Riviste UNIMI
Not a member yet
21278 research outputs found
Sort by
Cross-Country External Validation of a Multisource Comorbidity Score
No full textBackground
The increasing impact of multimorbidity is escalating clinical and economic demands on healthcare systems, underscoring the necessity for effective tools to assess clinical complexity and enhance management strategies [1-3]. The Multisource Comorbidity Score (MCS) is a population-based index based on regional healthcare Utilization databases - hospitalizations and drug prescriptions - of beneficiaries with age equal or greater than 50 years. The MCS was developed in the framework of the Monitoring and Assessing Care Pathways working group of the Italian Ministry of Health, and validated in four Italian regions, showing good performance in predicting mortality, hospitalizations, and healthcare costs [4].
Objective
This study aimed to externally validate and adapt the MCS within the Catalan healthcare system, assessing its predictive capability outside the original Italian setting. This validation seeks to determine whether the MCS can serve as transferable tool in other healthcare systems with different data availability.
Methods
An observational longitudinal study was performed on subjects aged 50 years or older, residing in the health district of Barcelona-Esquerra (ES) continuously during 2014-2015 and followed between January 1st, 2016, and December 31st, 2019.
Data were obtained from the Catalan Health Surveillance System [5] which integrates demographic, clinical and healthcare utilization information from several healthcare databases. For this study, we used Catalan healthcare system beneficiary’s, primary care, hospitalization, and pharmacy dispensation databases.
First, the MCS with Italian weights (MCS-1) was applied in the Catalonia setting using the same data sources (hospitalization and pharmacy dispensation databases) as in the original Italian version. Second, new MCS weights (MCS-2) were estimated in predicting one-year mortality (primary outcome) in the Catalonia setting using the methodology and data sources as described in [4]. Finally, a third MCS version (MCS-3) was developed estimating specific weights based on hospitalization, pharmacy dispensation and primary care databases to predict the primary outcome.
Secondary outcomes considered were four-year mortality, one- and four-year hospitalizations (≥1) and one- and four-year hyperfrequency primary care utilization (≥10 visits). To assess the performance of the three MCS versions, generalized linear models (GLMs) with a binomial distribution were used for each outcome. ROC curves and Area Under the Curve (AUC) with 95% Confidence Intervals (95% CI) were estimated to assess the discrimination ability of the three MCS versions. De Long\u27s method was used to compare the AUCs [6].
Net Reclassification Improvement (NRI) [7] was also calculated to assess improvements in risk classification by comparing new MCS versions with the MCS-1. The predicted GLM values, for each outcomes, of the three MCS versions were estimated and a threshold of 0.5 was used to distinguish between high and low risk individuals. The NRI estimates the percentage of individuals who were correctly reclassified into a higher risk category (according to the threshold) if they experienced the outcome, or into a lower risk category if they did not, minus those who were incorrectly reclassified when comparing the two models.
Results
As of January 1st, 2016, a total of 440,790 individuals had resided in the health district of ES for at least two years. Among them, 198,753 (45%) were aged 50 or older and formed the study cohort. They were mostly women (57%), with a median age of 66 years (IQR: 57–76).
Table 1 shows the MCS versions performances according to one- and four-year outcomes. All MCS versions demonstrated good discriminatory performance in primary and secondary outcomes.
For one-year mortality, the MCS-1 achieved an AUC of 0.742 (95% CI: 0.734-0.750) similarly to MCS-2 (AUC=0.756, 95% CI: 0.744-0.768), while MCS-3 version showed a significant improved AUC respect to MCS-1 (AUC=0.771, 95% CI: 0.760-0.783, p<0.001). Both new MCS versions performed better in predicting four-year mortality compared to MCS-1 (p= 0.012; p<0.001, respectively).
On the contrary, the MCS-1 showed better performance in predicting all secondary outcomes except one-year hospitalizations with respect to MCS-3.
In addition, significant improvements in risk reclassification for both one-year and four-year mortality were observed with the MCS-2 and MCS-3 compared to the MCS-1. For one-year mortality, the NRI increased by 0.63% (95% CI: 0.14–1.17) in MCS-2 and by 2.17% (95% CI: 1.39–2.97) in MCS-3. Similarly, for four-year mortality, the NRI increased by 1.8% (95% CI: 1.33–2.25) and 2.9% (95% CI: 2.35–3.45), respectively.
An increment in risk reclassification was found in four-year hospitalizations and hyper-frequency when comparing MCS-3 to MCS-1; the reclassification worsened in all secondary outcomes comparing MCS-2 to MCS-1.
Conclusions
The study supports the external validity of the Multisource Comorbidity Score in other healthcare systems with different data availability, such as Catalonia. The local adaptation slightly improved the ability of the score in predicting mortality, however this advantage was not maintained in the secondary outcomes, highlighting the importance of contextual adaptation of such tools.
