knitr::opts_knit$set(root.dir = 'C:/Users/jkempke/Box Sync/Side Projects/MCOD/Tidy Data')
knitr::opts_chunk$set(warning = F)
library(leaflet)
library(dplyr)
library(ggplot2)
library(knitr)The CDC WONDER data sets aggregate statistics from US death certificates. In the below descriptive analyses I start with the colloquial question: “are there states where there are higher incidences of sepsis mortality?”. While others have demosntrated sepsis mortality clusters in the US, one important method that was not accounted for was whether this reflected the elevated underlying all cause mortality or was specific to sepsis. Therefore in the below analyses I calculate mortality ratios of the age-adjusted cause specific mortality rate divided by the state’s all cause age-adjusted mortality rate to examine if this highlights where there may be excess sepsis mortality. Sepsis is ill-defined in ICD-10 coding for death certificates so I use alternative case definitions of septicemia only (A40-A41), influenza and pneumonia (J09-J18), and a combined definition.
In conjunction with analyses we published in Critical Care, these data can help to make decisions about which state’s HCUP SIDs to buy to examine the effect of SEP-1 and AHA variables on septicemia mortality. Teh HCUP 2017 SID catalog available for purchase is available here.
Data sets Construction
Select button on the right to expand code.
states.map <-
geojsonio::geojson_read("F:/Side Projects 2019/Medicare/data/json/gz_2010_us_040_00_5m.json", what = "sp")
sepsis2017 <- read.csv('MCOD 2017 Septicemia States.csv',
stringsAsFactors = T)
sepsis2017 <- sepsis2017 %>%
rename(
NAME = State,
sepsis.adjusted = Age.Adjusted.Rate,
sepsis.deaths = Deaths
) %>%
select(
NAME,
sepsis.adjusted,
sepsis.deaths)
all2017 <- read.csv("MCOD 2017 All Cause States.csv",
stringsAsFactors = T)
flupna2017 <- read.csv('MCOD 2017 flupna States.csv',
stringsAsFactors = T)
flupna2017 <- flupna2017 %>%
rename(
NAME = State,
flupna.adjusted = Age.Adjusted.Rate
) %>%
select(
NAME,
flupna.adjusted)
both2017 <- read.csv('MCOD 2017 Septicemia n Flupna States.csv',
stringsAsFactors = T)
both2017 <- both2017 %>%
rename(
NAME = State,
both.adjusted = Age.Adjusted.Rate
) %>%
select(
NAME,
both.adjusted)
all2017 <- all2017 %>%
rename(
NAME = State,
all.adjusted = Age.Adjusted.Rate
) %>%
select(
NAME,
all.adjusted)
mcod2017 <- left_join(both2017, all2017, by = 'NAME')
mcod2017 <- left_join(sepsis2017, mcod2017, by = 'NAME')
mcod2017 <- left_join(flupna2017, mcod2017, by = 'NAME')
mcod2017 <- mcod2017 %>%
mutate(
sepsis.adjusted.ratio = round(sepsis.adjusted/all.adjusted, digits = 2),
flupna.adjusted.ratio = round(flupna.adjusted/all.adjusted, digits = 2),
both.adjusted.ratio = round(both.adjusted/all.adjusted, digits = 2)
) %>%
arrange(flupna.adjusted.ratio)
states.map@data <- left_join(
states.map@data,
mcod2017,
by = "NAME")This is a state-level map of the Age-Adjusted Mortality Rates per100,000 persons in the year of 2017. This serves as a visualization of the denominator for each of the calculated ratios below.
pal0 <- colorNumeric(palette = "magma",
domain = states.map@data$all.adjusted)
leaflet(data=states.map)%>%
addProviderTiles(providers$Esri.WorldGrayCanvas)%>%
addPolygons(fillColor = ~pal0(all.adjusted),
color = "black",
weight = 1,
smoothFactor = 0.2,
fillOpacity = 0.75,
stroke = TRUE,
label = ~paste(NAME, all.adjusted, " per100K"),
highlight = highlightOptions(
weight = 5,
color="black",
bringToFront = TRUE))%>%
addLegend("bottomright",
pal = pal0,
values = ~all.adjusted,
title = "All Cause Age-Adj Deaths",
labFormat = labelFormat(suffix = " per100K"),
opacity = 1)%>%
setView(-90, 40, 4)These represent deaths with diagnostic codes for septicemia (A40-A41) among any of the listed causes of death. The first map is the age-adjusted death rates while the second map is the ratio of this age-adjusted septicemia death rate to the all-cause age-adjusted death rate o the state.
