The impact of socioeconomic status on changes in cancer prevention behavior during the COVID-19 pandemic

BACKGROUND

The impacts of socioeconomic status (SES) on COVID-19-related changes in cancer prevention behavior have not been thoroughly investigated. We conducted a cohort study to examine the effects of SES on changes in cancer prevention behaviors during the COVID-19 pandemic.

METHODS

We invited adult participants from previous studies conducted at Ohio State University to participate in a study assessing the impact of COVID-19 on various behaviors. Post-COVID-19 cancer prevention behaviors, including physical activity, daily intake of fruits and vegetables, alcohol and tobacco consumption, and qualitative changes in post-COVID-19 behaviors relative to pre-COVID levels, were used to construct a prevention behavior change index that captures the adherence status and COVID-related changes in each behavior, with higher index scores indicating desirable changes in prevention behaviors. Participants were classified into low, middle, or high SES based on household income, education, and employment status. Adjusted regression models were used to examine the effects of SES on changes in cancer prevention behaviors during the COVID-19 pandemic.

RESULTS

The study included 6,136 eligible participants. The average age was 57 years, 67% were women, 89% were non-Hispanic Whites, and 33% lived in non-metro counties. Relative to participants with high SES, those with low SES had a 24% [adjusted relative ratio, aRR = 0.76 (95%CI 0.72-0.80)], 11% [aRR = 0.89 (95%CI 0.86-0.92)], and 5% [aRR = 0.95 (95%CI 0.93-0.96)], lower desirable changes in prevention behaviors for physical activity, fruit and vegetable intake, and tobacco use, respectively. Low SES had a higher desirable change in alcohol consumption prevention behaviors, 16% [aRR = 1.16 (95%CI 1.13-1.19)] relative to high SES. The adjusted odds of an overall poor change in prevention behavior were adjusted odds ratio (aOR) 1.55 (95%CI 1.27 to 1.89) and aOR 1.40 (95%CI 1.19 to 1.66), respectively, higher for those with low and middle SES relative to those with high SES.

CONCLUSION

The adverse impacts of COVID-19 on cancer prevention behaviors were seen most in those with lower SES. Public health efforts are currently needed to promote cancer prevention behaviors, especially amongst lower SES adults.

Impact of the COVID-19 Pandemic on Cancer Screening Delays

PURPOSE

To examine delays in cancer screenings during the COVID-19 pandemic.

METHODS

Participants from previous studies (N = 32,989) with permissions to be recontacted were invited to complete a survey between June and November 2020. Participants (n = 7,115) who met the age range for cancer screenings were included. Participants were asked if they planned to have and then if they postponed a scheduled mammogram, Pap test, stool blood test, colonoscopy, or human papillomavirus (HPV) test. Logistic regression was used to determine the factors associated with cancer screening delays for each planned test.

RESULTS

The average age was 57.3 years, 75% were female, 89% were non-Hispanic White, 14% had public insurance, and 34% lived in rural counties. Those who planned cancer screenings (n = 4,266, 60%) were younger, more likely to be female, with higher education, had private insurance, and lived in rural counties. Specifically, 24% delayed a mammogram (n = 732/2,986), 27% delayed a Pap test (n = 448/1,651), 27% delayed an HPV test (n = 59/220), 11% delayed a stool blood test (n = 44/388), and 36% delayed a colonoscopy (n = 304/840). Age, race/ethnicity, education, and health insurance were associated with delays in cancer screenings (all P < .05). Compared with non-Hispanic White women, non-Hispanic Black women had lower odds of delaying a mammogram (odds ratio [OR], 0.60; 95% CI, 0.39 to 0.94), Hispanic women had higher odds of delaying Pap test (OR, 2.46; 95% CI, 1.34 to 4.55), and women with other race/ethnicity had higher odds of delaying both Pap test (OR, 2.38; 95% CI, 1.41 to 4.02) and HPV test (OR, 5.37; 95% CI, 1.44 to 19.97).

CONCLUSION

Our findings highlighted the urgency for health care providers to address the significant delays in cancer screenings in those most likely to delay. Strategies and resources are needed to help those with barriers to receiving guideline-appropriate cancer screening.

