Preventive behaviors and behavioral risk factors among gynecologic cancer survivors: Results from the 2020 Behavioral Risk Factor Surveillance System Survey

Abstract Background Maintaining a healthy lifestyle is an important factor in promoting positive outcomes for gynecologic cancer survivors. Methods We examined preventive behaviors among gynecologic cancer survivors (n = 1824) and persons without a history of cancer in a cross‐sectional analysis, using data from the 2020 Behavioral Risk Factor Surveillance System survey (BRFSS). BRFSS is a cross‐sectional telephone‐based survey of U.S. residents 18 years of age and older, which collects information about health‐related factors and use of preventive services. Results The prevalence rates of colorectal cancer screening were respectively 7.9 (95% CI: 4.0–11.9) and 15.0 (95% CI: 4.0–11.9) %‐points higher among gynecologic and other cancer survivors compared to that of 65.2% among those without any history of cancer. However, no differences were observed in breast cancer screening between gynecologic cancer survivors (78.5%) and respondents without any history of cancer (78.7%). Coverage of influenza vaccination among gynecologic cancer survivors were 4.0 (95% CI: 0.3–7.6) %‐points higher than that of the no cancer group, but 11.6 (95% CI: 7.6–15.6) %‐points lower than that of the other cancer group. Pneumonia vaccination rate among gynecologic cancer survivors, however, was not statistically different than that of other cancer survivors and respondents with no history of cancer. When examining modifiable risk behaviors, the prevalence of smoking among gynecologic cancer survivors was 12.8 (95% CI: 9.5–16.0) and 14.2 (95% CI: 10.8–17.7) %‐points higher than smoking prevalence among other cancer survivors and respondents without any history of cancer. The rate differentials were even higher in rural areas, 17.4 (95% CI: 7.2–27.6) and 18.4 (95% CI: 7.4–29.4) %‐points respectively. There were no differences in the prevalence of heavy drinking across the groups. Lastly, gynecologic and other cancer survivors were less likely to be physically active (Δ = −12.3, 95% CI: −15.8 to −8.8 and Δ = −6.9, 95% CI: −8.5 to −5.3, respectively) than those without any history of cancer. Conclusion Smoking prevalence among gynecologic cancer survivors is alarmingly high. Intervention studies are needed to identify effective ways to assist gynecologic cancer survivors to quit smoking and refrain from hazardous alcohol consumption. In addition, women with gynecologic malignancies should made aware of the importance of physical activity.


| INTRODUCTION
As a result of improvement in therapies and supportive care, the prevalence of gynecologic (uterine, cervical, and ovarian) cancer survivors continue to grow over the past few decades. Uterine cancer is the fourth most common cancer among women in the U.S., with an estimated 66,570 new cases in 2021. 1 There were an estimated 14,480 new cases of cervical cancer and an estimated 21,410 new cases of ovarian cancer in the U.S. in 2021. 1 As of January 1, 2019, it is estimated that there were 807,860 uterine cancer survivors, 283,120 cervical cancer survivors, and 249,230 ovarian cancer survivors in the U.S. 2 Maintaining a healthy lifestyle is an important factor in promoting positive outcomes for gynecologic cancer survivors. [3][4][5] Health behaviors, such as engaging in physical activity, not consuming large amounts of alcohol, abstaining from cigarette smoking, receiving routine immunizations, and undergoing recommended cancer screening tests, is known to improve health and quality of life. 6,7 For example, physical activity, healthy nutrition options, maintaining a healthy weight, and avoiding smoking is associated with reduced risk of cancer recurrence, secondary cancers, and chronic illness (e.g., obesity, cardiovascular disease, diabetes). 8,9 Gynecologic cancer survivors who have completed primary therapy for the disease often have complex health care needs that include close follow-up to detect recurrence and second primary malignancies, and the management of cancer-related morbidities. 10 Appropriate care during the post-initial treatment survivorship phase includes preventive care visits and screening for cancer recurrence. [11][12][13] Survivors of gynecologic cancer who live in rural areas often experience several barriers to receiving appropriate follow-up care. Residents of rural areas are more likely to be low-income and to lack health insurance than residents of urban areas. 14 Rural residence has been related to health disparities and greater mortality risk in gynecologic cancer patients. 15 However, there is a need for more studies of gynecologic cancer survivors in rural areas.
We examined preventive behaviors and modifiable behavioral risk factors among gynecologic cancer survivors, using data from the cancer survivorship module of the 2020 Behavioral Risk Factor Surveillance System (BRFSS). This manuscript considers survivors of gynecologic cancer from the time of diagnosis and initiation of treatment. We compared behavioral risk factors and preventive measures among gynecologic cancer survivors, female survivors of other types of cancer, and women without a history of cancer. Based on the existing evidence supporting greater likelihood of screening service utilization among cancer survivors, 16,17 our hypothesis was that the prevalence of preventive measures among gynecologic cancer survivors would be that of survivors of other types of cancer and higher than the prevalence among women without a history of cancer. In addition, in line with the existing evidence on rural-urban differences in health behaviors and outcomes, 18,19 we hypothesized that the prevalence of preventive measures and behavioral risk factors would vary by urban-rural residence. We assessed four preventive health measurescolorectal cancer screening (as per the U.S. Preventive Services Task Force recommendation) among women aged 45-75 years, 20 breast cancer screening (mammogram in past 2 years) among women aged 50-74 years, 20 vaccination for influenza, and vaccination for pneumonia among women age ≥65 years; and three behavioral risk factors-smoking, heavy drinking, and no leisure time physical activity. A respondent was determined to be a current smoker if she smoked at least 100 cigarettes in her entire life and smoked every day or some days at the time of the survey. Heavy drinking was defined as having more than 14 drinks per week for men and more than 7 drinks per week for women. Physical activity was defined as participation in physical activities or exercises other than regular job during the past 30 days preceding the survey. Of note, we did not have information on physical activities in the workplace; neither do we know about the intensity of the activity.

