|Home | About | Journals | Submit | Contact Us | Français|
Racial disparities in lung cancer have been well described in the literature, however little is known about the perceptions of lung cancer in the general population and whether these perceptions differ by race.
Data are from the 2005 HINTS survey. Sample design was random digit dialing of listed telephone exchanges in US. Complete interviews were conducted on 5491 adults, of which 1872 respondents were assigned to receive questions pertaining to lung cancer. All analyses were conducted on this subset of respondents. SAS callable SUDAAN was used to calculate X2 tests and perform logistic regression analyses to model racial differences in perceptions of lung cancer. All estimates were weighted to be nationally representative of US population; jack knife weighting method was used for parameter estimation.
Black and White patients shared many of the same beliefs about lung cancer mortality, and etiology. African Americans were more likely than Whites (1) to agree that it is hard to follow recommendations about preventing lung cancer (OR 2.05 1.19-3.53 95% CI), (2) to avoid evaluation for lung cancer due to fear of having the disease (OR 3.32 1.84-5.98 95% CI), and (3) to believe that patients with lung cancer would have pain or other symptoms before diagnosis (OR 2.20 1.27-3.79 95% CI).
African Americans are more likely to hold beliefs about lung cancer that could interfere with prevention and treatment
Lung cancer is the leading cause of cancer mortality for both men and women in the United States accounting for 161,840 deaths in 2008.1 Survival for lung cancer is poor, with only fifteen percent of lung cancer patients surviving five years past diagnosis.1 African-American men have the highest incidence of lung cancer as well as the highest mortality.2 SEER lists the 2001-2005 incidence rate of lung cancer for US White men to be 79.3 per 100,000 vs. 107.6 per 100,000 for African-American men, and the mortality rate for white men with lung cancer is 71.3 per 100,000 vs. 93.1 per 100,000 for African-American men.1,3 The survival gap by race in lung cancer started in the early 1980’s and has been sustained since.4 Tobacco use leads to most lung cancer diagnoses; overall, 80-90% of those diagnosed will be current or ex-smokers.5 Although African Americans have historically smoked at higher rates than White Americans, this alone is not enough to explain for the difference in mortality and African-American men are 2-4 times more likely to have lung cancer even when adjusting for tobacco use.2
Access to surgical care, hospital level factors, biologic differences to tobacco smoke, and patient preferences have all been suggested as possible contributors to the mortality gap in lung cancer. 6-14 It is well known that black patients present with later stage disease, and with a higher burden of symptoms than their white counterparts. 2,15,16 However, several studies have also suggested that Black patients may have different feelings regarding risk perception, fatalism, and fear of a cancer diagnosis.7,17-19 Few studies have examined these issues with regard to lung cancer. Lung cancer is unique among solid tumor cancers due to its relentless mortality rate, and social stigma. 20 Given the confluence of prevention messages, treatment improvements, and deadly nature of the disease, differences in the perception of lung cancer could contribute to racial disparities. Also, lack of knowledge of the outcomes and risks of lung cancer could also contribute to complacency regarding treatment recommendations.18
All patients regardless of race underestimate the lethality of lung cancer. In the previous HINTS study (2003), when asked what type of cancer would cause the most deaths in the US the majority of men and women picked gender-specific cancers (breast for women, prostate cancer for men) at 49%, while only 26% of those surveyed choose lung cancer.21 When asked what percentage of lung cancer patients will live past five years, only 17% picked the correct answer (less than 25%). The majority of patients (36%) indicated that they believed that greater than 50% of lung cancer patients will live past five years.22 There were few other lung cancer specific questions in the 2003 HINTS survey; however .this general lack of understanding of the impact of lung cancer on mortality in the US is indicative of the paucity of information about lung cancer in the general population. There have been few studies to examine the level of knowledge about lung cancer in the US.
Therefore, we conducted an analysis using The 2005 Health Information National Trends Survey (HINTS), to examine the perception of lung cancer risk in a non-patient population in order to determine if there were differences in the perception of lung cancer by race or gender in the U.S. population.
Data are from the National Cancer Institute’s 2005 Health Information National Trends Survey (HINTS). Detailed description of the methodology of the survey and sampling strategy has been published elsewhere.23 Briefly, HINTS is a cross-sectional survey of health communication among the U.S. adult population. The HINTS survey was designed by the National Cancer Institute to ask questions about cancer. While demographic data is captured, prior medical history was not available via the HINTS survey. The survey is conducted on a nationally representative sample using random-digit dialing of listed telephone exchanges in the United States. The 2005 data was collected from February 2005 through August 2005. The 2005 response rate for the household level screener interviews was 34% and 61.3% for the extended interviews. Data are weighted to be nationally representative. The HINTS survey was administered to 5491 study participants, of which 4532 met our study criteria and were not missing data on key variables. Likewise, 1872 answered the lung cancer question and 1530 were included in our study because they met study eligibility and were not missing key variables. Our analyses were conducted on this subset of respondents.
The mental model portion of HINTS, in which respondents were asked questions about cancer risk perceptions, contained fifteen questions. This analysis included responses to the following eight questions in which respondents were asked Do you agree or disagree with the following statement: if they agree or disagree that: (1) there’s not much you can do to lower your chances of getting lung cancer, (2) There are so many different recommendations about preventing lung cancer that it’s hard to know which ones to follow, (3)lung cancer is often caused by a person’s behavior or lifestyle, (4)You are reluctant to get checked for lung cancer because you fear you may have it, (5) getting checked regularly increases chances of finding lung cancer when it is easy to treat and (6) people with lung cancer have pain or other symptoms prior to diagnosis. In addition, they were asked (7) compared to average person, would you say that for lung cancer that you are: (answer choices: about as likely, less likely and more likely) to develop lung cancer and (8) Overall, how many people who develop lung cancer do you think survive at least 5 years (answer choices: <25%, about 25%, about 50%, about 75% and nearly all).
Respondents were asked to indicate their ethnicity (Hispanic or non-Hispanic) and a separate race question asked them to identify their racial group. Due to the nature of our inquiry and small numbers of people from racial/ethnic groups other than Black and White, our analysis is limited to those classified as Non-Hispanic Black and Non-Hispanic White. Smokers were defined as individuals who have smoked at least 100 cigarettes in their life and also currently smoke every day or some days. Other covariates included: gender, income (<$25,000, $25,000 - 74,999 or $75,000+) and health insurance coverage (insured or uninsured).
All data analysis were conducted using SAS-callable SUDAAN software (version 9). We first used chi-square tests to examine bivariate association between respondents’ sociodemographic characteristics and responses to all 15 questions related to perception of lung cancer risk. Guided by the results of the bivariate analyses in which eight questions showed crude differences by race and prior literature on racial disparities, we tested racial differences in the answers to eight of the 15 questions by performing eight separate multivariable logistic regression models with each question as the dependent variable. These multivariable models controlled for smoking status, insurance, income, and age. In order to calculate appropriate population weight and design adjusted population estimates, all models were implemented using replicate jack knife weighting estimate of variance to account for the multistage sample design of the HINTS survey. We excluded from analysis all responses of “refused” and “don’t know” to survey items and all respondents with missing values for relevant variables.
Table 1 describes the survey participants for both the overall survey and the lung cancer questions. Approximately one fourth of the respondent sample was asked questions about lung cancer. Overall, the lung cancer sample appeared to be similar to the general sample. More women than men responded to the survey with nearly 2/3 of the respondents being women, and this pattern was exacerbated in African Americans. African Americans made up eight percent of the total sample. Nearly one half of the survey participants were either current or past smokers. The population was also well earned with 49% earning more than $75,000 per year in household income. Only 15 % of respondents earned less than $25,000.
We examined racial and gender differences in participant responses to the survey items. There were no differences by gender for any of the questions even when stratified by race (data not shown), and eight questions showed crude differences by race. We used these eight questions, listed in Table 2, in multivariable analyses.
The percentages of white and black patients who agreed with each of the questions are listed in Table 3. On the whole, the majority of respondents indicated that they agreed that lung cancer was preventable, caused by behavior or lifestyle, and believed that screening for lung cancer could be successful. Black respondents were more likely to agree that there were too many recommendations for preventing lung cancer (53% versus 37%; p=0.02), be reluctant to be checked for lung cancer (22% versus 9%; p=0.003), and to expect symptoms prior to diagnosis of lung cancer (51% versus 32%; p=0.004). Black respondents were also less likely to agree that lung cancer is caused by behavior or lifestyles (73% versus 85%; p=0.013). There was little variation in the questions about perceived lung cancer risk between whites and blacks, with a majority of both indicating that they did not place themselves at high risk for lung cancer. Responses to the question about lung cancer mortality showed borderline differences, with African Americans being slightly more optimistic about the lung cancer prognosis (p=0.05), but with both groups overestimating the survival of lung cancer dramatically.
We performed logistic regression analyses comparing the likelihood to agree with each question by race. Age, smoking status, and education had no significant effect as individual predictors of racial difference in the univariate model. For face validity, (Data not shown) we adjusted for age, smoking status, insurance, and income in the multivariate models. Gender was not included given the distribution of the respondent sample. Compared to white respondents, black respondents had two times higher odds of agreeing that there were so many recommendations about preventing lung cancer that they don’t know which to follow (OR: 2.05; 95% CI 1.19-3.53). Black respondents also had higher odds of reporting that they are more reluctant to get checked for lung cancer (OR 3.32; 95% CI 1.84-5.98), and higher odds of endorsing that symptoms occur before lung cancer diagnosis (OR 2.20; 95% CI 1.27-3.79). Additionally, blacks were also less likely to agree that lung cancer was caused by behavior or lifestyle (OR 0.48; 95% CI 0.24-95), and were less likely to agree that getting checked for lung cancer increases chances of finding the disease early (OR 0.48; 0.24-0.95).
