Racial Differences in Cancer Specialist Consultation, Treatment, and Outcomes for Locoregional Pancreatic Adenocarcinoma

Ann Surg Oncol (2009) 16:2968–2977 DOI 10.1245/s10434-009-0656-5

ORIGINAL ARTICLE – HEALTHCARE POLICY AND OUTCOMES

Racial Differences in Cancer Specialist Consultation, Treatment, and Outcomes for Locoregional Pancreatic Adenocarcinoma Melissa M. Murphy, MD, MPH1, Jessica P. Simons, MD1, Sing Chau Ng, MS1, Theodore P. McDade, MD1, Jillian K. Smith, MD, MPH1, Shimul A. Shah, MD1, Zheng Zhou, MD, PhD1, Craig C. Earle, MD2, and Jennifer F. Tseng, MD, MPH1 1

Department of Surgery, Surgical Outcomes Analysis and Research, University of Massachusetts Medical School, Worcester, MA; 2Institute for Clinical Evaluative Sciences, Sunnybrook Health Sciences Centre, Toronto, ON, Canada

ABSTRACT Background. Blacks have a higher incidence of pancreatic adenocarcinoma and worse outcomes compared to whites. Identifying barriers in pancreatic cancer care may explain survival differences and provide areas for intervention. Methods. Pancreatic adenocarcinoma patients were identified in the Surveillance, Epidemiology, and End Results Registry (1991–2002). Treatment and outcome data were obtained from the linked Surveillance, Epidemiology, and End Results Registry–Medicare databases. Logistic regression was used to assess race as a predictor of specialist consultation/receipt of therapy. Kaplan–Meier survival curves were compared. Cox proportional hazard analyses were performed to estimate survival after adjustment for patient and treatment characteristics. Results. A total of 13,230 white patients (90%) and 1478 black patients (10%) were identified. Clinical/pathologic factors were compared by race. When we compared whites and blacks by univariate analyses, blacks had lower rates of specialist consultation (P \ .01), chemotherapy (P \ .01), and resection (P \ .01). On multivariate analyses predicting consultation with a cancer specialist, black race negatively predicted consultation with a medical oncologist (adjusted odds ratio [AOR] .74, P \ .01), radiation oncologist (AOR .75, P \ .01), and surgeon (AOR .71, P \ .01). For predicting receipt of therapy after consultation, blacks were less likely to undergo chemotherapy

Ó Society of Surgical Oncology 2009 First Received: 15 May 2009; Published Online: 11 August 2009 J. F. Tseng, MD, MPH e-mail: [email protected]

(AOR .59, P \ .01) and resection (AOR .79, P = .05). Blacks had worse overall survival on Kaplan–Meier survival curves (log rank, P \ .0001). On Cox proportional hazard modeling evaluating survival, black race was no longer independently associated with worse survival after adjustment for resection and adjuvant therapy (hazard ratio, 1.08; 95% confidence interval, .99–1.19). Conclusions. Racial disparities exist in pancreatic cancer specialist consultation and subsequent therapy use. Because receipt of care is fundamental to reducing outcome discrepancies, these barriers serve as discrete intervention points to ensure all locoregional pancreatic adenocarcinoma patients receive appropriate specialist referral and subsequent therapy.

Pancreatic adenocarcinoma continues to have a poor prognosis, with an overall 5-year survival of 5%.1 African Americans are disproportionately affected by pancreatic adenocarcinoma, as evidenced by the age-adjusted incidence rates of 11.7 per 100,000 for white patients and 15.6 per 100,000 for black patients, and race has been linked with poorer survival.2–4 Any opportunity for long-term survival in pancreatic cancer requires surgical resection; adjuvant therapy has been demonstrated to greatly improve outcomes for resected patients.5–10 Racial disparities have been described in the use of cancer therapies.11–15 Barriers may exist at many levels, including cancer specialist consultation and in subsequent receipt of therapy. Blacks have been reported to undergo surgery less often than whites for lung, esophageal, and colorectal cancers.14–18 For patients with pancreatic adenocarcinoma, our group has recently reported that blacks were far less likely to undergo pancreatic resection, despite similar rates of recommendation for surgery.19

Racial Differences in Pancreatic Cancer Care

In the current work, we used a two-step process to study disparities in cancer care.20,21 This approach evaluates both referral to specialists and treatment of referred patients. We used this methodology to determine whether racial differences exist in the rate of cancer specialist consultation (medical oncologist, radiation oncologist, or surgeon) or in the subsequent receipt of therapy (chemotherapy, radiotherapy, or resection) among patients with locoregional pancreatic adenocarcinoma. We also studied the effect of race on survival after adjustment for patient demographics and type of cancer therapy.

2969

METHODS

data in Medicare, we limited our cohort to patients aged 65 and older. We excluded patients for whom the date of death differed by [3 months between the SEER and Medicare databases, patients diagnosed from death certificate or autopsy, patients eligible for Medicare based on end-stage renal failure or disability, patients for whom the month of diagnosis was not available, and patients with health maintenance organization enrollment within 6 months after surgery, and those who did not have continuous Medicare enrollment during the study period and thus were likely to have important missing data.15 To prevent differences in the other multiple, smaller racial categories from obscuring the primary comparison in this investigation, we limited our analyses to black and white patients.