BRFSS is a cross
We divided the sample into three mutually exclusive groups-no history of cancer (other than non-melanoma skin cancer), gynecologic cancer, and other cancer (other than non-melanoma skin cancer). Of note, respondents who reported to have more than one type of cancer (including non-melanoma skin cancer) were grouped based on their most recent diagnosis of cancer. As such, there could be overlaps for some observations.
We examined how the prevalence of preventive health behaviors and behavioral risk factors vary across these groups. We fitted separate univariate and multivariable logistic regression models for each of the seven preventive measures and behavioral risk factors to obtain the unadjusted and adjusted odds in favor of complying with the preventive measure or having the risk factor for being a gynecologic cancer survivor or other cancer survivor compared to someone with no history of cancer. Our outcome variables are binary variables indicative of respective preventive measures or risk factors. Our key explanatory variables are dichotomous variables indicating gynecologic or other cancer survivorship; no history of cancer serves as the reference category. In the multivariable model we controlled for age (18- married, and divorced/ widowed/separated); employment status (employed [base category]-"employed for wages or self-employed", not employed-"out of work for <1 year or ≥1 year, homemaker, student, and unable to work", and retired); and fixed effects. Of note, in the multivariable model, we did not evaluate the sociodemographic predictors of preventive behaviors or risk factors. Rather we intended to examine how gynecologic cancer status was associated with certain behaviors after accounting for the socio-demographic correlates. Observations, for which sociodemographic data were missing, were excluded from the multivariable specification.
To check the robustness of our analysis, we further estimated the models for urban and rural sub-groups. Urban consists of counties that are classified as large central metro, large fringe metro, medium metro, small metro, and micropolitan statistical areas. 21 Rural consists of noncore counties outside of the micropolitan statistical areas. Additionally, to explore the relationships across different types of gynecologic cancer, we estimated the models (for full sample only) with the key explanatory variable having the following mutually exclusive categories: no history of cancer (base category), other cancer, cervical cancer, endometrial cancer, and ovarian cancer. The statistical analyses were conducted using Stata 17.0 and all estimates account for complex survey weights. Level of significance was set at 5% level.
This study utilized BRFSS public use data that include no personal identifiers and was therefore except from review by the Augusta University Institutional Review Board.

| RESULTS
Of the 110,820 women in our sample, 1824 were gynecologic cancer survivors, and 9843 were survivors of other types of cancer. Around 90% of the women in the sample had no history of cancer. Among respondents having gynecologic cancer, 904 had cervical-, 556-had endometrial-, and 364 had ovarian-cancer. The weighted prevalence of gynecologic cancer was higher in rural areas than in urban areas (Δ = 0.75%-points, 95% CI: 0.32-1.19). Table 1 shows the demographic and socioeconomic characteristics of the study participants by mutually exclusive categories and urban and rural residence. While the largest weighted share (37.35%) of gynecologic cancer survivors in the urban areas were elderly (age ≥ 65 years), the largest weighted share (38.37%) in rural areas were younger adults (age 18-49).
The weighted prevalence of colorectal cancer screening was significantly higher among gynecologic and other cancer survivors compared to those without any history T A B L E 1 Demographic and socioeconomic characteristics of study participants by urban-rural residence and type of cancer.

