Background
Higher mortality and morbidity burdens continue to disproportionately accrue to individuals from underrepresented racial and ethnic backgrounds into middle and old age [
1‐
4]. Chronic disease complications are largely preventable with timely and effective care, and by having responsive disease management plans in place. Smoking is widely recognized as the main modifiable risk factor to prevent exacerbations of existing chronic disease and avert the development of incident chronic conditions. As a result, effecting changes in smoking behavior are among the first line of recommended actions to avert onset of cardiovascular disease, pulmonary disease, and various forms of cancer [
5].
There are important differences in chronic disease morbidity and mortality for underrepresented racial and ethnic groups relative to whites. Both US-born Latinos and black Americans have greater smoking-related mortality than white Americans [
5], and contend with significantly higher rates of cardiovascular and metabolic disease than whites even into older age [
6]. Efforts to encourage behavior change after health problems have already set in (i.e., secondary prevention efforts) are critical to meeting public health goals of effective chronic disease management and health maintenance for older adult populations [
7]. In addition, promotion of smoking cessation programs in late life may help in averting adverse events such as cancer recurrence, repeated heart attacks, and exacerbations in disease severity for older adult smokers that remain resistant to quitting [
8‐
10]. A new chronic disease diagnosis could potentially serve as a warning, precipitating health behavior change and increased self-health vigilance [
11,
12], or a “teachable moment” to intervene on adults with newly diagnosed chronic conditions [
13‐
15]. Indeed, a small number of studies find that new health problems may actually improve the health and health habits of older adults [
7,
10,
16].
There is little information about whether cessation attempts at older ages are more or less likely across racial and ethnic groups, and these differences may be an important source of disparities in health, including disease prevalence, prognosis after diagnosis, and quality of life.
While health behaviors are widely recognized as influential contributors to morbidity and mortality for the adolescent and young adult population [
8], less attention has been focused on the effectiveness of smoking behavior modification among older adults [
17‐
19]. Further, the adoption and maintenance of healthful behaviors—such as smoking reductions and cessation—may be of particular importance in the prevention of disease and compromised function associated with aging [
18,
19].
The purpose of this study is to assess whether the onset of chronic disease elicits differential changes to smoking behaviors between black, Latino, and white older adults. To date, little is known about whether older adults from various racial and ethnic backgrounds are more or less likely to make lifestyle modifications in response to adverse changes to their health. These insights are critical in evaluating illness perception and behavior change as well as in assessing the adequacy and availability of secondary prevention efforts for underrepresented racial and ethnic older adults.
Discussion
The objective of this study was to investigate changes in smoking behavior between black, Latino, and white adults after a chronic illness diagnosis. Moderate-to-low percentages of middle-aged and older adults quit smoking after the onset of a new chronic disease diagnosis, and these varied widely by race/ethnic group and by chronic disease type. Similarly, we found modest reductions in the quantity of cigarettes smoked for middle-aged and older adults who continued to smoke after a newly diagnosed chronic disease, and this too varied widely by race/ethnic group and chronic disease type. Although observed reductions represent some positive change to smoking behavior among older adults, the majority of respondents continued to smoke after being diagnosis with a new illness.
The smoking patterns of older adults have been largely ignored by researchers, who have focused predominantly on younger adult smoking behaviors [
26]. Among non-elderly adults, whites are more likely to quit at any given time point, and importantly, black Americans—who predominantly smoke mentholated cigarettes—have lower cessation rates and experience higher rates of smoking-related health consequences relative to white Americans despite smoking fewer cigarettes, on average [
27‐
29]. This literature provides a useful counterpoint to our findings for middle-aged and older Americans of relatively low smoking cessation across race/ethnic groups in the face of health declines.
We found several interesting differences in smoking behavior change between black, Latino, and white older adults. Relative to whites, black and Latino respondents were less likely to smoke after suffering a stroke. Stated conversely, older white adults were less likely to quit after stroke than were black and Latino older adults. Comparing changes in smoking behavior between black and Latino middle-aged and older adults, Latinos were less likely to quit smoking after newly diagnosed heart disease. Further, white middle-aged and older adults smoked the greatest quantity of cigarettes despite health downturns. Relative to whites, black participants smoked fewer cigarettes after a new diabetes, heart disease, stroke, or cancer diagnosis, and Latinos smoked fewer cigarettes after a new diagnosis of diabetes or heart disease. These differences may have important health consequences for older adults—particularly older white adults—who continue to smoke after chronic disease onset, which could lead to exacerbations and complications of the disease, or other serious adverse events.
