The effect of changes in Medicare reimbursement on the practice of office and hospital-based endoscopic surgery for bladder cancer

Abstract

BACKGROUND:

Procedures performed in the office offer potential cost savings. Recent analyses suggest, however, that a fee-for-service system may incentivize subscale operations and, thus, contribute to excessive spending. The authors of this report sought to characterize changes in the practice of office-based and hospital-based endoscopic bladder surgery after 2005 increases in Medicare reimbursement.

METHODS:

All office and hospital-based endoscopic surgeries that were performed in a faculty practice from 2002 through 2007 were identified using billing codes for procedures, diagnoses, and procedure locations and then analyzed using the chi-square test and logistic regression. Costs were estimated based on published Medicare reimbursements for office and hospital-based surgeries.

RESULTS:

In total, 1341 endoscopic bladder surgeries were performed, including 764 in the office and 577 in the hospital. After 2005, the odds ratio (OR) for office surgery occurring among all cystoscopies and for surgery occurring in the office versus the hospital was 2.01 (95% confidence interval [CI], 1.71-2.37) and 2.29 (95% CI, 1.83-2.87), respectively. Among all treated lesions that were associated with a diagnosis of bladder cancer and nonbladder cancer, the OR for a procedure occurring in the office versus the hospital was 1.36 (95% CI, 1.07-1.73) and 1.99 (95% CI, 1.52-2.60), respectively. The likelihood of repeat surgery on the same lesion increased after 2005 (OR, 2.86; 95% CI, 1.46-5.62), and the likelihood of an office surgery leading to a bladder cancer diagnosis at the next visit declined (OR, 0.29; 95% CI, 0.16-0.51). The overall estimated expenditure increased by 50%.

CONCLUSIONS:

After 2005, more bladder lesions were identified and treated in the office. In a single group practice, office treatment of bladder cancer did not fully explain this new practice pattern, suggesting a lowered threshold for office intervention. Cancer 2010. © 2010 American Cancer Society.

The management of bladder lesions places a significant economic burden on the United States healthcare system. Analyzing the costs associated with life-long surveillance and treatment, 1 meta-analysis concluded that the cost per patient with bladder cancer from diagnosis to death is the highest of all cancers, ranging from $96,000 to $187,000.1 Despite this high cost, there have been only a few studies addressing the economics of bladder cancer care. Most predict that the overall cost of managing bladder cancer will continue to rise in concert with the aging population, the diffusion of surgical technology, and the development of novel, costly biomarkers.2, 3 Consequently, there are growing pressures to contain costs and improve clinical efficiency.

The use of office-based endoscopic procedures, such as transurethral resection (TUR), biopsy, or fulguration, instead of hospital-based endoscopic intervention offers potential cost savings to the healthcare system,4 provided that the office-based intervention is performed for the same indications, is equally efficacious, and is tolerable under minimal anesthesia. Recognizing this possibility, in 2005, the Centers for Medicare and Medicaid (CMS) increased physician reimbursement for office-based endoscopic procedures. We hypothesized that these changes altered physician incentives, leading to the increased use of office-based endoscopic surgery, a decline in hospital-based endoscopic surgery, and, consequently, a reduction in healthcare-related costs. To evaluate this hypothesis, we assessed treatment patterns in our faculty practice before and after CMS changes in physician reimbursement.

MATERIALS AND METHODS

From the billing database of a 10-physician faculty practice, we identified all office-based endoscopic surgeries (OBP) and all hospital-based endoscopic surgeries (HBP) performed between the years 2002 and 2007. By using Current Procedural Terminology (CPT) codes, we retrospectively compiled the number of procedures performed in the office and in the hospital for bladder biopsy (code 52204) and fulguration (code 52224) and for small (code 52234), medium (code 52235), and large (code 52240) tumor resections. Diagnosis codes associated with each procedure visit and subsequent office visits were available for evaluation. We then compared practice patterns from 2002 through 2004 (Period 1) with those from 2005 through 2007 (Period 2), after CMS alterations in physician reimbursement. To account for interim office growth from increased referrals or increased staffing, we used all office-based diagnostic cystoscopies (code 52000) as the “population at risk” in our analyses.

