More Information

Submitted: 07 April 2020 | Approved: 21 April 2020 | Published: 22 April 2020

How to cite this article: Jacobs M. Rural adolescent health: Issues, behaviors and self-reported awareness. J Community Med Health Solut. 2020; 1: 001-017.

DOI: 10.29328/journal.jcmhs.1001001

ORCiD: orcid.org/0000-0001-5943-8507

Copyright: © 2020 Jacobs M. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

 FullText PDF

Rural adolescent health: Issues, behaviors and self-reported awareness

Molly Jacobs*

Health Sciences BLDG, Greenville, NC 27858, USA

*Address for Correspondence: Molly Jacobs, 600 Moye Blvd, Mail Stop 668, 4340E Health Sciences BLDG, Greenville, NC 27858, USA, Tel: 252-744-6182; Email: jacobsm17@ecu.edu

Purpose: The purpose of the study was to examine the health status of rural adolescents and young adults in the United States through a comprehensive review of detailed health information, behavior and health awareness. The disparity in health awareness between rural and non-rural residents compared and evaluated.

Methods: Rural-Urban Commuting Area (RUCA) codes were combined with respondent-level data from the Longitudinal Survey of Adolescent to Adult Health (Add Health) to classify individuals as rural/non-rural residents. Health characteristics and perceived health awareness was tested for statistically significant differences using ANOVA. Differences in weight perception accuracy was compared for systematic differences controlling for self-selection into rural areas using a two-stage logistic selection model.

Findings: Analysis revealed that rural residents have a higher incidence of major health conditions including epilepsy, high cholesterol, high blood pressure and diabetes. Additionally, they have a higher prevalence of unhealthy behaviors including drinking and drug use. Rural residents are less likely to be insured, but more likely to be overweight or obese. While rural adolescents are more likely to mis-classify their body weight, this misclassification is a result of the higher incidence of overweight rather than the residential location.

Conclusion: The higher prevalence of chronic conditions combined with the income and education levels suggests the rural environment is a unique and potentially challenging context for adolescent health. Improving rural adolescent health will require innovative solutions appropriate for rural environments and changes in individual health literacy. Solutions must be multisectoral, engaging education, economic development, and other community perspectives to establish key drivers for health equity.

Since 2000, the rural population has grown less than urban and suburban, resulting a in smaller share of Americans living in rural counties [1]. A lower population base has led to a lack of health facilities [2]. Marginalized rural populations are particularly vulnerable to underrepresentation and policy neglect [3]. The difficulty in accessing quality health care combined with the rising cost of health care has put rural communities at risk for poor health outcomes [4]. A lack of information on the health status and risks of adolescent youth in rural areas undermines policymakers’ ability to justify budget expenditures for preventive care in rural areas. It is crucial to understand who they are and what contributes to health, chronic disease and conditions, to address the healthcare needs of rural communities.

Despite the difficulties faced by rural residents and evidence of disparate health, no comprehensive health assessments of rural adolescents and young adults in the US in recent decades [5] This study will supplement existing literature by providing an assessment of adolescent/young adult health in the US. This study uses the National Longitudinal Study of Adolescent to Adult Health (Add Health)-a longitudinal study of adolescents in grades 7-12 during the 1994-95 school year followed into young adulthood with four in-home interviews. This unique dataset with comprehensive health, clinical and biological outcomes to focus on three dimensions of adolescent health-chronic disease, health behavior and health self-awareness-in order to provide an understanding of the health issues faced by rural adolescents and possible avenues to health solutions.

Rural adolescents, particularly among poor and minority youth, are susceptible to significant risk behaviors and health concerns [3]. Studies have found that alcohol and drug use, pregnancy, and sexually transmitted disease rates are higher among rural adolescents [6,7]. Lack of employment, transportation, education, health services, and health insurance are associated with living in rural areas and increase rural health vulnerability [8]. Additionally, rural adolescents with substance abuse problems face the challenges of accessing adequate treatment and recovery [8].

One of the most frequently cited indicators of rural health is the disproportionately high rates of overweight and obesity rates. Rural adolescents have 26 percent greater odds of obesity, compared to urban adolescents (Johnson & Johnson, 2015). While most studies focus on those individual factors [3] that may contribute to obesity, but some examine associations with environmental characteristics such as food environment, physical changes, and social dynamics [10]. It is generally accepted that longer exposure to certain physical and social environments may contribute to differences in urban and rural obesity, but the mechanisms through which environmental aspects promote obesity warrants further study [11].

Rural areas suffer from a lack of physicians, specialists, nurses, and other healthcare practitioners, making it more difficult and cumbersome to obtain adequate preventative care [12]. Research estimates that an effective physician-to-population ratio is 1:1200 (Gale & Lambert, 2006), but the ratio is only 1:1910 in rural areas compared to 1:1300 in urban areas. National Rural Health Association reports that there are nearly 10 times more specialists per 100,000 urban residents compared to rural communities [13].

This study proceeds with a discussion of the data and methodology utilized, including the identification strategy and health outcomes selected, followed by a detailed outline of the primary significant differences between rural and non-rural youth and the most prevalent concerns among young adults in rural areas. Regression analysis attempts to explain some of these observed differences and most startling concerns. These ideas are then summarized with concluding remarks.

Identification

One primary explanation for variation in rural health outcomes research is the variable definition of rural. While the many definitions of the term rural seldom agree, the USDA Economic Research Service recommends that the choice of a rural definition be based on the purpose of the activity or the availability of information. This study utilizes the rural-urban commuting area (RUCA) codes which classify U.S. census tracts using measures of population density, urbanization, and daily commuting. RUCA codes are available in the data set used for this analysis. The most recent RUCA codes are based on data from the 2010 decennial census and the 2006-10 American Community Survey (ACS).

The Office of Management and Budget (OMB) uses RUCA codes to identify counties as metropolitan, micropolitan or neither. A metropolitan area contains a core urban area of 50,000 or more population while a micropolitan area contains an urban core of at least 10,000, but less than 50,000. All counties not part of a Metropolitan Statistical Area (MSA) are considered rural. Micropolitan counties are considered non-metropolitan or rural along with all counties not classified as metro or micro. After the 2010 Census, the non-metro counties contained 46.2 million people-15 percent of the US population and 72 percent of the land area of the country. This included all census tracts inside metropolitan counties with the codes 4-10 to be rural. Based on this assessment and review, this study classified respondents in areas with RUCA codes of 4-10 as rural aligning with OMB recommendations.

Data

Add Health Wave III data was collected when respondents were between 18 and 26 years old. Biological specimens, urine and saliva samples, were obtained from a subset of Wave III respondents for tests Chlamydia trachomatis (CT), Neisseria gonorrhoeae (GC), and other experimental STI testing. An oral mucosal transudate (OMT) specimen allowed for Human Immunodeficiency Virus Type-1 (HIV-1) testing along with other curable STDs. Saliva samples enabled DNA extraction, purification and subsequent genotyping of respondents.

In addition to biological and health outcomes data, Wave III contains information on parent-child and sibling relations, contact with friends from high school, the role of mentors and mentoring relationships, personal income, wealth and debt, civic and political participation, children and parenting, involvement with the criminal justice system, and religion and spirituality. Wave III also has extensive information on health and health related behavior including diet, physical activity, access and use of health services, sexual behavior, contraception, sexually transmitted infections, pregnancy and childbearing, suicidal intentions and thoughts, mental health and depression, substance use and abuse, injury, delinquency, and violence in addition to physical measurements of height and weight. Mean values for biological, demographic, social and behavioral characteristics are given in table 1.

