Predicting Factors that Lead to Drug and Alcohol Abuse
Akwellé (Q) Quaye, Andre Williams, Makayla Moore
2022-07-22
Introduction
Substance abuse is a very complicated issue. There are so many factors that lead to substance abuse. We looked at three different questions, two in which drug and alcohol abuse are the response variables, and one in which drug and alcohol abuse are used as predictors for another response variables. These three questions are detailed below.
Predictors of Drug Overdose
In 2021, there was a 28.5% increase in drug overdoses in the United States from the previous year (CDC). Which places in the United States had the highest rates of drug overdose deaths, and what are some characteristics of these places? We wanted to look at health behaviors and quality of life factors as well as socioeconomic factors for this section.
Drug Abuse and Teen Births
There are a variety of factors that can lead to teen births. Is excessive drinking and drug use one of these factors? We wanted to look at whether or not drug and alcohol abuse is correlated with teen births, or if other factors not related to excessive drinking and drug are better predictors than teen births alone.
Data
In order to answer our questions, we looked at a variety of datasets.
County Health Rankings Dataset
The County Health Rankings dataset, collected by the University of Wisconsin Population Health Institute, ranks every county in a given state on their Health Outcomes and Health Factors. This dataset also contains the measurements used to calculate the rankings for each county. We primarily focused on the measurements used to calculate the rankings, a table with 3,193 observations of 249 different variables.
| FIPS | State | County | Unreliable | Deaths | Years of Potential Life Lost Rate | 95% CI - Low | 95% CI - High | Quartile | YPLL Rate (AIAN) | YPLL Rate (AIAN) 95% CI - Low | YPLL Rate (AIAN) 95% CI - High | YPLL Rate (AIAN) Unreliable | YPLL Rate (Asian) | YPLL Rate (Asian) 95% CI - Low | YPLL Rate (Asian) 95% CI - High | YPLL Rate (Asian) Unreliable | YPLL Rate (Black) | YPLL Rate (Black) 95% CI - Low | YPLL Rate (Black) 95% CI - High | YPLL Rate (Black) Unreliable | YPLL Rate (Hispanic) | YPLL Rate (Hispanic) 95% CI - Low | YPLL Rate (Hispanic) 95% CI - High | YPLL Rate (Hispanic) Unreliable | YPLL Rate (white) | YPLL Rate (white) 95% CI - Low | YPLL Rate (white) 95% CI - High | YPLL Rate (white) Unreliable | % Fair or Poor Health | 95% CI - Low_1 | 95% CI - High_1 | Quartile_1 | Average Number of Physically Unhealthy Days | 95% CI - Low_2 | 95% CI - High_2 | Quartile_2 | Average Number of Mentally Unhealthy Days | 95% CI - Low_3 | 95% CI - High_3 | Quartile_3 | Unreliable_1 | % Low birthweight | 95% CI - Low_4 | 95% CI - High_4 | Quartile_4 | % LBW (AIAN) | % LBW (AIAN) 95% CI - Low | % LBW (AIAN) 95% CI - High | % LBW (Asian) | % LBW (Asian) 95% CI - Low | % LBW (Asian) 95% CI - High | % LBW (Black) | % LBW (Black) 95% CI - Low | % LBW (Black) 95% CI - High | % LBW (Hispanic) | % LBW (Hispanic) 95% CI - Low | % LBW (Hispanic) 95% CI - High | % LBW (white) | % LBW (white) 95% CI - Low | % LBW (white) 95% CI - High | % Smokers | 95% CI - Low_5 | 95% CI - High_5 | Quartile_5 | % Adults with Obesity | 95% CI - Low_6 | 95% CI - High_6 | Quartile_6 | Food Environment Index | Quartile_7 | % Physically Inactive | 95% CI - Low_7 | 95% CI - High_7 | Quartile_8 | % With Access to Exercise Opportunities | Quartile_9 | % Excessive Drinking | 95% CI - Low_8 | 95% CI - High_8 | Quartile_10 | # Alcohol-Impaired Driving Deaths | # Driving Deaths | % Driving Deaths with Alcohol Involvement | 95% CI - Low_9 | 95% CI - High_9 | Quartile_11 | # Chlamydia Cases | Chlamydia Rate | Quartile_12 | Teen Birth Rate | 95% CI - Low_10 | 95% CI - High_10 | Quartile_13 | Teen Birth Rate (AIAN) | Teen Birth Rate (AIAN) 95% CI - Low | Teen Birth Rate (AIAN) 95% CI - High | Teen Birth Rate (Asian) | Teen Birth Rate (Asian) 95% CI - Low | Teen Birth Rate (Asian) 95% CI - High | Teen Birth Rate (Black) | Teen Birth Rate (Black) 95% CI - Low | Teen Birth Rate (Black) 95% CI - High | Teen Birth Rate (Hispanic) | Teen Birth Rate (Hispanic) 95% CI - Low | Teen Birth Rate (Hispanic) 95% CI - High | Teen Birth Rate (white) | Teen Birth Rate (white) 95% CI - Low | Teen Birth Rate (white) 95% CI - High | # Uninsured | % Uninsured | 95% CI - Low_11 | 95% CI - High_11 | Quartile_14 | # Primary Care Physicians | Primary Care Physicians Rate | Primary Care Physicians Ratio | Quartile_15 | # Dentists | Dentist Rate | Dentist Ratio | Quartile_16 | # Mental Health Providers | Mental Health Provider Rate | Mental Health Provider Ratio | Quartile_17 | Preventable Hospitalization Rate | Quartile_18 | Preventable Hosp. Rate (AIAN) | Preventable Hosp. Rate (Asian) | Preventable Hosp. Rate (Black) | Preventable Hosp. Rate (Hispanic) | Preventable Hosp. Rate (white) | % With Annual Mammogram | Quartile_19 | % Screened (AIAN) | % Screened (Asian) | % Screened (Black) | % Screened (Hispanic) | % Screened (white) | % Vaccinated | Quartile_20 | % Vaccinated (AIAN) | % Vaccinated (Asian) | % Vaccinated (Black) | % Vaccinated (Hispanic) | % Vaccinated (white) | # Completed High School | Population | % Completed High School | 95% CI - Low_12 | 95% CI - High_12 | Quartile_21 | # Some College | Population_1 | % Some College | 95% CI - Low_13 | 95% CI - High_13 | Quartile_22 | # Unemployed | Labor Force | % Unemployed | Quartile_23 | % Children in Poverty | 95% CI - Low_14 | 95% CI - High_14 | Quartile_24 | % Children in Poverty (AIAN) | % Children in Poverty (Asian) | % Children in Poverty (Black) | % Children in Poverty (Hispanic) | % Children in Poverty (white) | 80th Percentile Income | 20th Percentile Income | Income Ratio | Quartile_25 | # Children in Single-Parent Households | # Children in Households | % Children in Single-Parent Households | 95% CI - Low_15 | 95% CI - High_15 | Quartile_26 | # Associations | Social Association Rate | Quartile_27 | Annual Average Violent Crimes | Violent Crime Rate | Quartile_28 | # Injury Deaths | Injury Death Rate | 95% CI - Low_16 | 95% CI - High_16 | Quartile_29 | Injury Death Rate (AIAN) | Injury Death Rate (AIAN) 95% CI - Low | Injury Death Rate (AIAN) 95% CI - High | Injury Death Rate (Asian) | Injury Death Rate (Asian) 95% CI - Low | Injury Death Rate (Asian) 95% CI - High | Injury Death Rate (Black) | Injury Death Rate (Black) 95% CI - Low | Injury Death Rate (Black) 95% CI - High | Injury Death Rate (Hispanic) | Injury Death Rate (Hispanic) 95% CI - Low | Injury Death Rate (Hispanic) 95% CI - High | Injury Death Rate (white) | Injury Death Rate (white) 95% CI - Low | Injury Death Rate (white) 95% CI - High | Average Daily PM2.5 | Quartile_30 | Presence of Water Violation | Quartile_31 | % Severe Housing Problems | 95% CI - Low_17 | 95% CI - High_17 | Severe Housing Cost Burden | Severe Housing Cost Burden 95% CI - Low | Severe Housing Cost Burden 95% CI - High | Overcrowding | Overcrowding 95% CI - Low | Overcrowding 95% CI - High | Inadequate Facilities | Inadequate Facilities 95% CI - Low | Inadequate Facilities 95% CI - High | Quartile_32 | % Drive Alone to Work | 95% CI - Low_18 | 95% CI - High_18 | Quartile_33 | % Drive Alone (AIAN) | % Drive Alone (AIAN) 95% CI - Low | % Drive Alone (AIAN) 95% CI - High | % Drive Alone (Asian) | % Drive Alone (Asian) 95% CI - Low | % Drive Alone (Asian) 95% CI - High | % Drive Alone (Black) | % Drive Alone (Black) 95% CI - Low | % Drive Alone (Black) 95% CI - High | % Drive Alone (Hispanic) | % Drive Alone (Hispanic) 95% CI - Low | % Drive Alone (Hispanic) 95% CI - High | % Drive Alone (white) | % Drive Alone (white) 95% CI - Low | % Drive Alone (white) 95% CI - High | # Workers who Drive Alone | % Long Commute - Drives Alone | 95% CI - Low_19 | 95% CI - High_19 | Quartile_34 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 01000 | Alabama | NA | NA | 88086 | 10350 | 10246 | 10454 | NA | 5967 | 4840 | 7094 | NA | 3411 | 2945 | 3876 | NA | 13245 | 13023 | 13467 | NA | 5244 | 4901 | 5588 | NA | 9563 | 9439 | 9686 | NA | 21 | 20 | 23 | NA | 4.8 | 4.5 | 5.1 | NA | 5.6 | 5.3 | 6.0 | NA | NA | 10 | 10 | 11 | NA | 10 | 8 | 12 | 9 | 9 | 10 | 16 | 16 | 16 | 7 | 7 | 8 | 8 | 8 | 8 | 21 | 20 | 23 | NA | 36 | 35 | 38 | NA | 5.3 | NA | 