Association Between Hospital and Surgeon Radical Prostatectomy Volume and Patient Outcomes: A Systematic Review

Article Outline

• Abstract
• Methods
  • Selection of Studies
  • Literature Search Strategy
  • Synthesis of Evidence
• Results
  • Hospital Volume Associated Outcomes
  • Cancer Control, Urinary Complications and Operative Quality
  • Surgeon Volume and Patient Outcomes
• Discussion
• Conclusions
• Acknowledgments
• Appendix 1
• Appendix 2
• References
• Copyright

Purpose

We examined the association between hospital and surgeon volume, and patient outcomes after radical prostatectomy.

Materials and Methods

Databases were searched from 1980 to November 2007 to identify controlled studies published in English. Information on study design, hospital and surgeon annual radical prostatectomy volume, hospital status and patient outcome rates were abstracted using a standardized protocol. Data were pooled with random effects models.

Results

A total of 17 original investigations reported patient outcomes in categories of hospital and/or surgeon annual number of radical prostatectomies, and met inclusion criteria. Hospitals with volumes above the mean (43 radical prostatectomies per year) had lower surgery related mortality (rate of difference 0.62, 95% CI 0.47–0.81) and morbidity (rate difference −9.7%, 95% CI −15.8, −3.6). Teaching hospitals had an 18% (95% CI −26, −9) lower rate of surgery related complications. Surgeon volume was not significantly associated with surgery related mortality or positive surgical margins. However, the rate of late urinary complications was 2.4% lower (95% CI −5, −0.1) and the rate of long-term incontinence was 1.2% lower (95% CI −2.5, −0.1) for each 10 additional radical prostatectomies performed by the surgeon annually. Length of stay was lower, corresponding to surgeon volume.

Conclusions

Higher provider volumes are associated with better outcomes after radical prostatectomy. Greater understanding of factors leading to this volume-outcome relationship, and the potential benefits and harms of increased regionalization is needed.

Key Words: prostate, prostatectomy, physician's practice patterns, outcome and process assessment (health care), professional competence

Abbreviations and Acronyms: RP, radical prostatectomy, SEER, Surveillance, Epidemiology and End Results

Prostate cancer is the most common nondermatological cancer in men.¹ It is the second most expensive cancer organ site for Medicare with an $8 billion annual expenditure.² RP is the most common treatment for localized prostate cancer with approximately 60,000 RPs performed annually.³ Variations in screening,⁴ mortality,⁵ treatment use and patient outcomes⁶, ⁷, ⁸ suggest the possibility for quality of care improvements in men with localized prostate cancer. The volume of RPs performed annually by individual surgeons and hospitals is a proposed prostate cancer quality care indicator.⁷, ⁹ To our knowledge valid quantitative estimates of the association between and patient outcomes and provider volume of RPs have not been evaluated. Evidence based hospital referral volume thresholds have been examined for several surgical procedures with similar complexity, including colorectal resections and cystectomy,⁹, ¹⁰, ¹¹ but not for RP. Conceptually one might expect that some outcomes would be more likely to reflect surgical volume and (related) skill, while others would more likely be related to hospital volume and (concomitantly) better organized perioperative care.⁹

We performed a systematic review of the association between surgeon and hospital RP volumes, and patient outcomes. Our review was derived from a report done for the Agency for Healthcare Research and Quality to examine the comparative effectiveness of therapies for clinically localized prostate cancer (Appendix 1).

Back to Article Outline

Methods 

Selection of Studies 

We performed the review according to the recommendations for Meta-analysis of Observational Studies in Epidemiology.¹² We included studies if they evaluated the associative hypothesis between provider characteristics and patient outcomes, and had a control group. We excluded studies not published in English, with no information regarding provider characteristics or with no control comparisons.

Literature Search Strategy 

We searched for original epidemiological studies published in English from 1980 to November 2007 in MEDLINE®, The Cochrane Library, the Centers for Disease Control website, the Catalog of U.S. Government Publications, the LexisNexis® Government Periodicals Index, Digital Dissertations and the Agency for Healthcare Research and Quality.

Synthesis of Evidence 

We evaluated studies and extracted data following a prespecified protocol.¹³ We compared the periods when patient events occurred with databases that the investigators obtained to select participants to avoid including the same patients in the analysis more than once. We scored study quality from 0—poorest to 5—highest (Appendix 1).

We assessed the association between provider/hospital characteristics and clinical outcomes, including surgery related mortality; postoperative complications; failure of cancer control; disease related adverse events; treatment related adverse events, including urinary and fecal incontinence; positive surgical margins; length of hospital stay; hospital costs; and the hospital readmission rate.¹⁴ Hospital and surgeon volumes were based on the annual number of RPs. We examined the strength and direction of the associations to determine whether they depended on year of publication, data source, sampling strategy, statistical adjustment for patient age, race or comorbidities, or tumor characteristics.

