Prostate cancer mortality in Connecticut, Iowa and New Mexico African American men

Cancer Detection and Prevention 28 (2004) 375–380 www.elsevier.com/locate/cdp

Prostate cancer mortality in Connecticut, Iowa and New Mexico African American men Luis G. Escobedo MD, MPHa,*, Saul D. Rivas BCHb, Michelle D. Holmes MD, DrPHc a

Public Health Division, New Mexico Department of Health, 1170 North Solano, Suite L, Las Cruces, NM 88001, USA b Department of Health Sciences, New Mexico State University, Las Cruces, NM, USA c Harvard Medical School and Brigham and Women’s Hospital, Channing Laboratory, Boston, MA, USA Received 25 May 2004; accepted 16 June 2004

Abstract We sought to assess trends in prostate cancer incidence, treatment and mortality in African American men by means of analysis of prostate cancer data from three states, Connecticut, Iowa and New Mexico, all participants in the Surveillance, Epidemiology, and End Results (SEER) Program. Compared with levels before prostate specific antigen (PSA) testing, prostate cancer incidence increased in all three states after widespread testing. For men diagnosed with localized or regional prostate cancer, the respective increases in radical prostatectomy in Connecticut, Iowa, and New Mexico were 3.2, 2.3, and 4.9 times pre-test levels. Age-standardized mortality in Connecticut and Iowa increased slightly; in New Mexico the 104.7 deaths per 100,000 in 1979–1986, 62.1 in 1987–1990, dropped to 47.6 in 1991–1998, an amount of decline that was statistically significant. Introduction of PSA testing influenced early detection and treatment of prostate cancer in all three states. Although decline in prostate cancer mortality in New Mexico over time may be linked with use of the PSA test and definitive therapy, the relationship among these factors, and thus the proper treatment for the early stages of this condition, is unclear on the basis of these data. # 2004 International Society for Preventive Oncology. Published by Elsevier Ltd. All rights reserved. Keywords: Mortality; Incidence; African American; Prostatic neoplasms; Mass screening

1. Introduction For men in the United States, prostate cancer is a leading cause of cancer death [1]. In 1997, age-adjusted rate for this condition per 100,000 men was 136 for incidence and 33.8 for mortality [2,3]. The burden of this disease is far greater for African American than it is for other men. In 1997, ageadjusted incidence and mortality per 100,000 African American men was 207.9 and 71.1, respectively [2,3]. Because this disease tends to occur in older men, who are likely to die of other conditions, it is often unclear whether definitive treatment should be pursued. In fact, an acceptable treatment option consists of observation alone [4–6]. However, the tumor can be aggressive and run a relentless

* Corresponding author. Tel.: +1 505 528 5153; fax: +1 505 528 6024. E-mail address: [email protected] (L.G. Escobedo).

course, underscoring the importance of prevention, early diagnosis, and appropriate therapy [4,5]. Introduction of the prostate specific antigen (PSA) has been associated with large increases in incidence, increases in early stage of disease, declines in advanced stages [7], and a slight initial increase and subsequent decrease in mortality [3]. As use of this test to screen for prostate cancer has become widespread, so have concerns about subjecting men who test positive to an array of diagnostic and treatment procedures [8–10]. These procedures represent a burden on the medical care system [9,11]. In the United States, of available health statistics, cancerrelated incidence, stage at diagnosis, and mortality contribute large amounts to known racial disparities in health status [12–14]. White men present at an earlier stage of prostate cancer than African American men do. From 1989 through 1995, the stage distribution for white versus African American men (80% versus 72%) for localized/

0361-090X/$30.00 # 2004 International Society for Preventive Oncology. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.cdp.2004.06.003

376

L.G. Escobedo et al. / Cancer Detection and Prevention 28 (2004) 375–380

regional disease and (8% versus 13%) for distant disease [15] clearly reflect this pattern. The five-year prostate cancer survival for white versus African American men (94% versus 86%) for those under 65 years of age and (94% versus 83%) for those older than 65 years [15] are also compelling. Later stage at presentation for African American men probably explains a large part of the racial differences in survival, although biologic factors probably play some role; we know that African American men generally present with more aggressive cancer grades than whites [16,17]. In studies across a variety of settings, African American men are less likely than white men to receive definitive treatment for prostate cancer [18–21]. We assessed trends in prostate cancer incidence, treatment and mortality in African American men in relation to introduction of the PSA test by means of analysis of data from ongoing cancer surveillance programs in selected regions of the country. Our purpose was to highlight issues of priority for future research in the light of uncertainties concerning early detection and treatment for localized prostate cancer.

