Background Correct diagnosis of food hypersensitivity (FHS) is important to ensure appropriate patient care and to accurately establish the population prevalence and incidence. Food challenges play a very important role in the diagnosis of FHS, but it is unclear when open food challenges (OFCs) opposed to double-blind placebo-controlled food challenges (DBPCFCs) should be used. This study investigated the use of OFCs and DBPCFCs when diagnosing FHS. Methods Children with a reported history of FHS or with sensitization to a food without known previous consumption were invited to undergo food challenges. Children of consenting parents underwent an OFC and those with a positive OFC were approached to undergo a DBPCFC. Food challenges were either performed as 1-day or 1-week challenges depending on sensitization status and clinical history. Results Forty-one children underwent both OFCs and DBPCFCs. The positive predictive values for 1-day and 1-week OFCs were 73% (8/11; 95% CI: 39–94%) and 57% (20/35; 95% CI: 39–74%) respectively. There was no evidence to indicate that the younger children were more likely to have a positive OFC confirmed by a DBPCFC compared to older children (Fisher’s exact P = 0.53). In the 1-day challenges parents indicated a preference for OFC rather than DBPCFC. By contrast, in the 1-week challenge parents indicated a preference for DBPCFC (P = 0.0192). Conclusion Open food challenge may be suitable for diagnosing immediate objective symptoms, whereas DBPCFC may be needed for the diagnosis of delayed and mainly subjective symptoms, irrespective of the child’s age.