2010 - 2020 saw the population of citizens 65 years and older become the largest and fastest growing age group in any decade since 1880 - 1890.(US Census)
Due to this monumental impact, the health and well-being of the older population is of great interest to data scientists.
Alzheimer’s Disease is the most common form of dementia in older adults, and it is important to collect data on any risk factors or treatments.
As a personal home care provider to the elderly, I am a collector of similar data on a much smaller scale. It is enriching to work with such a robust data set. It comes from the Centers for Disease Control and Prevention.
Source of this data
BRFSS stands for the Behavioral Risk Factor Surveillance System, which “is the nation’s premier system of health-related cell phone and landline surveys that collect state data about U.S. residents regarding their health-related risk behaviors, chronic health conditions, and use of preventive services.” (CDC)
It is conducted by individual state health departments. Each row represents a response to a survey question.
Source: CDC
Data Exploration
This dataset features 39 topics related to the health of older adults.
[1] "Lifetime diagnosis of depression"
[2] "Recent activity limitations in past month"
[3] "Cholesterol checked in past 5 years"
[4] "Arthritis among older adults"
[5] "Frequent mental distress"
[6] "Colorectal cancer screening"
[7] "Self-rated health (good to excellent health)"
[8] "Influenza vaccine within past year"
[9] "Binge drinking within past 30 days"
[10] "Fall with injury within last year"
[11] "Current smoking"
[12] "Diabetes screening within past 3 years"
[13] "Obesity"
[14] "Physically unhealthy days (mean number of days)"
[15] "Eating 2 or more fruits daily"
[16] "Prevalence of sufficient sleep"
[17] "Up-to-date with recommended vaccines and screenings - Men"
[18] "No leisure-time physical activity within past month"
[19] "High blood pressure ever"
[20] "Mammogram within past 2 years"
[21] "Oral health: tooth retention"
[22] "Taking medication for high blood pressure"
[23] "Eating 3 or more vegetables daily"
[24] "Self-rated health (fair to poor health)"
[25] "Fair or poor health among older adults with arthritis"
[26] "Severe joint pain among older adults with arthritis"
[27] "Intensity of caregiving among older adults"
[28] "Duration of caregiving among older adults"
[29] "Expect to provide care for someone in the next two years"
[30] "Provide care for someone with cognitive impairment within the past month"
[31] "Disability status, including sensory or mobility limitations"
[32] "Provide care for a friend or family member in past month"
[33] "Ever had pneumococcal vaccine"
[34] "Pap test within past 3 years"
[35] "Functional difficulties associated with subjective cognitive decline or memory loss among older adults"
[36] "Subjective cognitive decline or memory loss among older adults"
[37] "Talked with health care professional about subjective cognitive decline or memory loss"
[38] "Up-to-date with recommended vaccines and screenings - Women"
[39] "Need assistance with day-to-day activities because of subjective cognitive decline or memory loss"
Oral Health
Oral health is an interesting indicator for several reasons.
Dental insurance is not covered by Medicare.
Poor oral health is often a precursor to poor health overall
Oral health has historically correlated with economic status
Needing Assistance
This dataset measures the ever-important need for assistance with daily living in old age with a survey question inquiring what proportion of the older adult population needs assistance with day-to-day activities because of subjective cognitive decline or memory loss.
Providing care
On the flip-side, older adults are still often caregivers themselves. The survey provides five separate measurements:
“Intensity of caregiving among older adults”
“Duration of caregiving among older adults”
“Expect to provide care for someone in the next two years”
“Provide care for someone with cognitive impairment within the past month”
“Provide care for a friend or family member in past month”
A Comprehensive Look at Older caregivers
Linear Regression Analysis
Linear Regression - Tooth Loss vs Leisure Time Physical Activity
Call:
lm(formula = tooth_loss ~ no_leisure, data = aging_means)
Residuals:
Min 1Q Median 3Q Max
-6.9853 -2.0934 0.1149 1.8140 9.6744
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -6.0106 3.5082 -1.713 0.0931 .
no_leisure 1.1135 0.1123 9.920 3.3e-13 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 3.539 on 48 degrees of freedom
Multiple R-squared: 0.6721, Adjusted R-squared: 0.6653
F-statistic: 98.4 on 1 and 48 DF, p-value: 3.296e-13
Model equation:
tooth loss= 1.1135(lack of leisure time physical activity) - 6.0106
For each additional percent of adults who did not engage in leisure time physical activity in the past month, there is a predicted increase of 1.1135% in adults who lose 6 or more teeth due to decay.
The p-value is very small, suggesting it is a meaningful variable to to explain the linear increase in tooth loss.
66.53% of the variation in the data is likely explained by this model.
Linear Regression - Tooth Loss vs Leisure Time Physical Activity:
with residuals
Multiple Regression and Residuals:
Source: National Archives
Tooth loss and Lack of Leisure Time Physical Activity
The most intriguing relationship continues to be the one between oral health and leisure time physical activity. Can anything be deduced if the data is grouped by how much dental coverage is provided by the state in the absence of national dental insurance?
Medicaid Adult Dental Benefits Coverage by State
The following data was obtained from webscraping the Center for Health Care Strategies Medicaid Adult Dental Benefits Coverage by State fact sheet.
It puts each state into a category summarizing the amount of dental coverage it provides to Medicaid recipients: none, emergency, limited, or extensive.
The size of each point corresponds to the percentage of older adults in each state that report having frequent mental distress.
This plot attempts to illustrate a correlation between adults losing 6 or more teeth and having no leisure time physical activity. Each point represents a different state.
This dataset is ripe for more analysis. It would be interesting to compare various statistics stratified by race, gender, even age subdivisions.
Although the dataset includes geolocation measures down to latitude and longitude, each set of coordinates merely corresponds to the central point of whatever state the value was recorded in. It would be fascinating to explore this data on a more granular level, perhaps looking at counties, zip codes, or census designated areas.