Doi:10.1016/j.juro.2006.11.083

Economic Evaluation of Treatment Strategies for Benign Prostatic
Hyperplasia—Is Medical Therapy More Costly in the Long Run?
Christopher S. Saigal,* Mehran Movassaghi, Jennifer Pace, Geoffrey Joyce and the Urologic
Diseases in America Project
From the Department of Urology, University of California-Los Angeles Medical Center, Los Angeles, California
Purpose: Although medical therapy for newly diagnosed benign prostatic hyperplasia is initially less expensive than
surgery, to our knowledge the long-term costs of these treatments are unknown. We defined longer term costs of these
treatment strategies.
Materials and Methods: We examined spending on benign prostatic hyperplasia related services by examining health care
claims for a 5-year period subsequent to a new benign prostatic hyperplasia diagnosis. Expenditures for subjects treated
initially with surgery were compared to expenditures for those with initial medical treatment. Expenditures were projected
during longer periods and the net current value of these expenditures was calculated.
Results: Of the 970 subjects identified who received benign prostatic hyperplasia treatment 913 (94.1%) relied on medical
therapy as initial treatment. Of those subjects 832 (91.1%) were on ␣-blockers. The secondary treatment rates for surgery far
exceeded those for medical therapy (37% vs 8%). Average total expenditures were higher for subjects who initially received
surgery ($12,699, 95% CI 9,865–15,533) than for those initially treated with medication ($2,193, 95% CI 1,959 –2,428). If
future streams of spending were discounted at standard rates (3%), the costs of initial medical therapy as a treatment
strategy would always be lower than those of initial surgical therapy even at 40 years.
Conclusions: In a cohort of privately insured men with newly diagnosed benign prostatic hyperplasia monotherapy with
␣-blockers was the most common initial treatment. Surgical therapy was associated with higher treatment failure rates and
higher costs during 5 years. Increased expenditures related to initial surgical therapy were consistent when projected over
long time frames.
Key Words: prostate, prostatic hyperplasia, outcome assessment (health care), health expenditures Beforetheadventofeffectivemedicaltherapysurgery edgedataoncomplicationratesandlong-termexpenditures for BPH was the mainstay of treatment. However, associated with BPH treatment have not been reported. To ␣-adrenergic antagonist medications (␣-blockers), gain insight into the costs of differing initial treatment strat- which were introduced in the 1990s, have largely supplanted egies for BPH we evaluated the costs of initial medical or surgical therapy as first line treatment for BPH, as evi- surgical therapy for BPH in a cohort of privately insured denced by the rapidly decreasing rate of BPH related sur- men followed for 5 years. Our primary hypothesis was that The number of prostatectomies performed in Medi- for younger men with BPH initial surgical treatment would care beneficiaries decreased from 250,000 in 1987 to 88,000 be less expensive than lifelong medical therapy after ac- in This 65% decrease in TURP occurred despite a counting for treatment failure rates.
sustained increase in the number of male beneficiaries andit coincided with a dramatic increase in office visits forThis change in the national treatment paradigm for MATERIALS AND METHODS
BPH occurred with only limited analysis of the implications Data Source
We examined spending on BPH related services for noneld- Although initial surgical treatment is more costly than erly patients with employer provided insurance. The data initial medical therapy in the short term, to our knowledge included enrollment files as well as medical and pharmacy the longer term costs associated with these treatment strat- claims for 25 large employers in the United States covering egies are unknown, especially the costs associated with med- 40,895 male beneficiaries 40 years and older who were con- ical and surgical failures. Studies of this topic examined this tinuously enrolled in their employer provided plan from question using computer models, relying on various assump- 1997 to 2002. Claims files captured all health care claims tions drawn from experts and clinical To our knowl- and encounters, including prescription drugs. Medical claimsincluded date of service, diagnosis and procedure codes, andexpenditures, including billed charges, negotiated discounts, Submitted for publication July 12, 2006.
excluded expenses, deductibles, co-payments, payments made * Correspondence: Department of Urology, Center for Health Sci- by the employer and employee, and other third party coverage.
ences, University of California-Los Angeles Medical Center, Box Drug claims included information on drug type, place of pur- 951738, 10833 Le Conte Ave., Los Angeles, California 90095 (tele-phone: 310-206-8183; FAX: 310-206-5343; e-mail: csaigal@ucla.edu).
chase and expenditures. The enrollment files allowed us to Copyright 2007 by AMERICAN UROLOGICAL ASSOCIATION DOI:10.1016/j.juro.2006.11.083
ECONOMIC EVALUATION OF BENIGN PROSTATIC HYPERPLASIA TREATMENT STRATEGIES track who was eligible for services as well as basic demograph- ics, such as patient age, gender, marital status and relation-ship to the sponsoring employee.
