Introduction

For the past few assignments, I have been working with the Australian Marriage Law dataset provided in the class data folder. For homework 6, I have decided to branch out and work with a new dataset, which I will use for my final project.

Data

The data I will be working with comes from the Substance Abuse and Mental Health Services Administration (SAMHSA), which is an agency within the US Department of Health and Human Services. The agency’s mission is to “reduce the impact of substance abuse and mental illness on America’s communities”. [https://www.samhsa.gov/]

The data comes from the National Survey on Drug Use and Health (NSDUH). The collection of detailed tables within the report contains estimates on the number of people with specified characteristics, like substance abuse or mental illness. Data is collected by field staff through the administration of in-person surveys, which moved remote in 2020 due to COVID-19 (will discuss in later section of final project potential implications of change in methodology due to pandemic). Sample design and sample size are sufficient to allow for state and national estimates. The sample includes United States non-institutionalized residents aged 12 or older, including residents of households, shelters, dormitories, and boarding houses. Individuals that were excluded from the sample include those with no fixed address (but not utilizing shelters), active-duty military, and residents in institutional living (correctional facilities, nursing homes, mental institutions, and long-term care facilities. (In the final project I will discuss the implications of excluding these segments of the population in the study.)

The table I have gathered data from:

Table 8.18A – Received Mental Health Services in Past Year: Among People Aged 18 or Older; by Past Year Level of Mental Illness and Geographic and Socioeconomic Characteristics, Numbers in Thousands, 2019 and 2020

This data includes the number of people (in thousands) who had received mental health services in the past year. The years included in the data are 2019 and 2020. The counts are itemized by those with any mental illness, serious mental illness, or no mental illness. The geographic and socioeconomic characteristics included in the data are geographic region, county type, poverty level, education level, and type of health insurance. This research defined mental health services to include inpatient treatment, outpatient treatment (ie. counseling), and/or use of prescription medication for issues related to emotions, nerves, or mental health.

Variables of Interest

Presence/Classification of Mental Illness
- This variable represents the presence of mental illness in individuals and measures included any mental illness, serious mental illness, or no mental illness
- Measure of the presence of mental illness was based on DSM-IV criteria. Serious mental illness is a category within any mental illness, classified as mental illness resulting in “serious functional impairment”
- I will focus on the measures of either no mental illness or any mental illness when analyzing the data
Geographic Region
- This variable is defined by the US Census Bureau and includes the Northeast Region, the Midwest Region, the South Region, and the West Region. ++ Northeast Region: the New England Division (Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, Vermont) and the Middle Atlantic Division (New Jersey, New York, Pennsylvania) ++ Midwest Region: the East North Central Division (Illinois, Indiana, Michigan, Ohio, Wisconsin) and the West North Central Division (Iowa, Kansas, Minnesota, Missouri, Nebraska, North Dakota, South Dakota) ++ South Region: the South Atlantic Division (Delaware, District of Columbia, Florida, Georgia, Maryland, North Carolina, South Carolina, Virginia, West Virginia), the East South Central Division (Alabama, Kentucky, Mississippi, Tennessee), and the West South Central Division (Arkansas, Louisiana, Oklahoma, Texas) ++ West Region: the Mountain Division (Arizona, Colorado, Idaho, Montana, Nevada, New Mexico, Utah, Wyoming) and the Pacific Division (Alaska, California, Hawaii, Oregon, Washington)
Poverty Level
- This variable was determined based on an individual’s total family income being compared with the US Census Bureau’s definition of poverty and is classified as less than 100% (total family income is lower than poverty threshold), 100% to 199% (total family income is at or above poverty threshold but less than twice the threshold), and 200% or more (total family income is twice poverty threshold or greater)
Health Insurance
- This variable reveals the type of health insurance each respondent had
- The measures were Private insurance, Medicaid/CHIP, Other (Medicare, military-related health care, or other types), No Coverage
- Categories are not mutually exclusive, meaning respondents could report multiple types of health insurance

Read in & Tidy Data

Read in data:

menthealth <- read_excel("~/DACSS601_R/Data/US_mentalhealth.xlsx", skip = 1)

