Population-Based Validation of a Frailty Index Using Electronic Regional Healthcare Records for Public Health Use

Population-Based Validation of a Frailty Index Using Electronic Regional Healthcare Records for Public Health Use | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Population-Based Validation of a Frailty Index Using Electronic Regional Healthcare Records for Public Health Use Paola Rebora, Giuseppe Occhino, Giuseppe Bellelli, Olivia Leoni, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6769624/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 26 Sep, 2025 Read the published version in Scientific Reports → Version 1 posted 10 You are reading this latest preprint version Abstract Background Frailty indices derived from electronic health records offer an efficient approach to identify vulnerable individuals in the general population. We aimed to validate the electronic–Regional Healthcare Database Frailty Index (e–RHD–FI) in the general adult population, describe its distribution by sex and age, evaluate its predictive validity for mortality and hospitalization, and assess the performance of frailty cut–off points. Methods We conducted a population–based study of 8,404,004 adult beneficiaries of the Lombardy Regional Health System. The e–RHD–FI was calculated from 40 health deficits using electronic health records from 2008–2018. We assessed its distribution, predictive validity for 1–year mortality and hospitalizations through multi–state analysis and multivariable models evaluating performance across subgroups. Results The e–RHD–FI distribution was highly asymmetrical (median 0.0125, first–third quartiles 0–0.0375), with 45.8% of adults having no deficits. The index was higher in older adults. Each 0.1–point increment in e–RHD–FI was associated with doubled 1–year mortality risk (HR: 2.12, 99%CI: 2.11–2.14) and 2.5–fold increased hospitalization rate (IRR: 2.51, 99%CI: 2.49–2.53). The AUC for 1–year mortality was 0.883 (99%CI: 0.881–0.884). Conclusions The e–RHD–FI demonstrates strong predictive validity for adverse outcomes in the general population and can effectively identify at–risk individuals for targeted interventions. Health sciences/Diseases Health sciences/Health care Health sciences/Medical research Health sciences/Risk factors Frailty Index Electronic Health Records Mortality Hospitalizations Prediction Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Frailty is a clinical condition with multifactorial etiology, characterized by reduced strength, endurance, and physiological function, leading to increased vulnerability to external and internal stressors. Its assessment provides valuable insights from a public health perspective identifying people requiring increased healthcare resources or who would benefit from comprehensive geriatric evaluation. Multiple studies have shown that frailty effectively detects risk heterogeneity among individuals of the same chronological age 1–3 , including middle–aged subjects 4–6 . Based on this evidence, international guidelines propose frailty as a means to stratify patients according to their risk and to identify those in need of special care 7,8 . The frailty index (FI), based on the progressive accumulation of deficits model introduced by Rockwood and colleagues 9 , represents one of the most widely implemented methods for frailty assessment. It quantifies the number of health deficits (clinical signs, symptoms, diseases, and disabilities) that accumulate with aging and correlates strongly with adverse outcomes, such as mortality and hospitalizations in different contexts 4,10 . Defined as the ratio between the number of deficits identified across multiple domains and a predetermined number of potential deficits, the FI’s methodological simplicity facilitates large–scale application, enabling identification of at–risk individuals without extensive clinical resources. The FI has demonstrated utility in cross–national frailty comparisons 11 and in frailty assessment using administrative health data 12,13 . We previously developed and validated an Electronic–Regional Healthcare Database Frailty index (e–RHD–FI) in community–dwellers with SARS–CoV–2 using electronic regional health databases (e–RHD) 14 . The e–RHD–FI was constructed according to standardized criteria recommended by Searle et al. 15 and validated against both a clinically–derived FI and the Clinical Frailty Scale 16 , using a set of clinical data prospectively collected in a multicentre study of patients hospitalized for COVID–19 17 . This preliminary validation demonstrated that e–RHD–FI had good performance in predicting in–hospital and 30–day mortality, risk of hospital admission, and worsening on the WHO clinical scale in COVID–19 patients. We aimed to assess the feasibility and validity of applying the e–RHD–FI to the general adult population using electronic–Regional Healthcare Database of Lombardy. Specifically, our objectives were: 1) to apply the e–RHD–FI to the general population and describe its distribution by sex and age; 2) to evaluate its predictive validity with respect to risk of all–cause hospitalization and mortality; and 3) to analyze the predictive value of the frailty index across different subgroups and assess the performance of the frailty cut–off points in the general population. Among the subgroups we examined mortality following acute insults (urgent hospitalization) to evaluate the ability of e–RHD–FI to capture the reduced resilience characteristic of frailty persons. Results e–RHD–FI distribution On January 1st, 2019, the Lombardy Regional Health System provided care for a total of 8,404,004 adult adults, of whom 4,060,083 (48.3%) were male. The median age was 51 years (1st − 3rd quartile 38–66, Table 1). Table 2: Adjusted Cox proportional hazards models and negative binomial regression model results on 1–year mortality and hospitalizations and 30–day mortality after an urgent hospitalization by e–RHD–FI, age and sex. The dispersion parameter for the negative binomial regression on hospitalizations resulted in 0.337. The distribution of the e–RHD–FI is highly asymmetrical, as depicted in Fig. 1, with a median [quartiles] value of 0.0125 [0, 0.0375]. Almost half of the beneficiaries (3,846,242, 45.8%) had no deficit, with an e–RHD–FI equal to 0, while 3,150,964 (37.5%) had a e–RHD–FI greater than 0 but less than 0.056 (Table 1). A total of 1,119,904 (13.3%) presented an e–RHD–FI between 0.056 and 0.13, 2.7% had values between 0.13 and 0.25 and only 0.8% exceeded 0.25 with these pre–specific cut–offs defined in the methods section. The distribution of e–RHD–FI was similar in males and females; however, as expected, higher values were observed among individuals over 65 years of age. The prevalence of the 40 individual deficits within the cohort is reported in Supplementary Table 1, also differentiated for age class. The most common deficits were hypertension (9.2%), low income (8.3%), trauma (contusion with intact skin surface, 7.9%) and the use of transportation services, including ambulances (7%). Among older adults (≥ 65 years), in addition to those already mentioned above, heart diseases, diabetes and cancer were also common with a prevalence of approximately from 11 to 14%. The geographical distribution of the median value of e–RHD–FI is reported in Supplementary Fig. 1. Predictive validity for mortality and hospitalization A total of 99,571 (1.2%) deaths and 926,623 hospitalizations affecting 710,419 individuals (8.5% of the population, overall hospitalization rate of 111.4 per 1,000 inhabitants) were observed in 2019. Mortality was different according to frailty classes, with rates of 0.3%, 3.2%, 11.8% and 24.8% in individuals with e–RHD–FI below 0.056, between 0.056 and 0.13, between 0.13 and 0.25 and above 0.25, respectively. Figure 2 reports the state occupancy probabilities over the year 2019, as estimated by the multi–state model, which considered a common baseline state for all adults, first hospitalization of the year as a possible transition state, and mortality as the absorbing state, stratified by frailty class. The Figure clearly shows the increased burden of hospitalizations and mortality for increasing levels of e–RHD–FI. In particular, among individuals with an e–RHD–FI ranging from 0.13 to 0.25, 32.2% (71,916/223,477) had at least one hospitalization and 11.8% died (26,310/223,477) during 2019. In the very frail group (e–RHD–FI ≥ 0.25) these percentages raised to 39.6% (25,105/63,417) and 24.8% (15,701/63,417), respectively, resulting in more than half individuals in this category dying or experiencing at least one hospital admission during 2019. Mortality by e–RHD–FI, stratified by age and sex, is reported in Fig. 3 (as a continuous variable), and in Supplementary Fig. 2 (by e–RHD–FI classes). In both plots, mortality increased progressively with increasing e–RHD–FI values. Among individuals under 65 years of age, we observed a steeper slope after 0.2 points of e–RHD–FI both in males and females, likely due to the small number (3,824 over 6,115,611) of younger individuals in the very frail class (Fig. 3, left panel). In older adults (Fig. 3, right panel) the relationship between e–RHD–FI and mortality was approximately linear, with males showing higher mortality than females when the e–RHD–FI exceeded 0.1. Three–year survival estimates by e–RHD–FI classes were reported in Fig. 4. Survival at 3 years was 98.6 (99% CI: 98.6–98.6), 88.3 (99% CI: 88.3–88.4), 64.6 (99% CI: 64.4–64.9), and 41.3 (99% CI: 40.8–41.9) for beneficiaries classified as fit, pre–frail, frail and very frail, respectively, according to their e–RHD–FI status as of January 1st 2019. It is worth noting that a bump in survival is evident in 2020, corresponding to the first wave of SARS–CoV–2 pandemics. When accounting for sex and age in the multivariable Cox model, each 0.1–point increment of e–RHD–FI was associated with a doubled 1–year mortality hazard (HR: 2.12, 99% CI: 2.11–2.14, Table 2). The hazard ratio of mortality, when stratifying by sex and modeling age and e–RHD–FI flexibly, is depicted in Supplementary Fig. 3. The rate of hospitalizations increased with e–RHD–FI, with an incidence rate ratio (IRR) of 2.51 (99%CI: 2.49–2.53) each 0.1–point increase, adjusting for age and sex in the negative binomial regression model (Table 2). The dispersion parameter for the model was estimated 0.337. In order to assess the predictive ability of e–RHD–FI after an acute insult (resilience) we modeled 30–day mortality in a subpopulation who underwent a hospitalization classified as urgent (n = 274,629; events = 20,114) and still found an impact of e–RHD–FI (HR for 0.1–point increment, 1.33, 99% CI: 1.30–1.35, Table 2). Predictive performance across subgroups and validation of cut–off points The AUC of the ROC curve for 1–year mortality (Supplementary Fig. 4, panel a) resulted in 0.883 (99% CI: 0.881–0.884) for e–RHD–FI. The Brier score resulted in 0.011. The cut–off value of 0.056 showed relatively good performance with sensitivity of 78%, specificity of 84% and negative predictive value of 99.7%. The positive predictive value was quite low (5.5%) due to the low incidence of mortality, but was nevertheless higher than the overall mortality rate (1.2%). With increasing cut–offs, as expected, the sensitivity decreased while specificity and positive predictive value increased (Supplementary Fig. 4, panel a). Focusing on true positives (in our case frail individuals who died within one year) it might be useful to balance the positive predictive value and sensitivity (Supplementary Fig. 4, panel b). Sensitivity was higher at lower thresholds, although precision was low and increased with higher thresholds. The F1 score, i.e. harmonic mean between the positive predictive value and sensitivity, resulted higher (0.22) for the 0.13 cut–off, with a sensitivity of 42% and positive predictive value of 15%. ROC curves for 1–year mortality by sex and age class were reported in Supplementary Fig. 5, with similar values of AUC by sex (panel a). When restricting only to older individuals, AUC resulted in 0.775 (panel b). Discussion In this population–based study, we demonstrated the predictive validity of the Electronic–Regional Healthcare Database Frailty index (e–RHD–FI) using electronic regional health databases (e–RHD) in the adult population of a region of Italy. We successfully applied the e–RHD–FI to over 8.4 million adults in the Lombardy region, demonstrating its feasibility in a general population setting. The distribution analysis revealed that nearly half of the population were fit (e–RHD–FI = 0), while only a small percentage (3.5%) could be classified as frail or very frail according to established cut–offs. As expected, we observed higher FI values among older adults, confirming the age–related nature of frailty accumulation. However, the index detected varying degrees of deficit accumulation across all adult age groups. The e–RHD–FI demonstrated robust predictive validity for both mortality and hospitalization outcomes. Each 0.1–point increment in the index was associated with a doubled risk of 1–year mortality (HR: 2.12) and a 2.5–fold increase in hospitalization rate (IRR: 2.51), even after adjusting for age and sex. The multi–state model analysis clearly illustrated how increasing levels of frailty associated with progressively higher probabilities of hospitalization and death, with more than half of very frail individuals experiencing at least one of these outcomes within a year. Our analysis validated the previously established cut–off points for frailty classification, with the ROC curve analysis showing good discrimination ability for 1–year mortality (AUC: 0.883). The cut–off of 0.056 demonstrated good sensitivity (78%) and specificity (84%), while the F1 score suggested that 0.13 might represent a good threshold when balancing sensitivity and positive predictive value. The index maintained strong predictive performance across sex subgroups, though with some reduction in discrimination power when restricted to older adults (AUC: 0.775 and Brier Score: 0.034), likely due to the higher baseline risk in this group. Importantly, the e–RHD–FI also demonstrated predictive ability for mortality following acute insults (urgent hospitalizations), supporting its utility in assessing resilience—the capacity to recover from stressors. This finding is consistent with the theoretical understanding of frailty as a state of increased vulnerability and reduced physiological reserve. Our results align with international experiences where electronic health data–derived frailty indices are increasingly used for public health purposes. For instance, the electronic Frailty Index (eFI) in the UK, developed by Clegg and colleagues 12 , is integrated into NHS guidelines to identify