Concerns over analytical decisions in “Cat ownership and schizophrenia-related disorders and psychotic-like experiences: a systematic review and meta-analysis”

A meta-analysis of observational studies recently published in Schizophrenia Bulletin found a statistically significant association between cat ownership and the development of schizophrenia or psychotic experiences.¹ The findings of this study had significant resonance in Brazil, including news coverage in major newspapers and social media. We believe, however, that its conclusions are influenced by methodological issues that, if left unaddressed, could lead to dangerous immediate consequences in the real world, including the abandonment and mistreatment of animals. In this letter, we present issues that may have been overlooked during the peer-review process.

Although 17 articles were included in the review, upon examining its two forest plots, we only found 15 effect estimates, corresponding to 12 individual studies in both analyses. There is no rationale in the methods section explaining why five of the 17 studies (nearly 30%) were omitted from the meta-analyses, or even what criteria were specifically applied in deciding which studies would be combined, other than a brief sentence in the introduction “… to meta-analyze these estimates after sorting them into comparable study types.” Curiously, the meta-analysis represented in the first forest plot combines a cross-sectional study² and case-control studies, but does not include other cross-sectional studies.²,³ Assuming that the decision rule was to combine “comparable study types,” it is unclear why these studies were grouped in this way.

Unclear decision rules for combining studies have also left us unsure as to why Solmi et al.,⁴ the only study that prospectively evaluated the association between cat ownership and the development of psychotic experiences in adolescence, which found no association, was not included in any meta-analysis. Among all included studies, it scored best on the Newcastle-Ottawa Quality Assessment Scale, a tool for evaluating the quality of non-randomized studies. It also had the largest sample of individuals (n = 17,310). While the decision to exclude this study from the analysis may have been supported by reasonable explanations, the authors’ rationale is unknown, which leaves readers unable to understand or assess the adequacy of this analytical decision.

It is imperative for all studies that meet systematic review criteria to be adequately considered for inclusion in the meta-analyses – unless there is a compelling justification for their exclusion. Evidence-based psychiatry often relies on meta-analyses for clinical and public health decision-making; however, truly trustworthy meta-analyses are expected to provide a clear, explicit, and sound rationale for combining studies. The lack of transparency and questionable analytical decisions in this particular study are likely to have produced results that are unreliable at best, and dangerously misleading at worst.

In conclusion, we contend that the findings from McGrath et al.¹ should be considered with great caution until a scientifically valid correction of their manuscript is provided.

References

• ¹
  McGrath JJ, Lim CCW, Saha S. Cat ownership and schizophrenia-related disorders and psychotic-like experiences: A systematic review and meta-analysis. Schizophr Bull. 2023:sbad168.
• ²
  Bedwell JS, Spencer CC, Southwell AL. Childhood cat bites relate to increased adulthood severity of schizotypy, psychotic-like experiences, and social anhedonia in a transdiagnostic psychiatric sample. Psychopathology. 2020;53:36-47.
• ³
  Lindgren M, Torniainen-Holm M, Härkänen T, Dickerson F, Yolken RH, Suvisaari J. The association between toxoplasma and the psychosis continuum in a general population setting. Schizophr Res. 2018;193:329-335.
• ⁴
  Solmi F, Hayes JF, Lewis G, Kirkbride JB. Curiosity killed the cat: No evidence of an association between cat ownership and psychotic symptoms at ages 13 and 18 years in a UK general population cohort. Psychol Med. 2017;47:1659-67.

Edited by

Handling Editor: Andre Brunoni

Publication Dates

• Publication in this collection
  09 Sept 2024
• Date of issue
  2024

History

• Received
  29 Jan 2024
• Accepted
  31 Jan 2024