The impact of race and language concordance between patients and navigators on time to diagnostic resolution of breast and cervical cancer screening abnormalities

43 

Loading....

Loading....

Loading....

Loading....

Loading....

Full text

(1)

Boston University

OpenBU http://open.bu.edu

Theses & Dissertations Boston University Theses & Dissertations

2013

The impact of race and language

concordance between patients and

navigators on time to diagnostic

resolution of breast and cervical

cancer screening abnormalities

https://hdl.handle.net/2144/17149 Boston University

(2)

BOSTON UNIVERSITY SCHOOL OF PUBLIC HEALTH

Thesis

THE IMPACT OF RACE AND LANGUAGE CONCORDANCE BETWEEN PATIENTS AND NAVIGATORS ON TIME TO DIAGNOSTIC RESOLUTION OF

BREAST AND CERVICAL CANCER SCREENING ABNORMALITIES

by

MARJORY CHARLOT

MPH, George Washington University, 1999

M.D., University of Iowa Carver College of Medicine, 2004

Submitted in partial fulfillment of the requirements for the degree of

Master of Science 2013

(3)

© 2013 by

MARJORY CHARLOT All rights reserved

(4)

Approved by

First Reader

Karen M. Freund, MD, MPH Adjunct Professor of Medicine

Boston University School of Medicine, Professor of Medicine

Tufts University School of Medicine

Second Reader

Timothy Heeren, PhD Professor of Biostatistics

Boston University School of Public Health

Third Reader

Marianne Prout, MD, MPH Professor of Epidemiology

Boston University School of Public Health

Fourth Reader

Chris Andry, MPhil, PhD

Executive Director for Cancer Care Service Boston Medical Center

Associate Professor of Pathology and Laboratory Boston University School of Medicine

(5)

THE IMPACT OF RACE AND LANGUAGE CONCORDANCE BETWEEN PATIENTS AND NAVIGATORS ON TIME TO DIAGNOSTIC RESOLUTION OF

BREAST AND CERVICAL CANCER SCREENING ABNORMALITIES MARJORY CHARLOT

ABSTRACT

Background: Patient navigators have been shown to reduce cancer disparities among racial/ethnic minorities by improving timely diagnosis and treatment of cancer. We sought to determine if race/ethnicity and language concordance of patients and their navigator improved time to diagnostic resolution of breast and cervical cancer screening abnormalities.

Methods: Demographic data on patients and navigators from the Boston Patient Navigation Research Program were used to assess concordance by race,

ethnicity, and language. Kaplan-Meier survival curves and Cox proportional hazards regression models examined the association of race/ethnicity and language concordance on time to definitive diagnosis of cancer screening abnormalities. All analyses were performed separately for breast and cervical groups.

Results: There were 1257 patients and 23 navigators in this study. In the breast group (n=655), 44% of patient-navigator pairs were concordant by race/ethnicity and 75% were language concordant. In the cervical group (n=602), 70% of patient-navigator pairs were race/ethnicity concordant and 87% were language concordant. There was no association with race/ethnicity concordance and time

(6)

to diagnostic resolution for the breast group, aHR 1.19 (95% CI: 0.97, 1.46) or cervical group, aHR 1.23 (95% CI: 0.99, 1.53). However, in the stratified analysis, race/ethnicity concordance was associated with timelier resolution for minority women with breast and cervical cancer screening abnormalities but not for Whites. For cervical cancer screening abnormalities resolving in less than 90 days, language concordance was also associated with timelier resolution, aHR of 1.46 (95% CI: 1.18, 1.80) but there was no association in the breast group. In the subgroup analysis of Spanish concordance there was also an association of timelier resolution for those with cervical cancer screening abnormalities resolving in less than 90 days.

Conclusion: Patient-navigator race/ethnicity concordance is associated with timelier diagnostic resolution of breast and cervical cancer screening

abnormalities among minority women. Language concordance is also

associated with timelier resolution in participants with cervical cancer screening abnormalities despite the availability and use of interpreters. Given poorer cancer outcomes among minority women, the use of patient navigators that are diverse by race/ethnicity and multilingual may help address barriers to care and improve health outcomes among low-income minorities.

(7)

Table of Contents Background………. 1 Methods………... 4 Results……… 10 Discussion………... 16 Tables..………...22 Figures..………. 27 References…………... 29 Vita……….. 32

(8)

List of Tables

Table 1.1 Baseline characteristics of participants with breast cancer screening abnormalities by race and language concordance status (p22)

Table 1.2 Baseline characteristics of participants with cervical cancer screening abnormalities by race and language concordance status (p23)

Table 2.1 Navigator Characteristics (p24)

Table 3.1 Bivariate analysis of race and language concordance status by percent of breast cancer screening abnormalities resolved at 90 and 365 days (p25)

Table 3.2 Bivariate analysis of race and language concordance status by percent of cervical cancer screening abnormalities resolved at 90 and 365 days (p25)

Table 4.1 Univariate cox proportional hazard model predicting time to diagnostic

resolution (p26)

Table 5.1 Multivariate cox proportional hazard model predicting time to diagnostic

resolution (p26)

Table 6.1 Stratified Multivariate cox proportional hazard model predicting time to

(9)

List of Figures

Figure 1.1. Time to resolution by race concordance status among patients with

breast cancer screening abnormalities (p27)

Figure 1.2 Time to resolution by race concordance status among patients with

cervical cancer screening abnormalities (p27)

Figure 2.1 Time to resolution by language concordance status among patients with breast cancer screening abnormalities (p28)

Figure 2.2 Time to resolution by language concordance status among patients with cervical cancer screening abnormalities (p28)

(10)

