Objectives. We explored racial/ethnic disparities in reports of smoking cessation advice among smokers who had visited a physician in the previous year. Also, we examined the likelihood of receipt of such advice across Hispanic subgroups and levels of English proficiency.
Methods. We analyzed data from the 2000 National Health Interview Survey.
Results. Nearly half of the 5652 respondents reported receiving smoking cessation advice from their doctor. Compared with Hispanics, and after control for a range of other factors, respondents in the non-Hispanic White (adjusted odds ratio [OR]=1.57, 95% confidence interval [CI]=1.2, 2.0), non-Hispanic Black (adjusted OR=1.44, 95% CI=1.0, 2.0), and other non-Hispanic (adjusted OR=2.19, 95% CI=1.3, 3.6) groups were significantly more likely to report receiving advice. English proficiency was not associated with receipt of physician advice among Hispanic smokers.
Conclusions. Some 16 million smokers in the United States could not recall receiving advice to quit smoking from their physician in the preceding year. These missed opportunities, compounded by racial/ethnic disparities such as those observed between Hispanics and other groups and between Hispanic subgroups, suggest that considerably greater effort is needed to diminish the toll stemming from smoking and smoking-related diseases.
Recent World Health Organization estimates rank tobacco as the leading cause of preventable death worldwide, responsible for 1 in 10 adult deaths, or about 4.9 million deaths each year.1 In the United States, the percentage of smokers peaked in 1965 at 42.4%, decreasing to 21.6% as of 2003.2,3 Despite this impressive decline, morbidity and mortality attributed to smoking remain excessively high.4 Between 1995 and 1999, an average of 440000 Americans died annually from cigarette smoking.5
One in every 5 deaths in the United States is believed to be smoking related, with minority groups bearing the greatest health burden.6,7 For instance, middle-aged and older African Americans are more likely than members of other racial/ethnic groups to die from coronary heart disease, stroke, or lung cancer; lung cancer is the leading cause of cancer deaths among Hispanic men and the second leading cause among Hispanic women5; and rates of adverse infant health outcomes because of maternal smoking are especially high among African Americans, Native Americans, and US-born Mexican Americans, especially Spanish speakers.6,8,9
To reduce the disease impact and economic costs of smoking and eliminate health disparities, the US Department of Health and Human Services (DHHS) aims, through the program Healthy People 2010, to reduce the prevalence of cigarette smoking among adults to 12% or less10 and to increase the percentage of health professionals who counsel their at-risk patients about tobacco use cessation to 85% or more.10 As a means of achieving these objectives, specific guidelines have been developed, including the recommendation that physicians engage in a 5-step process—“ask, advise, assess, assist, and arrange”—with their patients at every health care visit.11
Simple behavioral interventions such as physician-provided advice are effective, particularly when they are personalized and made relevant to patients’ symptoms, concerns, and values.12–17 A recent Cochrane review provides evidence that even brief advice produces a small but significant increase in the odds of quitting relative to no advice (or usual care) and an absolute increase of approximately 2.5% in smoking cessation rates.15
Research on smoking cessation interventions has indicated significant reductions in death and disability across the life span among smokers who quit.12 Coronary heart disease risk drops by 50% in the first year after quitting, and within 15 years the relative risk of dying from coronary heart disease for an ex-smoker approaches that of a long-time (lifetime) nonsmoker.4
We sought to describe potential racial/ethnic disparities in reports of physician-provided smoking cessation advice among smokers who had visited a health professional in the past year. Specifically, we were interested in determining whether Hispanic respondents were less likely to report having received smoking advice than members of other racial/ethnic groups. We also estimated the odds of being advised to quit smoking across levels of English proficiency among Hispanic smokers, given that previous reports had suggested that language might represent a barrier in terms of the likelihood of patients receiving such advice from their physician.18
We used nationally representative data from the National Health Interview Survey (NHIS), an annual cross-sectional, face-to-face, computer-assisted household survey of the civilian, noninstitutionalized adult population (18 years or older) of the continental United States, Hawaii, and Alaska.19 We analyzed information gathered from 5652 cigarette smokers sampled in the 2000 NHIS (weighted sample = 35 407930) who had visited a doctor in the previous year. Respondents were interviewed in English or Spanish, according to their language preference.19
We used the original NHIS “doctor advice to quit smoking/using tobacco” outcome variable. Variables assessed as independent correlates and potential confounders included race/ethnicity (according to the NHIS classifications of Hispanic, non-Hispanic White, non-Hispanic Black, and non-Hispanic other), gender, age group, region of residence, educational level, annual family income (below vs at or above the poverty line, defined as an annual family income of $20000), place of birth (United States vs elsewhere), health insurance coverage, usual source of health care (none, clinic health center, doctor’s office/HMO, hospital emergency room/outpatient department, other), self-reported changes in health status in the past year (better, worse, about the same), and history of smoking-related conditions.
