An Evaluation of Breast and Cervical Cancer Screening Outcomes in an Education and Patient Navigation Program in Rural and Border Texas

Abstract

Purpose:

This study examines breast and cervical cancer screening uptake in a cancer education and patient navigation (PN) program for residents of rural and border counties in Texas by level of participation (education only, PN only, or education and PN).

Methods:

Data collected from March 1, 2012 to November 5, 2016 included 6,663 follow-up surveys from participants aged 21–74. Logistic regression models assessed program participation on the odds of completing breast or cervical cancer screening.

Results:

For women aged 40–74 years (N=4,942; mean age= 52 years), 58.4% reported a mammogram within 6 months on average from initial contact. In the breast cancer screening model, women who only received PN (OR: 6.06, CI: 4.87–7.53) or who participated in both the education plus PN program (OR: 3.33, CI: 2.77–4.02) had higher odds of mammogram screening compared to women who only received education. For women aged 21–64 years (N=6,169; mean age= 46 years), 37.7% received a Papanicolaou (Pap) test within 6 months on average from initial contact. In the Pap screening model, both education and PN (OR: 3.23, CI: 2.66–3.91) and PN only (OR: 2.35, CI: 1.88–2.93) groups had higher odds of screening for cervical cancer compared to those only receiving education. Graphed predicted probabilities examined significant interactions between race/ethnicity/language and program participation (P<0.0001) for both screenings.

Conclusions:

PN, solely or in combination with education, is an effective strategy to increase screening for breast and cervical cancer, beyond educational outreach efforts alone, among un-/underserved, racially/ethnically diverse women in rural and border Texas counties.

Introduction

There are persistent disparities in female breast and cervical cancer screening by race/ethnicity, educational attainment, and rural-urban residence. Research has found that Latina women (60%) and those with less than a high school education (52%) have the lowest rates of mammogram screening in the US [1]. Further evidence suggests that Papanicolaou (Pap) screening rates of approximately 65% are below national goals of 93% coverage and appear to be declining with fewer Latina and non-Latina black women receiving the test compared to non-Latina white women [2]. In addition, rural women screen for breast and cervical cancer at lower rates compared to urban women [3]. These findings may explain related trends in breast and cervical cancer mortality outcomes among these sub-populations. For example, non-Latina white women have the highest incidence of breast cancer, yet non-Latina black women suffer the highest mortality from breast cancer compared to any other racial/ethnic group [4]. While Latina women have the highest cervical cancer rates in the U.S. followed by non-Latina black women, the mortality statistics invert these rankings with non-Latina black women experiencing the highest cervical cancer mortality rates in the country followed by Latina women [4].

Given these disparities in cancer screening and outcomes, cancer education and patient navigation (PN) have focused on assisting disadvantaged sub-populations of women with screening for breast and cervical cancer [5, 6]. Reviews of educational interventions reported an increase in screening uptake for breast and cervical cancer screening by increasing knowledge and promoting decision making with clinical providers [7, 8]. However, education and PN interventions were generally lacking in rural contexts potentially adding to the barriers for these populations [6]. In a recent review of PN literature, the largest proportion of studies was dedicated to screening uptake of various cancers, but the use of cancer education for participants in conjunction with PN was less clear [9]. Some programs mentioned participant focused education as part of PN activities; however, the analysis lacked a description of the cancer education portion to determine how it was administered.

Analyses of cancer education programs for residents in rural and border areas demonstrated increases in knowledge and screening outcomes. For example, an education program for Arizona-Mexico border residents aged 50+ reported significantly higher odds of mammogram and Pap screening [10]. Also, a community education intervention addressed breast and cervical cancer education gaps for Latina farmer workers in rural Washington [11]. Increases in breast and cervical cancer screening have also been found in studies of cancer education [12, 13]. Meanwhile, a cervical cancer education program for Latina farmworkers in southeastern Georgia increased screening knowledge in a separate study [14].

