Purpose: This study investigated the effectiveness of algorithmic testing in hematopathology at the Brigham and Women's Hospital and Dana-Farber Cancer Institute (DFCI). The algorithm was predicated on test selection after an initial pathologic evaluation to maximize cost-effective testing, especially for expensive molecular and cytogenetic assays.
Materials and methods: Standard ordering protocols (SOPs) for 17 disease categories were developed and encoded in a decision support application. Six months of retrospective data from application beta testing was obtained and compared with actual testing practices during that timeframe. In addition, 2 years of prospective data were also obtained from patients at one community satellite site.
Results: A total of 460 retrospective cases (before introduction of algorithmic testing) and 109 prospective cases (following introduction) were analyzed. In the retrospective data, 61.7% of tests (509 of 825) were concordant with the SOPs while 38.3% (316 of 825) were overordered and 30.8% (227 of 736) of SOP-recommended tests were omitted. In the prospective data, 98.8% of testing was concordant (244 of 247 total tests) with only 1.2% overordered tests (3 of 247) and 7.6% omitted tests (20 of 264 SOP-recommended tests; overall P < .001). The cost of overordered tests before implementing SOP indicates a potential annualized saving of $1,347,520 in US dollars (USD) in overordered testing at Brigham and Women's Hospital/DFCI. Only two of 316 overordered tests (0.6%) returned any additional information, both for extremely rare clinical circumstances.
Conclusion: Implementation of SOPs dramatically improved test ordering practices, with a just right number of ancillary tests that minimizes cost and has no significant impact on acquiring key informative test results.