Patients and Data Sources

Patient Characteristics

We studied patients in the linked Surveillance, Epidemiology, and End Results Registry–Medicare databases (SEER-Medicare). The 2000–2002 diagnoses in the SEERMedicare linked data are from 16 SEER registries, approximating 26% of the U.S. population. The 1991–1999 diagnoses are from 12 SEER registries, which comprised approximately 14% of the U.S. population.22,23 Information contained within SEER includes patient demographics, disease-related data, and first-course of treatment. The SEER-Medicare database links SEER data to the Centers for Medicare and Medicaid Services Medicare claims database, as well as to census tract-level socioeconomic status (SES) information. Survival data are available through SEER via its linkage with the National Death Index, a centralized index of death records obtained through the state vital statistics offices. We identified pancreas as the organ site in SEER (site recode = 33), then merged the data with the linked Medicare claims files. We queried the merged SEERMedicare database for patients with pancreatic adenocarcinoma as defined by International Classification of Disease for Oncology, 3rd edition (ICD-O-3), codes (8021, 8140–8141, 8143–8144, 8211, 8230, 8255, 8260–8263, 8323, 8440–8441, 8450, 8453, 8470–8471, 8480–8481, 8500, 8503, 8521, 8550, 8560, 8570, 8576).24 Patients with pathologically confirmed pancreatic adenocarcinoma between January 1, 1991, and December 31, 2002, were identified. The mean and median age of patients diagnosed with pancreatic adenocarcinoma has been reported to range from 66 to 72 years of age (mean age, blacks 66 years and whites 69 years).3,4,19,25,26 Medicare includes complete information for patients C65 years of age or selected conditions (e.g., end-stage renal disease). Therefore, on the basis of the age distribution of patients diagnosed with pancreatic adenocarcinoma, coupled with availability of

We classified patients by age at diagnosis, sex, year of diagnosis, marital status, race, and SES. Stage of disease was based on ‘‘SEER historic stage,’’ which categorizes patient disease as localized, regional, distant, or unknown. After initial analyses demonstrated similarity between patients with localized and regional disease, we combined these patients in a composite category of ‘‘locoregional’’ disease. SES was determined by a consistent algorithm that uses a combination of median income and per-capita income in both the patient’s census tract and zip code using data from the 2000 census. First, median income in the patient’s census tract of residence was used. If this information was missing, zip code median income was substituted; if that was missing, census tract per-capita income was used; if that was missing, zip code–based percapita income was used. Finally, patients with all of the above data missing were classified in the lowest SES quintile. In total, 230 patients (1.6%) were classified in the lowest SES quintile as a result of missing data, including 205 white patients (1.6%) and 25 black patients (1.7%). These procedures have been well described and validated by Medicare research methodology.27 For dichotomous SES analyses, we defined low SES to be composed of the lowest three quintiles and high SES to include the highest two quintiles. Comorbidity was determined on the basis of inpatient and outpatient Medicare claims 13 months to 1 month before the date of diagnosis.28 Comorbidities were combined in the Charlson score.29,30 Of note, the presence of comorbidity could not reliably be determined for patients younger than 66 years plus 1 month, because Medicare claims were only available from age 65 years and older.28 Patients with no claims for comorbidities were assigned a Charlson score of 0. A total of 395 (2.7%) were age 65 at diagnosis, with a potential for missing comorbidity information that was based on the defined algorithm.

2970

Treatment Characteristics We identified patients seen by a cancer specialist by searching for claims from physicians with the following Medicare provider specialty codes (HCFA): medical oncologist (83, 90); radiation oncologist (32, 92); and surgeon (91, 02, 14, 19, 20, 24, 28, 33, 40, 77, 78, 85), as previously described.15 In addition, all patients who underwent pancreatectomy were by definition considered to have been seen by a surgeon. Information on radiotherapy was extracted from Medicare claims data.31 We developed an algorithm combining an in-hospital radiation indicator, International Classification of Diseases, 9th edition, Clinical Modification (ICD-9CM) diagnosis codes and procedure codes, with CPT codes, DRG codes, Berenson-Eggers type of service (BETOS) codes, and revenue center codes. An analogous procedure was performed for chemotherapy, with the exception of an in-hospital chemotherapy indicator. Pancreatic resection was determined by Medicare claims data. Resection was defined by the International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM), procedure codes for ‘‘radical pancreaticoduodenectomy,’’ (52.7) or Whipple procedure, ‘‘total pancreatectomy’’ (52.6), and ‘‘partial pancreatectomy’’ (52.5).32 In addition, we used Healthcare Common Procedure Coding System (HCPCS) codes for pancreatic resection (48140, 48145, 48146, 48150, 48152, 48153, 48154, 48155, 48160). Outcomes We analyzed survival as the time from the month of diagnosis until death from any cause. A censoring date was determined for patients last known to be alive on the basis of the last available Medicare coverage month.15 Statistical Analyses Patient demographics and associations between race and patient characteristics were compared by v2 tests for categorical variables and t-tests for continuous variables. We used logistic regression to calculate odds ratios for the effect of race on cancer specialist consultation (medical oncologist, radiation oncologist, or surgeon) and subsequent receipt of respective therapy (chemotherapy, radiotherapy, or pancreatic resection). Survival curves were constructed by using methods after Kaplan–Meier and compared by the log rank test.33 We compared survival between black and white patients overall and for those with locoregional disease. We then