No history of cancer N [%]
Other  (Table 2). For breast cancer screening, no differences were observed between gynecologic cancer survivors and respondents without any history of cancer. Pneumonia vaccination was higher among both gynecologic and other cancer survivors compared to those without a cancer history. However, no statistically significant differences were observed across these two groups. On the other hand, the weighted prevalence of smoking was significantly higher among gynecologic cancer survivors, particularly in rural areas. There were no significant differences in the weighted prevalence of heavy drinking across the groups. A total of 25.97% of gynecologic cancer survivors were current cigarette smokers, whereas the smoking prevalence among other cancer survivors and respondents with no history of cancer was 11.76% and 13.22%, respectively. Among gynecologic cancer survivors, 5.08% were heavy alcohol drinkers, which was 5.54% and 6.30%, respectively, among survivors of other cancer and respondents without any history of cancer. Lastly, both gynecologic and other cancer survivors were less likely to be physically active than those without any history of cancer, particularly in urban areas. Among gynecologic cancer survivors, colorectal cancer screening was significantly higher among endometrial cancer survivors compared to survivors of cervical cancer. While there was no significant difference in flu vaccination rates between survivors of cervical-and ovarian-cancer, the rate was significantly higher among survivors of endometrial cancer. Tobacco-use, on the other hand, was significantly higher among cervical cancer survivors compared to other endometrial-and ovarian-cancer survivors.
Though no differences in rates of heavy drinking were found between respondents without any history of cancer and gynecologic cancer survivors as a whole, the rate was significantly lower among survivors of endometrial cancer. No statistically significant differences were observed for mammogram, pneumonia vaccination, and leisure time physical activity across the three types of gynecologic cancer ( Table 2).
The unadjusted and adjusted odds ratios in favor of preventive health measures and behavioral risk factors are presented in Table 3. The odds of being screened for colorectal cancer for gynecologic cancer survivors were 1.5 (95% CI: 1.2-1.8) times that of those without any history of cancer. The odds of receiving influenza (OR = 1.2, 95% CI: 1.0-1.4) and pneumonia (OR = 1.9, 95% CI: 1.7-2.3) vaccination were also higher among gynecologic cancer survivors compared to women without cancer. The odds of receiving a mammogram for gynecologic cancer survivors, however, were not different (OR = 1.0, 95% CI: 0.8-1.2) than those with no history of cancer. Compared to individuals without any history of cancer, the odds of smoking were significantly higher among gynecologic cancer survivors (OR = 2.3, 95% CI: 1.9-2.7) and significantly lower among survivors of other types of cancer (OR = 0.88, 95% CI: 0.78-0.98). In rural areas, the odds of current smoking for gynecologic cancer survivors were 2.5 times (95% CI: 1.59-3.92) that of those with no history of cancer. The odds of heavy drinking, on the other hand, were not statistically different from zero (OR = 0.8, 95% CI: 0.6-1.1). Lastly, the odds of physical exercise were significantly lower among both gynecologic (OR = 0.56, 95% CI: 0.48-0.65) and other cancer survivors (OR = 0.71, 95% CI: 0.66-0.76) compared to their peers without any history of cancer, particularly in urban areas. Compared

Note:
Estimates were obtained using complex survey weights. 95% confidence intervals are in parenthesis. Regressions were separately estimated for each preventive behavior and behavioral risk factor for the full sample (i.e., All) and urban and rural sub-samples. Each row presents results of a separate regression (unadjusted and adjusted). The adjusted odds were obtained by controlling for age, race and ethnicity, educational attainment, household income, marital status, employment status, and fixed effects.

T A B L E 3 (Continued)
Lastly, the lower odds of physical activity among gynecologic cancer survivors were evident across all three types of gynecologic cancer.

