Healthy Behavioral Choices and Adherence to Recommended Cancer Screening among US Adults: National Health Interview Survey 2005

Abstract

Purpose: We investigated associations of healthy behaviors and cancer screening in a large sample from the US population

Methods: This analysis used the data from 2005 National Health Interview Survey and included adults at age≥18 years that completed the cancer questionnaires (15,995,240 cancer survivors and 201,590,979 respondents without cancer). Self–reported information on cancer history, healthy behaviors (body mass index [BMI], smoking, alcohol use, physical activity, fruit⁄vegetable consumption, and sunscreen use) was extracted from survey questionnaires. We examined associations of healthy behaviors and screening practices with survivorship status, overall and by gender.

Results: Overall, cancer survivors were more likely to be obese, to be current or former smokers, to use sunscreen regularly, and to follow recommended cancer screenings. Female cancer survivors were more likely to be former or current smokers as compared to their cancer–free counterparts (former: OR=1.43, 95%CI 1.22–1.69; current: OR=1.74, 95%CI 1.44–2.10), more likely to consume alcohol below the recommended limit (OR=1.20, 95%CI 1.01– 1.43) and to follow recommended cancer screenings. Male cancer survivors were more likely to be former smokers (OR=1.24, 95%CI 1.03–1.50) and to have inadequate fruit⁄vegetable intake (OR=1.36, 95%CI 1.03–1.80).

Conclusion: Certain cancer survivor groups can benefit from tailored preventive services addressing concerns related to selected healthy behaviors and screening practices.

Keywords: Healthy behaviors; Screening practices; Population–based survey; Cancer survivors.

Abbreviations

BMI: Body Mass Index; CI: Confidence Interval; NHIS: National Health Interview Survey; OR: Odds Ratio; PSA: Prostate–Specific Antigen; US: United States

Introduction

According to the American Cancer Society, over 1.7 million new cancer cases are expected in 2014 [1]. Recent advances in cancer detection and treatment resulted in improved survival for many tumor sites which increased the prevalence of cancer survivors [2]. Cancer survivors have a greater risk of second primary cancers and other co–morbidities including obesity, cardiovascular disease and hypertension, osteoporosis, and diabetes [3–7]. Healthy lifestyle choices help to manage long–term health consequences of cancer, by controlling and preventing the incidence of these conditions [8– 10]. Recommended healthy behaviors such as increased physical activity, proper nutrition, refraining from tobacco and alcohol use, maintaining a healthy weight, and adherence to recommended cancer screening guidelines have shown to reduce incidence of adverse health outcomes and to improve overall quality of life among cancer survivors [1,11–13].

Despite the potential benefits of healthy behaviors cancer survivors do not always adhere to these guidelines [13–15]. Further, the evidence on differences in healthy behaviors by gender is limited. We hypothesized that cancer survivors and respondents without a cancer history will differ with respect to selected healthy behaviors. The objective of this analysis was to describe the overall and gender–specific associations of healthy behaviors and adherence to recommended screening practices with cancer survivorship status in a population–based sample of US adults. Understanding these differences may help to develop tailored public health interventions to ensure long–term health benefits among cancer survivors.

Methods

Survey population and design

The National Health Interview Survey (NHIS) is a cross–sectional household survey administered to the non–institutionalized U.S. population by the National Center for Health Statistics of the Centers for Disease Control and Prevention to monitor the health of the U.S. population and to collect dsata for epidemiologic studies and policy analyses to address ongoing public health issues [16].

The survey utilizes a multistage probability sample design that has been previously described in detail [16]. The NHIS questionnaire collects information on basic health and demographics (core questionnaire), healthcare access and utilization, health insurance, and income and assets, and on current health topics. The data are collected separately from adults and children in each family. The 2005 NHIS survey includes a supplement that covers a variety of cancerrelated topics such as diet, physical activity, tobacco and alcohol use, and cancer screening.

Study population

This analysis uses 2005 NHIS data and is limited to adult participants (age 18 and older) who completed the cancer questionnaires. Participants who reported a history of cancer were referred to as cancer survivors. Participants reporting no prior cancer diagnosis were referred to as respondents without cancer. From 31,428 adults who completed the survey, we excluded 21(0.1%) individuals with missing cancer history. The final sample for this analysis included 2,428 cancer survivors and 28,969 respondents without cancer.

Healthy behaviors and cancer screening

We selected the following indicators of healthy behaviors to compare cancer survivors and respondents without cancer: body mass index (BMI), smoking status, alcohol use, physical activity, fruit⁄vegetable consumption, sunscreen use, and compliance with recommended screening. Self–reported height and weight were used to calculate BMI (kg⁄m²). BMI categories were defined as ≤24.9 (underweight or normal weight), 25–29.9 (overweight), 30– 34.9 (obese), and >35 kg⁄m² (morbidly obese) [17]. Smoking statusof the participants was categorized as never, former smoker, and current smoker. Dietary Guidelines for Americans for 2005 were usedto define the person’s alcohol use status as never use, former user, current use below recommended limit of >14 drinks per week if male or >7 drinks per week if female [18]. Use of sunscreen was categorized as always, sometimes, never, and unknown (including participants who reported not going into the sun).