## New names:
## * `Total (2019)` -> `Total (2019)...2`
## * `Total (2020)` -> `Total (2020)...3`
## * `Total (2019)` -> `Total (2019)...10`
## * `Total (2020)` -> `Total (2020)...11`

head(menthealth)

## # A tibble: 6 x 15
##   `Geographic/Socioecono~ `Total (2019)...~ `Total (2020)...~ `Any Mental Illne~
##   <chr>                               <dbl>             <dbl>              <dbl>
## 1 Total                               40154             42420              22953
## 2 Geographic Region                      NA                NA                 NA
## 3 Northeast                            7503              7533               4048
## 4 Midwest                              9740             10056               5682
## 5 South                               13503             15234               7650
## 6 West                                 9407              9597               5573
## # ... with 11 more variables: Any Mental Illness (2020) <dbl>,
## #   Serious Mental Illness (2019) <dbl>, Serious Mental Illness (2020) <chr>,
## #   No Mental Illness (2019) <dbl>, No Mental Illness (2020) <dbl>,
## #   Total (2019)...10 <dbl>, Total (2020)...11 <dbl>,
## #   Received Services (2019) <dbl>, Received Services (2020) <dbl>,
## #   Services Not Received (2019) <dbl>, Services Not Received (2020) <dbl>

Remove rows containing section labels geographic region, county type, poverty level, education level, and health insurance:

menthealth <- menthealth %>%
  filter(!str_detect(`Geographic/Socioeconomic Characteristics`, "Geographic Region"))%>%
  filter(!str_detect(`Geographic/Socioeconomic Characteristics`, "County Type"))%>%
  filter(!str_detect(`Geographic/Socioeconomic Characteristics`, "Poverty Level"))%>%
  filter(!str_detect(`Geographic/Socioeconomic Characteristics`, "Education Level"))%>%
  filter(!str_detect(`Geographic/Socioeconomic Characteristics`, "Health Insurance"))

Select columns that I want to keep:

colnames(menthealth)

##  [1] "Geographic/Socioeconomic Characteristics"
##  [2] "Total (2019)...2"                        
##  [3] "Total (2020)...3"                        
##  [4] "Any Mental Illness (2019)"               
##  [5] "Any Mental Illness (2020)"               
##  [6] "Serious Mental Illness (2019)"           
##  [7] "Serious Mental Illness (2020)"           
##  [8] "No Mental Illness (2019)"                
##  [9] "No Mental Illness (2020)"                
## [10] "Total (2019)...10"                       
## [11] "Total (2020)...11"                       
## [12] "Received Services (2019)"                
## [13] "Received Services (2020)"                
## [14] "Services Not Received (2019)"            
## [15] "Services Not Received (2020)"

recservices <- menthealth %>%
  select("Geographic/Socioeconomic Characteristics", "Total (2019)...2", "Any Mental Illness (2019)", "Serious Mental Illness (2019)", "No Mental Illness (2019)")%>%
  rename(Geographic_Socioeconomic_Characteristics="Geographic/Socioeconomic Characteristics", Total="Total (2019)...2", Any_Mental_Illness="Any Mental Illness (2019)", Serious_Mental_Illness="Serious Mental Illness (2019)", No_Mental_Illness="No Mental Illness (2019)")

Select values within Geographic/Socioeconomic Characteristics that I want to keep and rename them for clarity:

recservices <- recservices %>%
  filter(Geographic_Socioeconomic_Characteristics=="Northeast"|
           Geographic_Socioeconomic_Characteristics=="Midwest"|
           Geographic_Socioeconomic_Characteristics=="South"|
           Geographic_Socioeconomic_Characteristics=="West"|
           Geographic_Socioeconomic_Characteristics=="Less than 100%"|
           Geographic_Socioeconomic_Characteristics=="100% - 199%"|
           Geographic_Socioeconomic_Characteristics=="200% or More"|
           Geographic_Socioeconomic_Characteristics=="Private"|
           Geographic_Socioeconomic_Characteristics=="Medicaid/CHIP"|
           Geographic_Socioeconomic_Characteristics=="Other"|
           Geographic_Socioeconomic_Characteristics=="No Coverage")%>%
  mutate(Geographic_Socioeconomic_Characteristics = case_when(
    Geographic_Socioeconomic_Characteristics == "Northeast" ~ "Northeast Region",
    Geographic_Socioeconomic_Characteristics == "Midwest" ~ "Midwest Region",
    Geographic_Socioeconomic_Characteristics == "South" ~ "South Region",
    Geographic_Socioeconomic_Characteristics == "West" ~ "West Region",
    Geographic_Socioeconomic_Characteristics == "Less than 100%" ~ "Poverty Level less than 100%",
    Geographic_Socioeconomic_Characteristics == "100% - 199%" ~ "Poverty Level 100%-199%",
    Geographic_Socioeconomic_Characteristics == "200% or More" ~ "Poverty Level more than 200%",
    Geographic_Socioeconomic_Characteristics == "Private" ~ "Private Insurance",
    Geographic_Socioeconomic_Characteristics == "Medicaid/CHIP" ~ "Medicaid/CHIP Insurance",
    Geographic_Socioeconomic_Characteristics == "Other" ~ "Other Insurance",
    Geographic_Socioeconomic_Characteristics == "No Coverage" ~ "No Insurance Coverage"
  ))

Display tidied data in a table:

kable(recservices, col.names = c("Geographic/Socioeconomic Characteristics", "Total", "Any Mental Illness", "Serious Mental Illness", "No Mental Illness"), 
      align = c('c', 'c', 'c', 'c', 'c')) %>%
  add_header_above(c("Table 1: Characteristics of Individuals in the US who Received Mental Health Services in 2019 (in Thousands)"=5))%>%
    kable_styling(fixed_thead = TRUE)%>%
  scroll_box(width = "100%", height = "100%")

Table 1: Characteristics of Individuals in the US who Received Mental Health Services in 2019 (in Thousands)
Geographic/Socioeconomic Characteristics	Total	Any Mental Illness	Serious Mental Illness	No Mental Illness
Northeast Region	7503	4048	1406	3455
Midwest Region	9740	5682	2032	4058
South Region	13503	7650	2902	5853
West Region	9407	5573	2235	3835
Poverty Level less than 100%	5578	3860	1692	1718
Poverty Level 100%-199%	7403	4762	2000	2641
Poverty Level more than 200%	26957	14184	4847	12773
Private Insurance	26711	14061	4725	12650
Medicaid/CHIP Insurance	8131	5553	2540	2578
Other Insurance	10935	5633	2053	5301
No Insurance Coverage	2048	1419	665	629

The Geographic/Socioeconomic Characteristics column should be split up and have a column for each variable (poverty level, insurance type, and geographic region). However, since these variables are unrelated to each other, they cannot be made into different columns of the same table. I will make tables for each variable if needed throughout the research.

Variable 1: Geographic Region

I will make tables for each of the 3 geographic/socioeconomic variables I am interested in working with.

georegion <- recservices %>%
  filter(Geographic_Socioeconomic_Characteristics=="Northeast Region"|
         Geographic_Socioeconomic_Characteristics=="Midwest Region"|
         Geographic_Socioeconomic_Characteristics=="South Region"|
         Geographic_Socioeconomic_Characteristics=="West Region")%>%
  select(Geographic_Socioeconomic_Characteristics, Total, Any_Mental_Illness, Serious_Mental_Illness, No_Mental_Illness)%>%
  rename(Geographic_Region=Geographic_Socioeconomic_Characteristics)

georegion <- georegion[, c("Geographic_Region", "Any_Mental_Illness", "Serious_Mental_Illness", "No_Mental_Illness", "Total")]

kable(georegion, col.names = c("Geographic Region", "Any Mental Illness", "Serious Mental Illness", "No Mental Illness", "Total"), 
      align = c('c', 'c', 'c', 'c', 'c')) %>%
  add_header_above(c("Table 2: Individuals in the US who Received Mental Health Services in 2019 by Geographic Region (in Thousands)"=5))%>%
    kable_styling(fixed_thead = TRUE)%>%
  scroll_box(width = "100%", height = "100%")

Table 2: Individuals in the US who Received Mental Health Services in 2019 by Geographic Region (in Thousands)
Geographic Region	Any Mental Illness	Serious Mental Illness	No Mental Illness	Total
Northeast Region	4048	1406	3455	7503
Midwest Region	5682	2032	4058	9740
South Region	7650	2902	5853	13503
West Region	5573	2235	3835	9407

georegion1 <- georegion %>%
  pivot_longer(c(Any_Mental_Illness, Serious_Mental_Illness, No_Mental_Illness), names_to = "Mental_Illness_Type", values_to = "Number_Treated")%>%
  select(Geographic_Region, Mental_Illness_Type, Number_Treated)

ggplot(data = georegion1) +
  geom_bar(mapping = aes(x = Mental_Illness_Type, y = Number_Treated, fill = Geographic_Region), stat = "identity", position = "dodge") +
  labs(title = "Graph 1: Number of Individuals who Received Treatment in 2019", 
       subtitle = "By Geographic Region", x = "Type of Mental Illness", y = "Number of People Treated (in Thousands)")

Variable 2: Poverty Level

povertylvl <- recservices %>%
  filter(Geographic_Socioeconomic_Characteristics=="Poverty Level less than 100%"|
         Geographic_Socioeconomic_Characteristics=="Poverty Level 100%-199%"|
         Geographic_Socioeconomic_Characteristics=="Poverty Level more than 200%")%>%
  select(Geographic_Socioeconomic_Characteristics, Total, Any_Mental_Illness, Serious_Mental_Illness, No_Mental_Illness)%>%
  rename(Poverty_Level=Geographic_Socioeconomic_Characteristics)

povertylvl <- povertylvl[, c("Poverty_Level", "Any_Mental_Illness", "Serious_Mental_Illness", "No_Mental_Illness", "Total")]

kable(povertylvl, col.names = c("Poverty Level", "Any Mental Illness", "Serious Mental Illness", "No Mental Illness", "Total"), 
      align = c('c', 'c', 'c', 'c', 'c')) %>%
  add_header_above(c("Table 3: Individuals in the US who Received Mental Health Services in 2019 by Poverty Level (in Thousands)"=5))%>%
    kable_styling(fixed_thead = TRUE)%>%
  scroll_box(width = "100%", height = "100%")

Table 3: Individuals in the US who Received Mental Health Services in 2019 by Poverty Level (in Thousands)
Poverty Level	Any Mental Illness	Serious Mental Illness	No Mental Illness	Total
Poverty Level less than 100%	3860	1692	1718	5578
Poverty Level 100%-199%	4762	2000	2641	7403
Poverty Level more than 200%	14184	4847	12773	26957

povertylvl1 <- povertylvl %>%
  pivot_longer(c(Any_Mental_Illness, Serious_Mental_Illness, No_Mental_Illness), names_to = "Mental_Illness_Type", values_to = "Number_Treated")%>%
  select(Poverty_Level, Mental_Illness_Type, Number_Treated)

ggplot(data = povertylvl1) +
  geom_bar(mapping = aes(x = Mental_Illness_Type, y = Number_Treated, fill = Poverty_Level), stat = "identity", position = "dodge") +
  labs(title = "Graph 2: Number of Individuals who Received Treatment in 2019", 
       subtitle = "By Poverty Level", x = "Type of Mental Illness", y = "Number of People Treated (in Thousands)")

Variable 3: Type of Insurance

insurance <- recservices %>%
  filter(Geographic_Socioeconomic_Characteristics=="Private Insurance"|
         Geographic_Socioeconomic_Characteristics=="Medicaid/CHIP Insurance"|
         Geographic_Socioeconomic_Characteristics=="Other Insurance"|
         Geographic_Socioeconomic_Characteristics=="No Insurance Coverage")%>%
  select(Geographic_Socioeconomic_Characteristics, Total, Any_Mental_Illness, Serious_Mental_Illness, No_Mental_Illness)%>%
  rename(Insurance_Type=Geographic_Socioeconomic_Characteristics)

insurance <- insurance[, c("Insurance_Type", "Any_Mental_Illness", "Serious_Mental_Illness", "No_Mental_Illness", "Total")]

kable(insurance, col.names = c("Insurance Type", "Any Mental Illness", "Serious Mental Illness", "No Mental Illness", "Total"), 
      align = c('c', 'c', 'c', 'c', 'c')) %>%
  add_header_above(c("Table 4: Individuals in the US who Received Mental Health Services in 2019 by Insurance Type (in Thousands)"=5))%>%
    kable_styling(fixed_thead = TRUE)%>%
  scroll_box(width = "100%", height = "100%")

Table 4: Individuals in the US who Received Mental Health Services in 2019 by Insurance Type (in Thousands)
Insurance Type	Any Mental Illness	Serious Mental Illness	No Mental Illness	Total
Private Insurance	14061	4725	12650	26711
Medicaid/CHIP Insurance	5553	2540	2578	8131
Other Insurance	5633	2053	5301	10935
No Insurance Coverage	1419	665	629	2048

insurance1 <- insurance %>%
  pivot_longer(c(Any_Mental_Illness, Serious_Mental_Illness, No_Mental_Illness), names_to = "Mental_Illness_Type", values_to = "Number_Treated")%>%
  select(Insurance_Type, Mental_Illness_Type, Number_Treated)

ggplot(data = insurance1) +
  geom_bar(mapping = aes(x = Mental_Illness_Type, y = Number_Treated, fill = Insurance_Type), stat = "identity", position = "dodge") +
  labs(title = "Graph 3: Number of Individuals who Received Treatment in 2019", 
       subtitle = "By Insurance Type", x = "Type of Mental Illness", y = "Number of People Treated (in Thousands)")

Description of Relationship between Poverty Level and Receiving Treatment

After viewing Graph 2, I am interested in further exploring the relationship between poverty level and receiving mental health services. I will use the variables: total number of individuals who received treatment and poverty level.

povertytx <- povertylvl %>%
  select(Poverty_Level, Total)%>%
  pivot_wider(names_from = Poverty_Level, values_from = Total)

kable(povertytx, col.names = c("Less than 100%", "100% - 199%", "More than 200%"), 
      align = c('c', 'c', 'c')) %>%
  add_header_above(c("Table 5: Total Number of Individuals who Received Treatment by Poverty Level (in Thousands)"=3))%>%
    kable_styling(fixed_thead = TRUE)%>%
  scroll_box(width = "100%", height = "100%")

Table 5: Total Number of Individuals who Received Treatment by Poverty Level (in Thousands)
Less than 100%	100% - 199%	More than 200%
5578	7403	26957

Question:

Is poverty level independent from receiving mental health treatment?

Hypothesis:

Null: Poverty level and receiving treatment are independent of one another.
Alternative: Poverty level and receiving treatment are not independent of each other.

Chi Square Test of Independence:

chisq.test(povertytx)

## 
##  Chi-squared test for given probabilities
## 
## data:  povertytx
## X-squared = 21101, df = 2, p-value < 2.2e-16

Since the p-value is less than 0.05, the null hypothesis is rejected.

Disclaimers: I believe I used the incorrect statistical model to evaluate a potential relationship between the variables. I will continue to explore this in my final project. Also, I will need to look at the proportion of individuals below the poverty threshold and above it rather than just counts of those receiving treatment, as the lack of proportion may be skewing results.

Citations

Center for Behavioral Health Statistics and Quality. (2021). Results from the 2020 National Survey on Drug Use and Health: Detailed tables. Rockville, MD: Substance Abuse and Mental Health Services Administration. Retrieved from https://www.samhsa.gov/data/
Center for Behavioral Health Statistics and Quality. (2021). 2020 National Survey on Drug Use and Health (NSDUH): Methodological summary and definitions. Rockville, MD: Substance Abuse and Mental Health Services Administration. Retrieved from https://www.samhsa.gov/data/

HW 6: Bivariate Relationships, Modeling, and Functions

Megan Georges

12/13/2021