Currently Viewing:
Evidence-Based Diabetes Management March 2018
The Potential of a Population Health Strategy to Improve Healthcare Outcomes and Reduce Costs for Medicaid Programs
David J. Dzielak, PhD
How Technology Can Make CMS' Diabetes Prevention Program Viable
Dan Sheeran
Applying Digital Technology in Clinical Trials to Improve Real-World Outcomes
Henry Anhalt, DO
Currently Reading
Geographic Access to Endocrinologists for Florida's Publicly Insured Children With Diabetes
Ashby F. Walker, PhD; Jaclyn M. Hall, PhD; Elizabeth A. Shenkman, PhD; Matthew J. Gurka, PhD; Heather L. Morris, PhD; Michael J. Haller, MD; Henry J. Rohrs, MD; Kelsey R. Salazar, MPH; and Desmond A. Schatz, MD
Guidance That Allows for Higher A1C Misses the Mark
Robert A. Gabbay, MD, PhD, FACP

Geographic Access to Endocrinologists for Florida's Publicly Insured Children With Diabetes

Ashby F. Walker, PhD; Jaclyn M. Hall, PhD; Elizabeth A. Shenkman, PhD; Matthew J. Gurka, PhD; Heather L. Morris, PhD; Michael J. Haller, MD; Henry J. Rohrs, MD; Kelsey R. Salazar, MPH; and Desmond A. Schatz, MD
Enrollment, claims, and spatial data are used to demonstrate the importance of outreach strategies for families in rural areas who have children with diabetes. Spatial barriers, alone, do not fully elucidate racial/ethnic disparities in pediatric diabetes for street-level location. (For Tables and the Figure, please access the PDF on the last page.)
Enrollment files, eligibility files, and claims/encounter data were used to identify 7233 children with diabetes in Florida’s public insurance programs to examine driving times they encounter to reach in-network endocrinologists who serve publicly insured children with diabetes in Florida; the children are categorized by sociodemographic characteristics. Average driving times to pediatric endocrinologists were ≤30 minutes for children in urban areas but ≥70 minutes for children in rural communities. White children faced the longest driving times; only 56% were ≤30 minutes from a pediatric endocrinologist. These data reinforce the importance of outreach strategies for families in rural areas and demonstrate that spatial barriers, alone, do not fully elucidate racial/ethnic disparities in pediatric diabetes.for street-level location.


Youth with type 1 and type 2 diabetes (T1D and T2D) from low socioeconomic status (SES) households are at a greater risk than others for many negative health outcomes related to glycemic control, including higher hospitalization rates for very serious complications like diabetic ketoacidosis and elevated risk for diabetes-related morbidity and mortality.1-14 Moreover, race and ethnic minority status further compounds disparate outcomes in diabetes for non-Hispanic blacks and Hispanics.6-15 Despite the need for interventions to improve

health outcomes for economically vulnerable pediatric populations with T1D and T2D,16,17 there is a paucity of research that explicates barriers that may be unique to these children and their families.

In addition to basic primary care needs shared by all pediatric populations, a critical feature to achieving optimal health for children and adolescents living with diabetes is having regular access to pediatric endocrinologists. The American Diabetes Association recommends that children and adolescents with T1D and T2D visit a specialist at least four times a year.18 Children who do not meet these recommended guidelines for routine care with pediatric endocrinologists often have less-than-optimal glycemic control and higher rates of associated health risks.18,19 Moreover, though studies are limited in this area, public health insurance status has been identified as a risk factor for irregular pediatric endocrinology clinic attendance19,20 and for underuse of specialists in general, especially for non-Hispanic blacks.21

A rising scarcity of pediatric endocrinologists and a growing demand for their services22-24 compound difficulties that economically vulnerable families face in utilizing healthcare specialists who may be located considerable distances from their residences. Despite Family and Medical Leave Act protections, service sector jobs that are common among working-poor families rarely allow for adequate paid leave time; subsequently, a significant loss of income results when time away from work is taken to accommodate routine medical visits.25-27 Rural families are disproportionately poorer than urban families, and they are also at a greater disadvantage in important ways that could negatively affect their health.28,29

Adequate access to healthcare is significantly correlated with distance, an inability to obtain a driver’s license, and the lack of access to reliable transportation. All these factors negatively affect attendance of regular check-ups.30,31 Thus, the recommended standard of care of four visits to pediatric endocrinology a year presents a potential obstacle for low-SES families living with diabetes.

To better understand barriers of geographic access to pediatric endocrinologists, our study examined proximity to in-network providers of publicly insured children as a measure for access to endocrinology care among adolescents living with T1D and T2D in the state of Florida. This analysis also examined how socio-contextual factors such as urban versus rural location, race, and ethnic minority status shape geographic access as a key determinant in the complex construct of access to care. To our knowledge, there has not been a systematic attempt to document the distance that publicly insured children with diabetes in the state of Florida need to travel for access to potential endocrinologists.

