We investigated the consistency between the headache diagnosis based on medical history and three treated headache episodes diagnosed based on a diary. In a randomized double-blind study including individuals with either migraine or tension-type headache (TTH) we showed significant superiority of the fixed combination of acetylsalicylic acid + paracetamol + caffeine over the combination without caffeine, the single preparations, and placebo in the treatment of headache. A neurologist performed a classification of the usual headache episodes and each of the three treated ones in a blinded fashion based on a structured questionnaire. This was done for the 1734 patients included in the efficacy analysis who usually treated their episodic TTH or migraine attacks with non-prescription analgesics. The overall percentage of patients with migraine and TTH remained relatively stable. The treated headache episodes were between 75 and 77% migraine, 18–20% were TTH and 5–7% could not be classified. We observed some shift in headache type within patients from prior history and in treated attacks. In 60% of patients all three treated episodes were of the type initially diagnosed by the neurologist by history (56% migraine and 4% episodic TTH). Of those with an initial diagnosis of migraine, 24% had at least one attack meeting criteria for TTH. Of patients with an initial diagnosis of TTH, 54% had at least one attack meeting the diagnostic criteria for migraine. Our results demonstrate that an initial headache diagnosis does not accurately predict the headache type treated in a randomized trial. Symptom features of treated headaches should be captured to ensure that the attack is of the type targeted by the clinical trial. The International Headache Society Guidelines for controlled clinical trials should be updated accordingly.