Access to high-volume surgeons and the opportunity cost of performing radical prostatectomy by low-volume providers

Abstract

Background:

Evidence suggests that redirecting surgeries to high-volume providers may be associated with better outcomes and significant societal savings. Whether such referrals are feasible remains unanswered.

Methods:

Medicare Provider Utilization and Payment Data, SEER 18, and US Incidence data were used to determine the geographic distribution and radical prostatectomy volume for providers. Access was defined as availability of a high-volume provider within driving distance of 100 miles. The opportunity cost was defined as the value of benefits achievable by performing the surgery by a high-volume provider that was forgone by not making a referral. The savings per referral were derived from a published Markov model for radical prostatectomy.

Results:

A total of 14% of providers performed >27% of the radical prostatectomies with >30 cases per year and were designated high-volume providers. Providers with below-median volume (≤16 prostatectomies per year) performed >32% of radical prostatectomies. At least 47% of these were within a 100-mile driving distance (median = 22 miles), and therefore had access to a high-volume provider (>30 prostatectomies per year). This translated into a discounted savings of more than $24 million per year, representing the opportunity cost of not making a referral. The average volume for high- and low-volume providers was 55 and 13, respectively, resulting in an annual experience gap of 43 and a cumulative gap of 125 surgeries over 3 years. In 2014, the number of surgeons performing radical prostatectomy decreased by 5% while the number of high- and low-volume providers decreased by 25% and 11% showing a faster decline in the number of high-volume providers compared with low-volume surgeons.

Conclusions:

About half of prostatectomies performed by surgeons with below-median annual volume were within a 100-mile driving distance (median of 22 miles) of a high-volume surgeon. Such a referral may result in minimal additional costs and substantially improved outcomes.

1. Introduction

Evidence supports the role of experience in improving surgical and oncological outcomes of radical prostatectomy in early-stage prostate cancer [1–4]. This effect is not limited to prostate cancer and is observed in bladder cancer and even noncancer surgeries [5,6]. Despite these data, more than 80% of surgeons nationwide have an annual volume of less than 10, and perform approximately 40% of prostatectomies [7]. A cost-effectiveness analysis estimated that there is a $1,800 savings, per referral, associated with performing radical prostatectomies at high-volume centers based only on reduced downstream costs of management of treatment failure [8]. Better oncological outcomes and reduced societal costs make a persuasive case for referral; however, it is unclear what proportion of radical prostatectomy candidates has access to a more experienced surgeon, and therefore feasibility and costs of such referrals remain difficult to estimate.

This study aimed to investigate the geography of radical prostatectomies in the United States to determine the distribution of the providers and the patients in need of radical prostatectomy. Building on prior modeling for the societal benefits of performing radical prostatectomies at high-volume centers [8], this study estimates the effect of maintaining status quo for opportunity costs where there is access and where referral is possible within driving distance.

2. Methods

The opportunity cost [9] of performing a radical prostatectomy by a low-volume provider was defined as the value of benefits achievable by performing the surgery by a high-volume provider that was forgone by not making a referral. Assuming that the travel distance was the main determinant of access and the referral costs, access was defined as availability of a high-volume provider within a driving distance of 100 miles, and the referral cost for travel within this range was considered negligible. Therefore, the opportunity costs were calculated for radical prostatectomy cases performed by providers in the lower 50 and 75 percentiles of volume where there was a surgeon in the top 10 percentile of volume within a 100-mile driving distance.

Medicare Provider Utilization and Payment Data for 2012 to 2014 were used to determine the geographic distribution and radical prostatectomy volume for providers [10,11]. Geographic distribution was determined by provider zip code, and volume was determined by the total number of radical prostatectomies per provider per year. SEER 18 data were used to determine the number of radical prostatectomies in 2012 and 2013 in the United States by applying the ratio of surgery to incidence to the US incidence data provided by the American Cancer Society [12–14]. The SEER data for 2012 and 2013 were used to extrapolate (linear extrapolation) the expected numbers for 2014. Using the Medicare data, projections were made for the geographic distribution and provider volume in the United States.