Florida is one of the four largest states in the United States with significant racial and ethnic diversity, and it ranks among the top three states for the number of low-income children (those from households earning less than 200% of the Federal Poverty Level).32 Moreover, Florida has been identified as one of the top four states with persistent pronounced disparities in access to healthcare.33


This study relied on a cohort of publicly insured children from Florida’s Title XIX and XXI programs, which include Medicaid, MediKids, Children’s Medical Services Managed Care Plan (CMS), and the Florida Healthy Kids Program (FHKP or Florida’s Children’s Health Insurance Program, Title XXI), along with the 2015 provider directories for endocrinology of each program. All protocols in this study were approved by the Institutional Review Board-01 at the University of Florida and by the agencies represented in the research, including the Florida Agency for Health Care Administration and the FHKP. This study qualified as a retrospective review of existing data and operated under a waiver of informed consent. Enrollment files and eligibility files, along with claims/encounter data for each program, were used to identify children with diabetes using the following inclusion criteria: Children were defined as individuals aged 19 years or less who had any claims with International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis (primary or secondary)

code of T1D or T2D during the year. SAS 9.4 (SAS Institute, Inc; Cary, North Carolina) was used for analysis. Information about eligibility for each insurance program is provided in Table 1.

To our knowledge, a critical review of the provider directories available to publicly insured families has not been performed, nor are there studies in which directories have been carefully examined and verified prior to mapping. A systematic examination of the endocrinology provider directories available from health plans for each public health insurance

program was conducted to verify that the specialists listed were, indeed, endocrinologists, and to further confirm their credentialing. Each provider’s license and specialization were verified using both the Florida Department of Health (FL-DOH) practitioner profile search and the National Provider Identifier (NPI) registry. Providers were categorized as PE (pediatric endocrinology), GE (general/adult endocrinology), OE (other endocrinology; eg, reproductive endocrinology), or not applicable (providers whose licenses were not clear and active, who practiced in medical specialties other than endocrinology, or whose specialty could not be conclusively determined by the search[es]).

Provider addresses were entered into an online search for verification using a combination of the FL-DOH profile search and Google Maps. When registry searches yielded ambiguous results regarding addresses or specialty categorization, the U.S. News and World Report Find a Doctor34 search tool was used to make a final determination. Each provider entry was verified by a second coder who consulted both FL-DOH and NPI registries. In the few instances in which a discrepancy arose between coders, an arbiter was used to make a final call. For a point of general comparison, the total number of pediatric endocrinology providers listed as in-network for each program was compared with the number of pediatric endocrinologists practicing in each county, as determined by use of the FL-DOH directory.

Using data on race and ethnic minority status as reported by members’ families and addresses of the identified cohort available through the enrollment and eligibility files, members were geocoded using the industry-leading Navteq 2015 ESRI StreetMap Premium location software (Environmental Systems Research Institute, Inc. Redlands, CA). Geocoding is the process of determining the spatial location (latitude and longitude) of a residence from the written address.35 Of all the members in the cohort, 95% were able to be geocoded to

their street address number. Of the remaining 5%,which were able to be located only generally (at the centroid of their ZIP code), 77% of these addresses were post office boxes, and therefore not eligible for street-level location. Provider addresses were also geocoded using the verified provider directories, and all providers’ locations were successfully found geocoding.

ArcMap ESRI Network Analyst and StreetMap Premium data were used to measure the driving time from each member residence to both the closest in-network endocrinologist and to the closest in-network pediatric endocrinologist. Average drive times to a participating endocrinologist for members of each public health insurance program. Rural counties were identified as having less than 100 persons per square mile, based on the 2010 US Census, as defined by 2015 Florida Statute 381.0406 2015.37 Adequate proximity was considered to be no more than 30 minutes driving time to a provider; the 30-minute limit has been used to identify areas with poor healthcare coverage in other studies.36


A total of 7233 children in the identified cohort were mapped to available endocrinology providers:

• 4395 for Medicaid and MediKids

• 1562 for CMS

• 1276 for the FHKP

Demographic characteristics of the cohort are presented in Table 1. In brief, the cohort included the following children:

• 54% with T1D

• 46% with T2D

• Mean age of 12.2 years (±4.62)

• 47% male

• 24% white

• 31% Hispanic

• 21% black

• 3% other race/ethnicity

• 21% of unknown race/ethnicity.

The relatively high percentage of Hispanics is reflective of the overall population characteristics of the state of Florida, where approximately 29% of all children are Hispanic/Latino and the Hispanic population is the third-highest in the United States.32

Members of each program were mapped to available providers (Figures 1a-d). The distances to available providers were examined for each program, according to provider type (PE versus GE/OE, and then “any” representing the driving distance to any type of endocrinologist). The findings from the provider directory analysis were thus used to create a typology for geocoding output and to calculate proximity to a location with at least one provider. Key to this analysis was examining possible variations in driving distances depending on whether a member lived in a rural or urban location, and to see how this relationship between distance and location type varied by race and ethnic minority status (Table 2 and Table 3). As expected, rural populations represented a smaller proportion of the overall cohort and were more commonly non-Hispanic white. Ninety-five percent of non-Hispanic

black members resided in urban counties, and the members in densely urban, southeast Florida were 60% Hispanic.

In all, members in urban areas for all programs tended to have similar driving times to pediatric endocrinologists or to any endocrinologist. For urban members in the FHKP, driving time to any endocrinologist was an average of 12 minutes (13 to a pediatric endocrinologist); for CMS members, driving time was 29 minutes (30 to a pediatric endocrinologist);

Copyright AJMC 2006-2019 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
Welcome the the new and improved, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up