Premium
Diagnostic accuracy of the use of parental‐reported sleep bruxism in a polysomnographic study in children
Author(s) -
Restrepo Claudia,
Manfredini Daniele,
Castrillon Eduardo,
Svensson Peter,
Santamaria Adriana,
Alvarez Claudia,
Manrique Ruben,
Lobbezoo Frank
Publication year - 2017
Publication title -
international journal of paediatric dentistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.183
H-Index - 62
eISSN - 1365-263X
pISSN - 0960-7439
DOI - 10.1111/ipd.12262
Subject(s) - medicine , polysomnography , gold standard (test) , predictive value , kappa , correlation , cohen's kappa , ordinal data , diagnostic accuracy , pediatrics , psychiatry , statistics , electroencephalography , linguistics , philosophy , geometry , mathematics
Background Polysomnography ( PSG ) is the gold standard for sleep bruxism ( SB ) diagnosis. PSG / SB children's criteria are not available; thus, parental‐report SB is widely used. Aim Assessing the diagnostic accuracy of parental report of sleep tooth grinding ( STG ) with a PSG / SB diagnosis in children, adopting adult criteria. Design Thirty‐seven children from clinics of Universidad CES were included. Parents filled the Children's Sleep Habits Questionnaire ( CSHQ ) assessing the single‐observation report – CSHQ – of STG with a No/Yes answer and five ordinal answers. A 5‐day diary reporting the presence/absence of STG (multiple‐observation report) was also completed. Each child underwent a single‐night PSG study. Sensitivity, specificity, positive predictive value, negative predictive value, accuracy, likelihood ratios, Spearman correlation coefficient, and Cohen's kappa coefficient were used to analyze data. Results Single observation, using No/Yes answer, showed acceptable specificity and NPV , while low PPV and sensitivity. Accuracy and likelihood ratios were low. When using the five ordinal answers, weak correlation and fair agreement ( r = 0.34 and κ = 0.40) with PSG / SB adult criteria were found. Multiple‐observation evaluation of STG presented moderate correlation and agreement ( r = 0.50 and κ = 0.48). Conclusions Although multiple‐observation report achieved better agreement than single‐observation report, our results failed supporting the validity of report strategies for the diagnosis of SB in children, as an equivalent of PSG / SB adult criteria.