M. M. Murphy et al.

assessed the effect of therapy (chemotherapy alone, radiotherapy alone, resection alone, resection plus adjuvant therapy [chemotherapy and/or radiation]) on survival for blacks and whites with locoregional pancreatic adenocarcinoma. Cox proportional hazard modeling was used to study racial differences in survival, expressed as hazard ratios (HRs). We created five models that sequentially adjusted for type of therapy, as follows: no therapy, chemotherapy alone, radiotherapy alone, resection alone, and resection plus any adjuvant therapy. All data analyses were performed by SAS statistical software, version 9.1.3 (SAS Institute, Cary, NC). Significance was accepted at the P \ .05 level. The research protocol was reviewed and determined to be exempt from institution review board oversight. RESULTS Patient Demographics Over the study period 1991–2002, a total of 14,708 patients were identified with pancreatic adenocarcinoma after applying the described exclusion criteria (Table 1). Racial breakdown included 13,230 white (90%) and 1478 black (10%) patients. Mean age was 76 years (95% confidence interval [95% CI], 75.9–76.1). Most patients were women (54.1%), married (54.1%), and had a Charlson score of 0 (59.4%). Most patients were diagnosed with distant disease (52.9%), followed by regional (28.4%) and localized disease (8.3%). Black patients tended to be slightly younger at diagnosis, with a mean age of 75.3 years, in comparison to white patients, who had a mean age of 76.1 years. Overall, blacks were more likely to be women (black 59% vs. white 54%, P \ .0001), to be unmarried (black 63% vs. white 44%, P \ .0001), and to reside in low-SES areas (black 89% vs. white 57%, P \ .0001). Blacks were more likely to seek care with localized disease than white patients (black 8% vs. white 10%, P \ .01). A comparison of clinical and pathologic factors by race for pancreatic adenocarcinoma patients with localized and regional disease, and a composite category of locoregional disease are shown in Table 2. Of patients with locoregional disease (37% of total, n = 5400), 91% of patients were white and 9% black. Blacks were more likely to be younger (mean age, black 75.1 years vs. white 75.9 years, P \ .001), female (black 61.0% vs. white 54.7%, \.01), and unmarried (black 55.1% vs. white 40.5%). Blacks had more comorbid disease (Charlson score, black C2, 21.5% vs. white C2, 14.4%, P \ .0001) and resided in low-SES areas (black 88.8% vs. white 43.6%, P \ .0001).

Racial Differences in Pancreatic Cancer Care

2971

TABLE 1 Clinical and pathologic factors of patients with pancreatic adenocarcinoma with univariate analyses by race, SEER-Medicare 1991–2002

to those with potentially resectable disease (locoregional pancreatic adenocarcinoma).

Factor

Overall, n (%)

White, n (%)

Black, n (%)

P

Receipt of Therapy

Total

14708 (100)

13230 (90)

1478 (10)

–

After consultation with a cancer specialist, receipt of subsequent respective therapy was evaluated by race. Blacks were significantly less likely to receive chemotherapy (black 47.9% vs. white 62.8%, P \ .0001). After consultation with a radiation oncologist, no difference in receipt of radiotherapy was observed by race (black 88.5% vs. white 83.6%, P = .14). After surgical consultation, blacks underwent pancreatic resection significantly less often than whites (black 34.2% vs. white 41.2%, P \ .01). Multivariate analyses were performed for patients with locoregional pancreatic disease after a specialist consultation that evaluated race as a predictor of receipt of respective therapy. Patient covariates of age, sex, marital status, SES, and Charlson score were adjusted for in the model. Black race persisted as a negative predictor of receiving chemotherapy after medical oncologist consultation (AOR, black .59 vs. white; 95% CI, .45–.77). Race did not affect receipt of radiotherapy after radiation oncologist consultation (AOR, black 1.63 vs. white; 95% CI, .93–2.88). Blacks undergoing surgical consultation continued to be less likely to undergo resection compared with whites, with an adjusted point estimate similar to unadjusted, although the confidence interval reached the limit of statistical significance (AOR, black .79 vs. white; 95% CI, .63–1.0).

Age at diagnosis (y) Mean (95% CI)

76.0 (75.9–76.1)

76.1 (75.9–76.2)

75.3 \.0001 (75–75.6)

65–70

3807 (26)

3351 (25)

456 (31)

71–75

3963 (27)

3536 (27)

427 (29)

C76

6938 (47)

6343 (48)

595 (40)

Male

6747 (46)

6138 (46)

609 (41)

Female

7961 (54)

7092 (54)

869 (59)

7962 (54)

7412 (56)

550 (37)

6746 (46)

5818 (44)

928 (63)

8741 (60)

7989 (60)

752 (51)

Sex

Married Yes No

\.0001

\.0001

Charlson score 0 1

3706 (25)

3293 (25)

413 (28)

C2

2261 (15)

1948 (15)

313 (21)

\.0001

Socioeconomic status Low

8823 (60)

7503 (57)

1320 (89)

High

5885 (40)

5727 (43)

158 (11)

\.0001

SEER historic stage Localized

1221 (9)

1079 (8)

142 (10)

Regional

4179 (28)

3813 (29)

366 (25)

Distant

7783 (53)

6973 (53)

810 (55)

Unstaged

1525 (10)

1365 (10)

160 (10)

\.01

SEER, surveillance, epidemiology, and end results program; 95% CI, 95% confidence interval