The smoking-related characteristics assessed were smoking frequency (daily vs nondaily) and number of cigarettes smoked per day (less than 10, 10–20, more than 20). We assessed English proficiency according to the NHIS “language spoken” variable. His-panics who indicated that they spoke only or mostly English or spoke Spanish and English about the same were classified as highly proficient, and those who spoke only or mostly Spanish were classified as having low English proficiency. Hispanics were subdivided into 6 groups: Puerto Ricans, Mexicans, Mexican Americans, Cuban/Cuban Americans, Central/South Americans, and other (i.e., respondents from the Dominican Republic and those who identified themselves as being of multiple Hispanic, Latin American, Spanish, or unknown Spanish origin).
We calculated prevalence estimates using standard NHIS procedures that accounted for sampling probabilities as well as poststratification adjustments to compensate for variations in survey nonresponse. Estimated variances were based on Taylor series linearization conducted with SUDAAN (Research Triangle Institute, Research Triangle Park, NC), which was able to accommodate the complex NHIS sampling design.20 We used logistic regression modeling to account for the effects of potential confounders, including sociodemographic variables, health access/use variables, smoking behavior, health status, and smoking-related diseases, on the association between race/ethnicity and receipt of physician advice.
Stratified logistic analyses conducted among daily smokers and nondaily smokers incorporated all covariates other than smoking frequency. Here we express regression estimates from logistic models as odds ratios (ORs) and adjusted ORs. We conducted subsidiary analyses among Hispanics to examine the effects of language proficiency and subgroup on the likelihood of receiving advice from a physician to quit smoking.
In 2000, an estimated 23.1% of the US population (46.5 million individuals) smoked, of whom 75.2% reported at least 1 outpatient visit with a physician during the preceding 12 months. Selected characteristics of the study population (n = 5652) and the percentages who received smoking cessation advice are summarized in Table 1. As can be seen, 53% of the overall sample indicated having received advice from their doctor to quit smoking, with rates ranging from 34% among Hispanics to 59% among respondents in the “other” non-Hispanic group. In the case of all but 1 of the independent variables assessed (and all of the categories within these variables), Hispanics were significantly less likely than members of the other racial groups to report having received advice.
Results of the regression analyses revealed significantly reduced odds of having received advice among Hispanic smokers compared with members of the other race/ethnic groups after control for multiple potential confounders (Table 2). Specifically, in comparison with Hispanics, non-Hispanic Whites (adjusted OR = 1.57; 95% confidence interval [CI] = 1.2, 2.0), non-Hispanic Blacks (adjusted OR = 1.44; 95% CI = 1.0, 2.0), and other non-Hispanics (OR = 2.19; 95% CI = 1.3, 3.6) were significantly more likely to report having received smoking cessation advice.
Approximately 55% of daily smokers and 43% of nondaily smokers reported having received smoking cessation advice from their doctor (OR = 1.65; 95% CI = 1.4, 1.9; P < .001). Among daily smokers, non-Hispanic Whites (adjusted OR = 1.75; 95% CI = 1.3, 2.4), non-Hispanic Blacks (adjusted OR = 1.64; 95% CI = 1.1, 2.4), and other non-Hispanics (adjusted OR = 2.11; 95% CI = 1.2, 3.6) were significantly more likely than His-panic smokers to report receiving advice after control for all other covariates. Among nondaily smokers, only those in the other non-Hispanic group (adjusted OR = 3.61; 95% CI = 1.1, 11.5) were significantly more likely to report receiving advice than His-panic smokers.
No significant differences were found across levels of English proficiency in the likelihood of Hispanic respondents reporting having received advice to quit smoking (OR = 1.16; 95% CI = 0.7, 1.8; P > .05). The comparison across Hispanic subgroups revealed that members of all the subgroups were less likely than Puerto Ricans (the reference group) to report having been advised to quit smoking, although the difference was significant only among Central/South Americans (OR = 0.26; 95% CI = 0.1, 0.8; P < .05; data not shown).
