James M. Clay, Tim Stockwell, Su Golder, Keegan Lawrence, Jim McCambridge, Nicole Vishnevsky, Alexandra Zuckermann & Timothy Naimi
Addiction: 2025;1–10. Research Report
Abstract
Background and Aims: The International Scientific Forum on Alcohol Research (ISFAR),
many of whose members are linked to the alcohol industry, has published over 280 cri-
tiques on alcohol and health research. This study investigated whether ISFAR critiques
favour studies reporting alcohol’s health benefits while being more critical of those iden-
tifying harms. We also examined whether industry-funded studies are more likely to
report benefits, and whether ISFAR’s critiques reflect the methodological rigor of the
studies they assess.
Methods: We analysed 268 ISFAR critiques published between April 2010 and January
2024, manually coding each underlying study for its content (whether the original study
reported alcohol-related health benefits or harms) and each critique for its tone (positive
or negative). Sentiment analysis (SA) algorithms were applied to critique summaries to
assess tone using automated methods. Study authors were examined for prior receipt of
alcohol industry funding. AMSTAR-2 and ROBIS tools evaluated risk of bias in 36 system-
atic reviews and meta-analyses favoured (n = 24) or criticised (n = 12) by ISFAR.
Results: Studies reporting health benefits had higher odds of receiving positive reviews
from ISFAR [odds ratio (OR) = 6.50, 95% confidence interval (95% CI) = (3.62–12.00)], as
did studies minimising alcohol harms [OR = 2.47, 95% CI = (1.40–4.45)]. Studies report-
ing health harms had higher odds of receiving negative critiques [OR = 0.29, 95% CI =
(0.15–0.14)], as did studies minimising health benefits [OR = 0.21, 95% CI = (0.10–
0.41)]. Algorithmic SA replicated these patterns, though the correlation with manual cod-
ing was modest [r = 0.20, 95% CI = (0.08–0.32)]. Studies with industry ties had higher
odds of minimising alcohol-related harms [OR = 1.90, 95% CI = (1.04–3.50)], and those
co-authored by ISFAR members had higher odds of reporting a J-shaped relationship
between alcohol use and health [OR = 2.52, 95% CI = (1.00–6.48)]. No association was
found between ISFAR sentiment and study quality as independently assessed by
AMSTAR-2 and ROBIS (BF 01 = 6.13–6.21).
Conclusion: Critiques from The International Scientific Forum on Alcohol Research
(ISFAR) consistently promote alcohol’s purported health benefits while minimising evi-
dence of harm, regardless of study quality. These patterns provide a valuable resource
for industry actors to shape public perception, downplay risk and influence policy—using
strategies that closely resemble those historically employed by the tobacco industry.