes within the United States, individuals living in Puerto Rico or the American Virgin Islands (n = 27) were excluded from this study. There were also 179 other patients whose ZIP code was not included in the SDI database and thus they were excluded. For this study, we categorized SDI into quartiles (highest SES: 1–24, high SES: 25–46, medium SES: 47–69, and low SES: 70–100).

Follow-up and other covariates of interest

Patients were then followed until ostomy reversal, death, disenrollment in Medicare Part A or B, or for 365 days after their ostomy procedure, whichever came first (Fig 1). Other outcomes of interest included ostomy revision anytime during the 1-year follow-up and 30-day surgical complications (Supplemental Table S1). As a sensitivity analysis for our follow-up period, we also captured ostomy reversals that occurred >1 year after placement.

Complications of interest included myocardial infarction, surgical site infection, abscess, bloodstream infection, pneumonia, respiratory failure, venous thromboembolism, gastrointestinal bleed, ostomy complications (peristomal infections, fistulas, ulcerations, and perforations at the stomal site), parastomal hernia, ureteral injury, and blood transfusion. We assigned the date of diagnosis as either the date of admission (inpatient hospitalization records) or date of the service (inpatient and outpatient service records). For complications diagnosed during the index hospitalization, “present on admission” flags were used to help differentiate between pre-existing conditions and surgical complications. Diagnoses that were flagged as “present on admission” were assumed to have occurred before the ostomy procedure (and therefore not complications).

Other covariates of interest included patient age, sex, race/ethnicity, comorbidities, frailty, and type of ostomy. Comorbidities and frailty were measured using the lookback period. Charlson Comorbidity Index (CCI) score was calculated using Deyo et al method, and frailty score was calculated using Gautam’s method.^20,21

Statistical analysis

Descriptive statistics were used to report patient demographics and comorbidities, stratified by neighborhood SES. Kaplan-Meier curves were used to assess the cumulative incidence of ostomy reversal, stratified by SES and race/ethnicity.

Multivariable Cox proportional hazards regression was used to estimate the association between neighborhood SES and the 1-year incidence of ostomy reversal, after adjusting for patient sex, age, race/ethnicity, CCI score, frailty score, ostomy revision, and 30-day complications. Age, CCI score, and frailty score were modeled as restricted quadratic splines. Race/ethnicity was collapsed into non-Hispanic White, non-Hispanic Black, and other race/ethnicity owing to small cell sizes for analyses. Ostomy revision and 30-day complication rates were treated as time-varying confounders; patients were considered ‘exposed’ starting on the date of the revision or complication. Mortality was treated as a competing risk.

Similar methods were used to estimate the association between SES, 30-day complications, and mortality. We also performed additional analyses to assess whether the effect of SES on reversal and mortality was consistent across racial and ethnic groups. Separate models were run for each racial/ethnic group. We also combined race/ethnicity (non-Hispanic White and non-Hispanic Black only) into an 8-level variable and analyzed association with ostomy reversal; non-Hispanic White patients in the highest SES neighborhoods were the reference group.

All analyses were conducted using SAS software, version 9.4 (SAS Institute Inc., Cary, North Carolina). This study was compliant with the Health Insurance Portability and Accountability Act and classified as exempt by the University of North Carolina at Chapel Hill Institutional Review Board (IRB# 19–2857).

Results

Between 2010 and 2017, there were 14,543 stomas created among older adults (≥66 years old). After applying exclusions, 10,572 patients were included in the study. Overall, 75.5% (n = 7,981) of patients underwent colostomies and 24.5% (n = 2,591) underwent ileostomies. The median age was 76 (interquartile range [IQR] 71–82), 69.8% (n = 7,384) were female, and the racial/ethnic breakdown was 83% (n = 9,339) non-Hispanic White, 7.6% (n = 800) non-Hispanic Black, 0.8% (n = 80) Asian, 1.6% (n = 172) Hispanic, and 1.2% (n = 125) non-Hispanic other. The median CCI was 3 (IQR 2–5) and median frailty score was 0.2 (IQR 0.2–0.3). There were no differences in CCI or frailty across SES or race (Table I, Supplemental Table S2). Besides race/ethnicity, minimal differences in patient characteristics were seen across neighborhood SES (Table I).

Table I.