Introduction

Incidence and mortality rates of breast and cervical cancer in the United States vary by race and ethnicity. There is a higher incidence of breast cancer among White women compared to all races yet higher mortality rates among Black women.1 However, Hispanic and Asian women have the lowest breast cancer incidence and mortality rates.1 Cervical cancer is not as prevalent as breast cancer among American women and mortality rates have dramatically declined between 1955 and 1970 with the introduction of the Papanicolau (Pap) test.2 Despite overall lower incidence and mortality rates in cervical cancer, there is a disproportionate burden of disease among Black and Hispanic women

compared to Whites.3

Reasons for the disproportionate mortality rates among Black and Hispanic women with breast and cervical cancer are multifactorial and may be due to different tumor characteristics such as the aggressive estrogen,

progesterone, and Her-2 neu negative receptor type of breast cancer more common in Black women4, or later stage of disease at diagnosis for both breast and cervical cancer, which may be influenced by social factors such as lower socioeconomic status, less access to care, or cultural health beliefs.5 Prior studies examining these socio-cultural factors have found that regardless of race or ethnicity, access to cancer screening is increased with access to health

insurance5 and mammography and Pap test rates are similar among White, Black, and Hispanic women.6 However, despite similar screening rates,

(11)

racial/ethnic minorities still seem to have a delay in both the follow up of abnormal screening tests and in the initiation of cancer treatment.6,7

Patient navigation programs have been shown in several studies to reduce delays in cancer care8. Patient navigators work as patient advocates and have a wide range of prior professional training including training as lay

community health workers, licensed social workers, case managers and nurses. They are usually imbedded within the clinical care practice and conduct their work with access to clinical providers as well as scheduling and billing personnel. Navigators work to reduce barriers to care by helping patients acquire health insurance, gain access to care at Federally Qualified Health Centers or safety net hospitals, address logistical barriers such as transportation and child care,

educate patients to improve knowledge, facilitate communication between patients and their providers, and encourage patients to follow through with their care.6,7,9 Several studies also show that timelier resolution of abnormal screening mammography among minorities and low income women can be achieved when navigators coordinate care, provide health education and counseling to address barriers to care.10-12

Along with the important role of patient navigators, understanding and improving patient-provider interactions have also been shown in the literature to play an important role in improving health outcomes. Patient-provider

relationships are paramount to communicating and obtaining health information, delivery and acceptance of health services, as well as patient trust and

(12)

satisfaction.13-26 Race, ethnicity, and language are factors that are known to influence the patient-provider relationship and may play an important role in health disparities. These influences not only pertain to a patient’s perception of the type and quality of care received but have also been found to relate to the stereotypes held by health care providers who care for a diverse group of

patients.13,22,25 Prior studies report that minority patients experience lower quality interpersonal care with providers13,27, less utilization of health services15,19,20,26, and less satisfaction with their care14,18,27, all of which could contribute to poorer cancer outcomes.

Race, ethnicity and language concordance between patients and

providers may possibly mitigate health disparities by facilitating improved patient-provider relationships and better health outcomes. Research into the impact of patient-physician race and language concordance have shown mixed findings depending on the health outcome measured and have been primarily studied in retrospective observational or cross sectional studies.15-21,23-25,27 Up to now, there has been a gap in the literature on studies that examine the role of concordance by race or language in the health setting with health care staff other than

physicians. Little is known about the role of patient-navigator race and language concordance on follow up of cancer screening abnormalities. The goal of this study was to examine the impact of race/ethnicity and language concordance between patients and navigators on time to diagnostic resolution of breast and cervical cancer screening abnormalities.

(13)

Methods

Data Source and Study Population

This is a secondary data analysis of the Boston Patient Navigation

Research Program (PNRP), a quasi-experimental patient navigation intervention for approximately 3,000 women over the age of 18 with either breast or cervical cancer screening abnormalities diagnosed at one of six Boston neighborhood Community Health Centers (CHCs) from January 1, 2007 to December 31, 2008. The primary study was a non randomized group design with a new standard of care such that all patients at the intervention sites received navigation. Three CHCs were assigned to navigate patients with abnormal breast cancer screening results and the other three CHCs navigated patients with cervical cancer

screening abnormalities. Each site served as a control for the non-navigated cancer screening abnormality. Depending on the size of the health center, each was provided resources for a 0.5 – 1 full time equivalent navigator position. The Boston University Institutional Review Board approved the study.8,28

The patient navigation intervention consisted of female navigators hired as employees of the community health centers and were imbedded within the

clinical care team interacting with clinical providers and staff in their day to day work. They were trained by the PNRP research staff to identify and document barriers to care, develop individual plans to address barriers, and follow women to diagnostic resolution of their cancer screening abnormality or to the end of treatment if diagnosed with cancer. Their initial interaction with patients occurred

(14)

via telephone after the patient was informed of an abnormal cancer screening result by their clinician. Patient navigators had access to interpreter services if they could not communicate with the patient. Subsequent patient interactions were face to face, via telephone and/or mail. Patient navigators also attended bimonthly trainings on topics such as effective communication, cultural

competency, and patient empowerment. Each health center identified a clinical supervisor, who was also trained to provide navigation in addition to provide supervision, and who was available during the navigator’s absences or during transition periods in staffing. All interactions were documented in the electronic medical record. Data on follow up of abnormal screening findings were also collected through manual chart abstraction.8

Inclusion/Exclusion Criteria

By design all women with abnormal breast or cervical cancer screening abnormalities at the 6 CHCs were included in the primary study. For this secondary analysis, only women on the navigation arm of the study with one navigator were included. Participants with support from more than one navigator were excluded from this analysis given the primary exposure of interest was race/ethnicity and language concordance between each patient-navigator pair. Participants with more than one navigator had the possibility of interacting with navigators of various racial/ethnic backgrounds and/or language abilities potentially complicating measurement of the exposure. Women on the control

(15)

arm of the primary study were also excluded from this analysis since they did not have a navigator.