2. Methods 2.1. Data sources Since 1973, the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Program has gathered data about newly diagnosed cancers in nine separate sites in the country [15]. These sites are in five states (Connecticut, Iowa, New Mexico, Utah, and Hawaii) and four metropolitan areas (Atlanta, Georgia; Detroit, Michigan; Seattle, Washington; and San Francisco and Oakland, California). Registry abstractors visit hospitals, clinics, laboratories and treatment facilities to identify newly diagnosed cancer cases and gather data about all forms of invasive cancer. Data collected from medical records, pathology reports, and other documents include characteristics of the patient, type of cancer, treatment, and outcome [22]. Mortality data presented in this report were compiled by The National Center for Health Statistics (NCHS), and available at http://wonder.cdc.gov/mortICD9J.html. The database provides state-specific counts and rates from 1979 through 1998 for underlying cause-of-death by age, race/ethnicity and year. We present number of deaths and rates for males ages 25 years and older, collapsed as one age group. These mortality data originate from states and conform to Certificate of Death procedures for completion of certificates as recommended by the World Health Organization (WHO) and the NCHS. Procedures for ascertainment of cause of death include coding of underlying cause by NCHS-trained nosologists using WHO’s International Classification of Diseases (ICD), ninth revision.

2.2. Definitions Cancer data conformed to the third edition of the International Classification of Diseases of Oncology (ICDO-3) rubrics C61.9 for men diagnosed with prostate cancer. Mortality data conformed to version 9 of the International Classification of Disease, (ICD-9) rubrics 185 for prostate cancer. Use of the PSA as a screening test for prostate cancer began in 1988 [11]. For most analyses in this report, the years between 1973 and 1988 represent the period before use of the PSA and years from 1989 through 1998 the period of use. Age groups were created to examine prostate cancer patterns in younger men (less than 54 years), middle-aged (55–64 years) and older men (65–74 years and 75 years and older). Population denominator data for African American men that have been available from SEER [20] make possible calculation of incidence rates for the entire surveillance period. The SEER Program uses the historic stages of in situ (confined to epithelial layer with intact basement membrane), localized (confined to organ of origin), regional (presence of disease in proximal nodes or adjacent structures), and distant disease (distant lymph nodes or viscera, such as bone, brain, liver, lung) to classify cancer stages [15]. Since 1995 the SEER Program has combined localized and regional stages in one category. 2.3. Data analysis In assessment of trends in medical services and outcomes, SEER data from states are more useful than that from metropolitan areas. Since the 1940s, population growth in the United States has shifted from urban centers to surrounding areas [23]. Because these population shifts have had a stronger influence on denominator data for urban areas than for states, state data describe disease trends more consistently over time. Further, cancer patients often receive care outside their local community [24], and so state data probably provide more reliable information than local data about health services patterns for a defined resident population over time. Of the five states participating in the SEER Program, only Connecticut, Iowa and New Mexico reported sufficient number of deaths for African American men needed to estimate mortality rates. Denominator data used represent U.S. Bureau of the Census population estimates and inter-census estimates for the respective state for each year. African American-specific incidence rates in all three states were stratified by age and cancer stage for two surveillance periods, one before use of PSA tests, 1973–1988, and one to coincide with use of this test, 1989–1998. Radical prostatectomy, beam radiation, and their combination are generally considered definitive treatment for prostate cancer. Data about initial treatment were gathered