We calculated regression coefficients and the SE or 95% CI from reported means and SDs.¹³ We used pooled adjusted rates and ORs to estimate the association with hospital and surgeon volumes independent of cancer stage. Meta-analysis was done to test the consistency of the association from at least 2 studies.¹⁵ We calculated the OR weighted by sample size and the 95% CI from fixed and random effects models with a random intercept for each study.¹³ We used meta-regression models to analyze interactions with the year of data collection, databases to measure outcomes and adjustment for confounding factors.¹³ Calculations were performed using STATA®¹⁶ and SAS® 9.2.¹⁷

Back to Article Outline

Results 

Of the 792 articles identified 17 observational studies of a total of 235,763 men were eligible (table 1). We excluded 775 studies, of which 525 were articles without relevant information on provider characteristics, 166 were reviews or secondary data analyses, 34 were commentaries, 28 were observations of ineligible target populations and the remainder were case reports, editorials, expert opinions, reprints of original reports or articles that reported ineligible outcomes (Appendix 2). Average study quality was 65% of the maximum possible quality score. Of epidemiological investigations that examined associations between hospital RP volumes and patient outcomes¹⁸, ¹⁹, ²⁰, ²¹, ²², ²³, ²⁴ the highest quality studies were those that measured morbidity and urinary complications after RP.

Table 1. Studies of association between provider volume and patient outcome

In all studies the patient was the analysis unit.

Most studies analyzed patients selected from existing databases, including the North American Association of Central Cancer Registries (SEER),¹⁸, ²⁵ the Nationwide Inpatient Sample Database¹⁹, ²⁶ and the Medicare health claims national database.²⁰, ²¹, ²², ²⁷, ²⁸, ²⁹, ³⁰, ³¹, ³² Several single hospital³³, ³⁴, ³⁵, ³⁶, ³⁷ and multihospital²³, ³⁸, ³⁹ studies selected patients at clinics to analyze medical records. One study obtained the Quebec Healthcare Plan database to identify eligible subjects.²³ Few investigators reported a random sampling of study subjects.¹⁹, ²⁰, ²⁴ Investigators defined volume as an annual average of procedures¹⁸, ¹⁹ or the total number of procedures during the study period that were performed at each hospital.²⁰, ²¹, ²² Investigators compared volume measurements from linked SEER and Medicare hospital claims databases, and concluded that the 2 approaches yielded the same results.¹⁸

Studies that investigated provider volumes adjusted for patient age and comorbidity,²², ²⁵, ²⁷, ³⁶ race,¹⁸, ²⁰, ²¹, ²⁷, ³⁴ cancer stage and grade,¹⁸, ²⁴, ³⁴ provider location and teaching status,²⁰, ²¹, ²², ²⁵, ³⁸, ⁴⁰ and clustering of patients and providers.¹⁸ Investigators stated that the target population included patients with localized prostate cancer,⁴¹ reported the number of participants with localized cancer,²⁸, ¹⁹, ²⁴ adjusted for cancer stage and grade¹⁸, ²⁰, ²⁴, ³⁴ or assumed that all patients treated primarily with RP would have localized disease.⁴⁰ The investigators adjusted for comorbidities assessed using ICD codes for major confounding diseases²¹, ²⁷or comorbidity indexes.¹⁸, ¹⁹, ²⁰, ²², ²⁴, ²⁵, ³⁶

Hospital Volume Associated Outcomes 

Four retrospective cohort studies examined the association between hospital RP volumes and surgery related mortality. Investigators defined surgery related mortality as in-hospital death¹⁹ or postoperative death within 30 to 90 days after surgery.¹⁸, ²¹, ²² We combined these 2 measurements. Estimating surgery related mortality based only on in-hospital deaths may be influenced by hospital discharge practices⁴² and could bias the association with provider volume.

Investigators reported death rates in different hospital volume categories. We calculated the death rate corresponding to an increase by 10 procedures performed annually at hospitals for a pooled analysis. Three studies adjusted for patient characteristics when estimating the relative risk of surgery related mortality.¹⁹, ²¹, ²² Two series demonstrated a significant decrease in mortality with increased volume (fig. 1).¹⁹, ²⁰, ²¹, ²², ²⁴ Wennberg et al reported no association but they evaluated men who underwent RP in the mid 1980s.²² A recent study describing mortality showed a nonsignificant 1% decrease per 10 RPs performed.¹⁸

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  Relative risk of outcome corresponding to increase of 10 RPs performed annually at hospital. Ellison (2000), Ellison et al.¹⁹ Yao (1999), Yao and Lu-Yao.²¹ Wennberg (1987), Wennberg et al.²² Hu (2003), Hu et al.²⁰ Ellison (2005), Ellison et al.²⁴

Hospitals with 25 to 54 vs more than 55 RPs performed per year had a 1.71 increase (95% CI 1.20, 2.60) in the odds of death according to 1 study of 66,693 patients treated at a total of 1,334 hospitals.¹⁹ Another large study showed a 42% higher relative increase (RR 1.42, 95% CI 1.20, 1.70) in surgery related mortality at hospitals with fewer than 27 procedures performed per year (51st to 75th percentiles) vs more than 36 (greater than 75th percentile).²¹ A large cross-sectional analysis indicated a 2.60 to 2.90-fold increase in the in-hospital death rate at hospitals where fewer than 6 vs more than 20 RPs were performed per year (tertiles of volume distribution).²⁶