PN began as an intervention model in the 1990s to address barriers to care and connect lower socioeconomic (SES) and racially/ethnically diverse communities to supportive resources for women with breast cancer. PN has historically and traditionally used ethnically-matched minority navigators corresponding to their patient population [15]. Subsequent studies have demonstrated PN’s efficacy for increasing breast and cervical cancer screening uptake [5, 16]. PN roles vary broadly from specific roles in complex clinical environments that may be focused on one cancer site (e.g., breast or colorectal) to community-based outreach that links individuals to health care systems, complicating comparisons among studies [17–19]. However, no studies have evaluated the impact on breast and cancer screening of cancer education and PN programs individually compared to the combination of cancer education and PN, particularly among rural and border areas of Texas [9].

To address this gap, this study assesses breast and cervical cancer screening uptake among participants in a cancer education and PN program for residents of rural and border counties in Texas. Screening outcomes are compared by level of program involvement (education alone, PN alone, or education and PN) to evaluate the best approach for increasing receipt of mammogram and Papanicolaou (Pap) exams among a racially/ethnically and educationally diverse population of women.

Methods

Intervention

Friend to Friend plus Patient Navigation (FTF+PN) was established in 2012 through a grant from the Cancer Prevention Research Institute of Texas (CPRIT) to expand an existing evidence-based cancer education program with PN services [20]. FTF+PN aimed to increase breast and cervical cancer screening rates according to American Cancer Society (ACS) guidelines for breast and cervical cancer. These guidelines recommended annual mammograms for women aged 40–54 and biannual mammograms for those aged 55+ with average risk of breast cancer [21] and Papanicolaou (Pap) tests every 3 years for women aged 21–29 and every 5 years for women aged 30–64, with no additional screenings needed for women aged 65+ based on previously normal results [22]. All women were welcome to participate in the program; however, the staff attempted to recruit lower income, un-/underinsured women residing in more than 70 rural and border counties across Texas. The program also focused on women aged 40+ who may be disabled, self-employed, and/or have limited English proficiency.

The grants provided funding for clinical services and four trained, lay patient navigators to complement an existing team of five regional cancer specialists. The regional cancer specialists were employed by Texas Agrilife Extension and carried out the original FTF education program. The grant-funded lay navigators and regional cancer specialists were residents of the communities they served. The entire navigation and cancer specialist team consisted of five bilingual/bicultural Latina women, two non-Latina black women, and two non-Latina white women. They were tasked with follow-up for program participants and navigating participant barriers to screening, such as cost and transportation.

Recruitment Procedures

FTF recruitment

FTF leveraged county extension offices to recruit women to attend a “pink party” that educated them about cancer screening, shared a breast or cervical cancer survivor’s story about their cancer journey, and included a presentation by a medical professional (oncologist, nurse, radiologist, etc.), who answered questions from the women regarding screening. These events usually occurred once per program year and rotated within each region based on the capacity of the local navigators and the requests of the counties in the region. Attendees of the “pink party” completed the pre- and posttest surveys (available in English or Spanish) and could opt to include their contact information to receive a follow-up interview. Fliers were circulated in the county to invite residents to attend. Community partnerships with local health care providers also served as a source of referrals to the education events. Further, FTF organizers spoke with groups of women at churches, community gatherings, health promotion events, and other venues to invite women to participate in the event. In counties with large populations of women with limited English proficiency, the parties were also conducted in Spanish.

PN recruitment

Participants were engaged in navigation either (1) at the education event, or (2) through direct contact with the program staff outside of the annual education event. Participants from the “pink party” completed a simple paper form (help request) with their name and telephone number at the end of the event requesting help with screening. PN staff completed a PN intake form, if appropriate, based on recommended screening intervals and barriers faced by the participant.

Women who requested help outside of the education event provided their contact information directly to the program staff and received the same intake evaluation indicating their service needs and barriers to screening. For example, a woman was told by a local primary care physician of the program payment assistance to receive a mammogram. Rather than wait for the education event, she could immediately request PN services to get connected with a provider and receive a payment voucher. In this case, the participant would not complete a pre- and posttest as she did not attend the education event but still received PN services.