L.G. Escobedo et al. / Cancer Detection and Prevention 28 (2004) 375–380

from 1983 through 1997. We examined data about procedures associated with localized or regional disease during the earlier years of this surveillance period, since we were interested in assessment of survival subsequent to detection. Transurethral resection of the prostate gland is not considered treatment for localized prostate cancer; however because this procedure is used frequently and it likely resulted in many new diagnoses of prostate cancer, we decided to also examine trends in this procedure. The two periods were 1983 through 1988 (before screening was introduced) and 1989 through 1994 (during the early years of screening). Accordingly, procedures assessed included prostatectomy, transurethral resection of prostate, beam radiation, prostatectomy plus radiation, observation, and other forms of treatment for prostate cancer. We assessed African American-specific prostate cancer mortality for men in the three states for the period before (1979–1986), during (1987–1990) and after introduction of the PSA test (1991–1998). Prostate cancer incidence and mortality rates were adjusted to the age distribution of the 1970 United States population by the direct method of standardization. Calculation of 95% confidence intervals for incidence rates

377

is based on standard methods [22]; for age-standardized mortality, calculation of this interval is based on the sum of variability of age-specific rates [25]. The 95% confidence interval represents the range of values that would include the true value 95 times if 100 measurements were made, with limits set by the likelihood of chance [26]. Thus, when establishing level of statistical significance for differences of values between subgroups, lack of 95% confidence interval overlap for the respective values is equivalent to statistical significance at a level of P at 0.05.

3. Results Table 1 compares prostate cancer incidence rates for African American men by state before (1973–1988) and during (1989–1998) widespread use of the PSA test. After the introduction of this test, incidence rates nearly doubled. In fact, annual incidence data for all states show dramatic increases beginning in 1988 (data not shown). The least overall increase in incidence was observed for New Mexico. The largest increases in incidence were observed for younger age groups and those with localized or regional stage. The

Table 1 Prostate cancer incidence rates by age group and stage for two time periods, Connecticut, Iowa, and New Mexico African American men, 1973–1998, SEER Program Characteristic and period

Connecticuta Counts

Age (years) 1973–1988

1989–1998

Stage 1973–1988

1989–1998

Overall 1973–1988 1989–1998 a

Iowab

Incidence rated

<54 55–64 65–74 >75

40 211 367 243

3.8 210.3 721.5 1157.3

<54 55–64 65–74 >75

120 456 715 353

13.7 567.2 1450.0 1495.2

Localized/regional Distant Unstaged

493 287 80

63.0 36.0 10.9

Localized/regional Distant Unstaged

1272 195 175

158.1 26.1 23.5

Before PSA During PSA

861 1644

109.9 207.8

Rate ratioe

Counts

New Mexicoc Incidence rated

7 38 76 87

4.0 211.7 654.2 1275.1

19 71 112 77

13.0 512.1 1253.6 1589.9

125 69 14

66.3 36.7 7.5

2.51f 0.73f 2.16f

206 40 33

141.1 28.6 24.1

1.89f

208 279

110.6 193.8

3.61f 2.70f 2.01f 1.29f

Rate ratioe

Counts

Incidence rated

Rate ratioe

4 29 35 45

3.5 257.7 475.5 1195.4

8 34 85 42

6.7 295.0 1093.7 978.3

78 26 9

69.4 22.6 8.4

2.13f 0.78 3.21f

145 18 6

115.9 15.2 5.0

1.67f 0.67 0.60

1.75f

113 169

100.4 136.1

1.36

3.25 2.42f 1.92f 1.25

1.91 1.14 2.30f 0.82

In 1973–1988, out of 16,594 total cases, 15,676 white and 57 persons of other race/ethnicity were excluded, leaving 861. In 1989–1998, out of 23,954 total cases, 21,868 white and 442 persons of other race/ethnicity were excluded, leaving 1644. b In 1973–1988, out of 19,200 total cases, 18,978 white and 14 persons of other race/ethnicity were excluded, leaving 208. In 1989–1998, out of 21,944 total cases, 21,416 white and 249 persons of other race/ethnicity excluded, leaving 279. c In 1973–1988, out of 6693 total cases, 6365 white and 215 persons of race/ethnicity were excluded, leaving 113. In 1989–1998, out of 10,440 total cases, 10,024 white and 247 persons of other race/ethnicity were excluded, leaving 169. d Rates are cases per 100,000 and age-adjusted to the 1970 U.S. standard. e Incidence rate in 1989–1998 divided by incidence rate in 1973–1988. f Confidence interval for incidence rate in 1989–1998 does not overlap interval for rate in 1973–1988.