Pooled analysis of 3 cohorts for which the reported mortality risk was reported in categories of hospital volume showed a relative 13% decrease (relative risk 0.87, 95% CI 0.81, 0.94), corresponding to 10 additional RPs performed annually at hospitals (table 2). The relative risk of surgery related mortality at hospitals in the highest quartile (greater than 50 procedures) was almost half that at hospitals in the lowest quartile (fewer than 22) (relative risk 0.51, 95% CI 0.36, 0.71) and 39% lower than at hospitals with 23 to 39 operations per year (2nd quartile). Mean hospital volume appeared to be a reasonable criterion for evaluating the volume effect since it was consistent across studies with different design, quality, patient sampling and volume assessment (test for heterogeneity p = 0.24). The risk of surgery related death was lower at hospitals where more vs fewer than 43 RPs were performed per year (relative risk 0.62, 95% CI 0.47, 0.81).

Table 2. Pooled analysis of association between annual hospital RP volume and patient outcome

Measure (No. studies)	Hospital Vol (No. procedures)	Effect (95% CI)
Surgery related mortality RR (3):
Dose response	10	0.87 (0.81,0.94)
Mean	Greater than 43 vs less than 43	0.62 (0.47,0.81)
Quartiles	Greater than 85 vs less than 22	0.51 (0.36,0.71)
Quartiles	Greater than 85 vs 23–39	0.61 (0.48,0.78)
% Complication difference (3):
Dose response	10	−1.21 (−1.70,−0.71)
Mean	Greater than 43 vs less than 43	−9.70 (−15.80,−3.60)
Quartiles	Greater than 85 vs less than 22	−7.30 (−12.50,−2.10)
Quartiles	23–39 vs Less than 22	−2.85 (−4.60,−1.10)
Length of stay day difference (3):
Dose response	10	−0.32 (−0.44,−0.02)⁎
Quartiles	Greater than 85 vs less than 22	−1.49 (−2.2,−0.82)⁎
Quartiles	Greater than 85 vs 23–39	−0.9 (−1.46,−0.34)⁎
% Late urinary complication difference dose response (2)	10	−0.85 (−1.53,−0.17)⁎
% Long-term incontinence difference in rate dose response (2)	10	0.16 (0.01,0.3)⁎
% Late urinary complication symptom difference dose response (2)	10	−1.16 (−2.7,0.38),notsignificant
% Long-term incontinence symptom difference dose response (2)	10	−0.14 (−0.44,0.17),notsignificant

The absolute risk of hospital related mortality was 0.3% at hospitals with 54 or fewer annual procedures vs 0.2% at those with more than 54.¹⁹ We estimated that 44% of surgery related deaths were attributable to low annual hospital volume.¹⁹ Hospitals with more than 35 annual prostatectomies had an absolute 0.2% decrease in mortality compared to that of hospitals where fewer than 10 per year were performed (0.4% vs 0.6%).²¹ We estimated that 30% of surgery related deaths or 2 saved lives per 1,000 patients treated were attributable to such differences in annual hospital volume.²¹

Three cohort studies examined the association between hospital volume and surgery related morbidity, including cardiac, respiratory and vascular complications, bleeding, renal failure, shock and the need for reoperation.¹⁸, ²⁰, ²¹ For every 10 RPs performed annually at hospitals there was an absolute 1.21% decrease (95% CI −0.71, −1.70) in the rate of complications, which was consistent across studies despite differences in population and study characteristics (table 2). In 1 study the absolute risk of complications was 32% at hospitals with fewer than 34 annual surgeries vs 27% at those with more than 113 (after adjustment for case mix, and patient and provider clustering p = 0.03).¹⁸

Two studies evaluated the relative risk of complications after RP.²⁰, ²¹ Each demonstrated lower complication risks with increased volume (table 1 and fig. 1). The crude complication rate was 31% at hospitals with fewer than 10 annual procedures vs 26% at those with greater than 35.²¹

Cancer Control, Urinary Complications and Operative Quality 

One study of 5,837 men from a total of 348 hospitals examined the association between hospital volume and long-term cancer control for 10 years after RP.²⁴ Investigators defined cancer control failure as postoperative medical or surgical hormone ablation more than 6 months after surgery, adjusted for tumor stage and grade, and patient comorbidity. Patients who were operated on at low volume hospitals (fewer than 16 RPs per year) received ablation therapy 1.25 times more often (95% CI 1.14, 1.38) than those operated on at hospitals where more than 85 RPs per year were performed. The association was consistent across volume categories with a significant decrease in the risk of adjuvant therapy by 2% per 10 additional RPs per year (fig. 1).