Measurement of Outcome:

The primary outcome of interest was time to diagnostic resolution defined as the time from the initial screening abnormality to definitive imaging study, repeat Pap test, or biopsy after an abnormal screening test. All women included in the study were followed until the time of diagnostic resolution or up to 365 days from the date of the screening abnormality. Participants whose abnormalities were not resolved by 365 days were censored at 365 days and those who died were censored at date of death. The clinicians caring for the patient determined the time point for when the abnormality was considered resolved. Given that patients with a radiographic BIRADS 3 abnormality are recommended to have a repeat mammogram every 180 days for two years, we followed subjects to confirm that the first repeat imaging was completed. Participants with a

radiographic BIRADS 3 abnormality that resolved before 180 days were counted as having their screening abnormality resolved at day 0.5 for the purpose of statistical analyses while all others had 180 days subtracted from the diagnostic evaluation time in order to compare all the screening abnormalities in a similar time frame.8

(16)

Measurement of Exposure

The two main exposures of interest were race/ethnicity concordance and language concordance. We defined race/ethnicity concordance as a

dichotomous variable based on patient and navigator race. Race and ethnicity data were identified through health center registration records for all patients and were self-identified for navigators. Prior validation with a different sample of patients comparing registration data with self-identified race information showed high correlation and agreement measured by a Kappa score with values ranging from 81-91% for all racial/ethnic groups.29 We imputed race/ethnicity for 8% of subjects using in hierarchical order provider’s notes in the medical record, primary language, country of origin, or the patient’s surname.8

Race and ethnicity were combined into four mutually exclusive categories of Asian, Hispanic, non-Hispanic Black, and non-non-Hispanic White. Patients and navigators of the same race or ethnicity were classified as race concordant and pairs of different race or ethnicity were classified as race discordant. Language concordance was also a dichotomous variable based on patient and navigator language. For missing patient language, approximately 6% was imputed as other than English if there was documentation of another language in the medical record, need for

interpreter, or result letter sent to the patient in a language other than English.8 All navigators were English speaking and 5 navigators also spoke either Spanish Vietnamese, or Albanian. Patients and navigators that were able to communicate in the same language were classified as language concordant and pairs that

(17)

could not communicate in a common language were classified as language discordant.

Covariate Assessment

Patient age, race/ethnicity, language, socioeconomic status measured as percent of zip code under federal poverty level, insurance status (public, private, and no insurance including those with free care), and type of screening

abnormality were obtained from the administrative billing and electronic medical records. These covariates were assessed given this was a secondary analysis of a non-randomized study design and because of their potential to confound or modify the effect of time to diagnostic resolution of the screening abnormality with respect to race and language concordance.

Statistical Analyses

An a priori decision was made to conduct analyses separately for breast and cervical screening groups as done in the primary study given there were different demographic characteristics of the participants, particularly by age and race and because of the different recommendations for management of

screening abnormalities and varying severity of abnormalities. Descriptive

bivariate associations of demographic and clinical characteristics were examined to compare race concordant to race discordant patient-navigator pairs as well as language concordant to language discordant pairs. In addition, bivariate

(18)

associations of demographic and clinical characteristics were examined to compare those with screening abnormalities that ever resolved to those that never resolved. To assess the bivariate associations, chi-square tests and

Fisher’s exact tests were employed for categorical variables and Student’s t tests were used for continuous variables.

For the primary unadjusted analyses, Kaplan-Meier survival curves were generated to assess the effect of race and language concordance on time to diagnostic resolution of breast and cervical cancer screening abnormalities. The Logrank test was used for significance testing. Univariate and multivariate Cox proportional hazards regression models were used to assess the association between race and language concordance with time to diagnostic resolution. Proportional hazards assumptions were tested graphically using the survival curves for difference in the effect over time, and because of violation of the

proportionality assumptions multiplicative interaction terms between concordance and time were employed. Since differences in the effect over time occurred at around 90 days for most curves, hazard ratios were calculated for 0-89 days and 90-365 days where needed for significant time interactions.

Multivariate analyses were adjusted for age, race/ethnicity, language, insurance status, type of screening abnormality and clustering by patient

navigator. Adjustment for clustering by patient navigator was employed to control for the influence of the individual patient navigator. To assess the varying effect of race concordance by the different race/ethnicity groups, a stratified analysis

(19)

was performed comparing White concordant to White discordant, Black

concordant to Black discordant and Hispanic concordant to Hispanic discordant patient-navigator pairs. Asian patient-navigator pairs were excluded in stratified analyses for cervical screening abnormalities due to the small numbers in the study sample. We also performed a subgroup analysis of language concordance for Spanish speakers only given all of our navigators spoke English and therefore there were no English discordant pairs and the numbers were too small for

languages other than English or Spanish. Stratified analyses were adjusted only for age, insurance status and clustering by navigator given the subgroup analysis included smaller numbers and these covariates were important with respect to the study outcome of time to diagnostic resolution. All analyses were performed with SAS statistical software (SAS Institute, Cary, NC).

Results

Patient and Navigator Characteristics

There were a total of 1257 women with either breast (n= 655) or cervical (n=602) cancer screening abnormalities and 23 navigators. Over two thirds of women with breast cancer screening abnormalities were between the ages of 40 and 59 years, 56% were non-White, 63% spoke English, and 50% had public insurance. The most common breast cancer screening abnormality was BIRADS category 0 (66%) which refers to a finding that needs additional imaging

(20)

Women with cervical cancer screening abnormalities were younger than the breast group with over 80% under the age of 39 years. They were mostly Black and Hispanic (comprising 80% of cervical cancer screening group) and one third of these women spoke only Spanish while nearly 60% spoke English. Most were either uninsured (41%) or had public insurance (37%). Over 96% of the screening abnormalities were low-grade cervical lesions. (Table 1.2)

There were 23 female patient navigators and they were primarily White (57%). All patient navigators were able to communicate in English and 21% were also able to communicate in either Spanish or Vietnamese. Navigators with data available on age were young and mostly between the ages of 24 and 29 years. (Table 2.1)

Race/Ethnicity Concordance: Breast Cancer Screening Abnormalities

Forty four percent of the patient-navigator pairs were concordant by race or ethnicity. Race/ethnicity concordant groups compared to discordant groups were predominately White (75% vs. 20%). There were significant differences between the concordant and discordant groups in age, race/ethnicity, language, insurance status, socioeconomic status, and screening abnormality (Table 1.1). The rate of diagnostic resolution of breast cancer screening abnormalities at 90 days was over 85% for both the race/ethnicity concordant and discordant groups and at 365 days over 90% of screening abnormalities were resolved for

(21)

