Sleep Deprivation and Performance of Residents

To the Editor. —  The conclusion made by Deaconson et al¹ that sleep deprivation does not impair resident performance is not supported by their data. Mean scores were significantly worse in the sleep-deprived group (P<.05) on three of the five tests administered: grammatical reasoning, paced auditory addition, and the Purdue Pegboard.Furthermore, the study design had three deficiencies that probably minimized the real degree of impairment caused by sleep deprivation. It has been shown that physician performance on psychometric tests shows decrements when less than 3 hours of sleep are obtained.² Previous studies have defined sleep deprivation as less than 2 or 3 hours of total sleep within a 24-hour period.^3,4 Deaconson et al, however, liberally defined it as less than 4 hours of uninterrupted sleep in a 24-hour period. As judged by the ranges reported in Table 1 of their article, one subject classified as