Providers were grouped based on surgical volume. Using the provider zip codes and Google application programming interface [15], the travel distance required for referral to a provider in the top 10 percentile of surgical volume was calculated for providers with below-median volume (lower 50 percentile) as well as the lower 3 quartiles (lower 75 percentile) of volume.

2.1. Measuring the benefits of performing radical prostatectomy by high-volume providers

We previously measured and published the projected benefits of preferentially performing radical prostatectomy by high-volume providers using a Markov model to represent the natural history of prostate cancer after radical prostatectomy [8] (Fig.). Using projections on the number of surgeries within driving distance of a high-volume provider, the Markov model estimated the benefits if these surgeries were redirected to a high-volume provider.

Fig.

States of disease in the Markov model. NED = no evidence of disease; Met= metastatic disease; PSAR = PSA recurrence.

For the purpose of estimating the benefits of surgeries by high- vs. low-volume providers, the reduction in the rates of prostate specific antigen (PSA) recurrence (PSAR) (biochemical failure) was modeled using published data [1,2]. In keeping with standards of care, patients with PSAR were managed in a manner consistent with the rates of salvage therapy and development of metastatic disease [1,2,16–19]. Individual preferences for compliance with treatment were modeled using published data [16–19]. Mortality rates were modeled using disease-specific mortality for prostate cancer [20–22], and US life tables [23].

The total costs of care, exclusive of the costs of prostatectomy and management of short-term and long-term complications and side effects of radical prostatectomy, were calculated. The savings per referral to a high-volume provider were calculated. The reduction in the use of resources, such as shortened length of stay and transfusions, was calculated. Additional improvements in outcomes associated with greater experience, such as lower rates of impotence and incontinence, would give further advantage to radical prostatectomy by high-volume providers. These differences, although modeled, were excluded from monetary calculations [24].

The sums of these benefits (as projected by the Markov model) for referring surgeries performed by the lower 50 and 75 percentile providers within 100-mile driving distance to a top 10 percentile provider were calculated. The sum of the monetary value of these benefits was designated as the opportunity costs of performing radical prostatectomies by lower-volume providers where a higher level of experience was readily available.

2.2. Trends over time

Providers were followed over time with respect to the number of radical prostatectomies over the 3 consecutive years. Providers entering or exiting the prostatectomy segment of the Medicare Provider Utilization and Payment Data were identified and stratified by surgical volume. Experience gap between high- and low-volume providers was measured.

2.3. Study assumptions

A central assumption was that redirecting surgeries by the lower 50 and 75 percentile providers to top 10 percentile providers was associated with the benefits projected by the Markov model. The Markov model assumptions used a life surgical volume of >250 as the definition for high-volume providers; in this study, the annual surgical volume that placed the provider in the top 10 to 15 percentiles of volume was used to designate a provider as a high-volume provider.

In the Markov model, it was assumed that 80% of prostatectomies were minimally invasive. Costs of salvage radiation therapy (SRT), treatment of metastatic disease, and costs of last year of life were taken from the literature and were based on Medicare reimbursement rates [25–27] that tended to be lower than private payer rates [28]. Medicare reimbursement rates for robotic, laparoscopic, and open procedures were equal [25]. The reimbursement rates for prostatectomy by high- and low-volume providers were assumed to be equal (Table 1).

Table 1.

