Costly regional variations in primary health care test utilization in Sweden

Abstract

Objective: Laboratory tests are used increasingly in primary health care and they are thus associated with rapidly growing costs. Variations in clinical practice, an important determinant of expenditure for laboratory tests, could further increase the financial burden. The study's threefold objective was to determine the presence and extent of regional variations in test ordering between eight counties in Sweden, to investigate the influence on these variations by factors earlier described in the literature as explanatory, and to calculate the achievable savings that could be realized through optimized test ordering. Design: A retrospective study using test request data. Setting: A total of 223 primary health care centers in eight counties in Sweden. Main outcome measures: Thirteen ratios of commonly used laboratory tests, demographic data and the number of ordered tests per 1000 inhabitants served as the basis of comparison. The total savings per 100,000 inhabitants that could be achieved through optimized test ordering was estimated. Results: Large variations were found between all studied counties for all investigated ratios. However, none of the demographic variables investigated seemed to be able to explain the full extent of the variations. The range of achievable yearly savings per 100,000 inhabitants was €14,000–€185,000. Conclusion: The inter-county variations in Sweden are large and the savings associated with optimized test utilization are substantial. The investigated factors previously described as explaining the variations in test ordering only seem to explain a small part of the variation, and the variations are likely influenced by regional habits and traditions.

Key Words::

• Clinical chemistry tests
• cost savings
• diagnostic tests
• family practice
• health care costs
• physician's practice patterns

Acknowledgements

The authors would like to thank the laboratory representatives who provided production statistics from the studied counties.

Declaration of interest: No conflict of interest exists for any of the authors.