Race and Time to Receipt of Androgen Deprivation Therapy Among Men With Metastatic Prostate Cancer

INTRODUCTION

Prostate cancer represents the most common non-cutaneous malignancy for men, where in 2017 an estimated 161,360 new cases of prostate cancer were diagnosed in the United States.¹ Prostate cancer is the second-leading cause of cancer death in men.² African-American men are 2.4 times more likely to die from prostate cancer than Caucasian men.^1,3 The differential prostate cancer mortality rates in African-American compared to Caucasian men represent the largest disparity of any malignancy.^4,5 One of the contributing factors to higher mortality rates in African-American men are due to the differences in treatment received. Potentially curative therapies are less often recommended to African-American men.⁶

Androgen deprivation therapy (ADT) with a gonadotropin-releasing hormone or bilateral orchiectomy has been the mainstay of treatment for men with advanced prostate cancer.^7,8 The use of ADT has increased greatly since the early 1990s among American men with non-metastatic prostate cancer,^7,9,10 but few data are available about patterns of ADT in men with metastatic prostate cancer.⁷ Further, the Medicare Modernization Act (MMA) in 2005 drastically reduced the reimbursement for ADT,^11–13 and prior studies showed that this policy reduced ADT use.^11,12

Given the policy changes in ADT reimbursement by the MMA, it becomes increasingly important to assess racial disparities in timely receiving of ADT.¹⁴ Before changes in the ADT reimbursement policy by Medicare, African-American men with metastatic prostate cancer were significantly less likely to receive ADT, and when treated, had a slightly longer time to receipt than Caucasian men.¹⁴ One unintended consequence of the MMA may be an increase in the racial disparities in receipt of this treatment. Thus, the purpose of this study is to evaluate the associations between race and median time to receive ADT among men with metastatic prostate cancer before and after the MMA in 2005. We hypothesized that African-American men with metastatic prostate cancer are more likely to have longer median time to receive ADT following diagnosis, compared to Caucasian men after the implementation of MMA. To our knowledge, this is the first study to assess impact of race on receiving ADT treatment in metastatic cancer since the changes in Medicare ADT reimbursement.

MATERIALS AND METHODS

RESULTS

Table 1 summarizes the demographic and clinical characteristics of the analytic cohort before and after the MMA reimbursement policy changes for ADT, along with the p-values. Before the MMA, there were 1028 African-American and 5288 Caucasian men with metastatic prostate cancer in the cohort. The cohort after the MMA included 818 African-American and 4174 Caucasian men with metastatic prostate cancer. The cohorts, before and after the implementation of MMA had similar characteristics.

Overall, most men in both study cohorts had a Charlson Comorbidity score of 0, lived in areas of low educational level, had poorly differentiated tumors, and were from an urban region. Compared to Caucasian men, African-American men were younger, less likely to be married, less educated, more likely to reside in the South urban region, and having higher percentage of poorly differentiated/undifferentiated tumors. These differences between the cohorts were found to be statistically significant (P < 0.05) before and after the implementation of MMA. Since, the racial cohorts were somewhat dissimilar at baseline, an accelerated failure time regression model was used to control these differences to draw an empirical conclusion about the racial effect of receipt of ADT.

The median time to receipt of ADT for both Caucasian and African-American men increased after the MMA. Before the MMA, the median time to receipt of ADT for Caucasian men was 95 days compared to 172 days for African American men (P < 0.05). However, after the MMA, the median time to receipt of ADT for Caucasian men was 109 days compared to 205 days for African-American men (P < 0.05). Figures 2 and 3 presents the racial differences in the probability of receipt of ADT as a function of days since diagnosis for the men in the pre-MMA and the post-MMA cohort, respectively. African-American men were less likely compared to Caucasian men to receive ADT after diagnosis (P < 0.05).

Table 2 presents the results of an accelerated failure time regression model to determine the association between race and time to receipt of ADT, along with other covariates before and after the MMA. After adjusting for other covariates, African-American men had a longer median time to receipt of ADT both before the MMA (Time Ratio (TR): 1.15; 95% Confidence Interval (CI) [1.05, 1.27]) and after the MMA (TR: 1.29; 95% CI [1.10, 1.53]) as compared to Caucasian men. Unmarried men had significantly longer median time to receipt of ADT both before and after the MMA (TR: 1.23, 1.18; 95% CI [1.02, 1.42; 1.05, 1.33] respectively) compared to men who were married. Men in the age group of 70–74 years had significantly shorter median time to receipt of ADT both before and after the MMA (TR: 0.76, 0.84; 95% CI [0.56,0.97; 0.72,0.98] respectively) as compared to men aged 66–69 years.

Further, men residing in the South had longer median time to receipt of ADT (TR: 1.26, 1.52; 95% CI [1.01, 1.52; 1.24, 1.87] before and after MMA, respectively) compared to the Northeast region. Before the MMA, men diagnosed with cancer in 2000, 2001, 2002, and 2003 had significantly longer median time to receipt of ADT as compared to those diagnosed in 2004 (TR: 2.76, 2.13, 1.47, 1.28; 95% CI [2.13,3.42; 1.89, 2.36; 1.23, 1.74; 1.05, 1.56] respectively). After the MMA, men diagnosed with cancer in 2006, 2007, 2008, 2009, and 2010 had significantly longer median time to receipt of ADT as compared to those diagnosed in 2011 (TR: 4.36, 3.55, 3.37, 2.45, 1.62; 95% CI [4.15, 4.96; 2.85,4.34; 2.90,3.74; 2.03, 2.97; 1.36,1.91] respectively). In addition, compared to the men with well/moderately differentiated tumors, men with poorly differentiated tumors had shorter median time to receipt of ADT (TR: 0.73, 0.78; 95% CI [0.52, 0.94; 0.58, 0.96] before and after the MMA respectively).

