Analysis of Lifestyle Factors and Their Association with Physical Health

This short study delves into the relationships between selected lifestyle habits and self-reported days of poor physical health. Utilizing a dataset of 361,320 observations from the CDC Behavioral Risk Factor Surveillance System (BRFSS), I applied linear regression to uncover the associations of exercise, body mass index (BMI), smoking, and alcohol consumption with physical health.

Study Results Overview

The dependent variable, PHYSHLTH, was not normally distributed, so I decided to perform a bootstrapped regression model using the following code in R:

fit_model <- function(data, indices) {
fit <- lm(PHYSHLTH ~ EXERANY2 + BMI + SMOKE100 + ALCDAY4, data = data[indices, ])
return(coef(fit))
}
set.seed(5)
boot_results <- boot(data, fit_model, R=999)

This ran the linear models 999 times in order to get bootstrapped results to approximate a normal distribution.

The multiple linear regression model was significant at P<0.001, and an adjusted R-Squared value of 0.078.

  • Baseline Physical Health (Intercept): The model starts with an average of 1.74 days of poor health per month, establishing a baseline for individuals with default behavior (p < 0.001, 95% CI: [1.59, 1.89]).
  • Exercise (EXERANY2):
    • No: Not engaging in exercise was associated with a significant decrease of 4.61 days in reported poor health, reinforcing the positive impact of physical activity (p < 0.001).
    • Don’t know/Not sure: Uncertainty about exercise habits showed a non-significant association, with an increase of 4.15 days in reported poor health, at P<0.001 (95% CI: [3.11, 5.22]).
    • Refused: Refusing to disclose exercise habits also resulted in a non-significant increase of 4.04 days in reported poor health (P < 0.001, 95% CI: [2.10, 6.07]).
  • Body Mass Index (BMI):
    • A unit increase in BMI was associated with a 0.10-day increase in reported poor health days, highlighting a significant correlation between higher BMI and perceived poor health (P < 0.001, 95% CI: [0.09, 0.10]).
  • Smoking (SMOKE100):
    • No: Not smoking was associated with an decrease of 1.75 days in reported poor health, indicating a significant negative impact on perceived health (P < 0.001, 95% CI: [-1.81, -1.69]).
    • Don’t know/Not sure: Uncertainty about smoking habits ad a strong negative correlation with days of poor health, decreasing days of poor physical health by 1.10 (P<0.001, 95% CI: [-1.49, -0.70])
    • Refused: Those who refused to disclose their smoking status showed a non-significant decrease of 1.22 days in reported poor health (P = 0.05, 95% CI: [-2.43, 0.07]).
  • Alcohol Consumption (ALCDAY4):
    • The number of days drinking alcohol showed a surprising, slight but significant negative association, with a decrease of 0.05 days in reported poor health for each additional day of drinking (P< 0.001, 95% CI: [-0.06, -0.05]).
Coefficient Estimate P.value Lower.CI Upper.CI
(Intercept) (Intercept) 1.7407989 0.0000000 1.5896479 1.8922662
EXERANY2No EXERANY2No 4.6075077 0.0000000 4.5232760 4.6928286
EXERANY2Don’t know/Not sure EXERANY2Don’t know/Not sure 4.1513608 0.0000000 3.1137402 5.2218730
EXERANY2Refused EXERANY2Refused 4.0441483 0.0000313 2.0991206 6.0727065
BMI BMI 0.0983551 0.0000000 0.0933678 0.1034138
SMOKE100No SMOKE100No -1.7478107 0.0000000 -1.8083567 -1.6900012
SMOKE100Don’t know/Not sure SMOKE100Don’t know/Not sure -1.1022758 0.0000000 -1.4905564 -0.6982676
SMOKE100Refused SMOKE100Refused -1.2150366 0.0573350 -2.4345447 0.0665713
ALCDAY4 ALCDAY4 -0.0523532 0.0000000 -0.0559247 -0.0483179

Interpretation of Findings

The study’s findings illuminate the complex interplay between lifestyle choices and health perceptions:

Regular exercise significantly correlates with fewer self-reported days of poor physical health, underscoring its importance in maintaining good health. Accordingly, the positive association between BMI and poor health days suggests that higher body weight may be detrimental to health, although it is entirely possible to have a high BMI while being physically fit. Lastly, it is not surprising that, consistent with existing literature, smoking is associated with an increased number of days feeling physically unwell.

Lastly, the relationship between alcohol consumption and health perception is complex. According to these results, moderate drinking appears to be slightly associated with better health perceptions. There could be several explanations for this. One explanation could be that alcohol consumption often occurs in social settings, which can lead to psychological well-being and increased relaxation. Individuals who frequently engage in such activities may report fewer days of poor physical health. Additionally, people who consume alcohol may belong to socioeconomic groups with greater access to healthcare, or due to the “sick quitter” effect, where drinkers report fewer health issues due to their broader lifestyle and health management practices.

Statistical Significance and Model Strength

The statistical significance of the predictors, indicates robust associations.

The model’s adjusted R-squared value of 0.078 suggests that while the included variables explain a portion of the variability in reported poor health days, a substantial amount of variance remains unexplained, highlighting the complexity of health outcomes and the potential influence of unmeasured factors.


Conclusion

This analysis offers valuable insights into how certain lifestyle behaviors are associated with individuals’ perceptions of their physical health. The significant associations found align with the broader public health understanding that these factors are crucial to health outcomes. The nuanced relationship with alcohol suggests that its impact on health perception may vary with consumption patterns. These findings contribute to the ongoing dialogue on public health and underscore the importance of promoting healthy lifestyle choices to enhance overall well-being.

Similar Posts

Leave a Reply