Problem Statement

There are many known health issues associated with air pollution, often measured by Particulate Matter (PM), a term for solid or liquid particles or droplets that are found in the air (such as dust, smog, or smoke). PM of size 2.5 micrometers or smaller are very fine inhalable particles that are particularly damaging to lung health*. Our team seeks to investigate the potential links between Air Quality Levels and Chronic Lower Respiratory Disease (CLRD) Mortality in California.

Currently, PM2.5 levels of under 12.0 are considered “good,” and under 35.4 are considered “moderate” according to the US Environmental Protection Agency. It is not until we reach levels over 35.5 that air quality is considered “unhealthy for sensitive groups,” and over 55.5 to be considered “unhealthy.” Our team believes that this threshold of “safe” PM2.5 levels should be drastically lowered in California. We will investigate aggregate statewide PM2.5 measurements and CLRD Mortality counts 1) at a zip code level to determine where further research should be conducted, and 2) examine a potential correlation between zip codes with critically high PM2.5 levels and CLRD deaths from 2009-2018.

Methods

Data Sources

Exposure data: This dataset comes from CalEnviroScreen3.0 dataset from the California Environmental Protection Agency (CalEPA). In January 2017, the Office of Environmental Health Hazard Assessment (OEHHA), on behalf of the California Environmental Protection Agency (CalEPA) released this dataset. It contains data from three columns, county, zip code and average PM2.5 per zip code. Outcome data: This dataset contains counts of deaths for California residents by ZIP Code based on information entered on death certificates from 2009-2018.

Cleaning and Creating New Variable

We first removed the NA from our exposure data and obtained the 75th percentile (11.83) of the average PM2.5 per zip code. We then grouped our exposure data by zip code and created a new binary column to indicate whether or not the zip has PM2.5 measurement in the critical category of PM2.5 (the 75th percentile). We created a subset of our data which only included zip codes in the critical category. Duplicated zip codes were eliminated from the dataset.

We grouped our outcome data by zip code and created a new column called total count containing the total number of CLRD deaths in each zip code from 2009 to 2018. We then filtered zip codes so that only those that are both in our exposure dataset and outcome dataset were retained. Zip codes that were not in the exposure dataset were not included in this new subset.

We combined our exposure data, zip codes and corresponding PM2.5 measurements in the critical category, and the outcome data, zip codes and corresponding CLRD deaths, into one final dataset. Columns of county, zip code, PM2.5 measurements and total number of CLRD deaths were retained. NA were removed from this final dataset. Zip codes were stored as factors. PM2.5 measurements were rounded to two decimal places.

Analytic Methods

A plot containing a bar chart and a line chart was created from our final dataset. The bar chart represents the total number of CLRD deaths in each zip code while the line chart represents the PM2.5 measurements in each zip code. Then linear regression was used to test the relationship between PM2.5 pollution levels and total Chronic Lower Respiratory Disease Mortality among zip codes in the critical category of PM2.5.

Results

Table 1 provides a list of all of the zip codes with critical values of PM2.5 in California, their air pollution level, and the number of deaths caused by Chronic Lower Respiratory Disease(CLRD) in these regions between the years 2009-2018. Table 2 collapses this information to demonstrate average PM2.5 and total CLRD mortality on the county level. Table 1 and Table 2 facilitate the identification of zip codes with the highest levels of PM2.5 pollution or the largest CLRD mortality counts. These tables can be posted on our nonprofit’s website to help inform our audience on this issue. Because further research on all of these zip codes would be costly, we recommend a targeted investigation of the zip codes that ranked among the top 25 most burdened for both PM2.5 pollution and Chronic Lower Respiratory Disease Mortality. These zip codes are “93307” (Kern), “93306” (Kern), “93304” (Kern), “93308” (Kern), “93309” (Kern), “93277” (Tulare), “93274” (Tulare), and “93230” (Kings).

