|
|
||||||||
|
J Am Dent Assoc, Vol 139, No 12, 1657-1666.
© 2008 American Dental Association |
TRENDS |
A study of the Wisconsin Medicaid program
| ABSTRACT |
|---|
|
|
|---|
Methods. The authors analyzed all Medicaid dental claims in Wisconsin from 2001 through 2003 to examine factors associated with NTDC visits to EDs and POs. They performed bivariate and multivariable analyses. The independent variables they examined included race/ethnicity, age, sex, dental health professional shortage area (DHPSA) designation and urban influence code for county of residence.
Results. The authors evaluated 956,774 NTDC visits made during 1,718,006 person-years; 4.3 percent of visits occurred in EDs or POs. Native Americans, African-Americans and enrollees of unknown race/ethnicity had the highest unadjusted ED and PO visit rates for NTDCs. African-Americans, Native Americans, adults and residents in partial or entire DHPSAs had significantly higher adjusted rates of NTDC visits to EDs. The authors observed significantly higher adjusted NTDC visit rates to POs for Native Americans, adults and enrollees residing in entire DHPSAs, and a significantly lower adjusted rate among African-Americans.
Conclusions. Native Americans, those residing in entire DHPSAs and adults have significantly higher risks of NTDC visits to EDs and POs. African-Americans are at increased risk of making visits to EDs for NTDCs but at decreased risk of making visits to POs for NTDCs.
Clinical Implications. Reductions in Medicaid visits to EDs and POs and the associated costs might be achieved by improving dental care access and targeted educational strategies among minorities, DHPSA residents and adults.
Key Words: Medicaid; racial and ethnic disparities; nontraumatic dental conditions; emergency department; dental health professional shortage area
Abbreviations: DHPSA: Dental health professional shortage area. ED: Emergency department. ICD-9-CM: International Classification of Diseases, Ninth Revision, Clinical Modification. NTDC: Nontraumatic dental condition. PO: Physician office. UIC: Urban influence code.
The Medicaid program is the largest source of public funding for medical and dental services in the United States.1,2 It covers millions of people who are disabled, elderly, blind and pregnant, as well as children classified as poor on the basis of federal poverty level guidelines.1,2 In 2004, 41.3 million Americans were insured through Medicaid.3 Under Medicaid, the Early Periodic Screening, Diagnosis, and Treatment Program requires that basic dental services be covered for children1 but not for adults. However, basic dental services are covered for adults in some states such as Wisconsin.
Visits to emergency departments (EDs) in the United States increased by 26 percent from 1993 to 2003.4 Three million visits to EDs in the United States from 1997 through 2000 involved dental complaints, with an average of 738,000 visits annually.5 National data for 1999 through 2000 suggested that dental problems accounted for 0.2 percent of visits to physician offices (POs) and 0.9 percent of visits to EDs.6 Dental conditions are best treated in dental offices, where appropriate procedures can be performed and continuity of care can be provided regularly by a dental care provider. Most patients who go to EDs or POs for treatment of dental conditions receive temporary treatment, antibiotics or analgesics, and referrals for follow-up with a dental care provider.5 The transitory nature of dental care received in EDs and POs has important cost implications because the Medicaid system is made to pay for largely unnecessary visits to EDs and POs, and in states in which dental care for adults is covered, Medicaid is made to pay for subsequent visits to dental care providers.
Although dental care needs are best addressed in primary dental care settings,7 barriers exist to accessing primary dental care, especially for Medicaid enrollees. These barriers include geographic maldistribution of dentists (only 6 percent of dental needs are met in the 1,198 health professional shortage areas in the United States8), inadequate numbers of dentists treating Medicaid enrollees9,10 and low Medicaid reimbursement rates.9,11 These barriers persist despite overall improvement in the oral health of most Americans, as the poor and racial/ethnic minority groups (which include multiracial people) are disproportionately affected by dental disease.8 The use of EDs and POs for nontraumatic dental conditions (NTDCs) has received limited attention.12–14 In particular, not enough is known about factors associated with NTDC visits to EDs and POs among Medicaid enrollees. Therefore, we conducted a study to identify factors associated with NTDC visits to EDs and POs among Medicaid enrollees.
Definition of NTDC.