Participants answered questions about weekly frequency and average duration of moderate and vigorous physical activity. From these, we calculated total minutes per week of vigorous or moderate activity and categorized physical activity as inactive (total minutes=0), insufficiently active (total minutes more than 0 and less than 150), and sufficiently active (total minutes equal to 150 or more) [19].

To estimate the intake of 18 food categories, the 5–Factor Screening was administered as part of the 2005 NHIS questionnaire [20]. For this study, adjusted cup equivalents of vegetables (excluding French fries) per day and adjusted cup equivalents of fruit were used to determine if participants met the US Department of Agriculture 2005 guidelines for adequate fruit⁄vegetable consumption (4.5 cups daily) [18].

Participants were asked about their participation in cancer screening (PSA, Pap–smear, and mammography, any colorectal cancer screening modality) and to specify when they had their last screening. Based on the provided information, compliance with the screening for prostate, breast, and cervical cancer was categorized as screening within the last 2 years (compliant or regular), screening within the last 2–5 years, and screening more than 5 years ago. For colorectal cancer screening, the categories were defined as screening within the last 5 years (compliant or regular), screening within thelast 5–10 years, and screening more than 10 years ago.

Statistical analyses

All analyses were performed using SAS version 9.3 (SAS Institute, Inc., Cary, NC, USA) and SAS callable SUDAAN10 (Research Triangle Institute, NC, USA). All analyses were adjusted for the complex sampling design (multistage probability sampling) in the NHIS data to obtain population estimates [16]. Because of this complex sampling design, all the results are presented in the form of weighted estimates, which in this case, are more appropriate than the sample statistics [16]. After weighting, our analyses included 15,995,240 cancer survivors and 201,590,979 respondents without cancer, which correspond to 2,428 and 28,969participants in the sample, respectively.

We used Χ² test to compare distribution of the following sociodemographic characteristics among cancer survivors and respondents without cancer: age (18–39, 40–49, 50–59, and ≥60 years), race⁄ ethnicity (White non–Hispanic, Black non–Hispanic, other), gender,marital status (married or partner, other), education (less than high school level, high school graduate or equivalent, some college orassociate degree, bachelors degree or above), employment (employed, unemployed) and insurance (uninsured, private insurance, Medicare, and other plans). Χ² test was also used to compare the distribution of healthy behaviors and screening practices among cancer survivors and respondents without cancer: BMI (≤24.9, 25–29.9, 30–34.9, >35kg⁄m,sup>2), current user below recommended limit of >14 drinks per week if male or >7 drinks per week if female [18], current user above recommended limit), physical activity (none, insufficiently active or <150 min⁄week, sufficiently active or ≥150 min⁄week), sunscreen use (always⁄most of the time, sometimes or rarely, never), adequate fruit⁄vegetable intake (no or <4.5 cups⁄day, yes or ≥4.5 cups⁄ day[18], prostate cancer screening (PSA within past 2 years, within past 3–5, within >5 years), colorectal cancer screening (within past 5 years, within past 6–10, within >10 years), breast cancer screening (mammogram within past 2 years, within past 3–5, within >5 years), and cervical cancer screening (Pap–smear within past 2 years, within past 3–5, within >5 years).

We used multivariate logistic regression to describe the overall association of healthy behaviors with cancer survivorship status (cancer survivors vs. respondents without cancer) while controlling for age, gender, race⁄ethnicity, marital status, employment, education,and insurance. Cancer survivors were compared to respondents without cancer (reference group). For each of the behaviors, thecategory representing the healthy choice was used as the reference. In overall analysis, we modeled overall screening practices as categoricalvariable with three levels: having all recommended screenings (colorectal and prostate cancer screening for men; breast, cervical and colorectal cancer screening for women), having some screening, and having no screening. Younger participants (age 40 and younger) who had not yet reached the recommended screening age were combined into a separate category. To retain observations with missing data on covariates in the logistic regression analysis, we created an “Unknown” category for each of the variables with missing data.

In a secondary analysis, we examined these associations separately among male and female respondents. Statistical significance was assessed at α=0.05 level in all analyses.

Result

Characteristics of cancer survivors (weighted N=15,995,240) and respondents without cancer (weighted N=201,590,979) are presented in Table 1. Compared to respondents without cancer, cancer survivors were older, were more likely to be females and White non–Hispanic (Table 1). Among women, breast, non–melanoma skin, female reproductive (cervical, ovarian or uterine), colorectal cancer, and melanoma were the most common cancers. Among men, nonmelanoma skin, prostate, colorectal, bladder cancer, and melanoma were the most common cancers (Supplementary Table 1).