These findings provide a basis for expanding the use of the score and refining it in different health systems and population segments
Anni di Vita Persi Aggiustati per Disabilita’ per Malattie Amianto Correlate in Italia nel Periodo 2010-2020
No full textIntroduzione: Le stime degli anni di vita aggiustati per disabilità (Disability Adjusted Life Years: DALYs) rappresentano uno strumento fondamentale per quantificare il carico globale delle malattie e dei fattori di rischio, utili alla pianificazione delle politiche sanitarie [1]. L’amianto è riconosciuto come cancerogeno certo, responsabile di mesotelioma (M), tumori del polmone, dell’ovaio e della laringe, oltre che di asbestosi e ispessimenti pleurici. Nel 2019, a livello globale, si sono stimati 4.189.000 DALYs attribuibili a patologie derivanti dall’esposizione occupazionale all’amianto. In Italia, tra il 2010 e il 2020 si sono registrati 16.993 decessi per M, mentre dal 1993 al 2021 il Registro Nazionale dei Mesoteliomi (ReNaM) ha censito oltre 35.000 casi incidenti.
Obiettivo: Stimare i DALYs attribuibili ad asbestosi e mesotelioma, patologie con elevata frazione eziologica legata all’amianto (rispettivamente 80% e 100%), in Italia nel decennio 2010-2020.
Metodi: Sono stati presi in considerazione i decessi per M e asbestosi, i ricoveri ospedalieri per asbestosi e i casi incidenti di M. I DALYs sono stati calcolati come somma degli anni di vita persi per morte prematura (PYLLs) e degli anni vissuti con disabilità (YLDs). I decessi per M (ICD-10: C45) e asbestosi (ICD-10: J61), così come i ricoveri per asbestosi (ICD-9: CM 501), sono stati estratti dalle banche dati elaborate dall’ISS, utilizzando fonti ISTAT e Ministero della Salute. I casi incidenti di M sono stati stimati a partire dai dati ReNaM. Le informazioni sono state stratificate per genere, classe di età e anno di calendario. I PYLLs sono stati calcolati in valore assoluto e come tassi standardizzati (per 100.000 abitanti), a livello nazionale e regionale, applicando i valori di speranza di vita ISTAT, disaggregati per genere, età e anno. Gli YLDs sono stati stimati a livello nazionale considerando i ricoveri per asbestosi e i casi incidenti di M, utilizzando un peso di disabilità pari a 0,217 e una durata di 20-30 anni per l ’asbestosi, e un peso di 0,540 con un anno di durata per il M [2, 3].
Risultati: Nel periodo 2010-2020, in Italia si stimano 204.232 DALYs per patologie amianto-correlate tra gli uomini e 72.625 tra le donne, assumendo una durata di 30 anni per l’asbestosi. I PYLLs ammontano a 161.300 nei maschi (di cui il 96,7% per M) e 67.311 nelle femmine (99,1% per M). Le regioni settentrionali, in particolare Liguria, Piemonte e Friuli-Venezia Giulia, mostrano i tassi più elevati di PYLLs per M in entrambi i sessi. Gli YLDs sono pari a 7.075 per M e 35.857 per asbestosi negli uomini, e a 2.710 per M e 2.604 per asbestosi nelle donne.
Conclusioni: Per la prima volta in Italia viene proposta una stima del carico complessivo di mesotelioma e asbestosi in termini di DALYs. Questi dati, che includono gli anni vissuti con ridotta funzionalità, rappresentano un supporto rilevante per la programmazione sanitaria e l’attuazione di adeguate misure di welfare
Evaluation of Vaccination Status of Patients Diagnosis of Multiple Myeloma or Monoclonal Gammopathy of Uncertain Significance: Analysis of Coverage in the Province of Catania
No full textIntroduction : The patient diagnosed with multiple myeloma or monoclonal gammopathy of undetermined significance (MGUS) is considered a fragile patient, at risk of severe post-infectious outcomes [1] for which adequate vaccination prevention against pneumococcus and meningococcus, Herpes Zoster virus , influenza virus, SARS-CoV2 is recommended [2].
Despite existing recommendations, there is limited evidence regarding actual vaccination coverage of these patients [3].
Objectives: This study aims to evaluate the vaccination status in a cohort of haemato-oncology patients and to propose strategies to improve their vaccination coverage.
