Variation in Colectomy Rates for Benign Polyp and Colorectal Cancer

Abstract

Background

Removal of pre-cancerous polyps on screening colonoscopy is a mainstay of colorectal cancer (CRC) prevention. Complex polyps may require surgical removal with colectomy, an operation with a 17% morbidity and 1.5% mortality rate. Recently, advanced endoscopic techniques have allowed some patients with complex polyps to avoid the morbidity of colectomy. However, the rate of colectomy for benign polyp in the United States is unclear, and variation in this rate across geographic regions has not been studied. We compared regional variation in colectomy rates for CRC versus benign polyp.

Methods

We performed a retrospective population-based study of Medicare beneficiaries undergoing colectomy for CRC or benign polyp, using the 100% Medicare Provider Analysis and Review files from 2010–2015. We used multivariable linear regression to obtain population-based colectomy rates for CRC and benign polyp at the Hospital Referral Region (HRR) level, adjusted for age, sex, and race.

Results

Of 280,815 patients, 157,802 (65.8%) underwent colectomy for CRC compared to 81,937 (34.2%) for benign polyp. Across HRRs, colectomy rates varied 5.8-fold for cancer (0.32 to 1.84 per 1,000 beneficiaries). However, there was a 69-fold variation for benign polyp (0.01 to 0.69). While the rate of colectomy for CRC was correlated with the rate of colectomy for benign polyp (slope=0.61, 95% CI 0.48 – 0.75), HRRs with the lowest or highest rates of colectomy for CRC did not necessarily have similarly low or high rates for benign polyp.

Conclusions

The use of colectomy for benign polyp is much more variable compared to CRC, suggesting overuse of colectomy for benign polyp in some regions. This variation may stem from provider-level differences, such as endoscopists’ referral practice or skill or surgeons’ decision to perform colectomy, or from limited access to advanced endoscopists. Interventions to increase endoscopic resection of benign polyps may spare some patients the morbidity and cost of surgery.

Introduction

Endoscopic removal of pre-cancerous polyps found on screening colonoscopy is a mainstay of colorectal cancer (CRC) prevention. However, 2–15% of polyps may be technically challenging to remove on endoscopy due to large size, sessile configuration, or a difficult-to-access location.^1–4 Treatment of benign but “endoscopically unresectable” polyps has traditionally required colectomy surgery, an operation with a 17% risk of major complications and a 1.5% risk of mortality. As such, advanced endoscopic techniques have been developed to remove complex benign polyps, such as endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD).^1,5,6 EMR uses fluid injection into the submucosal space to facilitate polyp resection without injuring the underlying muscle, while ESD involves en bloc resection of large lesions that may extend into the submucosa.^7,8 While there are no population-based data on the success of these techniques, smaller studies have suggested that up to 90% of patients referred for surgery can instead be treated endoscopically with the use of EMR/ESD.^2,9–11 The use of EMR has increased in the United States in the past 10 years, while ESD, a more difficult technique, is still not widely adopted and performed primarily at specialized referral centers.¹²

However, the dissemination and adoption of advanced endoscopic techniques may not occur quickly or consistently by geographic area in the United States, leading to overuse of more traditional management strategies.¹³ The Dartmouth Atlas and academic studies on healthcare utilization by geographic region have demonstrated significant variation in the use of new treatments and invasive procedures.^14–17 Although the prevalence of advanced endoscopic resection among patients with complex polyps remains unknown, anecdotally we have observed that patients with benign polyps are often referred for surgical intervention, suggesting a lack of awareness or availability of alternative interventions. In fact, the rate of colectomy for benign polyp has been increasing in the United States, but no population-based studies have investigated whether this rate varies geographically.¹⁸ Identifying regional overuse of surgical resection for benign disease may present a target for interventions to increase endoscopic treatment, sparing some patients the morbidity and cost associated with surgery.^19,20

In this context, we used Medicare claims to examine variation in the rates of colectomy surgery for CRC and benign polyps at the population level. We hypothesized that there would be greater variation in resection rates for benign polyp compared to CRC.