Demographic and surgical characteristics, stratified by neighborhood SES,^* n = 10,572

	Highest SES	High SES	Medium SES	Low SES
Total, N	2,748	2,600	2,638	2,586
Age, median (IQR)	76 (71–82)	76 (71–82)	76 (71–81)	76 (71–82)
Female, n (%)	1898 (69.1)	1785 (68.7)	1874 (71.0)	1827 (70.7)
Race/ethnicity,† n (%)
Non-Hispanic White	2,626 (96.4)	2,444 (94.4)	2,389 (90.9)	1,880 (73.0)
Non-Hispanic Black	50 (1.8)	85 (3.3)	155 (5.9)	510 (19.8)
Non-Hispanic Asian	15 (0.6)	15 (0.6)	19 (0.7)	31 (1.2)
Hispanic	11 (0.4)	18 (0.7)	31 (1.2)	112 (4.4)
Non-Hispanic other	23 (0.8)	26 (1.0)	33 (1.3)	43 (1.7)
CCI score, median (IQR)	3 (1–5)	3 (2–5)	3 (1–5)	4 (2–6)
Frailty Score, median (IQR)	0.2 (0.2–0.3)	0.2 (0.2–0.3)	0.2 (0.2–0.3)	0.2 (0.2–0.3)
Type of ostomy, n (%)
Colostomy	2058 (74.9)	1963 (75.5)	2021 (76.6)	1939 (75.0)
Ileostomy	690 (25.1)	637 (24.5)	617 (23.4)	647 (25.0)

CCI, Charlson comorbidity index; IQR, interquartile range; SES, socio-economic status.

^*Neighborhood SES was estimated using ZIP code Social Deprivation Index (SDI) scores; scores were then split into quartiles.

^†Fifty-six patients were missing race/ethnicity.

The overall incidence of 1-year ostomy reversals ranged from 21.2% (low SES) to 31.8% (highest SES) (Fig 2). The majority of reversals occurred between 90 and 180 days after placement. Very few reversals occurred after 365 days (n = 194, 6.6%). Median time to reversal was also slightly shorter in the higher SES groups (Table II). After adjustment, patients with low SES were still substantially less likely to have their ostomies reversed, compared with the highest SES group (hazard ratio [HR] 0.83, 95% confidence interval [CI] 0.74–0.93) (Table II). It appeared that patients from high and medium SES neighborhoods were also slightly less likely to have their ostomy reversed, compared with the highest SES group, but estimates were imprecise and not statistically significant (high SES: HR 0.94, 95% CI 0.85–1.04; medium SES: HR 0.92, 95% CI 0.83–1.02).

Fig 2.

Rate of ostomy reversals over time, stratified by neighborhood SES. SES, socioeconomic status.

Table II.

One year incidence of ostomy reversal, stratified by neighborhood SES^* and race and/or ethnicity

	Highest SES	High SES	Medium SES	Low SES
Total, N	2,748	2,600	2,638	2,586
Ostomy reversal, n (%)	820 (29.8)	717 (27.6)	705 (26.7)	549 (21.2)
Days to reversal†, med (IQR)	125 (88–184)	135 (92–194)	133 (91–202)	143 (95–198)
HR (95% CI)‡	1.0 (ref)	0.89 (0.81–0.99)	0.86 (0.78–0.95)	0.66 (0.59–0.73)
Adjusted HR (95% CI)‡,§	1.0 (ref)	0.94 (0.85–1.04)	0.92 (0.83–1.02)	0.83 (0.74–0.93)
Non-Hispanic White
Adjusted HR (95% CI)ǁ	1.0 (ref)	0.95 (0.85–1.05)	0.93 (0.83–1.03)	0.83 (0.74–0.93)
Non-Hispanic Black
Adjusted HR (95%CI)ǁ	1.0 (ref)	0.72 (0.29–1.79)	0.53 (0.24–1.17)	0.62 (0.31–1.27)
Other race/ethnicity¶
Adjusted HR (95% CI)ǁ	1.0 (ref)	0.68 (0.27–1.71)	1.33 (0.62–2.84)	0.77 (0.38–1.56)

CI, confidence interval; HR, hazard ratio; med, median; IQR, interquartile range; SES, socio-economic status.

^*Neighborhood SES was estimated using ZIP code Social Deprivation Index (SDI) scores; scores were then split into quartiles.

^†Among those who had an ostomy reversal within 1 year of placement.

^‡Death was treated as a competing risk.

^§Adjusted for patient sex, age, race/ethnicity, Charlson Comorbidity Index (CCI) score, frailty score, ostomy revision, and 30-day complications; age, CCI score, and frailty score were modeled as restricted quadratic splines; 30-day complications and ostomy revision were treated as time-varying covariates.

