We are grateful for the insightful comments of Dr Peralta and colleague in highlighting the main finding that delusional disorder and schizophrenia do not differ in their clinical features or neurocognitive performance. In their letter (Peralta & Cuesta, Reference Peralta and Cuesta2015), several methodological issues relating to the representativeness and diagnostic stability of the sample of our study (Hui et al. Reference Hui, Lee, Chang, Chan, Lin, Xu and Chen2015) have been raised.

The first concern is about the comparison of delusional disorder with age-matched schizophrenia in the study. We understand control of confounding factor can be done in various ways. It has been suggested that matching has several advantages over mere adjustment of variables in data analysis, and both methods are in fact complementary and best used in combination (Stuart, Reference Stuart2010). As documented in our article, the major findings were similar irrespective of age-matching, and that the pros and cons of matching have also been thoroughly discussed.

A second concern raised is the use of a first-episode sample, in particular delusional disorder due to its potential diagnostic instability. Our first-episode sample enabled us to shed light on the missing puzzle in the literature by examining the neurocognitive performance and functioning differences during the early phase in the two disorders, where results were least contaminated by the effect of long-term medication. Acknowledging the issue on diagnostic instability, local data found a better 5-year prospective consistency of a DSM-IV diagnosis of schizophrenia (96%) and delusional disorder (50%) than other studies reported (Chang et al. Reference Chang, Pang, Chung and Chan2009). Existing diagnosis was made by the research team which reviewed all available clinical information until 6 months following an initial episode. We fully agree with the importance of ascertaining diagnosis in a longitudinal manner.

The third argument concerns the exclusion of a large proportion of schizophrenia cases aged under 25 years. In this paper (Nowrouzi et al. Reference Nowrouzi, Kamhi, Hu, Matmary and De Luca2015), the authors stated the diagnostic heterogeneity in schizophrenia and they suggested three patient subgroups according to their age at onset. The comparison of delusional disorder with age-matched schizophrenia can therefore facilitate the exploration of their genuine differences as a function of age, given it is a well-known confounding factor on outcome in patients with psychotic disorders.

To address the potential inconsistent data with those in the previous literature: (a) A major reason for the higher ratio of delusional disorder to schizophrenia in the original sample than in Western samples is because the population was drawn from an early intervention service specifically for adult-onset psychosis patients aged ⩾26 years, resulting in more patients with delusional disorder (mostly late onset) but fewer with schizophrenia (mostly early onset). (b) The Chinese population differs in some ways from the Western population. It has been suggested that the substance abuse rate is lower in the Hong Kong Chinese than in the West (Lam et al. Reference Lam, Stewart and Ho1996) or other Asian countries such as Japan and South Korea (WHO, 2004). A population-based survey conducted in Hong Kong has found that only 5% of males and 1.8% of females had used psychotropic substances in the last 12 months (Lau et al. Reference Lau, Kim and Tsui2005). Our local data from a randomized controlled trial on medication discontinuation in remitted first-episode psychosis patients aged 18–65 years show a low prevalence of co-morbid substance abuse (3%), personality disorders (1%) and affective disorders (4%) (Chen et al. Reference Chen, Hui, Lam, Chiu, Law, Chung, Tso, Pang, Chan, Wong, Mo, Chan, Yao, Hung and Honer2010). (c) The low scores of symptoms is because of the recruitment of patients after clinical stabilization. (d) Occupational and social functioning improvement may not follow immediately after clinical stabilization. Instead, it takes some time to re-establish social relationships with friends/families and return to the workforce; functional recovery is likely to happen upon symptoms remission which is not reflected in the time-frame of the reported findings.

We believe more research is needed to examine if delusional disorder and schizophrenia are distinct psychotic disorders rather than variants with different manifestation of psychopathology and distribution of age at onset. We respectfully argue that it is both appropriate and necessary to use the methods and samples we have employed. In light of the limitations discussed, we believe these definitive data have provided a step forward for future revisiting of the diagnostic classification system.