378

L.G. Escobedo et al. / Cancer Detection and Prevention 28 (2004) 375–380

Table 2 Prostate cancer treatment strategies in African American men diagnosed with localized or regional disease, 1983–1988 and 1989–1994, Connecticut, Iowa, and New Mexico, SEER Program Procedure

Connecticut 1983–1988

Prostatectomy TURPc Radiation only Prostatectomy plus radiation Observation All otherd

Iowa 1989–1994

a

New Mexico

1983–1988 a

Count

%

Count

%

Ratio

16 131 0 2 4 89

6.6 54.1 0.0 0.8 1.7 36.8

136 124 4 17 8 349

21.3 19.4 0.6 2.7 1.3 54.7

3.2 0.4 1 3.4 0.8 1.5

b

1989–1994 a

1983–1988 a

Count

%

Count

%

Ratio

6 39 0 0 1 8

11.1 72.2 0.0 0.0 1.9 14.8

24 15 0 9 1 47

25.0 15.6 0.0 9.4 1.0 49.0

2.3 0.2 0.0 1 0.5 3.3

b

1989–1994 a

Count

%

Count

%a

Ratiob

2 19 0 1 0 12

5.9 55.9 0.0 2.9 0.0 35.3

23 12 0 6 1 38

28.8 15.0 0.0 7.5 1.3 47.5

4.9 0.3 0.0 2.6 1 1.3

a

% denotes percent type of treatment among patients. Treatment percent in 1989–1994 divided by treatment percent in 1983–1988. c TURP denotes transurethral resection of the prostate. d All other denotes diagnostic categories, non-cancer directed surgery, subtotal/simple prostatectomy (including those not specified), pelvic surgery, node surgery, and other categories. Other categories include a few radiation therapies such as radioactive implants, radioisotopes, and otherwise not specified. b

Table 3 Prostate cancer age-adjusted mortality rates in African American for three time periods, 1979–1998, Connecticut, Iowa, and New Mexico, CDC Wonder Program Period

Counts

Crude ratea (95% CI)b

Connecticut 1979–1986 1987–1990 1991–1998

167 102 296

39.2 (33.5–45.6) 38.2 (31.2–46.4) 49.1 (43.6–55.0)

78.2 (65.9–90.6) 81.2 (65.1–97.3) 93.1 (82.3–103.8)

1.04 1.19

Iowa 1979–1986 1987–1990 1991–1998

41 31 57

53.0 (38.1–71.9) 69.8 (47.4-99.1) 52.7 (39.9–68.2)

79.5 (55.1–103.9) 111.1 (71.9–150.3) 93.5 (69.1–117.9)

1.40 1.18

New Mexico 1979–1986 1987–1990 1991–1998

34 12 27

63.8 (44.2–89.2) 35.6 (18.4–62.2) 29.7 (19.5–43.2)d

104.7 (69.4–140.0) 62.1 (26.9–97.4) 47.6 (29.6–65.5)d

0.59 0.45

a b c d

Age-adjusted ratea (95% CI)b

Rate ratioc

Rates are cases per 100,000; rates adjusted to the 1970 U.S. population. 95% CI denotes 95% confidence interval. Adjusted mortality rate in corresponding time period divided by adjusted mortality rate in 1979–1986. 95% confidence interval for mortality rate in this time period does not overlap interval for rate in 1979–1986.

exception to this pattern was that in New Mexico the greatest increase in incidence occurred in 65–74 year olds. Table 2 reports procedures associated with prostate cancer for African American men with localized or regional disease. When procedures are examined before and during the period of PSA testing, men in all three states received increasing amounts of definitive therapy, particularly radical prostatectomy. In New Mexico, men treated with prostatectomy increased nearly five-fold, an increase greater than that for men in the other two states. The number of men who received beam radiation was too few to assess trends. Table 3 compares crude and age-standardized prostate cancer mortality rates in African American men between the two time periods for the three populations. Whereas mortality increased slightly over time in Connecticut and Iowa, mortality decreased substantially in New Mexico. In this last group, mortality declined during 1987–1990, and with further follow up during 1991–1998 this amount of decline reached statistical significance.

