Brewer, D. D., Seddig, E. L., & Leigh, B.C. Comparison of daily reports and retrospective recall for eliciting drug injection partners. Presentation at the 24th International Sunbelt Social Networks Conference, San Diego, CA, March 10-15, 2009. Presentation available at SSRN: http://dx.doi.org/10.2139/ssrn.2542626.

To understand the spread of bloodborne viruses in injection drug users (IDUs), it is important to ascertain networks of injection contact. However, in retrospective interviews, IDUs often forget many injection partners. We compared retrospective recall with daily reporting.

We gave mobile phones to 14 IDUs who had participated in the first daily reporting study of the project. For four weeks, participants reported their injection partners in automated daily telephone interviews. Then, in a retrospective interview, they recalled their partners from the four-week period and received location and network cues to enhance recall. The 11 participants who reported partners in the daily interviews missed 3.4 (out of 28) such interviews on average. In the daily reports, they listed a mean of 6.4 unique partners. In the retrospective interview, they recalled a mean of 5.1 partners, including a mean of 0.8 in response to the recall cues (a 17% increase on average). One-third of cue-elicited partners were mentioned in the daily reports. In the retrospective interview, participants forgot 2.7 partners on average, and listed a mean of 1.5 partners not in the daily reports (perhaps partners on days with missing daily reports). The total number of partners reported at follow-up was moderately positively correlated with the total number of unique partners reported during the daily interviews (r = 0.66, intraclass correlation = 0.64). There was some under- and over-reporting of partnership risk at follow-up, relative to risk reported in the daily interviews, which could be due to missing daily reports and misreporting at follow-up, respectively. Other results mirrored prior research on correlates of forgetting. First, the number of partners freely recalled at follow-up correlated moderately positively with the number of partners elicited by the recall cues (r = 0.52). Second, forgotten partnerships involved, on average, slightly less risk and slightly greater injection frequency. Forgotten partnerships were moderately less recent, though still within the 28-day study period, and one-third of forgotten partnerships occurred within 14 days of the follow-up interview. In conclusion, daily reporting of injection partners is feasible and may be a useful adjunct to retrospective recall in some situations. The replication of the recall cues' effectiveness suggests that such techniques should be used as standard practice in injection network elicitation.

full presentation - pdf (ISR site)    full presentation (SSRN site)