Socioeconomic status for predicting COVID-19-related changes in cancer prevention behaviors

10547

Background: Disruptions of daily life activities during the CVOID-19 pandemic have adversely affected cancer-prevention behaviors. Socioeconomic status (SES) impacts on changes in cancer prevention behaviors have not been fully investigated. To tackle this gap, we examined the effects of SES on COVID-19 related changes in cancer prevention behaviors. Methods: We invited participants from previous studies conducted at the Ohio State University Comprehensive Cancer Center who agreed to be re-contacted to participate in a survey assessing the impact of COVID-19 on various behaviors between June and November 2020. Participants reported current cancer prevention behaviors, including physical activity, daily fruit and vegetable intake, alcohol consumption, and tobacco use. In addition, participants reported qualitative changes in current behaviors relative to pre-COVID levels. We combined current behaviors with COVID-related changes to construct a 24-point cancer prevention score. Participants were classified into low, middle, or high SES based on household income, education, and employment status. Adjusted multinomial logistic regression was used to examine the association between SES and COVID-19 related changes in cancer prevention behaviors. Results: The study sample included 6136 eligible participants. The average age was 57 years, 67% were female, 89% were non-Hispanic White, and 33% lived in non-metro counties. The proportion of participants in the lowest cancer prevention behavior quartile decreased significantly with higher SES [low SES vs. high SES; 32% vs. 28%; P-value <.001]. Relative to pre-COVID-19 levels, higher SES was significantly associated with increases in post-COVID-19 prevalence of more physical activity [low SES vs. high SES; 12% vs. 28%; P-value <.001], higher fruit and vegetable intake [low SES vs. high SES; 12% vs. 14%; P-value <.001], and more alcohol consumption [low SES vs. high SES; 15% vs. 22%; P-value <.001]. Higher SES was associated with lower tobacco use prevalence [low SES vs. high SES; 5% vs. 2%; P-value <.001]. Relative to the highest prevention score quartile, the adjusted odds of scoring in the lowest prevention score quartile were: adjusted odds ratio (aOR) 1.55 (95% CI: 1.27 - 1.89) and aOR 1.40 (95% CI: 1.19 - 1.66), respectively higher for low and middle SES. Low SES was significantly associated with higher odds of less frequent physical activity (aOR = 1.87; 95% CI: 1.49 - 2.35) and less fruit and vegetable consumption (aOR = 1.56; 95% CI: 1.15 - 2.12). Middle SES relative to high SES was associated with lower odds of more alcohol consumption (aOR = 0.64; 95% CI: 0.49 - 0.85) and higher odds of binge drinking (aOR = 1.32; 95% CI: 1.09 - 1.59). Conclusions: The adverse impacts of COVID-19 on cancer prevention behaviors were seen most in those with lower SES. Public health efforts are currently needed to promote cancer prevention behaviors, especially amongst lower SES adults.

Associations between cancer history, social distancing behaviors, and loneliness in adults during the COVID-19 pandemic