Efforts to promote health behavior change for individuals from socioeconomically and ethnically diverse backgrounds have generated a nuanced discussion in the literature. On one hand, high-risk and underserved populations should be a focus of prevention efforts given lower life expectancy in the US, greater likelihood of cancer mortality, and greater likelihood of engaging in risky health behaviors [
30]. In addition, high prevalence of risky health behaviors—smoking in particular—among disadvantaged older adults may contribute to maintaining inequalities in mortality and morbidity into middle and old age [
19]. Our study explicitly compared changes in smoking behavior between white, black, and Latino older adults, and we did not find that blacks or Latinos were less likely to quit or reduce smoking levels relative to whites. Our findings provide support for better access to smoking cessation counseling and therapies for middle-aged and older adults regardless of race/ethnic background, but may warrant particular emphasis on whites whom we found less likely to quit smoking after suffering a stroke, and who smoke 1.5 to nearly 2 times as many cigarettes as blacks after newly diagnosed diabetes, cancer, heart disease or stroke.
Our findings suggest there may be differences in smoking behavior change after the diagnosis of specific chronic diseases, but we could not ascertain why or how this occurred. Based on other research, we can speculate about several potential explanations. There may be differences in the importance assigned to the diffusion of smoking cessation counseling by clinicians in the face of critical health events, such as a stroke. In addition, the ability to react to adverse health events and assimilate health warnings brought on by a recent diagnosis or health decline may be constrained for some individuals. Lack of access to timely smoking cessation counseling may have important repercussions for secondary prevention efforts to avert poor health outcomes for older adults [
31]. While our results do not speak to the causes of continuing, stopping, or reducing smoking behaviors, our overall findings imply that middle-aged and older adults may have different patterns of smoking, may have other deterrents to stopping smoking, may not be effectively targeted by smoking interventions, or may have different perspectives on the effects of smoking on health [
31].
This study contributes to the understanding of health behaviors in important ways, and draws from a number of strengths. First, the prospective study design assesses health behaviors prior to the diagnosis of the condition and is less susceptible to recall and social desirability biases. Second, this study compares lifestyle changes among the most serious chronic conditions, an improvement over many earlier studies that examine health conditions in isolation. Third, the consideration of racial and ethnic group differences in smoking behavior change in response to chronic disease onset provides important insights on reducing subsequent health risks for white, black, and Latino older adults. Finally, this study contributes to the understanding of whether behavioral processes are similar or different for minority and white individuals and where improvements in chronic disease management are most needed.
Several limitations should also be noted. First, our measures of health conditions, including disease diagnosis and smoking behaviors were derived from self- report. To the extent that there is under-reporting of health conditions [
32], any bias would likely overestimate changes in smoking behavior, because individuals with pre-clinical indications of disease (e.g., pre-diabetes) would not have been included and might be less likely to receive repeated smoking cessation messaging or would be less motivated to change. However, several studies have shown reasonable concordance between self-reported chronic conditions and other methods of ascertainment [
33‐
35]. Second, the present study is not a matched-cohort controlled study, therefore we should be cautious in attributing changes in smoking behavior to new disease onset. However, the objective of our study was not to examine smoking cessation outcomes relative to controls who are not diagnosed with chronic disease. There is a higher likelihood of cessation that has been shown relative to controls in a previous study [
18]. In the present study, we instead compare smoking changes over a 2-year window around a new chronic disease diagnosis between black, white, and Latino middle-aged and older adults in the US. Third, our data do not permit examination of quit attempts in our study. Cessation attempts may vary for individuals from underrepresented racial/ethnic background due to compounding influences with poor socioeconomic status, low literacy levels, lack of access to environments that are conducive to healthful lifestyles, not receiving important health messages from their health providers that are delivered in appropriate ways, and lack of access to and use of smoking cessation therapies. Further, success in quitting (i.e., lack of re-initiation) is more likely to occur for adults with greater socioeconomic resources [
36]. While we cannot assess cessation attempts, our study examines cessation and changes in smoking amounts around a critical clinical diagnostic window, which represents useful and actionable information to intervene on older smokers after they have developed a chronic condition. Finally, our study involves new diagnosis of serious health conditions; however, it is unknown whether individuals made lifestyle changes prior to diagnosis. Some individuals may reduce or quit smoking after receiving health warnings such as repeated high blood pressure or cholesterol measurements. Inclusion of medical record information of pre-diagnosis risk factors along with subsequent diagnosis of major conditions in future studies would provide important new information about whether or when individuals change behavior at earlier points of disease development.
There may be important extensions to this study that should be addressed in future work. While our study examined differences in changes to smoking after chronic disease onset between white, black, and Latino Americans, it is also important to understand the predictors and determinants of changes to smoking behaviors for each of these racial and ethnic background groups. Future studies should address important factors that predict cessation and smoking reductions for black, white, and Latino older adults. In addition, there may be important dynamics to smoking cessation and smoking reductions for middle-aged and older men and women, and for men and women from different racial/ethnic backgrounds. These considerations should also be addressed in future studies.