See related article, p. 561 .
[Williams RM. Frequent emergency department use in Sweden: implications for emergency medicine in the United States. Ann Emerg Med. June 2001;37:627-629.]
In this issue of Annals, Hansagi et al1 compare the frequency of use of health care services at non–emergency department settings among categories of ED patients according to how many times per year they visit the ED. The authors were able to conduct such a study because of the availability of a national database in Sweden that tracks all health care use for about 95% of the population.
The study provides some interesting findings with regard to the health care status and total use of services in all settings among those patients who have a minimum of 4 ED visits per year (high users). On the basis of the findings in this study, there appear to be marked differences in patterns of use in Sweden compared with those in the United States among patients who make frequent ED visits. The policy goal of any health care system is to ensure access and to achieve the desired balance of incentives relative to the use of services, and there are some interesting policy implications from this study regarding the role of the ED relative to need and access.
The contrast in health care systems between the United States and Sweden is striking. Sweden has a national health plan that provides universal coverage for virtually all its citizens. All physicians, including those in private practice, are affiliated with the national health insurance plan, which is financed by taxes. Law regulates all fees for physician services, and these are only slightly higher for physicians in private practice.
As in a number of other European countries, there is a marked separation of services provided in the outpatient and inpatient settings. Other than a differential in the copayment ($27 versus $13), there is no discouragement of ED visits compared with the private office setting, and there is a cap of about $100 in out-of-pocket expenses per year for outpatient services.
Sweden maintains a national patient database that tracks the total use of health care services in the country by means of unique patient identification numbers. The database is available to researchers, and a special key code system is used to protect patient confidentiality. The Swedish database is extraordinary compared with data available in the United States. However, despite the substantial national database of total health care use in Sweden, it is interesting to note that the ICD-9 diagnosis codes were not consistently recorded for ED visits.
The authors examined ED use at a large university hospital with 70,000 annual visits. ED patients were categorized by frequency of use from rare visitors who had only 1 visit per year to frequent visitors with 4 or more visits per year. The study compared health care use in other settings according to the frequency of use of ED services. The study also examined standardized mortality ratios by category of frequency of ED use.
The results are indeed interesting. Nearly 75% of ED patients had only 1 visit during the year, and for 27% of patients in the study, 1 ED visit was the only medical care that they received during the entire year. Frequent users comprised 4% of total ED patients and accounted for 18% of total visits. Frequent ED users also had much higher levels of use in other outpatient settings, as well as markedly increased hospital admission rates (80% versus 36%) and elevated adjusted mortality ratios (standardized mortality ratio of 1.55).
The authors concluded that high ED users had higher non-ED use because they had lower health status that required more services. The authors also make the important and supported conclusion that the availability of primary care physicians does not alter rates of ED use.
This study by Hansagi et al1 provides a fascinating look at health care in a country with a well-established national program and universal insurance coverage. There are no comparable studies in the United States because of the lack of comprehensive interconnected databases.
Explaining ED use is the holy grail of public health research in the United States. Although theoretically one could construct a massive multiple regression equation that includes all possible independent variables (eg, age, race, sex, health status, socioeconomic status, geographic location), we are a long way from such an analysis, and fully 75% of health care use in the United States is still basically unexplained. About all we know for sure is that sicker patients use more services.2, 3, 4 As noted by Hansagi et al,1 several studies5, 6, 7 in the United States have related high ED use to use of nonurgent services, as well as identification of the ED as the primary source of care.
Economic disincentives are a component of almost all health insurance plans and are founded in the results of the Rand Health Insurance Experiment study.8, 9 Copayments and deductibles were found to be effective in reducing use of services, with a maximum effect occurring at about $1,000 in annual out-of-pocket expenses. The importance of the Rand study relative to the Hansagi et al1 findings is that the maximum out-of-pocket deductible for outpatient services in Sweden is only $100, which is well below the maximum effect level of $1,000 reported in the Rand study. This is significant because the high use group, who had at least 4 ED visits per year plus significant additional outpatient and inpatient use, would be virtually guaranteed of exceeding the annual $100 deductible. In such situations, copayments become moot and have no effect on the use of services.
It is unfortunate that the database does not reliably track diagnosis codes in the ED or track some type of severity level, such as what can be inferred from Current Procedural Terminology levels of service codes in the United States. These results could be useful in explaining high use by category of disease.
The study by Hansagi et al1 provides strong evidence that sick people do indeed seek increased health care services at all levels. The study cannot be generalized to the United States, however, because hospital EDs not only represent an important source of care for those who are sicker and require more services but also serve the important function of providing access to patients who may have no other source of care.10
What is really significant from a policy perspective about the study by Hansagi et al1 is that it clearly demonstrates that under a system of universal coverage, patients can seek and receive the services they require. Recall that three quarters of the ED patients in this study had only 1 visit during the year, and this sole ED visit was the only health service used during an entire year by 27% of patients. The danger in making generalizations from such studies is that we do not have universal health insurance coverage in the United States and that attempts to limit access to emergency services run the substantial risk of denying access to many patients who have no other source of care.
FULL TEXT