Two cohort studies using different definitions of low and high volume examined the association between urinary complications or urinary incontinence and hospital volume.¹⁸, ²⁰ The rate of overall urinary complications was 0.74% lower (95% CI −1.12, −0.36) in patients sampled from the SEER database¹⁸ and 1.83% lower (95% CI −3.57, −0.10) in Medicare beneficiaries,²⁰ corresponding to 10 additional RPs performed annually at hospitals. An 8% lower relative risk of urinary complications was reported in 1 study of hospitalized Medicare beneficiaries (fig. 1).²⁷

Pooled analysis estimated the association between hospital volume as a continuous variable and as rates of surgery related urinary complications (table 2). There was a small 0.85% decrease (95% CI −1.53, −0.17) in the absolute rate of late urinary complications. Men treated at hospitals with a volume above vs below the mean of 43 procedures had a 5.30% lower rate (95% CI −9.30, −1.30) of urinary complications. The threshold effect appeared at approximately this level and higher volume was not associated with a greater decrease in complications.

Assessing RP specimens is important for selecting adjuvant therapy, estimating prognosis and analyzing outcomes, and it has been proposed as an indicator of operative quality of care.⁴³ One study of 113 New York hospitals demonstrated that teaching hospitals and high volume hospitals had better compliance with the Cancer Committee of the College of American Pathologists and the Association of Directors of Anatomic and Surgical Pathology protocols to submit and examine RP specimens.²³ No studies examined positive surgical margins in relation to hospital volume.

Three studies analyzed length of stay or the hospital readmission rate in relation to hospital volume.¹⁹, ²⁰, ²¹ Differences in length of stay and readmission were calculated, corresponding to an increase by 10 procedures in annual hospital volume (fig. 2). All reported a small decrease in length of stay associated with an increase in volume. Pooled analysis detected a small 0.32-day decrease in length of stay (95% CI 0.02, 0.44), corresponding to 10 additional RPs and among quartiles of volume distribution but the effect was inconsistent across the studies (table 2). The decrease in length of stay was 0.9 days (95% CI −1.6, −0.3) at hospitals above vs below the mean of 43 RPs per year. While the decreased length of stay might be attributable to decreased complications, the investigators did not analyze this association.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 2. 

  Difference in length of stay and rehospitalization rate corresponding to increase of 10 RPs performed annually at hospitals. Ellison (2000), Ellison et al.¹⁹ Yao (1999), Yao and Lu-Yao.²¹ Hu (2003), Hu et al.²⁰

One study of changes in hospital volume with time showed that length of stay became shorter at all hospitals irrespective of changes in hospital volume.²¹ However, the decrease in length of stay was greater at hospitals with a relative increase in the volume of prostatectomies performed compared to that at hospitals with a decreased or unchanged volume (−1.6 vs −1.4 or −0.99 days, respectively).

Readmission rates adjusted for patient age, race and comorbidity, and hospital teaching status were 1.3 times higher in patients operated on at hospitals where an average of 9 RPs were performed per year compared with that in patients operated on at high volume facilities with greater than 36.²¹ An increase in hospital volume by 10 RPs per year was associated with a decrease in the relative risk of readmission by 10% (relative risk 0.90, 95% CI 0.85, 0.99) with an absolute rate decrease of 0.24% (Fig. 1, Fig. 2).

Hospital characteristics may confound the association between hospital volume and patient outcomes. Most investigators combined hospitals with different statuses.¹⁸, ¹⁹, ²¹, ²², ²⁷, ³⁹, ⁴⁰ Higher rates of surgery related mortality⁴⁰ and an increased relative risk of death were reported at nonteaching hospitals.²⁷ For profit institutions had an increase in surgery related mortality compared to that at teaching not for profit institutions (relative risk 1.18, 95% CI 1.1, 1.22) after adjustment for year of admission, patient demographics and comorbidities, and hospital volume.²⁷ Pooled analysis did not demonstrate a significant association with surgery related mortality rates¹⁸, ¹⁹, ²¹, ²², ³⁹, ⁴⁰ or relative risk¹⁹, ²¹, ²², ²⁷ and hospital status.¹⁸, ¹⁹, ²¹, ²², ²⁷, ³⁹, ⁴⁰ However, teaching hospitals had a 17.6% lower absolute rate (95% CI −9.0, −25.8) of surgery related complications¹⁸, ²¹, ³⁴, ³⁹ and higher operative quality scores.²³ Multivariate adjustment for surgeon and hospital volume, patient characteristics and residency attenuated the association between academic status and in-hospital complications.²⁰ However, patients treated at government hospitals vs nonprofit institutions were at higher risk for anastomotic stricture (adjusted OR 1.45, 95% CI 1.03 to 2.04) and longer length of stay (0.51 days, 95% CI 0.04 to 0.98).²⁰ Private institutions had a longer length of stay than academic centers.³⁵ Public nonteaching hospitals had a 1.5-day shorter length of stay (95% CI −0.3, −2.6) compared with academic centers.¹⁹, ²⁰, ²¹, ²⁷, ³⁵, ³⁶, ³⁹

Surgeon Volume and Patient Outcomes 

Three studies measured postoperative death within 30 to 60 days after RP in relation to surgeon experience.¹⁸, ²⁵, ⁴⁰ Two studies of the association with surgery related mortality did not show differences in the death rate.¹⁸, ⁴⁰ Pooled analysis of 2 studies of a total of 15,734 men operated on by a total of 1,103 surgeons did not demonstrate a significant increase in mortality with more operations (table 3). No groups reported an adjusted relative risk of surgery related death associated with surgeon volume.