4. Discussion Recent declines in prostate cancer metastatic disease and mortality in the United States may be the result of use of the PSA as a screening test [27,28], or to statistical artifacts such as lead-time and length-time bias associated with screening [9]. Because screening can result in earlier diagnosis of prostate cancer, the longer period of observation can appear to improve survival even when early diagnosis and treatment have no effect on the subsequent date of death (lead-time bias) [29]. Persons with indolent cancer are more likely than persons with more aggressive cancer grades to be detected through screening programs, because of slow evolution of disease and longer survival (length-time bias) [29]. In the SEER Program, it is difficult to untangle bias from screening effects because of the lack of a formal screening program in each of the states, that is, we do not know who was and who was not screened. Still, health outcomes for a population do emerge following introduction of a test intended to detect

L.G. Escobedo et al. / Cancer Detection and Prevention 28 (2004) 375–380

new cases. If after use of the PSA test the initial population survival advantage is short-lived, then this pattern may be consistent with lead-time and length-time bias. If on the other hand, the initial survival advantage is sustained over time, as it was for New Mexico, this pattern may be more likely to be consistent with an effect such as that seen following screening. Although medical care access [17], quality of medical care [30], and medical therapies [18,19], probably play a role in the risk of death from prostate cancer, data about these relationships are unavailable. In the present study, changes in prostate cancer treatment patterns emerged following widespread use of PSA testing. Faced with PSA and diagnostic test results, the initial treatment strategy chosen by men with early stage prostate cancer and their doctors probably signals an agreement to a defined course of treatment perceived to be the best available [31]. If so, the growing trend toward radical prostatectomy and beam radiotherapy indicates greater efforts to seek cure. Mortality data found in this study, which may result from these various screening and treatment decisions reached individually, may have also been observed elsewhere. In Olmsted County, Minnesota [32] and in Tyrol County, Austria [33] PSA screening may have played a role in reduction of prostate cancer mortality, although the influence of non-screening factors cannot be ruled out in these studies. Other investigators argue that the diminishing gap of the most significant predictor of prostate cancer mortality—stage at diagnosis—between white and African American men can probably be attributed to PSA screening [34]. Although the SEER Program is a valuable resource to measure cancer burden, lack of data not routinely gathered limits other applications. The SEER Program does not gather data about socio-economic status, and so it is difficult to assess cancer burden in relation to social class. In its earlier years, SEER gathered data to measure extent of disease using a classification system that has given way to more accurate methods [35]. Small sample sizes for key treatment categories, especially those considered curative, greatly limit our ability to assess impact of treatment patterns over time. Despite statistically significant declines in mortality in New Mexico, we are unable to make broad generalizations regarding underlying factors, other than to indicate that a temporal relationship with use of the PSA test exists. This limitation, especially in New Mexico, may be due to few deaths and lack of data regarding measures about the quality of cancer care. Although use of definitive therapies in the relatively few specialized treatment centers in the largest city of New Mexico (Albuquerque) may have influenced mortality patterns in that state over time, we are unable to assess the role that quality of cancer care may have played in this or the other two states. As future studies gather data about quality of prostate cancer care, for example experience of the surgeon and hospital staff [36], investigators will be in a stronger position to evaluate findings from treatment studies [37].