^ǁAdjusted for patient sex, age, Charlson Comorbidity Index (CCI) score, frailty score, ostomy revision, and 30-day complications; age, CCI score, and frailty score were modeled as restricted quadratic splines; 30-day complications and ostomy revision were treated as time-varying covariates; separate models were run for each race/ethnicity.

^¶Includes non-Hispanic Asian, Hispanic, and non-Hispanic other races; race/ethnicity was collapsed owing to small cell sizes.

Overall colostomy reversal rate was 27.4% (n = 2,188/7,981) and ileostomy reversal rate was 23.3% (n = 603/2591). When stratified by colostomy versus ileostomy, the association between SES and colostomy reversal rates mirrored the overall findings, while the impact of SES on ileostomy reversal rates was consistently more substantial (high SES: HR 0.78, 95% CI 0.62–0.97; medium SES: HR 0.75, 95% CI 0.60–0.94; low SES: HR 0.73, 95% CI 0.58–0.93) (Supplemental Table S3).

Interestingly, all-cause mortality was also higher among patients from low SES neighborhoods, but 30-day complication rates were similar (Table III and Supplemental Table S4). For example, patients with low SES, compared with highest SES, were 20% more likely to die within a year of their surgery (HR 1.21, 95% CI 1.10–1.33), but minimal differences were seen in 30-day complications (any complication: HR 1.04, 95% CI 0.98–1.10).

Table III.

One year mortality rates, stratified by neighborhood SES^*

	Highest SES	High SES	Medium SES	Low SES
Total, N	2,748	2,600	2,638	2,586
Deaths, n (%)	794 (28.9)	857 (33.0)	860 (32.6)	978 (37.8)
Days to death†, med (IQR)	365 (192–365)	365 (108–365)	365 (100–365)	365 (74–365)
HR (95% CI)	1.0 (ref)	1.17 (1.07–1.29)	1.17 (1.06–1.29)	1.38 (1.26–1.51)
Adjusted HR (95% CI)‡	1.0 (ref)	1.08 (0.98–1.19)	1.11 (1.01–1.22)	1.21 (1.10–1.33)

CI, confidence interval; HR, hazard ratio; med, median; IQR, interquartile range; SES, socio-economic status.

^*Neighborhood SES was estimated using ZIP code Social Deprivation Index (SDI) scores; scores were then split into quartiles.

^†Among those who died within 1-year of their ostomy placement.

^‡Adjusted for patient sex, age, race/ethnicity, Charlson Comorbidity Index (CCI) score, frailty score, and cirrhosis; age, CCI score, and frailty score were modeled as restricted quadratic splines; 30-day complications and ostomy revision were not included in the model because they were considered mediators.

The effect of SES also appeared to differ across patient race/ethnicity. Among non-Hispanic White patients, only small differences were seen between patients with high and medium SES, compared with the highest SES group (high SES: HR 0.95, 95% CI 0.85, 1.05; medium SES: HR 0.93, 95% CI 0.83–1.03) (Table II). However, among non-Hispanic Black patients, SES had a major impact on reversal rates at all levels of SES (high SES: HR 0.72, 95% CI 0.29–1.79; medium SES: 0.53, 95% CI 0.24–1.17; low SES: HR 0.62, 95% CI 0.31–1.27), although estimates were imprecise and not statistically significant owing to the smaller sample size (Table II). When race/ethnicity and SES were combined, non-Hispanic White patients at every SES level had higher reversal rates than any non-Hispanic Black patients (non-Hispanic White patients: highest 32.0%, high 29.9%, medium 29.4%, low 24.5%; non-Hispanic Black patients: highest 19.6%, high 15.9%, medium 14.9%, low 14.7%). Compared with non-Hispanic White patients in the highest SES group, White patients in the lowest SES group were almost 20% less likely to have their ostomy reversed (HR 0.83, 95% CI 0.74–0.93), whereas Black patients from both the medium and low SES neighborhoods were over 40% less likely (HR 0.51, 95% CI 0.33–0.78 and HR 0.58, 95% CI 0.45–0.74, respectively) (Table IV). White patients in the low SES group appeared to have the same reduction in ostomy reversal as Black patients in the highest SES group, although estimates were imprecise owing to small sample sizes (HR 0.83, 95% CI 0.74–0.93 and HR 0.83, 95% CI 0.45–1.56, respectively.) Overall, non-Hispanic Black patients had the lowest rate of ostomy reversals compared with non-Hispanic White and other race/ethnicity groups (Fig 3).