In comparing the number of office-based surgeries between periods, we adjusted for practice growth by calculating the fraction of office-based surgeries that occurred among all office cystoscopic procedures (OBP/diagnostic cystoscopies plus OBP). The chi-square test and logistic regression were used to compare the likelihood of office-based intervention with the likelihood of hospital-based intervention among all endoscopic surgeries and the likelihood of office-based intervention among all office cystoscopies. We performed secondary analyses to assess the impact of diagnosis code on the place of treatment, the likelihood of OBP among nonbladder cancer diagnoses, the likelihood of a change in diagnosis code at the office visit after an office intervention, and the likelihood of “redundant” office surgery followed by hospital surgery on the same bladder lesion. Statistically significant results had P values <.05, and all tests were 2 tailed.

A constant dollar healthcare cost analysis was also performed based on an estimate of total Medicare expenses incurred both from procedures performed in the office and from procedures performed in the hospital. The cost per office-based intervention was estimated based on the Medicare in-network reimbursement rate by CPT code in a given year. Between 2004 and 2005, the reimbursement for an office-based biopsy (code 52204) increased from $290.57 to $834.66, and reimbursement for office-based fulguration (code 52224) increased from $199.30 to $1969.86 (Table 1). For hospital-based surgery, based on previous experience, we assumed that 10% of operations resulted in an inpatient stay, whereas 90% occurred in the outpatient setting. The cost for hospital-based surgery was estimated based on Medicare payments reported in the CMS Healthcare Consumer Initiative and by using Diagnosis-Related Groups for the 10% of surgeries that resulted in inpatient stays.5 All calculated dollar amounts were adjusted for inflation to 2005 dollars. For hospital-based surgery, we estimated that Medicare payments totaled $4038 per operation in New York City during the period under study, and a fraction of this amount reflected physician reimbursement. Finally, to adjust for practice growth, we calculated the relative cost incurred per office cystoscopy performed (absolute cost/diagnostic cystoscopies plus OBP) in each time period.

Table 1. Centers for Medicare and Medicaid Reimbursement Changes for Office-Based Endoscopic Bladder Procedures

CPT Code	2004	2005	Factor Change
CPT indicates Current Procedural Terminology.
52204, Biopsy	$290.57	$834.66	×2.87
52224, Fulguration	$199.30	$1969.86	×9.88
52234, Small tumor	$290.48	$291.02	×1.00
52235, Medium tumor	$341.14	$341.52	×1.00
52240, Large tumor	$604.54	$601.97	×1.00

RESULTS

From 2002 to 2007, in total, 1341 TUR procedures were performed, including 764 in the office and 577 in the hospital. Comparing the 2 periods surrounding the changes in Medicare reimbursement, office-based surgery increased from 227 (Period 1) to 537 (Period 2), whereas the number of hospital-based surgeries increased only slightly from 284 (Period 1) to 293 (Period 2) (Fig. 1). Stratified by CPT code, we observed a decline in the use of office biopsy (code 52204) with an increase in the use of all other coded office procedures (codes 52224, 52234, 52235, and 52240) (Table 2). OBP was associated with a bladder cancer diagnosis code in 66.5% (357 of 537) of office procedures that were performed during Period 2 compared with 63.4% (144 of 227) of office procedures that were performed during Period 1. Likewise, a nonbladder cancer diagnosis code was coded in 33.5% (180 of 537) of office procedures performed during Period 2 compared with 36.6% (83 of 227) of office procedures performed during Period 1 (Table 3). The total number of diagnostic cystoscopies performed in the office increased marginally (2278 in Period 1 vs 2408 in Period 2) (Fig. 1). Consequently, after adjusting for practice growth, there remained a 201% increase in office-based surgery compared with a 12% decline in hospital-based intervention (Fig. 2).

Figure 1. The numbers of office-based endoscopic surgeries (OBP) and all hospital-based endoscopic surgeries (HBP) and the total numbers of office cystoscopies are shown relative to the date of service.

Figure 2. The probability of office-based endoscopic surgeries (OBP) and hospital-based endoscopic surgeries (HBP) when undergoing office cystoscopy is shown.

