“…Most studies used convenience sampling, but three studies used snowball sampling with female sex worker or client-client referrals [42], [47], [48], and one used stratified probability sampling [4]. Twenty-four studies sampled detainees [12], [36], [37], [39], [41], [49]–[52], [54], [57]–[70], but other studies sampled miners [46], clients at commercial sex venues [38], [41]–[43], [45], [47], [48], [56], [71], STI clinic attendees [38], [41], [47], [53], [55], [71], or the general population [4]. Four studies defined clients as men who reported purchasing sex in the previous year [4], [47], [48], [55].…”