Follow-up

Participants were contacted approximately 6 months after the education event or initial contact with program staff. Ideally, this occurred within 6 weeks; however, programmatic barriers such as establishing contracts with providers, identifying providers willing to screen participants, and variation in follow-up procedures by region delayed follow-up for some women. Women who did not provide contact information at the education event did not receive follow-up, and the program staff could not determine their screening status. For PN participants, contact information was inherent to the process as program staff had multiple conversations with participants via telephone, text, and email arranging services, payment, transportation, etc. Follow-up usually occurred shortly after screening as the navigators received confirmation from the providers of the screening date if the grant funds reimbursed the screening services. A follow-up survey recorded demographic information, receipt of the screenings, screening outcomes, and the need for additional services from participants providing contact information at the education events but not requesting PN and from participants who opted for PN, regardless of the attendance at the “pink party”. Follow-up surveys were also conducted in Spanish for Spanish-speaking participants. Details of the recruitment and data collection process have been previously described [23].

Measurement

Demographics

Demographic variables included age (year of birth), educational attainment (less than high school, high school graduate or those with a GED, or some college or more), race/ethnicity (Latina, non-Latina black, or non-Latina white), and language use at home for Latina women (English only, Spanish only, or Spanish and English equally). This delineation created four distinct racial/ethnic/language categories: English speaking Latina, Spanish speaking Latina, non-Latina black, and non-Latina white women.

Program classification

The variation in “pink party” attendance and PN participation meant that participation-level could be classified into three categories. Participants were identified as: 1) education only participants if they attended the FTF “pink party”, did not request PN services, and received a follow-up interview; 2) education and PN participants if they attended the FTF “pink party”, requested PN services, and received a follow-up interview; 3) PN only participant if they did not attend the FTF “pink party”, requested PN services, and received a follow-up interview.

Screening status

The follow-up surveys recorded the date of the screening and the result of the screening if known at the time of the follow-up interview based on self-report from the individual. If the screening required payment from the grant funds rather than other sources, the screening occurrence was also confirmed with the providers at time of payment. However, not all women used the grant funds to pay for screening (e.g., women with insurance that needed transportation to a provider). A dichotomous screening outcome variable (screened/not screened) was calculated based on the screening date for both mammogram and Pap screenings.

Data

Evaluation data collected from March 1, 2012 to November 5, 2016 included 6,663 follow-up surveys from unique program recipients aged 21–74 who had participated in (1) the education program only, (2) the education and PN program, or (3) the PN program only. Furthermore, only Latina, non-Latina black, and non-Latina white women were included in the analysis as other racial/ethnic categories had small sample sizes. Those who had not received follow-up (n=968) were excluded from the analysis. This category included women who did not provide contact information and women waiting for follow-up interviews at the end of the program. Although follow-up surveys were prioritized during an extension period of the final service award, a backlog existed of participants without follow-up interviews. Analyses of breast cancer screening outcomes were limited to women aged 40–74 (N=4,942), while cervical cancer screening outcomes were limited to women aged 21–64 (N=6,169). The Institutional Review Board (IRB) of The University of Texas at Austin reviewed and approved (FWA # 00002030) the proposed study prior to analysis.

Statistical Analysis

Analyses for breast and cervical cancer screening were completed separately; women participating in the program had the option to receive both screenings and could be included in both analyses based on their age. Univariate analyses described the distribution of the samples by demographic characteristics, geographic region, year of participation, and mammogram or Pap screening completion. Bivariate analyses included these same variables and compare the proportions by program participation using Chi-square tests. Finally, logistic regression models assessed the role of demographic characteristics, geographic region, program year, and program participation on the odds of breast and cervical cancer screening. Interaction terms among race/ethnicity/language and program participation were also examined in the regression models. Analyses were conducted using SAS software, version 9.4 (SAS Institute, Cary, NC).