Cohort Identification
We identified patients receiving treatment for BPH based on diagnostic codes (International Classification of Disease, 9th edition), procedure codes (Common Procedural Terminol- ogy), and National Drug Codes for BPH and BPH related medications (see Appendix). Subjects were initially selected based on primary or secondary International Classification of Disease, 9th edition diagnosis codes for BPH. Men whohad a claim as described in 1997 were excluded to create acohort of incident cases of BPH in 1998. Subjects were fur- were for generic formulations of the drugs. A total of 71 men ther stratified according to the type of medical intervention (7.8%) who were initially on medical therapy underwent a (␣-blocker alone, 5␣-reductase inhibitor alone or combina- surgical procedure at a later date. Of these men 92% initially tion therapy) using medication claims. These patients were used ␣-blocker monotherapy, 7% initially used 5␣-reductase then followed through their claims to determine the need for inhibitors and 1% initially used combination therapy. Of the continued or additional therapy as well as related expendi- subjects 57 had surgical intervention as their index event, tures during the 5 years of observation.
including transurethral prostatectomy in 56 and open pros- To accomplish this, subjects were stratified into 2 cohorts tatectomy in 1 (cohort 2). Seven of those men (12%) under- according to their initial BPH treatment selection. Cohort 1 went a second procedure and 14 (25%) received subsequent was composed of subjects who initially chose medical therapy with an ␣-blocker medication, a 5␣-reductase inhibitor medi- For subjects choosing initial medical therapy who did not cation or a combination of the 2 medications. We also examined undergo subsequent surgery average time on medical ther- costs in the subgroup composed of subjects who initially were apy was 3.2 years. Of these subjects 10% filled only 1 pre- treated using medication but who later underwent a procedure scription and excluding them increased average duration to to treat BPH (medication failure). Cohort 2 includes individu- 3.5 years. The mean time of medical therapy for subjects als who selected a BPH procedure for initial therapy. We also who started with medical therapy but subsequently under- examined costs in the subgroup of subjects who subsequently received medical therapy for BPH after initial surgical therapy lists expenditures for each cohort. In inflation and the subgroup of those who subsequently underwent repeat adjusted 2002 dollars average total expenditures during the 5 study years were higher for subjects who initially under- Annual medical and pharmacy expenditures were calcu- went surgery ($12,699, 95% CI 9,865–15,533) than for those lated for each cohort and subcohort. Expenditures consisted of initially treated with medication ($2,193, 95% CI 1,959 – total annual payments made by the enrollee (co-payments, 2,428). At 5 years therapy with ␣-blocker medication was deductibles and excluded expenses) and by all third party less expensive than therapy with 5␣-reductase inhibitor (ex- payers (primary and secondary coverage, and net of negoti- cluding subjects on combination therapy $1,397 vs $2,175, p ated discounts) for BPH related medical services and pre- Ͻ0.001). also lists expenditures by specific treat- scription drug claims. Expenditures were tabulated for each ment paths. A total of 71 subjects initially chose medical group during the 5 years and a direct comparison was made.
therapy but underwent surgical therapy at some point in the To determine total expenditures the expenditures for all next 5 years, which increased average total expenditures to BPH related office visits, inpatient stays, drugs and proce- $9,012 (95% CI 6,753–11,271). A total of 11 subjects under- dures were summed annually and after 5 years. Addition- went surgery initially but were subsequently placed on med- ally, we calculated the net current value of initial medical vs ical therapy, which resulted in 5-year average total expen- surgical treatment for 5, 10, 20 and 40-year horizons. A ditures of $11,745 (95% CI 6,857–16,634).
discount rate of 3% was applied to these calculations and we If future streams of spending were discounted at stan- assumed that prices were fixed in constant dollars.
dard rates (3%), the costs of initial medical therapy wouldalways be lower than initial surgical therapy. Even at 40 years the net current value of the initial medication strategywas $7,816, while the net current value of the initial surgical shows the age distribution of the sample. Of the subjects 1,952 had a new BPH diagnosis code in 1998 but noassociated claim for medical or surgical therapy.