Table 3. Pooled analysis of association between annual surgeon RP volume and patient outcome

Measure Difference (No. studies)	Surgeon Vol	Effect (95% CI)
% Surgery related mortality dose response (2)	10 Procedures	0.02 (−0.01,0.4),notsignificant
% Surgery related complication dose response (4)	10 Procedures	−1.92 (−3.13,−0.70)⁎
Length of stay days (4):
Dose response	Logarithmic scale	−0.97 (−1.45,−0.48)
Quartiles	Less than 3 vs greater than 15	−3.30 (−6.10,−0.50)
Median % long-term incontinence events (2)	Greater than 10 vs less than 10	−0.60 (−0.84,−0.34)
% Long-term incontinence symptom dose response (2)	10 Procedures	−1.2 (−2.5,−0.1)
% Late urinary complication symptom dose response (2)	10 Procedures	−2.4 (−5,−0.1)

Five studies examined the association between surgeon volumes and surgery related complications,¹⁸, ²⁰, ²⁵, ³⁴, ³⁶ including cardiac, respiratory or vascular events, the need for reoperation, bleeding, renal failure and shock. In pooled estimates of 4 studies providing the complication rate patients operated on by surgeons with a higher volume had lower rates of complications (−1.92% per 10 RPs per year, 95% CI −3.13, −0.70) with significant heterogeneity across studies (table 3). The relative risk of surgery related complications adjusted for patient age, race and comorbidity, and hospital type and location was lower (relative risk 0.53, 95% CI 0.32, 0.89) in patients treated by higher volume surgeons (greater than 40 vs fewer than 40 surgeries per year) in 1 study.²⁰ Blood transfusions were 8.6 times more common in patients operated on by surgeons who performed fewer than 15 RPs per year compared to surgeons with a higher annual volume.³⁴

Studies of the association between surgeon volume and quality measures showed rates of late urinary complications, long-term incontinence and positive surgical margins for different categories of surgeon volume.¹⁹, ²⁰, ²⁵, ³³, ³⁸ On pooled analysis the rate of late urinary complications was 2.4% lower (95% CI −5.0, −1.0) and the rate of long-term incontinence was 1.2% lower (95% CI −2, −1.0), corresponding to an increase of 10 RPs per year (table 3). Surgeon volume was not associated with positive surgical margins.

Length of hospital stay in relation to surgeon volume was assessed in 4 studies.²⁰, ³⁵, ³⁶, ³⁷ Pooled analysis showed a decrease in length of stay by −0.97 days (95% CI −1.45, −0.48), corresponding to surgeon volume expressed in the logarithmic scale, meaning that there was a significant dose response association, adjusted for skewedness of volume distribution with large case loads, at a few high volume hospitals. Patients operated on by surgeons performing more than 15 RPs per year (4th quartile) had a 3.3-day shorter hospital stay (95% CI −6.1, −0.5) than patients operated on by surgeons who performed fewer than 3 surgeries per year (1st quartile) (table 3).

Back to Article Outline

Discussion 

The current analysis confirmed previous reviews that patients have better outcomes after RP performed by high volume providers,⁷, ⁸ similar to the positive association between patient outcomes with other surgical procedures.¹⁰ The current quantitative synthesis detected a small significant and consistent decrease in surgery related mortality and morbidity at higher volume hospitals despite substantial differences in population and study characteristics.

We attempted to find an evidence-based threshold of provider volume, considering a wide variation in volume categories in individual studies. Between 33% and 70% of men were treated at low volume hospitals¹⁸, ²⁵ with a definition of fewer than 6²⁶ to fewer than 60²⁰ surgeries per year. The threshold in the current analysis (approximately 40 procedures for hospitals and around 13 for surgeons) was also suggested in the recently published cross-sectional analysis of the Nationwide Inpatient Sample of the Health Care Utilization Project²⁶ and in a representative survey of 2,000 urologists who treated prostate carcinoma in the United States.⁴⁴ The studies defined an annual RP threshold of 30 procedures at a hospital and more than 10 for an individual surgeon.

The role of changes in provider volume and patient outcome remains unclear. Evidence is limited to a decreased length of stay at hospitals that maintained a high volume for 3 years.²¹ Variation in outcomes occurred even in the category of high volume surgeons.¹⁸, ²⁵ Future analysis would clarify whether patient outcomes are sensitive to continuous annual volume measures or thresholds calculated as an average during several years.

The magnitude of the effect differed depending on the baseline risk of the outcomes. The absolute risk of surgery related mortality is low after RP compared to that after other cancer surgeries.¹¹ However, even modest absolute decreases in outcomes such as surgery related mortality can be considered clinically important when the relative attributable decrease is large (30% of deaths were attributable to low hospital volume). We could not determine how decreases in morbidity and surgery related complications were associated with decreased length of stay, although such a relationship seems likely. Cost benefits from avoided general morbidity, late urinary complications and long-term incontinence should be estimated.