Results

Univariate Analyses of Mammogram Screening

The analytical sample for mammogram screening (N=4,942) consisted of women 40–74 years of age, with 90% of the women aged 40–64; 58.4% of participating women received a mammogram within 6 months on average from initial contact (Table 1A). Latina women comprised the largest proportion of respondents (42.5% Spanish speaking Latinas and 9.7% English speaking Latinas), followed by non-Latina white women (39.0%) and non-Latina black women (8.8%). The sample was roughly equally divided into educational groups (34.7% less than a high school education, 27.3% high school education, and 38.0% college education). The East (38.5%) and South (31.6%) regions had the greatest number of participants followed by the North (18.7%) and West (11.2%) regions. PN only had the highest proportion of women that screened for breast cancer (74.3%), followed by education and PN (62.9%), and education only (28.4%).

Table 1.

Sample distribution for mammogram and Papanicolaou (Pap) screening by demographic characteristics and program participation among follow-up respondents

	Table 1A. Mammogram Screening				Table 1B. Pap Screening
	Full Sample (N=4,942)	Education Only (n=1,254)	Education and Patient Navigation (n=1,860)	Patient Navigation Only (n=1,828)	Full Sample (N=6,169)	Education Only (n=1,290)	Education and Patient Navigation (n=2,513)	Patient Navigation Only (n=2,366)
Age (years)
21–39	-	-	-	-	27.9	28.8	30.4	24.7
40–64	90.0	73.2	94.0	97.5	72.1	71.2	69.6	75.3
65–74	10.0	26.8	6.0	2.5	-	-	-	-
Race/ethnicity/language
English speaking Latina	9.7	7.7	3.6	14.9	11.1	19.9	12.4	4.8
Spanish speaking Latina	42.5	15.2	11.4	4.1	46.1	15.7	68.7	38.6
Non-Latina black	8.8	13.2	66.9	37.9	8.6	6.4	3.3	15.6
Non-Latina white	39.0	64.0	18.1	43.1	34.2	58.0	15.7	41.0
Education level
< high school	34.7	9.4	48.9	37.5	35.0	9.5	47.8	35.4
High school	27.3	25.0	27.4	28.8	27.6	23.0	28.2	29.5
Some college+	38.0	65.6	23.7	33.6	37.4	67.6	24.0	35.1
Program region
North	18.7	48.1	15.9	1.4	16.2	46.5	14.9	1.1
East	38.5	18.5	11.5	79.8	38.2	16.8	10.9	78.9
West	11.2	17.1	11.1	7.2	10.8	18.8	10.5	6.9
South	31.6	16.4	61.6	11.7	34.7	17.9	63.6	13.1
Program year
2012	10.3	24.9	8.7	1.9	8.6	23.9	7.0	1.9
2013	28.4	23.1	29.5	31.1	27.3	22.8	27.1	29.9
2014	30.4	20.0	30.2	37.6	30.6	19.7	28.4	38.8
2015	13.9	14.0	16.8	10.8	14.9	15.4	17.9	11.4
2016	17.1	18.1	14.9	18.5	18.7	18.3	19.5	18.0
Screening status
Not screened	41.6	71.6	37.1	25.7	62.3	83.3	44.8	69.6
Screened	58.4	28.4	62.9	74.3	37.7	16.7	55.2	30.4

Note: Statistics are % for each variable category of the column total; P-value was <0.0001 for Chi-square tests of all variables by program participation for both mammogram and Pap screenings.

Multivariate Models of Mammogram Screening

In the logistic regression model for breast cancer screening, women who only received PN (OR: 6.06, CI: 4.87–7.53) or who participated in both the education plus PN program (OR: 3.33, CI: 2.77–4.02) had higher odds of mammogram screening compared to women who only received education (Table 2A). English speaking Latina women experienced lower odds of receiving a mammogram (OR: 0.55, CI: 0.43–0.70) compared to non-Latina white women, and women residing in the South region had higher odds screening (OR: 1.51, CI: 1.22–1.86) compared to those in the North region. Compared to the first year of the program, women who participated in 2014 had higher odds of receiving a mammogram (OR: 1.37, CI: 1.08–1.73).

Table 2.

