I am grateful for the letters published by Mundt and Timms & Craig as they raise several important points. Regarding the comments by Mundt, I agree that the mental health of prison populations is of serious concern and it deserves to be urgently addressed by developing and implementing cost-effective services.

I also agree that in countries which underwent deinstitutionalisation and were included in our review, ^{Reference Winkler, Barrett, McCrone, Csémy, Janousková and Höschl1} excessively long-term hospital stays for psychiatric patients no longer commonly occur. After all, this was one of the main reasons that deinstitutionalisation was pursued. However, in the Czech Republic, for instance, 16% of in-patients with schizophrenia still stay in hospital for more than a year and hundreds remain in psychiatric hospitals for decades. ^{Reference Winkler, Mladá, Krupchanka, Agius, Ray and Höschl2} Therefore, unfortunately, our review is not just of historical value but conveys an important message for current mental health systems in the majority of Central and Eastern European countries.

I acknowledge that neither our review nor ecological studies can (dis)prove whether new cohorts of patients who became imprisoned in the era after deinstitutionalisation would have also become imprisoned if the mental care systems were still hospital based. We have also admitted that the cohort of patients followed or traced in studies included in our review are not representative of all deinstitutionalized patients. ^{Reference Winkler, Barrett, McCrone, Csémy, Janousková and Höschl1} However, what our study shows is that – contrary to some interpretations – there is scant evidence of adverse consequences for people who have been discharged from long-term institutional care. Our main point is that despite the importance of the data provided by ecological studies, these can be hardly helpful in showing whether there is a direct link between deinstitutionalisation and criminality. Moreover, it seems that ecological studies testing the Penrose hypothesis may have further important limitations, ^{Reference Tsai and Venkataramani3} and as such are arguably of inherently limited value. Indeed, linkage studies could be theoretically much more relevant, but, regrettably, Mundt does not cite any of them.

Our review ^{Reference Winkler, Barrett, McCrone, Csémy, Janousková and Höschl1} casts doubts on statements such as ‘the general prison population has increased in all the countries, and this may be linked to the processes of deinstitutionalisation and reinstitutionalisation' ^{Reference Priebe, Badesconyi, Fioritti, Hansson, Kilian and Torres-Gonzales4} or ‘changes in capacities of psychiatric hospitals and prisons appear to be linked’ ^{Reference Mundt, Chow, Arduino, Barrionuevo, Fritsch and Girala5} contained in the discussions and conclusions of some of the ecological studies. Our paper shows that at the individual level these statements have negligible empirical support, and they might be detrimental to mental health care reforms in countries of Central and Eastern Europe. ^{Reference Winkler, Barrett, McCrone, Csémy, Janousková and Höschl1} As Salisbury & Thornicroft ^{Reference Salisbury and Thornicroft6} argued, individual countries should focus on developing optimally balanced mental health care systems suitable to their setting.

There seems to be a clear consensus that substantial investment in community care is a condicio sine qua non of successful deinstitutionalization, which is why I suggest that cost-effective investments into mental health should replace the number of psychiatric beds as the ‘hydraulic’ in the updated Penrose hypothesis.

I would like to thank Timms & Craig for complementing our review with their depiction of some of the pressing issues related to the current homelessness among people with mental health problems in South London. Their insights are extremely valuable and should be considered when pursuing mental health care reforms in the countries of Central and Eastern Europe. Unfortunately, although we know that homelessness associated with mental illness is a serious problem in the Czech Republic as well, this issue is extremely under-studied and only anecdotal evidence is available. I agree with the suggestion that more research is needed to understand what enables community teams to engage effectively with ‘the most alienated and intractable’ patients. This might be especially important when it comes to the period immediately following a discharge from in-patient psychiatric care, which is associated with other concerning phenomena, such as re-admissions ^{Reference Winkler, Mladá, Krupchanka, Agius, Ray and Höschl2} and suicides. ^{Reference Winkler, Mladá, Csémy, Nechanská and Höschl7}