Table 1:
Covariate Descriptive Statistics
  Non-Rural Rural
  12875     1183    
  N Mean Std Error N Mean Std Error
Race/Ethnicity
White 6556 64.5822 3.0101 790 75.5688 5.611
Black 2487 14.6194 2.0185 316 19.2103 5.1117
Hispanic 1028 5.5378 0.8731 11 0.7987 0.2826
Indian 415 2.9525 0.4436 45 3.0867 0.8599
Asian 1117 4.5415 0.9181 8 0.3829 0.2047
Other 1250 7.7666 1.1086 13 0.9526 0.3549
Age/Gender
Age 9820 19.751 0.127978 948 19.8021 0.255906
Male 6764 48.8644 0.6768 646 51.4848 1.5323
Female 6099 51.1356 0.6768 537 48.5152 1.5323
School Enrollment
Not Enrolled in School 7908 63.1946 1.5119 896 74.5069 2.2166
Enrolled in School 4941 36.8054 1.5119 287 25.4931 2.2166
Highest Grade Completed
6 7 0.0441 0.0235      
7 10 0.1061 0.0437      
8 52 0.6961 0.1564 10 0.5262 0.2492
9 195 2.1295 0.2659 27 2.5964 0.58
10 412 4.2883 0.3771 70 6.4159 0.9813
11 823 6.553 0.4106 121 10.4783 1.4387
12 4108 32.2841 1.3947 507 41.8246 1.9295
13 1934 16.0101 0.9409 159 12.9007 1.2475
14 1951 14.3438 0.6021 141 11.6303 1.3514
15 1268 8.7915 0.6251 75 6.9965 1.304
16 1480 10.3058 0.9872 48 4.1176 1.0093
17 371 2.4449 0.283 17 1.3358 0.3812
18 123 0.992 0.153 3 0.3862 0.2271
19 69 0.5286 0.083 2 0.293 0.2754
20 49 0.377 0.082 2 0.335 0.2551
21 8 0.062 0.0301      
22 3 0.043 0.0302 1 0.1635 0.1676
Average Highest Grade Completed 12863 13.1308 0.090942 1183 12.5578 0.111956
Household Circumstance
Household size 9581 4.32943 0.033446 910 4.217 0.071898
Lives with mother 8983 93.3223 0.4339 846 91.8895 1.0438
Lives with father 7440 78.596 1.1292 664 75.2789 1.6555
Income Parental/Earned/Household
Parental Income 9707  $ 47,009 1.791871 911  $ 33,967 1.725714
Income from earnings 9708  $ 12,941 431.7947 816  $ 11,164 952.6318
Household income 3059  $ 62,142 2348.48 265  $ 37,641 2769.819
Current Health Insurance Situation
You have no health insurance. 2915 24.1231 0.9158 333 28.721 1.9823
You are covered by your parents' insurance. 3413 27.3756 1.603 218 20.4388 2.2762
You are covered by your husband's or wife's insurance. 535 3.8294 0.3502 95 8.178 1.3951
You get insurance through work. 4196 30.8677 1.2746 335 27.6481 1.7675
You get insurance through a union. 77 0.5254 0.0908 5 0.5697 0.2953
You get insurance through school. 327 2.4399 0.3286 12 0.8619 0.2733
You are covered because you are active-duty military. 198 1.5208 0.1582 9 0.5705 0.3031
You buy private insurance yourself. 278 2.2642 0.251 34 3.3388 0.7702
You are on Medicaid. 745 6.17 0.6812 126 8.6948 1.2351
You are covered through the Indian Health Service. 26 0.2481 0.1708 2 0.0792 0.0805
You don't know what your health insurance coverage is. 72 0.6359 0.113 9 0.8992 0.3607
Months Covered by Health Insurance Last Year
Months last year have health insurance 12806 8.76413 0.114088 1173 8.21798 0.220624
BMI Classification
Underweight 359 2.8528 0.1741 29 2.4978 0.6216
Normal Weight 5418 43.0547 1.1497 440 37.8984 1.9113
Overweight 3726 29.609 0.5067 312 26.8734 1.3118
Obese 3081 24.4835 1.0642 380 32.7304 2.0066
Weight Perception (Self-Reported)
Very Underweight 155 1.1956 0.1478 11 0.7126 0.2949
Slightly Underweight 1437 11.8442 0.4066 113 8.9982 0.8284
Normal Weight 6135 47.9518 0.8577 549 45.6863 1.9268
Slightly Overweight 4294 33.151 0.6998 415 37.0928 2.0026
Very Overweight 822 5.8574 0.4139 93 7.5101 0.9772
Weight Action (Self-Reported)
Lose Weight 4349 32.1654 0.6913 380 33.722 1.579
Gain Weight 2016 16.0536 0.6293 158 12.5693 1.3897
Stay the same weight 1976 15.2615 0.5407 147 12.9375 1.3546
Not trying to do anything 4500 36.5195 0.7521 496 40.7712 2.6214
General Health Status (Self-Reported)
Excellent 4242 32.388 0.6393 379 31.1397 2.025
Very Good 5238 41.1549 0.6395 462 39.7833 1.8905
Good 2814 21.7971 0.6092 271 22.2716 2.0622
Fair 524 4.2773 0.2955 63 6.3396 0.9266
Poor 44 0.3827 0.0731 8 0.4658 0.255
Behavior
Hours television watching weekly 12741 12.7835 0.266617 1170 13.992 0.805366
Times exercise in last week 12833 5.87526 0.103162 1181 5.94603 0.30212
Gets enough Sleep 193 73.6075 3.1331 952 80.5103 1.7397
Days drink in last 12 months 9835 2.93523 0.038049 834 2.58848 0.07968
Days smoke in last month 4027 24.8003 0.224249 479 25.5647 0.502792
Illegal Drugs
Taken sedative last 5 years 12670 0.1139 0.005676 1162 0.10173 0.011782
Taken tranquilizers last 5 years 12676 0.09118 0.005118 1163 0.08818 0.016073
Taken stimulants last 5 years 12675 0.07965 0.004248 1166 0.0833 0.011478
Taken pain killers last 5 years 12661 0.20137 0.007233 1165 0.18703 0.0146
Taken steroids last 5 years 12680 0.0193 0.001919 1166 0.01495 0.004042
Used marijuana last 5 years 12657 0.47693 0.012005 1166 0.36284 0.023749
Used marijuana 1 last year 5702 0.71624 0.008777 429 0.69876 0.026699
Frequency used marijuana last 30 days 3975 11.8377 0.533064 296 9.8266 0.96387
Chronic Health Conditions
Asthma 2168 16.907 0.563 163 15.0583 1.0631
Cancer/Leukemia 116 0.7617 0.1033 10 1.005 0.4273
Depression 1328 11.3926 0.4803 137 13.4567 1.4405
Diabetes 120 0.9076 0.1419 18 1.5876 0.4031
Epilepsy/Seizure Disorder 160 1.3949 0.1795 32 2.4436 0.5154
High Cholesterol 582 4.4489 0.289 38 3.2651 0.5594
High Blood Pressure 677 5.4189 0.3038 96 8.4615 0.9888
STD 12796 0.12535 0.008164 1174 0.09798 0.019206
Health Status
Baroreflex Sensitivity (ms/mmHg) 11022 0.70035 0.059876 1039 0.45785 0.044319
Pulse Rate Recovery (beats/min) 11022 1.05952 0.055284 1039 0.82731 0.034677
SBP Recovery (mmHg) 11022 -0.61804 0.069721 1039 -0.94966 0.044025
High Sensitivity C-RCTV Protein (hsCRP)(MG/L) 9888 4.68294 0.119603 969 5.63744 0.352647
Epstein Barr Viral Capsid Antigen (EBV)(AU/ML) 9951 151.025 1.639873 973 150.133 3.466914
Count of Infectious/Inflammatory Diseases 11021 0.45999 0.010425 1039 0.44052 0.029067
Count of Subclinical Symptoms 11021 0.45529 0.010258 1039 0.46329 0.032034
Glucose (MG/DL) 9889 107.27 0.480135 960 109.333 1.282897
Hemoglobin A1c (%) 10149 5.57889 0.014988 989 5.65618 0.050861
Diabetes Joint Classification 11022 0.06205 0.004078 1039 0.07411 0.013395
Anti-Diabetic Medication Use 11022 0.01266 0.001505 1039 0.01749 0.005455
Triglycerides Decile 9636 5.58072 0.057281 936 5.91673 0.134905
Total Cholesterol Decile 9852 5.58067 0.053905 962 5.54253 0.140337
HDL Cholesterol Decile 9692 5.45514 0.057367 942 5.49539 0.155284
LDL Cholesterol Decile 9253 5.58159 0.053302 893 5.42939 0.159837
Total Number of Medications Currently Using 4145 1.86364 0.0309 429 1.95759 0.09849
Source: National Longitudinal Survey of Adolescent to Adult Health, Wave III, Restricted Use File
Covariates-health related behaviors