pal1 <- colorNumeric(palette = "magma",
domain = states.map@data$sepsis.adjusted)
leaflet(data=states.map)%>%
addProviderTiles(providers$Esri.WorldGrayCanvas)%>%
addPolygons(fillColor = ~pal1(sepsis.adjusted),
color = "black",
weight = 1,
smoothFactor = 0.2,
fillOpacity = 0.75,
stroke = TRUE,
label = ~paste(NAME, sepsis.adjusted, " per100K"),
highlight = highlightOptions(
weight = 5,
color="black",
bringToFront = TRUE))%>%
addLegend("bottomright",
pal = pal1,
values = ~sepsis.adjusted,
title = "Sepsis Age-Adj Deaths",
labFormat = labelFormat(suffix = " per100K"),
opacity = 1)%>%
setView(-90, 40, 4)pal1 <- colorNumeric(palette = "magma",
domain = states.map@data$sepsis.adjusted.ratio)
leaflet(data=states.map)%>%
addProviderTiles(providers$Esri.WorldGrayCanvas)%>%
addPolygons(fillColor = ~pal1(sepsis.adjusted.ratio),
color = "black",
weight = 1,
smoothFactor = 0.2,
fillOpacity = 0.75,
stroke = TRUE,
label = ~paste(NAME, "<br>ratio =", sepsis.adjusted.ratio),
highlight = highlightOptions(
weight = 5,
color="black",
bringToFront = TRUE))%>%
addLegend("bottomright",
pal = pal1,
values = ~sepsis.adjusted.ratio,
title = "Sepsis to All Cause Age-Adj Ratios",
opacity = 1)%>%
setView(-90, 40, 4)These represent deaths with diagnostic codes for influenza and pneumonia (J09-J18) among any of the listed causes of death. The first map is the age-adjusted death rates while the second map is the ratio of this age-adjusted influenza and pneumonia death rate to the all-cause age-adjusted death rate o the state.
pal2 <- colorNumeric(palette = "magma",
domain = states.map@data$flupna.adjusted)
leaflet(data=states.map)%>%
addProviderTiles(providers$Esri.WorldGrayCanvas)%>%
addPolygons(fillColor = ~pal2(flupna.adjusted),
color = "black",
weight = 1,
smoothFactor = 0.2,
fillOpacity = 0.75,
stroke = TRUE,
label = ~paste(NAME, "ratio =", flupna.adjusted),
highlight = highlightOptions(
weight = 5,
color="black",
bringToFront = TRUE))%>%
addLegend("bottomright",
pal = pal2,
values = ~flupna.adjusted,
title = "Flu & Pna Age-Adj Deaths",
labFormat = labelFormat(suffix = " per100K"),
opacity = 1)%>%
setView(-90, 40, 4)pal2 <- colorNumeric(palette = "magma",
domain = states.map@data$flupna.adjusted.ratio)
leaflet(data=states.map)%>%
addProviderTiles(providers$Esri.WorldGrayCanvas)%>%
addPolygons(fillColor = ~pal2(flupna.adjusted.ratio),
color = "black",
weight = 1,
smoothFactor = 0.2,
fillOpacity = 0.75,
stroke = TRUE,
label = ~paste(NAME, "ratio =", flupna.adjusted.ratio),
highlight = highlightOptions(
weight = 5,
color="black",
bringToFront = TRUE))%>%
addLegend("bottomright",
pal = pal2,
values = ~flupna.adjusted.ratio,
title = "Flu & Pna to All Cause Age-Adj Ratios",
opacity = 1)%>%
setView(-90, 40, 4)These represent deaths with diagnostic codes for septicemia (A40-A41) or influenza and pneumonia (J09-J18)among any of the listed causes of death. The first map is the age-adjusted death rates while the second map is the ratio of this age-adjusted death rate to the all-cause age-adjusted death rate o the state.