We defined NTDCs as any claim with the following ICD-9-CM codes: 521.0–521.9 (diseases of dental hard tissues of teeth), 522.0–522.9 (diseases of pulp and periapical tissues), 523.0–523.9 (gingival and periodontal diseases), 525.3 (retained dental root), 525.9 (unspecified disorder of the teeth and supporting structures) and 873.63 (internal structures of mouth, without broken tooth).15 These ICD-9-CM codes for NTDC visits are identical to those used in other published studies in which dental visits to EDs and POs were analyzed.7,12
County-level information.
We used county of residence at the time of the claim to define two county-level variables: dental health professional shortage area (DHPSA) designation and urban influence code (UIC). DHPSA and UIC were classified according to county (there are 72 counties in Wisconsin) rather than by ZIP code. The federal government developed the DHPSA designation to meet the need for better programs in communities with high unmet dental needs.16,17 In our study, we used the classification from the Wisconsin Department of Health and Family Services oral health report.18 Under this classification, a county is considered an entire DHPSA when it meets the criteria established by the Bureau of Health Professionals, which are "that the area should be a rational area for the delivery of dental services; the population to full-time dentist ratio should be less than 5,000:1 but greater than 4,000:1 and has unusually high needs for dental services or insufficient capacity of existing dental care providers; or dental professionals in the contiguous area are overutilized, or excessively distant, or inaccessible to the population of the area under consideration."17
The 2003 UICs, published by the U.S. Department of Agriculture, were used as a measure of the rurality of the county of residence for each enrollee making a claim.19 The UICs use population and commuting data from the 2000 census to categorize the 3,141 U.S. counties and county equivalents into 12 classification levels. For the purposes of this study, we used only the three major classification levels: metropolitan, micropolitan (an area with at least one urban cluster of at least 10,000 residents but fewer than 50,000 residents) and rural.
Race/ethnicity, age and missing data.
Enrollees self-designated race/ethnicity was available from the state database in the following categories: white, African-American, Latino, Asian/Pacific Islander, Native American, other and unknown. We dichotomized age groups into two groups: children (
Database creation process.
The database for those enrolled in Medicaid in Wisconsin from 2001 through 2003 included 1,317,811 claims (Figure
![]()
METHODS
TOP
ABSTRACT
METHODS
RESULTS
DISCUSSION
CONCLUSIONS
References
Data source.
We used the electronic data systems of the Medicaid Evaluation and Decision Support database for Wisconsin from 2001 through 2003. This database contains all Medicaid claims for the state of Wisconsin and is managed by the Division of Health Care Financing in the Wisconsin Department of Health and Family Services. The database includes information on age; sex; date of service; International Classification of Diseases, Ninth Revision, Clinical Modifications (ICD-9-CM) codes for the dental condition; race/ethnicity; and county of residence.
18 years) and adults (> 18 years). We excluded visits with missing information on age or county of residence.
). Of these, 1,235,211 records were claims from dental office visits, and 82,600 were claims from visits to EDs and POs. We merged claims by patient identification number, date of service and site of visit to account for multiple charges during a single visit. For ED claims, we used the facility claim instead of the claim from the attending physician. After we removed multiple charges for the same visit, records with no information on county of residence or age and claims with ICD-9-CM codes not classified as NTDCs, the number of records was reduced to 915,190 dental office visits and 41,584 visits to EDs and POs for final analysis.
|
We also examined potential interactions of race/ethnicity with DHPSA, race/ethnicity with UIC, and DHPSA with UIC. Although we noted a few small, statistically significant interactions, they did not alter the main significant findings in multivariable models. In addition, these interactions were inconsistent, only accounted for a small amount of variance in models, may reflect large sample sizes, have limited dental public health implications and may produce models that are overfitted. Therefore, we chose to omit the findings of multivariable models that included interaction terms.
All analyses included a DHPSA classification and UIC based on county of residence at the end of 2003 for patients whose residence was missing at the time of the visit. Because these subjects could bias the results by having moved between the time of the claim and the end of 2003, we also performed all analyses excluding these subjects. The results were similar to those presented and are not shown. We used statistical software (SAS, Version 9.1, SAS Institute, Cary, N.C.) to perform all statistical analyses. We used an
level of 0.05 throughout to denote statistical significance.
