b iframe width height src https www
r> Methods We assessed the administrative database containing Acetylcysteine registration of all episodes (comprising hospitalizations and ambulatory episodes) occurred in public hospitals in Mainland Portugal between January 1, 2011 and December 31, 2015. This database was provided by the Portuguese Healthcare System Central Administration (Administração Central do Sistema de Saúde). For every episode, the database contains information regarding the respective main diagnosis (clinical condition responsible for patient’s admission) and accessory diagnoses; diagnoses had been coded after discharge with International Classification of Diseases, 9th Clinical Modification (ICD-9-CM) codes. In order to assess episodes with associated diagnosis of malignant neoplasm of skin, we identified all hospitalizations flagged with the ICD-9-CM codes 172.x and 173.x as main or accessory diagnosis. These codes correspond, respectively, to the diagnoses of “MM of skin” and “other malignant neoplasm of skin”. We compared MM and NMSC hospitalizations and ambulatory episodes over their annual frequencies, inpatients’ age and sex distributions, hospital costs (costs were indirectly calculated for each hospitalization and ambulatory episode, using a classification system based on Diagnosis Related Groups – All Patient (AP-) DRG Version 27 –, which mostly takes into account diagnoses, performed procedures, and inpatients’ demographic characteristics), length of stay, and in-hospital mortality (these latter two variables only concern hospitalizations, but not ambulatory episodes). Hospitalizations were defined as episodes with hospital stays lasting for at least 24 h, while ambulatory episodes encompass medical diagnosis and/or therapeutic procedures lasting less than 24 h . Subgroup analyses were performed for those episodes classified with “skin-cancer related DRG” (as those were episodes for which skin cancer was probably the main condition) - these DRG are listed in Supplementary Table 1. Additionally, for each type of skin cancer, we determined the frequency of episodes according to the anatomic location, occurrence and location of metastases, and performed procedures. For NMSC episodes, we performed separate analyses for BCC and SCC; nevertheless, information on the subtype of NMSC was only available for the period between 2013 and 2015. We were able to estimate the number of individual patients by identifying episodes which shared the same patient’s number, sex, birthdate and residence. We subsequently estimated the frequency of patients with MM and NMSC treated in public hospitals per 100,000 inhabitants (population data was provided by the National Institute of Statistics) , as well as average costs per patient (total costs – as calculated for hospitalization and ambulatory episodes – were divided by the number of patients). Additionally, we assessed the frequency of patients with skin cancer in each anatomical location according to their sex and age. Categorical variables were described using absolute and relative frequencies; continuous variables were described using means and standard deviations or medians and interquartile ranges. Categorical variables were compared using the chi-square test, while continuous variables were compared using the Mann-Whitney U test. We performed linear regressions to identify variables associated with increased costs of melanoma and NMSC – independent variables (namely, sex, age, type of episode, presence of metastases, and neoplasm location and subtype) were firstly tested individually with simple linear regressions. Independent variables with marginal association (p < 0.10) with hospital costs were subsequently introduced in multiple linear regression models. Statistical analyses were performed using IBM SPSS Statistics®, version 24 (IBM, Armonk, NY).