Specialist Consultation Consultation with a cancer specialist for patients with locoregional pancreatic adenocarcinoma was evaluated by race (Fig. 1). Blacks were significantly less likely than whites to see a medical oncologist (black 52.6% vs. white 60.2%, P \ .001), radiation oncologist (black 25.6% vs. white 32.5%, P \ .05), or a surgeon (black 72.1% vs. white 78.0%, P \ .01). Multivariate analyses were performed that evaluated race as a predictor of cancer specialist consultation for patients with locoregional pancreatic adenocarcinoma (Table 3). After adjustment for patient age, sex, marital status, SES, and Charlson score, black race persisted as a negative predictor for consultation with a medical oncologist (adjusted odds ratio [AOR], black .74 vs. white; 95% CI, .61–.90), radiation oncologist (AOR, black .75 vs. white; 95% CI, .60–.93), and surgeon (AOR, black .71 vs. white; 95% CI, .57–.88). Stage or severity of disease was accounted for by limiting analyses

Survival Univariate Kaplan–Meier analyses performed on all patients in the identified pancreatic adenocarcinoma cohort demonstrated blacks to have worse overall median survival (blacks 3 months vs. whites 3.9 months, log rank P \ .0001) (Table 4). Black patients with locoregional disease demonstrated a persistent survival disadvantage (locoregional disease median survival, blacks 6 months vs. whites 7.7 months, log rank P \ .0001) (Fig. 2). After chemotherapy or radiotherapy, blacks continued to have worse survival compared with whites (chemotherapy median survival, black 10.3 months vs. white 10.9 months, log rank P \ .05; radiotherapy median survival, black 9.6 months vs. white 10.9 months, log rank P \ .05). After pancreatic resection, no significant difference in survival between black and white patients with locoregional disease was detected (resection median survival, black 12.8 months vs. white 14.9 months, log rank P = .55) (Fig. 3). Multivariate Cox proportional hazard modeling was performed for locoregional pancreatic adenocarcinoma patients using the covariates of race, age, sex, marital

2972

M. M. Murphy et al.

TABLE 2 Comparison of clinical and pathologic factors by race for pancreatic adenocarcinoma patients with localized, regional, and composite locoregional disease, with associated univariate analysis, SEER-Medicare 1991–2002 Factor

Localized (n = 1221)

Regional (n = 4179)

White, n (%)

Black, n (%)

1079 (88)

142 (12)

65–70 71–75

217 (20) 266 (25)

41 (29) 36 (25)

C76

596 (55)

65 (46)

Male

485 (45)

52 (37)

Female

594 (55)

90 (64)

Yes

573 (53)

59 (42)

No

506 (47)

83 (58)

0

607 (56)

71 (50)

1

297 (28)

35 (25)

C2

175 (16)

36 (25)

Total

P

White, n (%)

Black, n (%)

3813 (91)

366 (9)

1065 (28) 1062 (28)

120 (33) 112 (31)

1686 (44)

134 (36)

1733 (45)

146 (40)

2080 (55)

220 (60)

2211 (58)

143 (39)

1602 (42)

223 (61)

2372 (62)

185 (50)

912 (24)

108 (30)

529 (14)

73 (20)

2134 (56)

331 (90)

1679 (44)

35 (10)

Locoregional (n = 5400) P

White, n (%)

Black, n (%)

4892 (91)

508 (9)

1282 (26) 1328 (27)

161 (32) 148 (29)

2282 (47)

199 (39)

2218 (45)

198 (39)

2674 (55)

310 (61)

2784 (57)

202 (40)

2108 (43)

306 (60)

2979 (61)

256 (50)

1209 (25)

143 (28)

704 (14)

109 (22)

2133 (44)

451 (89)

2759 (56)

57 (11)

P

Age (y) \.05

\.05

\.01

Sex .06

\.05

\.01

Married \.01

\.01

\.01

Charlson score \.05

\.001

\.01

Socioeconomic status Low

625 (58)

120 (85)

High

454 (42)

22 (15)

\.001

\.001

\.01

SEER, surveillance, epidemiology, and end results program FIG. 1 Locoregional pancreatic adenocarcinoma patients with associated univariate analyses by race of cancer specialist consultation and subsequent receipt of therapy

Locoregional pancreatic adenocarcinoma White: n = 4892 Black: n = 508 Consultation

Medical Oncologist White: 60.2%, n = 2944 Black: 52.6%, n = 267 p < 0.001

Radiation Oncologist White: 32.5%, n = 1588 Black: 25.6%, n = 130 p < 0.05

Surgeon White: 78.0%, n = 3815 Black: 72.1%, n = 366 p < 0.01

Therapy after Consultation

Chemotherapy White: 62.8%, n = 1850 Black: 47.9%, n = 128 p < 0.001

Radiation White: 83.6%, n = 1327 Black: 88.5%, n = 115 p = 0.14

Resection White: 41.2%, n = 1573 Black: 34.2%, n = 125 p < 0.01

status, SES, and Charlson score. We adjusted the model for type of therapy received with results shown in Table 5. Advanced age, male sex, marital status of unmarried, higher Charlson score, and low SES were independent predictors of worse survival in all models. Black race, adjusted for patient, tumor, and societal characteristics including age, sex, stage, comorbidities, and SES, predicted worse survival for locoregional pancreatic adenocarcinoma patients (HR, 1.14; 95% CI, 1.04–1.25). The negative association of black race persisted after adjustment for chemotherapy alone (HR, 1.11; 95% CI, 1.01–1.23), radiotherapy alone (HR, 1.13; 95% CI 1.03– 1.24), and surgical resection alone (HR, 1.11; 95% CI, 1.01–1.22). However, after adjustment for resection plus

any adjuvant therapy, black patients had a statistically equivalent survival compared with whites (HR, 1.08; 95% CI, .99–1.19). DISCUSSION In the current study, we examined whether racial disparities exist in cancer care after a diagnosis of locoregional pancreatic adenocarcinoma. We studied whether barriers exist at the level of cancer specialist consultation or in receipt of subsequent therapy. We found blacks were statistically significantly less likely to be seen by medical oncologists, radiation oncologists, and surgeons compared with white patients. This difference persisted

