Dental utilization in the MPW program ranged from 1 percent to 26 percent with a median of 8.5 percent. A major focus of this study was to determine whether characterization of the context of rurality identifies differences in health utilization rates for dental services. Therefore, we began by presenting basic statistics in Table 1. It is notable that there is no clear pattern when it comes to PUR by RUCC, except for the fact that the last category has the highest level of variability, (RUCC=9) rural areas not adjacent to a metro area. We note that all 95% confidence intervals overlap indicating that statistically, there are no pairwise mean differences.
Table 1 Percent Utilization Rate (PUR) for dental services by eligible women in the Medicaid for Pregnant Women Program by RUCC for North Carolina Counties 2014-2016.
|
Rural Urban Continuum Code Description
|
Mean PUR
|
Std
|
N (counties)
|
Min
|
Max
|
95% Confidence Interval
|
|
(1) Metro areas of 1 million or more
|
9.06
|
2.61
|
12
|
4.79
|
15.21
|
(7.40
|
10.72)
|
|
(2) Metro areas of 250,000 to 1 million
|
8.56
|
2.57
|
25
|
4.57
|
13.60
|
(7.50
|
9.62)
|
|
(3) Metro areas of fewer than 250,000
|
10.07
|
3.10
|
9
|
7.02
|
15.57
|
(7.69
|
12.46)
|
|
(4) Urban areas of 20,000 or more, adjacent to metro
|
9.83
|
4.33
|
15
|
4.72
|
18.65
|
(7.43
|
12.23)
|
|
(5) Urban areas of 20,000 or more, not adjacent to metro
|
8.05
|
3.32
|
2
|
5.70
|
10.39
|
(-21.73
|
37.83)
|
|
(6) Urban areas of 2,500 to 19,999, adjacent to metro
|
8.99
|
2.78
|
16
|
4.53
|
14.56
|
(7.51
|
10.47)
|
|
(7) Urban areas of 2,500 to 19,000, not adjacent to metro
|
8.83
|
4.54
|
5
|
5.82
|
16.71
|
(3.19
|
14.47)
|
|
(8) Rural areas of or less than 2,500, adjacent to metro
|
8.22
|
3.21
|
9
|
3.39
|
13.50
|
(5.57
|
10.69)
|
|
(9) Rural areas of or less than 2,500, not adjacent to metro
|
13.18
|
8.58
|
7
|
0.97
|
26.19
|
(5.24
|
21.11)
|
|
|
|
|
|
|
|
|
|
|
|
Note: RUCC is Rural Urban Continuum Code (USDA).
Figure 1 presents an overall spatial pattern of PUR in the state of North Carolina using a map of counties. It is notable that the spatial distribution is not random, with clusters of relatively high utilization rate in the western and south eastern parts of the state. On the other hand, central and north eastern counties have lower percent utilization rates with few exceptions.
In order to have a better understanding of RUCC we also present a map showing the spatial distribution of this classification in the state (Figure 2). We divided the sample into 2 groups based on metro status with Metro Counties (RUCC <=3) and Non-metro counties (RUCC >=4). Such classification is warranted because the RUCC scheme is intended to highlight proximity to services that are generally concentrated in metro areas.
Descriptive statistics for variables used in our analysis are presented in Table 2. Note that not all variables are available in all 100 counties. There are two observations missing for High School Graduation rate and Poor Mental Health days. These counties have small populations and the data sources did not provide values for them. Even though those missing values limit our sample, sensitivity analysis revealed that our baseline results remain essentially the same and there is no systematic pattern for those omitted observations reducing risk of bias in our statistical analysis.
Table 2. Descriptive Statistics in Full Sample (all counties) and by Non-Metro and Metro classification.
