The Lancet Cannabis and psychosis. Misinterpretations of the results and scaremongering headlines. They were pre-committed to causal language and failed to acknowledge important limitations of their study.

The Lancet Cannabis and psychosis. Misinterpretations of the results and scaremongering headlines. They were pre-committed to causal language and failed to acknowledge important limitations of their study.

https://www.thelancet.com/journals/lanpsy/article/PIIS2215-0366(15)00113-3/fulltext

Cannabis and psychosis
James Coyne
x
James Coyne
Search for articles by this author
Affiliations
University of Groningen, University Medical Centre Groningen, Department of Health Sciences, Health Psychology Section, 9700 AD Groningen, Netherlands
Email the author James Coyne
Published: May 2015
PlumX Metrics

Captures
Exports-Saves: 7
Readers: 49
Mentions
Comments: 22
Social Media
Shares, Likes & Comments: 425
Tweets: 50
Citations
Citation Indexes: 2
see details
Learn more about article metrics
DOI: https://doi.org/10.1016/S2215-0366(15)00113-3
hideArticle Info
Publication History
Published: May 2015
© 2015 Elsevier Ltd. All rights reserved.

Marta Di Forti and colleagues' retrospective case–control study1x1Di Forti, M, Marconi, A, Carra, E et al. Proportion of patients in south London with first-episode psychosis attributable to use of high potency cannabis: a case-control study. Lancet Psychiatry. 2015; 2: 233–238

Summary | Full Text | Full Text PDF | PubMed | Scopus (120) | Google ScholarSee all References distinguished between self-reported use of high-potency (skunk) versus lower potency (hash) cannabis. Scaremongering headlines in the media predictably followed. The Daily Mail screamed “Scientists show cannabis TRIPLES psychosis risk: Groundbreaking research blames ‘skunk’ for 1 in 4 of all new serious mental disorders.” Undoubtedly, politicians and policy makers who have already made up their minds about regulation of cannabis will seize on the study as support for their views.

The authors must share some of the blame for misinterpretations of their results. They were pre-committed to causal language and failed to acknowledge important limitations of their study. An earlier report2x2Di Forti, M, Sallis, H, Allegri, F et al. Daily use, especially of high-potency cannabis, drives the earlier onset of psychosis in cannabis users. Schizophr Bull. 2014; 40: 1509–1517

Crossref | PubMed | Scopus (119) | Google ScholarSee all References by the authors concerning the same cases but not cited in the present one1x1Di Forti, M, Marconi, A, Carra, E et al. Proportion of patients in south London with first-episode psychosis attributable to use of high potency cannabis: a case-control study. Lancet Psychiatry. 2015; 2: 233–238

Summary | Full Text | Full Text PDF | PubMed | Scopus (120) | Google ScholarSee all References declared “Daily use, especially of high-potency cannabis, drives the earlier onset of psychosis in cannabis users”. The title of the present Article announces “psychosis attributable to use.” The abstract reports a “population attributable fraction”, with obvious interest to journalists and politicians who will not recognise the inferential leap. Population attributable fraction or risk is a term that should be reserved for associations for which causality has been more firmly established and appropriate caveats should indicate uncertainty about its applicability.3x3Rockhill, B, Newman, B, and Weinberg, C. Use and misuse of population attributable fractions. Am J Public Health. 1998; 88: 15–19

Crossref | PubMed | Scopus (924) | Google ScholarSee all References

Both cases and controls were selected from South London. The area's distinctive character limits generalizability to other areas of the UK or the world. Cases and controls were poorly matched. Half of the cases were black individuals and only a third were white, with these proportions reversed among controls. There are also highly significant differences in gender, education, and ever being employed. Such differences cannot be overcome in analyses with inclusion of an incomplete selection of crudely measured control variables. Frequency of use of cannabis was established by retrospective self-report and the key distinction of skunk versus hash was not one that participants might be able to reliably make.

Results could conceivably be explained by protopathic bias, with people who developed psychosis using cannabis or skunk to self-medicate early symptoms. Note the doubling of probability of psychosis for “never” as compared with “less than once a week”.1x1Di Forti, M, Marconi, A, Carra, E et al. Proportion of patients in south London with first-episode psychosis attributable to use of high potency cannabis: a case-control study. Lancet Psychiatry. 2015; 2: 233–238

Summary | Full Text | Full Text PDF | PubMed | Scopus (120) | Google ScholarSee all References Confounding by indication 4x4Salas, M, Hotman, A, and Stricker, BH. Confounding by indication: an example of variation in the use of epidemiologic terminology. Am J Epidemiol. 1999; 149: 981–983

Crossref | PubMed | Scopus (252) | Google ScholarSee all References is suggested by the profound disorientation that daily use of skunk entails and the likelihood that people who are not already dysfunctional would find it aversive and intolerable.

Causal inferences from case–control studies are always hazardous, particularly with poor matching and incomplete specification and imprecise measurement of possible confounds. The unwanted influence of relevant group differences, whether measured or left unmeasured, might only be compounded by brute application of statistical controls. Awareness of the obvious political and policy implications of results, and the likely misuse to which they could be put, apparently failed to discourage the authors from inappropriate causal inferences and ignoring of obvious limitations of their study. This failure can only serve to distract readers from the important message concerning the need for a distinction between higher and lower potency forms of cannabis, even if it should be made with greater precision and in the context of a prospective study.

I declare no competing interests.

References
1Di Forti, M, Marconi, A, Carra, E et al. Proportion of patients in south London with first-episode psychosis attributable to use of high potency cannabis: a case-control study. Lancet Psychiatry. 2015; 2: 233–238
View in Article
| Summary
| Full Text
| Full Text PDF
| PubMed
| Scopus (120)|
Google Scholar
2Di Forti, M, Sallis, H, Allegri, F et al. Daily use, especially of high-potency cannabis, drives the earlier onset of psychosis in cannabis users. Schizophr Bull. 2014; 40: 1509–1517
View in Article
| Crossref
| PubMed
| Scopus (119)|
Google Scholar
3Rockhill, B, Newman, B, and Weinberg, C. Use and misuse of population attributable fractions. Am J Public Health. 1998; 88: 15–19
View in Article
| Crossref
| PubMed
| Scopus (924)|
Google Scholar
4Salas, M, Hotman, A, and Stricker, BH. Confounding by indication: an example of variation in the use of epidemiologic terminology. Am J Epidemiol. 1999; 149: 981–983
View in Article
| Crossref
| PubMed
| Scopus (252)|
Google Scholar

Commentaires

Ajouter un commentaire

Plain text

  • Aucune balise HTML autorisée.
  • Les adresses de pages web et de courriels sont transformées en liens automatiquement.
  • Les lignes et les paragraphes vont à la ligne automatiquement.