Human immunodeficiency virus partner notification. Cost and effectiveness data from an attempted randomized controlled trial

OBJECTIVE: Determine the cost and effectiveness of partner notification for human immunodeficiency virus (HIV) infection. METHODS: Persons testing HIV positive in three areas were randomly assigned one of four approaches to partner notification. Analysis plans changed because disease intervention specialists notified many partners from the patient referral group. We dropped the patient referral group and combined the others to assess the cost and effectiveness of provider referral. RESULTS: The 1,070 patients reported 8,633 partners. Of those 1,035 were located via record search or in person. A previous positive test was reported by 248 partners. Of the 787 others, 560 were tested: 438 were HIV negative and 122 were newly identified as HIV positive. The intervention specialist's time totaled 197 minutes per index patient. The cost of the intervention specialist's time, travel, and overhead was $268,425: $251 per index patient, $427 per partner notified, or $2,200 per new HIV infection identified. No demographic characteristic of the index patient strongly predicted the likelihood of finding an infected partner. CONCLUSION: We could not compare the effectiveness of different partner notification approaches because of frequent crossover between randomized groups. The cost of partner notification can be compared with other approaches to acquired immunodeficiency syndrome prevention, but the benefits are not easily measured. We do not know the number of HIV cases prevented or the value of fulfilling the ethical obligation to warn partners of a potential threat to their health.