Measuring the error in sleep estimation in normal subjects and in patients with insomnia

The aims of this study were to measure the error in sleep estimation in normal controls and subjects with primary insomnia to establish the minimum amount of sleep needed for reliable subjective estimation and to depict the distribution of the error in sleep estimation in both groups. A two-step retrospective (study 1) and prospective (study 2) validation study was carried out. Study 1 included 288 normal subjects [176 females and 112 males, mean age 58.5 years, standard deviation (SD) 7.23]. Study 2 included 159 patients (98 females and 61 males; mean age 49.1 years, SD 12.71) with primary insomnia. Participants underwent a full-night polysomnographic study, followed by a morning assessment of subjective sleep parameters. A misperception index (MI) was computed using the following formula: MI = [objective total sleep time (oTST)-subjective total sleep time (sTST)]/oTST. The statistical properties of this index were analysed in detail in both groups. In controls, the Bland-Altman test demonstrated the reliability of this index for values of oTST >120 min. Healthy subjects estimated their sleep time correctly, while insomniacs largely underestimated their sleep time. Statistical analysis of the distribution of MI in insomnia patients disclosed the presence of two subgroups, one with moderate sleep misperception (132 patients) and the other with high sleep misperception (27 patients). The latter presented MI values >or=0.9, exhibiting statistical properties different from those with MI