NHIS data from the year 2000 show that just over half (53%) of smokers in the United States who had visited a doctor during the previous year reported receiving advice at that visit to quit smoking. The increase from the 37% of respondents in the 1991 NHIS who reported receiving advice18 demonstrates that important improvements have been achieved during the past decade. These improvements, however, have been inconsistent across racial/ethnic groups. The 1991 rates of reported physician advice of 31% to 40% increased to 50% to 60% in 2000 in all groups other than Hispanics, among whom the rate was lowest and remained virtually unchanged between the 2 periods (31% vs 34%). These findings indicate that more effort must be made to promote the delivery of antismoking messages if the Healthy People 2010 proposed target of 85% among all racial/ethnic groups is to be met10 and racial/ethnic disparities in smoking cessation rates are to be reduced.21
The finding that Hispanic smokers were less likely than smokers in other racial/ethnic groups to report having been advised by their doctors to quit smoking could not be accounted for by racial differences in a range of demographic and health-related factors. There is evidence that racial/ethnic discordance in patient–physician relationships impairs communication and participatory decisionmaking and that, in addition to language, other physician or patient factors may pose barriers.22 Nationally, only 21% of Hispanics report receiving regular care from a racially concordant physician, compared with 88% of Whites and 23% of Blacks.23
Moreover, the traditionally held perception that Hispanics are not a group at high risk of smoking may influence physicians’ advice practices. Although rates of smoking among Hispanics indeed are generally lower than those among other racial/ethnic groups (with the exception of Asian Americans), estimates indicate that 16% of Hispanic adults in the United States are smokers.3 Furthermore, significant variations have been noted in smoking rates among Hispanic subgroups,24,25 ranging from 11% among Central/South Americans to 16% among Cuban Americans, 19% among Mexican Americans, and 27% among Puerto Ricans.25
At the same time, Hispanics have been shown to less frequently report intentions to quit smoking,26 which may impede internalization and memory of smoking cessation messages. Arguably, from a health behavior perspective it may be of little difference whether Hispanics are less likely than members of other racial/ethnic groups to receive smoking advice, less likely to understand the advice given, or less likely to remember the advice. In view of the recent intensive targeting of Hispanics by tobacco companies and the concomitant increase in cigarette smoking among Hispanic adolescents in the 1990s,6,27,28 along with the rising prevalence of obesity, diabetes, and other cardiovascular disease risk factors among Hispanics,4 effective antismoking messages should be a central feature of health promotion activities within the growing Hispanic population.
As mentioned, in this study likelihood of receipt of smoking cessation advice among Hispanics was not associated with level of English proficiency. Similarly, the Commonwealth 2001 Health Quality Survey (which involved a national sample of US residents) revealed that slightly fewer than half (44%) of Hispanics experienced communication problems with their doctor, regardless of language proficiency.29 These findings are somewhat counterintuitive and challenge the hypothesis that language discordance poses a barrier to provision of smoking cessation messages.18,30,31
However, NHIS 2000 data show that language proficiency does have an impact on the likelihood of receipt of physician advice about diet and physical activity (unpublished data), suggesting that the lack of an association between language proficiency and receipt of smoking cessation advice in the present study was not because of insufficient statistical power associated with sample size. Culturally influenced assumptions and expectations shape the doctor–patient relationship and may represent a barrier to effective care that surpasses linguistic congruency.23,32,33 This issue and the role of cultural competency in the provision of health promotion messages warrant more focused investigation than the NHIS data set allows.
Our results, derived from a well-recognized, nationally representative data set, provide information that can be useful in assessing the implementation and effectiveness of current guidelines and strategies recommended by DHHS and the Centers for Disease Control and Prevention. At the same time, these findings are subject to several limitations. For example, although the data collection techniques used in the NHIS are designed to minimize reporting biases, studies focusing on smoking cessation advice have shown that smokers tend to overreport receipt of such advice,34,35 and it is not known whether accuracy of reporting differs across racial/ethnic groups. Also, a small percentage of the respondents (in approximately the 2% range) may have initiated smoking subsequent to the physician visit on which their report was based.
Despite consistent evidence supporting the efficacy of behavioral counseling in promoting smoking cessation12–17 and specific DHHS recommendations, we found that approximately 16 million smokers who visited a physician during the previous year could not recall receiving advice to quit. Furthermore, significant racial/ethnic disparities were observed in receipt of advice. Innovative methods of overcoming barriers to provision of such messages (e.g., vital sign stamps that include smoking status along with the traditional vital signs, physician training, computer-tailored health educational materials)36–40 are warranted, and health care providers must remain aware of their key role in reducing the burden of smoking-related diseases by consistently providing culturally appropriate smoking cessation advice to their patients of all ethnic backgrounds.