frail older adults in primary care and guide geriatric assessments. Similarly, various Claims–Based Frailty Indices (CFIs) in the US, such as the VA–FI and the Kim–CFI, serve as validated tools for research and population health management 18,19 . The administrative data approach offers significant advantages in scalability and efficiency, enabling continuous, low–cost monitoring of the entire assisted population. As highlighted by Silan et al. in the Italian context, such indicators address the need to stratify the population based on health needs for better intervention planning and resource allocation 20 . Our study confirms the e–RHD–FI's potential for this purpose in Lombardy. While different CFIs might identify slightly different target populations, their similar predictive performance suggests that using a validated index like the e–RHD–FI provides a robust tool for population–level risk prediction and analysis 21 . A notable aspect of our study is the e–RHD–FI's ability to identify and stratify risk across the entire adult population, including middle–aged individuals (< 65 years). While frailty prevalence is significantly higher in older adults, our index also detected deficit accumulation in younger age groups. In fact, mortality risk progressively increased with higher e–RHD–FI values even among those < 65, suggesting frailty's negative impact begins well before the conventional threshold of 65 years. This observation is consistent with studies emphasizing a life–course perspective on frailty. In a large UK Biobank study of 493,737 participants aged 37–73 years, 3% of participants were frail and 38% pre–frail, with significant associations between frailty and mortality across all age strata, including middle–aged adults 5 . Similarly, in a Korean prospective cohort from age 53, frailty transitions and burden significantly impact mortality in middle–aged/younger–old adults 22 . A Taiwanese study found that clinically detected frailty predicts unplanned readmissions in patients aged ≥ 50 (17% of the study sample < 65) 23 . Frailty prevalence in younger and middle–aged adults is increasingly recognized, though variable depending on definition and population 6 . These studies highlight that vulnerability associated with frailty is not exclusive to advanced age. Using administrative data derived instruments like the e–RHD–FI allows for early identification of individuals showing deficit accumulation before age 65. While identifying frailty might present greater challenges in younger populations using administrative data, our study indicates risk signals can be detected 24 . Early identification of frailty trajectories in middle age could have significant implications in prevention. Targeted interventions implemented at this stage could potentially modify the aging trajectory, slowing or reversing progression towards severe frailty and reducing future disability and healthcare burden. The e–RHD–FI, applied routinely, could act as an early warning system, flagging middle–aged individuals at risk for further assessment and personalized preventive interventions. Beyond public health applications, the validation of e–RHD–FI offers significant implications for pharmacoepidemiology. Studies using large administrative health databases often face ‘confounding by frailty’, especially when examining drug effects in older or multimorbid populations. Frailty, as a state of increased vulnerability, influences both treatment propensity and outcome risk, acting as a strong confounder. Inadequate adjustment for frailty can bias estimates of drug effectiveness and safety 25 . The e–RHD–FI, developed and validated using the same regional administrative data frequently used for pharmacoepidemiologic studies in Lombardy, represents a particularly suitable covariate. Its strong association with mortality and hospitalization confirms its validity as a summary measure of health status and vulnerability. Incorporating the e–RHD–FI (preferably as a continuous variable or using meaningful thresholds) into pharmacoepidemiologic models can enable more effective control of confounding by frailty, leading to more accurate treatment effect estimates. A preliminary study on the cardiovascular safety of targeted therapies in rheumatoid arthritis in Lombardy already included e–RHD–FI as a covariate and possible effect modifier 26 . The use of electronic health databases for estimating frailty has several limitations. First, these databases primarily focus on administrative issues and lack comprehensive clinical information needed to directly assess all frailty domains. While we included deficits covering various domains, including proxies for socioeconomic status and functional impairments often missing in electronic FIs, capturing aspects like physical performance, psychological well–being, and social factors remains challenging. This was evident in Fig. 3 , where at the same level of frailty, males over 65 years of age showed higher mortality compared to females, likely due to socio–demographic, occupational, and lifestyle factors associated with male sex that are not captured by the index. Second, information might be inaccurate or missing due to coding errors, inconsistent reporting, or use of private healthcare, although Italy's universal health system mitigates the latter concern. Our previous validation found administrative data deficits are sensitive but can lack specificity, potentially leading to misclassification 14 . Third, the index might overlook frail individuals who rarely seek medical attention. Finally, we measured frailty at an index date and did not consider changes over time in this specific analysis, although the data source allows for longitudinal assessment. Despite these limitations, this population–based study demonstrates that the e–RHD–FI, derived from regional electronic health databases, is a feasible and valid tool for assessing frailty in the general adult population of Lombardy. It shows strong predictive validity for mortality and hospitalization, performs well across different subgroups, and corresponds with established frailty cut–offs. The e–RHD–FI offers valuable insights on a large scale and can identify subjects for more comprehensive clinical assessments or targeted public health interventions. Furthermore, it represents a valuable resource for improving the validity of pharmacoepidemiologic research by enabling better control for confounding by frailty. Its application supports evidence–based health planning and preventive strategies aimed at mitigating the impact of frailty on individuals and the healthcare system. Methods Data source and population Adult beneficiaries of the Lombardy Regional Health System are our target population. We retrieved adult beneficiaries on January 1st, 2019 by the Population Register (Demographics NAR) and we followed them up to death, emigration/end of assistance in the region or last available follow–up time (December 31st, 2021), whichever occurs first. The study was approved by the Ethical Committee of the University of Milano-Bicocca (approval number 920, protocol n. 0065413 dated 03/02/2025 – UOR:003406), which also waived the requirement to obtain informed consent. All methods were performed in accordance with relevant guidelines and regulations. Electronic–Regional Healthcare Database Frailty index (e–RHD–FI) The electronic–regional healthcare database frailty index (e–RHD–FI) previously developed 14 was applied considering the dates of January 1st, 2019 as the index date. The e–RHD–FI was developed, according to Rockwood’s approach, with 40–items/deficits and considering a recall period from 1 to 10 years, according to different types of health deficits and determinants. The e–RHD–FI was calculated as the ratio between the sum of health deficits and determinants and the total number of health deficits and determinants included in the index (theoretical range 0–1). For each patient, all deficits contributed to the index with 1 point, except income, which was classified into three categories: upper–middle/high income (no contribution to the index), lower–middle income (0.5 contribution), and low income (1 point contribution). Health deficits and determinants were retrieved by cross–referencing the different data sources in the e–RHD. Health deficits were defined as symptoms, signs, disabilities, and diseases, covering a range of systems and domains (i.e., physical, mental and social). The co–payment exemptions database, Integrated Home Assistance, and Intermediate Assistance Observation Sheet datasets were evaluated to obtain a proxy for socio–economic status and to identify nursing home residents. Physical disabilities were retrieved from exemptions, Integrated Home Assistance, and prosthesis delivery service. Mental illness data were used to assess the mental domain 20,27 . Full details on data source and algorithm development are reported elsewhere 14 . End–points The primary end–point is 1–year mortality. We will also consider the number of hospitalizations occurring in 1 year, a combined end–point of death and hospitalizations and a longer–term mortality end–point (3 years). Mortality was retrieved by the Population Register, while for hospitalizations we used the Hospital Discharge Form. We considered all hospitalizations except for those due to delivery (DRG codes between 370 and 391). Statistical methods The e–RHD–FI was computed on adult beneficiaries of the Lombardy Regional Health System on January 1st, 2019 by using electronic–regional healthcare databases from 2008 up to 2018. Variables included in the e–RHD–FI, along with their prevalence, were reported in the overall population as well as for adults younger than 65 years and those aged 65 years or older. The resulting e–RHD–FI was described by median and quartiles and represented by histogram. Previously identified/used cut–offs were used to classify adults as fit (< 0.056 14 ), pre–frail (0.056–0.13), frail (0.13–0.25) and very frail (≥ 0.25). The median value of e–RHD–FI was computed also across different territories of Health Protection Agencies in Lombardy and reported in a map to depict the geographical distribution. Proportions of 1–year mortality and first hospitalization were graphically represented by e–RHD–FI (< 0.056, 0.056–0.13, 0.13–0.25 or ≥ 0.25), age (< 65 or ≥ 65 years) and sex classes. To describe hospitalizations and mortality at different levels of frailty we used a multi–state (illness–death model) model with three states: 1) a baseline state event–free common to all adults, 2) first hospitalization and 3) mortality as an absorbing state. We accounted for age and sex and stratified the model on frailty class. The relationship between e–RHD–FI and 1–year mortality was modeled by a logistic regression model with a restricted cubic spline with three knots placed on 10th, 50th and 90th percentiles, respectively, and stratifying by age (< 65 vs. ≥ 65 years) and sex. Survival up to three years was estimated using the Kaplan–Meier method and compared across frailty classes using the log–rank test. The association between the e–RHD–FI and 1–year mortality was evaluated by Cox proportional hazards regression model adjusting by continuous age and stratifying by sex. Restricted cubic splines with three knots have been used to relax the linearity in the relationship between the e–RHD–FI and the logarithm of the mortality hazard. The Brier score was computed in the overall population, as well as for adults younger than 65 years and those aged 65 years or older. The number of hospitalizations occurring in 2019 was also derived, and rates of hospitalizations were computed at different levels of e–RHD–FI and compared by a negative binomial regression model including sex and continuous age with the length of follow–up used as offset. Furthermore, we restricted the population to those who underwent an urgent hospitalization during 2019, and evaluated the association between e–RHD–FI and 30–day mortality after that hospital admission by a Cox proportional hazards regression model adjusting by sex and age. The performance of the frailty index in predicting 1–year mortality was assessed by the area under the Receiver–Operating–Characteristic (ROC) curve (AUC) and the Precision–Recall curve. The performance of the three cut–offs previously used were reported by specific values of sensibility (also called recall) , specificity, Positive Predictive Value (PPV, also called precision ), and Negative Predictive Value (NPV). For each cut–off the harmonic mean between precision and recall was reported (F1 score). Furthermore, AUC values were also computed stratifying the population according to sex and age (< 65 vs. ≥ 65 years). Declarations Author contributions P.R. and C.A.S. conceived the study. P.R. and G.O. curated and analyzed the data. P.R., G.O., and C.A.S. developed the methodology and software. P.R. acquired funding and administered the project. P.R., G.B., M.G.V., S.G., and C.A.S. supervised the work and validated the results. P.R., G.O., and C.A.S. wrote the original draft. All authors reviewed and edited the manuscript. Data availability statement The data that support the findings of this study are available from Regione Lombardia but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the corresponding author upon reasonable request and with permission of Regione Lombardia. Funding Declaration P.R. and G.O. were partially funded by the European Union – Next Generation EU – NRRP M6C2 – Investment 2.1 Enhancement and strengthening of biomedical research in the NHS project code PNRR–MAD–2022–12376033, title “Evidence–based models for high impact chronic disease prevention and risk of progression management in outpatient community services and community hospitals: towards eHealth integrating stratification on individual history with predictive models of disease progression, using machine learning and artificial intelligence on administrative and clinical databases”. Additional Information The authors declare they have no conflict of interest. References Rutenberg, A. D., Mitnitski, A. B., Farrell, S. G. & Rockwood, K. Unifying aging and frailty through complex dynamical networks. Experimental Gerontology 107 , 126–129 (2018). COVID–19 Monza Team members et al. Frailty index predicts poor outcome in COVID–19 patients. Intensive Care Med 46 , 1634–1636 (2020). Ofori–Asenso, R. et al. Global Incidence of Frailty and Prefrailty Among Community–Dwelling Older Adults: A Systematic Review and Meta–analysis. JAMA Netw Open 2 , e198398 (2019). Galimberti, S. et al. Effect of frailty on 6–month outcome after traumatic brain injury: a multicentre cohort study with external validation. The Lancet Neurology 21 , 153–162 (2022). Hanlon, P. et al. Frailty and pre–frailty in middle–aged and older adults and its