Summary of assumptions

Assumptions	Value (range)	Varied in sensitivity analysis	Source
Share of ORP vs. MIRP
ORP	20%	No	Modeling assumption
MIRP	80%	No	Modeling assumption
Probability of 5-year PSA recurrence: volume
ORP: <50	27% (24%, 30%), N = 1402	Yes, lower and upper end of range	[1]
ORP: 50–99	23% (19%, 27%), N = 696	Yes, lower and upper end of range	[1]
ORP: 100–249	19% (17%, 22%), N = 1575	Yes, lower and upper end of range	[1]
ORP: 250–999	16% (14%, 17%), N = 2940	Yes, lower and upper end of range	[1]
ORP: >1000	8% (6%, 10%), N = 1152	Yes, lower and upper end of range	[1]
MIRP: <50	21% (17%, 26%), N = 793	Yes, lower and upper end of range	[2]
MIRP: 50–99	22% (17%, 29%), N = 611	Yes, lower and upper end of range	[2]
MIRP: 100–249	20% (16%, 24%), N = 946	Yes, lower and upper end of range	[2]
MIRP: 250–1100	13% (10%, 16%), N = 2,352	Yes, lower and upper end of range	[2]
Probability of PSA recurrence by year after surgery
1 –2	45%	No	[19]
3–5	32%	No	[19]
6–9	19%	No	[19]
>10	4%	No	[19]
Probability of metastatic-free survival
At 5 years	67% (61%, 72%)	No	[16]
At 10 years	48% (40%, 56%)	No	[16]
Probability of death
Natural causes	US life table	No	[23]
Operative mortality	0.5%	No	[34]
Stage IV cancer, at 5 years	69% (68%, 69%)	No	[20]
Stage IV cancer, at 10 years	43% (42%, 44%)	No	[20]
Prostatectomy complications, long term
Incontinence-HVC	18%	No	[34–37]
Incontinence-LVC	20%	No	[34–37]
Impotence-HVC	20%	No	[34–37]
Impotence-LVC	20%	No	[34–37]
Prostatectomy hospitalization
Transfusions-HV-ORP	16%	No	[38]
Transfusions-LV-ORP	20%	No	[38]
Transfusions-HV-MIRP	1.6%	No	[38]
Transfusions-LV-MIRP	2.4%	No	[38]
LOS-HV-ORP	2.8	No	[38]
LOS-LV-ORP	3.2	No	[38]
LOS-HV-MIRP	1.8	No	[38]
LOS-LV-MIRP	2.0	No	[38]
Salvage radiation
Probability of SRT after PSA recurrence	37% (32%, 42%)	No	[17,18]
Relative risk reduction for cancer death after SRT	64%	No	[39]
Costs
Cost of SRT	$18,000 ($11,300, $25,000)	Yes, lower and upper end of range	[27]
Cost of management for metastatic disease	$12,765 ($11,543, $13,986)	Yes, lower and upper end of range	[26]
Cost of management in the last year of life	$37,504 ($35,948, $39,060)	Yes, lower and upper end of range	[26]

HV = high volume; LV = low volume; OPR = open radical prostatectomy; MIRP = minimally invasive radical prostatectomy; LOS = length of stay; SRT = salvage radiation therapy.

The study took a societal perspective, and costs and effects were discounted at the standard rate of 3%. The model, assumptions and analysis, conformed to the best practices [29] and recommendations of the US Panel on Cost-Effectiveness in Health and Medicine [30] and the Agency for Healthcare Research and Quality [31].

2.4. Sensitivity analysis

The key assumption in this study that could affect the results was the probability of PSAR depending on the experience of the surgeon. A one-way sensitivity analysis around lower and upper ends of the range of the PSAR probabilities for high- vs. low-volume providers was performed.

One-way sensitivity analyses around the lower and upper ends of range for the costs of SRT, and management costs for metastatic prostate cancer and the last year of life were also performed.

3. Results

Analysis of the Medicare Provider Utilization and Payment Data showed that 10% to 14% of providers performed 27% to 32% of the radical prostatectomies with more than 30 cases per year between 2012 and 2014. This group was designated the top 10 percentile of volume for the purposes of referrals. Surgical volumes of ≤23 and ≤16 per year were designated the lower 75 and 50 percentile providers; these providers performed more than 55% and 33% of radical prostatectomies, respectively. When stratified for driving distance, more than 50% and 47% of radical prostatectomies performed by providers in the lower 75 and 50 percentiles were within a 100-mile driving distance with a median distance of 22 miles, and therefore were considered to have access to a top 10 percentile provider (Table 2).

Table 2.