In the unadjusted analysis of time to diagnostic resolution by race/ethnicity concordance status represented by Kaplan-Meier curves (Figure 1.1), there was no statistically significant difference between the race/ethnicity concordant and discordant groups (p=0.26). Univariate and multivariate Cox proportional hazards models for time to diagnostic resolution of breast cancer screening abnormalities (Tables 4.1 and 5.1) also did not show a statistically significant difference

between race concordant and discordant patient-navigator pairs. However, the multivariate analysis stratified by patient race/ethnicity and adjusted for age, insurance status, and clustering by navigator, showed that minority women in race concordant groups had timelier resolution of their screening abnormalities. Hazard ratio greater >1 in the Cox models represented timelier resolution. Black concordant patient-navigator pairs compared to Black discordant pairs were nearly three times as likely to have timelier resolution of screening abnormalities that resolved after 90 days (aHR 2.62, 95% CI: 1.56 to 4.41); however, at less than 90 days Black concordant patient-navigator pairs had less timely resolution of their screening abnormalities. Hispanic and Asian concordant patient-navigator pairs had a 30% or more likelihood of timelier resolution of breast cancer

screening abnormalities resolving before 90 days. There was no association with time to diagnostic resolution of breast cancer screening abnormalities and White race concordance status. (Table 6.1)

(22)

Race/Ethnicity Concordance: Cervical Cancer Screening Abnormalities

Among the cervical cancer screening group, 70% of patient-navigator pairs were concordant by race or ethnicity. Race/ethnicity concordant groups compared to discordant groups were mostly Black (40% vs. 19%) and Hispanic (55% vs. 26%). There were significant differences between the concordant and discordant groups in race/ethnicity, language, insurance status, and

socioeconomic status (Table 1.2). The rate of diagnostic resolution of cervical cancer screening abnormalities at 90 days was 65% for race/ethnicity concordant group and 62% for discordant groups. At 365 days 94% of race concordant groups and 87% of race discordant groups ever achieved diagnostic resolution of cancer screening abnormalities and this finding was statistically significant (Table 3.1).

The Kaplan-Meier curves in Figure 1.2 represents the unadjusted analysis of time to diagnostic resolution of cervical cancer screening abnormalities by race/ethnicity concordance status and show no statistically significant difference between the race concordant and discordant groups (p=0.07). The univariate and multivariate cox proportional hazard models for time to diagnostic resolution by race/ethnicity concordance status were also not statistically significant but there appeared to be a trend toward race/ethnicity concordant groups having timelier resolution compared to race/ethnicity discordant groups. In the stratified

multivariate analysis, again minority women in race/ethnicity concordant groups had timelier resolution of their screening abnormalities. Hispanic patients with a

(23)

Hispanic navigator had an 80% more likelihood of having timelier resolution of their cancer screening abnormality (aHR 1.82 95%CI: 1.39 to 2.39) and this association did not vary with time. Black patients with a Black navigator were 50% more likely to have timelier resolution of cervical cancer screening abnormality that resolved after 90 days (aHR 1.50, 95%CI: 1.35 to 1.67). However, there was a negative association between White race concordance status and time to diagnostic resolution (aHR 0.60 95%CI 0.48 to 0.76). (Table 6)

Language Concordance: Breast Cancer Screening Abnormalities

Over 75% of patient navigator pairs in the breast cancer screening group were language concordant. Since all navigators spoke English, there were no English discordant patient-navigator pairs. The largest non-English speaking groups were Spanish and Vietnamese speakers and the percentage of Spanish speakers in the language concordant and discordant groups was fairly similar (12% and 20%) while the percentage of Vietnamese speakers were unbalanced comprising only 4% of the language concordant group and 75% of language discordant groups. At 90 days close to 90% of language concordant patient-navigator pairs had diagnostic resolution of cancer screening abnormalities compared to only 80% in the language discordant groups and this finding was statistically significant (Table 3.1).

The unadjusted analysis of time to diagnostic resolution by language concordance status represented by the Kaplan-Meier curve in figure 2.1, shows a

(24)

suggestion of the language concordant group having timelier resolution than language discordant groups and this finding had borderline significance (p=0.05). Table 4.1 also shows that in the univariate Cox proportional hazards models of time to diagnostic resolution by language concordance status that language concordant groups are 30% more likely to have timelier resolution than language discordant groups (HR 1.30 95% CI:1.06 to 1.58). However, in the subgroup analysis of Spanish speaking study participants, there was no association between time to diagnostic resolution of breast cancer screening abnormalities and Spanish concordance status after adjusting for age, insurance status and clustering by navigator (Table 6.1). The subgroup analysis was limited to Spanish speakers given there were no English discordant groups for comparison to the English discordant group and the number of participants with languages other than English or Spanish were too small for comparison.

Language Concordance: Cervical Cancer Screening Abnormalities

Eighty seven percent of patient-navigator pairs in the cervical group were language concordant. One third of the women in the language concordant and discordant groups were Spanish speakers and racial/ethnic distribution was similar between the groups and comprised mostly of Black and Hispanic women. At 90 days, over 60% of cervical cancer screening abnormalities had diagnostic resolution in both the language concordant and discordant groups, and at 365

(25)

days over 90% of cancer screening abnormalities were resolved in both groups (Table 3.2).

In the unadjusted analysis, the Kaplan-Meier curve in Figure 2.2 shows no difference between language concordant and discordant groups in time to

diagnostic resolution of cervical cancer screening abnormalities. In the

multivariate proportional hazard models (Table 5.1), language concordant groups had timelier resolution of cancer screening abnormalities that resolved in less than 90 days (HR=1.46, 95% CI: 1.18, 1.80) after adjusting for race/ethnicity, age, language, insurance, socioeconomic status, screening abnormality and clustering by navigator. In addition, after adjusting for age, insurance status, and clustering by navigator, the subgroup analysis of Spanish speakers showed that Spanish concordant pairs were 43% more likely to have timelier resolution of cervical cancer screening abnormalities resolving in less than 90 days (aHR 1.43, 95% CI: 1.10, 1.84) with a marginal positive association for concordant groups with resolution occurring at 90 days or more.

Discussion

Delays in follow up of cancer screening abnormalities have been found to be associated with poor health outcomes.10 This study sought to determine the impact of patient and navigator race/ethnicity and language concordance on time to diagnostic resolution of breast and cervical cancer screening abnormalities. Many but not all of our findings indicated that minority women with race/ethnicity

(26)

concordant navigators had timelier resolution of breast and cervical cancer screening abnormalities compared to race/ethnicity discordant groups. In contrast, there was either no association or no benefit of race concordance for White participants in this study.