association with multimorbidity and mortality: a prospective analysis of 493 737 UK Biobank participants. The Lancet Public Health 3 , e323–e332 (2018). Loecker, C., Schmaderer, M. & Zimmerman, L. Frailty in Young and Middle–Aged Adults: An Integrative Review. J Frailty Aging 10 , 327–333 (2021). Società Italiana di Medicina Generale (SIMG). Linea Guida Inter–Societaria per La Gestione Della Multimorbilità e Polifarmacoterapia . https://www.simg.it/linea–guida–inter–societaria–per–la–gestione–della–multimorbilita–e–polifarmacoterapia–2021/ (2021). Farmer, C., Fenu, E., O’Flynn, N. & Guthrie, B. Clinical assessment and management of multimorbidity: summary of NICE guidance. BMJ i4843 (2016) doi:10.1136/bmj.i4843. Rockwood, K. & Mitnitski, A. Frailty in relation to the accumulation of deficits. J Gerontol A Biol Sci Med Sci 62 , 722–727 (2007). Liu, X. et al. Association between frailty index and mortality in depressed patients: results from NHANES 2005–2018. Sci Rep 15 , 3305 (2025). O’Donovan, M. R. et al. Comparing frailty prevalence between countries: validation of the Global Burden of Disease study Frailty Index (GBD–FI) in the survey of health, ageing and retirement in Europe. Age and Ageing 52 , afad214 (2023). Clegg, A. et al. Development and validation of an electronic frailty index using routine primary care electronic health record data. Age Ageing 45 , 353–360 (2016). Kim, D. H. Measuring Frailty in Health Care Databases for Clinical Care and Research. Ann Geriatr Med Res 24 , 62–74 (2020). Rebora, P. et al. Development and validation of an electronic database–based frailty index to predict mortality and hospitalization in a population–based study of adults with SARS–CoV–2. Front Med (Lausanne) 10 , 1134377 (2023). Searle, S. D., Mitnitski, A., Gahbauer, E. A., Gill, T. M. & Rockwood, K. A standard procedure for creating a frailty index. BMC Geriatrics 8 , 24 (2008). Rockwood, K. A global clinical measure of fitness and frailty in elderly people. Canadian Medical Association Journal 173 , 489–495 (2005). on behalf of the FRACoViD Team et al. The effect of frailty on in–hospital and medium–term mortality of patients with COronaVIrus Disease–19: the FRACOVID study. Panminerva Med 64 , (2022). Orkaby, A. R. et al. The Burden of Frailty Among U.S. Veterans and Its Association With Mortality, 2002–2012. J Gerontol A Biol Sci Med Sci 74 , 1257–1264 (2019). Kim, D. H. et al. Measuring Frailty in Medicare Data: Development and Validation of a Claims–Based Frailty Index. J Gerontol A Biol Sci Med Sci 73 , 980–987 (2018). Silan, M., Caperna, G., Boccuzzo, G. & Link to external site, this link will open in a new window. Quantifying Frailty in Older People at an Italian Local Health Unit: A Proposal Based on Partially Ordered Sets. Social Indicators Research 146 , 757–782 (2019). Sison, S. D. M. et al. Claims–Based Frailty Index and Its Relationship With Commonly Used Clinical Frailty Measures. J Gerontol A Biol Sci Med Sci 79 , glae094 (2024). Son, M. K. & Lee, K. Frailty transition and burden on mortality risk in middle–aged and older population: a prospective cohort study. Sci Rep 14 , 26498 (2024). Kan, S.–W., Yen, H.–Y., Chi, M.–J. & Huang, H.–Y. Influence of physical function and frailty on unplanned readmission in middle–aged and older patients discharged from a hospital: a follow–up study. Sci Rep 15 , 10003 (2025). Lahbib, H. et al. Development and indirect validation of a model predicting frailty in the French healthcare claims database. Sci Rep 15 , 11344 (2025). Kim, D. H. & Schneeweiss, S. Measuring Frailty Using Claims Data for Pharmacoepidemiologic Studies of Mortality in Older Adults: Evidence and Recommendations. Pharmacoepidemiol Drug Saf 23 , 891–901 (2014). Silvagni, E. et al. Frailty and cardiovascular safety of JAK inhibitors versus TNF inhibitors in rheumatoid arthritis: a real–world comparative study of drug effects and patient profiles. Front. Pharmacol. 16 , 1565909 (2025). Bertini, F. et al. Predicting Frailty Condition in Elderly Using Multidimensional Socioclinical Databases. Proceedings of the IEEE 106 , 723–737 (2018). Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterialtobesubmitted.docx Cite Share Download PDF Status: Published Journal Publication published 26 Sep, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 25 Jul, 2025 Reviews received at journal 24 Jul, 2025 Reviewers agreed at journal 24 Jul, 2025 Reviews received at journal 09 Jul, 2025 Reviewers agreed at journal 01 Jul, 2025 Reviewers invited by journal 26 Jun, 2025 Editor assigned by journal 26 Jun, 2025 Editor invited by journal 20 Jun, 2025 Submission checks completed at journal 04 Jun, 2025 First submitted to journal 04 Jun, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6769624","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":477986024,"identity":"c20c0995-18c7-4c23-9195-76c0d89db4e3","order_by":0,"name":"Paola Rebora","email":"","orcid":"","institution":"Bicocca Center of Bioinformatics, Biostatistics and Bioimaging (B4 Center), School of Medicine and Surgery, University of Milano-Bicocca","correspondingAuthor":false,"prefix":"","firstName":"Paola","middleName":"","lastName":"Rebora","suffix":""},{"id":477986025,"identity":"2d235205-734a-4009-b758-f2db7869af3e","order_by":1,"name":"Giuseppe Occhino","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/UlEQVRIiWNgGAWjYDACZhS2AQMDP5BxgMHgAAMb0VokG2BacOpB1w5UzAC2CJc18u48xh8+7rgjx8De/PBzQcEdOeMbuQcP3Si4w8An34BVi+FhHjPJmWeeGTPwHDOWnmHwzNjsRl7C4RyDZzgdZtjMY8bM23Y4sUEih42Zx+Bw4rYbOQZALYfxaTH+/LftcD1cy+YZBLTIA5VJM7YdTmCAadkgQUCLATNbmWTvmWeGbSC/ALUYS5x5A9LyjIeNLQG7Lf2HN3/4ueOOPD8oxHj+HJbjb88x/pzz546cfPMB7LaAhBkbsMQ1D3ZnAW1pgGoZBaNgFIyCUYATAABUnVqtJAFVYgAAAABJRU5ErkJggg==","orcid":"","institution":"Bicocca Center of Bioinformatics, Biostatistics and Bioimaging (B4 Center), School of Medicine and Surgery, University of Milano-Bicocca","correspondingAuthor":true,"prefix":"","firstName":"Giuseppe","middleName":"","lastName":"Occhino","suffix":""},{"id":477986026,"identity":"d7dddec8-9310-4fa5-9617-a5b8f217f45a","order_by":2,"name":"Giuseppe Bellelli","email":"","orcid":"","institution":"School of Medicine and Surgery, University of Milano-Bicocca","correspondingAuthor":false,"prefix":"","firstName":"Giuseppe","middleName":"","lastName":"Bellelli","suffix":""},{"id":477986027,"identity":"48dbc0e5-9956-4e98-86fa-2d4ac1bf167f","order_by":3,"name":"Olivia Leoni","email":"","orcid":"","institution":"Direzione Generale Welfare, Regione Lombardia","correspondingAuthor":false,"prefix":"","firstName":"Olivia","middleName":"","lastName":"Leoni","suffix":""},{"id":477986028,"identity":"a6955a7d-f14c-4350-9f38-01922c029e9b","order_by":4,"name":"Silvano Casazza","email":"","orcid":"","institution":"Fondazione IRCCS San Gerardo dei Tintori","correspondingAuthor":false,"prefix":"","firstName":"Silvano","middleName":"","lastName":"Casazza","suffix":""},{"id":477986029,"identity":"81290f2c-3539-4f1e-8c15-a7b6fc8b6896","order_by":5,"name":"Maria Grazia Valsecchi","email":"","orcid":"","institution":"Bicocca Center of Bioinformatics, Biostatistics and Bioimaging (B4 Center), School of Medicine and Surgery, University of Milano-Bicocca","correspondingAuthor":false,"prefix":"","firstName":"Maria","middleName":"Grazia","lastName":"Valsecchi","suffix":""},{"id":477986030,"identity":"a40a5e5b-e6b9-4b68-b2a0-ccf29546f06c","order_by":6,"name":"Stefania Galimberti","email":"","orcid":"","institution":"Bicocca Center of Bioinformatics, Biostatistics and Bioimaging (B4 Center), School of Medicine and Surgery, University of Milano-Bicocca","correspondingAuthor":false,"prefix":"","firstName":"Stefania","middleName":"","lastName":"Galimberti","suffix":""},{"id":477986031,"identity":"5da6028b-7701-4e24-93fd-052f74fa3761","order_by":7,"name":"Carlo Alberto Scirè","email":"","orcid":"","institution":"School of Medicine and Surgery, University of Milano-Bicocca","correspondingAuthor":false,"prefix":"","firstName":"Carlo","middleName":"Alberto","lastName":"Scirè","suffix":""}],"badges":[],"createdAt":"2025-05-28 15:38:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6769624/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6769624/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-18611-9","type":"published","date":"2025-09-26T15:56:52+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":85728661,"identity":"097c21b8-10fa-4dfd-8cc6-bf27f561673e","added_by":"auto","created_at":"2025-07-01 06:58:51","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":48089,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of e–RHD–FI in adult beneficiaries of the Lombardy Regional Health System cohort on January 1\u003csup\u003est\u003c/sup\u003e, 2019\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-6769624/v1/2432be5f125545d2662487f1.png"},{"id":85728663,"identity":"0d5182c6-c512-43ca-8434-8d6f6a08dd20","added_by":"auto","created_at":"2025-07-01 06:58:51","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":229089,"visible":true,"origin":"","legend":"\u003cp\u003eMulti–state model on 1–year health transitions (a) and related state occupancy probabilities by e–RHD–FI classes (b).\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-6769624/v1/35b237f1a0e1437dab4c6aeb.png"},{"id":85728665,"identity":"e0a7ff5c-482f-48a7-a23a-41a394eaf544","added_by":"auto","created_at":"2025-07-01 06:58:51","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":115646,"visible":true,"origin":"","legend":"\u003cp\u003eOne–year mortality by e–RHD–FI stratified by age and sex based on logistic regression model with restricted cubic spline with three knots placed on 10th, 50th and 90th percentiles\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-6769624/v1/06a0ed4c1f1970e50b58e51d.png"},{"id":85728669,"identity":"8496bf2c-de97-4d08-8b54-675e824ba78b","added_by":"auto","created_at":"2025-07-01 06:58:51","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":150863,"visible":true,"origin":"","legend":"\u003cp\u003eKaplan Meier curves for 3–year survival across the four e–RHD–FI classes.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-6769624/v1/02128c6e841ab4bc4ed6ee08.png"},{"id":92430411,"identity":"d26047ce-9193-4aff-b526-fdaf13189694","added_by":"auto","created_at":"2025-09-29 16:01:12","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1138439,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6769624/v1/8858c06d-8d6e-4d82-a54c-9169f1afdacb.pdf"},{"id":85728684,"identity":"c6589bc6-9b26-4582-a709-973a9282fb8c","added_by":"auto","created_at":"2025-07-01 06:58:52","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":7765881,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterialtobesubmitted.docx","url":"https://assets-eu.researchsquare.com/files/rs-6769624/v1/5ffd7b7044c698983a4b4f71.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Population-Based Validation of a Frailty Index Using Electronic Regional Healthcare Records for Public Health Use","fulltext":[{"header":"Introduction","content":"\u003cp\u003eFrailty is a clinical condition with multifactorial etiology, characterized by reduced strength, endurance, and physiological function, leading to increased vulnerability to external and internal stressors. Its assessment provides valuable insights from a public health perspective identifying people requiring increased healthcare resources or who would benefit from comprehensive geriatric evaluation.\u003c/p\u003e \u003cp\u003eMultiple studies have shown that frailty effectively detects risk heterogeneity among individuals of the same chronological age \u003csup\u003e1\u0026ndash;3\u003c/sup\u003e, including middle\u0026ndash;aged subjects \u003csup\u003e4\u0026ndash;6\u003c/sup\u003e. Based on this evidence, international guidelines propose frailty as a means to stratify patients according to their risk and to identify those in need of special care \u003csup\u003e7,8\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe frailty index (FI), based on the progressive accumulation of deficits model introduced by Rockwood and colleagues \u003csup\u003e9\u003c/sup\u003e, represents one of the most widely implemented methods for frailty assessment. It quantifies the number of health deficits (clinical signs, symptoms, diseases, and disabilities) that accumulate with aging and correlates strongly with adverse outcomes, such as mortality and hospitalizations in different contexts \u003csup\u003e4,10\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eDefined as the ratio between the number of deficits identified across multiple domains and a predetermined number of potential deficits, the FI\u0026rsquo;s methodological simplicity facilitates large\u0026ndash;scale application, enabling identification of at\u0026ndash;risk individuals without extensive clinical resources. The FI has demonstrated utility in cross\u0026ndash;national frailty comparisons \u003csup\u003e11\u003c/sup\u003e and in frailty assessment using administrative health data \u003csup\u003e12,13\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWe previously developed and validated an Electronic\u0026ndash;Regional Healthcare Database Frailty index (e\u0026ndash;RHD\u0026ndash;FI) in community\u0026ndash;dwellers with SARS\u0026ndash;CoV\u0026ndash;2 using electronic regional health databases (e\u0026ndash;RHD) \u003csup\u003e14\u003c/sup\u003e. The e\u0026ndash;RHD\u0026ndash;FI was constructed according to standardized criteria recommended by Searle et al. \u003csup\u003e15\u003c/sup\u003e and validated against both a clinically\u0026ndash;derived FI and the Clinical Frailty Scale \u003csup\u003e16\u003c/sup\u003e, using a set of clinical data prospectively collected in a multicentre study of patients hospitalized for COVID\u0026ndash;19 \u003csup\u003e17\u003c/sup\u003e. This preliminary validation demonstrated that e\u0026ndash;RHD\u0026ndash;FI had good performance in predicting in\u0026ndash;hospital and 30\u0026ndash;day mortality, risk of hospital admission, and worsening on the WHO clinical scale in COVID\u0026ndash;19 patients.\u003c/p\u003e \u003cp\u003eWe aimed to assess the feasibility and validity of applying the e\u0026ndash;RHD\u0026ndash;FI to the general adult population using electronic\u0026ndash;Regional Healthcare Database of Lombardy. Specifically, our objectives were: 1) to apply the e\u0026ndash;RHD\u0026ndash;FI to the general population and describe its distribution by sex and age; 2) to evaluate its predictive validity with respect to risk of all\u0026ndash;cause hospitalization and mortality; and 3) to analyze the predictive value of the frailty index across different subgroups and assess the performance of the frailty cut\u0026ndash;off points in the general population. Among the subgroups we examined mortality following acute insults (urgent hospitalization) to evaluate the ability of e\u0026ndash;RHD\u0026ndash;FI to capture the reduced resilience characteristic of frailty persons.