Despite great heterogeneity among the studies in this period, the outcomes studied, volume categories and methods of risk adjustment we found that surgery related mortality and morbidity were consistently associated with provider volume. No studies included a comprehensive adjustment for all likely confounders. Such a comprehensive risk model would include a better conceptual understanding of different patient and provider characteristics along the disease management pathway. Hospital ownership, teaching status and specialization may be associated with hospital/surgeon volume and reflect a higher commitment to overall quality of care and patient outcomes¹⁴, ²⁶ As expected, hospital volume was associated with the outcomes that depend on hospital organization and support, while surgeon volume demonstrated an association with procedure specific, late urinary complications. The quality of nonsurgical in-hospital and post-discharge care was not reported and it may contribute to differences in process of care and patient outcomes.²⁶ Adjustment for patient characteristics was performed in most studies that evaluated hospital but not surgeon volume. Precise tumor characteristics, including grade, stage and prostate specific antigen, were rarely reported.

A conceptual model would suggest that some outcomes are more directly related to surgical technique and skill and, hence, to surgical volume, while others are related to perioperative care and linked to hospital volume. Mortality rates might reflect the 2 factors. Our findings show that mortality and morbidity rates varied with hospital volume but not with surgical volume, while urinary complications were associated with surgeon volume.

The association between surgeon and hospital volumes, and the potential role of referral patterns as well as clustering between patients and providers deserves scrutiny.¹⁸, ⁴⁵, ⁴⁶, ⁴⁷ Patients select providers based on informal knowledge rather than on a comprehensive analysis of numerical probabilities of clinical outcomes.⁴⁸, ⁴⁹ Physician referral preferences and choices of treatment depend not only on patient prognosis, but also on available health care, hospital structure, surgical skills, clinician gender, specialty, training and experience (http://www.ahrq.gov/clinic/epcix.htm). Nested hierarchical models with random effects to incorporate clustering of hospital and surgeon characteristics, and referral patterns or propensity scores at the individual patient level with adequate adjustment for case mix and quality of nonsurgical care may provide more valid estimates of the true volume association with patient outcomes. Randomized trials to test volume associations are not feasible. Stratified randomization by provider volumes with assessment of patient outcomes among subgroups of hospitals and surgeons with different volumes is possible in multicenter trials of the comparative effectiveness of surgical treatment for prostate cancer. Future research is needed to examine the association between patient outcomes and other potential volume-outcome sensitive treatments, including external beam radiation, brachytherapy, cryoablation and high intensity focused ultrasound therapy.

Back to Article Outline

Conclusions 

Higher provider volumes are associated with better outcomes after RP. A greater understanding is needed of factors leading to this volume-outcome relationship, and the potential benefits and harms of increased regionalization.

Back to Article Outline

Acknowledgments 

Dr. David Jacobs assisted with conceptualization and methodology of the meta-analysis, Alisha Baines and Diane Smith with programming, and Jim Beattie, Lisa McGuire, Judy Stanke and Dr. Delbert Reed with the literature search.

Back to Article Outline

Appendix 1 

This study is part of a larger evidence report performed for the Agency for Healthcare Research and Quality to examine the comparative effectiveness of therapies for clinically localized prostate cancer. The full report is available on the agency website (Wilt TJ, Shamliyan T, Taylor B, MacDonald R, Tacklind J, Rutks I et al: Comparative Effectiveness of Therapies for Clinically Localized Prostate Cancer. Comparative Effectiveness Review No. 13 [Prepared by the Minnesota Evidence-based Practice Center under Contract No. 290-02-0009]. Rockville, Maryland: Agency for Healthcare Research and Quality, February 2008. Available at: www.effectivehealthcare.ahrq.gov/reports/final.cfm). It has not otherwise been published.

Back to Article Outline

Appendix 2 

The reference list with justifications for exclusions is available at www.effectivehealthcare.ahrq.gov/reports/final.cfm.

