Determining the value of a periodontal diagnostic test

The purpose of this analysis is to present a method to determine the value of diagnostic tests for evidence of bacterial activity related to periodontal disease. Five population studies are discussed: 1) screening the general population; 2) screening the patients of a general practice; 3) diagnosing initial patients referred to a periodontist; 4) monitoring periodontal maintenance patients; and 5) monitoring postsurgical refractory patients. A decision tree is constructed in which the primary decision branches are test and no test. Chance branches after the test branch were test positive and test negative. The treatment alternatives in the treatment branches were prophylaxis and oral hygiene, scaling and root planing, and advanced periodontal therapy (including surgical intervention and/or antibiotics or chemotherapies). Probabilities for the prevalence of disease in each population were obtained from epidemiological surveys. Disease activity and outcome states were estimated from clinical studies of periodontal treatment. For initial periodontal patients, using a sensitivity of .87 and a specificity of .84 the value of the test branch is .851 versus .566 for the no test branch. Assumptions on the diagnostic test's sensitivity and specificity and the utility of outcomes were tested with the decision analysis techniques of sensitivity and 2-way threshold analyses. These analyses show that using a test had a greater value than no test for older patients in the general practitioner's office, and for the 3 types of periodontal treatment patients. In contrast, using the test had a lower value than not using the test in younger dental patients in the general practitioner's office and for screening the general population. The diagnostic and treatment schemes proposed here are amendable to rigorous testing to evaluate their reliability and usefulness.