A variety of behavioral patterns are included in the Add Health survey. As with all surveys, patterns of omission, valid skip, non-response and refusal can impact the robustness of response data. In order to capture behavioral impacts on health and provide robust estimates, exercise frequency, sleep sufficiency, television watching, cigarette smoking, alcohol consumption and marijuana and illegal drug use are examined. Illegal drugs include sedatives, tranquilizers, stimulants, pain killers and steroids used by respondents anytime during the five years prior to their interview. Additionally, this study examines frequency of marijuana use in the last 12 months and last 4 weeks.

Alcohol consumption is measured as the number of days the respondent drank in the last 12 months, while smoking is measured as the number of days in the last month the respondents smoked. Binary indicators are added for having health insurance and receiving enough sleep, while variant terms measure the frequency of exercise and hours of television watching in an average week.

Covariates-health issues/indicators

Three measure of cardiovascular fitness are provided—Systolic Blood Pressure (SBP) Recovery, Pulse Rate Recovery (PRR) and Baroreflex Recover. First, SBP recovery after exercise represents an important index of cardiovascular and autonomic nervous system response to physical stress and has been shown to be a clinical tool applied toward diagnosing cardiovascular abnormalities. Second, PRR is a pulse measurement taken immediately following intense exercise. PRR is used in some fitness tests to evaluate the heart’s ability to recover from exercise and is used to evaluate the heart’s ability to recover from exercise. Finally, the baroreflex acts as an effective buffer of short-term blood pressure fluctuations that accompany daily life. Studies suggest that a diminished baroreflex recovery is an independent risk factor for sudden death after myocardial infarction. In hypertensive humans and animals, the baroreflex control of heart rate is diminished.

In addition to SBP Recovery, PRR and Baroreflex recovery, thirteen additional clinical measures are reported for each respondent. These measures indicate the existence, persistence or maintenance of health issues. 1) High Sensitivity C-reactive Protein (hsCRP) is a protein that increases in the blood with inflammation and infection as well as following a heart attack, surgery, or trauma. Studies have suggested that a persistent low level of inflammation is often associated with cardiovascular disease (CVD). The hs-CRP test accurately measures low levels of CRP to identify low but persistent levels of inflammation and helps predict a person’s risk of developing CVD.

The 2) Epstein Barr Viral Capsid Antigen (EBV) indicates that a person has or has had the Epstein Barr Virus. EBV is a member of the herpes virus family and one of the most common viruses to infect people around the world. According to the Centers for Disease Control and Prevention (CDC) [14] most people will contract EBV at some point. In adolescents and adults, it causes an illness called infectious mononucleosis, or mono, in about 35 to 50 percent of cases (2011). Also known as “the kissing disease,” EBV is usually spread through saliva and rarely through blood or other bodily fluids.

Additionally, 3) Glucose level, 4) Hemoglobin A1c level, 5) Triglycerides Decile, 6) Total Cholesterol Decile, 7) HDL Cholesterol Decile and 8) LDL Cholesterol Decile are provided in addition to four summary measure. The first summary measure, 9) Count of Common Subclinical Symptoms numerates the sources of infection or inflammation that have the potential to confound hsCRP-based estimates of cardiovascular disease risk. High hsCRP concentrations triggered searches for non-cardiovascular (e.g. infectious or inflammatory) diseases which were counted and categorized.

The second summary measure, 10) Count of Infectious/Inflammatory Diseases, therefore, counts and categorizes these conditions to enable investigators to control for potential confounding in hsCRP analyses. The third summary measure, 11) Diabetes Joint Classification, classifies respondents as having diabetes if they had a fasting glucose ≥ 126 mg/dl, non-fasting glucose ≥ 200 mg/dl, HbA1c ≥ 6.5%, self-reported history of diabetes except during pregnancy or used anti-diabetic medication in the past four weeks. Finally, 12) Anti-Diabetic Medication Use, flags those who report using medications in the past four weeks associated with one or more of the following therapeutic classification codes: antidiabetic agents, sulfonylureas, non-sulfonylureas, insulin, alpha-glucosidase inhibitors, thiazolidinediones, meglitinides, miscellaneous antidiabetic agents, antidiabetic combinations, dipeptidyl peptidase 4 inhibitors, amylin analogs or incretin mimetics. A final indicator, 13) Total Medications Currently using, captures drug use at the time of the survey.

In addition to these clinical measures, Add Health Respondents indicate whether they have ever been diagnosed by a doctor or nurse with any of the following conditions: asthma, cancer/leukemia, depression, diabetes, epilepsy/seizure disorder, high cholesterol, high blood pressure, bacterial vaginosis, cervicitis/ or mucopurulent cervicitis, chlamydia, genital herpes, genital warts, gonorrhea, hepatitis B, HIV/AIDS, human papilloma virus, pelvic inflammatory disease, syphilis, trichomoniasis, urethritis or vaginitis. For the purpose of this analysis, sexually transmitted diseases (STD) are collapsed into a single category indicating whether a respondent had been diagnosed with at least one STDs.

Covariates-self-reported health awareness

The CDC categorizes weight as (i) Underweight, (ii) Normal Weight, (iii) Overweight, and (iv) Obese based on their BMI level. Compared to other measure of body fatness, BMI appears to be correlated with various metabolic and disease outcomes. Despite criticisms of this generic scale, in general, BMI is an inexpensive and easy-to-perform method of screening for weight category. This analysis examines whether one’s own body perception aligns with their BMI classification. Add Health respondents classify their weight status as (i) Very Underweight, (ii) Slightly Underweight, (iii) Right Weight, (iv) Slightly Overweight, and (v) Very Overweight. Assuming that these categories represent self-assessments of BMI, they are aligned with the CDC categories as outlined in table 1a.