Materials and Methods

Study Design and Cohort

This was a study of national claims data from the 100% Medicare Provider Analysis and Review (MedPAR) files from 2010 to 2015. This study of deidentified data was considered exempt by the Institutional Review Board at the University of Michigan.

Medicare is the United States federal health insurance program and covers adults aged 65 and older, certain patients with disabilities, and patients with end-stage renal disease. All U.S. citizens aged 65 and older are eligible for Medicare, and certain premiums are covered for citizens who paid Social Security and Medicare taxes for at least 10 years.²¹

The Centers for Medicare & Medicaid Services (CMS) maintains the MedPAR files, which together make up an administrative database of patient-level claims submitted by medical facilities who serve Medicare beneficiaries.²²

Patients undergoing colectomy were defined using International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9) codes (45.73, 17.33, 17.32, 45.75, 45.76, 17.35, 17.36, 45.74, 17.34, 45.82, 45.83, 45.81, 48.50, 48.51, 48.52, and 48.53) and diagnosis-related group (DRG) codes (329, 330, and 331). We used all diagnosis and procedure codes associated with the claim, not just the primary codes. Included DRG codes were specific to colectomy. Patients younger than 65 or older than 99 years were excluded. We captured additional data on patient age, race, sex, and 27 Elixhauser comorbidities.²³ We then linked Medicare data to the American Hospital Association (AHA) Annual Survey data, yielding information on hospital attributes including bed size, profit status, and teaching status. We also captured information on the presence of advanced endoscopy services as a hospital characteristic, based on two AHA variables: presence of endoscopic retrograde cholangiopancreatograohy (ERCP) and presence of endoscopic ultrasound (EUS). Hospitals with either resource were considered to have advanced endoscopic services.

We determined the indication for surgery using ICD-9 codes (CRC codes 153 −154.8, 230.3, 230.4; benign polyp codes 211.3, 211.4).¹⁸ We excluded secondary malignant neoplasms from non-colonic sources, neuroendocrine neoplasms, lymphoma, or other codes to specifically identify colectomy performed for colonic adenocarcinoma. We performed two analyses: first, a population-level analysis comparing variation in regional colectomy rates between CRC and benign polyp, and second, a hospital-level analysis of the relationship between each hospital’s characteristics and its rate of colectomy for benign polyp.

Variable Definitions

The primary outcome of interest was the annual rate of colectomy per 1,000 Medicare beneficiaries for CRC and benign polyps. We used publicly available data from the Centers for Medicare and Medicaid Services to determine the number of Medicare beneficiaries assigned to each hospital referral region (HRR).^24,25 HRRs are defined geographic areas used to study regional healthcare utilization. Each of the 306 HRRs contains at least one city where major cardiovascular surgery and neurosurgery operations are performed, reflecting regional referral areas for tertiary healthcare.²⁶ A secondary outcome was at the hospital level—each hospital’s rate of colectomy for benign polyp (defined as the ratio between a hospital’s number of colectomies for benign polyp and the total colectomies performed for both indications).

Statistical Analysis

The purpose of this analysis was to estimate population-based colectomy rates for benign polyp and colon cancer diagnoses across HRR’s. We adjusted raw colectomy rates per 1,000 Medicare beneficiaries for age, sex, and race using multivariable linear regression as has been previously described.^24,25 We then normalized these adjusted estimates to the mean rate for each surgical indication. To do this, we divided the adjusted estimates by the mean, which eliminates skew from small numbers of high or low outliers.²⁶ We performed linear regression to determine the association between colectomy rates for polyp and cancer, using cancer rates as the independent variable. To obtain the trend in rate of colectomy for benign polyp over time, we used linear regression to determine the association between colectomy rate and year.

For the hospital-level analysis examining each hospital’s rate of colectomy for benign polyp, we used linear regression to evaluate the association between colectomy rates for benign polyps and structural characteristics of the hospital. For this analysis, we report regression coefficients (marginal effects) related to the presence of each hospital characteristic or resource.

All analyses were performed using StataSE version 15 (College Station, Texas).

Results

Characteristics of the Cohort

Our cohort included 280,815 Medicare beneficiaries who underwent colectomy for CRC or benign polyp. Overall, 157,802 (65.8%) patients underwent colectomy for CRC compared to 81,937 (34.2%) for benign polyp. Characteristics of the cohort are summarized in Table 1.

Table 1.

Characteristics of the Cohort.

	All	CRC	Benign Polyp	p-value
Patient
Procedures, No. (%)	239,739 (100)	157,802 (65.8)	81,937 (34.2)	-
Age, y, mean (SD)	75 (8)	75 (6)	72 (8)	<0.01
Male sex, No. (%)	114,597 (47.8)	73,969 (46.9)	40,628 (49.6)	<0.01
White race, No. (%)	204,567 (85.8)	134,985 (86.0)	69,582 (85.4)	0.24
Black race, No. (%)	23,470 (9.8)	14,476 (9.2)	8,994 (11.0)	<0.01
Comorbidities, mean (SD)	2.5 (1.8)	2.7 (1.8)	2.2 (1.7)	<0.01
Hospital
Number of beds, No. (%)
<200	66,191 (27.7)	44,214 (28.2)	21,977 (27.0)	0.04
200–349	65,927 (27.6)	43,611 (27.8)	22,316 (27.4)	0.53
350–499	47,238 (19.8)	30,846 (19.6)	16,392 (20.1)	0.02
>=500	59,259 (24.8)	38,382 (24.4)	20,877 (25.6)	0.22
Profit status, No. (%)
Not for profit	186,870 (78.3)	122,959 (78.3)	63,911 (78.4)	0.18
For profit	29,301 (12.3)	18,996 (12.1)	10,305 (12.6)	0.03
Other	22,444 (9.4)	15,098 (9.6)	7,346 (9.0)	0.01
Teaching hospitals, No. (%)	44,944 (18.5)	29,708 (18.9)	14,336 (17.6)	<0.01
Rural hospital, No. (%)	5,656 (2.4)	3,980 (2.5)	1,676 (2.1)	<0.01

Variation in Colectomy for Cancer and for Adenoma

The average national rate of colectomy overall was 0.86 cases per 1,000 beneficiaries, ranging from 0.41 – 2.4 across HRRs. Colectomy rates across HRRs varied 5.8-fold for cancer (range: 0.32 – 1.84 cases per 1,000 beneficiaries), compared to a 69-fold variation for benign polyp (range: 0.01 – 0.69 cases per 1,000 beneficiaries). The variation profiles in Figure 1 demonstrate the higher relative variation in colectomy for benign polyp across HRRs.

Figure 1. Variation Profile of Colectomy by Diagnosis.

. Each point represents one of 306 HRRs, with colectomy rate standardized to the national average (by diagnosis). Rates were adjusted for age, sex, and race. The relative variation (ratio of highest to lowest HRR) of colectomy for benign polyp was 69, compared to that of colectomy for CRC (5.8).

The rate of colectomy for CRC was positively correlated with rate of colectomy for benign polyp at the HRR level (slope=0.61, 95% CI 0.48 – 0.75) (Figure 2). However, HRRs with the highest rates of colectomy for CRC did not necessarily have the highest rates for benign polyp. This discordance is demonstrated in maps showing HRR colectomy rates by diagnosis (Figure 3). From 2010 to 2015, there was a slight but statistically significant trend towards lower rates of colectomy for benign polyp (slope= −0.004, p<0.01) and colon cancer (slope=−0.005, p<0.01).

Figure 2. Correlation Between Rates of Colectomy for CRC and Benign Polyp.

Mean-standardized rates of performing colectomy for CRC were positively correlated with mean-standardized rates of colectomy for benign polyp.

Figure 3. HRR rates of Colectomy by Diagnosis.