The benefit of race/ethnicity concordance among minority women varied with time with exception of Hispanic women in the cervical cancer screening group that benefited from having a Hispanic navigator for abnormalities that resolved before and after 90 days. Differences in the association of race/ethnicity concordance with time to diagnostic resolution of cancer screening abnormalities among minority women may be explained by differences in culture and level of acculturation among and within different racial/ethnic groups26. Given the younger population of those with cervical cancer screening abnormalities, they may be newly arrived immigrants and represent a less acculturated population. The negative association between race concordance and time to diagnostic resolution of breast cancer screening abnormalities resolving in less than 90 days among Black concordant patient-navigator group was an exception among minority study participants. Our sample of the Black race concordant group with breast cancer screening abnormalities included only 5 patient-navigator pairs and therefore it is difficult to draw any conclusions on this finding on such a small sample. However, the sample of race concordant Black women in the cervical group was larger and showed a benefit of race/ethnicity concordance for abnormalities that resolved in 90 or more days and no association with

(27)

abnormalities that resolved in less than 90 days. This suggests that the benefit of race concordance for Black women may occur later and a race concordant

navigator may mitigate the factors associated with delayed resolution of cancer screening abnormalities. Prior studies on patient-provider race concordance and health outcomes have had mixed findings with a few studies showing a modest association particularly for minorities.19,20,24,26 A literature review of 27 studies (all observational) found inconclusive evidence for the association of patient-provider race concordance and health outcomes and argued that these findings could be a result of small sample size, patient self report of health outcomes, and

evaluating race concordance with physician providers only.22 The design of our study differs from the existing literature in that we prospectively assessed the impact of patient and navigator race/ethnicity concordance using an objectively defined study outcome.

The finding of no association between race concordance and time to diagnostic resolution among White women with breast cancer screening abnormalities may be related to high diagnostic resolution rates for both race concordant and discordant groups thereby making it difficult to assess a difference between the groups. Also, White study participants may face fewer cultural barriers in a health system where providers and staff are predominately White.26 However, reasons for the negative association between White race concordance status and resolution of cervical cancer screening abnormalities are unclear but may be due to cultural discordance within the White race concordant

(28)

group, for example, White Albanian patients paired with White American

navigators, or unmeasured factors such as competing priorities resulting in these differences. It is also possible that minority navigators may be better at identifying and addressing barriers to care among White women and other minority women seeking care at urban community health centers.

Our findings on language concordance demonstrated timelier resolution of cervical cancer screening abnormalities for language concordant groups. The subgroup analysis of Spanish speaking patients also showed the benefit of language concordance among those with cervical cancer screening

abnormalities. Even though most participants in this study were English speaking about one third were non-English speakers suggesting that having a navigator that can communicate in the same language was beneficial. Our findings also suggest that despite interpreter use among the language discordant group that having a navigator that can communicate in the same language is ideal. Our finding is consistent with prior studies showing a benefit of language concordance among non-English speaking populations but these studies focused on patient-provider interactions.23,26 We were able to show that the use of

linguistically concordant navigators can also improve outcomes particularly among those with language barriers.

There was no association between language concordance and timely diagnostic resolution of breast cancer screening abnormalities. This lack of association particularly among the breast cancer screening group may again be

(29)

related to the high rate of diagnostic resolution. There may have also been fewer language barriers among this group given they were mostly English speakers and older women who may have been in this country long enough to understand English.

While this study is able to elucidate the impact of race and language concordance among 4 distinct racial/ethnic cohort, there are limitations. First, this was a secondary data analysis and race and language concordant and

discordant groups were not balanced for either the breast or cervical cancer screening groups. Second, given the study design, our subgroup analysis of language concordance was limited because the study population was primarily English speakers and all navigators spoke English therefore limiting our

understanding of the effect of language concordance for non-English speakers other than Spanish speakers.

Conclusion

Our data suggest that patient-navigator race/ethnicity and language concordance may be beneficial for minority women particularly for Hispanic women with cervical cancer screening abnormalities. Even with the use and availability of interpreters, Spanish language concordant patient-navigator pairs have timelier resolution of cervical cancer screening abnormalities. Given poorer cancer outcomes among minority women, the use of patient navigators that are diverse by race/ethnicity and multilingual may help address barriers to care and

(30)

improve health outcomes for a low-income minority population. Future research on patient-navigator relationships may further contribute to our understanding of the dynamics involved in improving delays in diagnostic resolution of cancer screening abnormalities.

(31)

Table 1.1 Baseline characteristics of participants with breast cancer screening abnormalities by race and language concordance status.

Total N=651 Col % Race Concordant N=286 Col % Race Discordant N=365 Col % P-value Total N=655 Concordant Language N=494 Col % Language Discordant N=161 Col % P-value Age 18-29 30-39 40-49 50-59 60-69 70-79 80+ 31 (5) 74 (11) 269(41) 168(26) 72 (11) 33 (5) 4 (<1) 22 (8) 39 (14) 111 (39) 60 (21) 33 (12) 17 (6) 4 (1) 9 (3) 35 (10) 158 (43) 108 (30) 39 (11) 16 (4) 0 (0) 0.001 32 (5) 74 (11) 272 (42) 168 (26) 72 (11) 33 (5) 4 (<1) 29 (6) 62 (13) 213 (43) 114 (23) 49 (10) 23 (5) 4 (<1) 3 (2) 12 (7) 59 (37) 54 (34) 23 (14) 10 (6) 0 (0) 0.01 Race/ Ethnicity White Black Asian Hispanic 289(44) 99 (15) 127(20) 136(21) 215 (75) 5 (2) 29 (10) 37 (13) 74 (20) 94 (26) 98 (27) 99 (27) <.0001 291 (44) 100 (15) 127 (19) 137 (21) 242 (49) 93 (19) 52 (11) 107(22) 49 (30) 7 (4) 75 (47) 30 (19) <.0001 Language English Spanish Vietnamese Other 412(63) 88 (14) 96 (15) 55 (8) 197 (69) 24 (8) 22 (8) 43 (15) 215 (59) 64 (18) 74 (20) 12 (3) <.0001 414 (63) 90 (14) 96 (15) 55 (8) 414 (84) 58 (12) 22 (4) 0 (0) 0 (0) 32 (20) 74 (46) 55 (34) <.0001 Insurance Public Private Uninsured 327(50) 220(34) 104(16) 113 (40) 128 (45) 45 (16) 214 (59) 92 (25) 59 (16) <.0001 330 (50) 221 (34) 104 (16) 236 (48) 198 (40) 60 (12) 94 (58) 23 (14) 44 (27) <.0001 % Zip Code Below Federal Poverty Level Mean ± SD 14.74 ± 6.9 13.59 ± 6.34 15.64 ± 7.19 0.0001 14.74 ± 6.9 15.06 ± 6.92 13.77 ± 6.77 0.04 SCREENING ABNORMALITY Clinical BIRADS 0 BIRADS 3 BIRADS 4/5 161(25) 431(66) 51 (8) 8 (1) 91 (32) 172 (60) 23 (8) 0 (0) 70 (19) 259 (71) 28 (8) 8 (2) 0.0002 164 (25) 432 (66) 51 (8) 8 (1) 133 (27) 318 (64) 37 (8) 6 (1) 31 (19) 114 (71) 14 (9) 2 (1) 0.28