Back to Article Outline

References 

1. Edwards BK, Brown ML, Wingo PA, Howe HL, Ward E, Ries LA, et al. Annual report to the nation on the status of cancer, 1975–2002, featuring population-based trends in cancer treatment. J Natl Cancer Inst. 2005;97:1407
   • View In Article
   • CrossRef
2. Brown ML, Riley GF, Schussler N, Etzioni R. Estimating health care costs related to cancer treatment from SEER-Medicare data. Med Care. 2002;40(IV-):104
   • View In Article
3. Healthcare Cost and Utilization Project (U.S.) (United States Agency for Healthcare Research and Quality). http://www.ahrq.gov/data/hcup/Accessed May 1, 2008
   • View In Article
4. Behavioral Risk Factor Surveillance System Survey Data, 2005 (National Center for Chronic Disease Prevention and Health Promotion). http://www.cdc.gov/brfss/technical_infodata/surveydata/2005.htmAccessed May 1, 2008
   • View In Article
5. United States Cancer Statistics: 1999–2002 Incidence and Mortality Web-based Report Version (United States Cancer Statistics Working Group). Atlanta: Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute; 2005;
   • View In Article
6. Chassin MR, Galvin RW. The urgent need to improve health care quality (Institute of Medicine National Roundtable on Health Care Quality). JAMA. 1998;280:1000
   • View In Article
   • MEDLINE
   • CrossRef
7. Nuttall M, van der Meulen J, Phillips N, Sharpin C, Gillatt D, McIntosh G, et al. A systematic review and critique of the literature relating hospital or surgeon volume to health outcomes for 3 urological cancer procedures. J Urol. 2004;172:2145
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (96 KB)
   • CrossRef
8. Joudi FN, Konety BR. The impact of provider volume on outcomes from urological cancer therapy. J Urol. 2005;174:432
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (172 KB)
   • CrossRef
9. Halm EA, Lee C, Chassin MR. Is volume related to outcome in health care? (A systematic review and methodologic critique of the literature). Ann Intern Med. 2002;137:511
   • View In Article
10. Birkmeyer JD, Siewers AE, Finlayson EV, Stukel TA, Lucas FL, Batista I, et al. Hospital volume and surgical mortality in the United States. N Engl J Med. 2002;346:1128
    • View In Article
    • CrossRef
11. Goff V. Advancing Patient Safety: The Leapfrog Group. In: Washington, D. C: National Health Care Purchasing Institute; 2001;p. 4
    • View In Article
12. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting (Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group). JAMA. 2000;283:2008
    • View In Article
    • MEDLINE
    • CrossRef
13. The Cochrane CollaborationHiggins J, Green S. The Cochrane Handbook for Systematic Reviews of Interventions. Chichester, United Kingdom: John Wiley and Sons; 2005;
    • View In Article
14. Gagliardi AR, Fleshner N, Langer B, Stern H, Brown AD. Development of prostate cancer quality indicators: a modified Delphi approach. Can J Urol. 2005;12:2808
    • View In Article
    • MEDLINE
15. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557
    • View In Article
    • CrossRef
16. Sterne JAC, Bradburn MJ, Egger M. Meta-analysis in Stata. In:  Egger M,  Smith GD,  Altman DG editor. Systematic Reviews in Health Care: Meta Analysis in Context. London: BMJ Books; 2001;p. 347–369
    • View In Article
17. Littell RC, Milliken GA, Stroup WW, Wolfinger RD. SAS System for Mixed Models. Cary, North Carolina: SAS Publishing; 1996;
    • View In Article
18. Begg CB, Riedel ER, Bach PB, Kattan MW, Schrag D, Warren JL, et al. Variations in morbidity after radical prostatectomy. N Engl J Med. 2002;346:1138
    • View In Article
    • CrossRef
19. Ellison LM, Heaney JA, Birkmeyer JD. The effect of hospital volume on mortality and resource use after radical prostatectomy. J Urol. 2000;163:867
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (265 KB)
    • CrossRef
20. Hu JC, Gold KF, Pashos CL, Mehta SS, Litwin MS. Role of surgeon volume in radical prostatectomy outcomes. J Clin Oncol. 2003;21:401
    • View In Article
    • MEDLINE
    • CrossRef
21. Yao SL, Lu-Yao G. Population-based study of relationships between hospital volume of prostatectomies, patient outcomes, and length of hospital stay. J Natl Cancer Inst. 1999;91:1950
    • View In Article
    • MEDLINE
    • CrossRef
22. Wennberg JE, Roos N, Sola L, Schori A, Jaffe R. Use of claims data systems to evaluate health care outcomes (Mortality and reoperation following prostatectomy). JAMA. 1987;257:933
    • View In Article
    • MEDLINE
23. Imperato PJ, Waisman J, Wallen M, Pryor V, Starr H, Rojas M, et al. The use of quality indicators for assessing radical prostatectomy specimens. Am J Med Qual. 2000;15:212
    • View In Article
    • MEDLINE
    • CrossRef
24. Ellison LM, Trock BJ, Poe NR, Partin AW. The effect of hospital volume on cancer control after radical prostatectomy. J Urol. 2005;173:2094
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (149 KB)
    • CrossRef
25. Bianco FJ, Riedel ER, Begg CB, Kattan MW, Scardino PT. Variations among high volume surgeons in the rate of complications after radical prostatectomy: further evidence that technique matters. J Urol. 2005;173:2099
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (340 KB)
    • CrossRef
26. Konety BR, Allareddy V, Modak S, Smith B. Mortality after major surgery for urologic cancers in specialized urology hospitals: are they any better?. J Clin Oncol. 2006;24:2006
    • View In Article
    • CrossRef