Table 1a:
BMI Value CDC Category Add Health Perception
< = 18.49 Underweight Underweight
18.50-24.99 Normal Weight About the right weight
25.0-29.99 Overweight Slightly Overweight
30.0+ Obese Very Overweight

Given the very small proportion of the sample classified as very underweight, both underweight categories are combined into a single underweight group. Analysis will compare individual’s assessment of their weight to the classification of their actual BMI to determine whether they systematically under, over or accurately estimate their body weight. The extent to which respondents over, under or accurately assess their weight is also examined and how mis-estimation varies by rural/urban status. These BMI classifications will also be compared to their reported intention to gain weight, lose weight, maintain weight or do nothing about their body weight, referred to as weight action. In addition to awareness of weight and necessary weight action, this study also examines individual assessment of their personal health which they classify as excellent, very good, food, fair or poor.

Statistical analysis

To accommodate the design of Add Health, statistical analysis needs to account for the sample weights, stratification and clustering. Failure to account for sampling weights affects the calculation of the point estimate while misspecification of the stratification or clustering impacts the calculation of the standard errors. Various procedures in SAS software package (SAS 9.4, Cary, NC) allow for correct estimation of variances/standard errors from complex samples. ANOVA tests for statistically significant differences between rural and non-rural samples.

Multinomial logit models evaluate the observed differences in weight and weight classification. Misclassification was identified as over (1), under (-1) or accurate (0) relative to the actual BMI classification (measured relative to their actual weight classification) and expressed as a function of age, BMI level, gender, income, general health, rural residence and school enrollment. Since individuals choose geographic location (urban, rural, suburban, etc.), residential self-selections could bias estimates by confounding observed differences. To ensure that estimates are robust to residential selection, a two-stage estimation selection procedure similar to the framework popularized by Heckman, [15] also estimates misclassification [16]. Stage one-selection-frames a binary indicator for rural residency as a function of age, adolescent school enrollment and income. Stage two-response-contends that misclassification is a function of age, gender, BMI and general health status. BMI serves as an explanatory variable to allow for variation in misclassification along the distribution.

Demographic characteristics

Results of listed in table 2. Few demographic differences between rural and non-rural residents exist. They appear to have similar age, gender, household size and household composition profiles. Respondents in both groups are equally distributed male and females, live in 3-4 person households and are between 18 and 24 years old. Surprisingly, the proportion living with their biological mother and/or father does not significantly differ, however, they do present significantly different education and income. Three income measurements-parent’s income, own earned income and own household income-were examined and show statistically lower income for rural residents who also have statistically lower educational attainment and fewer individuals enrolled in school. The racial/ethnic composition of rural and non-rural populations also differ significantly. Rural populations appear to be less diverse than others consisting of over 75% whites, compared to 65% in other areas. Minorities have a smaller representation in rural areas compared to non-rural communities.

Table 2:
Test of Statistically Significant Rural, Non-Rural Differences
Race
Effect F Value Pr > F  
Rural 5.84 <.0001  
Parameter Race Estimate Std Error
Intercept Black -1.4261*** 0.2043
Intercept Hispanic -3.5031*** 0.2242
Intercept Asian/Pacific Islander -3.1422*** 0.1879
Intercept American Indian/Alaskan Native -3.9704*** 0.3115
Intercept Other -3.2465*** 0.217
Rural Black -0.0565 0.1766
Rural Hispanic 1.0467*** 0.2142
Rural Asian/Pacific Islander 0.0557 0.1655
Rural American Indian/Alaskan Native 1.3145*** 0.2922
Rural Other 1.1271*** 0.2163
       
Gender
Effect F Value Pr > F  
Rural 2.42 0.1223  
Parameter Gender Estimate Std Err
Intercept Male -0.0072 0.0335
Rural Male 0.0522 0.0336
       
Age
Source Sum of Squares Mean Square F Value
Model 3464 3464.069 0.79***
Error 47382291 4401.104  
Effect F Value Pr > F  
Model 0.04 0.8458  
Intercept 9295.14*** <.0001  
Rural 0.04 0.8458  
       
Enrolled in School
Effect F Value Pr > F  
Rural 18.59*** <.0001  
Parameter   Estimate Std Error
Intercept Enrolled -0.8068*** 0.0712
Rural Not Enrolled 0.2654*** 0.0615
       
Highest Grad Completed
Effect F Value Pr > F  
Rural 50.04 <.0001  
Parameter Highest Grade Completed Estimate Std Error
Intercept 6 -5.7188*** 0.672
Intercept 7 -4.8401*** 0.6499
Intercept 8 1.9766** 0.6677
Intercept 9 3.3337*** 0.6355
Intercept 10 4.1361*** 0.615
Intercept 11 4.5934*** 0.6324
Intercept 12 6.0828*** 0.6226
Intercept 13 5.144*** 0.625
Intercept 14 5.0372*** 0.6183
Intercept 15 4.5383*** 0.6218
Intercept 16 4.3527*** 0.5977
Intercept 17 3.0705*** 0.6298
Intercept 18 1.9991** 0.6813
Intercept 19 1.5462* 0.7941
Intercept 20 1.4442** 0.7082
Intercept 21 -5.3777*** 0.6581
Rural 6 5.7432*** 0.6754
Rural 7 5.7432*** 0.6488
Rural 8 0.8077 0.6694
Rural 9 0.5687 0.6345
Rural 10 0.4664 0.6265
Rural 11 0.4332 0.6324
Rural 12 0.5384 0.6301
Rural 13 0.7758 0.6301
Rural 14 0.7727 0.6257
Rural 15 0.782 0.6415
Rural 16 1.1266* 0.6326
Rural 17 0.9701 0.6422
Rural 18 1.1395* 0.6754
Rural 19 0.9629 0.7894
Rural 20 0.7269 0.7201
Rural 21 5.7432*** 0.6645
       
Household Size
Source Sum of Squares Mean Square F Value
Model 16125 16125.13 8.86***
Error 19112564 1820.94  
Effect F Value Pr > F  
Model 2.21 0.1398  
Intercept 10703*** <.0001  
Rural 2.21 0.1398  
       
Lives with Biological Mother in Household
Effect F Value Pr > F  
Rural 1.81 0.1812  
Parameter   Estimate Std Error
Intercept   -2.5325*** 0.0774
Rural Mother not Present -0.1052 0.0782
       
Lives with Biological Father in Household
Effect F Value Pr > F  
Rural 3.48* 0.0646  
Parameter   Estimate Std Error
Intercept   -1.2067*** 0.0618
Rural Father not present -0.0932* 0.05
       
Parental Income
Source Sum of Squares Mean Square F Value
Model 2.32E+08 2.32E+08 71.54***
Error 3.45E+10 3246958  
Effect F Value Pr > F  
Model 29.72*** <.0001  
Intercept 971.15*** <.0001  
Rural 29.72*** <.0001  
       
Own Earned Income
Source Sum of Squares Mean Square F Value
Model 3.88E+12 3.88E+12 12.06***
Error 3.39E+15 3.22E+11  
Effect F Value Pr > F  
Model 3.44* 0.0658  
Intercept 457.54*** <.0001  
Rural 3.44* 0.0658  
       
