In this population-wide study, we found important differences in kidney disease-related mortality across Italian regions, even after accounting for individual sociodemographic covariates, with the largest southern regions (Campania, Sicilia, Puglia, Calabria) at greater risk in both sexes. These four regions were the ones with the lowest gross domestic product per capita in the country in the last decades [22]. In contrast, four regions showed significantly lower kidney disease-related mortality than the average (Emilia-Romagna, Friuli-Venezia Giulia, Trentino-Alto Adige, Veneto), all located in the North and among the wealthiest areas of Italy.

Adjusting for the regional prevalence of self-reported kidney disease slightly impacted the associations, whereas taking into account differences in regional healthcare funding produced substantial attenuation of their strength in both sexes, particularly for Campania and Sicilia, the regions with the highest kidney disease-related mortality. The adjustment for the regional number of nurses and beds in dialysis units resulted in a further attenuation of the RRs and a loss of their statistical significance for all regions except for Friuli-Venezia Giulia. These results suggest that regional differences in the availability of healthcare resources, in terms of both general expenditure and size of dialysis units, in terms of beds and nursing personnel, are factors that could explain the regional differences in kidney disease-related mortality across Italian regions. The number of nurses per capita clearly may affect the quality of care of kidney disease patients, thus impacting the natural history of the disease and mortality. It is worth considering that the five regions (Campania, Sicilia, Lazio, Calabria, Puglia) found to be at higher risk in the analysis adjusted only for individual sociodemographic covariates (Model 1) are among the seven regions that are still under the healthcare deficit recovery plan, a national economic program limiting their healthcare expenditure [23]. Their healthcare funding per capita in the last years is still 15–20% lower than the rest of the country and approximately 30% lower than the northern regions (which include all regions found to be at significantly lower risk of kidney disease-related mortality). It is also important to mention that the southern regions, compared to those in the North, started from a lower funding base due to their populations' age structure being younger, the main criterion used to allocate financial resources among regions.

In the five regions with the highest kidney disease-related mortality, the healthcare deficit recovery plan has been ongoing for over a decade (it started in 2007 for Campania, Sicilia, and Lazio, in 2009 for Calabria, and in 2010 for Puglia), covering the entire observational period of this study. Therefore, such a long-term reduction in the regional resources for healthcare may have impacted the adequate delivery of care to renal patients and may have resulted in higher kidney disease-related mortality. This possibility appears strengthened by the finding that the size of the nursing personnel in dialysis units in each region contributed to explaining regional differences in kidney disease-related mortality even after considering regional differences in healthcare funding. Of note, the number of nurses in dialysis units per million residents showed significant variability across regions, with the lowest values being observed for regions with the highest kidney disease-related mortality, particularly Campania, Sicilia and Lazio, where they were less than half the national average.

Kidney diseases represent a major public health challenge because their economic burden on the healthcare system is substantial [8, 24]. Although most of the costs per patient in the renal population are related to kidney failure, earlier and less severe stages also generate financial costs and impact on the death rate and kidney failure incidence [25, 26]. It is evident, therefore, how political decisions, like the healthcare deficit recovery plan, have a very short horizon, saving money in the short term but leaving a negative long-term health and budgetary legacy.

Several studies in the USA, Canada, and Norway, have demonstrated that having access to good-quality pre-dialytic care substantially impacts mortality, thus partially explaining geographic differences between less and more socially and economically deprived areas, which may also depend on the healthcare system organization (private insurance vs. public universal healthcare system, varying levels of coverage) [27,28,29,30,31]. Our study supports these findings and highlights the disadvantages that can persist even within a public universal healthcare system like the Italian one. A primary strength of the study is that it was conducted on the whole Italian population aged 30 years or over (more than 40 million people). Furthermore, the large population gave the study great statistical power to examine differences in kidney disease-related mortality by region, also stratifying by sex, thanks to the long follow-up covering almost a decade.

Access to various data sources provided a key advantage for this study, as it allowed for the availability of a large set of data (individual, organizational, administrative, economic) that is rarely found in other studies, enabling adjustments for a wide range of variables.

Another strength is that the regional indicators of healthcare that were used (general healthcare expenditure, number of nurses in dialysis units per million residents) were derived from administrative data. Therefore, they can be considered objective measures, unlikely influenced by potential information bias.

Furthermore, while there are numerous studies analyzing the negative effects of patient-centered variables (i.e. low personal income, comorbidities, marital status, racial differences, service connection) on renal insufficiency [32], our study has added the healthcare system characteristics to the analysis, testing their impact on kidney disease-related mortality.

The study's main limitation is the potential ecological bias related to the variables measured at a regional level. In fact, other regional characteristics that are correlated with general healthcare supply, and potentially associated with kidney disease-related mortality, could explain the attenuation of differences in kidney disease-related mortality, for example the accessibility to treatment for kidney diseases or the capability of healthcare operators to provide effective treatment. Furthermore, information on the regional prevalence of CKD was estimated from self-reported information, which may not be accurate, given that the majority of affected individuals are unaware of the disease and often are diagnosed after several years [33]. Moreover, possible differences in CKD awareness between Italian regions could have affected the differences in the estimates of regional CKD prevalence. Although this issue has never been investigated, it could contribute to explaining the lack of association between CKD prevalence and CKD mortality in fully adjusted models and its small impact on regional differences in CKD mortality.

Last, the lack of awareness has been reported among CKD patients also in the presence of diabetes or hypertension, with a consequent underestimation of mortality caused by CKD and an overestimation of mortality due to these disorders [33]. It appears however difficult to evaluate whether there are differences in the extent of misclassification of deaths due to CKD across regions and if they could have affected our results. Nonetheless, even though we cannot measure the difference in awareness between residents of different regions, we can assume that the level of education included in the models may represent a proxy for cultural differences.