Methods : We conducted a retrospective cohort study on 737 patients with multiple myeloma or MGUS diagnosed in the province of Catania in the period 2000-2023, coded by the Integrated Cancer Registry. Vaccination status was retrieved by cross-referencing cases with data in the Provincial Vaccination Registry. Medical records and vaccination data collected during outpatient visits, hospital admissions for possible hospitalization for infectious disease were examined and the presence of infectious disease diagnoses among the causes of death was verified. The data will be presented in aggregate form. The variables considered were: tumor diagnosis, last follow-up date, date of death, date and components of vaccinations performed.
Relative risk with 95% confidence intervals on deaths with respect to the vaccination stratum analyzed by type of disease was calculated. Analyses were performed with Stata 17. For all statistical tests, the significance level was set at a p value <0.05.
Results : Preliminary results indicate poor adherence to recommended vaccinations. Frequency distribution of variables of vaccination status, post-diagnosis vaccination, cycle completeness and timing of vaccinations were calculated.
The observed coverage is 30.9% for influenza, 6% for Varicella Zoster, 39% for SARS-CoV2, 1.5% for meningococcus and 10% for pneumococcus of vaccinable subjects.
The observed all-cause mortality rate among never-vaccinated subjects is 45.4%; among subjects with at least one dose of vaccine, it is 18.4%. Delays in meeting recommended intervals and incomplete cycles have been observed.
Relative risk (RR) analysis shows that vaccination is significantly associated with a reduction in mortality in the overall cohort (including patients with MGUS and multiple myeloma). In particular, subjects who have received at least one of the recommended vaccinations have a reduction in the risk of death of approximately 60% compared to unvaccinated subjects.
In detail:
In patients with MGUS, vaccination is associated with an approximately 74% reduction in the risk of death.
In patients with multiple myeloma, the risk reduction is around 42.7%.
Focusing the analysis on individual vaccines:
Vaccination against COVID-19 is associated with an approximately 63% reduction in the risk of death.
The one against the flu shows a reduction of 59.2%.
For pneumococcal, meningococcal, and Herpes Zoster vaccines, no statistically significant association with mortality was observed, either in the overall cohort or in subgroups. However, potentially relevant trends emerge:
In patients with MGUS, meningococcal vaccine showed a RR of 1.65 (95% CI: 0.91–3.02; p = 0.10), while herpes zoster vaccine showed a RR of 0.65 (95% CI: 0.35–1.21; p = 0.18). The wide confidence intervals suggest a possible lack of statistical power to detect significant effects.
Conclusions : Protection provided by vaccination is present in both subgroups, but is more marked in MGUS patients (RR ≈ 0.26) than in those with myeloma (RR ≈ 0.57). The study provides an overview of the vaccination status in haemato-oncology patients, highlighting the need for personalized strategies: vaccination recommendations in the discharge letter, promotion of training of hospital staff and the attending physician. The organization of vaccination sessions in the hospital outpatient setting has been activated in conjunction with periodic or follow-up checks
Validation Protocol of the Italian Version of the Internet Related Measures (IRM) Questionnaire and Assessment of Internet Use Among High School Students in Italy: A Multicenter Study
No full textIntroduction
Internet use has become an integral part of adolescents’ daily lives, a trend that intensified during and after the COVID-19 pandemic [1]. Alongside the increase in online activity, adolescents are now more exposed to digital threats such as harmful content (e.g., self-harm, radicalization), sexual exploitation, and cyberviolence [2]. Scientific literature highlights that some groups of adolescents—particularly those with psychological, familial, or social vulnerabilities—are more likely to encounter these risks [3,4].In the Italian context, however, validated tools to assess Internet use and related risks are often outdated and lack cultural and technological relevance. This gap hinders effective screening and prevention efforts in schools. The present study addresses this need by validating an internationally adopted tool—the Internet Related Measures (IRM) questionnaire—for use with Italian adolescents, ensuring both scientific accuracy and contextual relevance.
Objectives
The study aims to:
Validate the Italian version of the Internet Related Measures (IRM) questionnaire through linguistic, cultural, and psychometric adaptation;
Explore and describe patterns of Internet use and perceived online risks among high school students in Lombardy.
Methods
This study will adopt a multicenter design and will involve students attending upper secondary schools (scuole secondarie di secondo grado) across the Lombardy region in Northern Italy. Participants will complete a structured questionnaire comprising sociodemographic items (age, gender, ethnicity, school attended) along with the Italian version of the IRM questionnaire.Descriptive statistics will summarize participant characteristics and Internet use patterns. Means and standard deviations will be used for normally distributed variables; medians and interquartile ranges for non-normally distributed data; and frequencies and percentages for categorical variables.Psychometric properties of the IRM will be evaluated through both exploratory and confirmatory factor analyses. Internal consistency will be assessed using Cronbach’s alpha and corrected item-total correlations. To explore associations between IRM responses and sociodemographic variables, generalized additive models for location, scale, and shape (GAMLSS) will be employed. All statistical analyses will be performed using R software (version 4.4.1), with significance set at p < 0.05.
