Do they want to know?: analysis of the decision for presymptomatic testing
Is Version Of
This study examined the decision-making process involved when individuals consider being tested presymptomatically for two different diseases, Huntington's disease and HIV infection (including the Acquired Immunodeficiency Syndrome). Each disease has a long latency period between contracting the disease and developing obvious symptoms. Both illnesses follow a debilitating, catastrophic path often involving neurological and cognitive deterioration. No successful cure currently exists for either disease. Recent advances in medical diagnostic technology have presented patients with the opportunity to know if they will develop these illnesses. Positive test results may be devastating while no successful cure for the diagnosed disease is currently available. The circumstances under which an individual will choose to be tested for such an illness are examined in this study. Policy capturing techniques (PCT) were used to determine the contribution of the independent variables by covarying the five dichotomous variables to create 32 hypothetical diagnostic testing situations to assess the individual decision processes of 217 undergraduate subjects. These five independent variables were: transmission mechanism, health status of the subject, partner/parent carrier status, physician test recommendation, and reproductive plans. Results indicate significant main effects for each of the five factors for the whole group. The five variables combined to account for more than 65% of the variance in the diagnostic decisional choices from the hypothetical scenarios. Cluster analysis was performed on the individual patterns of beta weights to determine if there were groups of subjects with similar decisions strategies. Seven clusters were found to represent different decisional patterns. Subject variables were of some use in describing the membership of each cluster. History of previous serious illness, plans for diagnostic testing, native language, and income were found to predict cluster membership and hence decision strategy. The implications of inadequate descriptors for the clusters and possible limitations in PCT are discussed.