Altering Test-Ordering Behavior of Hospital Staff FREE

Several studies show that hospital staff often order too many laboratory and x-ray tests. The excessive test ordering adds unnecessary costs and sometimes results in false-positive findings. Researchers and managers have tried to change the test-ordering behavior of health care providers, but it is not easy. Providers are increasingly using computer-based systems to automate orders for tests, medications, and referrals. These systems are called care provider/computerized physician order entry (CPOE), and it ensures legible, standardized orders. It can be linked with software that provides staff with important information about tests or drugs, such as abnormal values, costs, and drug interactions. It may also include reminders or alerts to help assure that orders are appropriate on the basis of previous orders, medications, and patients' diagnoses. Given these features, hospitals are beginning to test ways of using CPOE to improve provider test-ordering behavior.

To see whether peer suggestions about limiting routine computerized orders can help hospital staff reduce unnecessary testing.

Hospital staff at an academic medical center.

The medical center formed a committee of peer leaders to study test orders for hospitalized patients. The committee reviewed past files from the medical center's care CPOE system to help identify patterns of excessive test ordering. The committee then chose and tested 2 different ways of using their CPOE system to reduce unnecessary test ordering. First, they targeted tests that were ordered repetitively in an open-ended manner beyond 3 days, such as “daily blood tests or x-rays.” They included a daily pop-up prompt that listed such tests and asked staff whether they wanted to discontinue any of the orders. Second, they changed the ways that tests could be ordered. They modified the CPOE software so that staff couldn't repetitively order tests in an open-ended manner. They also unbundled panels of blood tests that previously had been grouped as 1 order so that staff had to order each test separately. The software also showed staff graphs of past test results. The researchers then compared the numbers of tests ordered before and after the 2 interventions.

Computer prompts that questioned repetitive orders reduced the numbers of routine blood test and heart tracing (electrocardiograms) orders but not chest x-ray orders. Stopping open-ended repetitive orders and unbundling tests reduced the numbers of chest x-ray and blood test orders. Patient readmission rates, lengths of stay, transfers to intensive care units, and mortality rates remained stable despite reduced testing.

The study involved only 1 academic medical center. The study compared numbers of tests ordered before and after interventions. Other events or changes in the hospital occurring during the time of the interventions could have affected the staff's test-ordering behavior.

Peer-designed interventions using CPOE systems can improve the test-ordering behavior of hospital staff.