Expected Results
The Italian version of the IRM is expected to show robust psychometric properties, including high internal consistency and structural validity in line with its original version. The study is also expected to reveal a relevant prevalence of problematic Internet use among high school students in Lombardy.Additionally, the analysis is anticipated to identify associations between certain sociodemographic characteristics—such as age and gender—and increased perceptions of digital risk. These may include exposure to harmful or inappropriate content, compulsive or excessive use of the Internet, and experiences of cybervictimization. The findings will offer valuable insights for designing evidence-based digital health promotion strategies in the educational setting.
Conclusions
This study will provide an updated and in-depth picture of Internet use and digital risk perception among Italian adolescents. By validating the Italian version of the Internet Related Measures (IRM) questionnaire, it will offer a reliable and culturally adapted tool for use in educational and public health contexts. The results are expected to support the development of targeted prevention programs, digital literacy initiatives, and evidence-based policies aimed at promoting safer and healthier online behaviors among youth.
 
A Retrospective Analysis of Severe Injury Patients in The Hub and Spoke Centres in The Province of Alessandria
No full textTrauma is one of the leading causes of death in Western countries and results in severe health, economic, and social consequences due to residual disabilities. To improve the management of major trauma cases, integrated trauma care systems (SIAT) have been implemented in Italy, inspired by North American models. The goal is to centralize severe cases in the most equipped facilities (trauma centers), thereby reducing avoidable mortality [1]. These systems are based on a “Hub and Spoke” network, which includes a Hub center with a Level II Emergency Department (DEA) and several Spoke centers (with Level I DEAs), such as peripheral hospital facilities. In the Province of Alessandria, in Piedmont region, the Hub center is the hospital of Azienda Ospedaliero-Universitaria of Alessandria (AOU AL), and there are four Spoke centers representing the hospitals of the Local Health Authority of Alessandria (ASL AL), located in the cities of Acqui Terme, Casale Monferrato, Novi Ligure, and Tortona. Despite the public health relevance, epidemiological data on trauma in the Italian territory are scarce and inconsistent in the literature.
Objective
This study aims to give an epidemiological description of hospitalized severe trauma cases in the Spoke and Hub centers of the Province of Alessandria during the period 2017–2021, also analyzing potential differences between the pre-pandemic and pandemic periods. Additionally, the study aims to assess in-hospital and 30-day mortality for this patient category in both Spoke and Hub centers.
Methods
Subjects admitted to the Hub and Spoke centers were identified based on hospitalizations (ICD9-CM codes: 800.xx–904.xx; 925.xx–939.xx; 950.xx–959.xx) with discharge dates between 01/01/2017 and 31/12/2021.Trauma severity was assessed using the Injury Severity Score (ISS), based on ICD9-CM codes for principal and secondary diagnosis [2,3,4]: only admissions with ISS >15 (severe trauma) were included. In cases of multiple hospitalizations for the same patient, only the first was considered. Mann-Whitney and Chi-square tests were used to evaluate associations between variables. In order to find determinants of in-hospital and 30-day mortality, Odds Ratios (ORs) with 95% Confidence Interval (95% CI), estimated from multivariate logistic regression models, were used; gender, age, ISS, length of stay, access via Emergency Department, being hospitalized in Intensive Care Unit, being hospitalized during Covid-19 pandemic period (after 01/03/2020) and being hospitalized in Hub/Spoke center were put as covariates. The probability of in-hospital death for each subject was also calculated using the Trauma Mortality Prediction Model (TMPM) [5,6,7]. The predictive capacity of the model was tested on the study sample using ROC (Receiver Operating Characteristic) curves and the corresponding area under the curve (AUC) [8].