Odds of mammogram and Papanicolaou (Pap) screening among follow-up respondents

	Table 2A. Mammogram Screening	Table 2B. Pap Screening
	OR (95% CI)	OR (95% CI)
Age (years)
21–39	-	1.40 (1.23–1.59)***
40–64	1.05 (0.85–1.30)	Ref.
65–74	Ref.	-
Race/ethnicity/language
English speaking Latina	0.55 (0.43–0.70)***	0.51 (0.40–0.64)***
Spanish speaking Latina	1.10 (0.91–1.34)	1.04 (0.87–1.24)
Non-Latina black	0.92 (0.73–1.17)	1.00 (0.79–1.25)
Non-Latina white	Ref.	Ref.
Education level
< high school	Ref.	Ref.
High school	1.04 (0.87–1.23)	0.96 (0.83–1.13)
College+	1.05 (0.87–1.26)	0.97 (0.82–1.14)
Program region
North	Ref.	Ref.
East	1.17 (0.94–1.46)	0.98 (0.77–1.24)
West	0.98 (0.77–1.24)	1.03 (0.80–1.33)
South	1.51 (1.22–1.86)***	4.88 (3.96–6.02)***
Program year
2012	Ref.	Ref.
2013	0.98 (0.75–1.27)	0.83 (0.65–1.07)
2014	1.37 (1.08–1.73)***	0.88 (0.69–1.13)
2015	0.98 (0.75–1.27)	0.63 (0.48–0.82)***
2016	0.92 (0.71–1.18)	0.53 (0.41–0.70)***
Program
Education only	Ref.	Ref.
Education and patient navigation	3.33 (2.77–4.02)***	3.23 (2.66–3.91)***
Patient navigation only	6.06 (4.87–7.53)***	2.35 (1.88–2.93)***

^*P <0 .05,

^**P < 0.01,

^***P <0.001

The interaction among race/ethnicity/language and program participation was significant in a subsequent regression model (P<0.0001, results not shown), and predicted probabilities for this interaction were graphed to interpret the relationship (Fig. 1A). Probabilities for women receiving education ranged from 18% for English speaking Latinas to 38% for non-Latina white women. Among women receiving education and PN, English speaking Latina, non-Latina black, and non-Latina white women were clustered from 45–50%; however, Spanish speaking Latina women had a 70% probability of receiving a mammogram. Those receiving PN only were virtually indistinguishable by race/ethnicity/language at approximately 70%.

Fig. 1.

Predicted probabilities of mammogram and Papanicolaou screening by race/ethnicity/language and cancer screening program participation.

Univariate Analyses of Pap Screening

The sample for cervical cancer screening consisted of women aged 21–64 (N=6,169), and 37.7% received a Pap test within 6 weeks of initial contact (Table 1B). Spanish speaking Latina women represented the largest proportion (46.1%) of participants, while non-Latina black women were the fewest (8.6%). Most had some college education or more (37.4%), and 35.0% had less than a high school education. The East region had the largest proportion of participants (38.2%) followed by the South (34.7%), North (16.2%), and finally the West (10.8%). Only 16.7% of education only participants received a Pap test, while 55.2% of education and PN participants and 30.4% of PN only participants screened for cervical cancer.

Multivariate Models of Pap Screening

In regression models for Pap screening, both education and PN (OR: 3.23, CI: 2.66–3.91) and PN only (OR: 2.35, CI:1.88–2.93) groups had higher odds of screening for cervical cancer compared to those receiving education only (Table 2B). Women aged 21–39 had higher odds of cervical cancer screening (OR: 1.40, CI: 1.23–1.59) compared to older women, while English speaking Latina women (OR: 0.51, CI: 0.40–0.64) had lower odds compared to non-Latina white women. Participants in the South region (OR: 4.88, CI: 3.96–6.02) experienced increased odds of receiving a Pap test than the North region, and women in the last two years of the program had reduced odds of screening compared to those in the first year (OR: 0.63, CI: 0.48–0.82 in 2015 and OR: 0.53, CI: 0.41–0.70 in 2016).

A significant interaction was observed between race/ethnicity/language and program participation (P<0.0001, results not shown) and graphed for interpretation in Fig. 1B. Among women receiving education only, English speaking Latina women only had an 18% probability of receiving a Pap test, while non-Latina white women had a 50% probability. For those receiving education and PN, Spanish speaking Latina women had the highest probability at 70%, followed by non-Latina white (62%), non-Latina black (60%), and English speaking Latina women (50%). Women who received PN only ranged from 58% for non-Latina white women to 62% for English speaking Latina women.

