Exploring disparities in post-cancer treatment instructions: an analysis of rural vs. urban breast cancer survivors in Missouri using BRFSS data Authors: Allison B Anbari, Knoo Lee, Jennifer M Hulett, Lada Micheas, Zewei Ren, Anthony Belenchia, Suhwon Lee Published: 2024 ABSTRACT AbstractBackgroundThere are more than 4 million breast cancer survivors in the United States. With continuing improvements in early detection and treatment, the number of breast cancer survivors will only continue to increase. Breast cancer survivors face a lifetime risk of long-term or late-effects from cancer treatments, thus post-cancer treatment care, referred to as survivorship care, is critical. Social determinants of health, including where a breast cancer survivor lives, may impact cancer survivorship care. Our purpose was to evaluate the relationship between rural/urban status and receipt of cancer survivorship care and instructions for Missouri breast cancer survivors using state-level data from a nationwide telephone survey—the Behavioral Risk Factor Surveillance System (BRFSS).MethodsMissouri included the BRFSS Cancer Survivorship Module in 2014, 2016, 2018, and 2020. Module items ask cancer survivors about receipt of a “post-cancer-treatment summary” and “follow-up instructions.” We hypothesized chances of receipt of both components would differ among four level Rural–Urban Commuting Area (RUCA) Code groups (Rural, Small-Town, Micropolitan, Metropolitan). Data from 430 breast cancer survivors over four survey years were combined and grouped according to RUCA status. Using a logistic model with a multilevel approach (after sampling weights calibration), the relationship between receipt of survivorship instructions and RUCA group was examined.Results46% of Missouri breast cancer survivors reported receipt of complete survivorship instructions. Compared to rural breast cancer survivors, micropolitan breast cancer survivors had 5.9 times higher odds of receiving survivorship instructions (p < 0.0001).ConclusionsLess than half of respondents reported receiving complete post-cancer treatment instructions. More urban respondents were more likely to receive instructions than their rural counterparts which raises questions about the quality of post-treatment care between rural and urban survivors. A sophisticated and purposeful approach to mitigating potential disparities is warranted. Receipt of cancer survivorship care plans may be impacted by rurality/urbanicity. Clinicians caring for cancer survivors living in rural or urban settings must continue to be diligent with providing personalized survivorship care.Keywords: Breast cancer, Survivorship care planning, Behavioral Risk Factor Surveillance System, Rural health, Health disparities Background Breast cancer survivorship There are more than 4 million breast cancer survivors in the United States [1]. With continuing improvements in early detection and treatment, the number of breast cancer survivors will only continue to increase. Breast cancer survivors face a lifetime risk of long-term or late-effects from cancer treatments. As breast cancer survivors live longer without recurrence, they are encouraged to transition from oncological to primary care [2]. Survivorship care plans, traditionally comprised of a cancer treatment summary and a plan for follow-up care, have been recommended to help ease this transition and inform post-cancer treatment care. However, outcomes for use of survivorship care plans are inconsistent. Furthermore, primary care clinicians, perhaps especially those in rural settings, cite the need for additional information and resources about cancer survivorship care [3–5]. Context of Missouri including breast cancer and survivorship Our research team is in Missouri and focuses largely on the breast cancer care continuum there. As such, we prioritized the context of Missouri and breast cancer for this study, anticipating the study approach and results could be useful to research teams in other states. While Missouri’s breast cancer incidence and mortality rates are slightly above the national average this trend is common among Midwestern US states, where demographic and healthcare access factors often overlap [6]. This made Missouri an effective starting point, providing insights that could be broadly applicable across similar regions. Likewise, breast cancer is perhaps unique in terms of survivorship because of its high survival rate yet nonlinear trajectory of recovery and the potential for long term or late effects resulting from treatment [7, 8] Approximately 20% of the United States (US) population and 33% of the people in Missouri live in rural counties [9–11], where residents have poorer cancer outcomes than urban dwellers [12–14]. Survivorship concerns are often amplified for rural breast cancer survivors due to contextual influences such as distance from care, later cancer stage at diagnosis, and lower adherence to post-treatment recommendations [14–17]. The context for urban breast cancer survivors is also important to consider as previous work found symptom burden among this group can be higher than their rural