Results
In this study, a total of 1,337 patients were included: 705 hospitalized in the Hub center and 632 in the Spoke centers. Most patients in the whole sample were male (59.6%) and the median age was 76 (IQR: 60-85) years. The median length of hospital stay was 8 (IQR: 4–14) days, significantly longer in the Hub center: 9 (IQR: 5-16) days in the Hub center, 7 (IQR: 4-12) days in the Spoke center (p<0.001). In-hospital mortality was 10.3% in ASL AL centers and 6.4% in AOU AL (p=0.01). The emergency department was the primary point of access (79.4%) for these hospitalizations. The median ISS was 16 (IQR: 16–22). There were no significant differences in ISS between the hospitalizations during Covid-19 pandemic period and the ones that occurred before (p=0.95), nor in the distribution of in-hospital (p=0.48) and 30-day (p=0.90) deaths. TMPM ROC curves showed an AUC of 0.74 for in-hospital mortality and 0.71 for 30-day mortality. The results of the multivariate logistic regressions indicate that increasing of age (OR=1.08; 95% CI: 1.07–1.10), being male (OR=1.50; 95% CI: 1.06–2.14), higher ISS (OR=1.03; 95% CI: 1.02–1.05), and admission to Intensive Care Unit (OR=3.17; 95% CI: 1.74–5.70) are associated with increased odds of 30-day mortality. Furthermore, increasing of age (OR=1.06; 95% CI: 1.04–1.09), shorter length of hospital stay (OR=0.96; 95% CI: 0.94–0.99), higher ISS (OR=1.02; 95% CI: 1.00–1.03), admission to Intensive Care Unit (OR=5.72; 95% CI: 2.93–11.10), and being hospitalized in Spoke centers (OR=1.74; 95% CI: 1.10–2.80) are associated with increased odds of in-hospital mortality. Being hospitalized during the Covid-19 pandemic period was not significantly associated with increased odds of either in-hospital mortality (OR=0.83; 95% CI: 0.52–1.29) or 30-day mortality (OR=0.96; 95% CI: 0.66–1.37).
Conclusions
The TMPMs showed good discriminatory power in predicting mortality, supporting their reliability in clinical and operational settings. The absence of statistically significant differences in outcomes between the pre-pandemic and pandemic periods suggests that the Covid-19 pandemic had a limited impact on the care of polytrauma patients. The analysis of in-hospital mortality highlights the need to optimize the management of severely injured patients within the "Hub and Spoke" network. In 2024, the urgent transfer protocol for patients from ASL AL centers to the AOU AL center was revised; a future perspective will therefore be the evaluation of in-hospital mortality in the coming years.
 
Road Traffic Accidents Epidemiology: Evaluations on a 15 Years Surveillance Program from an Emergency Care Unit (ECU) in Northern Italy
No full textIntroduction
Road accidents cause 20-50 millions injuries every year and 2.35 millions deaths, representing the eighth cause of death in the general population and the first in people between 5 and 29 years of age [1]. Human and health care costs of the phenomenon have a significant impact on the population and the institutions, with about 280 billions of euros spent in Europe each year [2]. Legal regulations, educational interventions and new safety technologies have the potential of lowering the number and severity of road accidents [3]
Aims
To evaluate the prevalence, characteristics and severity of road accidents and their changes through the 16 years of observation and to estimate the impact of recent lifestyle and legislation changes on road accident epidemiology
Methods
We retrospectively evaluated data from the DATIS/SINIACA registry considering all ECU admissions to one of the main hospitals in Genoa due to road accidents from January 1st, 2008 to December 31th, 2023. The observation period was divided into subgroups considering legal regulation changes for further studies. We used the data software STATA 14.2 SE® for data analysis. We described the events using absolute and percentage frequencies. Chi-squared test (χ2) had been used to evaluate differences between categorical variables, p-values <0.05 were considered as statistically significant.
Results
During the 16 years of the study (2008-2023) we documented a total of 33.708 road accidents, with an overall lowering trend and a dramatic deflection in 2020 along with the COVID pandemic. Our local data followed a similar trend of national registries but with a greater percentage decline during the years. The majority of the accidents involved motorbikes (16.779) followed by cars (6.379), public transportation services (2.467), pedestrians (1.620), bicycles (1542), microcars (685), and heavy vehicles (225). Nearly half of the victims (49,78%) were motorcyclists and the drivers were the most frequently involved (55,08%); in total 19.104 (56,67%) of injured people were males and 14.604 (44,33%) females, most commonly in the 20-29 years (7041 patients, 20,89%) and e 40-49 years (6550 patients,19,43%) age-range. Triage codes had been classified as “urgent” in 20,71% of the events; female victims were more likely to show lower priority injuries compared with males. Accidents with bicycles, microcars and pedestrians had a greater chance of receiving higher priority triage codes. Pandemic effects on people’s habits and both national and local policies about sustainable mobility caused interesting changes in road accidents epidemiology for what concerns the kind of involved vehicles and the severity of the events.
Conclusions
Road safety is a crucial issue in public health. Collecting and interpreting data on road accidents is an important tool for evaluating and implementing preventive strategies such as changes in legal regulations. In the next future we are planning to analyze data on drugs and alcohol-related events aiming to quantify the effect of the recently approved changes in traffic laws.
 
Access to Emergency Department in the Marche Region During 2011-2023: Exploring Differences between Italian and Migrant Residents (MIGHTY PROJECT P2022ASXKR)
No full textIntroduction
Scientific literature reports high geographical variability in migrants\u27 health status and access to healthcare services in Europe and inconsistent evidence on their access to Emergency Department (ED) respect to natives [1-3]. Nevertheless, migrants’ accesses to ED for non-urgent conditions, presenting as “walk-ins”, and during unsocial hours seems to be higher than natives [4-5].
