Provider Discussion about Lifestyle by Cancer History: A Nationally Representative Survey

Health care providers have a role to play not only in treatment of disease, but also in promotion of a healthy lifestyle, particularly among those with a history of cancer who are at higher risk for recurrence and comorbidities. Cancer survivors, defined as such from the moment of diagnosis, numbered an estimated 16.9 million in the United States in 2019 (1). The American Institute for Cancer Research recommends that survivors maintain a healthy body weight, perform physical activity, consume a healthy diet (including fruits and vegetables, while limiting high fat foods, sugar sweetened beverages, and red and processed meat), and refrain from smoking (2). These recommendations have been further described by prominent professional societies, including the National Comprehensive Cancer Network (3) and the American Cancer Society (4). Furthermore, among all adults, including cancer survivors, the U.S. Preventive Services Task Force (USPSTF) suggests behavioral counseling for overweight or obesity, which can be based on the framework of assess, advise, agree, assist, and arrange for addressing other risk behaviors, like tobacco usage (5, 6). The American Society of Clinical Oncology (ASCO) has described risks associated with overweight status among survivors and has advocated for increased research and delivery of services in this area (7). However, despite recommendations from the federal government and professional societies, providers receive limited training in behavioral counseling, especially on topics of nutrition or physical activity, and may not feel sufficiently prepared to advise or refer patients for lifestyle behavior adoption (8, 9). A study using the National Health Interview Survey (NHIS) data from the year 2000 found that only 30% of cancer survivors reported that their provider discussed diet or offered recommendations on diet, 26% were offered exercise recommendations, and 42% were asked about smoking during provider visits in the prior year (10). Prior research suggests that meeting lifestyle recommendations could prevent 48% of cancer-related deaths among women and 44% in men (11). Cancer diagnosis is sometimes seen as a teachable moment, after which individuals may be more motivated to change behaviors (12). In addition to standard primary care visits, cancer survivors may have more frequent healthcare utilization as they continue to see their oncology team for monitoring hormonal therapies or regular cancer screening. These physician visits represent a critical opportunity to screen for health behaviors and refer patients to appropriate programs. Given the significant percentage of the population who do not meet recommendations for lifestyle behaviors, we aimed to update previous findings of the 2000 NHIS study (10). Specifically, we examined recent nationally representative data on provider discussions and recommendations for behavior change related to weight, physical activity, diet, and smoking in individuals with a history of cancer versus those with no cancer history.

We used the NHIS, a continuous cross-sectional interview survey that uses a probability design to create a representative sampling of U.S. households and noninstitutionalized adults. The U.S. Census Bureau trains and employs staff according to procedures set by the National Center for Health Statistics. For this study, we used the Sample Adult File, with information on demographic characteristics, health history, and lifestyle behaviors. We compiled data from 2016 to 2017, as these years had the relevant questions we were interested in on provider lifestyle discussions. The final response rate for the Sample Adult File was 54.3% for 2016 and 53% for 2017.

We identified 59,770 individuals, 6,755 of whom reported a diagnosis of cancer (classified as such by the question “have you ever been told by a doctor or health professional that you had cancer or a malignancy of any kind?”). We excluded individuals who reported nonmelanoma skin cancer only (n = 1,674) and those who did not identify what type of cancer they were diagnosed with (n = 28). After these exclusions, there were 52,966 individuals with no cancer history and 5,053 who reported a diagnosis of cancer. As our questions of interest focused on questions that asked about provider discussions in the prior year, we further limited our dataset to the 44,385 individuals with no cancer history and the 4,792 cancer survivors who reported seeing a doctor or other health care professional within the prior 12 months.

We categorized demographic and health information as follows: age at questionnaire (18–<40, 40–<65, and 65+ years), sex (male and female), race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, and other), education (<high school, high school graduate, some college, and Bachelor's degree or greater), family income (≥$45,000, $20,000–$45,000, and <$20,000), health insurance coverage over prior 12 months (yes and no), functional limitations (limited in any way and not limited), body mass index (BMI in kg/m², 15–<18.5, 18.5–<25, 25–<30, 30–<35, and 35+), and self-reported health status (excellent, very good, good, fair, and poor). We also categorized people as reporting ever/never for diagnosis of the following chronic conditions: heart disease, stroke, hypertension, emphysema, and diabetes. We chose these five common conditions that were uniformly queried in NHIS as diagnosis ever/never.