e24124

Background: Due to COVID-19, social distancing initiatives have been enacted, highlighting differences of social distancing practices and the resulting loneliness in various populations, especially in those with a history of cancer (survivors and those in active treatment).The objective of this study was to examine whether social distancing practices and loneliness differ between individuals with and without a history of cancer during the COVID-19 pandemic. Methods: This study is part of the “Impact of COVID-19 on Behaviors across the Cancer Control Continuum in Ohio” project conducted from June to November 2020. Participants from previous studies (N = 32,989) who had given permission to be re-contacted were invited to complete a survey online, by phone, or by mail. For the current analysis, participants who had complete data on history of cancer diagnosis, COVID social distancing behaviors (staying at home, not having visitors, staying 6 feet apart, wearing a masking indoors, and wearing a mask outdoors), attendance at social gatherings, contact with people outside of their household, and feelings of loneliness were included. Linear and logistic regression models were used to determine the associations between demographics, cancer history, social distancing, and loneliness. Results: Among the eligible participants (n = 5729), 54.9% had a cancer history (n = 3147), while 45.1% did not have a cancer history (n = 2582). The average age was 56.7 years, 35.6% were male, 89.4% were White, 74.7% were partnered, and 74.5% lived in metro counties. Out of all individuals, 69.3% participated in ≥4 of the 5 COVID social distancing behaviors, 31.5% did not attend any social gatherings, and 40.1% reported feeling lonely. Compared to individuals without a cancer history, individuals with a cancer history were more likely to contact no one outside of the household (49.0% vs. 41.9%, p < 0.01) and less likely to report feeling lonely (35.8% vs. 45.3%, p < 0.01). Among individuals with a cancer history, compared to their counterparts, those who were older, Black, Asian, and living in metro counties were adherent to more social distancing behaviors (mean differences = 0.02, 0.62, 0.71, 0.13, respectively, p all < 0.05) and attended fewer social gathering (mean differences = -0.01, -0.29, -0.49, -0.12, respectively, p all < 0.05). Higher adherence to social distancing behaviors was associated with higher odds of loneliness among individuals with (OR = 1.27, 95% CI: 1.72-1.38) and without a cancer history (OR = 1.15, 95% CI: 1.06-1.25). Conclusions: Social distancing practices and loneliness during the COVID-19 pandemic varied between individuals with and without a cancer history. These findings can inform efforts to support and address all aspects of health among individuals susceptible to loneliness during the pandemic, especially those with a history of cancer.

Feasibility of implementing a text-based symptom-monitoring program of endometrial, ovarian, and breast cancer patients during treatment

Purpose

To evaluate the feasibility of implementing systematic patient symptom monitoring during treatment using a smartphone.

Methods

Endometrial [n = 50], ovarian [n = 70] and breast [n = 193] cancer patients participated in text-based symptom reporting for up to 12 months. In order to promote equity, patients without a smartphone were provided with a device, with the phone charges paid by program funds. Each month, patients completed the Patient Health Questionnaire (PHQ-9), and 4 single items assessing fatigue, sleep quality, pain, and global quality of life during the past 7 days rated on a 0 (low) –10 (high) scale. Patients’ responses were captured using REDCap, with oncologists receiving monthly feedback. Lay navigators provided assistance to patients with non-medical needs.

Results

Patients utilizing this voluntary program had an overall mean age of 60.5 (range 26–87), and 85% were non-Hispanic white. iPhones were provided to 42 patients, and navigation services were used by 69 patients. Average adherence with monthly surveys ranged between 75–77%, with breast patients having lower adherence after 5 months. The most commonly reported symptoms across cancer types were moderate levels (scores of 4–7) of fatigue and sleep disturbance. At 6 months, 71–77% of all patients believed the surveys were useful to them and their health care team.

Conclusions

We established the feasibility of initiating and managing patients in a monthly text-based symptom-monitoring program. The provision of smartphones and patient navigation were unique and vital components of this program.

Provision of smartphones in a symptom monitoring program of gynecologic and breast cancer patients during active therapy

255

Background: With advances in technology, smartphones are being used for multiple research and clinical care functions. However, not all patients have these devices, leading to disparities in participation. We report on a quality improvement program that provided smartphones to patients without these devices. Methods: Gynecologic (n = 120) and breast (n = 193) cancer patients under active treatment were enrolled in a 12-month text-based symptom monitoring program to facilitate communication and optimize patient management. Patients without a smartphone were provided with an iPhone through a partnership with a U.S. wireless company. The company provided smartphone devices at zero cost, and program funds paid for 12 months of phone service. Program staff helped patients set up the iPhones, and provided basic education and ongoing phone support. After 12 months, patients were able to keep their iPhones, but had to secure their own phone plan for calling and texting functions. Results: iPhones were provided to 42 (13.4%) patients across all cancer types. Patients who received iPhones, compared with those who had a smartphone, had incomes below $50,000/year (p = 0.03) and an educational level of < high school (p < 0.0001). Program staff had few difficulties training patients to operate the phones or in patients’ adherence to symptom monitoring after receiving the iPhones. Phone service charges averaged $40 per month or $500 per person for 12 months. Greater than 90% of patients believed the phones enabled them to better communicate with their health care team and family/support networks, and 95% believed the phones had a positive impact on their life. However, only 70% planned on getting a phone service plan at the end of the 12 months, due to cost or believing a smartphone was not a necessity. Conclusions: Providing smartphones to patients enabled them to better communicate with their health care team and families, and participate in remote symptom monitoring during active treatment. Programs such as these are needed to reduce disparities in patient care, and support quality improvement efforts using electronic devices.

