Although data regarding the use of EDs for the treatment of dental problems are limited, African-Americans and the poor have been found to be more likely than other groups to use EDs for physician contacts.² Those people without access to private dental services are disproportionately the poor and minorities, because these groups most frequently face economic and other barriers in accessing the private practice delivery system.^3–8

Because these groups also experience a greater burden of oral disease than other groups,^1–4,9–11 it is likely that they will use EDs for dental treatment and relief of pain. Many hospital EDs lack readily available dental services and, thus, are often not capable of providing definitive treatment for oral conditions. Nevertheless, costs are incurred when patients are assessed standard charges for ED visits (that is, facility and physician charges). When definitive treatment is not provided, this pattern of care and its associated costs may be repeated if patients are forced to return for treatment of the unresolved condition.¹² The magnitude of this problem is unknown.

Much has been written about the inappropriate use of EDs for nonurgent primary medical care services. Numerous studies also have examined dental treatment provided in the hospital setting. These studies have involved the provision of services to both children^13–15 and adults,^16–18 and often have focused on oral trauma.^19,20 However, these studies generally have involved hospitals with dedicated departments of dentistry, and the studies describe services provided by dentists in dental clinical facilities. Few studies have examined the use of EDs (as opposed to departments of dentistry) for the treatment of dental pain and infection.

A pilot study conducted at the University of Maryland Medical System in 1995 examined ED use for the treatment of dental problems.^21,22 The focus of that study was an analysis of ED use before and after a change in coverage status for poor adults took place. Specifically, to reduce dental-related costs, the state of Maryland in February 1993 eliminated Medicaid reimbursement to dentists for adult emergency dental services (routine adult dental services were eliminated in 1976). Medicaid reimbursement for ED visits, however, was not eliminated. The purpose of the pilot study was to determine if this change in policy, which reduced access to private dental services for the poor, also resulted in a concomitant increase in the use of EDs.

To better understand the impact of this policy change, we analyzed hospital data to test the hypothesis that the elimination in 1993 of Medicaid reimbursement to dentists in private practice for the treatment of adult dental emergencies resulted in an increase in the use of the hospital’s ED by patients receiving Medicaid. In the period following the policy change, the rate of ED dental visits by patients receiving Medicaid increased by 21.8 percent, while dental visits by patients who did not receive Medicaid increased by only 10.5 percent. This increase in the rate of dental visits by patients receiving Medicaid was opposite to a decrease from the prechange period to the post-change period in the overall percentage of ED patient visits (medical and dental) made by all patients who received Medicaid.

This pilot study had several limitations. Dental ED visits at only one hospital were examined and, therefore, the generalizability of the findings is limited. Because of the study period and the date of the elimination of dental Medicaid reimbursement, we could not control for potential seasonal variation in ED visits. Furthermore, potential racial, sex- and age-specific differences in ED use that might have resulted from the elimination of Medicaid reimbursement could not be adequately studied owing to the relatively small sample size. We have attempted to overcome these deficiencies and expanded on the pilot study by examining statewide data gathered during a four-year period that described visits to EDs, physicians and dentists. This report focuses on visits to EDs.

	METHODS AND PROCEDURES

TOP ABSTRACT METHODS AND PROCEDURES RESULTS DISCUSSION CONCLUSION REFERENCES

 
The study plan involved the examination of the use of Maryland EDs for the treatment of mouth pain and infections associated with the teeth and periodontal tissues during the four-year period from Feb. 16, 1991 through Feb. 15, 1995. The study used a natural experiment—the change in Medicaid policy on Feb. 16, 1993, that eliminated dentist reimbursement and participation in the program—to establish two observation periods (that is, before Feb. 16, 1993 [prechange period], and after Feb. 15, 1993 [postchange period]). The study sample included all adults aged 21 years and older who received Medicaid and who used Maryland EDs on a fee-for-service basis during one or both of the study periods.

We abstracted the data analyzed in this study from the Maryland Medicaid Management Information System, which is managed by the Maryland Department of Health and Mental Hygiene. This database contains information on all claims for reimbursement made to Maryland Medicaid (including date; provider; International Classification of Diseases, 9th Revision, Clinical Modifications, or ICD-9-CM, code; claim payments; and many other items). Published reports were available that reflected the mean number of people who were eligible for Medicaid each month during the study period (categorized by age, sex and race). Using these sources of information, we were able to calculate the number of dental-related ED visits per "person-years" of Medicaid eligibility both before Feb. 16, 1993, and after Feb. 15, 1993, (for example, 10 people who are eligible for Medicaid for an entire year equates to 10 person-years).

