Population-based health surveys are increasingly including self-reported oral health measures. However, their validity is frequently questioned. This study aimed to review the diagnostic validity of self-reported oral health measures - regarding periodontal conditions, number of remaining teeth and use and need of prostheses - and to present prototypes of oral health items to assess periodontal conditions. Papers published between 1991 and 2011 were identified through PubMed database. The sample profile, the sample size and the methods used in each study were analyzed, as well as the sensitivity, specificity, positive and negative predictive values of the oral health items. Periodontists were contacted, using a standardized text, sent by e-mail, which asked them to provide self-reported items regarding periodontal conditions. We reviewed 19 studies; 13 assessed periodontal conditions; five, the number of remaining teeth and four, the use and need of prosthesis - some studies evaluated two or more conditions simultaneously. Five of the eight periodontists suggested questions to assess periodontal conditions. The maximum and the minimum sensitivity values to assess periodontal conditions, number of remaining teeth and use and need of prosthesis were 100 and 2%; 91 and 21%; 100 and 100%; respectively; the maximum and the minimum specificity values were 100 and 18%; 97 and 96%; 93 and 93%; respectively. In conclusion, there are acceptable sensitivity and specificity values for number of remaining teeth and use and need of prosthesis only. Finally, we consider there is the need for further studies in the national context, in order to assess the impact of the questions about self-reported oral health conditions in epidemiological analyses. Therefore, it will be possible to empirically verify if self-reported questions can be used in such studies.