Aim
We want to investigate whether people coming from High Migratory Pressure Country access to the ED differently than Italians and to explore differences according to their demographic characteristics and clinical presentation, in Marche Region between 2011 and 2023, using healthcare utilization databases.
Methods
In this cross-sectional study, we used the Emergency Department and the Regional Beneficiary databases of the Regional Healthcare System as data sources. In this study, residents in the Marche Region were distinguished by citizenship as Migrants, considering only those who came from High Migration Pressure Countries (HMPC) [6], and Italians. Sex, age, arrival mode (walk-in, ambulance, other emergency medical services), access time (daytime: 8:00-20:00; night-time: 20:00-8:00), and distribution of main access diagnosis (first three digits of ICD-9 CM code) were analyzed by citizenship.
ED admission rates per 1,000 person-years (py) were standardized by age and stratified by sex using the direct method and the Italian ISTAT population as of January 1, 2019, as the standard population [7]. The standardized rate ratios (SRRs) were calculated as the ratio of HMPC to Italian rates and estimated with a 95% confidence interval (95%CI). SRRs were also stratified by three age groups (0-19, 20-65, over 65 years) and by triage categories based on priority: emergent (red), urgent (yellow, orange, blu), less urgent (green), non-urgent (white), according to the new Regional Guidelines, DGR n. 1457/2019. All data were processed in compliance with the European (GDPR, EU 2016/679) and national privacy laws (D.lgs. 196/2003 and subsequent amendments).
Results
In the period 2011-2023, there were 5,189,603 visits to the ED for Italians and 442,514 for HMPCs, corresponding to 1,253,820 and 114,503 subjects, respectively. In the entire period, HMPC subjects accessing the ED were younger (mean age, SD: 33, 18 years) and more frequently female (54%) than Italians (mean age, SD: 50, 27 years; female: 49%). Both HMPCs and Italians presented to the ED more frequently as "walk-ins" (79% vs 73% respectively), in ambulance (12% vs 17% respectively), and at night (66% and 64% respectively). The most frequent main access diagnoses were “trauma or poisoning” (27% Italians vs. 21% HMPC), followed by “symptoms, signs, ill-defined conditions and unknown causes of morbidity” (16% Italians vs. 18% HMPC), in both populations.
Overall standardized ER access rates for Italians and HMPCs were 279.1 per 1,000 py (95%CI 279.1 - 279.2) and 275.6 per 1,000 py (95%CI 275.5 - 275.6), respectively. Excluding the pandemic years, the lowest values were observed in 2012 (266 per 1,000 py) in Italians and in 2015 (250 per 1,000 py) in HMPCs, while the highest values were observed in 2019 (321 and 317 per 1,000 py, respectively) for both populations. During the pandemic years, as expected, a decrease in access rates was observed in both populations (213 and 240 for Italians versus 212 and 249 for HMPCs in 2020-2021, respectively).
SRRs showed that HMPC men had lower access rates than Italian men between 2012 and 2020, while HMPC women had lower access rates than Italian women between 2014 and 2018; SRRs were above 1 in other years (Figure 1).
In the stratified analysis by age group, the standardized ED access rates and SRRs comparing the HMPC with Italians were: 341.7 versus 314.7 in the 0-19 age group (SRR 1.085, 95%CI 1.085-1.086); 272.6 versus 231.8 in the 20-65 age group (SRR 1.176, 95%CI 1.176-1.176); 243.0 versus 381.9 in the over 65 age group (SRR 0.636, 95%CI 0.636-0.637).
Overall, ED admission rates standardized by HMPC and Italian triage categories were, respectively: emergent 4.9 vs 6.4 per 1000 py (SRR 0.757, 95%CI 0.756-0.758); urgent 68.6 vs 75.1 per 1000 py (SRR 0.913, 95%CI 0.913-0.914); less urgent 169.7 vs 171.3 per 1000 py (SRR 0.991, 95%CI 0.991-0.991); non-urgent 24.9 vs 18.9 per 1000 py (SRR 1.317, 95%CI 1.316-1.318).
Conclusions
The use of healthcare utilization databases has allowed us to assess the heterogeneity of emergency department access in Migrant and Italian populations. However, analyses stratified by age and triage highlighted a higher emergency department access in Migrants younger than 65 years and for non-urgent conditions compared to Italian citizens. Further analyses are needed to identify factors associated with different emergency department access based on citizenship, in order to promote healthcare strategies for appropriate access to emergency care
Developing a Framework for Assessing the Applicability of the Target Condition in Diagnostic Research
No full textIntroduction: Assessment of the applicability of primary studies to the review questions is an essential but often challenging aspect of systematic reviews of diagnostic test accuracy studies (DTA reviews).
