Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Reducing diagnostic tests in primary care


Bandolier has highlighted that ordering of many diagnostic tests is unnecessary and wasteful (Bandolier 55), and that relatively simple strategies with information technology support can reduce unnecessary test ordering in primary care (Bandolier 87). We now have a randomised trial in primary care which further supports the view that primary care can substantially reduce the ordering of tests [1].


The setting was five regions of Holland where diagnostic centres serviced primary care physicians, where GPs can order tests without hospital referral. Between the five centres there were 37 local groups of GPs.

The intervention here was a multifaceted contact with GPs that included the following elements:

Most of these discussions occurred in small groups of GPs supervised by the medical coordinator of the diagnostic centre. GPs then took the results of the discussions back to their practices with plans for implementation of guidelines and change at individual and group level.

The local group of primary care physicians formed the unit of randomisation. One group had a strategy related to cardiovascular and abdominal complaints, and the other had a strategy related to COPD and asthma, general malaise and fatigue, and degenerative joint disease. The conditions and associated tests are shown in Table 1. Each group acted as a control for the other, thereby balancing the influence on nonspecific effects related to actually being in a trial.

Table 1: Tests monitored in the trial by condition

Appropriate tests
Inappropriate tests
Cardiovascular Cholesterol and subfractions, K, Na, creatinine, exercise ECG Urea
Upper abdominal SGPT, gamma-GT, ultrasound of hepatobiliary tract SGOT, LDH, amylase, bilirubin, alkaline phosphatase
Lower abdominal PSA, CRP, ultrasound of the kidney, IVP, double contrast barium enema, signoidoscopy
COPD/asthma Allergy screening, chest radiograph IgE
Malaise, fatigue, vague complaints ESR, Hb, Ht, TSH,monospot Leucocyte count
Degenerative joint disease ESR, uric acid, rheumatoid factors Radiographs of lumbar or cervical spine, shoulder, knee or hip

The outcome was the number of tests requested per physician over six months. Tests monitored for the study included simple and cheap laboratory tests (cholesterol, ESR, uric acid) and more complex and expensive tests (exercise ECG, ultrasound scan of the hepatobiliary tract, sigmoidoscopy and radiographs).


Twenty-six groups with 174 GPs entered the study, with no differences between the them after randomisation. Fewer tests were ordered by GPs in intervention groups than in control groups (Table 2). All reductions in cardiovascular tests were statistically significant, but none of the changes for respiratory, malaise and joints tests were significant. The overall reduction was relatively modest, with savings of 67 and 28 tests per six months for each of the groups.

Table 2: Results of using evidence and guidelines to target diagnostic tests in Holland

Percent change from baseline
Change in tests over six months
Cardiovascular and abdominal
Upper abdominal
Lower abdominal
Respiratory, malaise and joints
General complaints
Degenerative joint disorders
All reductions in cardiovascular were statistically significant, but none of the changes for respiratory, malaise and joints were significant


This is an interesting study. It is properly and cleverly designed and conducted, and shows that relatively straightforward education and management interventions can make a difference.

The interventions are not rocket science. They just asked GPs to look at guidelines and evidence, and discuss and plan together how they might overcome the difficulties in making change. The GPs were not told to make the change: rather they were invited to. A breath of fresh air, some may say, but actually just good management principles, using management in its real-world sense.

Despite this, some might question whether it was all worth it. Bandolier calculates the saving to be about 7,500 fewer tests per million population. No great shakes this in the vast scheme of things, but valuable nonetheless if some of those savings were for more complicated and expensive tests which might have long waiting times. And if the cost of each test was just £25 the savings would be getting on for £200,000 per million population per year.

Even so, the importance of the study is not the result, but the principle that relatively simple management interventions based on good evidence and good guidelines, implemented by GPs working together to find effective ways of implementation gets a job done. Might even be a way to run a complex organisation like a health service.


  1. WH Verstappen et al. Effect of a practice-based strategy on test ordering performance of primary care physicians. A randomised trial. JAMA 2003 289: 2407-2412.

previous or next story