Response to the Editorial from Micali and Herle on the research article entitled ‘Mortality and Care of Eating Disorders’

We are pleased to note that our study entitled ‘Mortality and Care of Eating Disorders’ received attention from other researchers and stimulated the debate in the field of clinical management of Eating Disorders (EDs). In their editorial entitled ‘Gone too soon: Studying mortality in eating disorders’, Micali and Herle highlighted the importance of studying mortality in EDs, given the number of psychiatric and medical comorbidities of these disorders, as well as their increasing burden. Furthermore, the authors considered possible explanations for the discrepancy of our study's low standardised mortality ratios compared with previous observations. First, the authors considered methodological differences between our study and that of Plana-Ripoll et al., which analysed data from the entire Danish population (7,378,598 individuals) between 2000 and 2018. PlanaRipoll found that EDs were associated with 3.8-fold increased mortality compared to age and sex-matched populations. These findings were coherent with several other observations demonstrating increased mortality in patients with EDs. Indeed, the mentioned study was based on a larger population with a longer follow-up period. As highlighted by Micali and Herle, sample compositions might account for discrepancies between studies in this field. Indeed, our study included patients from a clinical population, specifically those who were under treatment within the Tuscan Eating Disorder Treatment Network (EDTN); thus, the sample was not comparable to those derived from national registries, nor to those including only hospitalised patients (which obviously showed greater mortality rate as compared to our more heterogeneous sample). Patients included in our study were all under the same protocol of treatment, while subjects included in a national registry refer to heterogenous settings of care. While it is true that the average follow-up duration in our study was not long, as compared to other studies, this does not imply that the results are not valid. In particular, our finding that mortality rates are not increased in the first years upon clinical diagnosis is robust, and uncertainty is limited to long-term risk of death. Regarding the sample composition, our study was based on a population-based sample of patients, not selected based on severity (as it is often the case in hospital-based studies), as a further demonstration of the Tuscan health system's inclusiveness. Studies with large samples based on national registries are obviously important; however, these studies do not focus on treatment strategies and clinical management and often pool patients from different care services. Undoubtedly, a larger sample size and a longer follow-up would have ensured larger statistical power, less uncertainty, and eventually more confidence in the results (which is valid for all studies). However, what is essential in epidemiology is also the representativeness of the study sample regarding the question that is addressed. Concerning the criticism raised by Micali and Herle, it must be emphasised that the sample included in our study is likely to be representative of the entire population of patients with EDs in Tuscany. Finally, the inclusion of patients with Binge Eating Disorder should be considered as a strength of the study, rather than a bias for interpretation of mortality rates, considering the high cross-over rates between these patients and those with Bulimia Nervosa. We believe that different approaches should not be considered merely in a quantitative manner; rather, they reflect different specific aims of the studies. Indeed, the primary purpose of our study was not to compare mortality in the Tuscan region with the one obtained from larger datasets in other countries. Instead, we wanted to stimulate the analysis of factors associated with public health services that possibly influence the mortality of EDs. We realise that a relevant part of the message we wanted to promote with our study was not completely understood, and comprehensibly Micali and Herle were against a conclusion which might sound like ‘EDs are not as dangerous as previous studies reported’. Apart from the methodological differences, Micali and Herle hypothesised that low mortality rates in our study could be explained by the multidisciplinary approach or by socio-demographic peculiarities of the Received: 22 January 2023 Revised: 31 January 2023 Accepted: 5 February 2023