Methods
Study Sample
We obtained data from the 2010–2019 National Health Interview Surveys (NHIS) to examine self-reported cigarette smoking behaviors among adults aged 18 years or older with chronic disease. We chose to include adults aged 18 years or older on the basis of prior research related to the prevalence of multiple chronic diseases.[8] The NHIS is an annual, nationally representative, cross-sectional, household survey of the noninstitutionalized US civilian population that has previously been described in detail.[9] In 2019, NHIS underwent changes to nonresponse survey weighting methodology and questionnaire redesign.[10,11] Respondents with unreported age and missing cigarette smoking status were excluded (n = 878). Our analyses were conducted during 2020 through 2022. During 2010 through 2019, survey response rates for sample adults aged 18 years or older ranged from 53.0% in 2017 to 63.3% in 2011.[9,12]
Measures
Current cigarette smoking was defined as a person having smoked 100 or more cigarettes in their lifetime and smoking every day or some days at the time of interview.[2] Former cigarette smoking was defined as a person having smoked 100 or more cigarettes in their lifetime and not smoking at all at the time of interview.
Chronic Disease
Chronic diseases were assessed by self-report, asking participants if they had ever been diagnosed with any 1 of the 5 selected chronic diseases, or 2 or more. Chronic diseases were cancer (bladder, cervix, colorectal, esophagus, kidney, larynx, liver, lung, oropharynx, pancreas, stomach, trachea); COPD (emphysema, chronic bronchitis); diabetes; coronary heart disease (CHD); and stroke.[1] Participants were included in the analysis as having a chronic disease if they answered yes to "Have you ever been told by a doctor or other health professional that you had [disease]?", apart from CHD and COPD.
CHD and Myocardial Infarction (MI)
Separate questions were asked for CHD or heart attack and MI, an outcome of CHD. Respondents were coded as having CHD if they answered yes to having been told they had CHD or if they answered yes to having been told they had a MI, regardless of their response for CHD.
COPD
For COPD, participants were considered to have COPD if they answered yes when asked if they had ever been told by a doctor or other health professional that they had chronic obstructive pulmonary disease, also called COPD, 2) if they have ever been told by a doctor or other health professional that they had emphysema, or 3) if during the past 12 months they were told by a doctor or other health professional that they had chronic bronchitis.
Two or More Chronic Diseases
Participants were included as having 2 or more chronic diseases if they reported more than 1 of the aforementioned chronic diseases assessed in this study. Disease categories were not mutually exclusive.
NHIS questions did not allow us to distinguish between type 1 and type 2 diabetes, although cigarette smoking increases the risk of developing type 2 diabetes.[1] Data on kidney cancer was not accessible for 2019. Our analysis excluded acute myeloid leukemia because NHIS does not differentiate between leukemia and acute myeloid leukemia.
Statistical Analysis
We used 2019 NHIS data to calculate prevalence estimates and 95% CIs for current and former cigarette smoking among adults with chronic disease. We reported the prevalence of cigarette smoking by chronic disease for the following age groups: 18 to 44 years (young), 45 to 64 years (middle-aged), and 65 or older (older).[8]
We calculated the annual percentage change (APC) of current cigarette smoking from 2010 to 2019 for each chronic disease by age group using the National Cancer Institute's Joinpoint Regression Program version 4.8.01 (SEER*Stat), which uses log-linear models and a Monte Carlo permutation test for significant changes in trend.[13] Identification of a joinpoint at a given year indicates a significant change in trend. In the absence of joinpoints, APCs were considered constant and equal to average APC, which is a summary measure of APCs over a period of time.[14] Significance was defined as P < .05 for trends. To calculate prevalence estimates, we first used variance estimation variables to account for the multistage complex sampling design of the survey. Data were then weighted to provide nationally representative estimates. In accordance with the 2017 National Center for Health Statistics guidelines, statistically unreliable estimates were suppressed.[15] Analyses were conducted using SAS-callable SUDAAN software version 11.0.3 (RTI International).
Prev Chronic Dis. 2022;19(9):E62 © 2022 Centers for Disease Control and Prevention (CDC)