DISCUSSION
Of the 970 subjects identified who received BPH treat- ment 913 (94.1%) relied on medical therapy as initial treat- Our study has several significant findings. First, expendi- ment (cohort 1). Of those subjects 832 (91.1%) were on tures in a 5-year period were higher in the cohort of subjects ␣-blockers, 62 (6.8%) were on 5␣-reductase inhibitor therapy initially choosing surgery and they continued to be higher in and 19 (2.1%) were on combination therapy with the 2 types future time frames. Contrary to our hypothesis, there was no of medication Because of the small number of time frame at which surgical therapy was cumulatively less individuals on combination therapy, this group was ana- expensive than medical therapy in current value dollars.
lyzed with the cohort initially placed on 5 ␣-reductase inhib- Total average expenditures during 5 years were signifi- itors. Of total terazosin and doxazosin prescriptions 37% cantly higher for subjects undergoing surgery as an initial ECONOMIC EVALUATION OF BENIGN PROSTATIC HYPERPLASIA TREATMENT STRATEGIES TABLE 2. Initial and subsequent BPH treatments treatment strategy. The dramatically higher expenditures this sample of men with private insurance. This finding is associated with a surgical procedure during the initial study consistent with trends documented The num- year were not significantly mitigated by lower medication bers of subjects treated with 5␣-reductase inhibitor or sur- costs. Our results are consistent with computer model based gery was similar and far smaller than the number treated economic analyses of the direct treatment costs of ␣-blockers with ␣-blockers. Treatment with combination medical ther- vs 5␣-reductase inhibitors vs Using a 2-year time apy was relatively rare. During the study period sparse data frame Lowe et al found that costs for terazosin were about supported the effectiveness of 5␣-reductase inhibitors as 38% of costs for TURP based on costs in a privately insured monotherapy or combination therapy in most men with population, which is higher than our corresponding 11% BPH. However, subsequent publication of the MTOPS trial, This difference may be due to our longer time frame for which showed a significant decrease in the risk of BPH analysis (capturing more treatment failures) or differences in progression in men on combination may have in- the assumptions made in the model used by Lowe et al.
creased the initial use of combination therapy in men diag- It is interesting to note that, although medical therapy nosed with BPH in 2003 and later. Surprisingly only 33% of for BPH with ␣-blockers may be considered lifelong therapy, men with a new BPH diagnosis were treated in year 1 of the average duration of treatment in this study was shorter diagnosis, a finding similar to the 23% active treatment rate than the 5 years of available observation even when exclud- in year 1 of diagnosis documented by Black et al in a pro- ing subjects who only filled 1 prescription. This phenomenon prietary claims This finding may be partially the contributed to lower expenditures in the cohort treated ini- result of artifact since a physician claim for an office visit tially with medication. Expenditures for medical therapy in may include a diagnosis code for BPH if BPH is suspected as our study are smaller than those reported previously. Lowe the reason for patient complaints. This diagnosis may later et al performed a trial in men randomized to terazosin or change, obviating the need for BPH treatment. Our bias is placebo and reported health care costs from the perspective that this clinical scenario is not common enough to explain of the health care plan at 1 Expenditures were calcu- the high proportion of men on watchful waiting. We believe lated using a private insurance health care claims database.
that these data support the thesis that many patient/physi- Annual expenditures were $2,932 in the terazosin group.
cian dyads select watchful waiting as initial therapy. To our However, in that study all health care use was recorded. Our knowledge the true appropriate active treatment rate for data are specific to BPH related services. Medication costs in men with BPH is undefined. In practice the active treatment the terazosin group for year 1 were $327 more than in the rate may be influenced by several variables, including the placebo group, similar to annual costs in our ␣-blocker structure of patient health care benefits.