|
Variable
|
Mean
|
Std
|
N
|
Min
|
Max
|
95% Confidence Interval
|
|
|
FULL SAMPLE
|
|
|
|
|
|
|
|
|
|
Percent Utilization Rate (PUR) dental care
|
9.31
|
3.83
|
100
|
0.97
|
26.19
|
8.55
|
10.07
|
|
|
Rural Urban Continuum Code (RUCC)
|
4.25
|
2.54
|
100
|
1
|
9
|
3.74
|
4.75
|
|
|
Group 1 Predisposing Factors
|
|
|
|
|
|
|
|
|
|
Farming/Recreation
|
0.24
|
0.43
|
100
|
0
|
1
|
0.15
|
0.32
|
|
|
Persistent Poverty
|
0.10
|
0.30
|
100
|
0
|
1
|
0.04
|
0.16
|
|
|
High School Graduation Rate
|
81.59
|
4.97
|
98
|
68.00
|
92.50
|
80.59
|
82.58
|
|
|
Group 2 County Well-being
|
|
|
|
|
|
|
|
|
|
Abuse reporting past year
|
63.68
|
23.57
|
100
|
18.75
|
157.86
|
59.00
|
68.35
|
|
|
Change in Abuse reporting
|
-1.40
|
10.77
|
100
|
-52.03
|
40.46
|
-3.54
|
0.73
|
|
|
Infant Mortality
|
7.83
|
4.76
|
100
|
0.00
|
24.10
|
6.88
|
8.77
|
|
|
Length of Life County Rank
|
50.50
|
29.01
|
100
|
1
|
100
|
44.74
|
56.26
|
|
|
Poor Mental Health Days in past month
|
3.65
|
0.86
|
93
|
2.00
|
6.20
|
3.47
|
3.82
|
|
|
Group 3 Dental Health Services
|
|
|
|
|
|
|
|
|
|
Dentists per 100,000 pop
|
36.27
|
24.12
|
100
|
0
|
184.54
|
31.48
|
41.06
|
|
|
FQHC Dental in County
|
0.28
|
0.45
|
100
|
0
|
1
|
0.19
|
0.37
|
|
|
NON-METRO COUNTIES (RUCC >= 4)
|
|
|
|
|
|
|
|
|
|
Percent Utilization Rate (PUR) dental care
|
9.59
|
4.59
|
54
|
0.97
|
26.19
|
8.33
|
10.84
|
|
|
Rural Urban Continuum Code (RUCC)
|
6.22
|
1.76
|
54
|
4
|
9
|
5.74
|
6.7
|
|
|
Group 1 Predisposing Factors
|
|
|
|
|
|
|
|
|
|
Farming/Recreation
|
0.30
|
0.46
|
54
|
0
|
1
|
0.17
|
0.42
|
|
|
Persistent Poverty
|
0.17
|
0.38
|
54
|
0
|
1
|
0.06
|
0.26
|
|
|
High School Graduation Rate
|
80.99
|
5.46
|
52
|
68.00
|
92.50
|
79.47
|
82.51
|
|
|
Group 2 County Well-being
|
|
|
|
|
|
|
|
|
|
Abuse reporting past year
|
67.38
|
27.22
|
54
|
18.75
|
157.86
|
59.95
|
74.8
|
|
|
Change in Abuse reporting
|
-1.64
|
13.50
|
54
|
-52.03
|
40.46
|
-5.33
|
2.04
|
|
|
Infant Mortality
|
8.34
|
5.88
|
54
|
0.00
|
24.10
|
6.73
|
9.94
|
|
|
Length of Life County Rank
|
59.87
|
28.54
|
54
|
3
|
100
|
52.08
|
67.66
|
|
|
Poor Mental Health Days in past month
|
3.66
|
0.93
|
50
|
2.00
|
5.70
|
3.39
|
3.92
|
|
|
Group 3 Dental Health Services
|
|
|
|
|
|
|
|
|
|
Dentists per 100,000 pop
|
31.03
|
16.35
|
54
|
0
|
70.14
|
26.57
|
35.49
|
|
|
FQHC Dental in County
|
0.26
|
0.44
|
54
|
0
|
1
|
0.14
|
0.38
|
|
|
METRO COUNTIES (RUCC <= 3)
|
|
|
|
|
|
|
|
|
|
Percent Utilization Rate (PUR) dental care
|
8.99
|
2.69
|
46
|
4.57
|
15.5689
|
8.18
|
9.78
|
|
|
Rural Urban Continuum Code (RUCC)
|
1.93
|
0.68
|
46
|
1
|
3
|
1.73
|
2.13
|
|
|
Group 1 Predisposing Factors
|
|
|
|
|
|
|
|
|
|
Farming/Recreation
|
0.17
|
0.38
|
46
|
0
|
1
|
0.06
|
0.29
|
|
|
Persistent Poverty
|
0.02
|
0.15
|
46
|
0
|
1
|
-0.02
|
0.07
|
|
|
High School Graduation Rate
|
82.27
|
4.32
|
46
|
72.35
|
90.20
|
80.98
|
83.55
|
|
|
Group 2 County Well-being
|
|
|
|
|
|
|
|
|
|
Abuse reporting past year
|
59.34
|
17.71
|
46
|
31.18
|
107.67
|
54.08
|
64.6
|
|
|
Change in Abuse reporting
|
-1.12
|
6.35
|
46
|
-22.45
|
12.35
|
-3.00
|
0.76
|
|
|
Infant Mortality
|
7.23
|
2.90
|
46
|
0.00
|
12.40
|
6.37
|
8.09
|
|
|
Length of Life County Rank
|
39.50
|
25.76
|
46
|
1
|
90
|
31.85
|
47.15
|
|
|
Poor Mental Health Days in past month
|
3.65
|
0.79
|
43
|
2.10
|
6.20
|
3.40
|
3.89
|
|
|
Group 3 Dental Health Services
|
|
|
|
|
|
|
|
|
|
Dentists per 100,000 pop
|
42.42
|
29.89
|
46
|
0
|
184.54
|
33.54
|
51.3
|
|
|
FQHC Dental in County
|
0.30
|
0.47
|
46
|
0
|
1
|
0.17
|
0.44
|
|
Notes on Variables:
County Codes from US Department of Agriculture (USDA)