Objectives: To explore review authors’ applicability assessments for the QUADAS-2 reference standard domain within Cochrane DTA reviews, by highlighting applicability concerns, identifying potential issues with assessments. Using the findings to develop a framework for assessing the applicability of the target condition as defined by the reference standard.
Methods: Methodological review. DTA reviews in the Cochrane Library that used QUADAS-2 and judged applicability for the reference standard domain as “high concern” for at least one study, were eligible. One reviewer extracted the rationale for the “high concern”, this was checked by a second reviewer. Two reviewers categorized the rationale inductively into themes, a third reviewer verified these. Discussions in the QUADAS development group (steering group and expert panel) regarding the extracted information informed framework development.
Results: We identified 50 eligible reviews. We identified potential issues over half (n=28) of the included reviews. In 7, applicability assessments deviated from QUADAS-2 definitions and in 21 applicability concerns were insufficiently described. Five themes emerged in the remainder: study uses different reference standard threshold to define the target condition (6 reviews), misclassification by the reference standard in the study such that the target condition in the study does not match the review question (11 reviews), reference standard could not be applied to all participants resulting in a different target condition (5 reviews), misunderstanding QUADAS-2 applicability (7 reviews), and insufficient information (21 reviews).
The Figure shows the framework that was informed by our findings and discussions. Our framework outlines four potential applicability concerns for the assessment of the target condition as defined by the reference standard: different sub-categories of the target condition, different threshold used to define the target condition, reference standard not applied to full study group, and misclassification of the target condition by the reference standard (Figure).
Conclusion: Clear sources of applicability concerns are identifiable, but several Cochrane review authors struggle to adequately identify and report them. We have developed the first applicability framework to guide review authors in their assessment of applicability concerns for the QUADAS reference standard domain. The development of a framework for the QUADAS-2 domains Patient selection and Index Test are ongoing
Ventricular Volume as the most Informative Biomarker of TSPO-PET Binding Status in Multiple Sclerosis Patients
No full textIntroduction. Ability to identify multiple sclerosis (MS) patients burdened by smouldering inflammation is of great importance for therapeutic and clinical trial purposes [1]. PET imaging using 18 kDa translocator protein (TSPO)-binding radioligands can be used as an imaging biomarker for quantitation of glial activation in vivo [2]. However, large scale use of PET is challenging. Proxy biomarkers for TSPO-PET outcomes would therefore be helpful.
Objectives. The objective of the study was to identify key predictors of high TSPO-binding status in MS brain (HOT-PET phenotype) to provide widely usable tools for identification of patients with significant smouldering inflammation. Moreover, establishing a valid link between TSPO-PET and MRI measures would support the use of MRI as a proxy for microglial activation, given the limited accessibility of TSPO-PET.
Methods. The sample included 128 MS patients (92 RRMS, 36 progressive). 3T magnetic resonance imaging (MRI) was performed and the images were processed in MATLAB and segmented using FreeSurfer. PET imaging using the TSPO-binding radioligand [11C]PK11195 was performed. Patients were classified according to their TSPO binding status: threshold for HOT-PET was set at 9.6% of active voxels in the white matter (based on TSPO-binding in an SPMS cohort). Partial least squares discriminant analysis (PLS-DA) extracted independent subspaces of variables best explaining outcome variability (high or low TSPO-PET binding status). Based on variable importance in projection (VIP) scores, significant predictors of high TSPO binding were combined into a decision tree model to quantify misclassification error and refine the predictive framework [3,4].
Results. Based on the highest VIP scores, we constructed a decision tree using MRI-derived features alone. Ventricle parenchymal fraction (%PF) consistently emerged as a key predictor across all models, with a stable threshold around 2.12%. When combined with thalamus %PF, brain %PF, and age, and after tuning model parameters, we achieved an improved test accuracy of 0.84, outperforming the ventricle-only model (accuracy = 0.74). Key decision nodes included ventricle %PF ≥ 2.12 and age ≥ 38.4, which were strongly associated with the HOT-PET phenotype. Even among patients with smaller ventricles, low thalamic and brain %PFs contributed to identifying those with high glial activation.
To further characterize these patterns, we plan to stratify patients by MS subtype (RRMS vs PMS) and ventricle size, assessing whether enlarged ventricles in RRMS suggest a progressive-like inflammatory profile or whether some PMS patients with preserved brain volumes show lower microglial activation.