group. Given the results of the MTOPS trial, which showed Third, the secondary treatment rates for surgery far ex- that therapy with 5␣-reductase inhibitor decreased the need ceeded those for medical therapy. Approximately 8% of sub- for subsequent surgical therapy (or the medical therapy jects on medical therapy required subsequent surgical inter- failure rate) by medical therapy may be even more vention, while 25% undergoing surgery subsequently required favored from a cost perspective if combination medical ther- medical therapy. Additionally, 12% of the surgery cohort re- quired second surgery, including 5% who required second sur- Second, for those who were treated ␣-blocker medication was overwhelmingly the most common initial treatment in TABLE 4. Net current value of expenditures related to initial medical or surgical therapy projected over various TABLE 3. Five-year expenditures per individual by initial cohort ECONOMIC EVALUATION OF BENIGN PROSTATIC HYPERPLASIA TREATMENT STRATEGIES gery and received subsequent medical therapy, for a total sec- APPENDIX
ondary treatment rate of 37%. This secondary treatment rate Administrative Codes Used to Identify Subjects
is much larger than that reported in published clinical series With BPH, or Use of BPH Related Medication
(range to Our higher rates of secondary treatment or Procedures
may reflect the relatively long followup in our cohort. Studies ofTURP re-treatment rates describe outcomes at less than 2 1) Any male 40 years or older with
Additionally, subsequent treatment with BPH targeted 600.9 “Prostatism” not otherwise specified (median lobe) (prior to 2000, A strength of our study design was that we were able to capture all resource use in our cohort during 5 years. Ex- 600.1 “Nodular prostate” excluding prostate cancer (prior to 2000, was cluding TURP failures treated with medication only, our 600.2 Benign localized hyperplasia (eg adenoma of prostate, adenofi- TURP failure rate is similar to that in prior reports. Our bramatous hypertrophy of prostate) of prostate, excludes hypertrophy failure rate for initial medical treatment is similar to that reported in the literature. The MTOPS trial described an If occurring with any of the procedure or BPH medication codes 2) Any male 40 years or older with
approximately 8% incidence of surgical therapy at 5 years 788.20 Urine retention or stasis not elsewhere classified for BPH in subjects initially treated with an ␣-adrenergic blocking as were most subjects in our study.
788.29 Other “specified” retention of urine788.41 Urinary frequency This concordance may reflect similar followup in the 2 stud- ies. Differences in failure rates in our study cohorts may 788.43 Nocturia788.61 Intermittent or splitting urinary stream reflect underlying clinical differences in this nonrandomized group since patients undergoing TURP as initial BPH treat- But not carrying diagnosis code 185 (prostate cancer) as another diagno- ment in the current decade likely present more frequently 3) Any male undergoing 1 of the following BPH related procedure
with severe disease. Adverse risk factors, such as detrusor fibrosis due to longstanding high pressure voiding, may be International Classification of Disease-9 procedure codes more prevalent in the surgical arm, clouding definitive con- 60.2 Transurethral prostatectomy60.21 Transurethral ultrasound guided laser induced prostatectomy clusions regarding failure rates between our study cohorts.
60.29 Other transurethral prostatectomy (transurethral electrovaporiza- Nevertheless, the failure rates in this analysis provide a tion of prostate, excision of median bar) portrait of existing resource use for BPH treatment.
60.4 Retropubic open prostectomy60.94 Control (postoperative) hemorrhage of prostate60.95 Balloon dilation of prostatic urethra Limitations
60.96 Transurethral microwave thermotherapy (added 2000, was 60.29) Our study is a retrospective analysis and it has the limita- 60.97 Other transurethral destruction of prostatic tissue (eg transure- tions common to such studies. Our cohorts were subject to thral needle ablation) (added 2000, was 60.29) Common Procedural Terminology procedure codes selection bias since subjects presenting with more severe symptoms, including those in acute urinary retention, were 52510 Transurethral balloon dilation of prostatic urethra most likely over represented in the cohort treated initially 52601 Transurethral electrosurgical resection of prostate52606 Transurethral fulguration of postoperative bleeding occurring af- with surgery. Subjects who were suboptimal surgical candi- dates were most likely over represented in the cohort ini- 52612 Transurethral resection of prostate, first stage of 2-stage resection52614 Second stage of 2-stage resection tially treated with medication. This may have decreased 52620 Transurethral resection of residual tissue after 90 postoperative expenditures in the medication arm by limiting the number of subjects who subsequently underwent surgical therapy.
