I have some reservations regarding the conclusions drawn by Large et al ^{Reference Large, Farooq, Nielssen and Slade1} in their study on duration of untreated psychosis in low- and middle-income (LAMI) countries. This is because the samples are not really representative of the occurrence of psychosis. It seems, people with untreated psychosis who have recovered or remitted without antipsychotic or medical treatment are excluded from this study. There is enough evidence that in LAMI countries, a substantial proportion of patients with psychosis seek treatment from traditional healers, ^{Reference Razali and Mohd Yasin2} use indigenous methods based on their non-biomedical beliefs ^{Reference Sarvanan, Jakob, Deepak, Prince, David and Bhugra3} or pathways to care. ^{Reference Kulhara4,Reference Chong, Mythili, Lum, Chan and McGorry5} Perhaps, many of those who fail to respond to these methods seek psychiatric help. Thus, the sample which reaches psychiatric services is a biased one. In clinical practice, we do encounter patients who have had previous episodes of psychosis which remitted spontaneously or by indigenous methods. Studies on duration of untreated psychosis should be community or general population based to overcome the confounding effects of non-psychiatric treatments and biased sampling. This is true more so for LAMI countries where such non-medical services are popular, in contrast to high-income countries ^{Reference Temmingh and Oosthuizen6} with well-organised health services, where any patient with psychosis is likely to reach psychiatric services without the pathway to care through non-psychiatric methods. This limitation needs a mention by the authors. ^{Reference Large, Farooq, Nielssen and Slade1}