(32)

Table 1.2 Baseline characteristics of participants with cervical cancer screening abnormalities by race and language concordance status.

Total N=598 Col % Race Concordant N=420 Col % Race Discordant N=178 Col % P-value Total N=602 Concordant Language N=526 Col % Language Discordant N= 76 Col % P-value Age 18-29 30-39 40-49 50-59 60-69 70-79 357 (60) 143 (24) 73 (12) 19 (3) 3 (<1) 3 (<1) 248 (59) 108 (26) 46 (11) 14 (3) 1 (<1) 3 (<1) 109 (61) 35 (20) 27 (15) 5 (3) 2 (1) 0 (0) 0.21 361 (60) 143 (24) 73 (12) 19 (3) 3 (0.5) 3 (0.5) 328 (62) 122 (23) 57 (11) 13 (3) 3 (0.6) 3 (0.6) 33 (43) 21 (28) 16 (21) 6 (8) 0 (0) 0 (0) 0.004 Race/ Ethnicity White Black Asian Hispanic 115 (19) 201 (34) 7 (1) 275 (46) 23 (5) 168 (40) 0 (0) 229 (55) 92 (52) 33 (18) 7 (4) 46 (26) <.0001 115 (19) 202 (34) 7 (1) 278 (46) 98 (19) 172 (32) 6 (1) 250 (48) 17 (22) 30 (40) 1 (1) 28 (37) 0.34 Language English Spanish Vietnamese Other 355 (59) 195 (33) 1 (<1) 47 (8) 220 (52) 168 (40) 0 (0) 32 (8) 135 (76) 27 (15) 1 <1) 15 (8) <.0001 357 (59) 197 (33) 1 (<1) 47 (8) 357 (68) 168 (32) 0 (0) 1 (<1) 0 (0) 29 (38) 1 (1) 46 (61) <.0001 Insurance Public Private Uninsured 220 (37) 133 (22) 245 (41) 144 (34) 80 (19) 196 (47) 76 (43) 53 (30) 49 (28) <.0001 222 (37) 135 (22) 245 (41) 193 (37) 125 (24) 208 (40) 29 (38.16) 10 (13.16) 37 (48.68) 0.09 % Zip Code Below Federal Poverty Level Mean ± SD 14.45 ± 5.46 14.97 ± 5.33 13.23 ± 5.6 0.0003 14.48 ± 5.46 14.68 ± 5.37 13.11 ± 5.92 0.02 Cervical LGL HGL 575 (96) 23 (4) 403 (96) 17 (4) 172 (97) 6 (3) 0.69 578 (96) 24 (4) 508 (97) 18 (3) 70 (92) 6 (8) 0.06

(33)

Table 2.1. Navigator Characteristics Navigator Characteristics Total N=23 Col % Age 24-29 30-39 40-49 50-59 Missing 6 (26) 1 (4) 3 (13) 1 (4) 12 (52) Race/Ethnicity White Black Hispanic Asian Missing 13 (57) 4 (17) 3 (13) 1 (4) 2 (9) Language English only Spanish Vietnamese Albanian 18 (78) 3 (13) 1 (4) 1 (4)

(34)

Table 3.1 Bivariate analysis of race and language concordance status by percent of breast cancer screening abnormalities resolved at 90 and 365 days.

% Resolved at 90 days P-value % Resolved at 365 days P-value Race concordant Yes No Language Concordant Yes No 85% 86% 87% 80% 0.87 0.03 93% 96% 95% 95% 0.11 0.82

Table 3.2 Bivariate analysis of race and language concordance status by percent of cervical cancer screening abnormalities resolved at 90 and 365 days.

% Resolved at 90 days P-value % Resolved at 365 days P-value Race concordant Yes No Language Concordant Yes No 65% 62% 64% 63% 0.42 0.83 94% 87% 91% 94% 0.01 0.50

(35)

Table 4.1. Univariate cox proportional hazard model predicting time to diagnostic resolution

Breast Cervical

<90 days 90-365 days <90 days 90-365 days

Race concordant v. discordant 1.16 (0.98, 1.38) 0.62 (0.37, 1.05) 1.18 (0.98, 1.43) 1.18 (0.98, 1.43) Language concordant v. discordant 1.30 (1.06, 1.58) 0.67 (0.41, 1.12) 1.12 (0.83, 1.51) 0.82 (0.53, 1.27)

Table 5.1. Multivariate cox proportional hazard model predicting time to diagnostic resolution

Breast Cervical

<90 days 90-365 days <90 days 90-365 days

*Race concordant v. discordant 1.19 (0.97, 1.46) 1.19 (0.97, 1.46) 1.23 (0.99, 1.53) 1.23 (0.99, 1.53) *Language concordant v. discordant 1.13 (0.80, 1.59) 1.04 (0.70, 1.55) 1.46 (1.18, 1.80) 1.16 (0.65, 2.07) *Adjusted for race/ethnicity age, language, insurance, SES, screening abnormality, and clustering by navigator