379

We are uncertain about factors associated with differences in mortality trends in the three states. However, whereas adoption of the PSA test was greater in Connecticut and Iowa than in New Mexico, adoption of radical prostatectomy was greater in New Mexico. More limited use of the PSA test in New Mexico may reflect greater barriers to access a regular source of primary medical care in this state with a larger land mass, longer travel distances, and larger percent of persons who lack medical insurance than in the other two states [38]. The effects of the PSA test may have evolved along different time frames in the respective states. As incidence of localized or regional disease increased to greater levels in Connecticut and Iowa than in New Mexico, the number of newly diagnosed men known to be at risk of death would have had to have been greater in the two states. Because of the slightly more rapid uptake in use of definitive treatment in New Mexico than in the other two states, it is possible that the combined effect of the PSA test and aggressive treatment favored a more rapid mortality decline in New Mexico than Connecticut and Iowa. Despite the seemingly paradoxical data for Connecticut and Iowa, greater use of definitive therapy may have attenuated the natural tendency of mortality to have increased even further in the face of a surge in case detection. Moreover, as the impact of medical interventions matures so they are able to exert their full effect, we may see sustained drops in mortality in these two states. Although investigators have indicated that PSA screening had no influence on prostate cancer mortality in Canada and Europe [39,40], we must await data from randomized trials, and hope that methodological difficulties for doing such studies can be overcome. If, as data in this report show, communities tend to adopt varying patterns of definitive treatment following implementation of early detection programs, data about the quality of care becomes even more important. In the meantime and because of ambiguity regarding PSA screening-related health benefits and uncertainty regarding treatment efficacy for localized prostate cancer, medical providers and patients alike must continue to wrestle with questions about whether to screen for prostate cancer and what treatment makes the most sense for localized disease. In addition to measurement of mortality as health outcome, future studies that seek to evaluate screening effects also need to measure quality of life, in the light of known morbidity associated with definitive treatment [5,10].

References [1] Wingo PA, Cardinez CJ, Landis SH, et al. Long-term trends in cancer mortality in the United States, 1930–1998. Cancer 2003;97(Suppl. 12):3133–275. [2] Eberhardt MS, Ingram DD, Makuc DM. Health, United States, 2001. Urban and Rural Health Chartbook. Hyattsville (MD): National Center for Health Statistics; 2001 [Table 55].

380

L.G. Escobedo et al. / Cancer Detection and Prevention 28 (2004) 375–380

[3] Eberhardt MS, Ingram DD, Makuc DM. Urban and Rural Health Chartbook. Health, United States, 2001. Hyattsville (MD): National Center for Health Statistics; 2001 [Table 30]. [4] Gee WF, Hotgrewe HL, Blute ML, et al. American urological association gallup survey: changes in diagnosis and management of prostate cancer and benign prostatic hyperplasia, and other practice trends from 1994 to 1997. J Urol 1998;160:1804–7. [5] Middleton RG, Thompson IM, Austenfeld MS, et al. Prostate cancer clinical guidelines panel summary report on the management of clinically localized prostate cancer. J Urol 1995;154:2144–8. [6] Mettlin CJ, Murphy GP, Rosenthal DS, Menck HR. The national cancer data base report on prostate carcinoma after the peak in incidence rates in the U.S.. Cancer 1998;83:1679–84. [7] Hankey BF, Feuer EJ, Clegg LX, et al. Cancer surveillance series: interpreting trends in prostate cancer—Part I: evidence of the effects of screening in recent prostate cancer incidence, mortality, and survival rates. JNCI 1999;91:1017–24. [8] Lu-Yao GL, Greenberg ER. Changes in prostate cancer incidence and treatment in USA. Lancet 1994;343:251–4. [9] Brawley OW. Prostate carcinoma incidence and patient mortality. Cancer 1997;80:1857–963. [10] Mettlin CJ, Murphy GP, Sylvester J, McKee RF, Morrow M, Winchester DP. Results of hospital cancer registry surveys by the American College of Surgeons: outcomes of prostate cancer treatment by radical prostatectomy. Cancer 1997;80:1875–81. [11] Potosky AL, Miller BA, Albertsen PC, Kramer BS. The role of increasing detection in the rising incidence of prostate cancer. JAMA 1995;273:548–52. [12] Howard J, Hankey BF, Greenberg RS, et al. A collaborative study of differences in the survival rates of black patients and white patients with cancer. Cancer 1992;69:2349–60. [13] Bradkey CJ, Given CW, Roberts C. Disparities in cancer diagnosis and survival. Cancer 2001;91:178–88. [14] McDonald CJ. The American Cancer Society addressing disparities and the disproportionate burden of cancer. Cancer Suppl 2001;91: 195–8. [15] Ries LAG, Kosary CL, Hankey BF, Miller BA, Clegg LX, Edwards BK, editors. SEER Cancer Statistics Review, 1973–1996: Tables and Graphs. Bethesda (MD): National Cancer Institute; 1999. p. 422. [16] Fowler Jr JE, Bigler SA. A prospective study of the serum prostate specific antigen concentrations and gleason histologic scores of Black and White men with prostate carcinoma. Cancer 1999;86:836–41. [17] Robbins AS, Whittenmore AS, Van Den Eeden SK. Race, prostate cancer survival, and membership in a large health maintenance organization. JNCI 1998;90:986–90. [18] Mettlin CJ, Muphy GP, Cunningham MP, Menck HR. The national cancer data base report on race, age, and region variations in prostate cancer treatment. Cancer 1997;80:1261–6. [19] Harlan L, Brawley O, Pommerenke F, Wali P, Kramer B. Geographic, age, and racial variation in the treatment of local/regional carcinoma of the prostate. J Clin Oncol 1995;13:93–100. [20] Underwood W, De Monner S, Ubel P, Fagerlin A, Sanda MG, Wei JT. Racial/ethnic disparities in the treatment of localized/regional prostate cancer. J Urol 2004;171:1504–7.