Time since cancer diagnosis in years was calculated by subtracting reported age at diagnosis from age at interview. However, some individuals reported years since diagnosis rather than age at diagnosis (e.g., reported breast cancer diagnosis at age 1). Therefore, for cancer sites where <5% of diagnoses were among individuals under age 20 (using 2012–2016 Surveillance Epidemiology and End Results data), we assumed that the age at diagnosis was incorrectly reported as time since cancer diagnosis (n = 150).

Participants were asked the following questions: “during the past 12 months, have you been told by a doctor or health professional to do any of the following…” “…participate in a weight loss program?” and “…increase your physical activity or exercise?” Individuals were also asked, “during the past 12 months, has a doctor or health professional talked to you about…” “…your diet?” and “…your smoking?” (among those who reported current smoking).

The complex sample design in NHIS involves stratification, clustering, and multi-stage sampling. U.S. Census Bureau population estimates were used to assign person-level weights, which are adjusted quarterly by age/sex/race/ethnicity classes to provide national estimates. In tables and text, our reported Ns are unweighted and percentages are weighted. The R package “survey” was used to account for complex survey procedures. Weights were divided by two to account for combining 2 years of data. We used descriptive statistics and weighted χ² tests to compare lifestyle discussions by cancer history. We used generalized linear mixed models to examine predicted percentages of provider discussion, adjusting for age (continuous), sex, race/ethnicity (white and non-white), education (<high school, completed high school, 2- or 4-year college, and graduate education), comorbidities (continuous, 0–5), functional limitations (limited and not limited), and number of visits to an outpatient provider (continuous). We used mean values of covariates to calculate predicted percentages. We first examined predicted percentages of provider discussion and recommendations overall, and then stratified by those meeting the guidelines compared with those who did not meet guidelines for BMI (25–60 kg/m²) and physical activity (aerobic recommendations of 150 minutes/week). In sensitivity analyses, we examined differences in outcomes when we stratified BMI by only those reporting 30–60 kg/m², and examining those who met both aerobic and strength physical activity recommendations. In additional analyses, we examined counseling by specific cancer site compared with the population with no cancer history, limited by age. For each cancer site, we identified the youngest age at diagnosis and used only individuals of that age or older in our comparison group; we also limited the comparison group by sex for breast, prostate, cervical, and uterine cancers. Thus, breast cancer survivors were compared with women age 26+ with no cancer history, prostate cancer survivors were compared with males age 25+ with no cancer history, melanoma survivors were compared with those age 20+ with no cancer history, cervical cancer survivors were compared with females age 21+ with no cancer history, colon cancer survivors were compared with adults age 30+ with no cancer history, and uterine cancer survivors were compared with females age 27+ with no cancer history. We also examined prevalence of provider discussions by time since diagnosis, patient age, and race/ethnicity. R (version 3.6.0) was used to conduct descriptive statistics, χ² tests, and regression models. Python (Pandas, Python 3) was used for data management.

In our final dataset, we included 44,385 individuals with no cancer history and 4,792 cancer survivors who reported seeing a doctor or health care professional within 12 months of the questionnaire (Table 1). Those with a cancer history tended to be older (59.3% vs. 22.7%, over age 65), were more likely to be female (60.7% vs. 56.6%), more likely to be non-Hispanic white (81.6% vs. 67.3%), less likely to report an income more than $45,000 (13.6% vs. 24.9%), and more likely to report functional limitations (68.6% vs. 40.1%) than those with no cancer history (all P < 0.001). Cancer survivors were less likely to be never smokers than those with no cancer history (47.9% vs. 62.6%, respectively; P < 0.001) and were slightly less likely to be normal weight (BMI 18.5–25 kg/m², 29.9% vs. 32.3%; P = 0.002). In addition, those with a cancer history reported worse self-reported health status (26.8% vs. 13.8%, fair or poor) and higher prevalence of comorbidities (25.7% vs. 12.5%, reporting 2+ comorbidities) than the population with no cancer history (all P < 0.001).