Conclusion. MRI-based volumetrics offer a practical strategy to identify MS patients with high glial activation when PET is unavailable or difficult to perform. While ventricle PF% alone was a strong proxy biomarker, classification improved by including thalamus and brain volume PF%, as well as age. This multivariable approach can support better patient stratification for phase 3 trials targeting microglial activation and help clinicians screen those most likely to benefit from microglia-targeted therapies
Prediction of Risk of Disease in Children at Risk of Facioscapulohumeral Muscolar Dystrophy with Machine Learning Approach
No full textIntroduction
Facioscapulohumeral muscular dystrophy (FSHD) (MIM#158900) is one of the most prevalent forms of muscular dystrophy, characterized by progressive skeletal muscle weakness, primarily affecting the face, shoulders, and upper arms. Its genetic basis is complex and typically involves contractions of the D4Z4 repeat region on the 4q subtelomere, even though it might be still incompletely described [1].
As a hereditary disease, an accurate risk assessment is crucial for improving genetic counselling, especially in the context of pregnancy planning. However, due to the significant variability in clinical manifestation and progression and the age-dependent penetrance of the disease, predicting the probability and severity of FSHD in newborns poses several challenges [2], and no tools are available for clinicians for this purpose.
Objectives
The aim of this study was to develop a machine learning model aimed at enhancing FSHD disease risk prediction for child of D4Z4 alleles of reduced size (DRA) carriers. In particular, our study focused on designing a predictive tool which can estimate the probability of FSHD and the age of disease onset in newborns, given the information of parents and other family members.
Methods
This predictive model was estimated on the basis of genetic, clinical and socio-demographic data collected in the Italian National Registry for FSHD [3]. Clinical data includes presence and severity of FSHD symptoms, measured using the FSHD Score [4], and a standardized description of clinical phenotypes, obtained through the Comprehensive Clinical Evaluation Form (CCEF) [5]. The availability of detailed family trees allowed the model to include the information carried by each family member, weighted by the degree of kinship with respect to subject involved in the genetic counselling.
To be included in this study, each family must be composed by a child, a DRA carrier parent, and may include one or more relatives. Families which did not fit in this structure were excluded. Since the expected FSHD onset age lies between 15 and 30 years of age, subject with age at visit less than 30 were also exclude, to limit misclassification.
For the development of the predictive model, we relied on a stacking approach: 4 base learners (a generalized regression model (GLM), a random forest (RF), a support vector machine (DVM) and a Bayesian network (BN)) provided a first-level individual prediction. Subsequently, these first-level predictions were combined by a random forest meta-learner to obtain the final predictions. Figure 1 outlines the model structure.
Leave-One-Out cross-validation was used to train each base learner (except BN) and the meta-learner. The parameters and hyperparameters of GLM, RF and SVM were estimated via grid search within each cross-validation loop, using the classification error as performance measure.
The BN learning procedure articulated in two steps: in first place, the structure of the network is established by defining the causal connections among all features, relying on expert knowledge. Then, the probability parameters that describe how the variables influence each other are estimated with a Bayesian approach. Non informative prior distributions were assumed at each non-deterministic node of the network. Since all variables have been previously categorized, each node likelihood was a Multinomial distribution, and a Dirichlet prior was used [6]. The network was embedded also with a set of deterministic nodes, which were introduced to process family trees with different structure and depth, and to reduce the model complexity.
The prediction accuracy of the model for each outcome (occurrence of FSHD and age at onset of first symptoms) was estimated using Leave-One-Out cross validation.
Based on the probabilities estimated from the model, the child of each family was predicted as with FSHD phenotype or as asymptomatic/healthy and assigned to an estimated age at onset class (No Onset, ≤22 years, <22 years). Youden Index was used to estimate the optimal probability cutoffs.
Results
A total of 293 families were included in the study. Of these, 121 families contributed to the estimation of risk of disease base learners, whereas 104 families contributed to the age at onset base learners.
For the evaluation of risk of disease, the model showed an area under the ROC curve (AUC) equal to 0.89. With the selected probability cutoff, the sensibility was equal to 0.90 and the specificity 0.70. The accuracy was 0.75, with 91 out of 121 children correctly assigned to their actual clinical status.
For the estimation of age at disease onset, the model reached a multi-class AUC equal to 0.88, with an accuracy of 0.72 (75 out of 104 children’s age at onset correctly predicted).
Overall, 70 children (67.3%) were correctly assigned to both their actual clinical status and their onset age class.
Conclusion
The developed predictive model was able to provide accurate estimates of disease probability in children of patients characterised by FSHD symptomatology, even though it was not able to discriminate between finer clinical categories. These findings further support the hypothesis that additional elements, such as other genetic variants and environmental factors, must be consider for predictive purposes.
Nevertheless, this model can lead to significant advancements in FSHD genetic counselling and in implementing personalized medicine practices. Notably, our model is based on a very limited number of variables. So, it can be easily applied to provide tailored advice for families at risk of FSHD in real life scenarios.