In moderate alcohol use disorder, cognitive testing found nothing and predicted nothing

Design

Baseline neuropsychological assessment within a randomised trial of two psychological treatments, with a non-clinical reference sample

Population

147 adults with moderate alcohol use disorder and low psychiatric comorbidity in Stockholm; 72 non-clinical comparators

Primary outcome

Executive function across eight CANTAB tests, and its prediction of reduced alcohol consumption at 12 and 26 weeks

Effect

No test showed worse performance than the reference sample; no executive function measure significantly predicted reduction in drinking at either follow-up

One hundred and forty-seven adults with moderate alcohol use disorder and low levels of psychiatric comorbidity were recruited from three specialist addiction outpatient clinics in Stockholm into a randomised trial of two psychological treatments. Before treatment they completed eight CANTAB tests covering mental flexibility, sustained attention, visuospatial working memory, response inhibition and delay discounting. Seventy-two people from a non-clinical reference sample completed the same tests. Alcohol use was reassessed at 12 and 26 weeks.

The result is negative twice over. On none of the eight tests did the alcohol use disorder group perform worse than the reference sample. And no measure of executive function significantly predicted reduction in alcohol consumption at either follow-up.

The association between alcohol use disorder and executive impairment is well established - but the evidence for it comes overwhelmingly from convenience samples at the severe end, with heavy psychiatric comorbidity. This study deliberately sampled the other population: moderate severity, low comorbidity, the patients who actually fill an outpatient clinic. In that group the impairment is not there to find. Which means cognitive screening should not be used to decide who gets offered a psychological treatment, and a patient who relapses should not have it attributed to executive dysfunction that has not been demonstrated.

• Do not order neuropsychological testing to triage patients with moderate alcohol use disorder into treatments
• Do not explain relapse to a patient as a cognitive deficit unless one has actually been measured
• Reserve cognitive assessment for severe, long-standing or comorbid presentations, where the evidence base sits
• Record severity and comorbidity when reading any study of cognition in addiction - they determine the result
• Time psychological treatment on clinical grounds rather than waiting for cognition to 'recover'

Why it matters

The well-established link between alcohol use disorder and executive impairment was established in a population most clinics do not see.

Don't overread it

This was a Swedish specialist outpatient sample with low comorbidity; it does not speak to severe or highly comorbid presentations.

The statistics, in plain English

A negative finding across eight separate tests is stronger than a negative finding on one, because chance alone would be expected to produce a difference somewhere among eight comparisons. The more useful caution runs the other way: with 147 patients against 72 comparators, small differences could have been missed, so this establishes that there is no large impairment rather than that there is none. The prediction result is the more decisive of the two. A measure that does not predict outcome in the population where you would use it has no clinical role there, whatever its theoretical standing.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.