31 | 29 | 32 | NA | 57 | NA | 15 | 14 | 16 | NA | 1255 | 4848 | 26 | 25 | 27 | NA | 31228 | 636.9 | NA | 28 | 27 | 28 | NA | 14 | 11 | 17 | 4 | 3 | 5 | 34 | 33 | 35 | 52 | 50 | 54 | 23 | 23 | 23 | 457718 | 12 | 11 | 12 | NA | 3228 | 66 | 1519:1 | NA | 2429 | 49 | 2026:1 | NA | 5818 | 118 | 846:1 | NA | 4875 | NA | 8445 | 2704 | 6198 | 4263 | 4597 | 42 | NA | 37 | 35 | 41 | 32 | 42 | 42 | NA | 39 | 42 | 30 | 32 | 45 | 2905059 | 3344006 | 87 | 87 | 87 | NA | 761762 | 1237123 | 62 | 61 | 62 | NA | 131065 | 2230132 | 5.9 | NA | 21 | 20 | 22 | NA | 26 | 12 | 38 | 38 | 13 | 106150 | 20580 | 5.2 | NA | 339392 | 1090304 | 31 | 31 | 32 | NA | 5960 | 12.2 | NA | 23307 | 480 | NA | 21249 | 87 | 86 | 88 | NA | 37 | 28 | 48 | 28 | 23 | 33 | 86 | 84 | 88 | 43 | 39 | 46 | 92 | 91 | 94 | 9.0 | NA | NA | NA | 14 | 13 | 14 | 12 | 12 | 12 | 2 | 2 | 2 | 1 | 1 | 1 | NA | 85 | 85 | 85 | NA | 80 | 78 | 81 | 76 | 74 | 79 | 83 | 83 | 84 | 77 | 76 | 79 | 86 | 86 | 86 | 2095195 | 35 | 35 | 36 | NA |
| 01001 | Alabama | Autauga | NA | 836 | 8027 | 7198 | 8857 | 1 | NA | NA | NA | NA | NA | NA | NA | NA | 11549 | 9369 | 13728 | NA | NA | NA | NA | NA | 7333 | 6411 | 8254 | NA | 20 | 18 | 23 | 1 | 4.5 | 4.2 | 4.8 | 1 | 5.4 | 5.1 | 5.7 | 1 | NA | 10 | 9 | 11 | 2 | NA | NA | NA | NA | NA | NA | 14 | 12 | 17 | NA | NA | NA | 8 | 7 | 9 | 20 | 17 | 22 | 1 | 35 | 34 | 37 | 1 | 6.5 | 3 | 32 | 30 | 35 | 1 | 63 | 1 | 16 | 15 | 17 | 4 | 18 | 56 | 32 | 25 | 39 | 3 | 323 | 578.1 | 2 | 23 | 20 | 26 | 1 | NA | NA | NA | NA | NA | NA | 29 | 23 | 37 | NA | NA | NA | 22 | 19 | 25 | 4366 | 9 | 8 | 11 | 1 | 25 | 45 | 2235:1 | 2 | 19 | 34 | 2955:1 | 2 | 21 | 37 | 2674:1 | 3 | 4931 | 3 | NA | NA | 11278 | NA | 4356 | 42 | 2 | NA | NA | 39 | 15 | 43 | 42 | 2 | NA | 36 | 31 | 33 | 44 | 33587 | 37860 | 89 | 87 | 90 | 1 | 8733 | 14382 | 61 | 55 | 67 | 1 | 1262 | 25838 | 4.9 | 2 | 15 | 9 | 20 | 1 | NA | 17 | 50 | 14 | 16 | 109878 | 21396 | 5.1 | 3 | 3628 | 13143 | 28 | 22 | 33 | 2 | 72 | 12.9 | 2 | 149 | 272 | 2 | 191 | 69 | 59 | 78 | 1 | NA | NA | NA | NA | NA | NA | 63 | 44 | 87 | NA | NA | NA | 73 | 61 | 84 | 9.5 | 4 | No | 1 | 15 | 12 | 17 | 13 | 10 | 15 | 1 | 1 | 2 | 2 | 1 | 3 | 3 | 87 | 84 | 89 | 3 | NA | NA | NA | NA | NA | NA | 83 | 76 | 90 | NA | NA | NA | 83 | 78 | 89 | 24949 | 41 | 36 | 45 | 3 |
| 01003 | Alabama | Baldwin | NA | 3377 | 8118 | 7667 | 8570 | 1 | NA | NA | NA | NA | NA | NA | NA | NA | 11603 | 9971 | 13235 | NA | 4591 | 3232 | 6327 | NA | 8059 | 7555 | 8562 | NA | 17 | 15 | 20 | 1 | 4.2 | 3.9 | 4.5 | 1 | 5.2 | 4.8 | 5.5 | 1 | NA | 8 | 8 | 9 | 1 | NA | NA | NA | 9 | 5 | 13 | 15 | 13 | 17 | 6 | 5 | 8 | 8 | 7 | 8 | 20 | 17 | 23 | 1 | 30 | 28 | 31 | 1 | 7.4 | 1 | 28 | 25 | 30 | 1 | 75 | 1 | 22 | 21 | 22 | 4 | 57 | 177 | 32 | 28 | 36 | 3 | 750 | 336.0 | 1 | 24 | 22 | 25 | 1 | NA | NA | NA | NA | NA | NA | 31 | 26 | 35 | 42 | 34 | 50 | 22 | 21 | 24 | 19085 | 11 | 10 | 12 | 1 | 154 | 69 | 1450:1 | 1 | 112 | 49 | 2047:1 | 1 | 228 | 99 | 1006:1 | 2 | 3578 | 1 | NA | 4956 | 5689 | 2386 | 3503 | 44 | 1 | 25 | 44 | 44 | 31 | 44 | 46 | 1 | 43 | 50 | 31 | 35 | 47 | 140740 | 155563 | 90 | 89 | 91 | 1 | 33289 | 50798 | 66 | 62 | 70 | 1 | 5425 | 96763 | 5.6 | 3 | 12 | 7 | 18 | 1 | 31 | NA | 16 | 24 | 8 | 120479 | 27466 | 4.4 | 1 | 8613 | 46902 | 18 | 15 | 22 | 1 | 226 | 10.1 | 3 | 408 | 204 | 1 | 817 | 75 | 70 | 80 | 1 | NA | NA | NA | NA | NA | NA | 63 | 49 | 81 | 26 | 14 | 44 | 80 | 74 | 86 | 7.2 | 1 | No | 1 | 12 | 11 | 14 | 11 | 10 | 12 | 1 | 1 | 2 | 1 | 0 | 1 | 3 | 83 | 81 | 85 | 1 | NA | NA | NA | NA | NA | NA | 87 | 80 | 94 | 68 | 56 | 80 | 82 | 80 | 83 | 97098 | 38 | 35 | 41 | 2 |
| 01005 | Alabama | Barbour | NA | 539 | 12877 | 11150 | 14604 | 4 | NA | NA | NA | NA | NA | NA | NA | NA | 15534 | 12931 | 18136 | NA | NA | NA | NA | NA | 9950 | 7691 | 12209 | NA | 31 | 28 | 34 | 4 | 5.9 | 5.6 | 6.2 | 4 | 6.1 | 5.9 | 6.4 | 4 | NA | 12 | 10 | 13 | 3 | NA | NA | NA | NA | NA | NA | 16 | 14 | 18 | NA | NA | NA | 7 | 5 | 9 | 28 | 25 | 31 | 4 | 40 | 39 | 42 | 4 | 5.7 | 4 | 42 | 39 | 44 | 4 | 50 | 2 | 14 | 13 | 15 | 1 | 12 | 32 | 38 | 28 | 46 | 4 | 221 | 895.2 | 4 | 35 | 29 | 40 | 3 | NA | NA | NA | NA | NA | NA | 34 | 28 | 42 | 77 | 45 | 123 | 30 | 23 | 39 | 2194 | 13 | 11 | 15 | 3 | 9 | 36 | 2743:1 | 3 | 9 | 37 | 2732:1 | 2 | 6 | 24 | 4098:1 | 3 | 4548 | 2 | NA | NA | 5222 | NA | 4207 | 46 | 1 | NA | NA | 46 | NA | 46 | 36 | 3 | NA | NA | 30 | NA | 37 | 13300 | 17797 | 75 | 72 | 77 | 4 | 2566 | 6690 | 38 | 33 | 44 | 4 | 605 | 8587 | 7.0 | 4 | 38 | 24 | 51 | 4 | NA | NA | 59 | 88 | 15 | 79175 | 13207 | 6.0 | 4 | 2822 | 5222 | 54 | 48 | 60 | 4 | 18 | 7.3 | 4 | 106 | 414 | 3 | 103 | 82 | 66 | 98 | 1 | NA | NA | NA | NA | NA | NA | 82 | 61 | 108 | NA | NA | NA | 86 | 64 | 114 | 9.0 | 2 | No | 1 | 15 | 13 | 18 | 13 | 11 | 16 | 3 | 2 | 5 | 0 | 0 | 1 | 4 | 84 | 81 | 87 | 2 | NA | NA | NA | NA | NA | NA | 77 | 70 | 85 | NA | NA | NA | 87 | 83 | 90 | 8555 | 37 | 31 | 42 | 2 |
| 01007 | Alabama | Bibb | NA | 460 | 11191 | 9626 | 12757 | 2 | NA | NA | NA | NA | NA | NA | NA | NA | 11737 | 8711 | 15474 | NA | NA | NA | NA | NA | 11369 | 9516 | 13222 | NA | 25 | 22 | 27 | 3 | 5.2 | 4.9 | 5.4 | 3 | 5.8 | 5.6 | 6.1 | 3 | NA | 10 | 9 | 11 | 2 | NA | NA | NA | NA | NA | NA | 17 | 13 | 21 | NA | NA | NA | 8 | 7 | 10 | 25 | 23 | 28 | 4 | 41 | 40 | 42 | 4 | 7.6 | 1 | 38 | 35 | 40 | 3 | 11 | 4 | 16 | 15 | 17 | 3 | 6 | 28 | 21 | 12 | 32 | 1 | 121 | 540.3 | 2 | 34 | 28 | 40 | 3 | NA | NA | NA | NA | NA | NA | 24 | 14 | 38 | NA | NA | NA | 36 | 29 | 43 | 1824 | 11 | 9 | 13 | 1 | 13 | 58 | 1723:1 | 