Aim: Periodontitis proxy variables enable an expansion of periodontal research. The study aimed to estimate the validity of questionnaire items and registry data in relation to Stage III–IV periodontitis and having 50% bone loss.

Methods: Malmö Offspring Dental Study (MODS) participants (995) filled out questionnaires and underwent periodontal and panoramic radiography examinations. The questionnaire items, number of periodontal treatment procedures (PTP) in the Dental Health Register (DHR), and number of teeth with ≥ 6 mm probing depth in the Swedish Quality Register for Caries and Periodontal Disease (SKaPa) were evaluated as proxies for severe periodontitis. Stage III–IV periodontitis was the primary reference standard.

Results: For PTP‐based severe periodontitis proxy in DHR, positive predictive value (PPV) was 88% and negative predictive value (NPV) 87% for Stage III–IV. The SKaPa‐based proxy showed poor positive predictive values (PPVs, < 70%), but similar area under the curve (AUC), 0.74, compared with the DHR data (AUC 0.76). Sensitivity was < 70%, and specificity > 90% for the DHR and SKaPa proxies. Identification of cases with periodontitis by questionnaire combined with the demographic variables age, sex, smoking habits and education yielded good discriminatory ability (AUC > 0.75).

Conclusion: Register‐based data can effectively identify individuals with severe periodontitis in large cohort studies, thereby advancing periodontal research.