table1
Table 1: Chronic Lower Respiratory Disease (CLRD) Mortality by California zip codes with critically high levels of Particulate Matter 2.5 (PM2.5), 2009-2018
Zip Code County Average PM2.5 CLDRM Counts
90008 Los Angeles 12.23 151
90043 Los Angeles 12.09 194
90044 Los Angeles 12.10 228
90047 Los Angeles 12.05 204
90220 Los Angeles 12.02 183
90247 Los Angeles 12.05 165
90250 Los Angeles 12.05 219
90280 Los Angeles 12.05 156
90503 Los Angeles 12.05 187
90640 Los Angeles 12.05 159
90650 Los Angeles 11.87 403
90706 Los Angeles 11.98 280
91701 San Bernardino 12.79 158
91710 San Bernardino 12.89 230
91711 Los Angeles 12.26 185
91730 San Bernardino 12.89 180
91762 San Bernardino 12.89 188
91786 San Bernardino 12.81 243
91911 San Diego 11.94 312
92335 San Bernardino 12.94 199
92373 San Bernardino 12.12 222
92376 San Bernardino 12.86 233
92404 San Bernardino 12.05 396
92410 San Bernardino 12.30 171
92503 Riverside 12.99 319
92504 Riverside 12.51 265
92505 Riverside 13.35 151
92509 Riverside 13.79 265
92570 Riverside 12.41 177
93230 Kings 16.89 295
93257 Tulare 15.67 369
93274 Tulare 17.05 301
93277 Tulare 17.08 278
93292 Tulare 16.54 145
93304 Kern 18.92 254
93305 Kern 18.88 153
93306 Kern 19.21 325
93307 Kern 19.31 260
93308 Kern 18.76 469
93309 Kern 18.76 331
93312 Kern 18.76 189
93612 Fresno 15.40 208
93705 Fresno 15.40 163
93727 Fresno 15.40 212
95204 San Joaquin 13.44 196
95205 San Joaquin 13.55 159
95207 San Joaquin 13.44 257
95336 San Joaquin 12.85 255
95340 Merced 12.89 213
95350 Stanislaus 12.89 372
95351 Stanislaus 12.89 202
95355 Stanislaus 12.89 322
95380 Stanislaus 12.89 170
Interpretation
Counties with the most number of zip codes with both critical levels
of PM2.5 and CLRD deaths include Los Angeles (13), San Bernadino (10), and
Kern (7) counties. On the other hand, zip codes that hold the highest death
counts in the 75th percentile belong to many counties: Kern, Los Angeles,
Stanislaus, Tulare, Riverside, San Diego, and Kings counties, with one Kern
zip code (93308) holding the highest death count of 469. Kern zip codes also
hold the top 7 highest PM2.5 levels ranging between 18.76-19.31.
Legend
CLRD mortality counts highlighted in red represent death counts in the
75th percentile. Only zip codes with critical levels of PM2.5 are included in
this table, and zip codes with no reported deaths between 2009-2018 were excluded. 
table2
Table 2: Chronic Lower Respiratory Disease (CLRD) Mortality by California counties with critically high levels of Particulate Matter 2.5 (PM2.5), 2009-2018
County Average PM2.5 CLRD Mortality Counts
Los Angeles 12.22 2714
San Bernardino 12.67 2220
Kern 17.57 1981
Riverside 12.97 1177
Tulare 15.56 1093
Stanislaus 12.81 1066
San Joaquin 12.95 867
Fresno 15.24 583
San Diego 13.16 312
Kings 15.86 295
Merced 12.91 213
Imperial 13.23 0
Madera 12.67 0
Orange 12.05 0
Interpretation
In accordance with the table above, Los Angeles, San Bernardino, and
Kern counties hold the highest death counts. Counties with the highest average
PM2.5 levels are Kern, Kings, Tulare, and Fresno (17.57, 15.86, 15.56, and D
15.24, respectively). Despite having critical levels of PM2.5, the counties
of Imperial, Madera, and Orange had no reported CLRD deaths between 2009-2018.
Legend
Zip code level PM2.5 and CLRD death data are averaged by county.
Only counties with critical levels of PM2.5 are included in this table, and
zip codes with no reported deaths between 2009-2018 were excluded.

Figure 1 visualizes average PM2.5 levels and CLRD mortality among zip codes with a critical value of PM2.5 pollution. To identify the extent of association between the two, we used linear regression to measure a correlation value, as seen in Figure 2. Our intuition was that we would see CLDR mortality increase as levels of air pollution increase. Our linear regression model found a correlation value of 0.27, which indicates a small association but statistically significant correlation (p < 0.5).

figure1

figure2
## `geom_smooth()` using formula 'y ~ x'

Discussion

According to the US Environmental Protection Agency, levels of PM2.5 between 12.1 and 35.4 are considered “moderate”. While the EPA states that these levels of air pollution can be classified as “safe”, we argue that any level of PM2.5 above 11.84 should be reclassified as “unhealthy” in the state of California. Our group believes that truly “safe” levels of PM2.5 are those which demonstrate zero correlation between air pollution and CLRD mortality. However, our analysis suggests that there is a statistically significant correlation between “moderate” levels of air pollution and CLRD mortality (Figure 2). This is further supported by Figure 1, which demonstrates elevated levels of CLDR mortality for all levels of PM2.5 pollution greater than 11.84. We believe further research in the zip codes most burdened by both PM2.5 and CLDR mortality will provide insight into this relationship. We believe this research will also give us information on other potential safety measures California may consider implementing, such as a minimum residential distance from high emission structures.