| RESULTS |
|---|
|
|
|---|
|
Among ED and PO NTDC claims, unspecified disorder of the teeth and supporting structures was the most frequent ICD-9-CM code, accounting for 45.3 percent of visits; dental caries (24.8 percent) and periapical abscesses (19.2 percent) were the next most common codes (data not shown). We observed differences in diagnostic codes according to visit site and age (all of which were statistically significant; data not shown). Unspecified disorder of the teeth and supporting structures and dental caries constituted 31.6 percent and 22.4 percent of ED pediatric claims, but 16.5 percent and 43.0 percent of PO pediatric claims, respectively. Among adults, 55.4 percent of visits to EDs compared with 38.3 percent for visits to POs were for unspecified disorder of the teeth and supporting structures. The second most common diagnosis for adult visits to EDs was dental caries (24.8 percent), and the second most common diagnosis for adult visits to POs was periapical abscesses (27.3 percent).
Bivariate analysis of factors associated with NTDC visits.
NTDC visit rates were 242 per 10,000 person-years of enrollment to EDs or POs and 5,327 per 10,000 person-years to dental offices (Table 2
). Males had significantly lower unadjusted rates than did females for NTDC visits both to EDs and POs and to dental offices. Compared with children, adults had almost four times the NTDC visit rate to EDs and POs but only a slightly higher NTDC visit rate to dental offices. African-Americans, Latinos and Asian/Pacific Islanders had significantly lower NTDC visit rates to dental offices than did whites, but Asian/Pacific Islanders were the only minority group with an NTDC ED or PO visit rate that significantly differed from that of whites. Among NTDC visits to dental offices, significantly lower rates were found for enrollees residing in partial DHPSAs. We found higher NTDC visit rates to dental offices for micropolitan and rural enrollees, but we noted no other significant findings for any other comparisons.
|
|
Multivariable analyses of factors associated with NTDC visits to EDs and POs.
Several factors were significantly associated with NTDC visit rates to EDs after adjustment in multivariable analyses (Table 4
). Males had lower adjusted rates than did females, but the adjusted rate for adults was more than four times that of children. Compared with whites, Native Americans and African-Americans had double the adjusted rates of NTDCs to EDs, whereas Asian/Pacific Islanders had an adjusted rate that was five times lower. We noted higher adjusted rates for partial and entire DHPSAs and a lower rate for rural enrollees.
|
| DISCUSSION |
|---|
|
|
|---|
Racial/ethnic disparities. We identified specific, substantial racial/ethnic disparities in NTDC visits to EDs and POs, which persisted after we made adjustments for relevant covariates. Compared with whites, Native Americans had double the rate of NTDC visits to EDs and POs, African-Americans had about double the rate of NTDC visits to EDs, and enrollees of other race/ethnicity had about double the rate of NTDC visits to POs.
These findings expand the growing list of racial/ethnic disparities in oral health and dental care and are consistent with recent national data documenting that Native American, African-American and multiracial children have higher adjusted odds of unmet dental care needs and of having had no routine preventive dental visit in the past year.24 These new data are of particular concern, given the literature documenting that minorities have the highest risk of developing early childhood caries,25 dental caries,26 edentulism8 and oral cancer.8 Minority children have a higher risk of experiencing untreated dental disease,27 and this lack of treatment is associated with 52 million missed school hours per year,28 poor self-esteem and an impaired ability to learn.29
Two findings from our study merit further comment regarding lower rates of ED and PO use for certain minority groups. Compared with whites, African-Americans had lower adjusted rates of NTDC visits to POs but higher rates of NTDC visits to EDs, which could be related to findings that African-Americans are less likely to have a usual source of care and more likely to make visits to EDs.30 Asian/Pacific Islanders had significantly lower rates of NTDC visits to EDs and POs but also had lower NTDC visit rates to dental offices. Low use of EDs and POs among Asian/Pacific Islanders is consistent with findings from studies in the medical literature documenting lower use of medical services in general.31–33 Possible explanations for this phenomenon might be better health status, language barriers to care, cultural norms regarding the need for care and distrust of the Western medical system.32,33 These results suggest that more research is warranted on use of dental services among Asian/Pacific Islanders.
DHPSA findings. We found that DHPSA designation was associated with NTDC visit rates to EDs and POs, and we noted that there were significantly higher rates for both ED and PO visits in entire DHPSAs and for ED visits in partial DHPSAs. DHPSA designation is a prerequisite for participation in state and federal programs that are designed to improve access to dental care.17 Study results have shown that higher numbers of racial/ethnic minorities live in health profession shortage areas.34 Although DHPSA designation is based mainly on dentist-to-population ratios and unmet dental needs of the population,17 it appears to be a useful measure in deciding where to allocate resources aimed at reducing NTDC visits to EDs and POs. Our results suggest that a reduction in NTDC visits to EDs and POs might be achieved by establishing greater access to dental homes (sites providing care that is accessible, continuous, comprehensive, coordinated, compassionate, family-centered and culturally effective) in DHPSAs.