pal3 <- colorNumeric(palette = "magma",
domain = states.map@data$both.adjusted)
leaflet(data=states.map)%>%
addProviderTiles(providers$Esri.WorldGrayCanvas)%>%
addPolygons(fillColor = ~pal3(both.adjusted),
color = "black",
weight = 1,
smoothFactor = 0.2,
fillOpacity = 0.75,
stroke = TRUE,
label = ~paste(NAME, both.adjusted, " per100K"),
highlight = highlightOptions(
weight = 5,
color="black",
bringToFront = TRUE))%>%
addLegend("bottomright",
pal = pal3,
values = ~both.adjusted,
title = "Sepsis or Flu & Pna Deaths",
labFormat = labelFormat(suffix = " per100K"),
opacity = 1)%>%
setView(-90, 40, 4)pal3 <- colorNumeric(palette = "magma",
domain = states.map@data$both.adjusted.ratio)
leaflet(data=states.map)%>%
addProviderTiles(providers$Esri.WorldGrayCanvas)%>%
addPolygons(fillColor = ~pal3(both.adjusted.ratio),
color = "black",
weight = 1,
smoothFactor = 0.2,
fillOpacity = 0.75,
stroke = TRUE,
label = ~paste(NAME, "ratio =", both.adjusted.ratio),
highlight = highlightOptions(
weight = 5,
color="black",
bringToFront = TRUE))%>%
addLegend("bottomright",
pal = pal3,
values = ~both.adjusted.ratio,
title = "Sepsi or Flu & Pna to All Cause Age-Adj Ratios",
opacity = 1)%>%
setView(-90, 40, 4)sepsis2017 <- arrange(sepsis2017, desc(sepsis.deaths))
kable(sepsis2017, type="markdown", caption = "Total Number of Septicemia Deaths by State")| NAME | sepsis.adjusted | sepsis.deaths |
|---|---|---|
| California | 51.4 | 22301 |
| Texas | 66.1 | 17955 |
| Florida | 43.5 | 13351 |
| New York | 44.6 | 11190 |
| Pennsylvania | 50.8 | 8990 |
| Illinois | 52.3 | 7998 |
| Ohio | 53.0 | 7857 |
| North Carolina | 58.1 | 7032 |
| New Jersey | 57.4 | 6480 |
| Georgia | 59.7 | 6381 |
| Michigan | 48.3 | 6108 |
| Tennessee | 65.8 | 5252 |
| Virginia | 53.3 | 5135 |
| Indiana | 60.3 | 4702 |
| Kentucky | 83.7 | 4448 |
| South Carolina | 67.3 | 4141 |
| Massachusetts | 46.8 | 4081 |
| Washington | 48.5 | 4043 |
| Alabama | 65.7 | 3912 |
| Missouri | 48.3 | 3703 |
| Louisiana | 68.2 | 3593 |
| Maryland | 47.3 | 3341 |
| Arizona | 38.5 | 3311 |
| Oklahoma | 71.4 | 3267 |
| Wisconsin | 38.9 | 2861 |
| Mississippi | 78.3 | 2701 |
| Arkansas | 69.5 | 2557 |
| Colorado | 43.6 | 2537 |
| Connecticut | 50.5 | 2447 |
| Minnesota | 34.0 | 2295 |
| Nevada | 60.9 | 2023 |
| West Virginia | 77.1 | 1943 |
| Oregon | 36.0 | 1865 |
| Iowa | 45.1 | 1848 |
| Kansas | 52.2 | 1844 |
| New Mexico | 50.9 | 1270 |
| Utah | 45.2 | 1164 |
| Nebraska | 42.7 | 1004 |
| Idaho | 38.5 | 738 |
| Maine | 36.7 | 714 |
| New Hampshire | 38.7 | 693 |
| Rhode Island | 46.0 | 652 |
| Delaware | 48.5 | 621 |
| Hawaii | 31.6 | 599 |
| Montana | 41.6 | 571 |
| South Dakota | 48.1 | 515 |
| District of Columbia | 63.0 | 426 |
| North Dakota | 39.6 | 366 |
| Wyoming | 49.9 | 339 |
| Vermont | 30.2 | 264 |
| Alaska | 39.9 | 247 |