RUCC is the Rural Urban Continuum Code system.8 Codes range from 1 to 9. This is a classification scheme developed by the USDA. It attempts to arrange counties on a continuum from most metro to least metro based on population density and proximity of non-metro counties to urban areas.
County Typology Codes9 for Persistent Poverty, Farming, Recreation
Persistent Poverty is a county-level designation that identifies counties where at least 20 percent of the population is at or below the Federal Poverty Level in each of the census years of 1980, 1990, 2000 and the American Community Survey of 2013. In North Carolina, 10 counties are designated as areas of Persistent Poverty. Persistent Poverty is often a marker that alerts providers of direct and indirect services to the increased vulnerability of the population especially low educational attainment and poor health outcomes.
Farming/Recreation is derived to represent counties that were designated as either Farming or Recreation. Farming is the Farm-dependent county indicator, where 0=no 1=yes. Farming accounted for at 25% or more of the county's earnings or 16% or more of the employment averaged over 2010-2012.
Recreation defines counties (0=no 1=yes) based on computation from three data sources: 1) Percentage of wage and salary employment in entertainment and recreation, accommodations, eating and drinking places, and real estate as a percentage of all employment reported by the Bureau of Economic Analysis; 2) Percentage of total personal income reported for these same categories by the Bureau of Economic Analysis; and 3) Percentage of vacant housing units intended for seasonal or occasional use reported in the 2010 Census. The three variables measuring employment, earnings, and seasonal housing were converted to z-scores and combined into a weighted index (weights of 0.3 were assigned to income and employment and 0.4 to seasonal housing) to reflect recreational activity. Counties with index scores of 0.67 or higher were regarded as recreation counties. Seasonal housing was given a higher weight because in some areas employment and income may not reflect recreational activity because of the seasonality. The comparison group was all other counties and these were designated as either Manufacturing, Government, or Non-specialized.
Variables from NC Child (http://www.ncchild.org/)
Abuse and Neglect claims in 2015 and 2016. The rate per 1000 of children under age 18 who were assessed for abuse or neglect.
Infant Mortality in 2015. This represents the number of infant deaths per 1000 babies born alive.
North Carolina County Health Rankings 2015 variables (http://www.countyhealthrankings.org/rankings/data/NC)
Length of Life – County rank among counties in North Carolina (from 1 to 100, with 100 being worst) for age-adjusted Years of Potential Life Lost before age 75, calculated from 2010-2012 Mortality files National Center for Health Statistics.
Quality of Life - from BRFSS 2006-2012, measured as county-level average number of mentally unhealthy days (poor mental health days) in the past 30 days (age-adjusted). Poor mental health days ranged from 2.0 to 6.2 days per month, median of 3.6.
High School Graduation Rate = (percent) number students that graduated divided by the number of students expected to graduate from high school in 2015
Dentist rate = number of dentists per 100,000 population in 2015
North Carolina Department of Health and Human Services variable
(https://publichealth.nc.gov/oralhealth/services/safetynetclinics.htm)
FQHC with Dental in County = indicator variable for presence of a Federally Qualified Health Center (FQHC) that provides dental services located in the county.