Racial Differences in Pancreatic Cancer Care

2973

TABLE 3 Univariate and multivariate analyses of race as a predictor of cancer specialist consultation and subsequent receipt of therapy among patients with locoregional pancreatic adenocarcinoma Factor

Multivariate analysesa

Univariate analyses OR

95% CI

P

AOR

95% CI

P

Race White

Referent

Referent

Black Cancer specialist consultation Medical oncologist

0.73

0.61–0.88

\.01

0.74

0.61–0.90

\.01

Radiation oncologist

0.72

0.58–0.88

\.01

0.75

0.60–0.93

\.01

Surgeon

0.73

0.59–0.89

\.01

0.71

0.57–0.88

\.01 \.01

Receipt of therapy after consultation Chemotherapy

0.55

0.42–0.70

\.01

0.59

0.45–0.77

Radiotherapy

1.51

0.87–2.63

.15

1.63

0.93–2.88

.09

Resection

0.74

0.59–0.93

\.01

0.79

0.63–1.0

.05

OR, odds ratio; 95% CI, 95% confidence interval; AOR, adjusted odds ratio a

Multivariate analyses controlling for patient age, sex, marital status, socioeconomic status, and Charlson score

TABLE 4 Univariate survival analyses for patients with pancreatic adenocarcinoma patients by race Factor

Median survival (mo)

Log rank P

White Black p < 0.0001

80

White

Black

All patients

3.9

3.0

\.0001

Locoregional disease

7.7

6.0

\.0001

? Chemotherapy

10.9

10.3

\.05

? Radiotherapy

10.9

9.6

\.05

? Resection

14.9

12.8

.55

? Resection ? adjuvant therapya 17.0

16.8

.64

a

Survival (Percent) 100

Chemotherapy and/or radiotherapy

after adjustment for covariates including age, sex, marital status, SES, and patient comorbidities. After consultation, blacks were significantly less likely to receive chemotherapy, while no significant difference in receipt of radiotherapy was observed. After surgical consultation, black race was a negative predictor of undergoing pancreatic resection on univariate analyses, while on multivariate analyses, the point estimate was similar but reached the limit of statistical significance. Overall, blacks had worse survival. However, if both pancreatic resection and receipt of adjuvant therapy (chemotherapy and/or radiotherapy) were included in the model, black race no longer predicted a survival disadvantage. We and others have shown black patients undergo surgery for pancreatic adenocarcinoma at lower rates than white patients.19 Additionally, blacks have been reported to receive less adjuvant therapy. By using the National Cancer Database, Bilimoria et al. identified black race to negatively predict receipt of surgery for patients with pancreatic

60 40 20

0

20

60

80

100

175 13

91 11

56 8

Months

Number at Risk 4,892 508

40

915 72

376 24

FIG. 2 Kaplan–Meier survival curves comparing all identified black and white patients with locoregional pancreatic adenocarcinoma

cancer.25 In a state-based registry, Eloubeidi et al. reported black patients with pancreatic adenocarcinoma were less likely to receive either adjuvant therapy or surgical resection compared with whites.34 Investigating treatment of locally advanced pancreatic adenocarcinoma patients, Krzyzanowska et al. found advanced age and low SES to be associated with decreased use of adjuvant therapy, although the effect of race was not evaluated.35 Recently, Davila et al. reported black race, age C75 years, and a higher number of patient comorbidities correlated with lower use of adjuvant therapy after resection for pancreatic cancer.36

2974

M. M. Murphy et al.

Survival (Percent) 100

White Black p = 0.55

80 60 40 20 0

20

60

80

100

153 13

75 10

45 8

Months

Number at Risk 1,573 125

40

630 46

308 23

FIG. 3 Kaplan–Meier survival curves comparing black and white patients with locoregional pancreatic adenocarcinoma after pancreatic resection

Although several studies have observed that African American patients undergo pancreatic resection less often and receive less adjuvant therapy than whites, no published work has systematically investigated the steps toward these observed disparities. Our current work found two potentially modifiable steps at which barriers exist. First, a barrier may exist at the level of consultation with a cancer specialist. Disparities in cancer specialist consultation have been described for other cancers, including esophageal and lung cancer.14,20 Our work is the first to demonstrate that blacks with locoregional pancreatic adenocarcinoma are statistically significantly less likely to see a cancer specialist than whites. The explanation for this observation is not readily available. Referral patterns may contribute to this disparity. Advanced age has been shown to affect referral patterns for cancer screening and treatment.37,38 Although black patients with pancreatic cancer were younger than whites in the current study, it is possible that race may have an analogous effect on referrals for pancreatic adenocarcinoma patients. Conversely, patients may have received a referral but not attended as a result of personal reasons (beliefs, fear, culture), logistic reasons (day care, transportation, work), or specialist availability (limited schedules, few specialists).39–44 Insurance coverage has been reported to affect referrals, although among our cohort, all patients had Medicare insurance, making this a less likely contributor to the observed disparity.45 The second barrier we identified in cancer care for pancreatic adenocarcinoma patients lies in receipt of therapy after consultation. We found blacks received a lower percentage of both chemotherapy and pancreatic resection