[21] Klabunde CN, Potosky AL, Harlan LC, Kramer BS. Trends and black/ white differences in treatment for nonmetastatic prostate cancer. Med Care 1998;1337–48. [22] National Cancer Institute. A Guide to Using SEER*Stat, version 3.0. Bethesda, Maryland. [23] Elliott JR. Cycles within the system: metropolitanisation and internal migration in the US, 1965–90. Urban Stud 1997;34:21–41. [24] Guidry JJ, Aday LA, Zhang D, Winn RJ. Cost considerations as potential barriers to cancer treatment. Cancer Pract 1998;6:182–7. [25] Breslow NE, Day NE. Statistical methods in cancer research. The design and analysis of cohort studies, 59. Lyon, France: International Agency for Research on Cancer; 1987. p. 59 [volume II]. [26] Rothman KJ, Greenland S. Modern Epidemiology. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 1998. [27] Smart CR. The results of prostate carcinoma screening in the US as reflected in the surveillance, epidemiology, and end results program. Cancer 1997;80:1835–44. [28] Mettlin CJ, Murphy GP. Why is the prostate cancer death rate declining in the United States? Cancer 1998;82:249–51. [29] Hennekens CH, Buring JE. Epidemiology in Medicine. 1st ed. Boston/Toronto: Little, Brown and Company; 1987. [30] Jacobs EA, Lauderdale DS. Receipt of cancer screening procedures among Hispanic and non-Hispanic health maintenance organization members. Cancer 2001;91:257–61. [31] Hall JD, Boyd JC, Lippert MC, Theodorescu D. Why patients choose prostatectomy or brachytherapy for localized prostate cancer: results of a descriptive survey. Urology 2003;61:107–402. [32] Roberts RO, Bergstralh EJ, Katusic SK, Lieber MM, Jacobsen SJ. Decline in prostate cancer mortality from 1980 to 1997, and an update on incidence trends in Olmsted County, Minnesota. J Urol 1999;161:529–33. [33] Bartsch G, Horninger W, Klocker H, et al. Prostate cancer mortality after introduction of prostate-specific antigen mass screening in the Federal State of Tyrol, Austria. Urology 2001;58:417–24. [34] Bianco Jr JF, Wood Jr DP, Grignon DJ, Sakr WA, Pontes JE, Powell IJ. Prostate cancer stage shift has eliminated the gap in disease-free survival in black and white American men after radical prostatectomy. J Urol 2002;168:479–82. [35] Montie JH. Staging of prostate cancer: current TNM classification and future prospects for prognostic factors. Cancer 1995;75:1814–8. [36] Begg CB, Riedel ER, Bach PB, et al. Variations in morbidity after radical prostatectomy. N Engl J Med 2002;346:1138–44. [37] 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;325:740–3. [38] Centers for disease control and prevention. State specific prevalence estimates of uninsured and underinsured persons—behavioral risk factor surveillance system, 1995. MMWR 1998;47:51–5. [39] Perron L, Moore L, Bairati I, Beranrd PM, Meyer F. PSA screening and prostate cancer mortality. CMAJ 2002;166:586–91. [40] Oliver SE, Gunnell D, Donovan JL. Comparison of trends in prostatecancer mortality in England and Wales and the USA. Lancet 2000;355:1788–9.