Abstract
This study's goals were to (a) determine whether sending a survey by certified mail results in a higher response rate from physicians compared to sending by first-class mail and (b) evaluate the cost-effectiveness of this method. The study sample was 409 physicians who were nonrespondents to two previous mailings of a medical specialty society survey. Eligible physicians were designated at random to receive afinal mailing either by U.S. Postal Service certified mail including a return-receipt postcard or by first-class mail. There was a higher response rate from the certified mail group compared with the first-class mail group (41.3% versus 24.8%; relative risk = 1.66, 95% Confidence Interval 1.25, 2.21). A cost-effectiveness analysis showed that the cost per respondent was higher using certified mail versus first-class mail in the third mailing ($2.77 versus $2.34). Thus, use of certified mail is effective in increasing survey response but more costly.