Improving Contact Tracing With Directed Recall

Contact tracing is a key strategy for slowing the spread of infectious diseases. We evaluated the effect of a directed recall manipulation on the quantity of names produced during contact tracing interviews. Participants reported close contacts over the last 4 days prior to the experiment. We found that people report more contacts when they are guided in directed recall (i.e., backward or forward in time) than when given the freedom to recall as they please. A second recall opportunity led to the retrieval of additional contacts, regardless of recall direction. Finally, the direction of recall instructions did not affect the overall number of names retrieved but did result in different patterns of retrieval density with respect to time. Backward recall elicited more recall of recent contacts, making it appropriate for public health officials engaging in forward-tracing protocols. Forward recall elicited more retrieval of distant contacts, making it appropriate for backward-tracing protocols. General Audience Summary The spread of infectious diseases can be slowed if public health officials can identify individuals who are at risk of getting sick and passing the disease on to others around them. Contact tracing is the process by which a public health official identifies people who were in recent close contact with an infected person. The success of contact tracing efforts depends on how well people can remember their recent contacts and report them. In many cases, individuals are simply asked to list all of their recent contacts. In this article, we evaluated the effect of asking people to report their contacts in either a forward or backward direction in time while providing specific time and day cues to guide their recall. This directed recall instruction led people to report substantially more contacts. A second attempt at directed recall led to report of even more contacts. Reporting contacts backward in time led to a relatively greater recall of more recent contacts, and reporting contacts forward in time led to a relatively greater recall of more distant contacts. These different patterns make the two procedures appropriate for different situations, depending on whether public health officials care most about detecting the source or the potential destination(s) of the current case. These findings provide some initial guidance on how to design contact-tracing interview protocols effectively.