Statistical methods. We used descriptive and inferential statistical methods to analyze the data. The rate of claims from EDs for dental problems was estimated by dividing the total number of ED visits identified in the claims dataset by the total amount of person-time of Medicaid eligibility. Separate rates were estimated for the two years preceding the policy change and for the two years after the policy change. In addition, we calculated separate rates for demographic subgroups.

To determine the degree to which the rates of claims differed before and after the policy change, while controlling for changes in the age, racial and sex composition of the Medicaid population, we used multivariable Poisson regression models. These models also provided estimates of the independent associations between age and rates of claims, race and rates of claims, and sex and rates of claims while controlling for the effect of the policy change. Percentages do not always total 100 percent because of rounding error.

We identified dental ED visits through the use of ICD-9-CM codes.²³ The ICD-9-CM codes used in the study were as follows:

– 521-521.9 (diseases of hard tissues of teeth);

– 522-522.9 (diseases of pulp and periapical tissues);

– 523-523.9 (gingival and periodontal diseases);

– 525.3 (retained dental root);

– 525.9 (unspecified disorder of the teeth and supporting structures);

– 873.63 (internal structures of mouth, without mention of complication, broken tooth);

– 873.73 (internal structures of mouth, complicated, broken tooth).

We decided to use the ICD-9-CM codes in selecting subjects because of their wide use, standard definitions and required inclusion for Medicaid reimbursement. In addition, we used the hospital revenue codes to identify the ED as the clinic within the hospital setting in which the services were provided. Finally, provider location data were used to link physician claims to ED claims for the same visit where indicated (since a visit could have two associated claims: one for physician charges and another for facility charges).

	RESULTS

TOP ABSTRACT METHODS AND PROCEDURES RESULTS DISCUSSION CONCLUSION REFERENCES

 
Description of study population. A total of 3,639 people visited EDs for the treatment of dental problems sometime during the four-year study period. This included 1,831 people during the prechange period and 1,888 people during the postchange period (some people had visits during both periods). The racial, sex and age distributions of the users during the prechange and post-change periods were almost identical. The majority of patients were black (57 percent), female (72 percent during the prechange period and 73 percent during the postchange period) and aged 21 through 44 years (88 percent). Only 2 percent of the users were aged 65 years or older.

The demographic composition of the Medicaid-eligible population for inclusion in the study is presented in terms of person-years of eligibility. This was necessary because the eligibility status of participants could change during the course of the study. The person-years of eligibility were slightly greater in the prechange period than in the postchange period (406,903 vs. 399,953). As was the case with the actual users of emergency dental services, the demographic profile of the eligible population was very similar in the prechange and postchange periods.

Considering the entire four-year study period, we found that the largest percentage of person-years of eligibility was exhibited by women (73 percent), blacks (53 percent) and people aged 21 through 44 years (60 percent). The sex and racial profile of the eligible population generally was similar to those of users of ED services. The age profile, however, was quite different, with the eligible population composed of a larger percentage of older people, particularly those aged 65 years and older.

Description of ICD-9-CM codes. Tables 1 and 2 (page 719) describe the frequency distribution of ICD-9-CM codes used for ED claims and for physician claims linked to ED claims, respectively. Up to two ICD-9-CM codes could be submitted for each claim; therefore, there are more codes than claims.

The distribution of codes associated with physician claims linked to ED claims is provided in Table 2. The pattern of frequencies is consistent with those found in Table 1. "Unspecified disorder of the teeth and supporting structures" (525.9) again accounted for the largest percentage of total claims (40.3 percent), followed by "periapical abscess" (522.5) (28.6 percent) and "dental caries" (521.0) (21.56 percent). These three codesaccounted for more than 90 percent of the codes associated with physician claims linked with ED claims. Again, the frequency distribution of the codes did not vary greatly between the two periods.

Claims for emergency dental services. We examined claims for emergency dental services for ED facility claims and physician claims that were associated with the ED visit. The overall dental-related claims rate for the entire four-year period was .00536 per person-year of eligibility. The rates of ED claims before and after the policy change are presented in Table 3. We can see that the overall rate of claims significantly increased by 10 percent after the policy change. Comparisons between the prechange and postchange periods show significant increases in the rates during the postchange period for men, whites, blacks and patients aged 21 through 44 years and 45 through 64 years.

We also examined the claims pattern over the entire four-year study period for people with any dental-related ED visits. Approximately 44 percent of patients with at least one claim had one claim in the prechange period and no claims in the postchange period. Similarly, approximately 44 percent of patients had no visits in the prechange period and one visit in the postchange period. Approximately 2 percent of patients had at least one claim in both periods.

The pattern of claims from physicians in EDs, as might be expected, is quite similar to that seen for ED claims. In both the prechange and postchange periods, 89 percent of the claimants had one ED physician claim during the two-year periods. Eleven percent of the claimants had more than one claim in the prechange period, while 12 percent had more than one claim in the postchange period.

