Introduction: To generate indicators of the time to treatment for cancer patients using medical-administrative databases, we sought to determine whether the effective date of long-term illness (ALD) for cancer could be used as a "proxy" for the date of histological diagnosis. Objective: To study the concordance between cancer diagnosis dates and ALD effective dates. Materials and methods: Data were drawn from the cancer registries of Isère for 2011, and from Tarn and Lille and its surrounding area for 2009–2011, and pertained to colorectal, lung, and breast cancers. The analysis focused on the difference between the dates of diagnosis and the date of ALD effect. Results: The study sample consisted of 1,852, 1,514, and 3,437 cases, respectively, for colorectal, lung, and breast cancers. Only 19%, 8.5%, and 15%, respectively, of patients with colorectal, lung, and breast cancers had identical dates for cancer diagnosis and ALD effect. The absolute difference between the two dates was greater than one week in 54% of patients (colorectal and breast cancers) and 73% of patients with lung cancer. Discussion and conclusion: in the absence of information on the date of histological diagnosis, the ALD effective date cannot serve as a "proxy" date for calculating indicators of cancer treatment delay.