Table IV.

Cumulative incidence of Ostomy Reversal among non-Hispanic White and non-Hispanic Black patients, stratified by SES.

	Total, N	Reversals, % (IQR)	HR (95% CI)*
Non-Hispanic White
Highest SES	2,626	32.0 (30.2–33.9)	1.0 (ref)
High SES	2,444	29.9 (28.2–31.9)	0.95 (0.85–1.05)
Medium SES	2,389	29.4 (27.7–31.2)	0.92 (0.83–1.03)
Low SES	1,880	24.8 (23.0–26.8)	0.83 (0.74–0.93)
Non-Hispanic Black
Highest SES	50	19.6 (10.6–36.2)	0.83 (0.45–1.56)
High SES	85	15.9 (10.2–24.7)	0.65 (0.34–1.21)
Medium SES	155	14.9 (10.4–21.4)	0.51 (0.33–0.78)
Low SES	510	14.7 (11.8–18.2)	0.58 (0.45–0.74)

CI, confidence interval; HR, hazard ratio; IQR, interquartile range; SES, socio-economic status.

^*Adjusted for patient sex, age, Charlson Comorbidity Index (CCI) score, frailty score, ostomy revision, and 30-day complications; age, CCI score, and frailty score were modeled as restricted quadratic splines; 30-day complications and ostomy revision were treated as time-varying covariates; SES and race/ethnicity were combined into a single, 8-level variable for analysis.

Fig 3.

Rate of ostomy reversals over time, stratified by race/ethnicity.

Discussion

Overall, among universally insured older adults, the rate of 1-year ostomy reversal after partial colectomy for diverticulitis remains low (<30% across all SES levels and racial/ethnic groups). Neighborhood SES was highly correlated with ostomy reversal, with patients living in low SES neighborhoods being 20% less likely to have a reversal. Moreover, the impact of SES was disproportionate across race/ethnicity. While only non-Hispanic White patients living in low SES neighborhoods appeared to have a slightly lower rate of ostomy reversal (10% less), non-Hispanic Black patients living in high SES neighborhoods had a 30% decreased incidence, and medium/low SES patients had a 40% decrease. Race/ethnicity may play a larger role than SES in this universally insured population, as Non-Hispanic Black patients had lower rates of reversal compared with non-Hispanic White patients regardless of SES. SES was also associated with higher rates of mortality, but little difference was seen across 30-day complication rates. These findings highlight the ongoing racial and socioeconomic disparities in surgical care across the United States.

The inequalities revealed by this study are a symptom of a looming problem of inequality prevalent in the United States and mirror trends in disparities seen in other investigations. Counties in the United States with higher percentages of minorities have reduced access to surgical care and fewer surgical facilities and surgeons.²² Multiple studies have demonstrated worse surgical outcomes among Black patients.^23–27 There are a few other studies demonstrating an association between ostomy reversal, race, and SES; however, the majority of these studies evaluated smaller data sets limited to single institutions or specific states, do not specifically look at elderly patients, have a broad range of diagnoses, or evaluated race/ethnicity only.^{9,11,12,15,16} The most comprehensive study to date, Zafar et al, used the National Inpatient Sample and found correlations between ostomy reversal rates and uninsured status, lower income levels, and Black race.¹⁵ However, the National Inpatient Sample is a de-identified dataset so researchers were unable to follow patients over time or calculate the true reversal rates across subgroups.

Disparities exist secondary to a complex interplay between socioeconomic, environmental, political, and racial factors.²⁸ Ostomy reversal, in particular, may be related to issues such as individual financial hardship, health literacy and education, access to surgical care, and postoperative ostomy education. Unconscious bias may play a role as well, given that surgeons have few guidelines to follow. Given that this Medicare cohort all had identical insurance, equitable insurance coverage alone will not resolve access issues. Complication rates were similar across SES groups; thus, presence of complications does not explain the observed disparities.