Table 6.1. Stratified Multivariate cox proportional hazard model predicting time to diagnostic resolution

*Race Concordance Breast Cervical

<90 days 90-365 days <90 days 90-365 days

White concordant v. discordant 0.98 (0.83, 1.16) 0.98 (0.83, 1.16) 0.60 (0.48, 0.76) 0.60 (0.48, 0.76) Black concordant v. discordant 0.86 (0.75, 0.99) 2.62 (1.56, 4.41) 0.91 (0.47, 1.79) 1.50 (1.35, 1.67) Hispanic concordant v. discordant 1.37 (1.06, 1.77) 1.49 (0.83, 2.67) 1.82 (1.39, 2.39) 1.82 (1.39, 2.39) Asian concordant v. discordant 1.29 (1.07, 1.55) 0.98 (0.83, 1.16) --- --- *Language Concordance Spanish concordant v. discordant 1.10 (0.69, 1.75) 1.10 (0.69, 1.75) 1.43 (1.10, 1.84) 1.63 (1.00, 2.66) * Adjusted for age, insurance status, and clustering by navigator

--- Asian population not included in stratified analysis for cervical screening group due to small numbers

(36)

Figure 1.1. Time to resolution by race concordance status among patients with breast cancer screening abnormalities

Figure 1.2 Time to resolution by race concordance status among patients with cervical cancer screening abnormalities

0 10 20 30 40 50 60 70 80 90 100 0 50 100 150 200 250 300 350 400 % R e solv e d

Number of days to diagnostic resolution Race discordant Race concordant 0 10 20 30 40 50 60 70 80 90 100 0 50 100 150 200 250 300 350 400 % R e solv e d

Number of days to diagnostic resolution Race discordant Race concordant

Logrank p=0.26

(37)

Figure 2.1 Time to resolution by language concordance status among patients with breast cancer screening abnormalities

Figure 2.2. Time to resolution by language concordance status among patients with cervical cancer screening abnormalities

0 10 20 30 40 50 60 70 80 90 100 0 50 100 150 200 250 300 350 400 %Reso lved

Number of days to diagnostic resolution

Language Discordant Language Concordant 0 10 20 30 40 50 60 70 80 90 100 0 50 100 150 200 250 300 350 400 %Reso lved

Number of days to diagnostic resolution Language Discordant Language Concordant

Logrank p=0.05

(38)

References:

1. SEER Cancer Statistics Review, 1975-2008, National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/csr/1975_2008/, based on November 2010 SEER data submission, posted to the SEER web site, 2011.

2. American Cancer Society. What are the key statistics about cervical cancer? Revised 3/14/2012. http://www.cancer.org.

3. SEER Stat Fact Sheets: Cervix Uteri based on November 2011 SEER data submission. seer.cancer.gov.

4. Carey LA, Perou CM, Livasy CA, et al. Race, breast cancer subtypes, and survival in the Carolina Breast Cancer Study. JAMA: The Journal of the American Medical Association 2006;295:2492-502.

5. Selvin E, Brett KM. Breast and cervical cancer screening: sociodemographic predictors among White, Black, and Hispanic women. American Journal of Public Health 2003;93:618-23.

6. Natale-Pereira A, Enard KR, Nevarez L, Jones LA. The role of patient navigators in eliminating health disparities. Cancer 2011;117:3543-52. 7. Donelan K, Mailhot JR, Dutwin D, et al. Patient perspectives of clinical care

and patient navigation in follow-up of abnormal mammography. Journal of General Internal Medicine 2011;26:116-22.

8. Battaglia TA, Bak SM, Heeren T, et al. Boston Patient Navigation Research Program: the impact of navigation on time to diagnostic resolution after abnormal cancer screening. Cancer Epidemiology, Biomarkers and Prevention 2012;21:1645-54.

9. Hendren S, Chin N, Fisher S, et al. Patients' barriers to receipt of cancer care, and factors associated with needing more assistance from a patient

navigator. Journal of the National Medical Association 2011;103:701-10. 10. Battaglia TA, Roloff K, Posner MA, Freund KM. Improving follow-up to

abnormal breast cancer screening in an urban population. A patient navigation intervention. Cancer 2007;109:359-67.

11. Lobb R, Allen JD, Emmons KM, Ayanian JZ. Timely care after an abnormal mammogram among low-income women in a public breast cancer screening program. Archives of Internal Medicine 2010;170:521-8.

(39)

12. Ell K, Vourlekis B, Lee PJ, Xie B. Patient navigation and case management following an abnormal mammogram: a randomized clinical trial. Preventive Medicine 2007;44:26-33.

13. Cooper LA, Beach MC, Johnson RL, Inui TS. Delving below the surface. Understanding how race and ethnicity influence relationships in health care. Journal of General Internal Medicine 2006;21 Suppl 1:S21-7.

14. Cooper-Patrick L, Gallo JJ, Gonzales JJ, et al. Race, gender, and partnership in the patient-physician relationship. JAMA: Journal of the American Medical Association 1999;282:583-9.

15. Blanchard J, Nayar S, Lurie N. Patient-provider and patient-staff racial concordance and perceptions of mistreatment in the health care setting. Journal of General Internal Medicine 2007;22:1184-9.

16. Bleich SN, Simon AE, Cooper LA. Impact of patient-doctor race concordance on rates of weight-related counseling in visits by black and white obese individuals. Obesity 2012;20:562-70.

17. Eamranond PP, Davis RB, Phillips RS, Wee CC. Patient-physician language concordance and primary care screening among spanish-speaking patients. Medical Care 2011;49:668-72.

18. Jerant A, Bertakis KD, Fenton JJ, Tancredi DJ, Franks P. Patient-provider sex and race/ethnicity concordance: a national study of healthcare and outcomes. Medical Care 2011;49:1012-20.

19. Lasser KE, Mintzer IL, Lambert A, Cabral H, Bor DH. Missed appointment rates in primary care: the importance of site of care. Journal of Health Care for the Poor and Underserved 2005;16:475-86.

20. LaVeist TA, Nuru-Jeter A, Jones KE. The association of doctor-patient race concordance with health services utilization. Journal of Public Health Policy 2003;24:312-23.