Feasibility of text-based symptom monitoring of ovarian and endometrial patients during treatment

e18299

Background: Unreported symptoms during cancer treatment can lead to poorer patient care and quality of life. Newer technology enables effective means to track patients’ health in real time. We evaluated the feasibility of implementing systematic patient symptom monitoring during the first 12 months after diagnosis. Methods: Newly diagnosed endometrial and ovarian cancer patients were enrolled post-surgically to respond to monthly text message symptom surveys. Patients’ fatigue, sleep quality, pain, and quality of life during the past 7 days were rated on a 0 (worst) -10 (best) scale, and depressive symptoms were assessed using the Patient Health Questionnaire (PHQ-9). Patients’ responses were captured in REDCap and monitored by program staff, with patients’ oncologists receiving monthly feedback. Patient navigators were also engaged for patients needing assistance during treatment. We provide the results of the first 6 months of this program. Results: 134 patients were approached, and 120 patients (ovarian [n = 70] and endometrial [n = 50]) were enrolled among 5 physicians. The mean participant age was 63 (range: 35-87), 85% were non-Hispanic White, and 66% had education beyond high school. The most commonly reported monthly symptoms for both cancer types were moderate levels (scores of ≥ 4-7) of fatigue and sleep disturbance. 35 patients with PHQ-9 scores ≥ 10 and/or with suicidal ideation were reported to their oncologists for appropriate follow-up. At the 6 month survey, patients were asked to evaluate the text messaging program: 97% found the symptom surveys easy to complete on their smart phone or computer; 77% believed reporting their symptoms monthly was useful all or most of the time; 78% liked being monitored for symptoms all or most of the time; and 89% liked being asked if they needed any assistance prior to their next clinic visit. Patient navigators were used by 13 patients, and 17 patients dropped from the program over the 6 months due to death (n = 9) or lack of need/interest (n = 8). Average monthly compliance was 81%. Conclusions: We established the feasibility of enrolling patients in a monthly text-based monitoring program to facilitate symptom management during treatment. Patient follow-up is continuing.

Correlates of Rural, Appalachian, and Community Identity in the CITIES Cohort

PURPOSE

To determine correlates of rural, Appalachian, and community identity among a cohort of participants in the Community Initiative Towards Improving Equity and Health Status (CITIES) project.

METHODS

Mixed linear and logistic regression effects models were utilized to determine correlates of 3 outcomes: 1) community identity, 2) rural identity, and 3) Appalachian identity among participants in the Ohio CITIES project.

FINDINGS

Distinct demographic characteristics were found to be associated with each of the outcomes. For community identity, while no differences were found for rural or urban participants, those who were single or never married (P < .0001) as well as those who graduated from college (P = .0005) reported significantly lower community identity scores than married individuals with less than a college education. Those who resided in an Appalachian county reported higher community identity scores (P = .0009) than non-Appalachian residents. For rural identity, those who did not identify as Christian (P = .018) as well as those who identified as Democrat (P = .027) reported significantly lower rural identity scores than others. Lastly, for Appalachian identity, county-level percentage of families in poverty (P = .06), as well as gender (P = .05), were associated with self-reported Appalachian identity, but these effects were only marginally significant.

CONCLUSIONS

Although community, rural, and Appalachian identity may be viewed as similar due to their measure of attachment to a place, results from this study suggest that there are distinct individual and area-level correlates associated with community, rural, and Appalachian identity.

Financial burden among older, long-term cancer survivors: Results from the LILAC study

Background

Increasing attention is being paid to financial burdens of cancer survivorship, but little is known about the prevalence and predictors of these burdens in older, long‐term survivors.

Methods

We used data from 6012 participants diagnosed with cancer since enrolling in the Women's Health Initiative, and who participated in the Life and Longevity After Cancer (LILAC) ancillary study to estimate prevalence and identify predictors of financial burden. We used logistic regression to identify sociodemographic, socioeconomic, health‐ and cancer‐related factors associated with financial burden and backward selection to build a final multivariable model.