In Table 2 we present the distribution of cancer sites and average years since diagnosis. Numbers ranged from n = 111 (kidney) to n = 989 (breast) cancer survivors. Lung cancer survivors reported the fewest mean years since diagnosis (5.6 years; SE, 0.3) and cervical cancer survivors reported the longest mean years since diagnosis (20 years; SE, 0.4).

We examined lifestyle discussions with a doctor or health professional during the prior 12 months using weighted, predicted percentages (Table 3). In multivariable adjusted models, overall, 6% of cancer survivors were told to participate in a weight loss program compared with 7.7% of those with no cancer history (P < 0.001). Among those classified as overweight or obese (BMI, 25–<60 kg/m²), 9.2% of those with a cancer history and 11.6% of those without a cancer history reported being told to participate in a weight loss program (P < 0.001), and among those with a BMI, 30–<60 kg/m² the percentages were 15.2% [95% confidence interval (CI), 13.3–17.3] and 19.3% (95% CI, 18.5–20.2), respectively (P < 0.001; data shown in text only). Overall, 31.7% of cancer survivors and 35.3% of those with no cancer history were told to increase their physical activity (P < 0.001). Among those who did not report meeting aerobic physical activity recommendations, these percentages were 38% and 41.6%, respectively (P = 0.003). Results were similar when looking at those who did not meet the combined aerobic and strength recommendations. Only 27.6% of cancer survivors and 32.2% of those with no cancer history reported having a general discussion of diet (P < 0.001); these percentages increased to 34.6% and 40.5%, respectively, among those with a BMI of 25–60 kg/m² (P < 0.001). When we combined a recommendation to increase physical activity with a general discussion of diet, we found a significant difference by cancer history, such that 17.5% of those with a cancer history (95% CI, 16.2–18.8) and 21.1% of those with no cancer history (95% CI, 20.7–21.8) reported both topics were discussed (data shown in text only). Among smokers, 67.3% of cancer survivors and 69.9% of those with no cancer history reported receiving counseling on smoking (P = 0.309).

We also examined lifestyle discussions separately by cancer site, comparing those with a specific cancer type with adults in the same age range with no cancer history (Table 4). There were no statistically significant differences in recommendations to participate in a weight loss program, increase physical activity, or discussions of diet by cancer history for those with the following cancer histories: breast, prostate, cervical, colon cancer, or melanoma. Among those with uterine cancer, we observed differences in recommendations: 18.1% were told to participate in a weight loss program compared with 11.8% females of the same age with no cancer history (P = 0.02) and 64% were told to increase physical activity compared with 51.8% with no cancer history (P < 0.001). Women with a history of uterine cancer were also more likely to report a general discussion on diet (51.9% vs. 41.1%; P = 0.001). When we added BMI to the endometrial models, statistically significant differences between groups only remained for increasing physical activity, where 47.4% of survivors reported the recommendation compared with 39.3% of those with no cancer history (P = 0.036; data shown in text only). There were no differences in discussion of smoking among those reporting smoking by cancer site compared with adults with no cancer history.

We also separately stratified lifestyle counseling by time since diagnosis, age, and race/ethnicity (Table 5). There was no statistically significant difference by time since diagnosis and predicted lifestyle counseling (all P > 0.05), although recommendations for participating in a weight loss program and increasing physical activity were slightly more common in those with 3–10 years from diagnosis compared with those within 2 or 11 or more years from diagnosis. In analyses stratified by age, we found that those ages 40–64 years were more likely to receive recommendations to participate in a weight loss program, to increase physical activity, and to have a general discussion on diet than those ages 18–39 or 65+ years (all global P < 0.001). In analyses stratified by race, it appeared that a greater percentage of non-whites reported being told to participate in a weight loss program, increase physical activity, or discussed diet with their provider (all global P < 0.001). Younger individuals and those reporting a race other than non-Hispanic White or non-Hispanic Black were less likely to receive counseling on smoking (P < 0.001 and P = 0.010, respectively).