Own Household Income
Source Sum of Squares Mean Square F Value
Model 2.32E+14 2.32E+14 40.93***
Error 1.88E+16 5.66E+12  
Effect F Value Pr > F  
Model 48.25*** <.0001  
Intercept 737.27*** <.0001  
Rural 48.25*** <.0001  
       
Current Health Insurance Situation
Effect F Value Pr > F  
Rural 5.78*** <.0001  
Parameter Current Health Insurance Situation Estimate Std Error
Intercept You have no health insurance. 3.5277*** 0.2295
Intercept You are covered by your parents' insurance. 3.42** 0.2015
Intercept You are covered by your husband's or wife's insurance. 1.9782*** 0.2608
Intercept You get insurance through work. 3.631*** 0.2303
Intercept You get insurance through a union. -0.3471 0.3369
Intercept You get insurance through school. 0.6281** 0.264
Intercept You are covered because you are active-duty military. 0.1853 0.3609
Intercept You buy private insurance yourself. 1.2674*** 0.2283
Intercept You are on Medicaid. 2.249*** 0.2354
Intercept You are covered through the Indian Health Service. -1.7081** 0.8653
Rural You have no health insurance. 0.0638 0.2405
Rural You are covered by your parents' insurance. 0.2963 0.2117
Rural You are covered by your husband's or wife's insurance. -0.2297 0.2551
Rural You get insurance through work. 0.2052 0.2384
Rural You get insurance through a union. 0.1092 0.3387
Rural You get insurance through school. 0.6696 0.2609
Rural You are covered because you are active-duty military. 0.6398* 0.3766
Rural You buy private insurance yourself. -0.0444 0.2385
Rural You are on Medicaid. -0.02 0.2424
Rural You are covered through the Indian Health Service. 0.7198** 0.3058
       
Months Last Year with Health Insurance
Source Mean Square F Value Pr > F
Model 515404.1 14.16*** 0.0002
Error 36394.6    
Effect F Value Pr > F  
Model 4.87** 0.0291  
Intercept 4561.49*** <.0001  
Rural 4.87** 0.0291  
       
BMI
Source Sum of Squares Mean Square F Value
Model 2857920 2857920 49.06***
Error 7.74E+08 58252  
Effect F Value Pr > F  
Model 24.41*** <.0001  
Intercept 34220.7*** <.0001  
Rural 24.41*** <.0001  
       
Weight Perception
Effect F Value Pr > F  
Rural 3.3** 0.0131  
Parameter Weight Perception Estimate Std Error
Intercept Very Underweight -1.9701*** 0.2204
Intercept Slightly Underweight 0.4432*** 0.0897
Intercept Normal Weight 1.9544*** 0.0877
Intercept Slightly Overweight 1.6665*** 0.0821
Rural Very Underweight 0.3811* 0.2131
Rural Slightly Underweight 0.2623*** 0.0857
Rural Normal Weight 0.149* 0.0859
Rural Slightly Overweight 0.0695 0.0823
       
Weight Action
Effect F Value Pr > F  
Rural 3.68*** 0.0139  
Parameter Weight Action Estimate Std Error
Intercept Lose Weight -0.1573** 0.0561
Intercept Gain Weight -0.9996*** 0.0834
Intercept Stay the Same Weight -1.0107*** 0.0791
Rural Lose Weight 0.0326 0.0546
Rural Gain Weight 0.177** 0.0783
Rural Stay the Same Weight 0.1371* 0.0748
       
General Health Assessment
Effect F Value Pr > F  
Rural 1.62 0.1739  
Parameter General Health Assessment Estimate Std Error
Intercept Excellent 4.3212*** 0.2905
Intercept Very Good 4.5631*** 0.288
Intercept Good 3.9559*** 0.287
Intercept Fair 2.5152*** 0.3054
Rural Excellent 0.1186 0.3006
Rural Very Good 0.1155 0.2936
Rural Good 0.0885 0.2971
Rural Fair -0.0957 0.3109
       
Exercise Frequency
Source Sum of Squares Mean Square F Value
Model 7595 7595 0.13
Error 8.27E+08 58982.97  
Effect F Value Pr > F  
Model 0.04 0.8334  
Intercept 1297.9*** <.0001  
Rural 0.04 0.8334  
       
Hours of Television Viewing
Source Sum of Squares Mean Square F Value
Model 2483681 2483681 9.06***
Error 3.81E+09 274040  
Effect F Value Pr > F  
Model 2.04 0.1559  
Intercept 965.34*** <.0001  
Rural 2.04 0.1559  
       
Gets Enough Sleep
Effect F Value Pr > F  
Rural 0.37 0.542  
Parameter Gets Enough Sleep Estimate Std Error
Intercept 0 -1.4537*** 0.0625
Rural 0 -0.0352 0.0576
       
Frequency of Alcohol Consumption
Source Sum of Squares Mean Square F Value
Model 152432 152432 45.97***
Error 35381341 3315.7  
Effect F Value Pr > F  
Model 16.79*** <.0001  
Intercept 3645.13*** <.0001  
Rural 16.79*** <.0001  
       
Cigarette Smoking Frequency
Source Sum of Squares Mean Square F Value
Model 416658 416658.5 2.85***
Error 6.60E+08 146408.9  
Effect F Value Pr > F  
Model 2.04 0.1556  
Intercept 8126.33*** <.0001  
Rural 2.04 0.1556  
       
Used Sedatives in the Last 5 Years
Effect F Value Pr > F  
Rural 0.94 0.3346  
Parameter Sedative Use Estimate Std Error
Intercept No Use 2.1148*** 0.0746
Rural No Use -0.0633 0.0654
       
Used Tranquilizers in the Last 5 years
Effect F Value Pr > F  
Rural 0.03 0.8586  
Parameter Tranquilizer Use Estimate Std Error
Intercept No Use 2.3175*** 0.1072
Rural No Use -0.0181 0.1016
       
Used Stimulants in the Last 5 Years
Effect F Value Pr > F  
Rural 0.1 0.7547  
Parameter Stimulant Use Estimate Std Error
Intercept No Use 2.4227*** 0.0829
Rural No Use 0.0244 0.0779
       
Used Pain Killers in the Last 5 Years
Effect F Value Pr > F  
Rural 0.83 0.3646  
Parameter Pain Killer Use Estimate Std Error
Intercept No Use 1.4236 0.0555
Rural No Use -0.0458 0.0504
       
Used Steriods in the Last 5 Years
Effect F Value Pr > F  
Rural 0.75 0.3871  
Parameter Steriod Use Estimate Std Error
Intercept No Use 4.0581*** 0.1421
Rural No Use -0.1298 0.1496
       
Used Marijuana in the Last 5 Years
Effect F Value Pr > F  
Rural 18.71*** <.0001  
Parameter Marijuana Use Estimate Std Error
Intercept No Use 0.3277*** 0.0594
Rural No Use -0.2353*** 0.0544
       
Marijuana Use in the Last 1 Year
Effect F Value Pr > F  
Rural 0.44 0.5087  
Parameter Marijuana Use Estimate Std Error
Intercept No Use -0.8836*** 0.069
Rural No Use -0.0422 0.0637
       
Frequency Used Marijuana in the Last 30 Days
Source Sum of Squares Mean Square F Value
Model 1801777 1801777 1.63
Error 4.71E+09 1102393  
Effect F Value Pr > F  
Model 3.64* 0.0586  
Intercept 344.47*** <.0001  
Rural 3.64* 0.0586  
       