1 | 6 | 27 | 3689:1 | 3 | 6 | 27 | 3689:1 | 3 | 6329 | 4 | NA | NA | 7592 | NA | 6232 | 37 | 3 | NA | NA | 41 | NA | 37 | 33 | 4 | NA | NA | 22 | NA | 35 | 12931 | 15987 | 81 | 78 | 84 | 4 | 2392 | 6333 | 38 | 31 | 44 | 4 | 573 | 8640 | 6.6 | 3 | 22 | 14 | 30 | 2 | NA | NA | 42 | 8 | 25 | 87594 | 16239 | 5.4 | 3 | 1666 | 4539 | 37 | 25 | 49 | 3 | 20 | 8.9 | 3 | 20 | 89 | 1 | 116 | 103 | 85 | 122 | 4 | NA | NA | NA | NA | NA | NA | 95 | 60 | 142 | NA | NA | NA | 109 | 88 | 134 | 9.4 | 4 | No | 1 | 11 | 7 | 15 | 9 | 6 | 13 | 1 | 0 | 2 | 1 | 0 | 2 | 2 | 88 | 84 | 91 | 4 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 8107 | 55 | 47 | 63 | 4 |
| 01009 | Alabama | Blount | NA | 1143 | 10787 | 9825 | 11749 | 2 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 23 | 20 | 26 | 2 | 4.9 | 4.6 | 5.3 | 2 | 5.7 | 5.4 | 6.0 | 2 | NA | 8 | 7 | 9 | 1 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 8 | 6 | 10 | 8 | 7 | 9 | 22 | 19 | 25 | 2 | 39 | 37 | 40 | 3 | 7.8 | 1 | 33 | 31 | 36 | 2 | 23 | 4 | 16 | 15 | 17 | 4 | 14 | 93 | 15 | 10 | 21 | 1 | 234 | 404.7 | 2 | 30 | 26 | 33 | 2 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 31 | 23 | 41 | 30 | 26 | 33 | 6663 | 14 | 13 | 16 | 4 | 13 | 22 | 4448:1 | 4 | 11 | 19 | 5262:1 | 4 | 10 | 17 | 5788:1 | 4 | 4260 | 1 | NA | NA | 10161 | 2138 | 4231 | 37 | 3 | NA | NA | NA | 32 | 37 | 40 | 2 | NA | NA | 25 | 30 | 40 | 32976 | 39814 | 83 | 81 | 84 | 3 | 7843 | 13939 | 56 | 51 | 62 | 2 | 1008 | 24661 | 4.1 | 1 | 19 | 14 | 24 | 1 | NA | NA | 3 | 39 | 15 | 100031 | 20993 | 4.8 | 2 | 3012 | 13352 | 23 | 17 | 28 | 1 | 43 | 7.4 | 4 | 279 | 483 | 3 | 303 | 105 | 93 | 117 | 4 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 40 | 20 | 72 | 113 | 100 | 126 | 9.4 | 4 | No | 1 | 10 | 9 | 12 | 8 | 6 | 9 | 2 | 1 | 2 | 1 | 1 | 2 | 1 | 88 | 86 | 91 | 4 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 78 | 36 | 100 | 84 | 81 | 87 | 22636 | 60 | 54 | 65 | 4 |
Median Household Income by State
The World Population Review published a dataset with with the median household incomes for each state in the United States, for the year 2022. There are 50 observations with two variables: State and Median Household Income.
| State | HouseholdIncome |
|---|---|
| Maryland | 84805 |
| New Jersey | 82545 |
| Hawaii | 81275 |
| Massachusetts | 81215 |
| Connecticut | 78444 |
| Alaska | 77640 |
12 Month-ending Provisional Number and Percent Change of Drug Overdose Deaths
The National Vital Statistics System from the CDC published provisional counts for drug overdoses caused by a variety of different drugs, by state, for each month between January 2015 and February 2022. There were 50,052 observations of 12 variables over the years, but the variables of interest we used were state and total deaths.