In this study we found differences in lifestyle recommendations by cancer history, such that a higher percentage of those with no cancer history reported these lifestyle discussions than those with a cancer history. We thus identified a need to increase provider discussions among cancer survivors about weight loss, physical activity, and diet, particularly among those not meeting public health recommendations. Cancer survivors are at higher risk than the general population for health complications, including recurrence, cardiovascular disease, and mortality (13, 14); this risk could be mitigated by maintaining a healthy body weight, being physically active, and refraining from smoking.

Our findings are consistent with previously published literature on provider counseling prevalence. While we followed a similar approach to that conducted by Sabatino and colleagues using the 2000 NHIS data, both the questions asked in the survey and the model adjustment technique differed, precluding direct comparison between the predicted percentages (10). Still, it appeared that the percentage of individuals reporting provider counseling on diet improved slightly from the prior study (24.6% vs. 27.6% in our study), and the percentage of individuals who were offered exercise recommendations improved from 21.5% in the prior study to 31.7% in our study (10). While this is promising, there is still a significantly missed opportunity to provide these lifestyle recommendations to cancer survivors. In our study, the percentage of individuals who were told to participate in a weight loss program or increase physical activity increased only slightly among those who did not meet the U.S. recommendations, suggesting that perhaps some providers counsel on weight loss, physical activity, and diet regardless of reported habits, and others do not mention these factors among patients at all. In our previously published study of NHIS data from 2013 to 2017, a greater percentage of cancer survivors reported obesity than those with no cancer history (31.3% vs. 29.3%), underscoring the issue of obesity in this population (15). Another study conducted using the Medicare Expenditure Panel Survey found that 25% of cancer survivors reported that their provider did not discuss lifestyle recommendations at all, 24.6% reported a brief discussion, and 38.6% reported discussing it in great detail (16).

For their part, cancer survivors are already asking for help to adopt healthy lifestyle behaviors and looking to their oncology team for advice. A recent survey conducted by the National Coalition for Cancer Survivorship among a national sample of 840 cancer survivors found that among the top five listed survivorship needs were maintaining a healthy weight and getting enough exercise (17).

Studies in primary care suggest models for provider-led lifestyle counseling. One study of National Health and Nutrition Examination Survey data reported that lifestyle counseling among prediabetic individuals led to improved lifestyle (18), underscoring the potential impact of physician discussions on lifestyle. Recent studies have summarized primary care provider intervention impacts on physical activity levels, showing some success in achieved physical activity or dietary behaviors with physician counseling (19–21). USPSTF has also published recommendations for screening and management of obesity in adults, including tips for primary care providers and others to counsel on weight loss (22). These models may be considered in adapting interventions for cancer survivors.

Counseling on the preventive behaviors recommended for cancer survivors is not yet part of standardized clinical survivorship care, although it is a recognized component of comprehensive survivorship care planning (23). While the ASCO has generated a position statement on obesity and cancer and provided tools and resources to help providers address obesity with their patients (7), published studies still identify a gap in practice. A recent survey of 971 oncologists found that most recognized that overweight and obesity are detrimental to cancer treatment outcomes, but approximately 35% reported that they rarely or never refer overweight or obese patients for weight management after therapy (24). Still, 78.5% reported advising patients to eat a healthier, balanced diet and reported referring patients to dieticians. Other studies show mixed results on whether oncologists regularly counsel patients on physical activity, with estimates of <50%–82% (25, 24). A qualitative study of clinicians at four cancer hospitals found that clinicians felt that they lacked training and knowledge, and were unsure of when to introduce physical activity to survivors (25). In addition, they reported little knowledge of community-based programs for referrals (26). There is a paucity of literature on the impact of oncologist-delivered referral on patient behavior change. While one study found that breast cancer survivors who received an oncologist recommendation were more adherent to physical activity compared with usual care, this study also found that oncologist recommendation plus referral was no different than usual care in reported physical activity levels (27). More research is needed to better understand how to support both patients and providers in completing lifestyle assessment and referrals.