Diagnosed with Asthma
Effect F Value Pr > F  
Rural 2.1 0.1498  
Parameter Asthma Estimate Std Error
Intercept Not Diagnosed 1.662*** 0.0451
Rural Not Diagnosed -0.068** 0.0469
       
Diagnosed with Cancer/Leukemia
Effect F Value Pr > F  
Rural 0.36 0.5506  
Parameter Cancer/Leukemia Estimate Std Error
Intercept Not Diagnosed 4.7304*** 0.2151
Rural Not Diagnosed 0.1408 0.2353
       
Diagnosed with Depression
Effect F Value Pr > F  
Rural 2.3 0.1321  
Parameter Depression Estimate Std Error
Intercept Not Diagnosed 1.9567*** 0.0694
Rural Not Diagnosed 0.0956 0.0631
       
Diagnosed with Diabetes
Effect F Value Pr > F  
Rural 3.83** 0.0525  
Parameter Diabetes Estimate Std Error
Intercept Not Diagnosed 4.4107*** 0.1565
Rural Not Diagnosed 0.2838** 0.145
       
Diagnosed with Epilepsy/Seizure Disorder
Effect F Value Pr > F  
Rural 5.46** 0.021  
Parameter Epilepsy/Seizure Disorder Estimate Std Error
Intercept Not Diagnosed 3.9734*** 0.1295
Rural Not Diagnosed 0.2865** 0.1226
       
Diagnosed with High Cholesterol
Effect F Value Pr > F  
Rural 3.12* 0.0799  
Parameter High Cholesterol Estimate Std Error
Intercept Not Diagnosed 3.2283*** 0.099
Rural Not Diagnosed -0.1597* 0.0905
       
Diagnosed with High Blood Pressure
Effect F Value Pr > F  
Rural 11.66*** 0.0009  
Parameter High Blood Pressure Estimate Std Error
Intercept Not Diagnosed 2.6212*** 0.0705
Rural Not Diagnosed 0.24*** 0.0703
       
Number of STD Diagnoses
Source Sum of Squares Mean Square F Value
Model 1293 1292.614 2.06
Error 8767058 627.114  
Effect F Value Pr > F  
Model 1.73 0.1905  
Intercept 113.05*** <.0001  
Rural 1.73 0.1905  
       
Baroreflex Sensitivity (ms/mmHg)
Source Sum of Squares Mean Square F Value
Model 91940 91940.03 14.02***
Error 79087053 6558.34  
Effect F Value Pr > F  
Model 10.54*** 0.0015  
Intercept 241.58*** <.0001  
Rural 10.54*** 0.0015  
       
Pulse Rate Recovery (beats/min)
Source Sum of Squares Mean Square F Value
Model 84299 84298.78 15.09***
Error 67348012 5584.88  
Effect F Value Pr > F  
Model 12.45*** 0.0006  
Intercept 849.5*** <.0001  
Rural 12.45*** 0.0006  
       
Systolic Blood Pressure Recovery (mmHg)
Source Sum of Squares Mean Square F Value
Model 171923 171923.2 19.53***
Error 1.06E+08 8804.6  
Effect F Value Pr > F  
Model 16.46*** <.0001  
Intercept 353.76*** <.0001  
Rural 16.46*** <.0001  
       
High Sensitivity C-RCTV Protein (hsCRP)(MG/L)
Source Sum of Squares Mean Square F Value
Model 1318239 1318239 12.27***
Error 1.17E+09 107410  
Effect F Value Pr > F  
Model 6.87*** 0.0098  
Intercept 727.41*** <.0001  
Rural 6.87*** 0.0098  
       
Epstein Barr Viral Capsid Antigen (EBV)(AU/ML)
Source Sum of Squares Mean Square F Value
Model 1156102 1156102 0.07
Error 1.72E+11 15717576  
Effect F Value Pr > F  
Model 0.05 0.8163  
Intercept 6086.19*** <.0001  
Rural 0.05 0.8163  
       
Count of Infectious/Inflammatory Diseases
Source Sum of Squares Mean Square F Value
Model 593 592.9066 0.83
Error 8600782 713.2843  
Effect F Value Pr > F  
Model 0.43 0.5154  
Intercept 808.51*** <.0001  
Rural 0.43 0.5154  
       
Count of Subclinical Symptoms
Source Sum of Squares Mean Square F Value
Model 100 100.1764 0.12
Error 10104300 837.9748  
Effect F Value Pr > F  
Model 0.06 0.8109  
Intercept 744.41*** <.0001  
Rural 0.06 0.8109  
       
Glucose (MG/DL)
Source Sum of Squares Mean Square F Value
Model 6112566 6112566 4**
Error 1.66E+10 1529982  
Effect F Value Pr > F  
Model 2.31 0.131  
Intercept 24898.9*** <.0001  
Rural 2.31 0.131  
       
Hemoglobin A1c (%)
Source Sum of Squares Mean Square F Value
Model 8851 8851.242 8.8***
Error 11205436 1006.235  
Effect F Value Pr > F  
Model 2.19 0.1416  
Intercept 43154.4*** <.0001  
Rural 2.19 0.1416  
       
Diabetes Joint Classification
Source Sum of Squares Mean Square F Value
Model 228 227.5605 2.49
Error 1101430 91.3368  
Effect F Value Pr > F  
Model 0.82 0.3681  
Intercept 86.56*** <.0001  
Rural 0.82 0.3681  
       
Anti-Diabetic Medication Use
Source Sum of Squares Mean Square F Value
Model 36.5 36.51253 1.83
Error 240789.2 19.96759  
Effect Pr > F    
Model 0.73 0.3931  
Intercept 28.49*** <.0001  
Rural 0.73 0.3931  
       
Triglycerides Decile
Source Sum of Squares Mean Square F Value
Model 158263 158263 12.61***
Error 1.33E+08 12552.3  
Effect F Value Pr > F  
Model 5.64** 0.019  
Intercept 5698.14*** <.0001  
Rural 5.64** 0.019  
       
Total Cholesterol Decile
Source Sum of Squares Mean Square F Value
Model 2094.6 2094.6 0.16
Error 1.38E+08 12800.42  
Effect F Value Pr > F  
Model 0.07 0.792  
Intercept 5041.99*** <.0001  
Rural 0.07 0.792  
       
HDL Cholesterol Decile
Source Sum of Squares Mean Square F Value
Model 2286.9 2286.95 0.18
Error 1.37E+08 12872.31  
Effect F Value Pr > F  
Model 0.06 0.8014  
Intercept 4022.53*** <.0001  
Rural 0.06 0.8014  
       
LDL Cholesterol Decile
Source Sum of Squares Mean Square F Value
Model 30750.2 30750.18 2.41
Error 1.29E+08 12743.35  
Effect F Value Pr > F  
Model 0.85 0.357  
Intercept 4108.71*** <.0001  
Rural 0.85 0.357  
       
Total Number of Medications
Source Sum of Squares Mean Square F Value
Model 5775 5775.357 1.75
Error 15129349 3309.131  
Effect F Value Pr > F  
Model 0.76 0.3837  
Intercept 1469.12*** <.0001  
Rural 0.76 0.3837  
Health issues/indicators

There is a large difference in the health insurance status of the two groups. A higher percentage of rural residents have no health insurance, while less are covered by the insurance of a spouse or parent. They report that they held insurance for fewer months last year compared to non-rural young adults.While not directly related, it is likely that the lack of insurance coverage or full-year insurance coverage contributed to worse health outcomes by reducing the quantity and/or quality of care received [17]. Health disparities have also been linked to lack of preventative health services obtained [18].