21. Linsky A, McIntosh N, Cabral H, Kazis LE. Patient-provider language concordance and colorectal cancer screening. Journal of General Internal Medicine 2011;26:142-7.

22. Meghani SH, Brooks JM, Gipson-Jones T, Waite R, Whitfield-Harris L, Deatrick JA. Patient-provider race-concordance: does it matter in improving minority patients' health outcomes? Ethnicity & Health 2009;14:107-30.

(40)

23. Ngo-Metzger Q, Sorkin DH, Phillips RS, et al. Providing high-quality care for limited English proficient patients: the importance of language concordance and interpreter use. Journal of General Internal Medicine 2007;22 Suppl 2:324-30.

24. Saha S, Komaromy M, Koepsell TD, Bindman AB. Patient-physician racial concordance and the perceived quality and use of health care. Archives of Internal Medicine 1999;159:997-1004.

25. Strumpf EC. Racial/ethnic disparities in primary care: the role of physician-patient concordance. Medical Care 2011;49:496-503.

26. Traylor AH, Schmittdiel JA, Uratsu CS, Mangione CM, Subramanian U. Adherence to cardiovascular disease medications: does patient-provider race/ethnicity and language concordance matter? Journal of General Internal Medicine 2010;25:1172-7.

27. Street RL, Jr., O'Malley KJ, Cooper LA, Haidet P. Understanding concordance in patient-physician relationships: personal and ethnic

dimensions of shared identity. Annals of Family Medicine 2008;6:198-205. 28. Freund KM, Battaglia TA, Calhoun E, et al. National Cancer Institute Patient

Navigation Research Program: methods, protocol, and measures. Cancer 2008;113:3391-9.

29. Laiteerapong N, Sherman BJ, Kim K, Freund K. Validation of Racial Categorization in a Hospital Administrative Database. Journal of General Internal Medicine 2005;20 Suppl 1:130

30. Taplin SH, Ichiawa LE, Kerlikowskai K, et al. Concordance of Breast Imaging Reporting and Data System Assessments and Management

Recommendations in Screening Mammography. Radiology 2002;222:529-535.

(41)

V I T A M a r j o r y C h a r l o t , M D , M P H 1 0 R o b e r t s A v e n u e , N e w t o n v i l l e , M A 0 2 4 6 0 P h o n e ( 6 1 7 ) 8 0 3 - 9 5 0 9 E m a i l m a r j o r y . c h a r l o t @ b m c . o r g EDUCATION/TRAINING

09/2010-Present Boston University School of Public Health Master of Science in Epidemiology Candidate Expected graduation, September 2013

06/2011-06/2013 Boston University Medical Center

Post Doctorate Clinical Research Fellow

07/2010-06/2011 Boston University Medical Center Chief Hematology/Oncology Fellow 07/2008-06/2011 Boston University Medical Center

Hematology/Oncology Fellowship 06/2004-06/2007 Boston University Medical Center

Intern and Resident, Internal Medicine Residency Program

07/2000-05/2004 University of Iowa Carver College of Medicine, Iowa City, IA

MD May 2004

09/1997-01/1999 George Washington University, Washington, DC Concentration: International Health

Thesis: “Assessment of Adolescent Reproductive Health in Nigeria”

Data obtained from Family Health International, Washington DC

MPH, January 1999

08/1996-05/1997 Howard University, Washington, DC Post-Baccalaureate Pre-Med Program 08/1991-05/1995 Tufts University, Medford, MA

Major: Biology, Minor: Africa and New World Studies BS May 1995

(42)

BOARD CERTIFICATION

2007 American Board of Internal Medicine

2011 American Board of Internal Medicine – Medical Oncology

HONORS/AWARDS

2000-2004 University of Iowa College of Medicine Scholar: Four-year paid scholarship for Academic Excellence and Professional Promise.

1998 Delta Omega Honor Society for Public Health, George Washington University

PROFESSIONAL MEMBERSHIPS

American Society of Hematology American Society of Clinical Oncology American Association for Cancer Research

EMPLOYMENT

2011-Present Hematology/Oncology Attending, Boston Medical Center  Assist in diagnosis, evaluation, treatment of malignant solid

tumors and benign hematologic disorders 2007–2008 Hospitalist, Brigham and Women’s Hospital

 Responsible for the supervision and teaching of physician assistants and medical residents on the general internal medicine service.

1999-2000 Middle School Health Coordinator, Boston Public Health Commission

 Designed, implemented, and evaluated middle school health education and careers curriculum for Boston public school students.

1995-1996 Middle School Coordinator, Higher Achievement Program

 Designed, implemented, and evaluated after school and summer programs in Science, Math, and Reading for Washington DC public school students

(43)

RESEARCH

2012 Black Women’s Health Study. Slone Epidemiology Center at

Boston University. Diabetes and Breast Cancer Outcomes. Research Mentor: Julie Palmer, ScD, MPH.

2011 Boston Patient Navigation Research Program. Boston University

Medical Center. Secondary analysis of time to diagnostic resolution of abnormal breast and cervical screening exams in association with patient race/ethnicity language, and age concordance with patient navigators. Research Mentor: Karen Freund, M.D., MPH.

2001 Summer Medical Research Program, University of Iowa College of

Medicine. Investigation of Gene Expression in RNA of Adipose Tissue. Preceptor: Patricia Donohoue, M.D.

PUBLICATIONS

Obesity in relation to lung cancer incidence in African American women. Bethea TN, Rosenberg L, Charlot M, O’Connor GT, Adams-Campbell LL, Palmer JR. Cancer Causes Control. Epub 2013 June 7.

Triple negative breast cancers are increased in black women regardless of age or body mass index. Stead LA, Lash TL, Sobieraj JE, Chi DD, Westrup JL, Charlot M, Blanchard RA, Lee JC, King TC, Rosenberg CL. Breast Cancer Res. 2009;11(2):R18. Epub 2009 Mar 25.

Localized Amyloid of the Breast: A Case Series. Charlot M, Seldin DC, O’Hara C, Skinner M, Sanchorwala V. Amyloid. June 2011 18(2): 72-75. Poster presentation at ASCO Breast Cancer Symposium October 2010.

Figure

Updating...

Related subjects :