Aims: To assess the validity of the resting and chewing components of the recently developed observational diagnostic tool, the Orofacial Pain Scale for Non-Verbal Individuals (OPS-NVI). Methods: This cross-sectional observational study was carried out in two UK hospitals. A total of 56 participants with dementia who were admitted to the acute hospital were observed for 3 minutes during rest and during chewing, and the OPS-NVI was used to identify orofacial pain. Afterwards, the participants were asked about the presence of orofacial pain using self-report pain scales. The sensitivity, specificity, and area under the receiver operating curve (AUROC) of the OPS-NVI were calculated for each activity. Spearman coefficient was calculated to assess the correlation between the number of positively scored behavior items of the OPS-NVI and the presence of orofacial pain according to self-report. Results: According to the OPS-NVI, orofacial pain was present in 5.4% of participants during rest and in 9.1% during chewing. According to self-report, the prevalence of orofacial pain was 5.4% during rest and 10.7% during chewing. The specificity of the OPS-NVI was 98.1% to 100%, the sensitivity was 66.7% to 83.3%, and the AUROC was 0.824 to 0.917. The predictive validity showed a strong correlation (0.633 to 0.930, P .001) between the number of positive behavior items and the self-reported presence of orofacial pain. Conclusion: The resting and chewing components of the OPS-NVI showed promising concurrent and predictive validity. Nevertheless, further validation is required and highly recommended. Palabras clave: dementia, facial pain, hospital, observation, OPS-NVI, toothache, validation