The study by Malla et al^{Reference Malla, Iyer, Rangaswamy, Ramachandran, Mohan and Taksal1} explored the differences in the 2-year outcome of first-episode psychosis at two sites, one in Montreal, Canada, and the other in Chennai, India. The study concluded a better outcome for negative symptoms in low- and middle-income contexts compared with a high-income context, concurring somewhat with the long-held notion of a better outcome in psychosis, particularly schizophrenia.^{Reference Hopper and Wanderling2} More family support partly explained this outcome. Evidence against this axiom has also been published^{Reference Cohen, Patel, Thara and Gureje3} in light of methodological limitations of studies supporting this hypothesis, human rights abuses in people with mental illness prevalent in low- and middle-income contexts, and socio-cultural transformations occurring in this part of the world. Notwithstanding these debates, we wish to point out a few issues with the present study.^{Reference Malla, Iyer, Rangaswamy, Ramachandran, Mohan and Taksal1}

Primarily the way family support was evaluated and used as a statistical metric. The two items (support and family relationship) from the Wisconsin Quality of Life Index – Provider Version were scored on a Likert-type scale; support on a scale of 1–3, and family relationship on a scale of 0–5. For a single-weighted score of family support, both the scores were multiplied, thus ending up with zero total scores occasionally if the latter was scored zero despite a variable score on the former item. Its significance is related to the variation in environmental support and family relationship in the two sites.

Another essential variable of interest missing from the study is the aspect of income (or family income adjusted to the gross domestic product per capita) and controlling for it for site difference other than family support at month 3.

For the examination of predictors of negative symptoms, remission and remission status at month 24, the adherence to medication variable was dropped. We do not find any reason for doing so.

The high-income context site had one-third of participants with affective psychosis versus 10% in the low- and middle-income context site. Patients with affective psychosis are more prone to extrapyramidal symptoms from antipsychotics than those with non-affective psychosis.^{Reference Gao, Kemp, Ganocy, Gajwani, Xia and Calabrese4} The higher chances of categorising depressive and extrapyramidal symptoms as negative symptoms without an evaluation of side-effects results in the possibility of inflating the findings.

Finally, concerning individuals who were non-completers of the study, first, mortality in four participants (three because of suicide) in the India site, to us, needs greater emphasis (and may be interpreted as a unique aspect in outcomes research for psychiatric disorders). Second, the disproportionately small number of participants lost to follow-up in the India site is not well explained. The latter could probably be as a result of a combination of family support, aspect of the patient–provider relationship and the value of free medications (provided at Schizophrenia Research Foundation (SCARF), India) in improving treatment adherence in this context.