27. Yuan Z, Cooper GS, Einstadter D, Cebul RD, Rimm AA. The association between hospital type and mortality and length of stay: a study of 16.9 million hospitalized Medicare beneficiaries. Med Care. 2000;38:231
    • View In Article
    • MEDLINE
    • CrossRef
28. Bubolz T, Wasson JH, Lu-Yao G, Barry MJ. Treatments for prostate cancer in older men: 1984–1997. Urology. 2001;58:977
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (363 KB)
    • CrossRef
29. Brandeis J, Pashos CL, Henning JM, Litwin MS. A nationwide charge comparison of the principal treatments for early stage prostate carcinoma. Cancer. 2000;89:1792
    • View In Article
    • MEDLINE
    • CrossRef
30. Litwin MS, Pasta DJ, Stoddard ML, Henning JM, Carroll PR. Epidemiological trends and financial outcomes in radical prostatectomy among Medicare beneficiaries, 1991 to 1993. J Urol. 1998;160:445
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (500 KB)
    • CrossRef
31. Lu-Yao G, Albertsen PC, Stanford JL, Stukel TA, Walker-Corkery ES, Barry MJ. Natural experiment examining impact of aggressive screening and treatment on prostate cancer mortality in two fixed cohorts from Seattle area and Connecticut. BMJ. 2002;325:740
    • View In Article
    • CrossRef
32. Saigal CS, Pashos CL, Henning JM, Litwin MS. Variations in use of imaging in a national sample of men with early-stage prostate cancer. Urology. 2002;59:400
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (88 KB)
    • CrossRef
33. Hernandez DJ, Epstein JI, Trock BJ, Tsuzuki T, Carter HB, Walsh PC. Radical retropubic prostatectomy (How often do experienced surgeons have positive surgical margins when there is extraprostatic extension in the region of the neurovascular bundle?). J Urol. 2005;173:446
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (351 KB)
    • CrossRef
34. Dash A, Dunn RL, Resh J, Wei JT, Montie JE, Sanda MG. Patient, surgeon, and treatment characteristics associated with homologous blood transfusion requirement during radical retropubic prostatectomy: multivariate nomogram to assist patient counseling. Urology. 2004;64:117
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (113 KB)
    • CrossRef
35. Gheiler EL, Lovisolo JA, Tiguert R, Tefilli MV, Grayson T, Oldford G, et al. Results of a clinical care pathway for radical prostatectomy patients in an open hospital-multiphysician system. Eur Urol. 1999;35:210
    • View In Article
    • MEDLINE
    • CrossRef
36. Litwiller SE, Djavan B, Klopukh BV, Richier JC, Roehrborn CG. Radical retropubic prostatectomy for localized carcinoma of the prostate in a large metropolitan hospital: changing trends over a 10-year period (1984–1994) (Dallas Outcomes Research Group for Urological Disorders). Urology. 1995;45:813
    • View In Article
    • Abstract
    • Full-Text PDF (1027 KB)
    • CrossRef
37. Leibman BD, Dillioglugil O, Abbas F, Tanli S, Kattan MW, Scardino PT. Impact of a clinical pathway for radical retropubic prostatectomy. Urology. 1998;52:94
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (672 KB)
    • CrossRef
38. Eastham JA, Kattan MW, Riedel E, Begg CB, Wheeler TM, Gerigk C, et al. Variations among individual surgeons in the rate of positive surgical margins in radical prostatectomy specimens. J Urol. 2003;170:2292
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (68 KB)
    • CrossRef
39. Gaylis FD, Friedel WE, Armas OA. Radical retropubic prostatectomy outcomes at a community hospital. J Urol. 1998;159:167
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (580 KB)
    • CrossRef
40. Karakiewicz PI, Bazinet M, Aprikian AG, Tanguay S, Elhilali MM. Thirty-day mortality rates and cumulative survival after radical retropubic prostatectomy. Urology. 1998;52:1041
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (144 KB)
    • CrossRef
41. Krupski TL, Kwan L, Litwin MS. Sociodemographic factors associated with postprostatectomy radiotherapy. Prostate Cancer Prostatic Dis. 2005;8:184
    • View In Article
    • MEDLINE
    • CrossRef
42. Kahn KL, Brook RH, Draper D, Keeler EB, Rubenstein LV, Rogers WH, et al. Interpreting hospital mortality data (How can we proceed?). JAMA. 1988;260:3625
    • View In Article
    • MEDLINE
43. Recommendations for the reporting of resected prostate carcinomas (Association of Directors of Anatomical and Surgical Pathology). Hum Pathol. 1996;27:321
    • View In Article
    • Full-Text PDF (289 KB)
44. Denberg TD, Flanigan RC, Kim FJ, Hoffman RM, Steiner JF. Self-reported volume of radical prostatectomies among urologists in the USA. BJU Int. 2007;99:339
    • View In Article
    • MEDLINE
    • CrossRef
45. Luft HS. The relation between surgical volume and mortality: an exploration of causal factors and alternative models. Med Care. 1980;18:940
    • View In Article
    • MEDLINE
    • CrossRef
46. Luft HS, Bunker JP, Enthoven AC. Should operations be regionalized? (The empirical relation between surgical volume and mortality). N Engl J Med. 1979;301:1364
    • View In Article
    • MEDLINE
    • CrossRef
47. Luft HS, Hunt SS, Maerki SC. The volume-outcome relationship: practice-makes-perfect or selective-referral patterns?. Health Serv Res. 1987;22:157
    • View In Article
    • MEDLINE
48. Denberg TD, Melhado TV, Steiner JF. Patient treatment preferences in localized prostate carcinoma: the influence of emotion, misconception, and anecdote. Cancer. 2006;107:620
    • View In Article
    • MEDLINE
    • CrossRef
49. Denberg TD, Beaty BL, Kim FJ, Steiner JF. Marriage and ethnicity predict treatment in localized prostate carcinoma. Cancer. 2005;103:1819
    • View In Article
    • MEDLINE
    • CrossRef