In general, non-metro counties have a higher percent of counties classified as Farming/Recreational, classified as places with Persistent Poverty, more Child Abuse Reporting events, a higher Infant Mortality rate, worse rank for Length of Life among counties, and both a lower number of Dentists per population and a lower High School Graduation Rate.
Regression results
Our multivariate regression results are presented in Table 3. To highlight differences between metro and non-metro counties we estimated the same model for those two subpopulations.
The approach is validated by the fact that even though the R-square in full model is 0.488, the explained variation in the non-metro model is 0.618 versus metro with 0.390. Therefore, we conclude that metro versus non-metro areas may be influenced by two different processes, where non-metro populations have less random variation and thus any policy interventions based on our independent variables may be more effective.
ble 3. Multivariate Regression Assessing Determinants of Percent Utilization Rate Among MPW Eligible Women (Dependent variable) Calculated from Pooled Sample During Years 2014-2016, in North Carolina Counties, by Metro vs. Non-metro classification.
|
|
Non-metro Counties
|
Metro Counties
|
Full Sample
|
Variable
|
Coeff
|
95% Confidence Interval
|
Coeff
|
95% Confidence Interval
|
Coeff
|
95% Confidence Interval
|
Intercept
|
-18.26
|
-37.49, 0.97
|
-12.97
|
-29.16, 3.23
|
-16.26
|
-28.40, -4.12
|
Group 1 Predisposing Factors
|
|
|
|
|
|
|
Farming/Recreation
|
1.53
|
-1.37, 4.44
|
-0.002
|
-2.33, 2.32
|
1.61**
|
0.03, 3.20
|
Persistent Poverty
|
-4.29**
|
-7.38, -1.22
|
-0.81
|
-6.22, 4.59
|
-3.11**
|
-5.46, -0.76
|
High School Grad Rate
|
0.20*
|
-0.02, 0.43
|
0.18*
|
-0.009, 0.37
|
0.19**
|
0.04, 0.33
|
Group 2 County Well-being (need)
|
|
|
|
|
|
|
Abuse reporting past year
|
0.003
|
-0.05, 0.05
|
0.03
|
-0.02, 0.08
|
0.01
|
-0.02, 0.05
|
Change in Abuse reporting
|
0.16***
|
0.08, 0.25
|
-0.03
|
-0.19, 0.13
|
0.13***
|
0.07, 0.20
|
Infant Mortality
|
0.24**
|
0.07, 0.42
|
-0.09
|
-0.39, 0.22
|
0.18**
|
0.04, 0.32
|
Length of Life County Rank
|
0.05**
|
0.01, 0.09
|
0.03
|
-0.01, 0.07
|
0.04**
|
0.01, 0.07
|
Poor Mental Health Days
|
1.26**
|
0.15, 2.38
|
1.03*
|
-0.17, 2.23
|
1.35***
|
0.59, 2.11
|
Group 3 Dental Health Services
|
|
|
|
|
|
|
Dentists per 100,000 pop
|
0.05
|
-0.02, 0.12
|
0.02
|
-0.008, 0.05
|
0.03**
|
0.004, 0.06
|
FQHC Dental in County
|
1.28
|
-0.86, 3.42
|
1.31
|
-0.47, 3.09
|
1.28*
|
-0.09, 2.65
|
|
|
|
|
|
|
|
R2
|
0.63
|
0.43
|
0.51
|
Adj R2
|
0.54
|
0.25
|
0.45
|
N
|
50
|
43
|
93
|
|
|
|
|
|
|
|
Note: * P < .10; ** P < .05; *** P < .001.
|
Notes on Variables:
County Codes from US Department of Agriculture (USDA)
RUCC is the Rural Urban Continuum Code system.8 Codes range from 1 to 9. This is a classification scheme developed by the USDA. It attempts to arrange counties on a continuum from most metro to least metro based on population density and proximity of non-metro counties to urban areas.
County Typology Codes9 for Persistent Poverty, Farming, Recreation
Persistent Poverty is a county-level designation that identifies counties where at least 20 percent of the population is at or below the Federal Poverty Level in each of the census years of 1980, 1990, 2000 and the American Community Survey of 2013. In North Carolina, 10 counties are designated as areas of Persistent Poverty. Persistent Poverty is often a marker that alerts providers of direct and indirect services to the increased vulnerability of the population especially low educational attainment and poor health outcomes.