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\"\u003e\n \u003ch2\u003ee\u0026ndash;RHD\u0026ndash;FI distribution\u003c/h2\u003e\n \u003cp\u003eOn January 1st, 2019, the Lombardy Regional Health System provided care for a total of 8,404,004 adult adults, of whom 4,060,083 (48.3%) were male. The median age was 51 years (1st \u0026minus;\u0026thinsp;3rd quartile 38\u0026ndash;66, Table\u0026nbsp;1).\u003c/p\u003e\n \u003cp\u003e\u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAowAAAEzCAYAAAC7R1LrAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAAF+0SURBVHhe7Z0xbxXLlrbN9xcmQwihw80mJ0AHBwT4aGICICJC4kiToUtC4IDEVyQzwUEiIsI34AcAI6ERvkI6/AZ8ZSHL2SQzP+B8/TT9+iwvqqt77917e9t+H6n27q6urlpVXV21elV116U/GjaMMcYYY4zp4f91/8YYY4wxxhSxwmiMMcYYY6pYYTwF/vnPf25cunRp4x//+Efns7Hxl7/8ZeNvf/tbt7d8SIs0zyrzlhfl/ve//73bWw3ISbq4IZlLdWPZkBZpjkFhkXMMp5Ef4BqTLu7XX39t/8dcd503L2PTmZKc1yFOQ8ZFWWY9oizWtS1c1f1z2nWi1p6fVhty1ljJNWQOYx87OzvMb6w6wtR4/Pjxcdi9vb3O90/w03HCzkqMH7e/v98dGUc8N7tZ4xqL4i+VxyrY3d1t079+/XrnY5YFZa16rfupxqrrRrz/huB+UNix94bCr7Kuk9bW1la7rbq+rHv5tJk1r9zzhCHsIpCmri1uqB9YFKUzdT1apC2M9052yy6PKRmqEzoup3C57503z6pLfecr/lmuvdpaHHJy7qJ1fp2Z6r4eYtDC2BQ0PUn7L+TXCNf59PPbb79tNJnp9n7k559/3mguaLc3O8SPLPPCuU3l2mgqbbstxz5yL+OppmnQu63T4d69e22eTxOeGsdYQxbll19+6bZOh0+fPm1cu3at3f7rX/9arKs8WauerbpucP/Fe7vGTz/9VAxbK+PTqOufP39uZQXqOmWu/fPGmLzG+vX169f2f1HevXt3XBdIk7q9TJZVjyizMf1YCe4d9RW0p2zj2H769OngaMK6MFQnOE4Z0R+SP8oM6HtVB7g+89YB6lLWEegb6CNg1mtPXaf8dT02Nzc3bt261R2dnyjTujHVfR0ptetVhZEKwE3RBxVn2Q3FacHNADTIZnpevHjRbS0PzPPLuJFmYUwDQ+N2VlnHTvHg4KDbOv+MyetZrl9nEfWJF6keTglt5suXL7u92aHPRokX6CljH4r7WFSmswZK9/v377u9P5lsDiPaKGPocqWxdC6kjg9Zlzhei2sszI0gjlkthW/fvm3/79692/6D5MFF7Rv58ENmOlApCfzHc2rKA8eVT8UHOjd2zLPEm8PmcmBfx2I5x3NiXtmO14b44778BHLLX/OEiIObDxfD61rhJIvkUxr8QwwDOo+4SBN3//799ulUYVUWsQxK8oHS4zyl2UeUW2GVFjcdHTbbuewVBngCjuUMirfPHxfLIKM8yKmcIV73+FAkP8WrtErpED95I4+EyfmLHB4eHsedy7OUH8mHzDqez9M5OKXNdqxbxDfLdcc/7veVIfGyr+OcN5ROvI46j3hyvjKxfGJY9nNeI5IHSvWLfY7H/EJML8c5D4oLR3kIlbWuNY7tvB/pq0exnON5ikvlrXN0/XDfvn1r/VReckL7uvZDKI0HDx60/0Iy4KLsMb/RxWMxbV03XLymMayuYUwnpq9wy2CoztfSJw+yNvIf5QflK9flyNWrV9s2KZ6L4QvFMV53xaFy030gOXVv9smk8+QEx4k7pkVe836NeA/GeybG0RcPaXNM+RHyx8XyUz6Ub5wssmzH9DHZjqLR0I/nBLAdYdwcf+YKNB30cTjRFHK7z1wF0NwH5hnEff4Bf/ZJR+kSbw3ClMLJP8scUXrRIXOEMJJPMilOhVU5IAMuxsG52lcZxfPZ53zFLQeKF2rxZpQO/0BYXQPlOe7H7VJe8WNbsoDSEMiic2OcQDjyAvgrHHCe5FR+//u///s4PeISsbyAeFSWMSz/sWwUl8L2yae8Av995Qucr/RUFjFf8XgJnSOZtI9jW/v8Q6mctB/ReYLzJBdpxWPIGPfZVtkC52pf5ypN4oxlmJEcKkPtq0xK+YnXHcfxnC7bgrjiPjLFa6B4VMZ9113pZ1mFylCy6Dwhv5hOvPYcY5/jko9/pVdiqH6xHfczOkcyQZYzbpeuh/YznMPxGsSnMlJ48qO4cboW5CPva1v5yNdGZcO25OQcnMLEcEDasczZjvuEj+XFuX1lAKSldHJagryoHCSX9mN6nMs+RPmVfr7eHMt5VXiVN9s6LshvjkfylIjXq+QkH/LH/HNM5ce2yOmzr/Oi3KBz5ad9HS9B3DonygOKPyJZOKZt/lUvskzsq26C6hxpKV3Fo/oR97VdgjR1LVTukOUmDvYlE9s4wmlf24SNaXIMOVSWuFhOMd3Ijz49SFichCihMDExCiCep7h0MZRxZSierwzli57RObVK1AdxU3gRydR3E3FM+SmFixUnOuRTnmI5sl+qJKDyglq8GcLGShLhmMofSDPuR4hfsg5Vdo7pOPHFPEZiPMpfdsikYzl/Og7Ek68f5Dzmcu+Try/NjMJFdH0EcuHXR5Yp74P2lV52Koca8brwH2XK+chxUk7aV1iVTV/Zi1J+SJtzavnJ6Sge/jkew8spjZhXyDLU6mWuM5EYr2SIlNKRbHIx38pbHwoXyfUr5zWTZQL2S9dX6WUXw0ZK8kU4L5clssovl3Vtv5QPjvfd9/KXjLGsORbzlOXMctTKF4iPcyRjKTz+2RFO50i+vB/lz8cA2eWX81oKL0g7ykm4vusMpWsJOU3CsB9d6Rrl9DmPMoS+fOAv8n4JnScXywGZlD7+fWmL7M+5MW45HYt5HtqPkE6pnIFzJKcgTfmxHa8h8bCvcoh5qtUb0PHMZEPSMkM3iXY+3038i0B8Tabb7VXPB2EOY3OBNp49e9b5fEf5jDSF2w5/4i9zNXB+U8Yn3BST78fGu2iZlfLaB9easAyPiVgXxpDzpMnVQ3CtNCw6Vl7ok08vYlH3iI9hgBIMj2Vu3rzZ/i9a92vMUk6l6/Lhw4d22GZdmCU/gmuTz6vNt47MWi9LZTgW2oYoIxP8z1P9qqGh3ggvIMxa/rNAeQ7N2aStuHLlSrf3I8xBREaG6nDIPAba4KbzbusJ52WajvlEudJucQ7tuaZAffnypa0Xpfb86Oio2/qTGzdutP+lYyUWqctjKdV5sYr0BWVI+qpvd+7caf+Bfl0yUPaaejb23gTCxXziFqV0z4vSfHzqzlA/P0W9EZMojBQqNykFOAt5jD3DhdaF0Esoq0aVTXMHIFcMGlT8aBCohBrzzxcY/yka+rHx8nbuPOnV8lqCsNyMhI11gJuudgNkoqw0uLWbNUPaONIcqleiJh/1jfjoAHgYKHUA6nRKx0oN/lSMLae+6wKlznxV0MDF8pnnumelg3NK16HELPWyVoZjyOWsB8rzUL9qkE8eSrhOuQ2i/Kcg1iPN+VKZDjF0/bnWz58/33jz5s1MCjPKJp146a3cnKbqAsqL5jlTF3igK3H58uX2H6Uyo2M1Fq3LY+mr86tKn/4w3jPUEfrm2GagGFIPCZvbozH3JuR6HY1F88I9n9s2QR3/+PFjt/cn+gpHH4vWm8gkCmOsIDWNVTcMNyFk651QZdKbtFyYeS+GGpK+i94HFYknUcny+vXrtgLlt8JpTKWgUAnVWPHEwoWPcnOxF23oZ4mXsOQhNvhjyrGW1wx+3HhZicX/0aNH7Q0XUfqSV9eFOGhMBHVET0FDEKfiiY1ttKJxPTN98hGX5FQHUEKNzsOHDzuf73KP6bBEvG4lGSNKb0w5DV0XOiiVme5H7hPgPN3ThKe+UU5sZ9RYcY+Wjgvd+4TjoerJkycz5SeiDjxO3KbOEt8YavUyUivDMeRy5p/yOg/1qwbXhRc+dJ1i2XKdUMTmpVSPKE/qKJ18pO860aZz/dXhIxPnxwdN4qXtnAdZ1HQ/QU4T2WS55Poiu1xfH4E/dSXWXaxjxN13jli0Lo+lr87Pkn5UYobqbB9ZYace5/uMeois0YLcd29mmajf1I8on67tIugejPcMZYRc1JPYn5Me+7mfzsxTb6IF/sQ1aCroIE3E7Xh2dPgJjYHjGsGO5zGwDeyDwuA4B+K50V9x4HR+iSHZ5Kd4M/G87GI8GtOXk3xNQ9PmM/qLnDfC4qJfDvPv//7vJ/bz8b54+8hhyUfT4RzvI2/MW96XX/zHlco4hiENyNdHKA3VEYjhON6Xz+hHONKI14DzIJZ1qdyhJB/HYnwxryViuSjfOT1c33VSWvG64HL+JXP0I/99xHC166JtEa8/xzi373rIjzB95LLI5RCPldIppZvjLJUNLuYFp3Cl657vC4hhVIb6l4M+eWI6quscW7R+gfzklGZGaeWyYD/GzT7kMCWi/CUXif6KL5f10D7U6lHMh7ajHy6Gz+WvNCKE6StT4FiMHydiWSvuvmue5dTxHL/kL8WTw5bOjftKM5YzrnS9s3wKQ9rRX3Uz+iOriGFj+jF+xSG/WI64Uj4zxJHzG+WIcCxCfEPXSWlm2SDnfWi/D4XBRRlKaUL0I0wsU12vUr5K9USU/C7x03gaY4wxpgMrzyqmQmHByVaikp+ZnlVd4/PCZC+9GGOMMecBDaUuGxSW0ksLQy8ymMXR0LIZjy2MxhhjTAPzL5kXtrW1deLt3mUS5zoKd8vLQ+W9U5inb+pYYTTGGGOMMVU8JG2MMcYYY6pYYTTGGGOMMVWsMBpjjDHGmCorVRiZbBrdFB+6PA34cOai8vMWXiyLDG9w6RhhMzqeQSadhzurZWyMMcaY9WGlCiPv1zx+/Lh9A43toa/TryN8BiGvEjErKHF8ZZ4ywFEecZUBFNJXr14dH+er9VHx400+vlBf4vr168drlvLPvpVGY4wxxiyCh6RnhI98oogtAms6xk82sETi/v6fa66yH5fP4vV/LZMInLu7u9vt/QmKJsqnlkjjHwVdC9sbY4wxxszDKIUxDnEyFCrYxjKGoqPjWudwLFjsYvxSmjTsCzqmtLWvtLRetPZ1bg4nOXOaOK1Lq6HiaPEbA3HjSOt//ud/Ot8yeTF7rVOJxZU4UB7j+q2sidy3IH2E9X/zWp18fLa0YLkxxhhjzFgGFUYUJ6xZGuJkKBTFDcWIbZQbFqvnOJYwrGNjQTliAXkNvTJ8iiUNpU3DvihvHEMGDcOyH0FJ4lyBDBqWxcImmWIYpSlroSx+ssplxWuI//qv/9r4v//7v3bR7n/5l3/pfMeBxZGyg6Ojo/Y/DtcTJ+U8xM2bN9twUpDBKwYYY4wxZlGqCqMUD1nEpEwxvw4/lByUMClXWMJmAaUoKn8onkA6GnLVcZSmsSCPhmU3Nzfbf5DSFZcDIhzDuFgdASV2niWh/vVf/7VdIeB///d/2zhm4fXr15N8cZ68cE1QtmU5RSG/fft2F8IYY4wxZnaqCiNDnBkUsDHWrllAwZJyMyVYR8e8oIIFUmkz3+/u3bvt9iygrKH0/du//dtML/OgqMbFzzU8HZXOw8PDE9bRGsiQrafz5McYY4wxRlQVRiyGKIfZYjZWeekjKlQoihrSxno5BcxHJF7m/ZVeDsmg7JEnhtoZwp1F4VsE0kMBj+mxjSwamgYUd1lfZ+Hhw4dtma4qP8YYY4w5n1QVRg1Fa7gWGD6Nb/COQS+S6HuCGiZmyDsOaYs4B68PzpMFlPAotlgT//M//7MdGmY/KkqkHZWwDHlijmQcwl4meoEnvgCjckYWlTHKOnI9efKk3R+DLLYomdF6aYwxxhgzF3+MgGByu7u7rd/Ozs6xX6O8tf5xP6Njjx8/7ny+I3+dF//l9vb2TuxDTI84JQNsbW0dH6vFmcF/iJhvnNLMSIZSOlG+6GLYmE6Og/yWzuO/7xxjjDHGmHm5xE+jYJiGPJ9wUbD0MSdyihdajDHGGGNOCyuMHQwR81kavV29KCifzKGc9fM8xhhjjDHrxuB3GM87zPXDwVTKImCptLJojDHGmPOALYzGGGOMMabKhbcwGmOMMcaYOlYYjTHGGGNMFSuMxhhjjDGmihVGY4wxxhhTxQqjMcYYY4ypYoXRGGOMMcZUscJojDHGGGOqWGE0xhhjjDFVrDAaY4wxxpgqVhiNMcYYY0yV6tKA29vb3ZYxxpiLgtt+Y0zGa0kbY4wxxpgqHpI2xhhjjDFVrDAaY8wS+Pvf/95tGWPM2cdD0sYYswT+8pe/bHz9+rXbM8aYs80kFkaepC9dunTsaCjhb3/728Y///nPdruPX3/99fi8f/zjH53vdCCDMWZ1/PLLLyfaAxx+5wHaNuVJ7RzQzuFHe0Y7Rn739/dH5zuWWW6zFHdOE/rkEcSl48hmjDFzg4VxEba2trBQtv+iaShbPxzbQ1y/fr0Nu7e31/lMg+QwxqyW2AZMfV+fNuSHfNFuid3d3T92dnbaf3j8+PGJ4zUIq7KSUzwqR+IG2tnY1kJJHuAc/DlOfGyTljHGzMNCFkaeWN+/f7/RNFQb796963w3Nn766Se0tI2mcep8Toc7d+50W8YYszzu3bvX/r9+/br9v3btWvs/BNZDHO0lTnz69Kn9f/v2bft/8+bN9p+2lTZ3zGjMq1ev2v/Lly8fy/fy5cvBUR9jjCkxt8JIo0PjA8+fP2//M7/99lvbwEEeti41WoeHh8fH84TxOPQSh1bkzz9xahvXPJ23YfDz0LQxp4/uYTm1A3FINrYVagficc4pTWXJbUwcDu5rJ2J4tSvyz23QEH/96183bt++3Z539erVzrcO7WN82N7Z2Wn/pXB+/Pix/c98/vy52+pH7Z/gwR6Ojo7af2OMmYW5FcYvX750WxsbV65c6bbK0KDfv3+/bQx5it7a2jpuvCLPnj07buQIL2j48efcvb29VlGVAqgnesCiqAY3+nMejbkx5vSQQra7u9s6kF9Umrh3uc+BNgHiceBhNKM2Q20MljgpfX3tRFSqnjx50v5/+PChPS6r3BikrD59+nTjxo0bbZtI3PHhdgwHBwftv9orvzRjjFkXVvJZHVkg9dQtq2O2+sVGXaBsatgbGF4BGuYIjTONKw0t/wpnjFkPUPJQ5lCm4gNhpqQMjoG4cShptBklcjtBWyTlUcodw8CzPmCiXCp94vz555/b7VnzwsOwlOVFQWkGDWsLt43GmHmYW2GMVsUxwyMl9DTdR2nYOhKPlyyWxpj1gtECHiBlYRyyoEUL4BAaat7c3DxWlr59+9b+i1I7IeUQJXOozVkmPEBTLiibguHxecEqS355uKZcKEvKRQ/sxhgzC3MrjDRqapT75tkMzQEaOzG8Dzd8xqwP+oRL330vy9/Dhw87n2mRMjjLULKQlZE47t69226vEpWZZNc+cyJL6CWYIVDIsXQqf/Nab40xZqEhac0rohPIc3XY19O95iFpX0/xYxpmKaayNGjCdn4Du2aJID2/9GLMctGIAUpPfLFCFrNovcuWvzFE62BWSrNlsM9y2ddOqC1ahgWOaTUo0rgSmuMtSyBObabk0igO+UTGaIWsQXjiI26Guv2QbYyZm+bpc2Gap9f2G1/R8d2vCPvxeNNwt/75G2RNY3hiWzSdxbF//JZYnz8obsIYY5ZPvH9xus8hthPxvsc/nsf9WrqvYxsSz9f9nc/RNjKU4svgzzcLa8R45HJbl4ly57CkF+OSizIiv/yVV9EnTzyH8jXGmEXx0oDGGNPAKMSsL7vMAvMReQPbVj5jzFlkJW9JG2PMOsLUGYZs85D2lGhY2MqiMeYsY4XRGHNh0Yt3fJtxWdZFlEQGcqwsGmPOMh6SNsYYY4wxVWxhNMYYY4wxVaoWxu3t7W7LGGPMRcFtvzEm4yFpY4wxxhhTxUPSxhhjjDGmihVGY4wxxhhTxQqjMcYYY4ypYoXRGHOm0HrLOFZPGYPWc44f6GY9asWT18KPcOyXX37p9sooHlxGHwfH5TWwI0PHjTHmNLHCaIw5M0gh29/fbz+Gzf+QMge3bt3qtr6DYnb//v2Nvb29Np6XL1+2SwNmCMexGsj0+PHjNh7+oxKLssj5HCMt0kR5zYzJgzHGnCZWGI0xZ4JoHdSqKdevX9/4+vVru91HSRl7/fp1+//zzz+3/8Tz9OnTdluQHgpeDVkENzc3239WjkGJlT/K4tbWVrt9+fLl9v/58+ftv0BRff/+fbdnjDHrydwKI41cHGbRdm1oxxhj5iUurSclEOUsK2AR2qnbt293e3/Sp6BF6x/LBe7s7HR7ZT59+tRtnQT/kiURYtqE+fjx47FSaYwx68rcCiPrrqqR4ylcn3PkiXqZC/kbYy4uDOsCShcPqLu7uxv37t1r/TJSxvIa0WPaJx58UUSvXr3a+ZSpxXV4eNht9cNQ+bt377o9Y4xZXyYZkqbRNsaYZcMQMla/+LBamnsI8ypjGk7uU0SnAiupFGBjjFl3ljKH8ejoqNsyxpjp0JQXFEE9qOa5hyClT1NlBHMVh2B+IyMlnKc5jFg0S3Mh4zB55sqVK93Wj2CZJE6UWtLRMDXpSXZjjFkn/NKLMeZMgCKFIqdh4kUsgH1zBvUSzFj0sksG/764PF/RGHMWscJojDkTyGKnF030UokUMD5ng7UOxRJlknnVcoKXZLAKPnz4sN1XHPjrBReslzpPVkzSwB/LoKyWbEtplUwHBwetFVP+fGZH1kONvDx79qyVIcqnPNTmZBpjzKnSNFZz0TSutMLHrmnwjrebBrMLZYwx09EoVD+0O6LkJ3SsUQw7n5NxNYpd53sShVGce3t7x+fQBgr54TLErWPEV0LtZ99xY4w5bS7x0zRUxhhz5tELMPnN6KlhPiOWwlmHsI0x5qziIWljzLlAL6UsW1lkONrKojHmomELozHGGGOMqWILozHGGGOMqWKF0RhjjDHGVKkOSW9vb3dbxhhjLgpu+40xGc9hNMYYY4wxVTwkbYwxxhhjqlhhNMYYY4wxVawwGmPMEmGpQn1QfNmw1OGvv/7a7RljzHTMrTDykVytqYorNYg0XjEMa73WoGGtxbcO0BhLtiF5OZ7zHM835qLBvaL1m0vEtZprik/pvmNb/rU0oBY2HsNFYps21J4BcbHOdOlj4rShUX7Qeti4IcVP4XCUG/Ax8WvXrh1/xNwYYyaDl17mhXVPWTeaaErrR8fjcd3VGlpTdWz4VUJetKYsa9JKTq2rzTqzEfxKEIfiMeaioHs73yeReJw1mHM7oHstr7mstkYQJq4bHamF1VrR2keGeM/H+5btWjtFXH33ufIRz49pQSmfIspPHDkd/GqyGWPMrCw8JP3o0aP2v2lMB5/qzzJ62n/37l37/+XLl/b/6tWrrYPPnz+3/8ATftNhdHsnIY7379//YF0w5rxCXX/48GG3V4YwjeJzvOTekydPNp4+fdpuAxZ79pt2a+PevXud73dYqu/58+fd3sZGoyxtvHjxots7SS3s4eHhRqOMbfz000/t/ubm5sbXr1/bbdD9D7dv3+62ypBf0sqonSSdCFbCGCdl8e3bt27vJB8+fOi2NjZu3rzZbf0JFk3KSpZHY4xZlIUVRhorNXxv3rxp/4HGPTfqAuUrDqcMKU5x+LtPKY3DwzlcHOZRAxrDIo/ClIaBOOfly5fHyvEQ5IeGv7bW7OPHj090hsacV3Qv3rhxo/3v4+PHjycUJiltOh/lC+Uu3rfA/ckDa4yfh7ioVImhsPhzXG3S69evWweSB5Dp1atXvetWS+ZSG4CyWjqPvEvJw6Go9sUfZbl169bGb7/91u39CQrn27dvuz1jjFmMSV56kSKFUiX6noxpSAlHY4alAGpPwnQKWONoxHd3d9vGscT9+/fbzoQ4UcYEyibn4s/xO3futP7sS9Flzg+NOH6lhleNbnySV4dDPpVXjpMPdSR0OurcUGgjWC4g+xtz3uhTkDIoSLLWC+5RrH60G9zHwH2K9Z52hPvn6Oio9Y9K1JUrV47DR4bC4k/ctEnctyipWenj4bKvHRKMNpTmONIeRStlhDKi7SLPtFPRsllCD959oIAeHBx0e8YYsxiTKIx3797ttr4rQChNueEXNL40+DSapQY1Qjx0CjSgNOQ07FBSsghDI88xlD7SoZNB2UQ5BWSiY9DTvywHnMfE9D6wfGSQBwWWc3FskyYNPdYK0pA/iioKbUkp7lOsjTkP1BSkWUAB4z6W4sm9xr7u4anhnoWSYogiRxtGWzLUhkV4gCwNUWdIm7hLox0R2jnkoBxo/0rtS8nPGGPmYRKFEeVJVj0acCxyfcPRgFLHk3EcMtKTfwkaT8Kr8S4pWTTiNJooZrkRR2nkfI4BFgtQpwNYGWeFPNJg49imQ8DaSnnE+YxSnjXvMWILgDmv6IGNew/H/QncxxryjXDf5nube18PihkNX1++fLn9j8qR5iJmhsIiM3MPUUyljPUphVgikW8sPECSd5UH5+KHUg0oiLRDpM0xHpZL5ZSRQl5rQ40xZlEmURjhwYMH7T8dRE0JokFGcaMhRrEaixQzXN/wFkojFj0a2/h0TmcQz4/KrGSg4e5jjJzkC0tkn2zGXDQ0miAn5QpFq3Sf5CFUKXXEw3QP2pYM53B/co9HhQnFU9NPIkNh81AyVjzkLlnqpHyWKI2wxLLAIQfWRCl8PEDrPOTkWGl0o4QUXmOMWRaTKYw06mq0pDyCOgkh6x4MDZfQaMoCqCdtFLPSUzdP7IAyqHMkEzJoGJt/bRMPcwllHdWTfmaM9RHLQRx603xHOiNZTUqT/jWX0ZiLDkokVjW1Czz0aWhY97IeBAnDQ56mwzBPUm8+6xhvWZeohUVhQzGVDIwKkG7poZE3qyVfBqtoScGtgaIah9jzS0B90J5xLmUUGXu+McaMonnSnYtGKWu/E4ZrFK7Wr2k8j78Ptru7e3xcjmON8nZiX/Gwnc8hLEQ/wpdAhhhOxPRwkjXux+OSP6I4yF8JyZ6J+cnHla4xFwXdR3vddxZ1f+iehHi/Rn/Bvabjikdwf5aOqY2Jfn1hIR7DiexfuucjyJrjjnA8tykxfzH/ai8oH+JUmBwuovDGGDMFl/hpGhYzAJYNhoyG3lwcCxbRprPwELa58HBvlb5OMCVYDJlbvcr7jdEQLJlTvPQzKxqFcftijJkKK4wzwLAPc50W7dyIB3caHYkx6wIKFVM59vb2fhhOnZKpH/ZmAcWNoeFV3usMUTO07fbFGDMlVhhnhM5HbzLOw6LnG2POFihwzGNexT2PEs4CCsu22BpjLh5WGI0xxhhjTJXJ3pI2xhhjjDHnk6qFcXt7u9syxhhzUXDbb4zJeEjaGGOMMcZU8ZC0McYYY4ypYoXRGGOMMcZUscJojLkQ8HkbPpgvt47w2S3J17dUqTHGnAaTKYx8/4tGjn/zvXNaZoOv8sbxEXCR96ckdmal9bxXTZaBfCMbRFnnqZPErfOJC4jf9ft0yMpevg5cozH1/vr166z11zqhelOKN6erNaZnQXWxJF+8j9nm+4nI9rhb394YY9aFyRRGPhYb/y8ydBD3799f6koLrIzBChmnwf7+/ql/eJyO++nTp93ejyz64WLyRz4jrBTCyiTroCxfND59+nSs6OHiyjAoYi9fvuz2ZoOHOlaBIc7d3d2Nhw8fdke+1zHuY+4zjrOUJys9jYXzpWRyfl5phrSpT9SznCdjjFk3JlMY1WDP23CfF7BIUAZZ2VgVpY7pPFLquMk3+V8mKA8oqrY0rg7KenNzs9v7Ea47ytw88GDx008/tds3btxo/8XR0VH7L0Xu5s2bM93XWDNRQksPjrJaU1+VvjHGrDOTKIxYXKK1C6Upw9O0hl7kNESjJ/HoVyMO4+BIL8bBNsShJvlB9Mep84/xkicNS/JfSrPEs2fPNra2to47AQ1H6Rxt5zjZF7XyiPF9/vy58/1OzJfoy6vyhotyqSObBZ2LI36uNXnI8sQ6oOsR5cDF9HU+/yoTtnHquPEjjlLcmb5yjfnHqYxKoDygCDx//rzzMcuGssbSx7WZmqis8RDCGsxCiqLqJKMnKIBjoE4yrEx7oHolqIc8VN6+ffv4WF97Yowxa0PzhLswTcPY/jeKEuad9j/Ccfx3dnb+aDr6drvpdLujrUno+Bz+FV+NpuH+IR78GsW13ZYsQLoKx7ZkIWyOQ7Lip/P0L7nYJ62M4uN4BD/FqfzjkDHuCx2DWB7EyzHFrzxKfqWvuGL4Ul51vsJrG5lKqGzicfZL6RMm5y3ux+M6n3/2iQeizDjywX/OJ+S4QfuKj+1SueKvMsVP4RWnwolSOZjloXKO9TmjulGDe7YURvHidO1FrFele74P0qGOSXbtA+mxrXwo/VifqGMKb4wx68CfPe6c0MCqIeW/1PDS8MUGV2FoINVYqvFUZzwGxRsbXiBt/NXgSq4oU+wIYiei9LOSkNMqkfMilI7S1z4yRDkgxxHLAznZVjzKp+TXvsKLvrzWrksJyaLjpfzGOHLe4n5OQ7Lgcv5wkRwvlOLWPvGMKVeVg1CcuS7oXMlpVofqRK4/XNdYt0twfWthSteVdPBXHcnpllC9iWHVBuFHXY/3DHAs+pGm2i9jjFkHFh6SZphGw0X8i/jyi4Z9vn37dmK4MA4HMS+MODQHsm9YMcJwD7x69aodRmSOUeT9+/cn5Do8PGz/GWJi+KnpGNr9phFv/yPXrl3rtr7DPKSmAT+WcxnEPM9THiXG5DWjuVtDHBwcdFvzo2F5rmXTKbd+uk5C/lNQKlfmwJEG9SQOVQ+R5TTLh2Fi7sOxdXQWmM9IPdB1pW6wjz91pFHi5q6LeX5khjwZY8w6s5DCSIOKYtUonseueUpuj6lDBhpcGls6azW4hI1wPMYzZiK4Og+UIN5u1JwjQVoxznv37rVzi5Dt0aNHXajxoDRK8WLOXObq1avd1uLMUx6ZRfK6Cqg/vCXKdcrXbln0lSsKwe7ubluX5pnHaVbL5cuXu61piQ8MX758OaEg6s372hxXoE5xHudH8OMYcxc/fvzY+f5JfuA1xph1YiGFkY717t273d534r46XhQDXOyohcKj2BAGOI9tGm+sQbUJ4foMRlSKUD5onOn8dS7/uGgVq1mIsvUMWZCJuFE6Sly5cqXbmp9aeSiPst7ml14yY/M6Lw8ePGj/1flJXpGV3Ldv33Zb38lWoj7rZ80qSpooxkPUylUWYx4oZrH0DFmNzPRwD6NwzfMANQSKIA8O1APgfqbuSUHU/5iHG+7VOOJCXdP9S11k9CO2TaS7qocmY4yZi0Z5m4tGIWvn3eA096tpXI/95JoO+HhuUHaaK6T5PXKKT/vEUQNZMlkWZICYVpSLbeTRPk7yAcdjnom/BGGivDENHYv7MU6d11ceEM+P21k+9mt5Zb6U9nFZrhKKI+Y9xhPTV5h4PMqgvGo/H4ecn4jCStZcLjEuhekr1xgWJ1SHctrEJ/nNcum7jyPx2uP67k2ut+oC5LhL1zTXGcUt/5I8Itf9SE47Q3jXMWPMOjG3wjgLuXGUi413HzS6uBpDx1eJOpK+TussQyc2lDdd2/OQf9Xb2NnroSI+TJizQVYYp6CmMC6CFUZjzLox2Ye7azAc0yh1KKfHrmm8B18w0DzB0qoinMtxhnPWae4Pw1lNYz/TihDmbMDwNXMuqcsePrzY6GUtTcswxpjzzkoURuYZ6u1UOT6QO7R0Hsf7lqBDYWQeEEuGrVvnzeR4PjZcejHmPHD9+vXivEHNBYSzrjCTP/IZ0dvmp70sopmf/f394zZoEWhzePCduu3h4RrZmGtrjDHrxKWm0WOIzRhjjDHGmCJVhXF7e7vbMsYYc1Fw22+MydjCaIwxxhhjqqxkDqMxxhhjjDm7WGE0xhhjjDFVrDAaY8wS4dNfY1YjmgI+98Ob1sYYMzWTKYz6LpmWz1o3+MQN8sVvP8ovurMIcmu5u5gn+UU4HjuvuDSeMctAn4rBDSkztbCqq3JaWk9wb/edm4nx5PuE+0PH8qex1M7hhr4jC8TFp7/0KSZk7js/phtdHzGM2l0+88P6/lluY4xZlMkURq1vrP+poQEc6gRKqIFmrVbe7+Ef6CT4jmOED26fJdR58RF01talwyFPfGuObwjmbyFSFoSL3xFkn28Lxk7HmKmgTm5ubrb3HvWS7wviV2IoLPVZCwBQZ1mrWcoebQMKGMdwnJsVShHDEV9sVzjn1atXx8e5P6R8kRbfV9Ux4unLC3A/sc4632XVPsqjzoes2OkYjvu6r03KeeCD8ioL7m/W267JZowxM9M0OJNAVHLLguWylAbLfA2hZdxKYZtG9swv70beyIfQmrpxm+XtgP/asmjEo3ONmYp879WWvBsKS/2M92zcJ1y+F+J+RPcEcH5Mg+0sh+6jeB4Qf18awP0W5c3xsq97LscNyFJqo/DL4YmnT25jjJmCSTQEGk0aMRqoUsMFNH46LicFhkYt+w0R4+uDuHAKFzuGPv9MKZzSVmfBv8LEDiT64+gAheLlX/lnOx7D9cmmuGOHILnito4rnT4kQ5TfmKmhfsX7oEYOS53WPUIbE4/F+wGncEPke4jz8j2AX27TaO9qaQwdh1qYWfIAhCW+COXl+9kYMxWTKIxquKWk8B/huJQRKSaxMYzn8D+2QxGcX2rQc7ySQceyKym6oOOxQZaMnKNzlTfC5XzyH+OQDPjjkEv/+ANx5LIUOi+ic9XZ6Dj+OMCPMEojUorTmCmhPlM/x1AKq/pbaiPUzoypwwqLi2nIP8J+bBskQy0d7re+e1cQpq/Nwb+UxxJ97QTxj43DGGOGWFhhRPFRo8e/GuGoXNGYxUZXYWjoaNTYlkJTarBLKM6+sDleyVZqWBVPXwegc3VccYI6D3U6bOdGOsqaFUZcJKdVgrT6wuRyIR3lOZYt5+ayUF6MWQbUxXjv1OgLS51VPVZ7IvDXfZ/vwT50v0SlUfdBdPG4wL/vPkWOfH9lasc5prZiCPJakm+MDMYYM5aFX3rhJRcmn/PSBP8ivvzCxHH49u3biTcS5Q9Pnz5t49Ci+zGcYHI6YXAPHz5Es2ldiYODg25rGOJoGt12on2Je/futS+RcLzvxRCOIxdIdr2U8uzZs/Y4HB4etv9C/oK0mo6wTYtz+ybu9/Hu3bsT5UI5adL9hw8f2n9g0nx+6UeUyt6YReGFkfjCVY1SWOos9xL1mXuE9kb3I20DbwdzDvdO7eWaCPcLHB0dtf+gF+TULjRK14m2SjQKXW+bMQQvuyjtDPcfMvDG8xC0D7woVJLPGGOmZCGFkYaNRlqNK655qm2PSfEDGngaXpRCKUiEjXA8xpMbQL1NqOMoVjWQC3hLMdLXsD548KDtGPqgAwPeRrx582a7HaHzkGx0BJQNYcnvmIY/Qt6Ih3Pjm6BibOdAJ4rc7kzMaYOy16cgZUphuQdQzi5fvtzuc49wv37+/Lnd52Ho6tWr7Tb1nXYo3/t95Ic2gTJKO4aSWkKylJAsJbgv++KEt2/fbjx69Kjb60dvXQ+1hcYYMwmNYjI3DHc0jXi39x32iRbHUIn8+oZGYnjFxXk53lmJ8YKGqhjmwbHddBTtMRiTZj4HFG/MH9s5DbZxGkbTcVyE+HCgMCW5iDfLEiEdlb9gX/Fxbr4mQ3EaMw/UqViHqZvU7RK1sNRd3RvAvo5Rl/M9GMP2Qfz5PgDd131yAvdTXxq6dzOcozYAyGuOI5dBCeLPcuf7fWwZGGPMGOZWGGnUaBBxagBp5OQnR6MlRSU7NcacH/1jg7oIWZ7Y+Ef5cWPSpPEthSOPMS4R/WIZQEw/NvSkEcP2Nfj4c7zUsfR1VhBljais3MGYqSi1BzjqPui+p74PhYUcJt+LffeU6jz3he6NUjjQvRfTFbrn5IbaDOKotTly8R4mPPJmJBdhsxxy+d7FL8ZtjDGLMLfCOAtjOoMhslKW3RiFb1HWTZki31PJpE7ImFWTlbZlIEVrlfQpf6uAvK5be2WMOdusZC1p5uw0jRfayLFrFLx2rtJYmNMUz89uWfN4mDjPyyfMoZrlRZpV0HRI7bzQvhdxxsLEeeIhPmNWhV4KY/7wMqH9YZWYsS/cTAVzl1lxJa/msmy4n5m/uer8GmPON5caZQvL0lKhAYtvUEPz5D16EvxpEmVfQVHNBZ3u/v7+XC+3oAhfv359bfNmzFmHNoQvRKxCgUMJ5wsV+jKCMcZMxUoURmOMMcYYc3ZZyZC0McYYY4w5u1QtjNvb292WMcaYi4LbfmNMxkPSxhhjjDGmioekjTHGGGNMFSuMxhhjjDGmihVGY8yFhO8j8kkqHN9qXAeQQzKt+vuNxhhTY26FMTa2OL7nJ/ggt/yX0RDHtGO6q0Yf9ZarEcONkVkfNZZTR0KaQ6h8YjpjzuuDeIhvmSifUe5SPqaCb+MpvXXomMkjMgnVLfyirPNcR10/OSCedVGSxhAVqaH6EMsLV/uwvRYUiN8tjO3LrCjtMXU2pqNFDJADeR4/ftzuG2PMujC3wshHt1mtRfDxZ/H169e2IWblkGV8QHYdPvhNJ8TqKHyAnLxCrZOY9d2iW7dutf9afWVMB9QH5yLrPNABcm2XvQrMab17RR1eh/pUU95YxYh6Ni980J3yjXHoI9KzrLZ0WqDcUofJAx+oj21NiU+fPrVh5VhxZSxcB5UXdX6W8uHcvEBBH+SHVWAkI22mMcasMwsPSUtZgmipuXr16sbly5e7vfPH58+fu63vnS+N/jwrrZSIFhE6O+LW0ohjVotQWMnDsmjzQKdGB4hSNUunOxU5H+cVFKL37993e99RnVrWkpegh7l1tzTysPPs2bN2m7qA9a3P0sq9s7m52e3NBvX95cuXG0+ePGn3qfMojDw0jYHyHPtg9eLFi5Ws/GKMMVOxsMJ48+bNY6WRTi93PjTCGnaRQpmHYThHfnn4LQ7NljqJt2/f9h4nfh1DDohp4acwOh6JaeOkyCGjLHbkuW9IM+clE8tGZYGfrIugdBUulm/MCy77cR7xYpUB/CSHwuBinBE6NZDSkvOjssvXNZcH+zqmMoRa+eR8AGHkh5PcsRxJK547K1FW0sMhZ0yb/ZLsUQ45Ec8nP0qHf9UllHPKtFYukb5ylR+u79qKR48etUpSqf6vA8pXfGC5du1au1ZyiefPn7flSN5n5cuXL631Mj6gsI3Fckooa8p8zPUxxpi14Y8F2N3d/aN5om63m6d+xhRbhz+uUVTaY2zjv7W1dWK/aZzbfdC5+HGe9jkn7gvtky4Qjn3ijvvQKLS9aUkWySqUpuJX/vrylKFc4nFkYD/GEY/zr7RK+VX6CqP4kEOy4Ac6FxkkB04Qh8qD44ozwzk5f+wrPuQknhi/9nO9IKzkhLHlgyOsykRy53QUN45tpasyySg8/0J5y+krjI7nfaWhfWRSfmLZSmb+Y/o5XuiLW/t95Yq/ygg5YvqKI6Lr0FdOpw15U35EyU/o2qn8a/miPOJxtvGLlPxqqDwlR4l4THUih+e6zZKuMcYsm8nekmY4pmn82m2e8Od9Kn/9+vWJJ/yhOZAaPnr48GH7z/lYJbD8NQ1u68fweNMgn7DCAGGxnjXl8MOwp6xrWDPiv/yHePPmTfvPPCXIw0+yHOk46WN1GMurV6/a/xs3bhznoTTEVZoWQF4oDyxUWG5KZZzLKtN0pK3Msow2nXj7H5ElhXpB2CtXrrT+WNCGyifD+eSRuV7IjfwluObzDOPGOqO6oPozFg2hHx4eVueMYgXTNZtH1lq5jrm2JQ4ODrqts42uHfWpUd7a67Bu1lPJCNTnRjm0pdEYs/ZMpjBCnLg9i/IzC7M0/igADPtoIjodeWSeOZZTdz50aMio8hobf5/CNAY6UxQ+lc+QcjgPsayRlXQ01P7t27f2f1aQk3iYz6aHk3xNM2MVoaF4xoICzcMS5Qul6ykFb1FK5TrvtV03pUpQVrmuk089qNRAYUbpPzo66nzq8GCZXz6h/kQFbxnw0OuXXowx686kCiMsoshMQWzcUSqw4sjNY83JTN15YF2IMs4aP/Ou5gHFgvQgzpkUU7+wFPM4ZE0sgUKDnFzTOJ9tnZDl88GDB53P8imV69C1PUvoWkeFFiVOlukxjK3LWOtz+0W6875EMwtjFGBjjDlNJlcYUXjy8GS2qDAUPBWyHihOntbpZFAsaPwZpgP+tS1qlgcNdctCpf+xyoDCaeg4W3ru3r3b/mNZVGfIsNRYSw+KJujtUc4bGtYiDEPhhFNZlIaSYQrFmDg0rKsheMqB7aHyyeRrNfWDiR4msMyJbPWJZUJZxrAQw9csmzVr5pihyVq5jrm2JWZRwFYNFlNNBSGf3DNjHjooB/I1ti4TjvtKZcr5XNMpHjRrcM3GTh0wxphT4485aTqsdrI2jgnamaaRPzGRO4fXNuHifqPonQjLPk77HAP8iF/+uJhePiYZc1o1NIFdjn1g0n30l0wZ8hbDaFvp5njYz3Lj+uQoxVnKn/zifgxHuiUURmR5Yxy4KA9ORL9YVrXyiXFLbu3j4vFcZlkOlVdEeYl5z+VMujFMTIdjMR2ORZniNufHvOIiMV22czn/x3/8x4l90oLop3Il3Zy2kLwRyaU415V8vYX88cMpDI5jNSgP8p+J1zVS8ovka5zrjeTJ4Ur1k7C6psYYsw5c4qdptIz5AaxojWLUWqqWbWVZNViP9I3JvrwxxIwV8bzkX/mJtzxDoXyn8yJauCgPLJCzTJGg3jBisuxpEVgduf/W4aPyxhgDkw9Jm/MDQ3QoSyhWdF7mfIHCBB4OHYdeaFrXObTGGLNMrDCaKljW9veHl2M7q6AMS3GKyBoHZ11hRnbemI5zLpmnxwPBRX87V18pGDN3lLJatnKNHMizrK9MGGPMvHhI2hhjjDHGVLGF0RhjjDHGVKlaGLe3t7stY4wxFwW3/caYjIekjTHGGGNMFQ9JG2OMMcaYKlYYjTHGGGNMFSuMxpgLCZ9O4hM2Yz+rc5bR53pwpc9IGWPMECtXGPn4rRquktP37lhRIR/jXL4fV/KvxUtcYymdH13sZKKrrYWs7+AtkyiLylCyLuMbgvH6rEMHRPnHa6AOEr8oq9YJngXi0PnUMyD+eeI670TFZKjeDYXluMo7EtuAUt3LbUft3tzZ2WlXvonfV4z3+KzovDH3RF/YUn0TpXYRF8svyq/zyR/5fNytQW+MMbOycoWRj9/SSMPW1lbbiOHkxweiafz4YHR8H4fjnMsyXtGflUjwj/GC4iU+Prxc6nhKZFkUD44PWAPbangVpm/1Bxp4wuzt7XU+ywEZTgPKfx2WL3v48GG39SPUJeravHBtKV+uo6AD/vjx4yjF4KKAIse9q3sllldmKCz3a+nj1dxPwHk4PkaOYhlhNRYdx82yMgtx8UFzzuOeHdtuAGG5HziXJQezXJG+sJTJ8+fPj2UnXHww+fbt2/ExHDJSt5EZOJ/4dJx20RhjpmBthqTv3r3bbW1sHB0ddVuLowaTTmlRixCN8izKEY231is+jeXEkJVOQ53JeYXOVsq8kEVlmeVO+aKwLFqvzgusmvLs2bN2mzrHQ1Vf2QyFzQ+Akbj2M2E+fPjQ7X23zm1ubnZ7s8H9ipL65MmTdp+6g8ImJbWGwmjNcWQkLuLMDIWNbQzKXyS2k/DmzZsTD0svXryYaW1sY4wZy8IKYxwWWYS3b9+2/zwtT93Jy7r06tWr9n8e5rEk0XiDOgY6CpUVnSOdEduKW8dyWnGIKQ6v5fgiKFH5HMLIDyerRhwCw0/pId+sRFlJD4ecWZ4oizrQKAcuph/Pp2ONZSdL1K1bt9r9UtwlFIfiBP7lhyOuGig6KD8XHdWxeO9eu3attcJmZgmb0b0krl69eqKeYJ3jIY1rNytfvnxpLZ3xAYttLJZDEObOnTvd3neIizgztbAxbcqJNisqgPE4UPdVJtRd9sm77m1jjJmKhRRGGmoaOiw5dJx01rOAdUYd88HBQRtPnwVPa77KzYIa2WyFGoLw86QnaLzjUCgNu/bJD5YRyk/lQP61r06Vhp99ZMFSiVIEHKdjJD7Oy+Q1b+lMSFPXi3/kIx46bllz8MPyQ7ykWVO4Mlx/yUoapCdlCnlIU9AJxn2QpYTzqU/8S2GL+aGzRTEAZFU8DM9Rf0pxZ5BVcpJ3deCUt8qU8h5C1qxZyuk8cnh4+EOZo8yVhkRnCTsEypcslSDLMteUe2pI4Y8w3BuVT0CR1cNEDcIQNkJcxJkZE5Z93et9cO/G9gXIO462hfyPkd0YY8Ywt8JIB0mHq46WBhBlYZYGisZOc/tQVGqdLh2AGkPcKqBTmzc9KXx9kB8UWXVQJeWEsqRckIOwV65caf0pJ4aiQENWQ8NQnE8+6JSlLJVAUZvHwkt+uf5cUynouTMbAtmQEWsz+e7j9evXrfJN2EVlBZQVygN/ZOcYiiNpjB3eKykGZrmorYl1QHWP60bbwgPLWVSadC9QL7MSK/JwtPIOnM+9bEujMWYqFh6SpmPnSVaWJA0H9bnceNPY07AB5y4DpYniReNbkkuujz7L57LACiPoNJBNFod5lRMUIuKJVrmYTomxaQ3FMxZdA9UJLM+Zy5cvd1uLgWJIeqp35AELFWmrXo9VNkpyXiR4mMkPIdSdkrIzS9gaKEPZkh6hbeGhYOyc6JKVk+saFbE+CJPrAHERZ2aWsCi9uawEdTQP0UeYi5nzY4wx87KwwkiDzJOw3O+//35iP7tS4xsb/WU8EaMYwKNHj9oGtCSX3FRMpdSIKONYq1cExQeFE0VxHqvcKpDCkCf2L4toQcap86U+qpPOc81MGdWpqGCjFOWXNmCWsH1gJa8pi5Gx9+KNGzd+UM6QccxLNITJDxfERZyZWcL2yc4ogx6qasyqhBtjTB9zK4zqXFHGNPzKfCFtz8rYoelZkQKKYjuPojUvY6wSQxCHhk01F4vyZfvBgwftvl7kGSr3bGXps1rMS6wPIls3YueFvFmGuJ871EjNYjTmgQOFBWWR9FTX+MehiCAbZa95nWOY983c8wTlpRe9KEPu5b57bpawGa4RVnLdY5xfajPwQwkdey8SDiVM9xrnU4drVjxBGMIiC1APiauU9ixhKaNSPWRahtqAPoh3rFJtjDGD/LEAjZKHSe7YNQ1bd6SfpqM+cQ6u6bjbY02j+cOx6DiXNEr+pXjldnd32/jHMOb8UhjKIqP8COKJ5+T8NsrhiX0R/QgjYlnEcymLGDf7oH1cPJ6vY5ZD1yeivMSyyfHomihMPE4a8ZpBTDfLV8qPiOWKrLmO5HInDC76ET/wH+WI15V0c9qSy3wnXqdYb+Rf8sv+0FcHo1900HdN+yCNUpsV046U/DIKQxxCcmV5SmFLdTdDfLkeQj431l2BDDE9Y4wZyyV+msbFLAGsZE3D3r7QMsZKcZbA+sLcv1resARhcTwv+ZeFNFpOmefYdNQrtV6baaB+YoGc5dpR75mDua7TOobA6ki7tOo52caYs8/CcxhNPwwxoSyhWNWGWM3ZBAVya8VTHczpwfXmMz5nVVk0xphFsMK4ZLCs7e/Xl0k7y6AMY6nJyLoIZ11hZq4ZlkSuo8BSw0tUttScbfi6A9eW6zkEluWzOieQ/JFP5ooaY8w8eEjaGGOMMcZUsYXRGGOMMcZUqVoYt7e3uy1jjDEXBbf9xpiMh6SNMcYYY0wVD0kbY4wxxpgqVhiNMcYYY0wVK4zGGLNE+Ni3lhtcNnwCaswngowxZlZWrjDy8Vu+B9bn9L0+Gtl8jHNpeEv+tXiJayyl86Pj+4Ilf60Luyj6XhpuFrnnhfTi2rlKu9TBcZxyjlAeq5DTnF1inR5SnPL93Qdx9ilGukdL3wcVMY3avZuP65ucckOQHz72Xfq4O/Ll8sjx1+6t2OapLPio+LVr16p5N8aYueCll1WjNU/71lBlrVQhP45H5B/XWo1xCNZcZb+09mofiiefg1ySWevgzhLvEIoTtO5yaT3YqUB25Ye8kR55V/5z2pItQzxD6/aaiwl1Sfeo6li8vyOq8zret+4x8REu1zmdP1QX4z2rul6SibQ5pvuAMFEetmv3P+eV5Aely39k7H2Uy4a4cluY4zbGmEVYmyHpu3fvdlsbG0dHR93W4mjd36axH7RuDMFSf8tc2YNVGJpOoN2+fPly+//8+fP2f2pkkVB+vnz50v5fvXq1dfD58+f2H7BYNB1gt3eSDx8+tLLb0mgy3NdaR5z7p1Gw2u0Sh4eH7XHCwebm5ol1u4ERCCx2uk8E/rdu3WrraG01Fix41Fchy5/qv6CtePjwYbf3HcLE+//Zs2dtu9K3ihHnEyYji2UuC+LBOjgGwrIOtqA8vn371u19zxer2PTJZowxs7KwwhiHTxbh7du37T8N39RrtapzefXqVfs/D7MO8cRhuJIrDUWV0PJ6NZBN8cZh5T4Fjk4EBY+l7caArHROfdeFDp4yLnWO5mIj5Q+4J6hz0S9y48aNEw92r1+/bl2EOEoK4YsXLzZ2dnZapZG6n6dOCOpwTh/F7cqVK93en/ci8kSk+Ao91JXyozhK9wwPgaUhavKgpQqHFD3uRymEOBTrHCf3pNpVY4xZlIUURhplGts//vhj4/HjxzMrVShDUm4ODg7aePoseGpI5WZBDTqd0SwQfp70gE6N/PS53LhjXZmXWGZ0sLIE9ilw6kRu3rzZ/oM6R6wUslRwnM4IRRt56chVHlkZpQOjvPoUX3NxoU5QZ4bWMeY+pe7qXqf+RoWrT1kE4v748ePx/UVdlBW9BvWbdiym06fQZbA4oqSWwDpfUlppI/vauCdPnrSy7+7utu1qTX7ko80l3J07d36wxAL3JO2qMcZMwdwKIwoDjTKNFTCUggI49GQc4QlYys3QkCYNszoD3CqQMryq9KagNhwn6FgzdNZ0VHTWOLbpRLm+DOHR6cufa3H//v3itV5E8TXnE+oR95DqV7auZ6SEYS0UtA0MUdeseVERI44hBRWw6sV7pqbQZXg4G6NYCvJds8Irb1gyaVuRf+gBjHzWlONZ2mNjjKmx8JA0jRrWADoCQJGQFarkcgNGZ8KTMnDuMlCaKIDxzcKS62NsJyJowEvxy+VOMw6JTQUdySzQUUlBZhsZNYQY5zNqjmOe9wVxHpUxEeoU93qf1QvliHl/KGHUQR4oZaVD0YptCw+ntD0lKx5E63kfWQkl/TjqQXsBKK75fq1ZO/ugjdSQOY77E7/SyAwyUVZ9D2Ckz0M6ZSXlckgRN8aYRVhYYaRRl5KB+/3330/sZ1eyEMSGtzYMMy+aB4jyw9BNSS65qSBPpfjlsmUiDolFKN+pKV2DDJ0nlshZLCjGDFF7qSMP43IPoQzxwJfvW+4LFCr8de9ka1zt3iEsL8/EeYmyhMrpgYtRkHgfoJj1WTuFHqoiMW4cCikWwtrDaN+DJFZ/pYEcxFMaOTDGmKmYW2FUQ4sypoaahjQ32mMZOzQ9K1JA6TzWXfmhA5RyqzfFNYSlF1umsCKMeRMTS0jsyGSxiXMc80sBMMayYy4uWNSYq1cCBYj6rxEBLNgoVWMecLh34lvNbOe3nAVtFPMU53lQ1f0XFc3SuSh6upfnQe1o34MkinV8IQhlMb41DSU/Y4yZm+ZJd24aJa/9/pdc85TbHemn6QBOnINrnuTbY02j/8Ox6DiXNEr+pXjldsP3yYYYc34pDGUxBbEMYrrKH/+ZRhk+PieXBfFlKG+O9V0vzi+VGX6KNx9HhpJs5mIT6yIuovoU62i+v/ugvuW6He+dWLd1f3CPltoPXOle0H2iezveZ9H13fvkvdYucDymG+MkrYjyprYSYtmW7vMc3hhjFuESP03DYs4AWF4ZRpt17lQJrCIMa5XerpwVLEJN59W+1BAtL8aMgbo4RZ2uQR3l6wCrHGWQJXPW+c9TIEuop5QYY6bCCuMZQcNeU3asDGvxFvSicTJU/vjx46V3+uZ8gULF1Ie9vb3eodcpmPLhaFZQ3BgaXqXSyIMlw9WnoagaY84vVhgvOHSmettyHphbyVwxWxaNKYMCx7zfVVj7UMLfvHnjhzdjzORYYTTGGGOMMVUW/qyOMcYYY4w531QtjNvb292WMcaYi4LbfmNMxkPSxhhjjDGmioekjTHGGGNMFSuMxhhjjDGmihVGY4wxxhhTZeUKIx+L5kPPfU7ryPLtsnyMc/kQbsm/Fu8sa1OXzo9Oazpnp7VfxxBlXWeQU/nie421vHJcHxcXhNX1NMYYY8zZZeUKI6st7OzstNtbW1ssFts6+bHEHEoGH4KO7+NwnHP5+G30Zzk6/GO8oHiJ7/79+63yM4Ysi+LB7e/vt/5ss7IJKMwsK1WwCsM6Q/mj7D169KjNFwr3y5cv2xU5uGaszhFBgeSc/LFgyovymUVhN8YYY8z6sTZD0nfv3u22NjaOjo66rcXRcmAoL1pfdV5++umnC7HcFssFohBrZQrWr4bLly+3ZQDRyogCWSoXwqLQo7Db0miMMcacXRZWGDVMiVuEt2/ftv9YsKZeV5Y44dWrV+3/PDAUPQ9x+LkWRx5SjwpZHAaPltIYPg4Hx+HjaN3riydCeJTrBw8edD51iBPLYx9aMvDFixftvzHGGGPOHgspjCgdGpLFIjWrUvX+/ftjBebg4KCNp8+C9/Tp0+OwuFmQVUxDymMh/DzpCcqDOHCUD/ktzf/D8kkYhsKlfLE+M6DAcR7HKR+BUqiyjwobcWn4GCfrXl88GQ2XR6V9c3Oz/cfyK0uhhqopW7ajMpqtiSjsyGSMMcaYs8ncCqMsUQxfwrVr11qFZJahRxQJKTsoFMTZB8oUio7cKohzGGeFcqA8iAOlivl9xFOynmpeJsPyeX7glStX2n/iAQ2xU96UP4oacWr+IIo14McQMmC97Ysng8wZrIQovMgmpZP8PXv2rE0XJVX+XFPVicwsdcMYY4wx68PCQ9IoeliVpKhg0ZKlqeSy0oBioxdIOHcZKE2UpTz0m10fs85dnHUeJhZDFC0p0ChfQPnID/k0DxMlEyUaRQ3/bLnET8oh1tu+eMYihReHAoy8skZ+/Pix/QeOSaHMTDk31RhjjDGrY2GFEYuSFAnc77//fmI/Ow0PR+LbtXEu3lTIasZbv1jWSnLJTU2f8hTRMDLylUDZQzbKGsVcymF8YzxbJklXeVL59sUTkZJZA3lv377dxmeMMcaY88/cCqNeZkAZk+KBIlFSQsYwdmh6VqSAoiTprd9VgDIl5UsvfJCvUt6wAIrDw8Nu6zuEVx6yYq24eBNZyFqrNFFWFbYUT6bvZRjB9cWiGMsS5RE0x5F8lx4MrGAaY4wxZ5Q/FqBR8jBvHbudnZ3uSD+NMnHiHNz+/n57rFF2fjgWHeeSRsm/FK9co1C18Y9hzPmlMJRFiRiG/EGUFT/ij/v5WPYD/qN/lDHGzzb0xZORLKX8cJ04pusViXHn47X0jDHGGLP+XOKn6dCNOQYrI/Mpa5bIsWDZZG4qQ+Qlq6Mxxhhj1p+F5zCa88eHDx8mmRrA8DXK4u7urpVFY4wx5gxjC6PpBUsjb0LPO/eQt7FtWTTGGGPOPlYYjTHGGGNMFQ9JG2OMMcaYKlUL4/b2drdljDHmouC23xiT8ZC0McYYY4yp4iFpY4wxxhhTxQqjMcYYY4ypYoXRGGOWCN8zZdnUVcC3T7UEqDHGTIkVRnOhoYPle5G4uI42na78ax8w/+WXX+ZeP514lcZQJ18Ly/rdOtYXBmKeUGCG0pyamAdkLoFcCkPZRnI+YxzxvOj6iGHy9Yvp1MpIYWrKIMc+ffp0vPZ6rG+4nHY81ldGohSWb6Zeu3bth7IzxpiF4aUXYy4qWg89roOudbFBx/vWCt/a2uo9ViOv2Z1liAyFLa2vHmViW/4R8qm1xpeN1jevQT6jPMhH+YqYL+VJ65bnsiOuvvXLYxoqu7j+eUyHOHLcOoc0ahBHlB/iebquIsuVz40MhcUvy22MMYtghdFcaKR4xM6VfXXAKBJxP4O/lItZ4LysLMT9yFDYPtlA8vcpD8TVp1hNBfKNSYNwWQlDdvKAf86nlLmo7AnClq4Lfjk8aShd4ovpqPxEVvJqULZRBsWl9Ev7gvNyfiNjwsa4jTFmUTwkbUygb3j5/fv33dY09MVXSr8WFsdxDU3m4XOGVBvFZePp06ft8TjsDo8ePWrXDR8a/pwX5Pv69Wu7LRn7yphw37596/a+g+xfvnxp/RWPYOj148ePPyw9SV4IW1rSEr8cnjSuXLnSbhPf7du3221QWMn87NmzjUapPM5L35C1wkcZiIu0dM7bt2/buJRGlOvWrVsbv/32W7f3I2PCNkpkm4YxxkyBFUZjAoeHh93W8phFORsK++bNm27rO/fv328VGeBcKZvNw+HG48eP27W9o5Jz8+bN9n9ZigXyoaRubm4ey4CCU+LOnTutYhtBXkBOtqNCfHBw0G2dBAWTuMZAGSGfFDsUzatXr7bbAiWPeoESKHnIy97eXqtsZyUdPn/+/INyDsQvBR80t1FwbXRsiKGwKL59ZWSMMbNihdGYMwyWJZQX3O7ubuf7/WULFCeQ8vTgwYP2/8OHD+1/ZFmKBWk9fPhw4969e+3+kydP2v+SkkVeUM5QguTgxo0brUKHNU4KMQ5lLVoDxevXr4/zOsSLFy+qlrwISiBWOyl5yMQ+6c0C5+BQjrO1VdeT45RF7YFhTNihBw5jjBmLFUZjAhqanBWsSVHRyS4ShxOHmCUsShkKBKAA5uHdGqtSLMgPyk0fWOCkAGONRBlSGaCo6RjWPbh79277L8hH33B0BqUVy2csY65jLjesin31oqSw1qAuoOi9e/euVYCxtpbKnuNwdHTU/teYJawxxsyLFUZjAn2KBopLjajolFymL75S+rOEBRQtFCENrWaLYmmodFlg3eSzMhmshjWwvGFBZM5gCayW5DMr1AytMy9zCOJHLlk+RR7GlTJHWTMsXppPWlIa87A2yJoYFWCUZ1mCMzXFOjNLWGOMmYumMzPmwsIbptwGvB0rGkWk9QMd579Eo8z1Hquht211bpSBf/aRA2phJWuUH5kEx3D7+9/fNGabf6H8xfOnRPFLdtKJ8pVQ/nVOhHxwTGWTaRSnwTeDiTfLEOMjfsVBuFg2xK+wkqWUnvIdUXiVf+18wgyVk+gLm2U3xphFsMJoLjTq2HPHKkUsdvAl6JRLis0YiFdpRIVFClNUAvrCSv7oMvFYzovSKiktUxFlROESypPyozKPYUSMo6+88e9TnHSe8ptdvP5S5HCxrAXy6XifLEC4fJx9nRvPz/45XZUNsg2FFQpvjDFTcImfpmEx5kLCMCHzyBqF4Yc3VsfAihoMm46ZM7eOMDzNsPHYFz+WAW/7Ljt9hpYZrp7nGs8Ldev58+fHcwxXiVafWWV+jTHnG89hNOaCouXjTktZRKHiJZCxbzTPCwopSvGqlSceIpjfuOpl+niZh+9JWlk0xkyJFUZjGvjEySpfBjltsEDx8gUv65wWKFQMcCzbOotCfFr5RGnjBZ3aetNTopd5TsOqaYw533hI2hhjjDHGVLGF0RhjjDHGVKlaGLe3t7stY4wxFwW3/caYjIekjTHGGGNMFQ9JG2OMMcaYKlYYjTHGGGNMFSuMxhizRPgu4io/q8N3J40xZmqsMJozgT7yjIvfS6RzlD8dc41aWPZ1rC8M8BFmjvV9s5FjUxFl6YMVTGK4krIgmXHKEwrMKhUL0iP9MdTCKr+la8M1yfnsoxY2lmmtjDg2VIbkhe8ixo9oqy7zn1G6OOSoUQrLNy2vXbu28o+FG2MuALz0Ysy6o/Vz45q/Wl8XdFxr82aGwmq94egiWve4tM5xJJ83L8SjNYL5H0pXayTndYWV7+i0vvCYeKcglu0QtbBxPe28JjbnxbpRCiOGwrKvukEZxbBCsuTyjhAHaUV0nWIaIl4LwuVzI0Nh8SvJbYwx8zJN72bMkpGSFztB9tVRogTF/UwtLPtDHT/h+xSQCOEWRcqI0iPPQ+krTMxHzhfHcVFRQfGo5X0qlKcxDIXlWCwLXc8Ix0t1YSgs5RjPK4VXuRKuVnaUbVYKQXHmY/gLjpXkF2PCkkYMZ4wxi+AhaXMmKQ3nwfv377utPxkK+/bt242XL1/2DgWytFvT+W/cv3+/Pb7s4T6GMEv0+dfQOtEqg0axOLEU36NHj9q8Dw1/rjNHR0ftf8zDlStXinVhKCxrMLP+s2D5RIh1iGHoofW3FX6WZQ+VFty6dauaxpiwXGvqtjHGTIEVRnMmOTw87LaGGQr76tWrbus7KIdSCun49/f32zlvzQNW2wmjXCzzJYaplDcpFSg4KBUlbt682f6fZcVCStmLFy/af+i75kNhWXP66tWr3d53qA8KM0ZZhM+fP/fOc61B/DyUjGEoLIrvwcFBt2eMMYthhdFceFASUAZxjx8/bv1QClEW6fhBViesjYAl6qyAgrO3t9duky8UjcxZVyzIX7QSYw1GuS8xS9gIL8dsbm6esO5NDddKDyYoqrWHhzFhp3r4MMYYK4zmTMIw4lhmCRuVK6xKUyhSeuu3z+W3dJehkGBZU77Oo2JB/qT040DKfaYWFqvgt2/f2m2BhZk69OzZs+NpCTiUbxTPeSyJQ7x796791xB6jVnCGmPMvFhhNGcSDS1mSpaiWcIC4bHY3Lhxo/1ECWSL4ixKHZ9UiQpKdvfu3etCfgcrVok+/7GoHJZpIVsHsKBy/XK5lshh8zCuFGnKLlqicdQfLNL4Z/Kw9jwg11hmCWuMMfNghdGcWeis9bKCrCtYgUDfHpT1rhZW3+TTiwr88zIIitXdu3dbP50r69ODBw/a/2Ug5UUvuaDASKlBNmQdY9VSGcR8wZMnT9r/SHzR4yxDuXz48KGoxGVKYVHu40tAKJQ7Ozvt9iz0vXQzFuot8vU97ET6wuYXeIwxZiGaJ2Vj1p69vR8/qwN81gR/XPzUytbW9+/5xfB9YfVJGjnOjfBpkngcWfrg+FTENIXKQX5ZNjn8yWP0a5TO9pyI8k74ZRHLHafyk3wcF31hIeYdF8+TX64foLqguGphIZZpTCNDvLXjlHeUH/I1kQy1vIHKBdmGwgqFN8aYKbjET9OwGLPWYB3jTd+mgz2xasa6gUXvLN1SWKbu3Lkz6s3fZYEVb9npYzHkTfBV1h3q7PPnz4/nGK4SvcW/zveKMeZs4SFpYy4o+nTQaSmLGl5f5vA+oJCiFK9aeWKImCHhZX+3M8MQNcPRVhaNMVNiC6M5E8jCCMznGzNH7TQ4KxZGLFDMjTxNy+JFAQWOua+rUOC4T968eePraoyZHCuMxhhjjDGmioekjTHGGGNMlaqFcXt7u9syxhhzUXDbb4zJeEjaGGOMMcZU8ZC0McYYY4ypYoXRGGOMMcZUscJojDFLgE/c8A3IqeDzPPogtzHGrBorjOZcwIolfAMxr50sv9rHk+mIFQ7HfqYUfwmOT4XSG4ozyo+cmSi7wqB4TKnMLArySL58rVilJcqvdZ6F1szGla5dCcUZ41Ic0UlBo6zysVJZC+RglRd9DzHWRVyNmB85/FhLnG9nrtN1M8ZcIHjpxZizDuv2xrWStd6u1ozmWG3dX9Caw3GdaZB/dH3Ujs0C8Uhe/mPeInldYVxcC7t0XOsX1+JdJZR3XNcZGeO1ijLiz3HBPnkE5XUMxElYrbVcWleaMBzH5TpB+NI5QPgoM/vxmrDdV+7kQfkR5DGmn/eNMWYVWGE05wI64NgJS7GQQiClT/slFCZ2xnTeUgw4l+O1eDi2KKQf5SD9LJdAZsmiPEcZ8KvlmTIjzGmS80V+dS3zMV0DKVU5bzHvfUjhIqzOz/EQv5S8UvkhX8kfOC/KnfNA3DHtSMkv54kwY/JpjDFT4iFpcy758OFDt3WSL1++dFvjuHz58vGSbjq3UTg2fvrpp3Z7GXz69KnbOknJ/927d8eyPHnypP1vlJn2H16+fNnuM6xZGsp89OhRGyYP864ShlojV69ePR7uzceUV64LxOvAsG2jnHV7ZRgq3tzc3Lhy5Urn8518PVle7+HDh+12PsbwMvJlf6Ac379/f0LunIeS7CL7IS/1LUIYrunY4XdjjJkCK4zmXLK/v99tLYY6cJSR+/fvt9vLZl7l7ejoqP1n7hxkhQLFEMUxzsG8efNm+//27dv2fx1AMX727Fm3dxJkzwo7+SRfKGo1KFfizgpcCR44bty40e2d5PPnzxu3b9/u9k7CQ0VU2EsQZmdnp9urg7wouJk7d+6061MbY8yqsMJozAiw5Ml6heK1jm+rYhVDmZJCxP8f36ednFCgpVBGeJliHZCy/PPPP7f/mfgiiVA+UcJQ6vssb1hY87klahZEePXq1cbdu3e7vdl5/fr1sdV6COpan4K7LtfMGHMxsMJoziVDVp55oOOW0jhLZx3fAC65rOD0KSo1ULRwfQoRcaJUbW1tbXz9+rXz/ZN5rZpTU1PqKMc+yyOghJG/kuWNMsYCqTK/detW618a2o3D0ZkhZXKIsUorIFcejjbGmNPCCqM5lzBkV6JvmHEsOv/atWvt/xhQZGTpK7lsQSoNQUKfP5BfrKCiz8qGItRXNqcNw/59ypTy02d5FH2KXLS24vb29lp/LK+5/GtWvZoyCXluZASFl2s4VtnEEvngwYNuzxhjThcrjOZcoo5WL6pgVdPcN5QPrEwoKENgTSKsLHCKb+yQ4jxIWdFLLlgzsYThj4ULefRSCLCN4iPrGQ7FpZRPlJCSUtY3J29VICPWQylT5FNKIv+URSzz0gs8XCOUvUWuDWnVrHo1ZRJ4oYVrkS22msIQzy3lQXA+dbZPQWaOZe0BwhhjJqd52jbmzNMoVCc+qwP6fAlOn0gBPnOCn8JrPzod0ydt5GI8JQgzFTFdEfMEyBnDxWOQ/Tk/ovw1Sk7ns3qyjHLQKG/FY1wziH75+utchY30fdqG65vLSBAPcQ5BHDFN9qOcckpHxyNcF1wJZM7hjTFm2Vzip2l8jDnTyOJWmp+3SrDonaVbinJjiHrsvLqzyCzzBqcA6yBluqy6SH6wLtYsncYYMzUekjbmgqKh6vOqLOplo1Xnj2F13uaO0wamojS0bYwxq8AKozk3aB4f899MHRQPFJvTtsguE71sdBqg0DFftDZPcVY0p/M8W4ONMeuLh6SNMcYYY0yVqsK4vb3dbRljjLkouO03xmRsYTTGGGOMMVU8h9EYY4wxxlSxwmiMMcYYY6pYYTTGmImY+lM6y/g0jzHGzIMVRmMq0GHzqZ74uR4tz4cbWl6Qz6oobOkTKzom1xcfx+Yhyh+VDz6rM+UnX5aFvqUY8x/9ZskD+denaSLEo+8biniNo6t9sonjLNknSEvnldKNcN0VNtYBPs2DnzHGnDZWGI0Z4Pr16+33/FjXF4Xh1q1bG7u7u60f3zHsU1pQQlh7WLAdFZOSEvHw4cNuaxqQb29vr9v7E623vM4WLBSlV69eteWsd/NUZvKjTGtKnOAa8Z3OiJTCEp8/fz5OA8e51IO+tZ1R8ijnuBb2/fv3j89nnew+OakTXCeFZVv1hPSId+jBxBhjlo0VRmNm4M2bN+3/jRs32n8ULpQWloPLRGUHZQM+fvzY/gPWIx2XW+UKHvoA9DpaGinXnZ2dHz4sTrlL2YWtra1uqx+UTJbSy6CMxWsTuXv3brf1nbdv3248evSo2zuJlNioTLLSC/IL9nElDg4O2qUEBdv4CcU7ZKU0xphlYoXRmBmIQ46RL1++dFt/EpUdKQuyQKFgvn///ngYMg+JrgqUoD6F97RAMUJhRLlW+Ug+lR9QZuz3Wf2A8z59+jSzIh7TAZT/rEQKFP9sGeba3rx5s9vb2Lhy5UrrVwJlNlpK2c6rudy+fbtNxxhjTgsrjMbMQB7WHMu3b9/a/ydPnrT/WKwiT58+PaEYrQopNVme00SKEcO4WACxIkYLHGVEWVFmQ2A9XXQpvZKyGkERRCEUCn/58uX2H7Rdur4os1gjmepAvshzhuuUra3GGLNKrDAaswKwUDHvUUqH1jnGxTmGL1686LZWSxwCPW1QwFAWZTlkG0VdFjjKUOWGNa5vSH0KZRFqw9FToiHsvvmK8z6sGGPMFFhhNGYGSvPdhmCIFYWjb1g0zqWrWRj1kkafW2QuYi3d06ZvyBl/lKyS7PihTFKmKh/gRZRZy6k2HF1CDwVHR0ftP2i7ZKVkaB2FXQ8RKMyLXEtjjFkGVhiNmYE4NBrRSzAZlDzm0OlFDfb7lDOUSuaq9SHFss9NYU1bBxiC5i3lTElxvHr1ard1ElkhowOsvLOUE9eL+ZQlRU+UHiLIw+HhYbe30W73vaDDXM1r1651e99l7Jsra4wxp4UVRmNm4MGDB+2/XnJhXtnjx49bhQJLIpYsDSmiGDIvDUuXrFzsExarEvvRkoTiEN8AXiU1RXXV8AIJ8xOlWFNGlHEJhqtxy4K34oc+dYRCmRXcLBeWzT45qQ9YMQUPGPnBhPjHvBFujDFLo3nyNsb0cP369dZF9vb2MFe1runEO98//tjd3W39CL+/v38cJjqFLx3Hrw+OzwOy5HSQU+zs7LR+tbRPA5UlLpYx2/LPcuu6xPARjsW852tQOg//IYizdG7MQ0xX/tEv5qsUF34xvDHGrJpL/DSNlDGmANYjOO03VLFGLuNWJX9Ys87LcDZgkVx1frAqY0Hsm2+5CAyL81mmd+/edT7GGLN6PCRtzAVFQ+fnRVlkCBvFuvSR7mWDMsd0A5S7KdG0BiuLxpjTxgqjMQPs7++3isjUysBpoo9en6dv+5EfrLCrXC0nQtpTL+2I9deDQMaYdcBD0sYYY4wxpootjMYYY4wxpkrVwri9vd1tGWOMuSi47TfGZDwkbYwxxhhjqnhI2hhjjDHGVLHCaIwxxhhjKmxs/H969+6YtcGBGwAAAABJRU5ErkJggg==\"\u003e\u003c/p\u003e\n \u003cdiv\u003e\n \u003cp\u003e\u003cstrong\u003eTable 2:\u003c/strong\u003e Adjusted Cox proportional hazards models and negative binomial regression model results on 1\u0026ndash;year mortality and hospitalizations and 30\u0026ndash;day mortality after an urgent hospitalization by e\u0026ndash;RHD\u0026ndash;FI, age and sex. The dispersion parameter for the negative binomial regression on hospitalizations resulted in 0.337. \u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/p\u003e\n \u003c/div\u003e\n \u003cp\u003eThe distribution of the e\u0026ndash;RHD\u0026ndash;FI is highly asymmetrical, as depicted in Fig. 1, with a median [quartiles] value of 0.0125 [0, 0.0375]. Almost half of the beneficiaries (3,846,242, 45.8%) had no deficit, with an e\u0026ndash;RHD\u0026ndash;FI equal to 0, while 3,150,964 (37.5%) had a e\u0026ndash;RHD\u0026ndash;FI greater than 0 but less than 0.056 (Table 1). A total of 1,119,904 (13.3%) presented an e\u0026ndash;RHD\u0026ndash;FI between 0.056 and 0.13, 2.7% had values between 0.13 and 0.25 and only 0.8% exceeded 0.25 with these pre\u0026ndash;specific cut\u0026ndash;offs defined in the methods section. The distribution of e\u0026ndash;RHD\u0026ndash;FI was similar in males and females; however, as expected, higher values were observed among individuals over 65 years of age. The prevalence of the 40 individual deficits within the cohort is reported in Supplementary Table 1, also differentiated for age class. The most common deficits were hypertension (9.2%), low income (8.3%), trauma (contusion with intact skin surface, 7.9%) and the use of transportation services, including ambulances (7%). Among older adults (\u0026ge;\u0026thinsp;65 years), in addition to those already mentioned above, heart diseases, diabetes and cancer were also common with a prevalence of approximately from 11 to 14%. The geographical distribution of the median value of e\u0026ndash;RHD\u0026ndash;FI is reported in Supplementary Fig. 1.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003ePredictive validity for mortality and hospitalization\u003c/h3\u003e\n\u003cp\u003eA total of 99,571 (1.2%) deaths and 926,623 hospitalizations affecting 710,419 individuals (8.5% of the population, overall hospitalization rate of 111.4 per 1,000 inhabitants) were observed in 2019. Mortality was different according to frailty classes, with rates of 0.3%, 3.2%, 11.8% and 24.8% in individuals with e\u0026ndash;RHD\u0026ndash;FI below 0.056, between 0.056 and 0.13, between 0.13 and 0.25 and above 0.25, respectively. Figure\u0026nbsp;2 reports the state occupancy probabilities over the year 2019, as estimated by the multi\u0026ndash;state model, which considered a common baseline state for all adults, first hospitalization of the year as a possible transition state, and mortality as the absorbing state, stratified by frailty class. The Figure clearly shows the increased burden of hospitalizations and mortality for increasing levels of e\u0026ndash;RHD\u0026ndash;FI. In particular, among individuals with an e\u0026ndash;RHD\u0026ndash;FI ranging from 0.13 to 0.25, 32.2% (71,916/223,477) had at least one hospitalization and 11.8% died (26,310/223,477) during 2019. In the very frail group (e\u0026ndash;RHD\u0026ndash;FI\u0026thinsp;\u0026ge;\u0026thinsp;0.25) these percentages raised to 39.6% (25,105/63,417) and 24.8% (15,701/63,417), respectively, resulting in more than half individuals in this category dying or experiencing at least one hospital admission during 2019.\u003c/p\u003e\n\u003cp\u003eMortality by e\u0026ndash;RHD\u0026ndash;FI, stratified by age and sex, is reported in Fig.\u0026nbsp;3 (as a continuous variable), and in Supplementary Fig.\u0026nbsp;2 (by e\u0026ndash;RHD\u0026ndash;FI classes). In both plots, mortality increased progressively with increasing e\u0026ndash;RHD\u0026ndash;FI values. Among individuals under 65 years of age, we observed a steeper slope after 0.2 points of e\u0026ndash;RHD\u0026ndash;FI both in males and females, likely due to the small number (3,824 over 6,115,611) of younger individuals in the very frail class (Fig.\u0026nbsp;3, left panel). In older adults (Fig.\u0026nbsp;3, right panel) the relationship between e\u0026ndash;RHD\u0026ndash;FI and mortality was approximately linear, with males showing higher mortality than females when the e\u0026ndash;RHD\u0026ndash;FI exceeded 0.1.\u003c/p\u003e\n\u003cp\u003eThree\u0026ndash;year survival estimates by e\u0026ndash;RHD\u0026ndash;FI classes were reported in Fig.\u0026nbsp;4. Survival at 3 years was 98.6 (99% CI: 98.6\u0026ndash;98.6), 88.3 (99% CI: 88.3\u0026ndash;88.4), 64.6 (99% CI: 64.4\u0026ndash;64.9), and 41.3 (99% CI: 40.8\u0026ndash;41.9) for beneficiaries classified as fit, pre\u0026ndash;frail, frail and very frail, respectively, according to their e\u0026ndash;RHD\u0026ndash;FI status as of January 1st 2019. It is worth noting that a bump in survival is evident in 2020, corresponding to the first wave of SARS\u0026ndash;CoV\u0026ndash;2 pandemics.\u003c/p\u003e\n\u003cp\u003eWhen accounting for sex and age in the multivariable Cox model, each 0.1\u0026ndash;point increment of e\u0026ndash;RHD\u0026ndash;FI was associated with a doubled 1\u0026ndash;year mortality hazard (HR: 2.12, 99% CI: 2.11\u0026ndash;2.14, Table\u0026nbsp;2). The hazard ratio of mortality, when stratifying by sex and modeling age and e\u0026ndash;RHD\u0026ndash;FI flexibly, is depicted in Supplementary Fig.\u0026nbsp;3.\u003c/p\u003e\n\u003cp\u003eThe rate of hospitalizations increased with e\u0026ndash;RHD\u0026ndash;FI, with an incidence rate ratio (IRR) of 2.51 (99%CI: 2.49\u0026ndash;2.53) each 0.1\u0026ndash;point increase, adjusting for age and sex in the negative binomial regression model (Table\u0026nbsp;2). The dispersion parameter for the model was estimated 0.337.\u003c/p\u003e\n\u003cp\u003eIn order to assess the predictive ability of e\u0026ndash;RHD\u0026ndash;FI after an acute insult (resilience) we modeled 30\u0026ndash;day mortality in a subpopulation who underwent a hospitalization classified as urgent (n\u0026thinsp;=\u0026thinsp;274,629; events\u0026thinsp;=\u0026thinsp;20,114) and still found an impact of e\u0026ndash;RHD\u0026ndash;FI (HR for 0.1\u0026ndash;point increment, 1.33, 99% CI: 1.30\u0026ndash;1.35, Table\u0026nbsp;2).\u003c/p\u003e\n\u003ch3\u003ePredictive performance across subgroups and validation of cut\u0026ndash;off points\u003c/h3\u003e\n\u003cp\u003eThe AUC of the ROC curve for 1\u0026ndash;year mortality (Supplementary Fig.\u0026nbsp;4, panel a) resulted in 0.883 (99% CI: 0.881\u0026ndash;0.884) for e\u0026ndash;RHD\u0026ndash;FI. The Brier score resulted in 0.011. The cut\u0026ndash;off value of 0.056 showed relatively good performance with sensitivity of 78%, specificity of 84% and negative predictive value of 99.7%. The positive predictive value was quite low (5.5%) due to the low incidence of mortality, but was nevertheless higher than the overall mortality rate (1.2%). With increasing cut\u0026ndash;offs, as expected, the sensitivity decreased while specificity and positive predictive value increased (Supplementary Fig.\u0026nbsp;4, panel a).\u003c/p\u003e\n\u003cp\u003eFocusing on true positives (in our case frail individuals who died within one year) it might be useful to balance the positive predictive value and sensitivity (Supplementary Fig.\u0026nbsp;4, panel b). Sensitivity was higher at lower thresholds, although precision was low and increased with higher thresholds. The F1 score, i.e. harmonic mean between the positive predictive value and sensitivity, resulted higher (0.22) for the 0.13 cut\u0026ndash;off, with a sensitivity of 42% and positive predictive value of 15%. ROC curves for 1\u0026ndash;year mortality by sex and age class were reported in Supplementary Fig.\u0026nbsp;5, with similar values of AUC by sex (panel a). When restricting only to older individuals, AUC resulted in 0.775 (panel b).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this population\u0026ndash;based study, we demonstrated the predictive validity of the Electronic\u0026ndash;Regional Healthcare Database Frailty index (e\u0026ndash;RHD\u0026ndash;FI) using electronic regional health databases (e\u0026ndash;RHD) in the adult population of a region of Italy.\u003c/p\u003e \u003cp\u003eWe successfully applied the e\u0026ndash;RHD\u0026ndash;FI to over 8.4\u0026nbsp;million adults in the Lombardy region, demonstrating its feasibility in a general population setting. The distribution analysis revealed that nearly half of the population were fit (e\u0026ndash;RHD\u0026ndash;FI\u0026thinsp;=\u0026thinsp;0), while only a small percentage (3.5%) could be classified as frail or very frail according to established cut\u0026ndash;offs. As expected, we observed higher FI values among older adults, confirming the age\u0026ndash;related nature of frailty accumulation. However, the index detected varying degrees of deficit accumulation across all adult age groups.\u003c/p\u003e \u003cp\u003eThe e\u0026ndash;RHD\u0026ndash;FI demonstrated robust predictive validity for both mortality and hospitalization outcomes. Each 0.1\u0026ndash;point increment in the index was associated with a doubled risk of 1\u0026ndash;year mortality (HR: 2.12) and a 2.5\u0026ndash;fold increase in hospitalization rate (IRR: 2.51), even after adjusting for age and sex. The multi\u0026ndash;state model analysis clearly illustrated how increasing levels of frailty associated with progressively higher probabilities of hospitalization and death, with more than half of very frail individuals experiencing at least one of these outcomes within a year.\u003c/p\u003e \u003cp\u003eOur analysis validated the previously established cut\u0026ndash;off points for frailty classification, with the ROC curve analysis showing good discrimination ability for 1\u0026ndash;year mortality (AUC: 0.883). The cut\u0026ndash;off of 0.056 demonstrated good sensitivity (78%) and specificity (84%), while the F1 score suggested that 0.13 might represent a good threshold when balancing sensitivity and positive predictive value. The index maintained strong predictive performance across sex subgroups, though with some reduction in discrimination power when restricted to older adults (AUC: 0.775 and Brier Score: 0.034), likely due to the higher baseline risk in this group.\u003c/p\u003e \u003cp\u003eImportantly, the e\u0026ndash;RHD\u0026ndash;FI also demonstrated predictive ability for mortality following acute insults (urgent hospitalizations), supporting its utility in assessing resilience\u0026mdash;the capacity to recover from stressors. This finding is consistent with the theoretical understanding of frailty as a state of increased vulnerability and reduced physiological reserve.