Results

Average age at LILAC participation was 79 and 9.2 years had elapsed since cancer diagnosis. Overall, 6% experienced some form of financial burden, including having an insurance company refuse a claim (2.6%), being denied loans or insurance due to cancer history (2.2%), or experiencing significant indebtedness (1.8%, including facing large debts or bills or declaring bankruptcy). Eight predictors remained associated (P < 0.05) with financial burden in the fully‐adjusted model: younger age, shorter time since diagnosis, African‐American race, household income <$20 000/year, modified Charlson comorbidity score ≥2, receipt of chemotherapy, regional stage at diagnosis, and no private health insurance. Education, cancer site, social support, receipt of radiation, and receipt of hormone therapy were not associated with financial burden. Predictors differed between types of financial burden experienced and age at diagnosis (<65 vs 65+).

Conclusion

Cancer‐related financial burden was rare in this population of older, female long‐term cancer survivors. The identification of several socioeconomic, health‐related and demographic predictors of financial burden may suggest targets of intervention to reduce financial burdens.

Precis

Financial burden was uncommon in older, female, long‐term survivors. Predictors of financial burden included age, race, income, comorbidities, time since diagnosis, stage, insurance, and receipt of chemotherapy.

Telomere Length and Neighborhood Circumstances: Evaluating Biological Response to Unfavorable Exposures

Background: Multilevel frameworks suggest neighborhood circumstances influence biology; however, this relationship is not well studied. Telomere length (TL) shortening has been associated with individual-level and neighborhood-level exposures and disease and may provide insights into underlying biologic mechanisms linking neighborhood with biology. To support neighborhood-biology investigations, we sought to determine the independent effect of neighborhood exposures on TL using standard multilevel linear regression models and quantile regression, a nonlinear, social science method applicable for testing the biologic hypothesis that extremes of the TL distribution are related to poor outcomes.Methods: In a multicenter, cross-sectional study, blood TL was measured in 1,488 individuals from 127 census tracts in three U.S. regions using terminal restriction fragment assays. Multilevel linear and quantile regression models were adjusted for individual-level race, education, perceived stress, and depression. Neighborhood exposures included population density, urban/residential crowding, residential stability/mobility, and socioeconomic status.Results: TL was not associated with any neighborhood variable using linear models, but quantile regression revealed inverse associations between population density and urban crowding at the lower tails of the TL distribution [5th (population density P = 0.03; urban crowding P = 0.002), 50th (both P < 0.001), 75th percentiles (both P < 0.001)]. TL was related to residential stability at the upper tail (95th percentile P = 0.006).Conclusions: Findings support the use of nonlinear statistical methods in TL research and suggest that neighborhood exposures can result in biological effects.Impact: TL may serve as an underlying example of a biologic mechanism that can link neighborhood with biology, thus supporting multilevel investigations in future studies. Cancer Epidemiol Biomarkers Prev; 26(4); 553-60. ©2017 AACRSee all the articles in this CEBP Focus section, "Geospatial Approaches to Cancer Control and Population Sciences."

Race, Ethnicity, Psychosocial Factors, and Telomere Length in a Multicenter Setting

Background

Leukocyte telomere length(LTL) has been associated with age, self-reported race/ethnicity, gender, education, and psychosocial factors, including perceived stress, and depression. However, inconsistencies in associations of LTL with disease and other phenotypes exist across studies. Population characteristics, including race/ethnicity, laboratory methods, and statistical approaches in LTL have not been comprehensively studied and could explain inconsistent LTL associations.

Methods

LTL was measured using Southern Blot in 1510 participants from a multi-ethnic, multi-center study combining data from 3 centers with different population characteristics and laboratory processing methods. Main associations between LTL and psychosocial factors and LTL and race/ethnicity were evaluated and then compared across generalized estimating equations(GEE) and linear regression models. Statistical models were adjusted for factors typically associated with LTL(age, gender, cancer status) and also accounted for factors related to center differences, including laboratory methods(i.e., DNA extraction). Associations between LTL and psychosocial factors were also evaluated within race/ethnicity subgroups (Non-hispanic Whites, African Americans, and Hispanics).

