Michael Laposata, M.D., Ph.D., Massachusetts General Hospital in Boston
Of course, new tests are meant to help clinicians make quicker and more accurate diagnoses. Consequently, patients should receive appropriate treatment sooner resulting in better, less risky, less inconvenient, and possibly less expensive healthcare. But clinicians are already snowed under trying to keep up with new treatments and changes in practice and procedures associated with managed care. So just making a test available does not mean it is being used appropriately, or even being used at all.

Take blood tests, for example, currently there are more than 100 coagulation tests available. Some, like prothrombin time (PT) and partial thromboplastin time (PTT), have been around for years and are standard elements of every physician’s training. So most clinicians know when to request PT and PTT tests and how to interpret the results. However, knowing what additional tests to perform if the PT and PTT are elevated is a big challenge for most physicians. In addition, even these familiar tests are changing. Recent advances in clot waveform analysis now mean PT and PTT tests can provide more valuable clinical information. These advances mean interpreting the test results is more complicated.

Avalance of new tests
An obviously flawed strategy for coping with this avalanche of tests is for clinicians to order them all. This approach is like trying to ski ahead of a snow slide — success depends more on luck than skill. Some tests are bound to be inappropriate and clinically irrelevant and performing them subjects patients to unnecessary risk, anxiety, inconvenience, potential treatment delays and increased costs. And that assumes the clinician knows how to interpret all the results. Ironically, not ordering a test that could provide an accurate diagnosis quickly also could mean greater inconvenience and risk for the patient, not to mention prolonged and exacerbated illness. So how can clinicians wend their way through the maze of tests available?

One way, at least in regard to coagulation tests, is to use the many algorithms developed by Michael Laposata, M.D., Ph.D., and colleagues at Massachusetts General Hospital in Boston. Laposata started developing these algorithms nearly ten years ago, when he said, “It became apparent that new tests were appearing faster than the clinicians could understand how they were appropriate. This increased the possibility that they might order a test not knowing whether it was reasonable or not order a test because they had never heard of it.” Since the Mass General Lab evaluates about 15 to 20 cases a day just for hypercoagulability, the need was obvious.

Algorithm uses ‘reflex’ testing
The hypercoagulation algorithms Laposata and company developed are now available at the College of American Pathologists web site ( to any laboratory interested in using them. The algorithm Laposata and his colleagues developed for the diagnosis of activated protein C resistance and the factor V Leiden mutation uses a system called “reflex” testing. One of the first issues the Mass General team had to clarify while developing the algorithms was the difference between an algorithm with reflex testing and a panel. “A panel involves a series of tests, some of which may not necessarily be associated with the clinical condition, whereas, reflex-test based algorithms involve only tests associated with the patients clinical condition — for us in coagulation, that means bleeding or clotting,” said Laposata.

Like an “if then” statement that leads to different actions based on differences in conditions in a computer program, the coagulation algorithms use reflex tests to guide laboratory technologists in selecting the next test and the next algorithm depending on the results of the previous test. Many reflex algorithms are in use in essentially all laboratories already. These “universally applied reflex algorithms” include procedures such as manual examination of a peripheral blood smear to get a CBC if an automated system is unable to accurately perform the white blood cell differentiation count. Doing the manual test is a “reflex” because it does not have to be ordered. But what the Mass General laboratorians were developing were “local” reflex algorithms — that is algorithms based on their determination of which tests and in what order to include in the algorithms. Local reflex algorithms must be approved by the institution’s Medical Policy Committee to ensure that their tests and procedures are clinically justified and will provide clinically beneficial information. “We’ve tried to associate each algorithm with a clinical situation or an abnormal result for a screening test,” said Laposata. “But,” he warns,” the algorithms are still only guidelines. Variations in the specific situation may mean you have to break out of the algorithm.”

The lab only does the right tests
“Convincing the Medical Policy Committee to institute the use of the hypercoagulability algorithms was not difficult,” said Laposata “because the alternatives of checking the patients with one test at a time and having to call them back to the hospital if more tests are needed or doing all the tests available are both unacceptable.” Using the algorithm means the laboratory only does the right tests, and the patient only has to come to the hospital once. “The Committee had little difficulty seeing that the use of the algorithms was in everyone’s best interest,” said Laposata, “the patient, the hospital, and the insurance companies.” “And,” he added, “the hospital clinicians have indicated that using the algorithms has greatly reduced the number of unnecessary tests. Plus, indications are that we are saving about a million and a half dollars annually by getting a correct diagnosis sooner and shortening length of stay.”

The Mass General lab also provides an expert narrative interpretation of the more specialized coagulation test results on every case without the need for a special order to do so. The consult is billable and most insurance companies pay for the service.

“I hope other laboratories will take our algorithms and show them to their committees so they can start using them too,” said Laposata. Because much of the groundwork has been done and the algorithms are already in use at Mass General, Laposata believes it should not take long for other laboratories to get approval from their medical policy committees. As for clinician acceptance, he said some were skeptical at first, especially in teaching institutions where the emphasis is on the residents solving their own problems (that is occasionally order the wrong test). However, with environmental changes such as increased patient load, the time and cost savings the algorithms can bring are a strong argument for their use.

Laposata would like to expand the use of algorithms to guide laboratory testing. “We are trying to put algorithms for test selection in place wherever it makes clinical and economic sense, and the clinicians keep asking us for them. For example, I was recently asked if I could develop a “first-time anemia” algorithm. So, the clinical need is there. Now it is our turn to implement the needed services.”