\u003c/p\u003e \u003cp\u003eOur results align with international experiences where electronic health data\u0026ndash;derived frailty indices are increasingly used for public health purposes. For instance, the electronic Frailty Index (eFI) in the UK, developed by Clegg and colleagues \u003csup\u003e12\u003c/sup\u003e, is integrated into NHS guidelines to identify frail older adults in primary care and guide geriatric assessments. Similarly, various Claims\u0026ndash;Based Frailty Indices (CFIs) in the US, such as the VA\u0026ndash;FI and the Kim\u0026ndash;CFI, serve as validated tools for research and population health management \u003csup\u003e18,19\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe administrative data approach offers significant advantages in scalability and efficiency, enabling continuous, low\u0026ndash;cost monitoring of the entire assisted population. As highlighted by Silan et al. in the Italian context, such indicators address the need to stratify the population based on health needs for better intervention planning and resource allocation \u003csup\u003e20\u003c/sup\u003e. Our study confirms the e\u0026ndash;RHD\u0026ndash;FI's potential for this purpose in Lombardy. While different CFIs might identify slightly different target populations, their similar predictive performance suggests that using a validated index like the e\u0026ndash;RHD\u0026ndash;FI provides a robust tool for population\u0026ndash;level risk prediction and analysis\u003csup\u003e21\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eA notable aspect of our study is the e\u0026ndash;RHD\u0026ndash;FI's ability to identify and stratify risk across the entire adult population, including middle\u0026ndash;aged individuals (\u0026lt;\u0026thinsp;65 years). While frailty prevalence is significantly higher in older adults, our index also detected deficit accumulation in younger age groups. In fact, mortality risk progressively increased with higher e\u0026ndash;RHD\u0026ndash;FI values even among those\u0026thinsp;\u0026lt;\u0026thinsp;65, suggesting frailty's negative impact begins well before the conventional threshold of 65 years.\u003c/p\u003e \u003cp\u003eThis observation is consistent with studies emphasizing a life\u0026ndash;course perspective on frailty. In a large UK Biobank study of 493,737 participants aged 37\u0026ndash;73 years, 3% of participants were frail and 38% pre\u0026ndash;frail, with significant associations between frailty and mortality across all age strata, including middle\u0026ndash;aged adults \u003csup\u003e5\u003c/sup\u003e. Similarly, in a Korean prospective cohort from age 53, frailty transitions and burden significantly impact mortality in middle\u0026ndash;aged/younger\u0026ndash;old adults \u003csup\u003e22\u003c/sup\u003e. A Taiwanese study found that clinically detected frailty predicts unplanned readmissions in patients aged\u0026thinsp;\u0026ge;\u0026thinsp;50 (17% of the study sample\u0026thinsp;\u0026lt;\u0026thinsp;65) \u003csup\u003e23\u003c/sup\u003e. Frailty prevalence in younger and middle\u0026ndash;aged adults is increasingly recognized, though variable depending on definition and population \u003csup\u003e6\u003c/sup\u003e. These studies highlight that vulnerability associated with frailty is not exclusive to advanced age.\u003c/p\u003e \u003cp\u003eUsing administrative data derived instruments like the e\u0026ndash;RHD\u0026ndash;FI allows for early identification of individuals showing deficit accumulation before age 65. While identifying frailty might present greater challenges in younger populations using administrative data, our study indicates risk signals can be detected \u003csup\u003e24\u003c/sup\u003e. Early identification of frailty trajectories in middle age could have significant implications in prevention. Targeted interventions implemented at this stage could potentially modify the aging trajectory, slowing or reversing progression towards severe frailty and reducing future disability and healthcare burden. The e\u0026ndash;RHD\u0026ndash;FI, applied routinely, could act as an early warning system, flagging middle\u0026ndash;aged individuals at risk for further assessment and personalized preventive interventions.\u003c/p\u003e \u003cp\u003eBeyond public health applications, the validation of e\u0026ndash;RHD\u0026ndash;FI offers significant implications for pharmacoepidemiology. Studies using large administrative health databases often face \u0026lsquo;confounding by frailty\u0026rsquo;, especially when examining drug effects in older or multimorbid populations. Frailty, as a state of increased vulnerability, influences both treatment propensity and outcome risk, acting as a strong confounder. Inadequate adjustment for frailty can bias estimates of drug effectiveness and safety \u003csup\u003e25\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe e\u0026ndash;RHD\u0026ndash;FI, developed and validated using the same regional administrative data frequently used for pharmacoepidemiologic studies in Lombardy, represents a particularly suitable covariate. Its strong association with mortality and hospitalization confirms its validity as a summary measure of health status and vulnerability. Incorporating the e\u0026ndash;RHD\u0026ndash;FI (preferably as a continuous variable or using meaningful thresholds) into pharmacoepidemiologic models can enable more effective control of confounding by frailty, leading to more accurate treatment effect estimates. A preliminary study on the cardiovascular safety of targeted therapies in rheumatoid arthritis in Lombardy already included e\u0026ndash;RHD\u0026ndash;FI as a covariate and possible effect modifier\u003csup\u003e26\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe use of electronic health databases for estimating frailty has several limitations. First, these databases primarily focus on administrative issues and lack comprehensive clinical information needed to directly assess all frailty domains. While we included deficits covering various domains, including proxies for socioeconomic status and functional impairments often missing in electronic FIs, capturing aspects like physical performance, psychological well\u0026ndash;being, and social factors remains challenging. This was evident in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, where at the same level of frailty, males over 65 years of age showed higher mortality compared to females, likely due to socio\u0026ndash;demographic, occupational, and lifestyle factors associated with male sex that are not captured by the index. Second, information might be inaccurate or missing due to coding errors, inconsistent reporting, or use of private healthcare, although Italy's universal health system mitigates the latter concern. Our previous validation found administrative data deficits are sensitive but can lack specificity, potentially leading to misclassification \u003csup\u003e14\u003c/sup\u003e. Third, the index might overlook frail individuals who rarely seek medical attention. Finally, we measured frailty at an index date and did not consider changes over time in this specific analysis, although the data source allows for longitudinal assessment.\u003c/p\u003e \u003cp\u003eDespite these limitations, this population\u0026ndash;based study demonstrates that the e\u0026ndash;RHD\u0026ndash;FI, derived from regional electronic health databases, is a feasible and valid tool for assessing frailty in the general adult population of Lombardy. It shows strong predictive validity for mortality and hospitalization, performs well across different subgroups, and corresponds with established frailty cut\u0026ndash;offs. The e\u0026ndash;RHD\u0026ndash;FI offers valuable insights on a large scale and can identify subjects for more comprehensive clinical assessments or targeted public health interventions. Furthermore, it represents a valuable resource for improving the validity of pharmacoepidemiologic research by enabling better control for confounding by frailty. Its application supports evidence\u0026ndash;based health planning and preventive strategies aimed at mitigating the impact of frailty on individuals and the healthcare system.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eData source and population\u003c/h2\u003e \u003cp\u003eAdult beneficiaries of the Lombardy Regional Health System are our target population. We retrieved adult beneficiaries on January 1st, 2019 by the Population Register (Demographics NAR) and we followed them up to death, emigration/end of assistance in the region or last available follow\u0026ndash;up time (December 31st, 2021), whichever occurs first.\u003c/p\u003e \u003cp\u003eThe study was approved by the Ethical Committee of the University of Milano-Bicocca (approval number 920, protocol n. 0065413 dated 03/02/2025 \u0026ndash; UOR:003406), which also waived the requirement to obtain informed consent. All methods were performed in accordance with relevant guidelines and regulations.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eElectronic–Regional Healthcare Database Frailty index (e–RHD–FI)\u003c/h3\u003e\n\u003cp\u003eThe electronic\u0026ndash;regional healthcare database frailty index (e\u0026ndash;RHD\u0026ndash;FI) previously developed \u003csup\u003e14\u003c/sup\u003e was applied considering the dates of January 1st, 2019 as the index date.\u003c/p\u003e \u003cp\u003eThe e\u0026ndash;RHD\u0026ndash;FI was developed, according to Rockwood\u0026rsquo;s approach, with 40\u0026ndash;items/deficits and considering a recall period from 1 to 10 years, according to different types of health deficits and determinants. The e\u0026ndash;RHD\u0026ndash;FI was calculated as the ratio between the sum of health deficits and determinants and the total number of health deficits and determinants included in the index (theoretical range 0\u0026ndash;1). For each patient, all deficits contributed to the index with 1 point, except income, which was classified into three categories: upper\u0026ndash;middle/high income (no contribution to the index), lower\u0026ndash;middle income (0.5 contribution), and low income (1 point contribution). Health deficits and determinants were retrieved by cross\u0026ndash;referencing the different data sources in the e\u0026ndash;RHD. Health deficits were defined as symptoms, signs, disabilities, and diseases, covering a range of systems and domains (i.e., physical, mental and social). The co\u0026ndash;payment exemptions database, Integrated Home Assistance, and Intermediate Assistance Observation Sheet datasets were evaluated to obtain a proxy for socio\u0026ndash;economic status and to identify nursing home residents. Physical disabilities were retrieved from exemptions, Integrated Home Assistance, and prosthesis delivery service. Mental illness data were used to assess the mental domain \u003csup\u003e20,27\u003c/sup\u003e. Full details on data source and algorithm development are reported elsewhere \u003csup\u003e14\u003c/sup\u003e.\u003c/p\u003e\n\u003ch3\u003eEnd–points\u003c/h3\u003e\n\u003cp\u003eThe primary end\u0026ndash;point is 1\u0026ndash;year mortality. We will also consider the number of hospitalizations occurring in 1 year, a combined end\u0026ndash;point of death and hospitalizations and a longer\u0026ndash;term mortality end\u0026ndash;point (3 years).\u003c/p\u003e \u003cp\u003eMortality was retrieved by the Population Register, while for hospitalizations we used the Hospital Discharge Form. We considered all hospitalizations except for those due to delivery (DRG codes between 370 and 391).\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eStatistical methods\u003c/h2\u003e \u003cp\u003eThe e\u0026ndash;RHD\u0026ndash;FI was computed on adult beneficiaries of the Lombardy Regional Health System on January 1st, 2019 by using electronic\u0026ndash;regional healthcare databases from 2008 up to 2018. Variables included in the e\u0026ndash;RHD\u0026ndash;FI, along with their prevalence, were reported in the overall population as well as for adults younger than 65 years and those aged 65 years or older. The resulting e\u0026ndash;RHD\u0026ndash;FI was described by median and quartiles and represented by histogram. Previously identified/used cut\u0026ndash;offs were used to classify adults as fit (\u0026lt;\u0026thinsp;0.056 \u003csup\u003e14\u003c/sup\u003e), pre\u0026ndash;frail (0.056\u0026ndash;0.13), frail (0.13\u0026ndash;0.25) and very frail (\u0026ge;\u0026thinsp;0.25). The median value of e\u0026ndash;RHD\u0026ndash;FI was computed also across different territories of Health Protection Agencies in Lombardy and reported in a map to depict the geographical distribution. Proportions of 1\u0026ndash;year mortality and first hospitalization were graphically represented by e\u0026ndash;RHD\u0026ndash;FI (\u0026lt;\u0026thinsp;0.056, 0.056\u0026ndash;0.13, 0.13\u0026ndash;0.25 or \u0026ge;\u0026thinsp;0.25), age (\u0026lt;\u0026thinsp;65 or \u0026ge;\u0026thinsp;65 years) and sex classes.\u003c/p\u003e \u003cp\u003eTo describe hospitalizations and mortality at different levels of frailty we used a multi\u0026ndash;state (illness\u0026ndash;death model) model with three states: 1) a baseline state event\u0026ndash;free common to all adults, 2) first hospitalization and 3) mortality as an absorbing state. We accounted for age and sex and stratified the model on frailty class.\u003c/p\u003e \u003cp\u003eThe relationship between e\u0026ndash;RHD\u0026ndash;FI and 1\u0026ndash;year mortality was modeled by a logistic regression model with a restricted cubic spline with three knots placed on 10th, 50th and 90th percentiles, respectively, and stratifying by age (\u0026lt;\u0026thinsp;65 vs. \u0026ge; 65 years) and sex. Survival up to three years was estimated using the Kaplan\u0026ndash;Meier method and compared across frailty classes using the log\u0026ndash;rank test. The association between the e\u0026ndash;RHD\u0026ndash;FI and 1\u0026ndash;year mortality was evaluated by Cox proportional hazards regression model adjusting by continuous age and stratifying by sex. Restricted cubic splines with three knots have been used to relax the linearity in the relationship between the e\u0026ndash;RHD\u0026ndash;FI and the logarithm of the mortality hazard. The Brier score was computed in the overall population, as well as for adults younger than 65 years and those aged 65 years or older.\u003c/p\u003e \u003cp\u003eThe number of hospitalizations occurring in 2019 was also derived, and rates of hospitalizations were computed at different levels of e\u0026ndash;RHD\u0026ndash;FI and compared by a negative binomial regression model including sex and continuous age with the length of follow\u0026ndash;up used as offset. Furthermore, we restricted the population to those who underwent an urgent hospitalization during 2019, and evaluated the association between e\u0026ndash;RHD\u0026ndash;FI and 30\u0026ndash;day mortality after that hospital admission by a Cox proportional hazards regression model adjusting by sex and age.