Table 2. Summary of the Number of Procedures and Estimated Medicare Expenditure

CPT Code	No. of Procedures
	Office		Hospital		Medicare Expenditure, 2005 Dollars
	2002-2004	2005-2007	2002-2004	2005-2007	2002-2004	2005-2007
CPT indicates Current Procedural Terminology.
52204, Biopsy	56	32	29	16	$133,373.92	$91,317.12
52224, Fulguration	162	449	72	57	$323,022.60	$1,114,633.14
52234, Small tumor	7	16	58	60	$236,237.36	$246,936.32
52235, Medium tumor	2	39	72	97	$291,418.28	$405,005.28
52240, Large tumor	0	1	53	63	$214,014.00	$254,995.97
Total	227	537	284	293	$1,198,066.16	$2,112,887.83

Table 3. Office-Based Endoscopic Procedures Stratified by Treatment Session Diagnosis Code

Treatment Session Diagnosis	No. of Patients		Likelihood of OBP if Undergoing Cystoscopy Stratified by Preprocedural Diagnosis Code: Period 2 vs Period 1
	2002-2004	2005-2007	OR (95% CI)	P
OBP indicates office-based endoscopic procedure; OR, odds ratio; CI, confidence interval; UTI, urinary tract infection; TCC, transitional cell carcinoma. a Statistically significant.
Hematuria	26	40	1.44 (0.87-2.39)	.155
UTI	4	5	0.96 (0.25-3.64)	.954
Voiding dysfunction	6	10	1.34 (0.48-3.73)	.574
Neoplasm of unspecified nature	22	40	2.90 (1.59-5.31)	.0004a
Upper tract TCC	1	31	8.39 (1.10-63.73)	.0049a
Other	24	54	2.33 (1.42-3.81)	.0005a
Nonbladder cancer	83	180	1.99 (1.52-2.60)	<.00001a
Bladder cancer	144	357	1.71 (1.38-2.14)	<.00001a
Overall	227	537	2.01 (1.71-2.37)	<.00001a

The odds ratio (OR) of having an office-based surgery when undergoing any office cystoscopy in the latter period was 2.01 (95% confidence interval [CI], 1.71-2.37; P < .001). In addition, the OR of surgery occurring in the office versus the hospital among all endoscopic bladder surgeries was 2.29 (95% CI, 1.83-2.87; P < .001) in the latter period. When stratified by procedure type, similar changes were observed in the use of bladder fulguration (code 52224) and TUR for medium-sized tumors (code 52235). However, the likelihood of undergoing a bladder biopsy (52204) actually declined (OR, 0.49; 95% CI, 0.31-0.75; P = .0012) (Table 4).

Table 4. The Likelihood of Undergoing Office-Based Endoscopic Procedures During Period 2 (2005-2007) Versus Period 1 (2002-2004)

CPT Code	Likelihood of OBP if Undergoing Cystoscopy: Period 2 vs Period 1		Likelihood of OBP if Undergoing Endoscopic Bladder Surgery, Any Location: Period 2 vs Period 1
	OR (95% CI)	P	OR (95% CI)	P
OBP indicates office-based endoscopic procedure; CPT, Current Procedural Terminology; OR, odds ratio; CI, confidence interval; NA, not applicable. a Statistically significant.
52204, Biopsy	0.49 (0.31-0.75)	.0012a	1.04 (0.49-2.18)	.9
52224, Fulguration	2.36 (1.95-2.84)	<.00001a	3.50 (2.4-5.18)	<.0001a
52234, Small tumor	1.94 (0.80-4.73)	.1431	2.21 (0.85-5.76)	.095
52235, Medium tumor	16.59 (4.00-68.76)	<.001a	14.47 (3.38-61.91)	.0003a
52240, Large tumor	NA		1.86	.27
Overall	2.01 (1.71-2.37)	<.00001a	2.29 (1.83-2.87)	<.0001a

On secondary analysis, redundant surgery, which we defined as OBP followed by HBP within 60 days, accounted for 4.8% (11 of 227) of OBPs in Period 1 and 6.9% (37 of 537) of OBPs in Period 2. The OR for redundant surgery in the latter period among all patients who underwent office cystoscopy (OBP plus diagnostic cystoscopy) was 2.86 (95% CI, 1.46-5.62; P = .001). Among all (OBP plus HBP) treated bladder cancers (International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM] 188.x or 233.7), the OR for an office procedure versus a hospital procedure in the latter period was 1.36 (95% CI, 1.07-1.73; P = .0107). Among all office cystoscopies for a nonbladder cancer diagnosis (code 52000 and OBP), the OR for OBP in Period 2 compared with Period 1 was 1.99 (95% CI, 1.52-2.60; P < .001). Finally, among all OBPs performed with a nonbladder cancer diagnosis, the OR for a procedure leading to a bladder cancer diagnosis (ICD-9-CM 188 or 233.7) at the next visit was 0.29 (95% CI, 0.16-0.51; P < .001), favoring a decline in the likelihood that a biopsy in the latter period would lead to a diagnosis of cancer.

Cost analysis demonstrated a 1.76-fold increase in total CMS expenditure to the practice in the latter period (Table 2), reflecting an absolute cost increase for OBP by a factor of 18.13 and for HBP by a factor of 1.03. The cost incurred per office cystoscopy performed (diagnostic cystoscopies and OBP) was $478.27 for Period 1 and $717.45 for Period 2, reflecting an overall adjusted expenditure increase of 50% after the CMS increase in physician reimbursement.

DISCUSSION

Healthcare spending in the United States has grown to exceed the amount spent on food, representing 16% of the national gross domestic product (GDP), or an estimated 1.9 trillion dollars annually. On the basis of an assessment of national wealth and the average healthcare expenditure of 13 other industrialized nations, the McKinsey Global Institute (MGI) estimates that excessive spending on healthcare in the United States totals $478 billion annually. Thirty-six percent of this spending excess is attributable to inefficiencies in outpatient care. According to the MGI, the practice shift from the hospital to the outpatient setting has not resulted in overall cost savings, because providers are incentivized to treat more patients as a result of the fee-for-service reimbursement system. The desire to capture as many patients as possible has lead to a proliferation of outpatient treatment facilities, resulting in excess capacity that, in turn, has increased overall operational expenses. These increased operational expenses reinforce the incentive for providers to expand their patient base, resulting in a “vicious circle.”6 Other investigators have echoed similar concerns about physician incentives in a fee-for-service payment system.7-9 The Congressional Budget Office estimates that healthcare costs will rise to 25% of the national GDP by 2025.10 To stem this impending crisis, the economic stimulus bill approved by Congress in February 2009 allocated $1.1 billion to study and promote value-conscious treatment of specific conditions.

Bladder cancer is the most expensive cancer to treat from diagnosis until death,1 and its management is prone to the same financial incentives and practice-based inefficiencies described in the MGI report. We have demonstrated that the cost associated with office-based endoscopic surgery for bladder cancer has risen in concert with US healthcare spending without a clear clinical evidence base to support such a transition.

Adjusting for the growth of our faculty, we observed 1) a dramatic increase in the number of office-based endoscopic procedures for bladder lesions after the increase in Medicare reimbursement and 2) a modest decline in the number of procedures performed in the operating room during this period. An office intervention, such as a biopsy, fulguration, or TUR, was twice as likely to occur in Period 2 whenever cystoscopy was performed. In addition, when an intervention occurred in the latter time period, it was more than twice as likely to take place in the office as in the hospital. This practice trend applied not only to patients with a known history of bladder cancer but also to all patients with an indication for office cystoscopy. Office use increased for nearly all coded procedures. The use of fulguration (code 52224), the most frequently billed procedure, increased nearly 3-fold during the study period. Because the number of hospital-based surgeries did not decline significantly as office use increased, we observed a 50% increase in overall Medicare expenditure after adjusting for practice growth (Fig. 3).

Figure 3. Trends in office-based endoscopic surgeries (OBP) and all hospital-based endoscopic surgeries (HBP) and total Medicare expenditures are shown for 2002 through 2007.

We believe that there are 3 potential explanations for this observed change in practice related to financially induced incentives: 1) an increase in bladder cancer referrals to the faculty practice, 2) redundant use of office surgery at the time of diagnosis for patients who ultimately undergo hospital-based surgery for the same lesion, and 3) a reduced threshold for office intervention. Each of these possible explanations has distinct cost implications for the healthcare system.

First, the increase in office-based procedures may reflect an increase in bladder cancer referrals to the practice in the latter period. If these additional patients with bladder cancer were treated definitively in the office rather than in the hospital, then the observed increase in Medicare expenditure might represent an overall cost savings, because treating those same patients in the operating room would have been more costly. To assess the significance of this practice, we performed a secondary analysis of our data to determine whether increased bladder cancer referrals might account for the observed increase in office-based procedures. Among all treated bladder lesions (OBP plus HBP) that were associated with a bladder cancer diagnosis (ICD-9-CM 188.x or 233.7), the OR for a procedure occurring in the office versus the hospital in Period 2 compared with Period 1 was 1.36 (95% CI, 1.07-1.73); whereas, for all treated bladder lesions regardless of diagnosis code, the OR for a procedure occurring in the office versus the hospital comparing the 2 time periods was 2.29 (95% CI, 1.83-2.87). The discrepancy between these 2 ORs, both favoring office over hospital intervention, suggests that an increase in bladder cancer referrals alone is not sufficient to account for the entire observed change in the practice of office-based endoscopic surgery.

Second, our findings may reflect the redundant use of office surgery on patients who ultimately undergo hospital-based surgery for the same condition. Such superfluous use of healthcare resources would lead to an overall cost increase. In our practice, redundant surgery, defined as OBP followed by HBP within 60 days, occurred infrequently and accounted for only 4.8% of OBP in Period 1 and 6.9% of OBP in Period 2. The OR for redundant surgery in Period 2 compared with Period 1 among all patients who underwent office cystoscopy (OBP and diagnostic cystoscopy) was 2.86 (95% CI, 1.46-5.62). Although this represents a significant increase in the likelihood of redundant surgery in Period 2, the 26 additional redundant procedures that were performed in Period 2 accounted for less than $50,000 of the observed Medicare cost increase. Therefore, the increase in redundant use of HBP after OBP did not completely account for the great increase in overall Medicare expenditure; moreover, some of these repeat procedures may have been clinically indicated and unavoidable.

Finally, the additional office-based surgeries performed in Period 2 may be indicative of a reduced threshold for office intervention driven by indications that either were not appreciated or were nonexistent in Period 1. For example, superior technology and improved operator comfort in Period 2 may have allowed for the management of suspicious bladder lesions that went unrecognized or that were chosen for observation in Period 1. Alternatively, by incentivizing office surgery, CMS may have influenced clinical practice by lowering the threshold to perform surgery in Period 2 on lesions that would have warranted only observation in Period 1. If the latter were true, then the increased cost would reflect an increase in the number of interventions performed with unproven benefit. This possibility is concerning not only from an economic standpoint but also because it reveals that, despite the best intentions, clinical decision-making may be influenced by factors other than a perceived benefit or an evidence-based benefit to the patient.

Our analysis suggests that, generally, there was a lower threshold for performing office-based surgical procedures after changes in reimbursement. Among all office cystoscopies that were performed for a nonbladder cancer diagnosis, the OR for OBP was 1.99 (95% CI, 1.52-2.60) for Period 2 compared with Period 1. On the basis of changes that occurred in diagnosis codes from visit to visit, we determined that the OR for OBP leading to a new bladder cancer diagnosis in the latter period was 0.29 (95% CI, 0.16-0.51), indicating that the clinical yield of OBP performed in Period 2 was far inferior to the yield in Period 1. In other words, for a patient with a nonbladder cancer indication for cystoscopy in Period 2, the likelihood that a lesion would be biopsied increased, and the likelihood that that biopsy would change clinical management by leading to a bladder cancer diagnosis declined.

In creating an incentive to shift the management of bladder cancer from the hospital to the office, the goal of CMS was to contain costs while providing comparable efficacy. On the basis of our findings, it is unlikely that these standards have been met in all cases. Patients with bladder cancer were more likely to be treated in the office during Period 2; therefore, increasing reimbursement for office-based procedures did provide some cost savings to the healthcare system. However, the likelihood of incomplete office resection necessitating a second hospital-based surgery on the same lesion increased in Period 2. In addition, whereas office surgery on patients with a nonbladder cancer diagnosis was more likely in Period 2, the surgical outcome was less likely to lead to a change in clinical management, suggesting that some office-based endoscopic procedures may be unnecessary and may lead to inflation of healthcare costs.

