To the Editor: The rate of growth of pathology and radiology testing over the last decade has surpassed the average growth of most other medical services.1 Pathology Medicare items processed per capita between 1996–97 and 1998–99 demonstrated the largest increase (8%) of all item types. The interventional program for reducing diagnostic testing reported by Stuart, Crooks and Porton shows promise in addressing this increase in the emergency department, and has significant potential across other hospital departments.2

The program's focus on initiating behavioural change among test-ordering staff as a precursor to effecting significant long-term reduction of test use is important. Views differ as to the reasons for excessive clinical testing among hospital staff. Medicolegal issues, level of experience, fear of the consequences of inadequate testing and the desire to diagnose within a single presentation have been previously described.3 These issues were dealt with to some extent via the described educational component of the authors' intervention program and seem to explain the apparent sustainability of the intervention.

The authors report a "40% decrease in ordering of tests in the emergency department, with test utilisation falling from a mean of $39.32/patient to $23.72/patient." Other measures reported include reduced time taken for result review, with a resultant availability of additional resources for "other critical areas of service delivery". An assertion is made that "improvements to quality care" are "likely". Issues relating to improvement in quality of care, however, still remain:

• Although "no adverse patient outcomes relating to underutilisation of investigations" were identified, what follow-up was performed to ascertain "adverse outcomes"?

• What proportion of patients for whom further testing was requested via a general practitioner or outpatient clinic did not follow through with these investigations?

• Is there the potential for sufferers of undiagnosed chronic disease to develop more serious disease, requiring eventually more expensive therapies?

• What is the cost of patients' attending GPs and outpatient clinics for further investigations?

• Are the investigators' "evidence-based list of clinical indicators for ordering . . . tests" appropriate and are they rigid enough to prevent operator bias?

Test utilisation measured by cost was the primary outcome measure for this study. Based on this measure, the results appear promising; however, important quality-of-care issues need investigation before more widespread implementation is considered.