\u003c/p\u003e \u003cp\u003eThe performance of the frailty index in predicting 1\u0026ndash;year mortality was assessed by the area under the Receiver\u0026ndash;Operating\u0026ndash;Characteristic (ROC) curve (AUC) and the Precision\u0026ndash;Recall curve. The performance of the three cut\u0026ndash;offs previously used were reported by specific values of sensibility \u003cem\u003e(also called recall)\u003c/em\u003e, specificity, Positive Predictive Value (PPV, \u003cem\u003ealso called precision\u003c/em\u003e), and Negative Predictive Value (NPV). For each cut\u0026ndash;off the harmonic mean between precision and recall was reported (F1 score). Furthermore, AUC values were also computed stratifying the population according to sex and age (\u0026lt;\u0026thinsp;65 vs. \u0026ge; 65 years).\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eP.R. and C.A.S. conceived the study. P.R. and G.O. curated and analyzed the data. P.R., G.O., and C.A.S. developed the methodology and software. P.R. acquired funding and administered the project. P.R., G.B., M.G.V., S.G., and C.A.S. supervised the work and validated the results. P.R., G.O., and C.A.S. wrote the original draft. All authors reviewed and edited the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data that support the findings of this study are available from Regione Lombardia but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the corresponding author upon reasonable request and with permission of Regione Lombardia.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding Declaration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eP.R. and G.O. were partially funded by the European Union \u0026ndash; Next Generation EU \u0026ndash; NRRP M6C2 \u0026ndash; Investment 2.1 Enhancement and strengthening of biomedical research in the NHS project code PNRR\u0026ndash;MAD\u0026ndash;2022\u0026ndash;12376033, title \u0026ldquo;Evidence\u0026ndash;based models for high impact chronic disease prevention and risk of progression management in outpatient community services and community hospitals: towards eHealth integrating stratification on individual history with predictive models of disease progression, using machine learning and artificial intelligence on administrative and clinical databases\u0026rdquo;.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAdditional Information\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare they have no conflict of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eRutenberg, A. D., Mitnitski, A. B., Farrell, S. G. \u0026amp; Rockwood, K. Unifying aging and frailty through complex dynamical networks. \u003cem\u003eExperimental Gerontology\u003c/em\u003e \u003cstrong\u003e107\u003c/strong\u003e, 126\u0026ndash;129 (2018).\u003c/li\u003e\n\u003cli\u003eCOVID\u0026ndash;19 Monza Team members \u003cem\u003eet al.\u003c/em\u003e Frailty index predicts poor outcome in COVID\u0026ndash;19 patients. \u003cem\u003eIntensive Care Med\u003c/em\u003e \u003cstrong\u003e46\u003c/strong\u003e, 1634\u0026ndash;1636 (2020).\u003c/li\u003e\n\u003cli\u003eOfori\u0026ndash;Asenso, R. \u003cem\u003eet al.\u003c/em\u003e Global Incidence of Frailty and Prefrailty Among Community\u0026ndash;Dwelling Older Adults: A Systematic Review and Meta\u0026ndash;analysis. \u003cem\u003eJAMA Netw Open\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, e198398 (2019).\u003c/li\u003e\n\u003cli\u003eGalimberti, S. \u003cem\u003eet al.\u003c/em\u003e Effect of frailty on 6\u0026ndash;month outcome after traumatic brain injury: a multicentre cohort study with external validation. \u003cem\u003eThe Lancet Neurology\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 153\u0026ndash;162 (2022).\u003c/li\u003e\n\u003cli\u003eHanlon, P. \u003cem\u003eet al.\u003c/em\u003e Frailty and pre\u0026ndash;frailty in middle\u0026ndash;aged and older adults and its association with multimorbidity and mortality: a prospective analysis of 493 737 UK Biobank participants. \u003cem\u003eThe Lancet Public Health\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, e323\u0026ndash;e332 (2018).\u003c/li\u003e\n\u003cli\u003eLoecker, C., Schmaderer, M. \u0026amp; Zimmerman, L. Frailty in Young and Middle\u0026ndash;Aged Adults: An Integrative Review. \u003cem\u003eJ Frailty Aging\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 327\u0026ndash;333 (2021).\u003c/li\u003e\n\u003cli\u003eSociet\u0026agrave; Italiana di Medicina Generale (SIMG). \u003cem\u003eLinea Guida Inter\u0026ndash;Societaria per La Gestione Della Multimorbilit\u0026agrave; e Polifarmacoterapia\u003c/em\u003e. https://www.simg.it/linea\u0026ndash;guida\u0026ndash;inter\u0026ndash;societaria\u0026ndash;per\u0026ndash;la\u0026ndash;gestione\u0026ndash;della\u0026ndash;multimorbilita\u0026ndash;e\u0026ndash;polifarmacoterapia\u0026ndash;2021/ (2021).\u003c/li\u003e\n\u003cli\u003eFarmer, C., Fenu, E., O\u0026rsquo;Flynn, N. \u0026amp; Guthrie, B. Clinical assessment and management of multimorbidity: summary of NICE guidance. \u003cem\u003eBMJ\u003c/em\u003e i4843 (2016) doi:10.1136/bmj.i4843.\u003c/li\u003e\n\u003cli\u003eRockwood, K. \u0026amp; Mitnitski, A. Frailty in relation to the accumulation of deficits. \u003cem\u003eJ Gerontol A Biol Sci Med Sci\u003c/em\u003e \u003cstrong\u003e62\u003c/strong\u003e, 722\u0026ndash;727 (2007).\u003c/li\u003e\n\u003cli\u003eLiu, X. \u003cem\u003eet al.\u003c/em\u003e Association between frailty index and mortality in depressed patients: results from NHANES 2005\u0026ndash;2018. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 3305 (2025).\u003c/li\u003e\n\u003cli\u003eO\u0026rsquo;Donovan, M. R. \u003cem\u003eet al.\u003c/em\u003e Comparing frailty prevalence between countries: validation of the Global Burden of Disease study Frailty Index (GBD\u0026ndash;FI) in the survey of health, ageing and retirement in Europe. \u003cem\u003eAge and Ageing\u003c/em\u003e \u003cstrong\u003e52\u003c/strong\u003e, afad214 (2023).\u003c/li\u003e\n\u003cli\u003eClegg, A. \u003cem\u003eet al.\u003c/em\u003e Development and validation of an electronic frailty index using routine primary care electronic health record data. \u003cem\u003eAge Ageing\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 353\u0026ndash;360 (2016).\u003c/li\u003e\n\u003cli\u003eKim, D. H. Measuring Frailty in Health Care Databases for Clinical Care and Research. \u003cem\u003eAnn Geriatr Med Res\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, 62\u0026ndash;74 (2020).\u003c/li\u003e\n\u003cli\u003eRebora, P. \u003cem\u003eet al.\u003c/em\u003e Development and validation of an electronic database\u0026ndash;based frailty index to predict mortality and hospitalization in a population\u0026ndash;based study of adults with SARS\u0026ndash;CoV\u0026ndash;2. \u003cem\u003eFront Med (Lausanne)\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 1134377 (2023).\u003c/li\u003e\n\u003cli\u003eSearle, S. D., Mitnitski, A., Gahbauer, E. A., Gill, T. M. \u0026amp; Rockwood, K. A standard procedure for creating a frailty index. \u003cem\u003eBMC Geriatrics\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 24 (2008).\u003c/li\u003e\n\u003cli\u003eRockwood, K. A global clinical measure of fitness and frailty in elderly people. \u003cem\u003eCanadian Medical Association Journal\u003c/em\u003e \u003cstrong\u003e173\u003c/strong\u003e, 489\u0026ndash;495 (2005).\u003c/li\u003e\n\u003cli\u003eon behalf of the FRACoViD Team \u003cem\u003eet al.\u003c/em\u003e The effect of frailty on in\u0026ndash;hospital and medium\u0026ndash;term mortality of patients with COronaVIrus Disease\u0026ndash;19: the FRACOVID study. \u003cem\u003ePanminerva Med\u003c/em\u003e \u003cstrong\u003e64\u003c/strong\u003e, (2022).\u003c/li\u003e\n\u003cli\u003eOrkaby, A. R. \u003cem\u003eet al.\u003c/em\u003e The Burden of Frailty Among U.S. Veterans and Its Association With Mortality, 2002\u0026ndash;2012. \u003cem\u003eJ Gerontol A Biol Sci Med Sci\u003c/em\u003e \u003cstrong\u003e74\u003c/strong\u003e, 1257\u0026ndash;1264 (2019).\u003c/li\u003e\n\u003cli\u003eKim, D. H. \u003cem\u003eet al.\u003c/em\u003e Measuring Frailty in Medicare Data: Development and Validation of a Claims\u0026ndash;Based Frailty Index. \u003cem\u003eJ Gerontol A Biol Sci Med Sci\u003c/em\u003e \u003cstrong\u003e73\u003c/strong\u003e, 980\u0026ndash;987 (2018).\u003c/li\u003e\n\u003cli\u003eSilan, M., Caperna, G., Boccuzzo, G. \u0026amp; Link to external site, this link will open in a new window. Quantifying Frailty in Older People at an Italian Local Health Unit: A Proposal Based on Partially Ordered Sets. \u003cem\u003eSocial Indicators Research\u003c/em\u003e \u003cstrong\u003e146\u003c/strong\u003e, 757\u0026ndash;782 (2019).\u003c/li\u003e\n\u003cli\u003eSison, S. D. M. \u003cem\u003eet al.\u003c/em\u003e Claims\u0026ndash;Based Frailty Index and Its Relationship With Commonly Used Clinical Frailty Measures. \u003cem\u003eJ Gerontol A Biol Sci Med Sci\u003c/em\u003e \u003cstrong\u003e79\u003c/strong\u003e, glae094 (2024).\u003c/li\u003e\n\u003cli\u003eSon, M. K. \u0026amp; Lee, K. Frailty transition and burden on mortality risk in middle\u0026ndash;aged and older population: a prospective cohort study. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 26498 (2024).\u003c/li\u003e\n\u003cli\u003eKan, S.\u0026ndash;W., Yen, H.\u0026ndash;Y., Chi, M.\u0026ndash;J. \u0026amp; Huang, H.\u0026ndash;Y. Influence of physical function and frailty on unplanned readmission in middle\u0026ndash;aged and older patients discharged from a hospital: a follow\u0026ndash;up study. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 10003 (2025).\u003c/li\u003e\n\u003cli\u003eLahbib, H. \u003cem\u003eet al.\u003c/em\u003e Development and indirect validation of a model predicting frailty in the French healthcare claims database. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 11344 (2025).\u003c/li\u003e\n\u003cli\u003eKim, D. H. \u0026amp; Schneeweiss, S. Measuring Frailty Using Claims Data for Pharmacoepidemiologic Studies of Mortality in Older Adults: Evidence and Recommendations. \u003cem\u003ePharmacoepidemiol Drug Saf\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 891\u0026ndash;901 (2014).\u003c/li\u003e\n\u003cli\u003eSilvagni, E. \u003cem\u003eet al.\u003c/em\u003e Frailty and cardiovascular safety of JAK inhibitors versus TNF inhibitors in rheumatoid arthritis: a real\u0026ndash;world comparative study of drug effects and patient profiles. \u003cem\u003eFront. Pharmacol.\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 1565909 (2025).\u003c/li\u003e\n\u003cli\u003eBertini, F. \u003cem\u003eet al.\u003c/em\u003e Predicting Frailty Condition in Elderly Using Multidimensional Socioclinical Databases. \u003cem\u003eProceedings of the IEEE\u003c/em\u003e \u003cstrong\u003e106\u003c/strong\u003e, 723\u0026ndash;737 (2018).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Frailty Index, Electronic Health Records, Mortality, Hospitalizations, Prediction","lastPublishedDoi":"10.21203/rs.3.rs-6769624/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6769624/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eFrailty indices derived from electronic health records offer an efficient approach to identify vulnerable individuals in the general population. We aimed to validate the electronic\u0026ndash;Regional Healthcare Database Frailty Index (e\u0026ndash;RHD\u0026ndash;FI) in the general adult population, describe its distribution by sex and age, evaluate its predictive validity for mortality and hospitalization, and assess the performance of frailty cut\u0026ndash;off points.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eWe conducted a population\u0026ndash;based study of 8,404,004 adult beneficiaries of the Lombardy Regional Health System. The e\u0026ndash;RHD\u0026ndash;FI was calculated from 40 health deficits using electronic health records from 2008\u0026ndash;2018. We assessed its distribution, predictive validity for 1\u0026ndash;year mortality and hospitalizations through multi\u0026ndash;state analysis and multivariable models evaluating performance across subgroups.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe e\u0026ndash;RHD\u0026ndash;FI distribution was highly asymmetrical (median 0.0125, first\u0026ndash;third quartiles 0\u0026ndash;0.0375), with 45.8% of adults having no deficits. The index was higher in older adults. Each 0.1\u0026ndash;point increment in e\u0026ndash;RHD\u0026ndash;FI was associated with doubled 1\u0026ndash;year mortality risk (HR: 2.12, 99%CI: 2.11\u0026ndash;2.14) and 2.5\u0026ndash;fold increased hospitalization rate (IRR: 2.51, 99%CI: 2.49\u0026ndash;2.53). The AUC for 1\u0026ndash;year mortality was 0.883 (99%CI: 0.881\u0026ndash;0.884).\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThe e\u0026ndash;RHD\u0026ndash;FI demonstrates strong predictive validity for adverse outcomes in the general population and can effectively identify at\u0026ndash;risk individuals for targeted interventions.\u003c/p\u003e","manuscriptTitle":"Population-Based Validation of a Frailty Index Using Electronic Regional Healthcare Records for Public Health Use","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-01 06:58:46","doi":"10.21203/rs.3.rs-6769624/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-07-25T04:26:44+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-24T21:07:07+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"308885981419366872204286866174493590077","date":"2025-07-24T11:37:44+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-09T15:37:13+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"205631565388563841787778707363287229690","date":"2025-07-01T20:56:00+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-06-26T22:10:30+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-06-26T09:00:21+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-06-20T06:04:19+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-06-04T13:42:11+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-06-04T13:39:29+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"dd8e0690-6270-48ac-ac46-a840ff0842ad","owner":[],"postedDate":"July 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":50749271,"name":"Health sciences/Diseases"},{"id":50749272,"name":"Health sciences/Health care"},{"id":50749273,"name":"Health sciences/Medical research"},{"id":50749274,"name":"Health sciences/Risk factors"}],"tags":[],"updatedAt":"2025-09-29T15:58:51+00:00","versionOfRecord":{"articleIdentity":"rs-6769624","link":"https://doi.org/10.1038/s41598-025-18611-9","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-09-26 15:56:52","publishedOnDateReadable":"September 26th, 2025"},"versionCreatedAt":"2025-07-01 06:58:46","video":"","vorDoi":"10.1038/s41598-025-18611-9","vorDoiUrl":"https://doi.org/10.1038/s41598-025-18611-9","workflowStages":[]},"version":"v1","identity":"rs-6769624","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6769624","identity":"rs-6769624","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}