Results

Beyond adjustment for age, gender, and cancer status, additional adjustments for DNA extraction and clustering by center were needed given their effects on LTL measurements. In adjusted GEE models, longer LTL was associated with African American race (Beta(β)(standard error(SE)) = 0.09(0.04), p-value = 0.04) and Hispanic ethnicity (β(SE) = 0.06(0.01), p-value = 0.02) compared to Non-Hispanic Whites. Longer LTL was also associated with less than a high school education compared to having greater than a high school education (β(SE) = 0.06(0.02), p-value = 0.04). LTL was inversely related to perceived stress (β(SE) = -0.02(0.003), p<0.001). In subgroup analyses, there was a negative association with LTL in African Americans with a high school education versus those with greater than a high school education(β(SE) = -0.11(0.03), p-value<0.001).

Conclusions

Laboratory methods and population characteristics that differ by center can influence telomere length associations in multicenter settings, but these effects could be addressed through statistical adjustments. Proper evaluation of potential sources of bias can allow for combined multicenter analyses and may resolve some inconsistencies in reporting of LTL associations. Further, biologic effects on LTL may differ under certain psychosocial and racial/ethnic circumstances and could impact future health disparity studies.

Effect of a low-fat or low-carbohydrate weight-loss diet on markers of cardiovascular risk among premenopausal women: a randomized trial

BACKGROUND

Low-fat and low-carbohydrate weight-loss diets can have a beneficial effect on longitudinal measures of blood pressure and blood lipids. We aimed to assess longitudinal changes in blood pressure and blood lipids in a population of premenopausal women. We hypothesized that results may differ by level of adherence to the respective diet protocol and baseline presence of hypertension or hyperlipidemia.

METHODS

Overweight or obese premenopausal women were randomized to a low-fat (n=41) or low-carbohydrate (n=38) diet. As part of the 52-week Lifestyle Eating and Fitness (LEAF) intervention trial, we fit linear mixed models to determine whether a change in outcome differed by treatment arm.

RESULTS

Within-group trends in blood pressure and blood lipids did not differ (p>0.30). Across study arms, there was a significant decrease in systolic blood pressure (SBP, 3 mm Hg, p=0.01) over time, but diastolic blood pressure (DBP) did not change significantly over the course of the study. Blood lipids (total cholesterol [TC], low-density lipoproteins [LDL], and high-density lipoproteins [HDL]) all exhibited nonlinear trends over time (p<0.01); each decreased initially but returned to levels comparable to baseline by study conclusion (p>0.20). We observed a decline in SBP among women who were hypertensive at baseline (p<0.01), but hypercholesterolemia at baseline did not affect trends in blood lipids (p>0.40).

CONCLUSIONS

Our results support that dietary interventions may be efficacious for lowering blood pressure and blood lipids among overweight or obese premenopausal women. However, a decrease in SBP was the only favorable change that was sustained in this study population. These changes can be maintained over the course of a 1-year intervention, yet changes in blood lipids may be less sustainable.

Advancing the science of health disparities research

Research to eliminate health disparities in the United States is best approached from the perspective of population health. The objectives of this paper are to: (a) describe how ongoing research at the eight national Centers for Population Health and Health Disparities (CPHHD) is using a population health perspective and a community-based approach to advance the field of health disparities research; and (b) to discuss potential implications of such research for health policies that target some of the determinants of population health.

The recruitment of African-Americans to cancer prevention and control studies

BACKGROUND

African-Americans have the highest overall age-adjusted cancer incidence and mortality rates of any population group in the United States. Despite this fact, this group remains underrepresented in cancer prevention and control research studies, primarily because most recruitment strategies result in limited access to African-American populations.

METHODS

As part of three large-scale cancer prevention and control studies, effective strategies for recruiting African-American participants were developed and implemented.

RESULTS

Eight strategies have been identified as successful recruitment strategies for involving African-Americans in cancer prevention and control studies. Utilizing these strategies resulted in recruiting a representative number of African-American participants, in relation to the local population, into the three studies.

CONCLUSIONS

African-Americans can be recruited to participate in cancer control and prevention studies if plans include special strategies targeted to addressing unique barriers, beliefs, and concerns.