EVIDENCE LIBRARY
Does Professional Motor Skill Protect Against Alcohol Effects? Nine Neurosurgical Residents
Nine neurosurgical residents completed repeated manual tasks before and after drinking. Faster completion sometimes accompanied poorer quality, but the small uncontrolled sequence did not quantify driving impairment or prove tolerance.
Professional training did not prevent poorer manual-task quality after alcohol in a nine-person experiment. Some tasks were completed faster despite more errors or lower quality. The findings challenge the idea that speed or practiced skill alone establishes unimpaired performance, but they do not quantify a surgeon’s—or a driver’s—individual risk.
Experienced participants, repeated tasks
Holzer and colleagues studied nine neurosurgical residents, seven men and two women, with surgical experience ranging from 0.6 to six years. Participants practiced seven task stations before completing them sober, after drinking, and later during declining alcohol concentrations.
The stations included suturing, controlled drilling, screw removal, a navigation task and pegboard work. Across the three conditions, the study generated 189 task performances from nine people, not 189 independent participants.
Blood was collected after each condition’s task series. Median measured alcohol concentration was zero at baseline, 0.96 g/kg after drinking and 0.64 g/kg during the later condition. These are grams per kilogram of blood; they should not be silently substituted for g/L or a breath value. Sampling after a task series also differs from continuously measuring concentration during each task.
Speed and quality did not recover together
Median drilling time fell from 106 seconds sober to 60 seconds after drinking and 65 seconds later. However, the egg membrane used as a precision target remained intact in three of nine sober performances, none after drinking and one during the later condition.
The drilling-time comparison was significant across conditions, p = 0.013. The membrane-intact comparison was not, p = 0.097. Thus, the descriptive error pattern should not be presented as a separately confirmed effect merely because task speed changed significantly.
For microsurgical suturing, sufficient closure occurred in all nine sober performances, three after drinking and seven later. The across-condition sufficiency comparison was significant, p = 0.009, whereas completion time was not, p = 0.459. Different outcomes within the same task again supported different conclusions.
A lower later BAC did not establish return to baseline
Median completion times for timed tasks were shorter than baseline in the later condition. Some quality measures improved relative to the earlier alcohol condition, but several remained descriptively poorer than sober results. A lower measured concentration therefore did not uniformly restore the baseline performance profile.
This was not a matched-concentration comparison between rising and falling BAC. It cannot isolate acute tolerance from dose, elapsed time, practice or fatigue. Nor was the later condition an alcohol-free hangover: the median BAC remained 0.64 g/kg.
Apparent tolerance was not established
Only two participants were classified as having at-risk drinking patterns. They appeared to have more unchanged or improved task results, but the authors limited that subgroup comparison to description. No statistically supported tolerance conclusion can be drawn from two people.
The observers knew the experimental condition, and some quality ratings were subjective. Tasks were repeatedly performed in the same sober-then-alcohol-then-later sequence without a repeatedly tested sober control group. Familiarization reduced but could not remove practice effects.
Food intake was available but not systematically recorded. The study did not assess planning, communication or actual patient outcomes, and the professional participants were still residents rather than a representative sample of all experience levels.
The DUI relevance
A practiced action completed promptly can coexist with reduced precision. That principle can guide scrutiny of behavioral evidence, but a surgical-task result is not a validated substitute for driving assessment. This experiment cannot convert manual skill or occupational experience into a personal BAC, braking distance or probability of a crash.
The site’s real-car parking experiment provides a separate driving setting in which collision count and completion time did not behave alike. The laboratory-to-driving prediction analysis explains why performance endpoints require their own validation.
Primary source
Holzer A and colleagues. Impact of moderate alcohol consumption on visuo-motor skills in neurosurgical residents. Scientific Reports. 2026;16:18691. https://doi.org/10.1038/s41598-026-58320-5.