Peterman 1997.

Methods	Setting: public health services in Broward County, Florida; Tampa, Florida; Patterson, New Jersey, US Enrolment: participants were enrolled from December 1990 to March 1993 Follow up: no follow‐up recorded of index patients
Participants	1966 index patients with syphilis were randomised, 1042 male and 924 female Inclusion criteria Primary, secondary or early latent syphilis infection Exclusion criteria Criteria not specified
Interventions	After syphilis diagnosis all index patients were interviewed by DIS to identify sexual partners Contract referral (n = 586) Index patient to notify partners within 2 days, or a DIS would notify them on the third day Provider referral (n = 742) Partner notified immediately by DIS and referral of partner for testing Provider referral and field test (n = 638) Partner notified immediately by DIS who could draw blood for testing in the field, if it seemed unlikely for partner to come in for testing
Outcomes	Numbers of partners coming for syphilis testing, treatment or prevention (name and locating information of all partners were recorded in interview before randomisation, record searching) Cost per partner treated (clinic records)
Notes	Details on ethical approval not given
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Individual index patients were randomly assigned. Every day the study co‐ordinator at each site generated a list of assignments by using a random number table. The total number of patients in each arm differed significantly from 742, 638 and 586, this raises suspicion about whether randomisation was performed appropriately
Allocation concealment (selection bias)	High risk	The assignment was known to the interviewer before contact with the patient and the method was sequentially adapted by the interviewer
Incomplete outcome data (attrition bias) All outcomes	Low risk	Outcome data were available for all partners of index patients
Selective reporting (reporting bias)	Low risk	Same outcomes in methods section compared to results section. Trial registries not searched
Other bias	High risk	Deviation from protocol was reported by authors Some contamination was reported by the authors and this would have reduced the difference between the 3 groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding was not possible
Blinding of outcome assessment (detection bias) All outcomes	High risk	The DIS was not blinded, DIS did the interview before randomisation and also the intervention. No blinding of data entry personnel or data analyst No control in place to ensure 2‐day waiting period