The regions with highest rates of colectomy for CRC do not necessarily have the highest rates of colectomy for benign polyp.

Hospital Characteristics and Rate of Colectomy for Benign Polyp

Certain hospital characteristics were associated with the rate of colectomy for benign polyp performed at that hospital (i.e., the proportion of all colectomies that were performed for benign polyp). Effect sizes of each characteristic after multivariable adjustment are shown in Table 2. Each effect size represents an absolute percentage change in the proportion of colectomy performed for benign polyp at a given hospital. Increasing hospital bed size was associated with a small increase in colectomy for benign polyp, with the largest hospitals predicted to have a 2% increase in their proportion of colectomy for benign polyp (95% CI 1.9 – 2.1). By contrast, rural hospitals were predicted to have a 4.7% decrease in this proportion (95% CI −4.9 - −4.4). Hospitals that had advanced endoscopy services had a 1.1% increased proportion of colectomy for benign polyp (95% CI 1.0 – 1.3).

Table 2.

Association of Hospital Characteristics and Proportion of Colectomy for Benign Polyp.

Resource	Mean Effect (95% CI)*
Hospital bed size
<200	Ref
200–349	0.5 (0.4 to 0.7)
350–499	1.5 (1.4 to 1.6)
>=500	2.0 (1.9 to 2.1)
Teaching status	−1.9 (−2.0 to −1.8)
Business model
Not for profit	Ref
For profit	−1.0 (−1.2 to −0.8)
Other	−2.3 (−2.5 to −2.1)
Location
Urban	Ref
Rural	−4.7 (−4.9 to −4.4)
Advanced endoscopy services
Absent	Ref
Present	1.1 (1.0 to 1.3)

^*Results derived from linear regression models where hospitals’ proportion of benign polyp cases is outcome. Positive effect suggests an increase in proportion of colectomy surgery for benign polyps. Units of the effect size are in percentage points for the proportion of colectomy for benign polyp at that hospital.

Discussion

In this study, we found that the rate of colectomy for benign polyps varied far more widely than the rate of colectomy for CRC across geographic regions. Although rates of colectomy for both diagnoses were correlated, regions with the highest rate of colectomy for CRC did not necessarily have the highest rate of colectomy for benign polyps. While we do not have information on the use of specific advanced endoscopic techniques in this study, the extensive variation in use of colectomy for benign polyp suggests there may be underuse of non-surgical treatment in some regions.

The present study adds to the literature on regional variation in the United States healthcare system.²⁷When a medical practice or procedure varies substantially by geographic area, the implication is that there is underuse and/or overuse. In 1998, Birkmeyer and colleagues showed that rates of common surgical procedures vary the most for procedures considered “discretionary,” such as back surgery and prostatectomy (compared to hip fracture repair and colectomy for cancer).²⁶ The mechanisms responsible for variation include the supply of healthcare resources, such as specialists and hospital beds, as well as local providers’ beliefs and training.²⁸ Another mechanism responsible for variation in healthcare services is the uneven adoption of new technologies into practice.²⁹ Often there is concern for overuse of expensive new technologies, contributing to ballooning healthcare expenditures. However, our study highlights the potential downside of delayed adoption of new technology—that is, persistence of a more invasive, morbid, and costly treatment.

The regional variation we have shown in colectomy for benign polyps may stem from hospital- and provider-level differences. We found that urban hospitals and larger hospitals did perform a larger proportion of colectomies for benign polyp. Surprisingly, we found that hospitals that offered advanced endoscopy techniques also had larger proportions of colectomies for benign polyp, although this increase was modest. This may be related to how we defined advanced endoscopy services, which were identified by the availability of ERCP and EUS (not EMR or ESD). Nevertheless, these associations may be related to factors such as a higher level of care at larger tertiary referral centers. Although we did not investigate provider-level differences in this study, factors such as endoscopic skill or referral practice may also drive regional variation. For example, a population-based study in France found wide variation in surgical referral rate among endoscopists.³⁰ A less experienced endoscopist may have a lower threshold for surgical referral than a more experienced one. Surgeon practice may also differ, with some surgeons referring patients with benign polyps to a second endoscopist before proceeding with surgery. These patterns may be influenced by a surgeon’s practice setting—surgeons within a tertiary referral center may have professional relationships with advanced endoscopists, facilitating patient access to non-surgical treatment.

There is a growing body of evidence exploring the use of advanced endoscopic techniques or surgery for complex benign colorectal polyps. While there is no widely-accepted definition for a complex polyp, polyps that are greater than 33% of the circumference of the colon, occupy two or more haustral folds, or are located at the appendiceal orifice are considered difficult to resect.³¹ Patients referred for surgical resection for benign-appearing polyps tend to have polyps larger than 30 millimeters, sessile or flat polyps, or polyps located at the thinner-walled right colon, hard-to-reach cecum, or the ileocecal valve.^4,32–34 In several series, the minority—8–18%--of these benign-appearing polyps are ultimately found to harbor invasive CRC after colectomy.^33–37 Performing a second colonoscopy before surgical resection can eliminate the need for colectomy for over 90% of patients referred to a surgeon for benign polyp.^{4,11,32,38–42} Compared to colectomy, endoscopic resection has lower complication rates and is cost-effective.^20,40 As such, guidelines from the American Society for Gastrointestinal Endoscopy (ASGE)/American College of Gastroenterology (ACG) encourage referral to an advanced endoscopist for benign-appearing polyp prior to surgical referral, and recommend that polyps deemed “endoscopically unresectable” be photographed and reviewed to ensure appropriateness of surgical referral.^43,44 Finally, there are emerging hybrid techniques combining laparoscopy and endoscopy, where laparoscopic stabilization or traction can aid endoscopic resection of a difficult polyp and suture repair can be performed if a perforation occurs.^45,46

The overall volume and geographic distribution of endoscopic resection for complex benign polyp has not been reported in the United States. Unfortunately, our study—which uses a database of surgical patients—does not allow for this knowledge gap to be filled. Existing studies of advanced endoscopic techniques mainly include single- and multi-institutional cohort studies of patients undergoing endoscopic polyp resection, and many describe the effectiveness and safety of endoscopic resection compared to colectomy.^7,47–50 Further investigation in the United States, including clinician survey, prospective data collection, or use of administrative databases, may be helpful in characterizing the spread of these endoscopic techniques. Importantly, while endoscopic resection is likely the most appropriate first-line treatment for complex benign polyps,^35,43,44,51 a clearer understanding of current advanced endoscopic practice is needed before interventions can be designed and implemented to increase access to and adoption of advanced techniques.

This study has several limitations. As mentioned above, this study utilizes a Medicare database limited to surgical patients, and thus the denominator of patients treated non-surgically is unknowable with our data. Therefore, we are unable to measure use of advanced endoscopic techniques in the regions with high or low use of colectomy for benign polyp. In addition, we were unable to establish a convincing link between the presence of “advanced endoscopic services” and decreases in hospital benign colectomy rates. It may be that availability of ERCP and EUS does not adequately reflect availability of techniques used for complex polypectomy. There is currently no formal training for EMR or ESD during advanced endoscopy fellowship, and gastroenterologists who perform ERCP and EUS may not necessarily routinely perform advanced resection.^51,52 Finally, we had limited ability to adjust for patient-level factors, including the presence of comorbid conditions that may have influenced clinician decision-making. For example, in a patient with high anesthetic risk, a clinician might have a higher threshold for surgical referral and attempt multiple colonoscopies under lighter sedation.

Conclusions

In conclusion, the use of colectomy for benign polyp is highly variable across geographic areas. This variation could be influenced by underuse of non-surgical treatments, such as advanced endoscopic resection. Further investigation on advanced endoscopy access, adoption, and barriers is warranted. Interventions to increase endoscopic resection of benign polyps may spare some patients the morbidity and cost associated with surgery.