52630 Of regrown tissue after 1 year post op52640 Of resultant bladder neck contracture We were unable to identify any cases in which minimally 52647 Noncontact laser coagulation of prostate invasive surgical modalities were used, such as transure- 52648 Contact laser vaporization with or without transurethral resection thral microwave thermotherapy. These therapies avoid the 53850 Transurethral destruction of prostate tissue by microwave ther- surgical costs derived from use of the hospital operating 53852 By radio frequency thermotherapy (added 1998) room, and so they may alter the costs of surgical therapy if 55801 Prostatectomy, perineal, subtotal55821 Prostatectomy, suprapubic, subtotal used widely. However, re-treatment rates for these technol- 55831 Prostatectomy, retropubic, subtotal ogies have yet to be clearly defined. We were unable to 4) Or any male taking 1 of the following medications
describe out of pocket costs for herbal medicines used by Hytrin®, Proscar®, terazosin, terazosin HCl, tamsulosin, dutasteride subjects for BPH treatment. Because our expenditure anal-ysis is from the point of view of the health care system, weonly present economic data. We do not consider the impact oftreatment on quality of life or indirect costs of care. We also Abbreviations and Acronyms
cannot account for a critical deciding factor in BPH treat- CONCLUSIONS
In a cohort of privately insured men with newly diagnosed REFERENCES
BPH monotherapy with ␣-blockers was the most common ini- Wei JT, Calhoun E and Jacobsen SJ: Urologic Diseases in America tial treatment. Surgical therapy was associated with higher project: benign prostatic hyperplasia. J Urol 2005; 173: 1256.
treatment failure rates and higher costs during 5 years. In- Holtgrewe HL: Current trends in management of men with creased expenditures related to initial surgical therapy were lower urinary tract symptoms and benign prostatic hyper- consistent when projected over long time frames.
plasia. Urology 1998; 51: 1.
ECONOMIC EVALUATION OF BENIGN PROSTATIC HYPERPLASIA TREATMENT STRATEGIES Lowe FC, McDaniel RL, Chmiel JJ and Hillman AL: Economic EDITORIAL COMMENT
modeling to assess the costs of treatment with finasteride,terazosin, and transurethral resection of the prostate for This is a retrospective analysis of administrative claims men with moderate to severe symptoms of benign prostatic from a privately insured cohort of men with BPH. The hyperplasia. Urology 1995; 46: 477.
authors conclude that initial surgical therapy for BPH Manyak MJ, Ackerman SJ, Blute ML, Rein AL, Buesterien K, was more costly and had a higher treatment failure rate Sullivan EM et al: Cost effectiveness of treatment for benign than initial medical management for BPH in the short- prostatic hyperplasia: an economic model for comparison ofmedical, minimally invasive, and surgical therapy. J En- term and projected long-term time frames. These findings dourol 2002; 16: 51.
are interesting and well presented. However, the data are McConnell JD, Roehrborn CG, Bautista OM, Andriole GL Jr, derived from a select cohort of men with private insur- Dixon CM, Kusek JW et al: The long-term effect of doxazo- ance, calling into question the ability to extrapolate these sin, finasteride, and combination therapy on the clinical findings to other cohorts of men, for example those with progression of benign prostatic hyperplasia. N Engl J Med Medicare or Medicaid coverage. Furthermore, stratifica- 2003; 349: 2387.
tion of provider specialty was needed in these data to Black L, Naslund MJ, Gilbert TD Jr, Davis EA and Ollendorf DA: An examination of treatment patterns and costs of care address potential treatment biases that may have existed among patients with benign prostatic hyperplasia. Am J among the different types of physician providers who sub- Manag Care 2006; 12: S99.
mitted these claims. Specific knowledge of provider spe- Floratos DL, Kiemeney LA, Rossi C, Kortmann BB, Debruyne cialty, particularly as it pertained to medical management FM and de La Rosette JJ: Long-term followup of randomized in this cohort, would allow the reader to more fully scru- transurethral microwave thermotherapy versus transure- tinize the clinical and economic conclusions reached by thral prostatic resection study. J Urol 2001; 165: 1533.
Riehmann M, Knes JM, Heisey D, Madsen PO and Bruskewitz RC: Transurethral resection versus incision of the prostate:
a randomized, prospective study. Urology 1995; 45: 768.
Chris M. Gonzalez
Hoffman RM, MacDonald R, Monga M and Wilt TJ: Transure- thral microwave thermotherapy vs transurethral resection for treating benign prostatic hyperplasia: a systematic re- view. BJU Int 2004; 94: 1031.

Source: http://update-udaonline.ctrl.ucla.edu/uda-publications/file/view/uda-publications/Economics_BPH_therapy_Saigal2007.pdf