The novel coronavirus (Covid-19 or SARS-CoV-2) has gripped the world since its believed outbreak in late 2019. Governments and people around the world have since had to endure personal, economic, and social hardships in an unprecedented scale. The vaccine rollout for Covid-19 began in in January, 2021 for ages 65 and older. By August, around 70% of the US population are reported to have been vaccinated with at least the first dosage. However, this rate is still 8.5 million people short of its previous goal to fully vaccinate 165 million individuals. In order to reach this goal, this research aims to identify the driving factors which increase the likelihood of an individual getting innoculated. The research looks at previous data regarding the H1N1 Influenza outbreak in 2009 as well as data regarding seasonal flu shots. By identifying features that have the most impact in vaccination rates for these two viruses, the CDC can implement new campagins to improve vaccination rates to combat further outbreaks of Covid-19.
- What features determine the likelihood of vaccination against H1N1 Influenza?
- What features determine the likelihood of vaccination against the Seasonal Flu?
- What strategies can the CDC implement in improving Covid-19 vaccination rates using analyses about previous corovirus outbreaks?
- h1n1_concern - Level of concern about the H1N1 flu. 0 = Not at all concerned; 1 = Not very concerned; 2 = Somewhat concerned; 3 = Very concerned.
- h1n1_knowledge - Level of knowledge about H1N1 flu. 0 = No knowledge; 1 = A little knowledge; 2 = A lot of knowledge.
- behavioral_antiviral_meds - Has taken antiviral medications. (binary)
- behavioral_avoidance - Has avoided close contact with others with flu-like symptoms. (binary)
- behavioral_face_mask - Has bought a face mask. (binary)
- behavioral_wash_hands - Has frequently washed hands or used hand sanitizer. (binary)
- behavioral_large_gatherings - Has reduced time at large gatherings. (binary)
- behavioral_outside_home - Has reduced contact with people outside of own household. (binary)
- behavioral_touch_face - Has avoided touching eyes, nose, or mouth. (binary)
- doctor_recc_h1n1 - H1N1 flu vaccine was recommended by doctor. (binary)
- doctor_recc_seasonal - Seasonal flu vaccine was recommended by doctor. (binary)
- chronic_med_condition - Has any of the following chronic medical conditions: asthma or an other lung condition, diabetes, a heart condition, a kidney condition, sickle cell anemia or other anemia, a neurological or neuromuscular condition, a liver condition, or a weakened immune system caused by a chronic illness or by medicines taken for a chronic illness. (binary)
- child_under_6_months - Has regular close contact with a child under the age of six months. (binary)
- health_worker - Is a healthcare worker. (binary)
- health_insurance - Has health insurance. (binary)
- opinion_h1n1_vacc_effective - Respondent's opinion about H1N1 vaccine effectiveness. 1 = Not at all effective; 2 = Not very effective; 3 = Don't know; 4 = Somewhat effective; 5 = Very effective.
- opinion_h1n1_risk - Respondent's opinion about risk of getting sick with H1N1 flu without vaccine. 1 = Very Low; 2 = Somewhat low; 3 = Don't know; 4 = Somewhat high; 5 = Very high.
- opinion_h1n1_sick_from_vacc - Respondent's worry of getting sick from taking H1N1 vaccine. 1 = Not at all worried; 2 = Not very worried; 3 = Don't know; 4 = Somewhat worried; 5 = Very worried.
- opinion_seas_vacc_effective - Respondent's opinion about seasonal flu vaccine effectiveness. 1 = Not at all effective; 2 = Not very effective; 3 = Don't know; 4 = Somewhat effective; 5 = Very effective.
- opinion_seas_risk - Respondent's opinion about risk of getting sick with seasonal flu without vaccine. 1 = Very Low; 2 = Somewhat low; 3 = Don't know; 4 = Somewhat high; 5 = Very high.
- opinion_seas_sick_from_vacc - Respondent's worry of getting sick from taking seasonal flu vaccine. 1 = Not at all worried; 2 = Not very worried; 3 = Don't know; 4 = Somewhat worried; 5 = Very worried.
- age_group - Age group of respondent.
- education - Self-reported education level.
- race - Race of respondent.
- sex - Sex of respondent.
- income_poverty - Household annual income of respondent with respect to 2008 Census poverty thresholds.
- marital_status - Marital status of respondent.
- rent_or_own - Housing situation of respondent.
- employment_status - Employment status of respondent.
- hhs_geo_region - Respondent's residence using a 10-region geographic classification defined by the U.S. Dept. of Health and Human Services. Values are represented as short random character strings.
- census_msa - Respondent's residence within metropolitan statistical areas (MSA) as defined by the U.S. Census.
- household_adults - Number of other adults in household, top-coded to 3.
- household_children - Number of children in household, top-coded to 3.
- employment_industry - Type of industry respondent is employed in. Values are represented as short random character strings.
- employment_occupation - Type of occupation of respondent. Values are represented as short random character strings.
- Logistic Regression
- K-Nearest Neighbor
- Decision Tree
- Bagged Tree
- Random Forest
The top three features determining the probability of an individual getting vaccinated for both viruses are based on opinion. For the H1N1 vaccine doctor reccomendation, the individual's opinion on whether they at risk of contracting the virus, and their opinion on the efficacy of the vaccine are the top three predictors. Similarly, seasonal flu vaccination rates are driven by the same features.
People's opinions often dictate their behavior as illustrated by this analysis. Although the distribution of each vaccine differed, the top three important features remained relatively the same (opion based columns regarding the H1N1 vaccine and opinion based columns regarding the seasonal vaccine respectively).