While studies in primary care suggest that provider counseling on lifestyle improves behaviors, there is limited literature on the impact of oncologist recommendations to improve diet and physical activity or primary care–delivered counseling specific to cancer survivors (13). There is additional debate around the role of the oncologist in behavioral counseling for physical activity, and the extent to which various providers should be involved (28, 29). A recently funded study aiming to integrate physical activity data into the electronic health record for breast and endometrial cancer survivors (NCT04262180), may shed some light on strategies for improving provider feedback and monitoring of physical activity, building on published formative work (30). Meanwhile more active strategies are needed to give oncologists and primary care providers tools to assess, advise, and refer on strategies for health promotion specific to cancer survivors. Furthermore, studies are needed to understand patient perspectives, including knowledge, attitude, and behaviors about cancer survivorship care.

There are some existing tools intended to help guide oncologists in assessing and managing obesity, physical activity, and smoking. The American College of Sports Medicine's (ACSM) Exercise is Medicine initiative provides an action guide to providers to consider physical activity as a vital sign and to screen, provide a brief intervention, and refer to treatment as needed (31, 32). Furthermore, the ACSM is building a Moving Through Cancer registry to geographically document all physical activity programs developed for cancer survivors and to serve as a resource for both patients and providers (33). ASCO has long promoted tobacco cessation programs and provided resources to clinicians (34) and patients (35). Nononcology providers may rely on Centers for Disease Control's clinical tools, the USPSTF Cessation Recommendations, or tools from professional societies (6, 36, 37).

Strengths of our study include the nationally representative sample of the noninstitutionalized U.S. population. We were able to examine counseling by providers on multiple health behaviors overall and by cancer site, which has implications for targeting provider counseling on lifestyle. However, there are also limitations to our analysis. All NHIS data are self-reported, which may be subject to inaccuracies in recall, particularly over the time period of 12 months. We only included those who reported visiting a healthcare professional in the prior year; thus these results may not generalize to populations who do not have regular healthcare access. We were limited to the wording and questions collected through NHIS, and thus, were not able to distinguish by provider type, or further patient characteristics (e.g., former vs. never smokers). We were also unable to look at treatment among the cancer survivors, although we were able to examine differences by time since diagnosis. The cross-sectional design also precludes examination of change over time. NHIS includes only the noninstitutionalized population; therefore, we cannot generalize to those who are in nursing homes or other care facilities.

In this study, only 7.3% of cancer survivors who were overweight or obese were told to participate in a weight loss program, only 35.9% were told to increase physical activity levels, and only 29.2% had a general discussion of diet, which underscores a critical gap in delivery of lifestyle counseling from health professionals. While physicians cannot be expected to deliver ongoing counseling due to both time constraints and lack of training, lifestyle factors are critical to improving both late and long-term effects of cancer treatment, metabolic and cardiovascular health, and overall survival outcomes (38, 39). Importantly, as the cancer survivor population continues to grow, oncology teams need to make helping individuals live well with and beyond their cancer, not simply survive treatment, a priority. Furthermore, more research is needed to understand observed differences in provider discussions by age and race/ethnicity, and whether these differences may be due to implicit bias (40, 41). Health researchers, oncologists, and other health care providers may utilize the presented data to prioritize health behavior change in the cancer survivor and general population, with the goal of improving quality of life and multiple health outcomes. Providers still need additional tools, incentives, training in lifestyle counseling, and referral both for cancer survivors and for the general public, who also report high levels of obesity and physical inactivity, and a need for smoking cessation.

No disclosures were reported.

H. Arem: Conceptualization, resources, formal analysis, supervision, investigation, methodology, writing–original draft, project administration, writing–review and editing. X. Duan: Data curation, software, validation, methodology, writing–review and editing. D.K. Ehlers: Conceptualization, writing–review and editing. M.E. Lyon: Writing–review and editing. J.H. Rowland: Conceptualization, methodology, writing–review and editing. S.K. Mama: Conceptualization, writing–review and editing.

H. Arem, S.K. Mama, and D.K. Ehlers participated in the TREC Training Workshop R25CA203650 (principal investigator, Melinda Irwin). S.K. Mama was supported by a Career Development award from the NCI (K07 CA222335).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.