Farming/Recreation is derived to represent counties that were designated as either Farming or Recreation. Farming is the Farm-dependent county indicator, where 0=no 1=yes. Farming accounted for at 25% or more of the county's earnings or 16% or more of the employment averaged over 2010-2012.
Recreation defines counties (0=no 1=yes) based on computation from three data sources: 1) Percentage of wage and salary employment in entertainment and recreation, accommodations, eating and drinking places, and real estate as a percentage of all employment reported by the Bureau of Economic Analysis; 2) Percentage of total personal income reported for these same categories by the Bureau of Economic Analysis; and 3) Percentage of vacant housing units intended for seasonal or occasional use reported in the 2010 Census. The three variables measuring employment, earnings, and seasonal housing were converted to z-scores and combined into a weighted index (weights of 0.3 were assigned to income and employment and 0.4 to seasonal housing) to reflect recreational activity. Counties with index scores of 0.67 or higher were regarded as recreation counties. Seasonal housing was given a higher weight because in some areas employment and income may not reflect recreational activity because of the seasonality. The comparison group was all other counties and these were designated as either Manufacturing, Government, or Non-specialized.
Variables from NC Child (http://www.ncchild.org/)
Abuse and Neglect claims in 2015 and 2016. The rate per 1000 of children under age 18 who were assessed for abuse or neglect.
Infant Mortality in 2015. This represents the number of infant deaths per 1000 babies born alive.
North Carolina County Health Rankings 2015 variables (http://www.countyhealthrankings.org/rankings/data/NC)
Length of Life – County rank among counties in North Carolina (from 1 to 100, with 100 being worst) for age-adjusted Years of Potential Life Lost before age 75, calculated from 2010-2012 Mortality files National Center for Health Statistics.
Quality of Life - from BRFSS 2006-2012, measured as county-level average number of mentally unhealthy days (poor mental health days) in the past 30 days (age-adjusted). Poor mental health days ranged from 2.0 to 6.2 days per month, median of 3.6.
High School Graduation Rate = (percent) number students that graduated divided by the number of students expected to graduate from high school in 2015
Dentist rate = number of dentists per 100,000 population in 2015
North Carolina Department of Health and Human Services variable
(https://publichealth.nc.gov/oralhealth/services/safetynetclinics.htm)
FQHC with Dental in County = indicator variable for presence of a Federally Qualified Health Center (FQHC) that provides dental services located in the county.
The variables for Pre-disposing factors (Group 1) have expected signs and magnitudes with notable differences between metro versus non-metro counties. Farming/Recreation classification only matters for the non-metro counties and same is true for the Persistent Poverty indicator (it is notable that in metro counties those coefficients are effectively zero). High School Graduation rate appears to have slightly more explanatory power in the metro counties.
Factors that represent county-level need are measures of well-being that are relevant for the maternal-child population (Group 2). These factors present very interesting differences between two subpopulations. The variable Change in Abuse Reporting, which in combination with Abuse Reporting Past Year, provides an indication of how the dynamics of abuse reporting affects PUR while holding levels of Abuse Reporting the same. It only matters for the non-metro counties. The same is true for the Infant Mortality variables. On the other hand, Length of Life Ranking and Poor Mental Health Days appear to be important in both models.
The indicators for Dental Health Services (Group 3) show that the impact of the Dentist per population variable appears to be larger in non-metro areas but since it is not statistically significant and the confidence intervals overlap, we cannot determine in which subpopulation it matters more. Whether a county has an FQHC that provides dental services has a similar positive impact in all three models.
The non-metro model shows some strong relationships. Counties with dominant Farming/Recreation economies have a PUR that is higher by about 1.5 percentage points, compared to counties classified otherwise. Counties characterized by Persistent Poverty have PUR that is 4.3 percentages points lower than average which is relatively large since the mean PUR in our sample is about 9.3. A Dentist per population coefficient of 0.05 means that having one more dentist per 100,000 population improves PUR by about 0.05 percentage points. To put this into perspective, if one would double the number of dentists in a non-metro county of 150,000 population from 30 to 60 dentists (which seems unlikely) this would improve utilization by about 1 percentage point. The High School Graduation Rate coefficient is 0.20 which means that the improvement between the worst (68%) and the best county (92.5%) in North Carolina would amount to an improvement in PUR by about 5 percentage points- which is substantial.