Our examination of the claims pattern over the entire study period for patients with any physician visits linked to dental-related ED visits revealed that about 44 percent of patients had one claim in the prechange period and no claims in the postchange period. Similarly, about 43 percent of patients had no visits in the prechange period and one visit in the postchange period. About 2 percent of patients with physician-linked ED claims had at least one claim in both the prechange and postchange periods.

	DISCUSSION

TOP ABSTRACT METHODS AND PROCEDURES RESULTS DISCUSSION CONCLUSION REFERENCES

 
This study has several limitations. Although the results are representative of the state of Maryland, they are not generalizable to the country at large. In addition, we cannot be certain that the change in the dentist reimbursement Medicaid policy was responsible for all changes in ED use that were observed. Factors other than the elimination of Medicaid reimbursement also may be responsible for the observed results. However, the inclusion of all Maryland hospitals and the examination of patient visits during an extended four-year period help to ensure the validity of the findings.

The change in Medicaid policy on Feb. 16, 1993, which eliminated dentist reimbursement and participation in the program, appears to have increased the use of EDs for the treatment of dental problems. After controlling for age, race and sex, we found that the rate of ED claims was 12 percent higher in the postchange period. Although statistically significant, the practical impact of this change was quite modest. Within demographic groups, patients aged 21 through 44 years, whites and men exhibited the highest rates of ED claims.

Although the rate of ED use increased after the elimination of reimbursement to dentists, the magnitude of the increase paled in comparison to the reduction in dentist-provided emergency services that resulted from the policy change. A total of 41,143 patients made at least one emergency visit to a dentist’s office during the two-year period preceding the policy change.²⁴ Dentist visits before the policy change resulted in approximately 62,000 claims for tooth extractions (L.C., R.M., L.M., C.D.M. Adult Medicaid Patients’ Dental Visits in EDs. Unpublished data, RO1-HS10129, Rockville, Md., Agency for Healthcare Research and Quality, October 2001). By comparison, 1,831 patients had ED claims during the period before the policy change, and 1,888 had claims during the period after the policy change.

It seems probable that adult recipients of Medicaid in Maryland experienced oral disease burdens during the period after the policy change that were comparable to those experienced before the change. How were these needs met after dentist reimbursement was eliminated? Patients faced with the loss of Medicaid reimbursement for dental visits might continue to seek care and pay for treatment out of pocket, obtain free care from dentists or from public clinics, receive care at EDs, receive care from office-based physicians or forgo treatment.

As mentioned above, the increase in ED use was not sufficient to substitute for the eliminated dentist-delivered services. The rate of visits to physicians also declined after the policy change.²⁴ Although no data are available, it seems extremely unlikely that services donated by dentists or free clinics, few of which exist in Maryland, could have substituted in a meaningful way for the magnitude of previously reimbursable dental services. Similarly, significant increases in out-of-pocket payments by disadvantaged patients themselves seem unlikely. The purpose of the policy change was to decrease Medicaid costs associated with the dental program. From this perspective, the policy change was effective, reducing payments to dentists from approximately $7.5 million in the prechange period to $0 in the postchange period.²⁴

	CONCLUSION

TOP ABSTRACT METHODS AND PROCEDURES RESULTS DISCUSSION CONCLUSION REFERENCES

 
Although we found that costs were dramatically reduced by Maryland’s elimination of Medicaid dentist reimbursement, any economic evaluation of a health care program should take into consideration other factors, such as improvements in health outcomes and quality of life.²⁵ Research has shown that disadvantaged patients are more likely to experience untreated oral health problems and associated pain.^1–4 In addition, Vargas and colleagues²⁶ reported that disadvantaged patients are more likely to forgo dental care, even when in pain. Untreated dental disease and associated pain may adversely affect a patient’s well-being and general quality of life.^27,28 Dental diseases are, in general, not self-limiting. Consequently, further research is needed to determine the overall impact of this policy change on disadvantaged adults.

View larger version (110K): [in this window] [in a new window]   Dr. Cohen is a professor and chairman, Department of Oral Health Care Delivery, University of Maryland Dental School, 666 W. Baltimore St., Baltimore, Md. 21201, e-mail "lac001{at}dental.umaryland.edu". Address reprint requests to Dr. Cohen.

View larger version (121K): [in this window] [in a new window]   Dr. Manski is a professor, Department of Oral Health Care Delivery, University of Maryland Dental School, Baltimore.

View larger version (118K): [in this window] [in a new window]   Dr. Mullins is an associate professor, Department of Pharmacy Practice and Science, University of Maryland Pharmacy School, Baltimore.