| State | Year | Month | Period | Indicator | Data.Value | Percent.Complete | Percent.Pending.Investigation | State.Name | Footnote | Footnote.Symbol | Predicted.Value |
|---|---|---|---|---|---|---|---|---|---|---|---|
| AK | 2015 | April | 12 month-ending | Cocaine (T40.5) | 100 | 0 | Alaska | Numbers may differ from published reports using final data. See Technical Notes. Data not shown due to low data quality. | ** | ||
| AK | 2015 | April | 12 month-ending | Heroin (T40.1) | 100 | 0 | Alaska | Numbers may differ from published reports using final data. See Technical Notes. Data not shown due to low data quality. | ** | ||
| AK | 2015 | April | 12 month-ending | Methadone (T40.3) | 100 | 0 | Alaska | Numbers may differ from published reports using final data. See Technical Notes. Data suppressed (<10). Data not shown due to low data quality. | ** | ||
| AK | 2015 | April | 12 month-ending | Number of Deaths | 4,133 | 100 | 0 | Alaska | Numbers may differ from published reports using final data. See Technical Notes. | ** | |
| AK | 2015 | April | 12 month-ending | Percent with drugs specified | 88.0952381 | 100 | 0 | Alaska | Numbers may differ from published reports using final data. See Technical Notes. | ** | |
| AK | 2015 | April | 12 month-ending | Psychostimulants with abuse potential (T43.6) | 100 | 0 | Alaska | Numbers may differ from published reports using final data. See Technical Notes. Data not shown due to low data quality. | ** |
Methods
Predictors of Alcohol-Related Driving Deaths
After cleaning the County Health Ranking Dataset to build models with this specific dataset, we were left with 52 different variables, as shown below:
| % Children in Poverty | % Excessive Drinking | Annual Average Violent Crimes | # Unemployed | |
|---|---|---|---|---|
| 1 | 21 | 15 | 23307 | 131065 |
| 2 | 15 | 16 | 149 | 1262 |
| 3 | 12 | 22 | 408 | 5425 |
| 4 | 38 | 14 | 106 | 605 |
| 5 | 22 | 16 | 20 | 573 |
| 6 | 19 | 16 | 279 | 1008 |
Looking at Alcohol-Related Driving Deaths plotted against other variables in the County Health Rankings Dataset, there was no clear linear relationship present.
The first model I decided to use is a principal component regression. Principal components are weighted linear combinations of the original variables, made to capture variability in the data. Each of the arrows on this graph show how much an original variable contributed to the weight of each of the first two principal components, which together describe 57% of the variability in the data. I chose this model because it is good for dimension reduction, which I believed would help the model.
## [1] NAWe decided to use 32 components for the PCR, as this value minimized the mean squared error.
To see if we could achieve similar accuracy on a more interpretable model, we used lasso regression to select certain variables that the model deemed specific, compared the model to prove whether or not the interpretability of a model meant a loss in accuracy, in this specific case. We used a lambda tuning parameter of 0.017.
Predictors of Drug Overdose
Looking at this US map, we can see the total deaths caused by drug overdoses in each state.
Here, we can see that California and Florida have the highest total drug
overdose deaths of all the states. We can also see that Texas and Ohio
also have high drug overdose deaths From this, we decided to look at the
most common drug that caused overdose in California, Florida, Ohio, and
Texas.
## State Indicator Data.Value
## 1 CA Opioids (T40.0-T40.4,T40.6) 3324
## 2 CA Opioids (T40.0-T40.4,T40.6) 3249
## 3 CA Opioids (T40.0-T40.4,T40.6) 3103## State Indicator Data.Value
## 1 FL Opioids (T40.0-T40.4,T40.6) 3981
## 2 FL Opioids (T40.0-T40.4,T40.6) 3923
## 3 FL Opioids (T40.0-T40.4,T40.6) 3872## State Indicator Data.Value
## 1 TX Opioids (T40.0-T40.4,T40.6) 1452
## 2 TX Opioids (T40.0-T40.4,T40.6) 1431
## 3 TX Opioids (T40.0-T40.4,T40.6) 1412## State Indicator Data.Value
## 1 OH Opioids (T40.0-T40.4,T40.6) 3484
## 2 OH Opioids (T40.0-T40.4,T40.6) 3463
## 3 OH Opioids (T40.0-T40.4,T40.6) 3394Based on this information, we decided the best way to look at the data was via clustering, in order to further detect patterns in the data. We used k-means clustering and model-based clustering.
Drug Abuse and Teen Births
First, we did some exploratory data analysis on the different variables.
We also decided to do a ridge regression on the County Health Rankings Dataset to answer the questions pertaining to drug abuse and teen births. Then, we did a simple linear regression to regress teen births against excessive drinking and smoking adults, to see if these specific behaviors had any relationships with the number of teen births in a specific county.
Results
Predictors of Alcohol-Related Driving Deaths
After running the PCR, it was shown that the number of components that gave us the smallest mean squared error was 32 principal components.
## [1] 55.73038After running the model using a test set of data, it was shown that the root mean squared error (RMSE) was 55.40335.
In comparison, when running the lasso regression, we found that the lambda value that gave us the lowest cross-validation error was 0.01798226.
Running the model on test data gave us an RMSE of 250.6965.
This shows us that the PCR was a much better tool to use than lasso regression when it comes to predicting the number of alcohol-related driving incidents in counties across the country. However, this does not quite help us understand which factors contribute to these types of driving incidents, as we have a model buit with 32 principal components made up of fractions of the tool.
Predictors of Drug Overdose
After standardizing the variables, we clustered the data to display the relationship between total drug overdoses and three different socioeconomic factors: high school completion, unemployment, and some college experience, as showcased below:
Drug Abuse and Teen Births
## # A tibble: 6 x 13
## Name Teen.births.raw.… Teen.births..Asia… Teen.births..Bl… Teen.births..Hi…
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 Autaug… 23.8 0 28.9 0
## 2 Baldwi… 26.0 0 32.9 45.0
## 3 Barbou… 37.1 0 38.0 56.6
## 4 Bibb C… 37.8 0 24.2 0
## 5 Blount… 31.2 0 0 30.2
## 6 Bulloc… 45.5 0 47.3 63.3
## # … with 8 more variables: Teen.births..White. <dbl>,
## # Excessive.drinking.raw.value <dbl>, Unemployment.raw.value <dbl>,
## # Children.in.poverty.raw.value <dbl>,
## # Children.in.single.parent.households.raw.value <dbl>,
## # Poor.mental.health.days.raw.value <dbl>, Adult.smoking.raw.value <dbl>,
## # Total <dbl>## # A tibble: 3,142 x 13
## Name Teen.births.raw.… Teen.births..Asia… Teen.births..Bl… Teen.births..Hi…
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 Autau… 23.8 0 28.9 0
## 2 Baldw… 26.0 0 32.9 45.0
## 3 Barbo… 37.1 0 38.0 56.6
## 4 Bibb … 37.8 0 24.2 0
## 5 Bloun… 31.2 0 0 30.2
## 6 Bullo… 45.5 0 47.3 63.3
## 7 Butle… 36.9 0 37.5 0
## 8 Calho… 32.3 0 32.1 35.3
## 9 Chamb… 42.3 0 44.5 0
## 10 Chero… 32.7 0 0 0
## # … with 3,132 more rows, and 8 more variables: Teen.births..White. <dbl>,
## # Excessive.drinking.raw.value <dbl>, Unemployment.raw.value <dbl>,
## # Children.in.poverty.raw.value <dbl>,
## # Children.in.single.parent.households.raw.value <dbl>,
## # Poor.mental.health.days.raw.value <dbl>, Adult.smoking.raw.value <dbl>,
## # Total <dbl>
## Discussion
Predictions of Drug Overdose
Based on the clustering, we discovered that drug overdose deaths tend to occur more in states with a lower unemployment rate, higher college experience, and higher high school completion rate. We also discovered that opioids were the most used substance in drug overdose deaths.
With this information, we want to look further into areas where opioid usage is more common, and see if opioids have different effects on specific racial and age groups.
There were some limitations we encountered when working on this project: there was no drug data in the county health rankings dataset, which made it difficult to examine patterns on such a granular level. Also, due to time contraints, we could not build and tune as many models as we would’ve liked.
Based on the work achieved so far, in the future, we wish to potentially obtain more county-level data, look more into the ooioid model, and experiment with more clustering and model building for each question.
Acknowledgements
We want to thank Dr. Ron Yurko, our professor during this summer, as well as Wanshan Li, YJ Choe, and all the other TAs for supporting us through this journey. We would also like to thank everyone in the Carnegie Mellon Statistics department who supported the SURE 2022 program. Thanks to Danita Kiser and everyone in Optum who advised us, including our guest speakers and our mentors.