Ke Xu, SzePui Tsang, Wanxin Qi, Ziyue Yang, Lesi He
It is known that smoking causes a lot of health problems, but according to the Centers for Disease Control and Prevention (CDC), there were about 34 million U.S. adults smoke cigarettes in 2019, and 58 million nonsmoking people were exposed to secondhand smoke that severely negatively affected their health. Especially since the first laboratory-confirmed case of Covid-19 appeared in the United States on January 20, 2020, bad addictions such as smoking have been highlighted by health organizations looking at their relationship with Covid-19. The Centers for Disease Control and Prevention pointed out, “Being a current or former cigarette smoker can make you more likely to get severely ill from COVID-19. If you currently smoke, quit. If you used to smoke, don’t start again. If you’ve never smoked, don’t start.” Besides, people with certain medical conditions are more likely to get severe Covid-19, and smoking could make their situation worse, because smoking harms the immune system and can make the body less successful at fighting disease. With these information, we examined the trend of smoking in the U.S. in recent years.
Our main objective is to provide information to health organizations and healthcare facilities about the characteristics of smokers and people who quit smoking in the United States, as well as the specific populations which need more attention.
The datasets of 2016 to 2020 were downloaded from the websites as mentioned above and were imported. We selected several potential variables for the project and created a YEAR variable to combine the datasets for analysis. The value of the data was given in number indicating the specific response of the participants, so we mutated and recoded the variables based on the codebook. For example, the disability
variable was renamed from the disab3_a
variable in the original dataset, and we transferred “1”, “2”, “9”, to “yes”, “no”, “don’t know”, respectively, indicating if the respondent has disability. Then we mutated the variables into factors, except for variables like year
and age
which are continuous. There were several variables containing missing value, and we filled them with “missing”. Lastly, we filtered the dataset based on the smoke
variable that we kept only current every day smoker, current some day smoker, former smoker, and never smoker, excluding those with unknown status.
The dataset was downloaded from BRFSS. It contains the tobacco use information of participants from all age, race, sex, and states. Since we are focusing years between 2017 to 2020, filter was applied to filtered out year equals to 2017-2020. There are 27 columns in the data, we only selected few columns useful to our analysis, which were locationabbr
(State Name), geo_location
(latitude and longitude for map), sample_size
(sample size), topic
(only question related to current smoking status is selected). Then we summed up the number of smokers in each state and calculated the smoking proportion with respect to their states.
Tax data was obtained from CDC open data website. We have three legislation categories (Combustible Tobacco Tax, Non-Combustible Tobacco, and Stamp Tax. Cigarette tax is the most common tax among 58 regions and takes a majority of available data (>90%) in this dataset (Some region did not specify tax for less common types of combustible and non-combustible tobacco). In this case, we selected cigarette tax value ($ per pack) under Combustible Tobacco category in 2021 and drop out meaningless columns as well as NA values. After that, we ranked 58 regions by cigarette tax value.
The dataset was downloaded from the CDC open data website. It was filtered by locations
(states in the U.S.) because it contains data from other regions. Also, it was filtered by the categorical variable Counseling and/or Medications
for studying smokers who seeks for help in counseling and/or medication. Then, we filtered the data by year
because we want to focus on smokers in 2020. Since the original data value was the number of calls per 1000 tobacco users, we divided the value by 1000 for better readability. Finally, the dataset is ready to be visualized.
Statistical analysis is not applicable to our project, because the circumstances that lead a person to smoke or quit are complex. Smoking or quitting is a personal choice that may be influenced by factors such as education level, sexual orientation, and social networking, but such analyses are too large. It is also difficult to extract a few or dozens of factors from hundreds of questions in the health surveys to produce a statistically significant model of smoking. If to analyze and predict smoking rates in the United States, more datasets are needed to calculate the smoking rate in the previous year, which is a huge amount of processing, and the variables can change from year to year, especially since Covid-19 started. Therefore, we just used the data to show the recent situation of smoking in the United States.
There is a slightly decreasing smoking trend from 2017 to 2020. Men and people aged 56 to 65 consistently make up the majority of smokers. Smoking rates have dropped among 18-to-25-year-old. Through comprehensive comparison of smoking rates, AIAN and other groups have the highest smoking rate, regardless of male or female. Smoking rates are lowest among Asians, especially women. The southeast has a higher smoking rate from 2017 to 2020 and based on the lower cigarette taxes in the southeast in 2020, we infer that there is some correlation. If there are more suitable data sets, we can do further investigation. West Virginia, Kentucky, and Tennessee kept the top three smoking rates for four years out of more than 50 states in the U.S. Utah and Puerto Rico kept the top three lowest smoking rates for four years, and California has done well in the past two years. Guam has the largest decrease and Texas has the largest increase of proportion of smoking comparing 2020 to 2017.
In 2020, minority groups defined by race, sexual orientation, and disability, and people with lower education levels still had a severe problem of high smoking rate. People with diabetes and stroke, who should quit smoking by 2020, did not pay much attention to this problem. The proportion of smokers increased among pregnant women and dementia patients. Health organizations and medical facilities should focus about these patients and vigorously publicize the dangers of smoking to their fight against disease.
We observed that there are less than half of smokers had make attempts to quit smoking in 2020. Among those who tried to quit smoking, more than 70% of them had stop for smoking for only one month or less. However, more than half of them responded that a doctor, dentist, or other health professional have advised them to stop smoking in 2020. Across the U.S., many smokers have called the quit line to for counseling and/or medication related to quit smoking, especially in the Florida, New York, and California. This is a positive outcome suggesting that many smokers are willing to quit smoking. Tobacco tax remains steady throughout these years, and we consider tobacco tax only has a slight influence on proportion of quitting smoking. Even though several regions with higher taxes have a higher decrease of smokers, more influencing factors are related to people themselves as well as environment.
The strength of our study is that people can get a sense of trends in smoking and quitting in recent years. We also looked at the smoking situation in the context of the COVID-19 pandemic. These could provide information to health organizations and healthcare facilities about which populations should be highlighted and noticed.
Our study has several limitations. First, as discussed in the statistical analysis section, we find it difficult to come up with a statistically significant model due to the complexity of variables. These survey data may be biased because respondents respond to the questionnaire based on their own situation and feelings. Questions such as “smoking frequency” were answered by self-assessment, which may vary from person to person. In the study, there was no clear experimental data to reasonably justify these manifestations. It leads to the second limitation, as there were no strict rules to define a smoker. Some people may smoke only occasionally for social purposes, while others may smoke packs or even packs of cigarettes a day. We can only categorize roughly by “current every day smoker”, “current some day smoker”,“former smoker”, and “never smoker”. Third, due to limited data, we focused only on cigarettes, while there are many other forms of tobacco smoking, such as small cigarettes, cigars and other non-combustible smoking, such as chewing tobacco and e-cigarettes.
All in all, smoking is not a healthy behavior for the smokers as well as the people around the smoker because smoking will cause many health issues. According to the CDC, quitting smoking has health benefits at any age, no matter how long or how much you have smoked. We hope that more policies can be enforced to encourage current smokers to quit smoking and further reduce smoking rates in the United States.