The most positive aspect of research on ED use for oral problems is that it is being done, and being done fairly widely. There is an abundance of research and reports from local, state, and national levels. Researchers have investigated a wide variety of research questions including different aspects of care, costs, target populations and population subgroups, predictors, etc. There have been studies of both traumatic and non–traumatic reasons for dental care in EDs. When investigating presenting conditions, researchers have generally used different sets of ICD-9-CM codes. Though there is variation in the codes used to define conditions of interest, this variation is somewhat small, with the major differences relating primarily to whether trauma related oral conditions are included. Some codes are not used frequently and likely have a small impact on comparability of results between studies.
The problematic aspects of research methods addressing ED use for oral problems primarily relate to inconsistencies across studies. To some extent methodology will vary depending on the interests of the researchers. Researchers vary in their populations of interest, definition of ED treatments they are concerned about, the predictors of ED use they want to investigate, and factors related to potentially effective intervention strategies. Some of this variation may be due to low interest among researchers to simply determine rates and costs of ED dental care, and high interest in exploring unique and unstudied relationships of outcomes and predictors.
Another problematic aspect of research is coding at the EDs. Studies have assessed oral health training and knowledge, or lack thereof, of medical professionals providing care in EDs. Dentists are rarely on staff in EDs. Physicians typically address the presenting symptoms of pain and infection, often without a good understanding of the causes and appropriate treatment of the oral problems underlying these symptoms. Care usually involves providing prescriptions for pain medications and/or antibiotics, along with advice to see a dentist. This lack of ability to precisely determine the nature of the oral problem is reflected in part by use of the ICD code “dental disorder unspecified” (ICD-9 code 525.9, also related codes 521.8, 521.9, and 525.8), the most commonly used dental related code in hospital ED data. When more specific dental codes are used, they may be used inaccurately due to ED physicians’ lack of oral health knowledge. The problem of inaccurate and imprecise ICD-9 dental code use by physicians was pointed out by Lee et al. explaining their use of NHAMCS “reason for visit” codes instead of ICD-9 codes.(28) There is no easy way to address these issues.
Another code related issue is the variation in codes used by different researchers in different studies. The major difference in selected dental codes is whether there is interest in all dental related ED visits or interest in ED dental visits specifically related to conditions that might be preventable through regular dental care access, or NTDCs. However, even when researchers indicate an interest in all dental related visits or in NTDC visits, there still is variation in the codes used within these two defined categories. This lack of standardized code use can affect the comparability between studies.
Another problem associated with many available datasets is that identifiers are often associated with an ED visit, not a specific person. The result is that records for a person cannot be linked. Without knowing if a specific ED visit is for a first-time or a repeat patient, the extent of repeat visits to the ED for the same oral problem cannot be quantified. This was pointed out as a shortcoming in NHAMCS by Lee et al.(28) As much of the perceived problem with ED dental care is usually non-definitive treatment, and symptoms are addressed but the source of the problem is not resolved, a major shortcoming exists in our ability to assess unnecessary treatment and costs that would potentially not occur if there was a source of regular dental care, or at least some level of definitive treatment for those presenting at EDs.
A related problem is the inability to link medical and ED data for a given patient to dental claims data, hindering the ability to explore follow-up dental care in the primary care dental setting after ED visits for oral problems, and the impact of regular dental care on dental related ED visit rates. The utility of linked medical and dental data even when linking is possible is limited by the long-standing use of treatment codes rather than diagnostic codes in dentistry. However, initiatives for developing and implementing dental diagnostic codes, developing and promoting use of electronic health and dental records, and for linking medical and dental data are ongoing.
Some variability in research findings is introduced because some researchers limit their analyses to ED care resulting in discharge; some include ED care resulting in hospital admission; others limit their investigation to only ED treatment for oral conditions that result in hospital admission.
Summary and Conclusions
While many researchers have studied dental related ED care, there is a great deal of variation in target populations and different aspects of study methods. Study populations vary from the local to national level and are sometimes limited to subpopulations based on patient characteristics or certain care processes or outcomes. Research questions and outcomes of interest vary from general assessments of access to dental; care to specific rates, changes, or trends in ED usage for NTDCs. Studied predictive factors include basic demographics, insurance, environmental factors, other concurrent health conditions, and changes in dental care coverage or policies. Data sources also vary, though some national and state datasets are commonly used. Diagnostic codes used and the specific analysis methods employed also vary substantially, even when investigators were supposedly studying the same defined oral conditions. Although some problems with ED oral care research will be difficult to address, such as non-specific and incorrect use of oral diagnosis codes by physicians, efforts to develop and promote more standardized methods of study should be undertaken, especially in the area of basic surveillance.