counterparts [18]. In general, breast cancer survivors engage in screening activities, such as mammograms and Pap-smear tests, more often than those with histories of other cancers and other women [19]. Rural breast cancer survivors also express similar concerns to urban survivors [16, 20], with both groups desiring education and continuing support during survivorship [20]. However, both contexts in Missouri are unique [14]. Cancer burden in Missouri The cancer death rate in Missouri is the 10th highest in the US [21]. In 2016, just 57% of Missouri women diagnosed with breast cancer received treatment within the National Comprehensive Cancer Network recommendation of 30 days [22]. Cancer is the second leading cause of death for both rural and urban Missourians [11, 21]. Although there are downward trends in cancer mortality, the rural cancer death rate is significantly higher than the goals presented in Health People 2020 [11, 23]. Of Missouri’s 115 counties, 99 are considered rural (when defining rural as having a population density of less than 150 persons per square mile [11], estimated Missouri rural residents = 2.07 million). This translates to approximately one in three Missourians residing in a rural county where the poverty rate is 16%, while the urban poverty rate is 12%. The per capita income is 29% lower in Missouri rural counties than in urban counties and 84 of the state’s 106 Primary Medical Care Health Professional Shortage Areas (HPSA) are in rural areas [11]. The Missouri Comprehensive Cancer Control Program has a goal within its strategic plan to assure the highest quality of life possible for people with a history of cancer. Rural health care faces unique challenges when it comes to the priorities of cancer prevention, control, and survivorship [12, 13, 23–26]. Where a person lives invariably guides their interactions with the health care system and health outcomes [27–31]. Additional exploration into how context should be considered in coordination with post-cancer (survivorship) care in Missouri is warranted. Purpose Given the importance of providing personalized cancer survivorship care, the purpose of this analysis was to begin examining the contextual factors that impact receipt of post-cancer treatment care instructions for breast cancer survivors in Missouri. Our research question was: Does the receipt of survivorship instructions have a relationship with the context of where a breast cancer survivor lives (rural–urban status)? We report analysis of the Missouri Behavioral Risk Factor Surveillance System (BRFSS) data and results related to urban and rural groups. Methods Compliance with ethical standards Through communication and vetting with lead analysts at the Missouri Department of Health and Senior Services, Office of Epidemiology, the study was deemed exempt from full ethical review board approval as the BRFSS is a de-identified, public-use dataset. Analysis of BRFSS data is not approached as human subjects research. We did not merge or manipulate the dataset to make data identifiable. A data use agreement was signed by the Department and the University’s Institutional Review Board. Dataset description The BRFSS is a health-related random-digit-dialed telephone (both landline and cell phone) survey that has been used by state health agencies since 1984. Now in its 42nd year, the BRFSS is the gold standard of behavioral surveillance and collects interview data from more than 400,000 adults living in the 50 United States, the District of Columbia, and three United States territories [32]. The BRFSS findings are used to identify health-related risk behaviors, health conditions, and the use of preventive services [33]. The state and local level data provide guidance in creating programs and health promotion activities. BRFSS data are weighted to adjust for the probability of being surveyed, nonresponses, and any deficiencies in the sampling frame. Motivated by the recommendations surrounding cancer survivorship care reported in the Institute of Medicine’s 2006 from Cancer Patient to Cancer Survivor: Lost in Transition [34], the Centers for Disease Control and Prevention with BRFSS introduced the Cancer Survivorship Module as an optional module to be used during data collection. Since the module first became available in 2009, Missouri has opted to use it in 2010, 2014, 2016, 2018, and 2020. It is one of the only states to have used the module this frequently. We analyzed data from the 2014, 2016, 2018, and 2020 BRFSS survey, forgoing inclusion of 2010 data because the cancer survivorship module items were different this year. We also limited the data to Missouri as it is a somewhat representative state with average land area, a few urban cores, and many rural counties (Missouri Department of Health and Senior Services) [35]. Variables A sample of breast cancer survivors in our study using the BRFSS dataset were determined using the following three items and answers: YES answer to the item “Have you ever been told you had cancer?” YES answer to the item “Have you ever been told you had cancer?” BREAST answer to the item “What type of cancer was it?’ and BREAST answer to the item “What type of cancer was it?’ and “NO, I’ve completed treatment” to the item “Are you currently receiving treatment for cancer?” “NO, I’ve completed treatment” to the item “Are you currently receiving treatment for cancer?” Respondents answering this way were asked a series of items referred to as the “Cancer Survivorship Module.” The module includes items about 1) receipt of a post-cancer treatment summary, 2) receipt of follow-up instructions, 3) whether those instructions were verbal or printed, 4) whether they had health insurance, and 5) whether they were experiencing pain resulting from cancer treatments. We concentrated on two of the Cancer Survivorship Module items in which respondents were asked: “Did any doctor, nurse, or other health professional EVER give you a written summary of all the cancer treatments that you received?” (cancer treatment summary) and a combined variable that included cancer treatment summary with the question “Have you EVER received instructions from a doctor, nurse, or other health professional about where you should return or who you should see for routine cancer check-ups after completing your treatment for cancer?” (complete plan). Initially, we used the cancer treatment summary as the outcome variable (n = 436 breast cancer survivors). After reviewing these results, we did an additional round of analysis using the combined variable we refer to as “complete plan” which combined the “receipt of a post-cancer treatment summary” variable with the “receipt of follow-up instructions” variable (n = 430 breast cancer survivors). We used Rural Urban Commuting Area (RUCA) codes to delineate rural and urban areas of Missouri. RUCA codes classify rurality and urbanicity at sub-county levels using census tracts and commuting flow [36]. RUCA codes are based on zip codes which represent areas smaller than counties and are recommended for use during rural–urban studies [13, 36]. RUCA codes factor in population density, urbanization, and commuting distances. RUCA status variables were computed using the primary codes (1 (most urban) to 10 (most rural)) and conversions from the United States Department of Agriculture website [36]. We grouped RUCA codes into four groups—Metropolitan (RUCA of 1 to 3); Micropolitan (RUCA 4 to 6); Small Town (RUCA 7 to 9); and Rural (RUCA = 10). See Fig. 1 [37]. Map of Missouri by RUCA Code (from https://allthingsmissouri.org/missouri-map-room/) Data analysis The BRFSS is a complex survey study designed with survey probability weights, geographical strata and primary sampling units that account for restriction to randomization and/or noncoverage and nonresponse during data collection [38]. This means that survey weights, strata and PSU are encouraged to be used in the data analysis to better represent the actual population. To increase the sample size for this analysis, we combined BRFSS data from years 2014, 2016, 2018 and 2020 and the final survey weight variable was adjusted proportionally for the combined data set [38]. On the other hand, the existing literature [30] also suggests that a multilevel approach can be used to analyze cancer data as there can be some additional variation in responses due to zip code. For example, zip code can be considered an extra macro level in the model.”Macro-level factors will often be the Level II factors that characterize an individual’s geographic context (e.g. census tract, zip code, county) in multilevel models" [30]. We proceeded with converting zip codes to RUCA codes and used those statuses as zip code level predictors in our multilevel model. We did not focus on estimating variability due to zip code, but rather, we used the multilevel model to adjust for possible variation to get accurate standard errors when estimating fixed effects such as RUCA codes. To analyze our hypothesis that chances of receiving a cancer treatment summary or a complete plan would differ among the four level RUCA groups, we used the multilevel approach addressing complex survey data issues as well as possible extra macro levels due to zip code. To test the multilevel model with complex survey data we used guidelines outlined in Carle (2009), Diaz-Ramirez et al. (2020), and Zhu (2014) [39–41] using SAS 9.4 PROC GLIMMIX with adaptive quadrature. Adaptive quadrature approximation technique is one of the methods for parameter estimation available in SAS 9.4 PROC GLIMMIX and is recommended by for weighted multilevel models [41]. Survey weights were scaled “so that new weights sum to the cluster sample size” [39]. To predict the odds of receiving a complete plan we used a logistic model designed to predict a binary outcome [42]. The random effects of primary sampling units were nested under zip code and under geographical strata to account for variation due to complex survey design as well as possible clustering due to zip code. Preliminary testing on all available combined data not restricted to breast cancer survivors suggested that up to 7% of the variation in log odds of receiving a cancer treatment summary was due to the same zip code under the same geographical strata. For the complete plan variable, that variation was also 7%. Figure 2 provides a variation map for probabilities of receipt of a complete plan from the multilevel mixed model due to zip code random effect (not restricted to breast cancer survivors). Dark green indicates higher magnitudes (correlations) of receipt of complete plan and lighter green indicates lower magnitudes. Because the subpopulation command is not available in the PROC GLIMMIX procedure in SAS 9.4, we used a non-conditional approach by including a dummy variable for breast cancer survivorship in the model and tested the interaction of RUCA status and the dummy variable for breast cancer survivorship to avoid manually recomputing weights for the subpopulation of breast cancer survivors. Missouri County Map showing variations in chances of receiving a complete plan and support for use of multilevel modeling (generated using R 4.3.3 software’s tmap feature and BRFSS data) Results Description of sample A total of 436 breast cancer survivors were included from the four years of Missouri survey data collected to assess the receipt of a cancer treatment summary. For the receipt of a complete plan variable, we included 430 breast cancer survivors (428 female) from the four years of survey data. Description of the demographic composition of the sample is found in Table 1. The weighted proportion of respondents who were ages 18 to 54 at the time of the survey was 12.3% (95% CI = 7.41–17.2) and the remaining 87.7% proportion were over 55 years (95% CI = 82.8–92.6). The weighted proportion of respondents who were Non-Hispanic White was 92.7% (95% CI = 89.9–95.4). The weighted percentage of respondents who were retired or unable to work was 64.1% (95% CI = 57.1–69.7); 63.4% were married (95% CI = 57.1—69.7). RUCA status groups in the weighted sample were delineated as follows: Urban 68.6% (95% CI = 62.9—74.3); Micropolitan 9.5% (95% CI = (5.49—13.6); Small Town 8.4% (95% CI = 5.68—11.1); and Rural 13.5% (95% CI = (8.73—18.2). Of our sample, 196 or a weighted percentage of 51.4% with 95% CI = (44.5—58.3) of the respondents reported receiving a cancer treatment summary and 234 or 48.6% with 95% CI = (41.7—55.5) reported they did not. Table 1Sample DemographicsGroupAll(2014–2020, n = 430)N(weighted %, 95%CI)Receipt of Survivorship Instructions = Yes(2014–2020, n = 196)N(weighted %, 95%CI)Receipt of Survivorship Instructions = No(2014–2020, n = 234)N(weighted %, 95%CI)Sex Male2(0.89, 0.00—2.30)2(1.72, 0.00—4.46)0 Female428(99.11, 97.70—100)194(98.28, 95.54—100)234(100,100—NA)Age Ages 18–5432(12.31, 7.41—17.21)19(13.99, 6.76—21.23)13(10.52, 4.04—17.00) Ages 55 + 397(87.69, 82.79—2.59)177(86.01, 78.77—3.24)220(89.48, 83.00—5.96)Race Non-Hispanic White378(92.67, 89.92—5.42)160(89.94, 85.31—4.58)218(95.54, 92.38—98.69) Other48(7.33, 4.58—10.08)33(10.06, 5.42—14.69)15(4.46, 1.31—7.62)Four level RUCA status Metropolitan (1 to 3)253(68.61, 62.94—74.28)123(73.85, 65.24—82.45)130(63.07, 54.77—71.37) Micropolitan (4 to 6)55(9.54, 5.49—13.60)26(10.36, 3.12—17.60)29(8.68, 5.27—12.09) Small town (7 to 9)58(8.38, 5.68—11.09)25(7.04, 3.70—10.39)33(9.80, 5.39—14.20) Rural (10)64(13.46, 8.73—18.19)22(8.75, 3.40—14.10)42(18.46, 10.73—26.18) Sample Demographics 2 (0.89, 0.00—2.30) 2 (1.72, 0.00—4.46) 428 (99.11, 97.70—100) 194 (98.28, 95.54—100) 234 (100,100—NA) 32 (12.31, 7.41—17.21) 19 (13.99, 6.76—21.23) 13 (10.52, 4.04—17.00) 397 (87.69, 82.79—2.59) 177 (86.01, 78.77—3.24) 220 (89.48, 83.00—5.96) 378 (92.67, 89.92—5.42) 160 (89.94, 85.31—4.58) 218 (95.54, 92.38—98.69) 48 (7.33, 4.58—10.08) 33 (10.06, 5.42—14.69) 15 (4.46, 1.31—7.62) 253 (68.61, 62.94—74.28) 123 (73.85, 65.24—82.45) 130 (63.07, 54.77—71.37) 55 (9.54, 5.49—13.60) 26 (10.36, 3.12—17.60) 29 (8.68, 5.27—12.09) 58 (8.38, 5.68—11.09) 25 (7.04, 3.70—10.39) 33 (9.80, 5.39—14.20) 64 (13.46, 8.73—18.19) 22 (8.75, 3.40—14.10) 42 (18.46, 10.73—26.18) Group results by RUCA status To test the hypothesis that chances of breast cancer survivors receiving a complete plan differed for the four level RUCA groups we used a binary logistic model with a multilevel approach (four levels) with a non-conditional approach by including a dummy variable for breast cancer survivors in the model. The interaction term of RUCA group by breast cancer survivor dummy variable was significant in both models (F = 641.73, df = 3, p < 0.0001; F = 580.30, df = 3, p < 0.0001). The results obtained from the binary logistic model for survey data using SAS 9.4 proc SURVEYLOGISTIC (assuming no additional macro level due to zip code) were similar with the significant effects of status for breast cancer survivors on receipt of post-cancer treatment summary (F = 2.98, df = 3, p = 0.0301) and complete plan (F = 3.04, df = 3, p = 0.0278). Given these similarities, we report here the results of the complete plan analysis for our four-level model. The odds of receiving a complete plan for breast cancer survivors were the highest in the Micropolitan group (RUCA codes 4 to 6) (See Table 2). In particular, the odds of receiving a complete plan in this area were significantly 1.78 times higher than in the Metropolitan area (95%CI, 1.563,2.028), significantly 3.32 times higher than that of Small Town (95%CI,2.794,3.941) and significantly 5.90 times higher than that of the Rural group with (95%CI,4.971,6.934). Breast cancer survivors living in Rural RUCA areas had the lowest chances of receiving a post-cancer treatment summary and a complete plan. Table 2Post-hoc test results for the interaction term of RUCA group by breast cancer indicator variable for complete planOdds ratio for Complete Plan (Adjusted 95%CI)t-stat, degrees of freedomTukey adjusted P valueMicropolitan vs Metropolitan1.78**** (1.56—2.02)13.33, 262,000 < .0001Metropolitan vs Small Town1.86**** (1.63—2.12)14.28, 262,000 < .0001Metropolitan vs Rural3.31**** (2.91—3.73)29.10, 262,000 < .0001Micropolitan vs Small Town3.31**** (2.79—3.94)21.07, 262,000 < .0001Micropolitan vs Rural5.90**** (4.97—6.93)31.86, 262,200 < .0001Small Town vs Rural1.78**** (1.50—2.09)10.33, 262,200 < .0001**** indicates p < 0.001 Post-hoc test results for the interaction term of RUCA group by breast cancer indicator variable for complete plan **** indicates p < 0.001 Discussion The findings support our study’s hypothesis that receipt of survivorship care instructions for breast cancer survivors in Missouri varies by location within the state and that rural breast cancer survivors receive different (potentially less) information than their more urban counterparts. Additional analysis is warranted; however, our results also support a signal towards the status quo – people living with a history of cancer in rural areas may be at risk of receiving different or suboptimal care after completion of acute cancer treatment. Several factors may contribute to this phenomenon. Cancer screening rates in rural areas are lower [17, 26], which often leads to a more advanced cancer stage at diagnosis [43] and more intensive treatments. Such treatments often lead to increased health needs during survivorship which sheds additional light on the notion that rural breast cancer survivors should indeed receive thorough cancer survivorship care plans. Rural breast cancer survivors face challenges related to their inability to receive the correct care at the correct time [17, 24] and to access mental health care and health promotion activities [15–17]. Rural women with breast cancer also experience disparate financial impacts compared to urban women, including changes in household income and employment adjustments [44] which would impact their ability to pursue and carry out survivorship recommendations. Perhaps one of the most compelling findings was that 46% of breast cancer survivors in our analysis reported receiving a complete plan – meaning under half reported they received both treatment summaries and follow-up care instructions. This finding calls attention to a potentially a missing aspect of the standardized care that should be part of long-term surveillance following breast cancer treatment. This is concerning because the risk of breast cancer recurrence reportedly peaks in 15% of survivors during the first two years after initial breast cancer treatment [45]. This missing component of standardized care is also observed on the national level with only 9% of survey respondents (survivors of all cancers) in the most recent National Coalition for Cancer Survivorship Survey (n = 1,303) reported having discussions with their oncologists about a survivorship care plan [46]. A second key finding was that breast cancer survivors living in Micropolitan settings (i.e., typically more suburban areas) consistently reported receiving survivorship care plans more so than their Urban, Small Town, and Rural counterparts. This is likely a concern too, because as aforementioned, approximately 20% of the U.S. and 33.2% of the Missouri population live in rural counties [9–11]. Yet an estimated 5.6% of oncologists practice in these areas [47] which creates a potential barrier to ongoing communication/education about follow-up planning. This finding signals a potential disparity between the quality of post-treatment survivorship care for rural and urban breast cancer survivors. Along the same lines, it is known that there is a lower utilization of end-of-life indicators (e.g., emergency room visits, inpatient admissions, and days spent in intensive care units), among metropolitan populations compared to others, including micropolitan populations [48]. Additionally, there is an observed increase in cigarette and e-cigarette usage among metropolitan cancer survivors compared to micropolitan areas, suggesting potential disparities in healthcare knowledge and access [49]. However, there is a lack of comprehensive research focusing specifically on micropolitan or small town areas compared to other regions. This gap raises questions about the underlying causes of these disparities and how they can be effectively addressed. Additional research is required to better understand the nature of this finding and identify solutions to address this gap in survivorship care between differently located breast cancer survivors. A final factor to consider alongside these findings is that during the BRFSS survey years we included, survivorship care plans moved from being required to recommended [50, 51]. How this transition impacts our results is beyond the scope of our analysis, however it is important to consider that survivorship care in its most traditional format (i.e., treatment summary plus follow-up instructions) is in flux. There are ongoing discussions about: 1) the best way to deliver cancer survivorship care, and 2) which frameworks or delivery models optimize outcomes and are feasible from a sustainability perspective [52–54]. Context of delivery, such as rural or urban setting, will be important to consider as these discussions continue. Limitations The utilization of statewide data spanning numerous years in our study is significant as it lays the groundwork for findings that may be applied to a wider demographic regarding cancer survivorship and RUCA status in the future. However, the study does have limitations. Our purpose was to report on Missouri breast cancer survivors and thereby our sample size was small. And as such, our decision to limit analysis to Missouri-only and breast cancer-only may limit our findings’ applicability to other states and other cancers. We do maintain that because Missouri is one of 12 midwestern US states and is centrally located, our analysis and results may resonate with other states in our region. Additional analysis using multiple states and multiple cancers may heed even more robust and usable findings. The BRFSS is a telephone survey and represents the experiences of those breast cancer survivors willing to respond to the survey items. Likewise, those choosing to respond do so by self-reporting, so there may be recall and/or social desirability bias impacting the generalizability of our results. Within the survey items, there is not an opportunity to gather additional details about cancer treatment or survivorship care, e.g., where respondents received cancer treatment and/or did respondents choose not to attend follow-up appointments. Conclusions The clinical implications from our findings support important questions about health care policy and whether an important aspect of post-treatment care or follow-up is being overlooked following initial breast cancer treatment. Although survivorship care plans are no longer mandated, long-term surveillance is a well-established component of standardized care for breast cancer survivors. Post-treatment instructions are an expected component of this standardized care. Further research is needed to determine best practices for ensuring that survivors receive personalized and contextually derived post-treatment care instructions. Continued consideration of breast cancer survivors’ context (e.g., rurality or urbanicity) will be essential to improving quality of life and preventing new or recurring cancers. In conclusion, our findings offer evidence of a potential gap in post-treatment survivorship care quality for breast cancer survivors with less than half of survivors reporting receipt of a complete survivorship care plan. We also provide evidence that rural breast cancer survivors are less likely to receive post-treatment survivorship instructions compared to urban survivors. Although our results are from data from a single state, additional research is warranted to further examine this potential gap in survivorship care provision in other states. Acknowledgements We would like to acknowledge the biostatistics contribution of Chelsea DeRoche, PhD, in memoriam. Abbreviations Behavioral Risk Factor Surveillance System Rural-Urban Commuting Area Codes Authors’ contributions All authors contributed to the study conception and design. Study design and purpose were written by AA. Material preparation, data collection and analysis were performed by AA, KL, JH, LM, ZR, AB, and SL. Analysis was performed by LM, ZR, and SL. The first draft of the manuscript was written by AA, KL, and JH. Tables were prepared by LM and ZR. Discussion was prepared by AA, KL, JH, and AB. All authors commented on versions of the manuscript. All authors read and approved the final and revised manuscript. Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript. Data availability No datasets were generated during this study. BRFSS data is publicly available at https://www.cdc.gov/brfss/annual_data/annual_data.htm or upon request to the Missouri Department of Health and Senior Services. Please see https://health.mo.gov/data/brfss/index.php for additional details. Declarations Ethics approval and consent to participate Through communication and vetting with lead analysts at the Missouri Department of Health and Senior Services, Office of Epidemiology, the study was deemed exempt from full ethical review board approval as the BRFSS is a de-identified, public-use dataset. Consent for publication Not Applicable Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement No datasets were generated during this study. BRFSS data is publicly available at https://www.cdc.gov/brfss/annual_data/annual_data.htm or upon request to the Missouri Department of Health and Senior Services. Please see https://health.mo.gov/data/brfss/index.php for additional details.