The prevalence of shift work disorder (SWD) has been studied using self‐reported data and the International Classification of Sleep Disorders, Second Edition (ICSD‐2) criteria. We examined the prevalence in relation to ICSD‐2 and ICSD‐3 criteria, work schedules and the number of non‐day shifts (work outside 06:00–18:00 hours) using objective working‐hours data. Secondly, we explored a minimum cut‐off for the occurrence of SWD symptoms. Hospital shift workers without (n = 1,813) and with night shifts (n = 2,917) and permanent night workers (n = 84) answered a survey (response rate 69%) on SWD and fatigue on days off. The prevalence of SWD was calculated for groups with ≥1, ≥3, ≥5 and ≥7 monthly non‐day shifts utilizing the working hours registry. ICSD‐3‐based SWD prevalence was 2.5%–3.7% (shift workers without nights), 2.6%–9.5% (shift workers with nights) and 6.0% (permanent night workers), depending on the cut‐off of non‐day shifts (≥7–1/month, respectively). The ICSD‐2‐based prevalence was higher: 7.1%–9.2%, 5.6%–33.5% and 16.7%, respectively. The prevalence was significantly higher among shift workers with than those without nights (p‐values <.001) when using the cut‐offs of ≥1–3 non‐day shifts. Shift workers with nights who had ≥3 days with ICSD‐3‐based SWD symptoms/month more commonly had fatigue on days off (49.3%) than those below the cut‐off (35.8%, p < .05). The ICSD‐3 criteria provided lower estimates for SWD prevalence than ISCD‐2 criteria, similarly to exclusion of employees with the fewest non‐day shifts. The results suggest that a plausible cut‐off for days with ICSD‐3‐based SWD symptoms is ≥3/month, resulting in 3%–6% prevalence of SWD.