SES and race/ethnicity are intricately entangled in the United States. Non-Hispanic Black patients had lower ostomy reversal rates regardless of SES, yet the combination of Black race and lower SES also significantly reduced reversal rates. Discrimination and marginalization of racial and ethnic minorities can lead to a cycle of poverty and lack of access to care, education, and equal opportunities in employment. For example, Non-Hispanic Black children are more likely to attend high poverty schools (45% of non-Hispanic Blacks compared with 8% of non-Hispanic Whites) and are more likely to live in poverty (31% of non-Hispanic Black children compared with 10% of non-Hispanic White children).^29,30 Subtle cultural differences may play a role as well. Non-Hispanic Black patients may choose surgery less frequently than non-Hispanic Whites and may have long-standing distrust in the health system based on historical experiences such as the Tuskegee experiments, vesicovaginal fistula operations on female slaves, and grave robbing of Black cemeteries for cadaveric dissections in the late 19th century.^31–34 Haider et al have suggested multiple overlapping issues leading to racial disparities, including systemic factors like hospital capacity in largely minority populations; patient factors such as comorbidities, willingness to undergo surgery, advanced presentation, insurance status, and SES; and provider factors involving fewer consultations and referrals for Black patients and treatment of Black patients by lower volume, lower-tier surgeons.²³

Acknowledging the presence of disparities is the first step toward addressing them. However, modifying these disparities can be complicated given the nuance of contributing factors. More easily modifiable measures that may reduce racial and socioeconomic disparities in ostomy reversal include standardizing follow-up care, discharge and education protocols, follow-up telephone calls and protocols to prevent loss to follow-up, language assistance, both written and verbal education, and financial assistance programs. Protocols around ileostomy care have been shown to reduce readmissions.³⁵ Studies reveal that surgical patients perceive the quality of their discharge instructions to be poor, and readmissions can be linked to improper discharge planning.^36–38 Patients with ostomies are particularly vulnerable and may need more information about stoma care and postoperative expectations.³⁹ In lower SES groups and vulnerable populations with lower health literacy and language barriers, improved patient education and protocols could make the difference.

While not every patient is a candidate for reversal, the elderly population in the United States continues to grow and one should not assume that age is a contraindication to reversal. Estimates of stoma reversal rates among the general population range from 46% to 83%, but among older patients stoma reversal rate is closer to 48% to 51%.^8,15,40 The majority of these studies were small, did not specifically evaluate older patients, and encompassed a broad range of diagnoses. Stoma reversal rates in this Medicare population were less than 30%, which was lower than expected, even though patients appeared to have few comorbidities and low frailty. Age may be a contributing factor; however, further study is needed as to why ostomy reversal rates are so low among the aging population.

This study is not without limitations. First, we restricted our older adult population to those with Medicare fee-for-service plans (specifically Parts A and B). It is possible that these results may not generalize to older adults with other insurance types, including Medicare Advantage, privately insured, and the uninsured; however, the vast majority of older adults are covered by Medicare and over 70% are covered by fee-for-service. Second, we cannot be sure that we accounted for all potential confounders and we were unable to assess reasons for lack of stomal reversal. Additionally, some of the 30-day postoperative complications were more common than anticipated, possibly related to the age of the population. However, this issue may also be owing to the ICD-9-CM and ICD-10-CM diagnosis codes that were used to identify them, especially when the codes appeared during the index hospitalization. We used the “present on admission” flags to try and differentiate between concurrent/secondary conditions caused by diverticulitis and true postoperative complications, but this approach is imperfect and potentially misclassified secondary conditions incorrectly as complications. We do believe that these misclassifications would be nondifferential across SES and are therefore expected to bias results toward the null hypothesis. The relatively small number of non-Hispanic Black patients and wide CIs make it difficult to draw definitive conclusions regarding outcomes related to race/ethnicity, although our findings are consistent with the extensive literature on racial disparities in surgery. The Medicare dataset does not include information on disease severity, such as Hinchey Classification, which might affect timeliness of ostomy reversal. Charlson index, however, should have captured those patients who were severely ill from their disease process. Finally, reversal rates are generally higher for ileostomies; thus, we cannot account for why ileostomy reversal rates were actually lower in this population, however, perhaps this finding is related to follow-up rates.

Despite these limitations, to the best of our knowledge, this study is the largest and most comprehensive analysis evaluating diverticulitis ostomy reversal rates among older adults in the United States. The study included patients from across the entire United States with the ability to complete follow-up, assess complications after surgery, and capture mortality.

In conclusion, among Medicare recipients undergoing ostomy creation for diverticulitis, patients with higher SES are more likely to undergo stoma reversal within 1 year of surgery than their lower SES counterparts and are less likely to die. Non-Hispanic Black patients had the lowest rate of reversals, and SES also appeared to have a more substantial impact on non-Hispanic Black patients, compared with non-Hispanic White. Overall, ostomy reversal rates are low among all older adults, which can have major implications on quality of life. This study adds to the growing evidence of surgical disparities in the United States, and future studies should focus on how to diminish these inequalities so that all patients receive equivalent care.