after specialist consultation. The individual contributors to this discrepancy are unclear. In other cancers, barriers in physician-patient communication and inadequate access to culturally sensitive materials have been reported to affect treatment strategies.46,47 Individual patient beliefs and practices have been proposed to influence treatment acceptance.48 And for some cancers, black race has been associated with a higher rate of refusing cancer-directed therapy.34 Differences in physician recommendations have been reported for operations including kidney transplantation and cardiac procedures, and they may exist in recommendation for pancreatic resection.49,50 As a result of database limitations, we were unable to evaluate individual patient-physician discussions or assess individual treatment recommendations. Interestingly, there was no difference in receipt of radiotherapy by race after consultation. Standard externalbeam radiotherapy is typically schedule intensive, with several weeks of daily therapy, requiring much logistical planning. One possibility to explain the variance between radiotherapy and other treatment modalities is that once a patient has actually seen a radiation oncologist, the impetus on the part of the clinician and the patient to at least initiate radiotherapy may be greater. It is also possible that radiotherapy may be viewed differently than chemotherapy among the African American community. Differences in communication, patient education, and treatment recommendations may vary between radiation oncologists and medical oncologists. Jemal et al. recently published national cancer statistics from the SEER Registry from 1975 to 2003.1 The authors report 5-year relative survival rates for patients with all types of pancreatic cancer to be greatly improving over time for all races, with whites and blacks having equivalent 5-year relative survival rates of 5% during the most recent time period (1996–2003). Our current work includes patients with pancreatic adenocarcinoma, rather than all types of pancreatic cancer, with analyses limited to those with locoregional disease; therefore, the two cohorts are not directly comparable. Without adjustment for treatment, we identified black race to be an independent predictor of worse survival. After adjusting for pancreatic resection and adjuvant therapy, black race no longer conferred a survival disadvantage on Cox proportional hazard analyses. These results suggest that survival for black patients with pancreatic adenocarcinoma may be improved if they are provided recommended therapy, including resection and adjuvant therapy. Additionally, because blacks had a higher likelihood of being diagnosed with localized disease, more patients may be amenable to pancreatic resection. Therefore, because the only way for patients to undergo pancreatic resection is after consultation with a surgeon, it may be inferred that improving surgical

Racial Differences in Pancreatic Cancer Care

2975

TABLE 5 Cox proportional multivariate survival analyses with incremental adjustment for the effect of therapy on survival for patients with locoregional pancreatic adenocarcinoma Factor

No therapy

Chemotherapy

Radiotherapy

Resection

Resection ? adjuvant therapy

White

Referent

Referent

Referent

Referent

Referent

Black

1.14 (1.04–1.25)

1.11 (1.01–1.23)

1.13 (1.03–1.24)

1.11 (1.01–1.22)

1.08 (0.99–1.19)

65–70

Referent

Referent

Referent

Referent

Referent

71–75

1.07 (0.99–1.15)

1.04 (0.96–1.12)

1.03 (0.96–1.11)

1.02 (0.95–1.10)

0.98 (0.91–1.06)

1.42 (1.33–1.52)

1.31 (1.22–1.40)

1.32 (1.24–1.42)

1.25 (1.16–1.33)

1.13 (1.05–1.21)

Male

Referent

Referent

Referent

Referent

Referent

Female

0.94 (0.89–0.99)

0.93 (0.88–0.99)

0.94 (0.89–0.99)

0.93 (0.87–0.98)

0.92 (0.87–0.98)

Yes

Referent

Referent

Referent

Referent

Referent

No

1.16 (1.09–1.23)

1.13 (1.06–1.20)

1.13 (1.07–1.20)

1.13 (1.07–1.20)

1.09 (1.02–1.16)

0

Referent

Referent

Referent

Referent

Referent

1

1.13 (1.06–1.20)

1.13 (1.06–1.21)

1.13 (1.06–1.21)

1.10 (1.03–1.18)

1.11 (1.04–1.18)

C2

1.35 (1.25–1.46)

1.35 (1.25–1.46)

1.34 (1.24–1.45)

1.28 (1.18–1.38)

1.26 (1.17–1.37)

Race

Age (y)

C76 Sex

Married

Charlson score

Socioeconomic status Low

1.11 (1.05–1.18)

1.10 (1.04–1.16)

1.11 (1.04–1.17)

1.09 (1.03–1.16)

1.06 (1.01–1.13)

High

Referent

Referent

Referent

Referent

Referent

– –

1.46 (1.38–1.55) Referent

– –

– –

– –

None

–

–

1.42 (1.34–1.51)

–

–

Any

–

–

Referent

–

–

Chemotherapy None Any Radiotherapy

Adjuvant therapya None

–

–

–

–

1.61 (1.53–1.71)

Any

–

–

–

–

Referent

No

–

–

–

2.44 (2.29–2.60)

2.63 (2.46–2.80)

Yes

–

–

–

Referent

Referent

Resection

Data are expressed as hazard ratio (95% confidence interval) a

Chemotherapy and/or radiotherapy

consultation rates for blacks (and all patients) might allow more patients to be considered for resection and thus potentially improve overall survival. SEER-Medicare provides a wealth of information regarding cancer cases at the national level, including information on patient demographics, disease characteristics, and treatment information. However, there are some limitations to the use of this type of observational and administrative data. The causality of race on treatment and outcome cannot be proven. The data do not readily capture all medical factors relevant to pancreatic adenocarcinoma patients, including patient performance status, laboratory values (e.g., CA 19-9), specific operative course

information, and whether patients completed their adjuvant therapy course. Medicare data are limited to patients C65 years of age or selected conditions. However, because pancreatic adenocarcinoma is primarily a disease of the elderly—as evidenced by the median age of patients diagnosed in the United States with pancreatic cancer between 1975 and 2006 being 72 years (median age, whites 72 years and blacks 68 years)—most patients with pancreatic adenocarcinoma are captured in our cohort. SES is measured at the census tract level instead of the individual level, making direct inferences regarding SES limited. We were only able to capture those comorbidities serious enough to result in the use of medical services