Note. Values are weighted percentages. The reference category is Hispanics. aAmong Hispanics only. *P < .05; **P < .01. Note. OR = odds ratio; CI = confidence interval. aAdjusted for gender, age group, region of residence, educational level, marital status, annual family income, place of birth, health insurance coverage, usual source of health care, number cigarettes smoked per day, health status during the previous year, and history of smoking-related conditions. bAdjusted for all of the variables just listed, along with smoking frequency. *P < .05; **P < .01.Characteristic Smokers With Physician Visit (n = 5652), % Smokers Receiving Smoking Cessation Advice (n = 2835), % Hispanics(n = 640), % Non-Hispanic Whites (n = 4019), % Non-Hispanic Blacks (n = 839), % Other Non-Hispanics (n = 154), % Overall 100 52.9 33.8 54.7** 49.9** 59.1** Gender Male 46.9 49.1 33.0 50.9** 43.7** 58.9** Female 53.1 56.3 34.8 58.0** 54.7** 59.5** Age group, y 18–24 15.0 43.8 28.9 45.4** 41.0** 47.7** 25–44 45.2 51.5 34.9 53.3** 46.5** 60.8** 45–64 32.2 58.7 34.6 60.9** 54.4** 63.3** > 64 7.6 55.5 37.6 55.5** 65.5** 46.3** Region of residence Northeast 19.3 54.5 39.0 55.7** 56.5** 56.1** Midwest 28.7 54.9 40.2 56.4** 47.5** 44.6** South 36.7 50.3 31.2 52.3** 47.1** 54.7** West 15.3 53.7 30.8 55.8** 61.6** 68.6** Educational level, y 0–8 5.2 45.9 29.4 53.6** 50.2** 41.0** 9–12 53.6 51.6 32.5 53.3** 46.8** 68.4** ≥13 41.2 55.7 39.5 56.6** 55.9** 56.0** Family income status At/above poverty line 75.0 53.2 34.6 54.7** 50.9** 55.9** Below poverty line 25.0 52.9 32.3 55.6** 49.8** 69.6** Place of birth United States 92.2 54.1 37.2 55.0** 51.2** 63.5** Elsewhere 7.8 39.4 30.4 47.3** 20.0** 51.6** English proficiency levela High 66.3 34.9 Low 33.7 31.7 Past year health status Better 16.8 57.2 33.5 60.2** 56.2** 55.0** Worse 11.8 61.8 48.3 63.2** 55.0** 76.4** About the same 71.4 50.4 31.2 51.9** 48.0** 58.5** Health insurance coverage Yes 81.9 54.6 39.2 55.8** 52.8** 58.8** No 18.1 45.1 20.7 48.9** 39.9** 62.2** Usual source of health care None 12.9 36.3 12.6 41.3** 28.6** 34.4** Clinic/health center 16.3 53.8 43.6 54.2** 56.0** 56.5** Physician’s office/HMO 66.3 55.4 37.9 56.8** 51.2** 61.3** Emergency room/outpatient 3.4 57.0 38.8 57.7** 56.0** 92.0** Other 1.3 64.5 73.1 64.1** 36.9** 100.0** History of cardiovascular disease Yes 27.7 63.3 45.7 63.7** 61.7** 85.0** No 72.3 48.8 30.8 51.3** 42.8** 45.7** History of respiratory problems Yes 18.7 69.7 55.9 71.3** 64.0** 65.6** No 81.3 49.0 30.7 50.7** 46.8** 57.3** History of diabetes Yes 6.6 68.1 30.5 73.0** 66.0** 66.6** No 93.4 51.8 34.1 53.6** 47.9** 58.2** History of cancer Yes 6.0 65.1 60.1 65.7** 58.2 65.5** No 94.0 52.1 33.2 53.9** 49.6** 58.9** Smoking frequency Daily 82.7 55.1 35.6 56.6** 52.6** 57.9** Less than daily 18.3 42.6 30.5 44.5** 39.8** 64.9** No. of cigarettes per day < 10 25.7 43.2 31.1 45.2** 42.2** 61.0** 10–20 58.1 54.9 36.9 61.3** 55.1** 53.6** > 20 16.3 61.6 47.2 61.9** 50.2** 88.4** Overall Sample Receiving Advice Adjusted ORa (95% CI) Racial/Ethnic Group OR (95% CI) Adjusted ORb (95% CI) Daily Smokers Receiving Advice Nondaily Smokers Receiving Advice Hispanics (reference) 1.00 1.00 1.00 1.00 Non-Hispanic Whites 2.36 (1.9, 2.9)** 1.57 (1.2, 2.0)** 1.75 (1.3, 2.4)** 1.17 (0.7, 1.9) Non-Hispanic Blacks 1.95 (1.5, 2.6)** 1.44 (1.0, 2.0)** 1.64 (1.1, 2.4)** 1.03 (0.6, 1.8) Other non-Hispanics 2.83 (1.8, 4.5)** 2.19 (1.3, 3.6)** 2.11 (1.2, 3.6)** 3.61 (1.1, 11.5)*
Catalina Lopez-Quintero was funded by a Milstein Doctoral Training Fellowship.
Human Participant Protection No protocol approval was needed for this study.