Implications for policy and practice. Our study findings indicate possible problems in the Medicaid system, as well as how, when and where enrollees seek appropriate dental care services for NTDCs. Although it may be impossible to eliminate NTDC visits to EDs and POs completely, even a small reduction in such visits might result in substantial cost savings and improvements in both dental and health care. For example, reducing only the ED component of Wisconsin Medicaid NTDC visits to EDs and POs by 1 percent from 4.3 percent to 3.3 percent of all claims (equivalent to a reduction of 10,010 claims in the three-year study period [41,584 – 31,574]) could result in estimated savings of more than $6.1 million for Medicaid (if one uses published national data documenting an average expense of $637 per visit to an ED with one or more nonsurgical services35 and if one deducts the Medicaid expense of $25 for a dental office examination and periapical radiograph36), while simultaneously increasing the use of dental homes and decreasing the inappropriate use of EDs.
A small reduction in NTDC visit rates to EDs and POs could yield considerable cost savings and improvements in delivery of health and dental care, but additional studies are needed to identify specific mechanisms that are effective in achieving such rate reductions. Although it can be argued that the NTDC visit rate to EDs and POs of 4.3 percent is low, this rate is equivalent to one in 25 Medicaid enrollees with NTDC visits to EDs and POs. Further research is warranted regarding whether such NTDC visits to EDs and POs can be reduced by means of using interventions such as targeted patient education about use of dental offices for NTDCs and increasing Medicaid reimbursement rates to dentists for NTDC visits and regular dental care.
One practical step health professionals and policymakers might take to reduce NTDC visits to EDs and POs is to develop and implement programs that target expansion of the dental work-force and dental homes in minority and DHPSA communities. This step could include establishing oral health triage centers with expanded hours within primary care practices, especially in inner-city communities with the greatest unmet dental care needs. Expanded-duty auxiliary dental personnel, who are supervised by dentists, could staff these centers. Another measure is to provide tax benefits or start-up funds to dentists who are willing to set up practices in underserved communities that have high unmet dental needs. In addition, the Wisconsin Department of Health and Family Services should establish Medicaid dental outreach hotlines to link enrollees with a dental home in their community. These opportunities allow patients, health professionals and policy-makers to work together to improve access to dental care on a systemwide basis. Finally, public health agencies could explore the use of social marketing tools in teaching Medicaid enrollees about the importance of seeking care for NTDC in dental offices rather than in EDs or POs.
Limitations. Certain study limitations should be noted. We only examined patients enrolled in Wisconsin Medicaid, so the results may not pertain to other states. Nevertheless, we believe that the study findings may be generalized to apply to other states that provide coverage to adults for dental conditions, because Wisconsin, like many other states, has a racially/ethnically diverse population that includes both urban and rural poor.
We used Medicaid claims data, which can be subject to coding errors, multiple sites of residence, multiple visits, inconsistently recorded provider information and missing data caused by each of these factors. Of the original 1,317,811 claims considered for the analyses, there were 320,021 dental office visit exclusions and 41,016 ED or PO exclusions because of missing data or multiple claims for a single visit. This finding also must be viewed as a potential limitation, because inclusion of these visits (if we were able to obtain this missing information) might have yielded different findings.
We used DHPSA designation as a proxy for access to care. A better measure might have been the availability of dentists who accept Medicaid in the patients coverage area. Additional research with this measure is needed to examine NTDC visits to EDs and POs.
Another limitation is that our analyses could not account for the potentially clustered nature of the data. It was not possible for us to conduct an individual-level analysis that would account for the correlation between visits made by the same person because enrollment data were available only in an aggregate form within each county. Therefore, it was not possible for us to track the enrollment of each person from month to month. This limitation could lead to an underestimation of standard errors and to an overstatement of statistical significance, and caution must be exercised in interpreting results that are at the lower limit of the 95 percent CI that is close to the null value.
| CONCLUSIONS |
|---|
|
|
|---|
| FOOTNOTES |
|---|
| References |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |