Differences in racial/ethnic disparities in patient care experiences between prostate cancer survivors and males without cancer: A SEER-CAHPS study

Abstract

Introduction: Our purpose was to evaluate whether racial/ethnic disparities in patient care experiences (PCEs) differ between males with prostate cancer ("PCa group") and males without cancer ("non-cancer group").

Materials and methods: This retrospective study used 2007-2015 National Cancer Institute's Surveillance, Epidemiology and End Results registry data linked to Consumer Assessment of Healthcare Providers and Systems surveys. PCa and non-cancer groups were propensity score matched 1:5 on demographic and clinical characteristics. Differences in racial/ethnic disparities (DRD) (non-Hispanic Black [NHB], Hispanic, non-Hispanic Asian [NHA], and Other Races compared to non-Hispanic White [NHW]) in PCEs (getting needed care, getting care quickly, doctor communication, customer service, and getting needed prescription drugs) were compared between matched PCa and non-cancer groups. Per prior literature, DRD in PCE scores were categorized as small (<3), medium (≥3 but <5) or large (≥5).

Results: There were 7312 males in the PCa group and 36,559 matched males in the non-cancer group. Within each group, all racial/ethnic minority categories reported worse scores compared to NHW individuals (p < 0.05) for ≥3 PCE composite measures. Compared to PCa group, a larger NHA-NHW difference was observed in non-cancer group for getting needed care (-4.65 in PCa vs. -7.77 in non-cancer; DRD = 3.11, p = 0.029) and doctor communication (-2.46 in PCa vs. -4.85 in non-cancer; DRD = 2.38, p = 0.023).

Discussion: In both PCa and non-cancer groups, racial/ethnic minorities reported worse experiences compared to NHW individuals for several PCE measures. However, the difference in getting needed care and doctor communication between NHA and NHW individuals were more pronounced in non-cancer group than PCa group.

Keywords: CAHPS; Care experiences; Disparity; Prostate cancer; Quality of care; Retrospective; SEER; Survivorship.

Figure 1:. Sample selection flow chart

PCa: Prostate cancer CAHPS: Consumer Assessment of Healthcare Providers and Systems

Figure 2:. Least-square mean (LSM) estimates for patient care experiences composite measures among matched sample obtained from adjusted multivariate linear regression analyses^# for each race/ethnicity category among groups of PCa survivors and non-cancer controls

* Indicates LSM estimate for a race/ethnicity category statistically significantly different from LSM estimate for NHW category at p<0.05 within respective groups of PCa survivors and non-cancer controls. #LSM estimates were generated using linear regression after propensity score (PS) matching. Prostate cancer and non-cancer groups were PS matched on race, education, proxy answering questions for respondent, low-income subsidy, dual eligibility, plan type, prescription drug plan, geographic region of residence at the time of CAHPS survey, urban/rural residence status, current smoking status, survey year, comorbidity count, general health status, mental health status, age when responded to survey. Linear regression models adjusted for race/ethnicity, cancer status, and a variable indicating interaction between race/ethnicity and cancer status. Exact match on survey year was applied. PCa- Prostate cancer NHA- Non-Hispanic Asian NHB- Non-Hispanic Black NHW- Non-Hispanic White

Figure 2:. Least-square mean (LSM) estimates for patient care experiences composite measures among matched sample obtained from adjusted multivariate linear regression analyses^# for each race/ethnicity category among groups of PCa survivors and non-cancer controls

* Indicates LSM estimate for a race/ethnicity category statistically significantly different from LSM estimate for NHW category at p<0.05 within respective groups of PCa survivors and non-cancer controls. #LSM estimates were generated using linear regression after propensity score (PS) matching. Prostate cancer and non-cancer groups were PS matched on race, education, proxy answering questions for respondent, low-income subsidy, dual eligibility, plan type, prescription drug plan, geographic region of residence at the time of CAHPS survey, urban/rural residence status, current smoking status, survey year, comorbidity count, general health status, mental health status, age when responded to survey. Linear regression models adjusted for race/ethnicity, cancer status, and a variable indicating interaction between race/ethnicity and cancer status. Exact match on survey year was applied. PCa- Prostate cancer NHA- Non-Hispanic Asian NHB- Non-Hispanic Black NHW- Non-Hispanic White

Figure 3:. Differences in racial/ethnicity disparity between matched PCa survivors and non-cancer group^#.

*Indicates values that are significant at p<0.05 ⁺Small difference (<3), ⁺⁺Medium difference (3 to <5), ⁺⁺⁺Large difference (>=5) per Quigley, Denise D., Marc N. Elliott, Claude M. Setodji, and Ron D. Hays, “Quantifying Magnitude of Group-Level Differences in Patient Experiences with Health Care,” Health Services Research, Vol. 53, Suppl. 1, 2018, pp. 3027–3051. Only statistically significant differences-in-differences were categorized as small, medium, and large differences. ^#Difference-in-difference in racial/ethnicity disparity between matched PCa survivors and non-cancer group was estimated as the mean differences in Least Square Mean (LSM) estimates of each race/ethnicity category and NHW among non-cancer group subtracted from respective mean differences in LSM estimates of each race/ethnicity category and NHW among PCa group. LSM estimates were generated using linear regression after propensity score (PS) matching. Prostate cancer and non-cancer groups were PS matched on race, education, proxy answering questions for respondent, low-income subsidy, dual eligibility, plan type, prescription drug plan, geographic region of residence at the time of CAHPS survey, urban/rural residence status, current smoking status, survey year, comorbidity count, general health status, mental health status, age when responded to survey. Linear regression models adjusted for race/ethnicity, cancer status, and a variable indicating interaction between race/ethnicity and cancer status. Exact match on survey year was applied. PCa: Prostate cancer NHA: Non-Hispanic Asian NHB: Non-Hispanic Black NHW: Non-Hispanic White LSM: least square mean

Figure 3:. Differences in racial/ethnicity disparity between matched PCa survivors and non-cancer group^#.

*Indicates values that are significant at p<0.05 ⁺Small difference (<3), ⁺⁺Medium difference (3 to <5), ⁺⁺⁺Large difference (>=5) per Quigley, Denise D., Marc N. Elliott, Claude M. Setodji, and Ron D. Hays, “Quantifying Magnitude of Group-Level Differences in Patient Experiences with Health Care,” Health Services Research, Vol. 53, Suppl. 1, 2018, pp. 3027–3051. Only statistically significant differences-in-differences were categorized as small, medium, and large differences. ^#Difference-in-difference in racial/ethnicity disparity between matched PCa survivors and non-cancer group was estimated as the mean differences in Least Square Mean (LSM) estimates of each race/ethnicity category and NHW among non-cancer group subtracted from respective mean differences in LSM estimates of each race/ethnicity category and NHW among PCa group. LSM estimates were generated using linear regression after propensity score (PS) matching. Prostate cancer and non-cancer groups were PS matched on race, education, proxy answering questions for respondent, low-income subsidy, dual eligibility, plan type, prescription drug plan, geographic region of residence at the time of CAHPS survey, urban/rural residence status, current smoking status, survey year, comorbidity count, general health status, mental health status, age when responded to survey. Linear regression models adjusted for race/ethnicity, cancer status, and a variable indicating interaction between race/ethnicity and cancer status. Exact match on survey year was applied. PCa: Prostate cancer NHA: Non-Hispanic Asian NHB: Non-Hispanic Black NHW: Non-Hispanic White LSM: least square mean