DISCUSSION

Racial disparities exist in the receipt of ADT among men with metastatic prostate cancer, after the implementation of the MMA. Interestingly, the MMA substantially changed the drug reimbursement policy of Medicare Part B, and was intended to reduce non-evidence based ADT use (defined as clinically localized, stage T1–T2 patients with low- to intermediate-grade prostate cancer).¹² In men with metastatic prostate cancer, use of ADT is considered primary therapy.²⁰ However, we found that African-American men had significantly longer median time to receipt of evidence based ADT as compared to Caucasians. Therefore, one unintended consequence of the MMA may be an increase in the racial disparities in receipt of ADT. Similar findings were reported by Carson and colleagues, who assessed trends and factors associated with time to receive ADT in men with metastatic prostate cancer.¹⁴ The authors found that African-American men with metastatic prostate cancer were significantly less likely to receive ADT and, when treated, had a slightly longer time to receipt than Caucasian men. However, this study was conducted on the SEER cancer registry with Medicare claims data from 1991 to 2002, which was before the changes in Medicare reimbursement policies for ADT in 2005.¹⁴ The existence of statistically significant racial differences in the receipt of ADT for metastatic prostate cancer after policy changes in ADT reimbursement by the MMA indicates that there could be factors other than race that may be associated with the differences in receipt of ADT.

Our study also explored other clinical or socio-demographic factors that may impact the receipt of ADT. We found that unmarried men had significantly longer median time to receipt of ADT both before and after the MMA compared to men who were married. This might be due to difficulty in treatment decision-making or lack of social support.⁷ Men residing in the South had longer median time to receipt of ADT compared to men in the Northeast region. Further, before the MMA, men diagnosed with cancer in 2000, 2001, 2002, and 2003 had significantly longer median time to receipt of ADT as compared to those diagnosed in 2004. Whereas, after the MMA, men diagnosed with cancer in 2006, 2007, 2008, 2009, and 2010 had significantly longer median time to receipt of ADT as compared to those diagnosed in 2011. These findings might be due to reduction in an inappropriate use of ADT over time after the changes in Medicare reimbursement policies, particularly among men for whom the benefits of such therapy were unclear.²¹

Specifically, in 2004, reimbursement was generally reduced to 85% of the average wholesale price (and to 85% of the average price for gonadotropin-releasing hormone agonists).²¹ Whereas, in 2005 reimbursement was set a 106% of the average sale price (in contrast to the average wholesale price) based on actual transactions reported quarterly by pharmaceutical companies.²¹ These changes represented a reduction of 50% or more in the reimbursement for gonadotropin-releasing hormone agonists.²¹ Our findings are in line with another study reporting on the use of ADT for metastatic prostate cancer in older men; specifically, unmarried men or those residing in certain geographic region in the United States were less likely to receive ADT.⁷

There are several limitations of this study. Our study was based on Medicare claims from SEER-Medicare linked database. Hence, our findings may not be generalizable to other patient populations such as Medicaid, Health Maintenance Organizations, or other insurance plans. In addition, SEER-Medicare captures information from men who are above 65 years of age; hence our findings may not be generalizable to those younger than 65 years of age. There might be a possibility of selection bias due to non-randomized nature of the study. Also, much of differences in the characteristics of the analytic cohort before and after MMA reimbursement policy may be due to a truncation bias. Since we restricted the initial analytic cohort to patients who received ADT; therefore, the later cohorts had a shorter median waiting times (because the longer median waiting times which could have been observed, was not since the data was truncated). In traditional survival analysis, this is handled by including all cases and keeping track of which cases were censored. However, because we restricted the analytic cohort to patients who received ADT, there is no censoring in our data set.

Another limitation is that we were unable to assess certain factors, such as patient’s preferences or provider recommendations which could have been a driver in treatment choices. Further, the estimates for time until receipt of ADT are, on average, over-estimated by 15-days. This is because diagnosis dates are recorded only to the month, and arbitrarily assigned to the first day. However, this over-estimation does not impact the comparison between racial groups (i.e., Caucasian vs. African-American), since both groups had the same bias. Lastly, while the estimated time ratios decrease monotonically for men who received ADT for metastatic prostate cancer from 2006 to 2011; it is likely that prostate cancer diagnosis has improved over time and the health system has become more efficient. Despite these limitations, our study enhances the existing literature on racial differences in ADT treatment initiation in men with metastatic prostate cancer after changes in Medicare ADT reimbursement.

CONCLUSIONS

Considering the palliative benefits of ADT in reducing pain or other side-effects,²² and aggressiveness of cancer in African American men,²³ timely receipt of ADT among this population may be important. Further research is needed to assess factors beyond policy changes for ADT reimbursement that could contribute racial differences in ADT treatment for metastatic prostate cancer.