2976

before diagnosis, which likely underestimates the true presence of coexistent disease.28 Cancer specialist consultation was assessed if a claim for services was present in the Medicare database; however, consultations may have been missed if a bill was never submitted or recorded. Despite these limitations, the linked SEER-Medicare database provides a mechanism to study population-based cancer in a diverse patient population. We have identified two potentially modifiable points at which barriers to care exist for black patients with locoregional pancreatic adenocarcinoma: cancer specialist consultation and receipt of subsequent therapy. Additional work needs to be performed to further characterize these barriers and to allow intervention strategies to be designed, implemented, and evaluated. Disparities in cancer care for African American patients with locoregional pancreatic adenocarcinoma clearly exist. For a disease with historically poor survival, racial differences in the availability and delivery of care may have a substantive negative effect on outcomes for minority patients. Conversely, providing recommended, evidencebased care, including access to specialists and subsequent therapies to all patients with pancreatic adenocarcinoma, may have benefits, including improved survival. Because receipt of guideline-directed care is fundamental to reducing outcome discrepancies, ensuring that all patients have access to recommended pancreatic cancer care components may assist in closing the racial disparity gap. ACKNOWLEDGMENT We thank Bridget A. Neville, Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, for her statistical expertise. This work was supported by the Evans-Allen-Griffin Fellowship (M.M.M.), the Pancreatic Cancer Alliance (J.P.S. and J.F.T.), the American Surgical Association Foundation, the Howard Hughes Early Career Award, and an American Cancer Society Institutional Research Grant (all to J.F.T.).

REFERENCES 1. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2008. CA Cancer J Clin. 2008;58:71–96. 2. National Cancer Institute. SEER cancer statistics review 2005: age-adjusted SEER incidence and US death rates and 5-year relative survival rates. http://seer.cancer.gov/csr/1975_2005/. Accessed 4 Aug 2009. 3. Horner M, Ries LAG, Krapcho M, et al. SEER cancer statistics review, 1975–2006. Bethesda, MD: National Cancer Institute; 2009. http://seer.cancer.gov/csr/1975_2006. Accessed 4 Aug 2009. 4. Fesinmeyer MD, Austin MA, Li CI, De Roos AJ, Bowen DJ. Differences in survival by histologic type of pancreatic cancer. Cancer Epidemiol Biomarkers Prev. 2005;14:1766–73. 5. Warshaw AL, Fernandez-del Castillo C. Pancreatic carcinoma. N Engl J Med. 1992;326:455–65. 6. Bertagnolli MM. Surgical prevention of cancer. J Clin Oncol. 2005;23:324–32. 7. Kalser MH, Ellenberg SS. Pancreatic cancer. Adjuvant combined radiation and chemotherapy following curative resection. Arch Surg. 1985;120:899–903.

M. M. Murphy et al. 8. Neoptolemos JP, Dunn JA, Stocken DD, et al. Adjuvant chemoradiotherapy and chemotherapy in resectable pancreatic cancer: a randomised controlled trial. Lancet. 2001;358(9293): 1576–85. 9. Chua YJ, Cunningham D. Adjuvant treatment for resectable pancreatic cancer. J Clin Oncol. 2005;23:4532–7. 10. Oettle H, Post S, Neuhaus P, et al. Adjuvant chemotherapy with gemcitabine vs observation in patients undergoing curative-intent resection of pancreatic cancer: a randomized controlled trial. JAMA. 2007;297:267–77. 11. Shavers VL, Brown ML. Racial and ethnic disparities in the receipt of cancer treatment. J Natl Cancer Inst. 2002;94:334–57. 12. Ayanian JZ, Zaslavsky AM, Fuchs CS, et al. Use of adjuvant chemotherapy and radiation therapy for colorectal cancer in a population-based cohort. J Clin Oncol. 2003;21:1293–300. 13. O’Malley CD, Le GM, Glaser SL, Shema SJ, West DW. Socioeconomic status and breast carcinoma survival in four racial/ethnic groups: a population-based study. Cancer. 2003;97:1303–11. 14. Steyerberg EW, Neville B, Weeks JC, Earle CC. Referral patterns, treatment choices, and outcomes in locoregional esophageal cancer: a population-based analysis of elderly patients. J Clin Oncol. 2007;25:2389–96. 15. Steyerberg EW, Earle CC, Neville BA, Weeks JC. Racial differences in surgical evaluation, treatment, and outcome of locoregional esophageal cancer: a population-based analysis of elderly patients. J Clin Oncol. 2005;23:510–7. 16. McCann J, Artinian V, Duhaime L, et al. Evaluation of the causes for racial disparity in surgical treatment of early stage lung cancer. Chest. 2005;128:3440–6. 17. Bach PB, Cramer LD, Warren JL, Begg CB. Racial differences in the treatment of early-stage lung cancer. N Engl J Med. 1999;341:1198–205. 18. Rogers SO, Ray WA, Smalley WE. A population-based study of survival among elderly persons diagnosed with colorectal cancer: does race matter if all are insured? (United States). Cancer Causes Control. 2004;15:193–9. 19. Murphy MM, Simons JP, Hill J, Tseng JF. Pancreatic resection: a key component to reducing racial disparities in pancreatic adenocarcinoma. Cancer (2009). [Epub ahead of print]. 20. Earle CC, Neumann PJ, Gelber RD, Weinstein MC, Weeks JC. Impact of referral patterns on the use of chemotherapy for lung cancer. J Clin Oncol 2002;20:1786–92. 21. Tseng JF, Kronowitz SJ, Sun CC, et al. The effect of ethnicity on immediate reconstruction rates after mastectomy for breast cancer. Cancer 2004;101:1514–23. 22. Warren JL, Klabunde CN, Schrag D, Bach PB, Riley GF. Overview of the SEER-Medicare data: content, research applications, and generalizability to the United States elderly population. Med Care. 2002;40 Suppl 8:IV-3–18. 23. Potosky AL, Riley GF, Lubitz JD, Mentnech RM, Kessler LG. Potential for cancer related health services research using a linked Medicare-tumor registry database. Med Care. 1993;31:732–48. 24. World Health Organization. International classification of diseases for oncology. 3rd ed. Geneva: World Health Organization; 2000. 25. Bilimoria KY, Bentrem DJ, Ko CY, et al. National failure to operate on early stage pancreatic cancer. Ann Surg 2007;246: 173–80. 26. Zell JA, Rhee JM, Ziogas A, Lipkin SM, Anton-Culver H. Race, socioeconomic status, treatment, and survival time among pancreatic cancer cases in California. Cancer Epidemiol Biomarkers Prev. 2007;16:546–52. 27. Bach PB, Guadagnoli E, Schrag D, Schussler N, Warren JL. Patient demographic and socioeconomic characteristics in the SEER-Medicare database applications and limitations. Med Care. 2002;40 Suppl 8:IV-19–25.