Sensitivity Analyses

Sensitivity analyses tested for variation in the sample by program participation and follow-up status. Women attending the education program provided the main source of PN participants (64%); however, other outreach activities conducted by the program staff identified women in need of PN who may not have been able to attend the initial educational portion (36%). A comparison of these two groups showed significant differences by demographic variables, region, and year of participation (Appendix A). Significant differences were also observed between education participants with and without follow-up interviews (Appendix B). Participants from the first year had a much longer timeframe to receive follow-up compared to participants in subsequent years, and follow-up interviews had not been completed for all intervention participants at the time of the analysis.

Discussion

This study aimed to evaluate three strategies for increasing breast and cervical cancer screening in a naturalistic manner among an un-/underserved population of racially/ethnically diverse women in rural and border Texas: cancer education alone, PN alone, or the combination of education and PN. For both mammogram and Pap, women who received education and PN and PN alone significantly benefitted compared to only receiving education. For mammogram, receiving PN alone accounted for greater screening. Meanwhile, the results were similar for women in the education and PN and PN only program for Pap screening except for Spanish speaking Latina women who benefitted most from the combination of education and PN.

The findings for PN alone illustrate a significant variation in the sample. Often, these women were actively seeking screening independently and self-selected into PN. Considering those who participated in the education program, the hypothesized improvement in screening rates was evident in the results. For both cancer screenings, women who received education alone were least likely to be screened, especially those with the lowest education level. This suggests these women needed the additional support of patient navigation to overcome their barriers to screening, which were often financial.

The language used at home was an important distinguishing characteristic in terms of cancer screening uptake. English speaking Latina women were less likely than non-Latina white women to screen for both breast and cervical cancer. This disparity was not observed when comparing Spanish speaking Latina women to non-Latina white women. Although place of birth was not ascertained in this study, the preferred language at home was used a proxy measure for level of acculturation and a possible indication of country of origin in a state where 87% of Latina women originate from Mexico [24]. Single-item responses to preferred language at home have been used as an indication of acculturation in previous studies [25]. Thus, use of Spanish versus English at home provides greater insights into the differences among Latina women in this sample. Spanish speaking Latina women also represented the largest share of the participants and reported some of the highest screening rates for both mammograms and Pap tests. These results may stem from the many FTF events conducted in Spanish, as well as five of the nine program staff being bilingual/bicultural.

Also, these findings highlight the relationship among geography, poverty, access to insurance, health care infrastructure, race/ethnicity, and language use in Texas. The South region covered the Texas-Mexico border counties where the largest proportion of the Spanish speaking population resides. These counties also represent the most impoverished areas in the state with the highest rates of uninsured residents in Texas and limited or nonexistent health care facilities to perform the screenings [26]. FTF+PN provided services regardless of documentation status, which may have attracted participation from a larger group of Spanish speaking Latina women, many of whom may have not qualified for other screening programs due to eligibility barriers and may not be covered by Medicaid, Medicare, or private sources of insurance.

This analysis offers a unique perspective into the cancer screening behavior of a large, yet understudied population of women in rural and border counties in Texas. The evaluation benefitted from diverse participation across a large area of the state that has many challenges in receiving adequate access to high quality cancer care. As Texas did not expand Medicaid coverage available from the Affordable Care Act, FTF+PN offered an important mechanism to broaden access to breast and cervical cancer prevention unavailable to many un-/underserved individuals and communities through traditional health insurance coverage [27]. Moreover, the lack of available health care infrastructure to service these communities made providing access to this vulnerable population challenging. Regardless, the program staff persevered in broadening access to breast and cervical cancer screening to a population that greatly needs these services considering the disparate cancer outcomes observed by geography, educational attainment, and race/ethnicity/language.