| DISCUSSION
The results of this population-based study indicate that adherence with colorectal cancer screening and pneumonia vaccination recommendations among gynecologic cancer survivors compares favorably with that of persons with no history of cancer. Adherence to breast cancer screening and influenza vaccination, on the other hand, were not different across the two groups. In addition, adherence with colorectal cancer screening test and influenza immunization recommendations among gynecologic cancer survivors is generally lower than that of survivors of other types of cancer. In this study, 25.97% of gynecologic cancer survivors are current cigarette smokers and 5.08% are heavy alcohol drinkers, thus at higher risk of cancer recurrence, secondary malignancies, and other chronic diseases. Of note, though alcohol consumption does not directly impact the risk of ovarian and endometrial cancer, 22 heavy alcohol use may increase the risk of cervical cancer recurrence. 23 Gynecologic cancers include both smoking-related cancer (cervical cancer) and non-smoking-related cancers (ovarian cancer, uterine cancer). 24 Oncologists and primary care providers have important roles in encouraging cancer survivors T A B L E 4 Unadjusted and adjusted odds ratios in favor of preventive measures and behavioral risk factors for gynecologic cancer types. Note: Estimates were obtained using complex survey weights. 95% confidence intervals are in parenthesis. Regressions were separately estimated for each preventive behavior and behavioral risk factor for the full sample (i.e., All). Each row presents results of a separate regression (unadjusted and adjusted). Results for the "other cancer" group were not reported as they were identical to those reported in Table 3. The adjusted odds were obtained by controlling for age, race and ethnicity, educational attainment, household income, marital status, employment status, and fixed effects.
to reduce their risk of serious illness or death through healthy lifestyle changes. The prevalence of cigarette smoking was especially high (35.49%) among gynecologic cancer survivors who live in rural areas. Gynecologic cancer survivors in the current study were less likely to be physically active than persons without a history of cancer and having other types of cancer. Of note, as per data availability, the measure of physical activity analyzed in this paper, included leisure time physical activities only and did not take into account of any physical activities in the workplace. Further, the BRFSS data did not provide information on whether the recommended guidelines for weekly physical exercise were met. As such, this measure may not reflect the true rate of recommended physical exercise. However, given the data constraint it can be considered as a close proxy to the actual rate as used in other studies. 25 Our results are consistent with those reported by Hawkins et al. 26 in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening trial who found that cancer survivors were less likely to engage in physical activity than those who were cancer free. As such, it is important to make women with gynecologic malignancies aware about the importance of physical activity. 4 Adopting and maintaining healthy behaviors, however, is subject to some structural barriers at individual, social, and environmental levels. 27 For example, lack of equipment or facility as well as time constraints are some barriers to physical exercise for cancer survivors. 28 Additionally, perceived structural barriers among healthcare professionals impedes promotion of physical activity in cancer patients. 29,30 Psychosocial factors, such as lack of knowledge and advice, external motivation, and monitoring are other barriers of healthy eating and physical exercise among cancer patients. 31 Addressing these structural barriers can significantly improve healthy behaviors among survivors of gynecologic cancers.
With respect to limitations, the results may not be generalizable to all gynecologic cancer survivors in the United States because of the low response rate in the BRFSS. In addition, misclassification bias may have occurred due to the use of self-reported information. Misclassification refers to the classification of an individual, a value or an attribute into a category other than that to which it should be assigned. We did not perform single or multiple imputation to deal with missing data in the analysis. However, because the amount of missing data in BRFSS is small, this is unlikely to have resulted in any important bias.
We were limited by lack of data on temporality, including pre-diagnosis behaviors. Behaviors such as smoking, alcohol consumption, and physical activity may change following a cancer diagnosis. Due to the cross-sectional nature of the data, we were unable to assess any behavioral changes during pre-and postdiagnosis phases. Furthermore, we did not have information on cancer characteristics (e.g., stage at diagnosis), phase of care, or treatment patterns, which could influence the preventive and risky health behaviors. Future research using appropriate longitudinal data may address these issues.
In conclusion, the results of this study have substantial implication for healthcare providers, members of the healthcare team, and researchers. A recent study reported a higher risk of cardiovascular diseases and obesity among gynecologic cancer survivors. 32 As such, promotion of preventive health behaviors can play a critical role in improving health and wellbeing of women with gynecologic malignancies. Though adherence to certain cancer screening (i.e., colorectal cancer) and immunization (i.e., pneumonia vaccination) recommendations among gynecologic cancer survivors compared favorably with that of those with no history of cancer, the prevalence of smoking is markedly higher among gynecologic cancer survivors. Intervention studies are needed to identify effective ways to assist gynecologic cancer survivors to quit smoking and refrain from alcohol consumption. In addition, increased efforts should be made to make women with gynecologic malignancies aware of the health risk behaviors, emphasizing the adverse effects that behaviors like smoking could have on cancer treatments. In addition, the importance of physical activity, and the impact of preventive practices should be stressed to these patients by all members of the healthcare team.

DATA AVAILABILITY STATEMENT
This study was based upon anonymous, public use data provided by the Centers for Disease Control and Prevention.