We recognize that there are unmeasured variables, other than financially driven incentives, which could have influenced the shift of surgery from the hospital to the office. Such variables include decreasing operating room availability and emerging recognition of publications supporting the use of OBP. Without commensurate growth in hospital-based resources, surgeons experiencing practice growth may have been forced by limited operating room availability to perform bladder surgery in the office. However, from 2002 to 2007, the use of urology operative block time declined in our department, with surgeons filling 66.77% of the allotted time in 2002 and 59.56% in 2007. Furthermore, the time from preoperative visit to operative date actually declined in the latter period (mean, 6 days; median, 3 days), suggesting that our hospital did not reach maximum operating room capacity and, thus, that the growth of our hospital-based practice could have been accommodated.

In addition, several influential publications supporting the use of office fulguration in the setting of recurrent, small, superficial bladder tumors4, 11-14 became available before 2005. The majority of patients in those series had a previous diagnosis of superficial bladder cancer, and those who were chosen for office fulguration had comparable survival, progression, and risk of recurrence compared with those who underwent formal hospital-based TUR.4 In the current study, financial incentives incompletely explain all of the findings, because we did observe an increase in medium-sized tumor resections (code 52235) in the office during Period 2 without a change in reimbursement for this procedure. One potential reason for this increase could be the influence of the published literature, perhaps allowing frail patients at high risk for anesthesia complications to be treated safely in the office. Although this phenomenon may have some influence over the increased use of office procedures, evidence-based practice does not explain the lower threshold for performing a biopsy on an undiagnosed lesion, nor does it account for the decline in the likelihood that an office-based procedure will lead to a bladder cancer diagnosis.

There are several limitations to our study. First, because we used a billing database to identify procedures that were performed in a given period, we lack certain clinical information, including pathology. Consequently, our findings demonstrate a dramatic change in practice but cannot address the clinical efficacy of office verses hospital endoscopic surgery. Similarly, although we demonstrated a 50% increase in Medicare expenditure during the later period that was not explained sufficiently by an increase in bladder cancer referrals alone, the degree to which cancer was treated definitively in the office—and, thus, saved a more expensive visit to the operating room—is not known for individual patients in our dataset. In addition, because the ultimate decision to perform a biopsy in the office setting rests with the individual practitioner, the specific indications for intervention are not always obvious in a retrospective analysis. Finally, this is a single-center experience that may not be reflective of more global practice trends. Future research in nationally representative datasets is needed to confirm that our findings reflect national practice trends. Our study also has several important benefits, including a dataset of complete billing documents and the quality control inherent in the faculty practice of a major tertiary care medical center.

Finally, it is worth noting that, despite the cost implications of the current study, there may be additional benefits to widespread adoption of office endoscopic bladder surgery. Schrag et al15 have demonstrated that distinct patient and physician characteristics influence adherence to clinical guidelines in the Medicare population with bladder cancer. Perhaps the convenience of office-based surgery ultimately will influence the long-term management of patients with bladder cancer, encouraging stricter adherence to those guidelines and surveillance protocols. Much controversy remains, however, regarding whether the intensity of surveillance and treatment of superficial bladder cancer ultimately has an impact on survival.16

In conclusion, the current study demonstrated that the use of office-based endoscopic bladder surgery in a single-institution faculty practice increased subsequent to an increase in Medicare reimbursement. In evaluating the potential causes for this practice shift, we observed that it may be caused in part by the transfer of bladder cancer surgery from the operating room to the office, in part because of an increase in redundant surgical procedures, and in part because of an overall reduced threshold for performing office-based surgery on patients who have an indication for diagnostic cystoscopy. We believe these trends are disturbing, because they may reflect both diagnostic and therapeutic overuse of office-based endoscopic bladder surgery. Further evaluation of this trend in nationally representative datasets is needed to determine the extent and the efficacy of this practice and to provide data useful for the development of clinical guidelines and policy measures for office-based endoscopic bladder surgery.

CONFLICT OF INTEREST DISCLOSURES

Dr. Makarov was supported by the Veterans Health Administration and The Robert Wood Johnson Foundation. Dr. Taneja has acted as a consultant for Envisioneering Medical; as a consultant and scientific study investigator for Gtx; as a scientific study investigator for GlaxoSmithKline; as a member of the advisory board for US HIFU; and as a member of the advisory board for Pfizer.