One of the most unique aspects of the Add Health data is the large amount of medical diagnosis and clinical information available. Comparing diagnosis data between non-rural and rural adolescent and young adults show higher rates of asthma, epilepsy/seizure disorders, diabetes, high cholesterol, high cholesterol and high blood pressure in rural residents. Diabetes, high cholesterol and high blood pressure are known comorbidities of overweight and obesity and higher rates of excess weight among rural residents’ likely attributes to the higher rates of related comorbidities [19,20]. Additionally, rural adolescents have higher triglycerides and hs-CRP indicating high levels of these fatty particles in the blood and greater risk of heart disease. Rural residents also show statistically higher rates of seizure disorders—a condition that has been growing in prevalence over the last decade, according to the CDC. Studies attributed these higher rates to the increased prevalence of untreated traumatic head injuries (Engel et. al., 2003).

SBP Recover, PRR and BRS differ between rural and non-rural residents suggesting lower relative cardiac fitness, increased tendency towards cardiovascular disease (CVD) and higher likelihood of coronary issues or disorders. While detailed medical review of these factors lies outside the scope of this paper, they can be impacted by a variety of factors including our age, medical conditions, medications, diet, and fitness level.

Health related behaviors

There is no difference in exercise frequency, sleep or hours of television. The proportions of those who reported having used sedatives, pain killers, stimulants, tranquilizers or steroids in the last five years do not differ significantly. Rural residents consume alcohol and marijuana more frequently. These trends are supported by literature showing large differences were exhibited in marijuana use, both across nonmetropolitan‐metropolitan status and across youth from metropolitan and nonmetropolitan counties, but that rates of illicit drug use were essentially the same regardless of location [21,22].

Self-reported health awareness

Many of these health conditions are the result of excess body weight or obesity. Examination of BMI showed higher BMI among rural youth. While these BMI levels are highly unhealthy, it does not appear that rural respondents are aware of their situation or report an intention to change. Roughly equal proportions of both rural and non-rural residents report that they are overweight, despite a greater prevalence of overweight and obesity among rural residents. This indicates that either rural respondents are not aware of their BMI status or refuse to report themselves as such. Furthermore, they do not appear to be any more likely to report wanting to lose weight than their non-rural counterparts.

Finally, given the results presented above, rural and urban residents report similar self-assessments of their overall health. This lack of health awareness among rural residents has been found by other researchers as well [23]. The lack of awareness or refusal to accept their status is often perpetuated by the community at large and ignorance regarding the detrimental health effects of excess weight [24].

Misclassification selection model

Multinomial logit model estimates of weight misclassification show that misclassification type varies by age, gender, school enrollment, and general health status, but not by income or rural residency (Table 3). As individuals age and increase BMI they are less likely to underestimate and more likely to overestimate their weight. As adolescents leave school and experience health declines more likely to overestimate and less likely to underestimate their body weight. Blacks and females tend to overestimate weight. The multinomial showed that misclassification does not differ significantly for rural and non-rural residents when controlling for age, BMI, gender and other factors.

Table 3:
Multinomial Logit Estimates of BMI Miscalculation
Model Fit Statistics Dependent Variable: Misclassification
Criterion Intercept Intercept, Covariates Category Code N
AIC 20746022 17467953 Underestimate -1 1113
SC 20746051 17468180 Accurately Estimate 0 4751
-2 Log L 20746018 17467921 Overestimate 1 2878
Analysis of Maximum Likelihood Estimates Odds Ratio Estimates
Parameter Comparison Estimate Std Err Estimate 95% Confidence Limits
Intercept Underestimate 9.6725*** 0.8929      
Intercept Overestimate -12.4379*** 0.8528      
Age Underestimate -0.0679** 0.0205 0.934 0.897 0.973
Age Overestimate 0.012 0.0124 1.012 0.987 1.037
Female Underestimate -0.7809*** 0.1071 0.458 0.371 0.566
Female Overestimate 1.1303*** 0.0969 3.097 2.556 3.751
Health Underestimate 0.155** 0.0658 1.168 1.025 1.33
Health Overestimate -0.1931*** 0.0442 0.824 0.755 0.9
lBMI Underestimate -3.2074*** 0.283 0.04 0.023 0.071
lBMI Overestimate 3.6044*** 0.2524 36.759 22.301 60.589
School Underestimate 0.1894 0.1503 1.209 0.897 1.628
School Overestimate -0.214** 0.1086 0.807 0.651 1.001
lIncome Underestimate -0.0205 0.0181 0.98 0.945 1.015
lIncome Overestimate -0.0164 0.0195 0.984 0.946 1.023
Black Underestimate 0.5646*** 0.1123 0.74 0.558 0.982
Black Overestimate -0.3006** 0.1427 1.759 1.408 2.196
Rural Underestimate 0.0295 0.1267 1.03 0.801 1.324
Rural Overestimate 0.1349 0.0891 1.144 0.959 1.365
Reference: 0=Accurately Estimate Weight; Dependent Variable: Misclassification= -1=Underestimate, 0=Accurately Estimate, 1=Overestimate 
Estimates are weighted to account for survey sampling.

Multinomial odds ratio estimates suggest that BMI is the largest and most important driver of weight misclassification. Estimates suggest that the probability of overestimation increases as BMI increases with an odds ratio of 36.054. Estimates suggest that BMI is the primary driver of misclassification. A two-stage sample selection model tests the robustness of these results. This technique controls for self-selection into rural areas before estimating the misclassification model. Two-stage estimates (Table 4) suggest that those factors associated with weight misclassification are similar for rural and non-rural residents. Controlling for residential self-selection, model results show that BMI is the primary determinant of misclassification and misclassification type.

Table 4:
2 Stage Residential Selection Model of Weight Misclassification
Selection: Rural=0 Selection: Rural=1
Heckman First Stage Discrete Selection Response Profile
Index Value   Index Value    
N: Non-Rural 6114   N: Non-Rural 7040    
N: Rural 2140   N: Rural 1894    
Log Likelihood -4649   Log Likelihood -4566    
AIC 9307   AIC 9142    
Schwarz Criterion 9342   Schwarz Criterion 9178    
Likelihood Ratio (R) 150.2   Likelihood Ratio (R) 98.185    
Stage I: Parameter Estimates
Parameter Estimate Std Err Marginal Effect Estimate Standard Marginal Effect
Intercept 0.706606*** 0.130202   0.189208 0.129179  
Age -0.049711*** 0.004796 -0.0090389 -0.03159*** 0.00479 0.0090389
lIncome -0.027847*** 0.006366 -0.008311 -0.030036*** 0.006267 0.008311
Highest Grade 0.053768** 0.016805 0.0149391 0.05399** 0.01686 -0.0149391
School Enrollment -0.410627*** 0.043737 -0.092755 -0.335219*** 0.043354 0.092755
Dependent Variable: Rural- 1=Rural, 0=Non-Rural      
Heckman Second Step Model Fit Summary
Log Likelihood -5239     Log Likelihood -1508  
AIC 10493     AIC 3032  
Schwarz Criterion 10547     Schwarz Criterion 3077  
Stage II: Parameter Estimates
Parameter Estimate Std. Err Marginal Effect Estimate Standard Marginal Effect
Intercept -2.991752*** 0.12232   -3.408589*** 0.236078  
Age 0.01595*** 0.002035 0.0231966 0.023197*** 0.003042 0.0159503
Female 0.313844*** 0.014654 0.3095502 0.30955*** 0.024843 0.3138435
lBMI 0.952606*** 0.035012 1.1461962 1.146196*** 0.056458 0.9526056
Black 0.112400*** 0.017956 0.1123999 0.193228*** 0.029395 0.1932281
General Health -0.07035*** 0.008804 -0.0740746 -0.074075*** 0.015325 -0.07035
Lambda -0.160641** 0.06013   -0.380529** 0.118997  
Sigma 0.569997*** 0.005155   0.536512*** 0.008717  
Reference: 0=Accurately Estimate Weight; Dependent Variable: Misclassification= -1=Underestimate, 0=Accurately Estimate, 1=Overestimate 
Estimates are weighted to account for survey sampling.

