APPENDIX

Details of the Procedure

     Validation of referred partners.     We took several steps to minimize the possibility that reported partnerships involving successfully referred partners were fabricated. First, for index individuals who indicated they could locate >5 partners, only 5 were randomly selected for referral. Therefore, such index individuals could not plan in advance who their fabricated partners would be. Second, we collected detailed physical/identifying information from index individuals on the partners whom they were to refer. We compared this information with the persons presenting as index individuals' partners, and we rejected those whose physical/identifying information did not match that reported by the index individual (this occurred infrequently). Third, we asked referred partners how they got their vouchers and information about the persons who referred them. If persons presenting as partners either reported not knowing the index previously or reported information inconsistent with the description of the index, they were rejected.

     Laboratory procedures.     Oligonucleotide primers used in the first round of PCR amplification were 5′-GCGTCCGGGTTCTGGAAGACGGCGTGAACTATGCAACAGG-3′ (corresponding to nucleotides 802–841 of the HCV1 genome) and 5′-AGGCTTTCATTGCAGTTCAAGGCCGTGCTATTGATGTGCC-3′ (corresponding to nucleotides 1600–1639 of the HCV1 genome). The 805-bp product of the second round of PCR was generated using primers 5′-AAGACGGCGTGAACTATGCAACAGGGAACCTTCCTGGTTG-3′ (corresponding to nucleotides 821–856 of the HCV1 genome) and 5′-AGTTCAAGGCCGTGCTATTGATGTGCCAACTGCCGTTGGT-3′ (corresponding to nucleotides 1626–1587 of the HCV1 genome). PCR products were purified by agarose-gel electrophoresis with the Qiaex gel-purification kit (Qiagen). The purified PCR products were directly sequenced using the second-round-PCR primers.

     Identification of reported partners.     We identified index individuals and reported partners as uniquely as possible, by (1) participants' unique anonymous codes and interviewers' familiarity with participants' identities; (2) nicknames that were mentioned multiple times and that we judged to be uncommon (e.g., names similar to "Seahorse," "Mink," and "Grime") and therefore very likely to refer to the same individual in this local setting; (3) uncommon first names mentioned multiple times (< 0.5 persons expected to have that first name among the total number of partners mentioned by index individuals, as estimated from the Social Security Administration's first-name database stratified by decade of birth [http://www.ssa.gov/OACT/babynames/], weighted by the frequency of interviewed partners by birth decade, and accounting for whether a name was used for females, males, or both sexes); and (4) any links between these types of information in the data.