April 22, 1996

Trends in Diagnostic Testing Following a National Guideline for Evaluation of Dyspepsia

Author Affiliations

From the Department of Veterans Affairs Medical Center, White River Junction, Vt (Drs Schwartz, Woloshin, and Welch); and the Center for the Evaluative Clinical Sciences, Dartmouth Medical School, Hanover, NH (Dr Welch).

Arch Intern Med. 1996;156(8):873-875. doi:10.1001/archinte.1996.00440080069008

Background:  The extent to which national guidelines affect actual clinical practice is largely unknown.

Objective:  To gauge the impact of a national guideline about dyspepsia on subsequent clinical practice.

Design:  We compared trends in the rates of the two principal diagnostic tests used to evaluate dyspepsia: upper gastrointestinal tract series (UGI) and endoscopy.

Methods:  We used the Health Care Financing Administration's (Baltimore, Md) 100% National Medicare part B file to determine annual counts for the two procedures in each of 10 years (1984 through 1993). To calculate procedure rates, we divided the number of procedures in each year by the total population enrolled in Medicare part B in that year.

Results:  In 1984, before publication of the guideline, the UGI rate was more than triple the rate of endoscopy (UGI rate, 59 per 1000 beneficiaries; endoscopy rate, 17 per 1000). Subsequent to publication of the guideline in 1985, UGI use slowly decreased while endoscopy utilization slowly increased. The rate of change for both procedures was steady over the 10-year period. Endoscopy rates finally exceeded UGI in 1993 (endoscopy, 37 per 1000; UGI, 36 per 1000).

Conclusions:  Despite a prominent national guideline, UGI utilization remains high and is slowly decreasing. Because utilization data before 1984 are incomplete, we cannot isolate the precise impact of the guideline. However, the slow decline in UGI utilization suggests that, at best, the guideline had limited impact on clinical practice. If this response is representative of other guidelines, alternative approaches to change clinical practice will need to be found.(Arch Intern Med. 1996;156:873-875)