Results of the analysis of medicare PUP data. Results extrapolated to the US population using American Cancer Society and SEER data. Referrals modeled for the entire US population

Group			Provider			Prostate cancer
						Medicare PUP		SEER 18		ACS	Modeled
		N (%)	Volume	N < 100 m	Median distance	N, RP (% total)	N, RPs ≤ 100 m (% group)	Registry incidence	N, RP (% registry)	US incidence	N, RP	N, RPs ≤ 100 m
2012	All	500 (100%)	11–255			10,285 (100%)		50,277	15,941 (32%)	241,740	76,647
	Top 10 percentile	68 (14%)	>30			3,317 (32%)					24,719
	Bottom 50 percentile	277 (55%)	≤16	147	18m	3,611 (35%)	1,912 (53%)				26,910	14,249
	Bottom 75 percentile	383 (77%)	≤23	204	19m	5,662 (55%)	3,023 (53%)				42,195	22,528
2013	All	470 (94°%)	11–236			9,551 (100%)		42,993	14,678 (34%)	238,590	81,456
	Top 10 percentile	65 (13%)	>30			3,035 (32%)					25,884
	Bottom 50 percentile	260 (52%)	≤16	135	22 m	3,378 (35%)	1,912 (53%)				28,809	14,891
	Bottom 75 percentile	358 (72%)	≤23	199	19m	5,268 (55%)	3,023 (53%)				44,928	25,415
2014	All	444 (89%)	11–210			9,098 (100%)		36,764a	13,515 (37%)	233,000	85,654
	Top 10 percentile	49 (10%)	>30			2,492 (27%)					23,461
	Bottom 50 percentile	231 (46%)	≤16	109	21 m	2,998 (33%)	1,423 (47%)				28,225	13,397
	Bottom 75 percentile	338 (68%)	≤23	166	17m	5,094 (56%)	2,549 (50%)				47,958	23,998

PUP = provider utilization and payment data; ACS = American Cancer Society; RP = radical prostatectomy; 100m = 100-mile radius from a provider in the top 10 percentile of annual radical prostatectomy volume.

^aThe data for 2012 and 2013 were used to extrapolate the registry incidence and prostatectomy figures for 2014.

There were 241,740, 238,590, and 233,000 prostate cancer cases [12,13], and 32%, 34%, and an estimated 37% of these, that is, 76,647, 81,456, and an estimated 85,654 cases underwent radical prostatectomy in 2012, 2013, and 2014, respectively. It was estimated that for providers in the lower 50 percentile of volume, 14,249, 14,891, and 13,397 surgeries could be referred to a top 10 percentile provider within a 100-mile driving distance in 2012, 2013, and 2014, respectively. These numbers for the lower 75 percentile of volume were 22,528, 25,415, and 23,998 surgeries, respectively.

The Markov model projected that over a 20-year post-operative follow-up period, referral to a high-volume provider resulted in $1,807 to $1,839 in discounted savings as a result of fewer treatment failures linked to lower level of experience. This translated into a discounted savings of $24 to $27 million, if the lower 50 percentile surgeries within a 100-mile driving distance of a top 10 percentile provider were referred in each of the years 2012 to 2014. Similarly, these numbers were $40 to $46 million for the lower 75 percentile providers in each of the years 2012 to 2014. These estimates reflected the opportunity costs of not making a referral where access to a high-volume provider within a 100-mile driving distance was available (Table 3).

Table 3.

Summary of the referral benefits. Referrals modeled for surgeries performed by bottom 50 and 75 percentile providers to top 10 percentile providers. Transfusions and long-term complications refer to complications present beyond 12 months postoperatively. Costs of prostatectomy, hospitalization, and management of complications are excluded from analysis

Benefit		Referrals modeled for year 2012				Referrals modeled for year 2013				Referrals modeled for year 2014
		Bottom 50 percentile		Bottom 75 percentile		Bottom 50 percentile		Bottom 75 percentile		Bottom 50 percentile		Bottom 75 percentile
		14,249		22,528		14,891		25,415		13,397		23,998
		Minimum	Maximum	Minimum	Maximum	Minimum	Maximum	Minimum	Maximum	Minimum	Maximum	Minimum	Maximum
Reduction in hospitalization measures	Length of stay	3,333	3,557	5,269	5,624	3,483	3,717	5,944	6,344	3,133	3,344 141	5,613	5,990
	Transfusions	111	150	175	238	116	157	198	268	104		187	253
Reduction in long-terms complications	Incontinence	96	117	152	184	101	122	172	208	90	110	162	196
	Impotence	–	–	–	–	–	–	–	–	–	–	–	–
	Both	10	22	16	36	11	23	18	40	10	21	17	38
Reduction in cancer related events	PSA recurrence	653	679	1,033	1,074	683	710	1,166	1,212	614	639	1,101	1,144
	Metastatic disease	362	382	572	604	378	399	645	681	340	359	609	643
	Salvage radiation therapy	316	369	499	583	330	385	563	657	297	347	532	621
Reduction in mortality	Operative	–	–	–	–	–	–	–	–	–	–	–	–
	Cancer	248	253	392	400	259	264	442	451	233	238	417	426
	All cause	174	199	275	315	182	208	310	356	163	187	293	336
Survival in months	Biochemical failure-Free	76,512	79,563	120,971	125,795	79,960	83,148	136,472	141,914	71,938	74,807	128,862	134,001
Reduction in discounted costs of care	Salvage radiation therapy	$5,324,473	$6,201,355	$8,418,348	$9,804,758	$5,564,358	$6,480,747	$9,497,014	$11,061,068	$5,006,166	$5,830,627	$8,967,475	$10,444,320
	Metastatic disease	$13,194,873	$13,514,096	$20,861,978	$21,366,691	$13,789,346	$14,122,951	$23,535,081	$24,104,464	$12,406,059	$12,706,199	$22,222,800	$22,760,436
	Year of death	$6,810,976	$6,904,398	$10,768,609	$10,916,315	$7,117,833	$7,215,464	$12,148,421	$12,315,053	$6,403,803	$6,491,640	$11,471,043	$11,628,384
	All	$25,742,967	$26,207,204	$40,701,355	$41,435,345	$26,902,773	$27,387,926	$45,916,532	$46,744,570	$24,204,005	$24,640,489	$43,356,296	$44,138,164

The modeled benefits also included 107 to 266 fewer transfusions and 3,133 to 6,344 fewer hospitalization days and 90 to 208 fewer cases of long-term incontinence for each of the years 2012 to 2014 (Table 3). Referral was also associated with 5.4 to 5.6 months of biochemical recurrence-free survival benefit per referred surgery.

The volume of prostatectomies declined steadily between 2012 and 2014 at a rate of 7.1% to 8.5% annually. The number of providers performing radical prostatectomies in the Medicare database also declined at a rate of 6% annually. However, high-volume providers in 2014 declined by 25% compared to 2013, whereas low-volume providers declined by only 11%. Moreover, although the share of prostatectomy for high-volume providers declined by 5% in 2014, the share of prostatectomies by low-volume providers declined only by 2% and the lower 75 percentile providers increased their share by 1%. Conversion from low volume to high volume occurred at a negligible rate of 2 to 3 per year. The experience gap between providers at the median volume in each group (high vs. low) ranged from 37 to 56 per year, resulting in an expanding experience gap of 94 to 125 surgeries per provider over the 3-year period (Tables 2 and 5).

Table 5.

Provider trends over 3 years

			2012a	2013	2014	Any yearb	All yearsc
Number of providers performing prostatectomies for Medicare	All		500	470	444	724	247
	Top 10 percentile (HV)		68	65	49	131	39
	Bottom 50 percentile (LV)		277	260	231	628	35
	Bottom 75 percentile (NHV)		383	358	338	268	111
	Volume range		11–255	11–236	11–210	11–255	11 –255
Entry and exit to medicare prostatectomy service provider list	All	New entry		134	153
		Exit		164	209
		Balance		30	56
	Top 10 percentile (HV)	New Entry		1	1
		Exit		3	8
		Balance		3	19
	Bottom 50 percentile (LV)	New Entry		1	1
		Exit		3	8
		Balance		17	46
Conversions	High-to-low crossover	HV->LV		3	2
		LV->HV		0	1
	Top 10 percentile (HV)	HV->NHV		17	18
		NHV->HV		16	6
	Bottom 50 percentile (LV)	LV->NLV		44	36
		NLV->LV		52	39
Experience gap per providerd	Average volume	HV	49 (56)	45 (54)	37 (55)
	Average volume	LV	13 (13)	13 (13)	13 (13)
	Annual gap	HV-LV	36 (43)	32 (42)	24 (41)
	Cumulative gap	HV-LV	36 (43)	68 (84)	94 (125)

HV = high volume; LV = low volume; NHV = not high volume (includes low volume); NLV= not low volume (includes high volume).

^aBase year for comparison.

^bProviders practicing in at least one of the 3 years listed.

^cProviders practicing in all 3 years listed.

^dFigures for providers designated HV and LV in all 3 years are shown in parentheses.

3.1. Sensitivity analysis

When the lower ends of the range for PSAR probabilities for various prostatectomy settings were used, the modeled referral of prostatectomies was associated with $20 to $22 million in each of the years 2012 to 2014 for lower 50 percentile providers. These numbers for the lower 75 percentile providers were $35 to $40 million.

Using the lower ends of the range for costs of SRT, treatment of metastatic disease and costs of prostate cancer care in the year of death at the same time also resulted in smaller values for savings in modeled referrals: $20 to $21 million in each of the years 2012 to 2014 for the lower 50 percentile and $36 to $39 million for each of the years 2012 to 2014 for the lower 75 percentile (Table 4).

Table 4.

Sensitivity analysis results

Sensitivity Analysis	Benefit		Referrals Modeled for Year 2012				Referrals Modeled for Year 2013				Referrals Modeled for Year 2014
			Bottom 50 percentile		Bottom 75 percentile		Bottom 50 percentile		Bottom 75 percentile		Bottom 50 percentile		Bottom 75 percentile
			14,249		22,528		14,891		25,415		13,397		23,998
			Minimum	Maximum	Minimum	Maximum	Minimum	Maximum	Minimum	Maximum	Minimum	Maximum	Minimum	Maximum
PSA Recurrence Effect Set at Upper End of Range	Survival in Months	Biochemical Failure Free	92,899	96,618	159,676	166,069	93,036	96,761	169,694	176,489	90,319	93,936	161,788	168,265
	Reduction in Discounted Costs of Care	Salvage Radiation Therapy	8,156,879	8,399,450	14,020,193	14,437,129	$8,168,928	$8,411,857	$14,899,880	$15,342,975	$7,930,409	$8,166,245	$14,205,631	$14,628,080
		Metastatic Disease	15,881,042	8,399,450	27,296,627	14,437,129	$15,904,500	$8,411,857	$29,009,332	$15,342,975	$15,440,116	$8,166,245	$27,657,664	$14,628,080
		Year of Death	8,133,456	8,436,942	13,979,934	14,501,571	$8,145,470	$8,449,404	$14,857,094	$15,411,461	$7,907,636	$8,202,696	$14,164,838	$14,693,375
		All	32,501,492	32,850,961	55,864,162	56,464,835	$32,549,502	$32,899,486	$59,369,316	$60,007,677	$31,599,112	$31,938,877	$56,603,047	$57,211,664
PSA Recurrence Effect Set at Lower End of Range	Survival in Months	Biochemical Failure Free	60,770	64,592	104,453	111,022	60,860	64,687	111,006	117,988	59,083	62,799	105,834	112,490
	Reduction in Discounted Costs of Care	Salvage Radiation Therapy	4,901,774	5,140,580	8,425,260	8,835,724	$4,909,015	$5,148,174	$8,953,896	$9,390,115	$4,765,680	$4,997,856	$8,536,696	$8,952,589
		Metastatic Disease	10,190,178	5,140,580	17,515,065	8,835,724	$10,205,230	$5,148,174	$18,614,034	$9,390,115	$9,907,255	$4,997,856	$17,746,727	$8,952,589
		Year of Death	5,699,241	5,901,761	9,795,960	10,144,056	$5,707,660	$5,910,479	$10,410,600	$10,780,537	$5,541,006	$5,737,903	$9,925,526	$10,278,226
		All	20,791,194	22,144,240	35,736,286	38,061,928	$20,821,905	$22,176,951	$37,978,531	$40,450,093	$20,213,941	$21,529,421	$36,208,950	$38,565,352
	Survival in Months	Biochemical Failure Free	73,993	76,943	127,180	132,252	74,102	77,057	135,160	140,550	71,938	74,807	128,862	134,001
Treatment Costs Set at Upper End of Range	Reduction in Discounted Costs of Care	Salvage Radiation Therapy	7,151,567	8,329,351	12,292,244	14,316,642	$7,162,130	$8,341,655	$13,063,511	$15,214,928	$6,953,008	$8,098,093	$12,454,827	$14,506,000
		Metastatic Disease	13,980,895	8,329,351	24,030,620	14,316,642	$14,001,547	$8,341,655	$25,538,402	$15,214,928	$13,592,726	$8,098,093	$24,348,460	$14,506,000
		Year of Death	6,859,952	6,867,379	11,791,011	11,803,777	$6,870,085	$6,877,523	$12,530,829	$12,544,396	$6,669,490	$6,676,711	$11,946,965	$11,959,899
		All	28,424,747	29,170,198	48,856,977	50,138,273	$28,466,734	$29,213,287	$51,922,470	$53,284,160	$27,635,554	$28,360,309	$49,503,182	$50,801,424
Treatment Costs Set at Lower End of Range	Survival in Months	Biochemical Failure Free	73,993	76,943	127,180	132,252	74,102	77,057	135,160	140,550	71,938	74,807	128,862	134,001
	Reduction in Discounted Costs of Care	Salvage Radiation Therapy	3,232,508	3,764,867	5,556,094	6,471,122	$3,237,283	$3,770,428	$5,904,707	$6,877,148	$3,142,760	$3,660,338	$5,629,582	$6,556,712
		Metastatic Disease	11,538,787	3,764,867	19,833,079	6,471,122	$11,555,831	$3,770,428	$21,077,490	$6,877,148	$11,218,421	$3,660,338	$20,095,400	$6,556,712
		Year of Death	6,313,404	6,320,239	10,851,594	10,863,343	$6,322,730	$6,329,575	$11,532,469	$11,544,955	$6,138,117	$6,144,762	$10,995,122	$11,007,026
		All	21,450,453	21,617,057	36,869,434	37,155,796	$21,482,139	$21,648,989	$39,182,777	$39,487,107	$20,854,897	$21,016,875	$37,357,085	$37,647,235

4. Discussion

This analysis provides insights into the distribution of the radical prostatectomies in the United States by provider volume, and demonstrates that with minimal change, referrals to high-volume providers could be made that would result in substantial improvements in outcomes and reduction in costs of care.

Evidence suggests that while at 250 open radical prostatec-tomies, outcomes reach their maximum; for minimally invasive radical prostatectomy, this number may be as high as 750 [2], although robotic surgery may need a lower number to reach peak outcomes. It is difficult to overlook the effect of volume on outcomes [3]; whether it is a result of surgical excellence or patient selection, it comes with experience. In this analysis, snapshots of the Medicare prostatectomy land-scape were used to divide the providers into high- and low-experience groups based on annual volume. Although surgical caseloads may vary year to year, the groups’ designations here of high volume with more than 30 radical prostatectomies vs. low volume with 16 or less (below-average volume) provided a view into the geographic distribution of radical prostatectomies and surgical experience and whether a referral can be reasonably made. Using a driving distance of 100 miles or less as the threshold for a referral with negligible costs, it was demonstrated that 31% of surgeries for the Medicare population were done by providers with below-average volume, and almost half of these occurred within 100 miles of a provider with almost twice the annual volume. The median driving distance for these potential referrals was 22 miles. In our view, this indicated that neither location nor ease of access was an impediment to referral.

Although the volume of surgeries was declining, the rate of decline was faster for surgeries performed by high-volume providers compared with non–high-volume providers. The experience at the median volume for high- and low-volume providers showed a diverging pattern. For those providers that remained high volume (39 providers) or low volume (35 providers) during the 3-year period of this study, the experience gap expanded to 125 surgeries per provider. Although it is possible that over time low-volume providers may reach the lifetime volumes cited in the literature, it is evident that the experience gap would continue to widen. These trends must be further studied from a national health care system’s perspective and in a programmatic manner to build up experience where needed and to also ensure the availability of the best surgical expertise to patients in need of radical prostatectomy. Currently, Medicare does not use specific policy tools to influence providers in or out of radical prostatectomy service. Therefore, it is reasonable to believe that these trends are also present in the private insurance sector.

In the absence of high-quality data from individual payers other than Medicare, using the SEER registry data and the American Cancer Society incidence data, the number of radical prostatectomies that could be referred from a low-volume provider to a high-volume provider as defined here was calculated to be more than 13,000 per year for 2012 to 2014. The projections for monetary and non-monetary benefits of referrals from a Markov model, when applied to the 13,000 cases per year that could be referred to a high-volume surgeon, demonstrated a potential savings of more than $24 million per year. This represented the opportunity cost of foregoing referrals where a high-volume surgeon was available within 100 miles. This figure was also calculated for providers in the lower 75 percentile of volume who do 50% of radical prostatectomies and was in excess of $40 million per year.

Nonmonetary benefits were similarly impressive. A longer average biochemical failure-free survival of more than 5 months and reduced rate of incontinence were viewed as proxies for higher quality of life for the patient. Although, from a societal perspective, the length of stay and the number of transfusions did not affect the costs of surgery and were not included in the monetary calculations, they would free up valuable and limited societal resources and should be regarded as tangible benefits of referral.

Centralization of prostatectomy has been a controversial topic in oncologic urology [32,33]. Although the evidence existed that outcomes did improve with higher experience [1,2], and that these improved outcomes translated to reduced costs of care [8], questions remained as to how these could be realized in practice. Our view, as supported by these results, is one of incremental shift in where prostatectomies are done, starting with geographies where higher experience is readily available. We do not prescribe a specific volume cutoff, as this is a moving target affected by multiple factors such as training and technology. However, it is reasonable to attempt to improve the outcomes for the patients and the health care system by redirecting some of the surgeries as described here.

The sensitivity analysis revealed that using the lower limit of costs of treatment, the results remained stable. Similarly, using the smallest difference in the rate of PSAR between high- and low-volume providers continued to be associated with more than $20 million in savings for each of the years 2012 to 2014.

The savings were reflective of the reduced discounted costs of care achieved over a 20-year follow-up period for a given year when the referral plan was in effect. As such, for a 5-year budgetary period, such a referral program could result in over $100 million in societal savings.

This study is limited by the fact that it is a model and uses extrapolations where empiric data were not available. The providers were categorized into high- and low-volume based on their Medicare patients, and their caseload with private insurance carriers was not included. Annual case-load was used as a proxy for lifetime caseload, and the results must be interpreted with this limitation in mind. These results represent the natural history of prostate cancer, as it is understood today. Therefore, conclusions drawn here may significantly change by new technological and training advances that would narrow the experience gap, reduction in the costs of care, and the development of new and more effective therapies for prostate cancer.

The authors have exercised great care when selecting the evidence for assumptions required in the setup of this study. Although admittedly imperfect, this study provides a glimpse of what may be possible in optimizing the care for patients in need of radical prostatectomy and to the extent possible outlines the downstream financial and nonfinancial consequences of such a referral program. As such, it provides a rationale to support referral to high-volume providers for radical prostatectomy to achieve what may be a win-win-win scenario where patients, society, and providers will benefit from the program.

5. Conclusions

Almost 50% of all radical prostatectomies performed by surgeons with below-average annual caseload are within a 100-mile driving distance (with a median of 22 miles) of a high-volume surgeon in the top 10 percentile of volume. Such a referral is associated with minimal additional costs and substantially improved outcomes and lower costs.