Limitations

The interpretation of the results of this study are limited in several ways. First, the evaluation is based on women who opted into various types of program participation and may or may not have had the opportunity to participate in both education and patient navigation. Program implementation varied by region to account for local differences in the population and available resources. While this approach favored adaptation based on participants’ needs, the lack of fidelity to a single protocol limits comparability of the results across regions.

Follow-up procedures varied from region to region with different roles for patient navigators and cancer prevention specialists, whereby leading to differences in follow-up screening rates. For example, the additional patient navigator in the South region increased their capacity to perform follow-up interviews at higher rates than the other regions. Variation in participant recruitment led to many women participating in the PN intervention but not the FTF portion. While this encouraged screening among women who could not or would not attend the education intervention, participants did not complete pre- and posttest surveys during the events that would have contributed greater insights. Timely follow-up was also problematic because navigators had to establish contracts with a number of clinical providers for screening. This process delayed initial screening and follow-up for early participants.

Finally, all responses were based on self-report with possible recall bias. Both SES and acculturation were measured using a single item based on educational attainment and language use at home. The item for educational attainment was limited to three categories that may not have best represented the full range of educational attainment in the sample. The findings would be strengthened with greater details on number of years of education and degrees completed. Additional items surveying income, wealth, and occupation would strengthen the validity of the SES findings, while more comprehensive measures of acculturation would refine the interpretation of screening behavior among Latina women. For example, data on location of birth, time of residence in the U.S., education history in the U.S., and generation were not available.

Conclusion

This study suggests that women in rural and border Texas screen at higher rates with supportive services compared to only receiving cancer education. PN also serves as an effective strategy to reach un-/underserved, racial/ethnic minority women in these areas. FTF+PN serves a model of health care engagement in areas with limited health care resources that not only connect women to cancer screening, but also engage disadvantaged populations in other aspects of health care. Health care systems, both fragmented and integrated, should include supportive care services that address social determinants of health to improve outcomes for vulnerable populations. Future studies should assess implementation and adaptation of PN in other resource-limited areas to assess the efficacy of cancer education and PN in a wider range of settings and more rigorous study designs.

Appendix A. Chi-square tests comparing the proportions of participants aged 21–74 who received cancer education and those who only received patient navigation (n=6,663), 3/1/12–11/5/16.

	Education recipient (n=4,251)	Patient Navigation Only recipient (n=2,412)	P-value
Age			<0.0001
21–39	26.8	24.2
40–64	62.7	73.9
65–74	10.5	1.9
Race/ethnicity/language			<0.0001
English speaking Latina	4.6	15.6
Spanish speaking Latina	14.2	4.9
Non-Latina black	47.5	38.6
Non-Latina white	33.7	41.0
Education level			<0.0001
< high school	32.8	35.5
High school	26.6	29.5
College+	40.6	35.0
Program region			<0.0001
North	27.4	1.2
East	13.8	78.6
West	13.5	7.0
South	45.4	13.1
Program year			<0.0001
2012	14.0	2.0
2013	25.6	30.4
2014	25.0	38.6
2015	16.5	11.2
2016	18.9	17.9

Note: Statistics are % for each variable category of the column total

Appendix B. Chi-square tests comparing the proportions of program participants aged 21–74 without follow-up interviews to participants with follow-up interviews (n=7,631), 3/1/12–11/5/16.

	No follow-up (n=968)	Follow-up (n=6,663)	P-value
Age			<0.0001
21–39	23.8	25.8
40–64	55.7	66.8
65–74	20.6	7.4
Race/ethnicity/language			<0.0001
English Speaking Latina	8.7	10.8
Spanish Speaking Latina	18.2	44.3
Non-Latina black	6.6	8.6
Non-Latina white	66.5	36.4
Education level			<0.0001
< high school	11.5	33.8
High school	23.7	27.7
College+	64.9	38.6
Program region			<0.0001
North	34.1	17.9
East	33.8	37.2
West	20.1	11.1
South	12.8	33.7
Program year			<0.0001
2012	26.7	9.8
2013	21.9	27.3
2014	21.1	29.9
2015	10.9	14.6
2016	19.5	18.5

Note: Statistics are % for each variable category of the column total