While demographically similar, rural and non-rural youth have vastly different health profiles, behaviors and self-awareness. This study utilizes RUCA codes to classify adolescents as rural based on the OMB county classifications. Adolescents within these non-metropolitan, rural areas have higher incidence of all major health conditions including epilepsy, high cholesterol, diabetes and high blood pressure. Not only are these health concern more prevalent among rural individuals, but their health concerns extend beyond measurable conditions to include a higher prevalence of unhealthy behaviors including drinking and marijuana use.

Rural adolescents are more likely to be overweight or obese than urban. However, rural adolescents do not appear to be aware of the severity of their excess weight or the adverse health conditions that it causes—high cholesterol, high blood pressure and diabetes—which disproportionately impact rural youth. Disparate health outcomes could be partially attributed to the lack of preventative care. These findings that speak to the complexity of adolescent health. Rural areas have a higher prevalence of overweight compared to non-rural. Individuals in rural areas are also more likely to misclassify their body weight. Regression analysis indicated that as BMI increases, individuals are more likely to underestimate their weight status. Results transcend self-selection into rural areas showing that BMI misclassification is primarily determined by BMI level irrespective of residential location.

These results reinforce the notion that rural areas are a unique area with distinct challenges related to health. While the most prevalent health conditions are not surprising, they are becoming more difficult to treat as the number of rural hospitals has decreased over recent decades and the number of accessible physicians has decreased. Given the lack of health infrastructure, geographic isolation, insufficient financial resources and lack of available services, conventional public health solutions would likely not be effective. In order to attain health equity, alternatives such as school-based or community driven healthcare should be explored.

While the limited access to care in rural areas is often cited as a reason for poor health, few studies have examined the differential health literacy among rural residents. Health literacy is the ability to obtain, read, understand, and use healthcare information in order to make appropriate health decisions and follow instructions for treatment. In addition to greater provision of care, programs that increase awareness of individuals health needs and proper preventative lifestyle measures could also assist in improving health in rural areas.

  1. Parker K, Horowitz JM, Brown A, Fry R, Cohn D, et al. What Unites and Divides Urban, Suburban and Rural Communities. 2018. https://www.pewsocialtrends.org/2018/05/22/what-unites-and-divides-urban-suburban-and-rural-communities/
  2. Rechel B, Džakula A, Duran A,Fattore G,Edwards N, et al. Hospitals in rural or remote areas: An exploratory review of policies in 8 high-income countries. Health Policy. 2015; 120: 758-769. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27312144
  3. Curtis AC, Waters CM, Brindis C. Rural adolescent health: the importance of prevention services in the rural community. J Rural Health. 2011; 27: 60-71. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21204973
  4. Levine SB,Coupey SM. Adolescent substance use, sexual behavior, and metropolitan status: is 'urban' a risk factor? J Adolescent Health. 2003; 32L 350-355. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/12729984
  5. Catalano RF, Fagan AA, Gavin LE, Greenberg MT,Irwin CE Jr, et al. Worldwide application of prevention science in adolescent health. Lancet. 2012; 379: 1653-1664. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22538180
  6. Eaton DK, Kann L, Kinchen S, Shanklin S, Flint KH, et al. Youth risk behavior surveillance – United States, 2011. Morbidity and Mortality Weekly Report: Surveillance Summaries. 2012; 61: 1-62. PMid:22673000
  7. Raingruber B. Contemporary health promotion in nursing practice. Burlington, MA: Jones & Bartlett. 2016; PMid:27504500.
  8. Santrock JW. A topical approach to life-span development. New York: McGraw-Hill Education. 2014.
  9. Miller AS. CE FEATURE Adolescent Alcohol and Substance Abuse in Rural Areas: How Telehealth Can Provide Treatment Solutions. J Addic Nurs . 2005; 16: 107-115.
  10. Salois MJ. The built environment and obesity among low-income preschool children. Health Place. 2012; 18: 520–527.
  11. Akinbami LJ, Ogden CL. Childhood overweight prevalence in the United States: The impact of parent‐reported height and weight. Obesity. 2009; 17: 1574–1580. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19629061
  12. Rosenblatt RA, Chen FM, Lishner DM, Doesher MP. WWAMI Rural Health Center. The future of Family Medicine and Implications for Rural Primary Care Physician Supply. Grand Forks, ND. 2010.
  13. Warshaw R. Health Disparities Affect Millions in Rural U.S. Communities. 2017. https://www.aamc.org/news-insights/health-disparities-affect-millions-rural-us-communities
  14. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance. 2011. http://www.cdc.gov/std/stats11/adol.htm
  15. Heckman JJ. Sample Selection Bias as a Specification Error. Econometrica. 1979; 47: 153-161.
  16. Borjas GJ, Sueyoshi GT. A two-stage estimator for probit models with structural group effects. J Econometrics. 1994; 64: 165-182. PubMed:
  17. Frenzen PD. Health insurance coverage in US urban and rural areas. J Rural Health. 1993; 9: 204-214. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/10128159
  18. Casey MM, Thiede Call K, Klingner JM. Are rural residents less likely to obtain recommended preventive healthcare services? Am J Prev Med. 2001; 21: 182-188. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/11567838
  19. Bennett KJ, Probst JC, Pumkam C. Obesity among working age adults: the role of county-level persistent poverty in rural disparities. Health Place. 2011; 17: 1174-1181. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21665515
  20. Befort CA, Nazir N, Perri MG. Prevalence of obesity among adults from rural and urban areas of the United States: findings from NHANES (2005‐2008). J Rural Health. 2012; 28: 392-397. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23083085
  21. Cronk CE, Sarvela PD. Alcohol, tobacco, and other drug use among rural/small town and urban youth: a secondary analysis of the monitoring the future data set. Ame J Public Health. 1997; 87: 760-764. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/9184502
  22. Donnermeyer JF, Scheer SD. An analysis of substance use among adolescents from smaller places. J Rural Health. 2001; 17: 105-113. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/11573460
  23. Hartley D. Rural health disparities, population health, and rural culture. Ame J Public Health. 2004; 94: 1675-1678. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15451729
  24. Jilcott SB, Laraia BA, Evenson KR, Ammerman AS. Perceptions of the community food environment and related influences on food choice among midlife women residing in rural and urban areas: a qualitative analysis. Women Health. 2009; 49: 164-180. PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19533508
  25. Agency for Healthcare Research and Quality. National Healthcare Quality and Disparities Report chartbook on rural health care. Rockville, MD: Agency for Healthcare Research and Quality. Rockville, MD: AHRQ. 2017.