Racial Differences in Pancreatic Cancer Care 28. Klabunde CN, Warren JL, Legler JM. Assessing comorbidity using claims data: an overview. Med Care. 2002;40 Suppl 8: IV-26–35. 29. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40: 373–83. 30. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–9. 31. Virnig BA, Warren JL, Cooper GS, et al. Studying radiation therapy using SEER-Medicare-linked data. Med Care. 2002;40 Suppl 8:IV-49–54. 32. International classification of diseases. 9th rev., clinical modification. Salt Lake City: Medicode Publications; 2001. 33. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assn. 1958;185:1457–81. 34. Eloubeidi MA, Desmond RA, Wilcox CM, et al. Prognostic factors for survival in pancreatic cancer: a population-based study. Am J Surg. 2006;192:322–9. 35. Krzyzanowska MK, Weeks JC, Earle CC. Treatment of locally advanced pancreatic cancer in the real world: population-based practices and effectiveness. J Clin Oncol. 2003;21:3409–14. 36. Davila JA, Chiao EY, Hasche JC, et al. Utilization and determinants of adjuvant therapy among older patients who receive curative surgery for pancreatic cancer. Pancreas 2009;38:e18–25. 37. Sewitch MJ, Fournier C, Dawes M, et al. Do physician recommendations for colorectal cancer screening differ by patient age? Can J Gastroenterol. 2007;21:435–8. 38. Samet J, Hunt WC, Key C, Humble CG, Goodwin JS. Choice of cancer therapy varies with age of patient. JAMA. 1986;255:3385– 90. 39. Rathore SS, Krumholz HM. Differences, disparities, and biases: clarifying racial variations in health care use. Ann Intern Med. 2004;141:635–8.

2977 40. Dein S. Explanatory models of and attitudes towards cancer in different cultures. Lancet Oncol. 2004;5:119–24. 41. Whittle J, Conigliaro J, Good CB, Joswiak M. Do patient preferences contribute to racial differences in cardiovascular procedure use? J Gen Intern Med. 1997;12:267–73. 42. Oddone EZ, Horner RD, Diers T, et al. Understanding racial variation in the use of carotid endarterectomy: the role of aversion to surgery. J Natl Med Assoc. 1998;90:25–33. 43. Lannon C, Brack V, Stuart J, et al. What mothers say about why poor children fall behind on immunizations. A summary of focus groups in North Carolina. Arch Pediatr Adolesc Med. 1995;149:1070–5. 44. Bach PB, Pham HH, Schrag D, Tate RC, Hargraves JL. Primary care physicians who treat blacks and whites. N Engl J Med. 2004;351:575–84. 45. Ayanian JZ, Kohler BA, Abe T, Epstein AM. The relation between health insurance coverage and clinical outcomes among women with breast cancer. N Engl J Med. 1993;329:326–31. 46. Perez-Stable EJ, Sabogal F, Otero-Sabogal R, Hiatt RA, McPhee SJ. Misconceptions about cancer among Latinos and Anglos. JAMA. 1992;268:3219–23. 47. Fox SA, Stein JA. The effect of physician-patient communication on mammography utilization by different ethnic groups. Med Care. 1991;29:1065–82. 48. Womeodu RJ, Bailey JE. Barriers to cancer screening. Med Clin North Am. 1996;80:115–33. 49. Ayanian JZ, Cleary PD, Keogh JH, et al. Physicians’ beliefs about racial differences in referral for renal transplantation. Am J Kidney Dis. 2004;43:350–7. 50. Ibrahim SA, Whittle J, Bean-Mayberry B, et al. Racial/ethnic variations in physician recommendations for cardiac revascularization. Am J Public Health. 2003;93:1689–93.