A psychosis percentage from compulsory rehabilitation is not automatically the probability that anyone using methamphetamine will develop psychosis. A 2018 Shanghai study also shows why large headline samples do not eliminate the need to inspect subgroup counts and table labels.
Who supplied the 37.1% estimate?
Gan and colleagues recruited 1,430 people meeting DSM-IV methamphetamine abuse or dependence criteria from three compulsory rehabilitation centers between October 2012 and June 2014. The methods identify 530 with methamphetamine-induced psychosis and 900 without it. Dividing 530 by 1,430 produces the reported 37.1%.
This was a selected rehabilitation population, with medical exclusions, not a probability sample of all people using methamphetamine. It was also not a prospective study following initially symptom-free participants to measure incident psychosis. The estimate describes the classification within the recruited sample.
How were symptoms and exposure established?
Participants were interviewed by trained psychiatrists, with diagnostic verification by a senior psychiatrist. Drug-use history and psychosis symptoms relied on retrospective reports. The study also collected craving, depression and anxiety ratings and a much smaller cognitive-testing subset.
The investigators acknowledged recall error, other substance exposure and the lack of precise withdrawal-duration data. Participants were interviewed within a month of admission, which does not establish identical time since last use or an identical clinical state for everyone.
Why do the table headings need checking?
The methods and main results consistently describe 530 psychosis cases and 900 noncases. Tables 1 and 2 instead label the 900-person column as the psychosis group and the 530-person column as the nonpsychosis group. The 403 male participants described in the text as psychosis cases appear in the table’s 530-person column.
That internal conflict means a reader should not infer the direction of a comparison from the table heading alone. It is appropriate to flag the inconsistency and seek clarification before using the tabulated group contrasts for a consequential individual opinion.
Were duration categories available for all 530 cases?
The article reports 420 transient episodes lasting less than one month and 97 persistent episodes lasting more than one month. Those counts total 517, not 530. Their reported percentages, 81.2% and 18.8%, fit the smaller total approximately. The thirteen-case gap is not explained by those two counts.
It would therefore be inaccurate to say that the paper fully assigned every psychosis case to one of the two duration groups. The categories also leave the handling of exactly one month unclear. Neither the missing classifications nor the retrospective design supports a precise persistence probability for another population.
Did all 1,430 participants show cognitive deficits?
No such conclusion follows. The text says 56 completed cognitive testing, while the reported comparison contains 24 psychosis cases and 30 noncases, totaling 54. The discrepancy matters, and this small subset cannot be treated as the entire cohort.
The study reported differences in verbal and visual learning tasks, with p = 0.011 and p = 0.023, but six other listed task comparisons were nonsignificant. A broad statement that every measured cognitive domain was impaired would erase those null findings. Selection into testing and multiple comparisons also constrain interpretation.
Can the correlates establish causes or individual predictions?
No. The authors explicitly acknowledged that the cross-sectional design could not prove causation. Anxiety, craving, dependence and reported use patterns can be associated with psychosis without establishing temporal direction or a validated prediction rule.
The article provides historical evidence from a clinically important selected population. Its useful interpretation preserves the recruitment setting, retrospective ascertainment, actual denominators and reporting conflicts. It cannot identify a person’s mental state from drug presence, establish that methamphetamine alone caused a particular episode or assign the headline percentage to every exposed individual.
Source
Gan H et al. A Research of Methamphetamine Induced Psychosis in 1,430 Individuals With Methamphetamine Use Disorder: Clinical Features and Possible Risk Factors. Frontiers in Psychiatry. 2018;9:551. doi:10.3389/fpsyt.2018.00551. Historical retrospective, cross-sectional human evidence.
Leave a Reply