Lung cancer remains the most lethal of all cancers in the US 1, and yet it lags not only in research funding,24 but also in public awareness of causes and prognosis. Public health messages regarding smoking cessation have been effective in pushing the smoking rate down to its current level of 20%;5 however, we will continue to see high volume of lung cancer in women, and the slow decline of lung cancer mortality in men.5 25 1 Smoking rates for both African American and White men have decreased dramatically, with rates in 2006 at 26.1 percent for Black men, and 23.5 percent for White Men.
Unfortunately, due to the lag between tobacco exposure and incidence, lung cancer will continue to be a major health concern as even smokers who have ceased tobacco use 20 or even 30 years ago are still at risk for lung cancer.5,18,19,26 The literature has very little information on what the public knows and thinks about lung cancer, and our current study extends previous finding by demonstrating the pervasive lack of knowledge in both racial groups.
We found that all races and genders grossly underestimate the lethality of lung cancer. There were no differences by gender in the lung cancer specific questions, but we did find some differences by race. Initial evaluation of the lung cancer specific questions showed that Black patients appeared to be more likely to expect more symptoms, be more reluctant to seek care due to fear of disease, be confused about preventative recommendations, and are more likely to doubt the association with lifestyle. Given the presentation of lung cancer, these beliefs could have an effect on prevention messages, seeking appropriate medical care for symptoms, and the physician patient interaction in regards to seeking and accepting treatment. If patients doubt the link between smoking and lung cancer, this may translate to difficulty with smoking cessation.27 If patients are reluctant to go the physician for fear of diagnosing the disease, this could lead to difficulty in accepting care for the disease. Margolis et. al. noted that African Americans were more likely to believe that surgery would worsen their disease, when compared to their white counterparts.7 Previous studies using the 2003 HINTS data had shown that many had fatalistic beliefs about cancer but indicated that there were few differences by race.18
Studies have detailed the racial gap in lung cancer incidence and mortality, 9,28-31 but few have examined the role of the patient’s perceptions of the disease and its possible effects on evaluation and treatment of lung cancer. Racial disparities in health are multifactorial; with access to care, patient factors such as lifestyle choice and lifetime exposures, and physician factors all having some impact
The differences observed in the present study indicate that we must better communicate information about lung cancer to the public. 85% of lung cancer patients are current or ex-smokers, and as such, prevention should be one of the most important public health messages that we send about lung cancer. However, it cannot be the only message. Patients of color must also know about the risks of lung cancer. A discussion with community leaders about risk factors, prevention and treatment, as well as the natural history of lung cancer could go a along way in eliminating the misperceptions seen in our evaluation of this database.
There are many limitations to our study. The 2005 HINTS survey, even though designed to be representative of the population, had a low response rate. This possibly contributed to bias in those who comprise our dataset. The study had a majority of women, and as such we might have underestimated gender difference due to the make up of the sample. Likewise the economic distribution was skewed towards a higher income than we expected. Thus the voices of men, and poor were not as evident here, and these are populations that are more likely to be affected by lung cancer. Low power limits the subgroup and sensitivity analysis of the effects of age, smoking status and income in this study.
Although the prognosis of lung cancer remains poor, early stage lung cancer is curable. There are effective treatments for locally advanced lung cancer, and palliative treatments for later stage disease that can extend and improve quality of life.5 Lung cancer does not have early clinical warnings signs. Once symptoms such as cough, or hemoptysis appear, most patients will already have advanced disease.32 There is no known screening mechanism for lung cancer, but reluctance to seek a physician for fear of a diagnosis of cancer is a grave concern, as Black patients present with later stage disease.16 In addition, if indeed African American patients are confused about the warnings for preventing lung cancer, then this is an important finding. The effect of tobacco on lung cancer risk is well documented, and unlike other cancers, there are few other avoidable risk factors that have similar impact. The idea that some are confused by the lung cancer prevention messages is surprising. Likewise, doubting that lung cancer can be linked to behavior might grossly undermine prevention efforts. Further efforts should involve discussion of the role of lung cancer in all communities including communities of color, and addressing the knowledge gap.
Funding/Support: Dr. Lathan is supported by NCI KO1 CA124581, and by the NCI funded Program in Cancer Outcomes Research Training (5R25CA092203)
Presented in part at the 2007 Annual Meeting of the American Society of Clinical Oncology