French National Health Insurance Database Research Articles

Survival Outcomes of Primary versus Interval Cytoreductive Surgery for FIGO IV Ovarian Cancer (SOFI-4): a Nationwide Population-based Target Trial Emulation

The effect of primary cytoreductive surgery versus interval cytoreductive surgery on FIGO IV ovarian cancer outcomes remains uncertain, and may vary depending on the stage and the extraperitoneal metastasis location. Emulating target trials through causal assessment combined with propensity score adjustment has become a leading method for evaluating interventions using observational data. To assess the effect of primary versus interval cytoreductive surgery on progression-free and overall survival in patients with FIGO IV ovarian cancer using target trial emulation. Utilizing the comprehensive French national health insurance database, we emulated a target trial to explore primary versus interval cytoreductive surgery causal impacts on FIGO IV ovarian cancer prognosis (Surgery for Ovarian cancer FIGO 4: SOFI-4). The clone method with inverse probability of censoring weighting was used to adjust for informative censoring and balance baseline characteristics between the groups. Subgroup analyses were conducted based on FIGO stages and extraperitoneal metastasis locations. The study included patients under 75 years of age, in good health condition, diagnosed with FIGO IV ovarian cancer between January 1, 2014, and December 31, 2022. The primary and secondary outcomes were respectively five-year progression-free survival and seven-year overall survival. Among the 2,772 patients included in the study, 948 (34.2%) were classified as FIGO IVA and 1,824 (65.8%) as FIGO IVB at inclusion. Primary cytoreductive surgery was performed on 1,182 patients (42.6%), while interval cytoreductive surgery was conducted on 1,590 patients (57.4%). The median progression survival for primary cytoreductive surgery was 19.7 months (interquartile range [IQR]: 19.3-20.1), compared to 15.7 months (IQR: 15.7-16.1) for those who underwent interval cytoreductive surgery. The median overall survival was 63.1 months [IQR: 61.7-65.4] for primary cytoreductive surgery, in comparison to 55.6 months [IQR: 53.8-56.3] for interval cytoreductive surgery. The findings of our study indicate that primary cytoreductive surgery is associated with a 5.0-month increase in five-year progression-free survival (95% Confidence Intervals [CI]: 3.8-6.2) and a 3.9-month increase in seven-year overall survival (95% CI: 1.9-6.2). These survival benefits of primary over interval cytoreductive surgery were observed in both the FIGO IVA and IVB subgroups. Primary cytoreductive surgery demonstrated improved progression-free survival and overall survival in patients with pleural, supra-diaphragmatic, or extra-abdominal lymph node metastasis. SOFI-4 advocates for the benefits of primary cytoreductive surgery over interval cytoreductive surgery for patients with FIGO IV ovarian cancer, suggesting extraperitoneal metastases like supra-diaphragmatic or extra-abdominal lymph nodes should not automatically preclude primary cytoreductive surgery consideration in suitable patients.

Association of positive airway pressure termination with mortality and non-fatal cardiovascular events in patients with obstructive sleep apnoea

Background and aimsThe recurrence of obstructive sleep apnoea (OSA) after positive airway pressure (PAP) therapy termination has physiological consequences that may increase cardiovascular (CV) risk. We aimed to determine whether...

A retrospective observational study of ibrutinib in chronic lymphocytic leukaemia in a real-life setting in France using the national claims database (OSIRIS)

BackgroundIbrutinib is a Bruton’s tyrosine kinase inhibitor indicated for the first-line treatment and relapse of chronic lymphocytic leukaemia (CLL), Waldenström’s macroglobulinemia (WM) and mantle cell lymphoma (MCL). This study aimed to describe the characteristics of CLL patients treated with ibrutinib and its effectiveness, safety, and treatment pattern in real life.MethodsAll patients covered by the general health scheme (approximately 80% of the French population) with a first ibrutinib dispensation from August 1, 2017 (date of reimbursement in France) to December 31, 2020, were identified in the French National Health Insurance database (SNDS). An algorithm was developed to identify the disease (CLL, MCL or WM) for which ibrutinib was prescribed. This article focused on CLL patients. The time to next treatment (TTNT) was plotted using Kaplan‒Meier curves.ResultsDuring this period, 6,083 patients initiated ibrutinib, among whom 2,771 (45.6%) patients had CLL (mean age of 74 years; 61% of men). At ibrutinib initiation, 46.6% of patients had a cardiovascular comorbidity. Most patients (91.7%) were not hospitalized during the exposure period for one of the cardiovascular or bleeding events studied. Hospitalizations were more frequent in patients with a cardiovascular comorbidity (5.9% versus 11.0%, p-value < 0.0001) and aged over 70 (5.9% versus 9.4%, p-value < 0.0001). The median TTNT was not reached.ConclusionThis is one of the largest cohorts of ibrutinib-treated patients in the world. The profile of CLL patients treated with ibrutinib was in accordance with the marketing authorization and reimbursement. This study confirmed effectiveness and safety data.

Impact of pembrolizumab treatment duration on overall survival and prognostic factors in advanced non-small cell lung cancer: a nationwide retrospective cohort study

The efficacy of front-line pembrolizumab has been established in studies that limit treatment duration to 2 years, but decision to stop pembrolizumab after 2 years is often at physician's discretion. ATHENA is a retrospective cohort study using a comprehensive administrative database aimed firstly at exploring the optimal duration of pembrolizumab and secondly real-life prognosis factors in patients with advanced non-small cell lung cancer (NSCLC). Using the French National Health Insurance database (SNDS), we identified patients with incident lung cancer in France from 2015 to 2022. Treatments and patients' characteristics were extracted or inferred from hospital, outpatient care, pharmacy delivery reports. The duration's hazard ratio (HR) was estimated with Cox model weighted by inverse of propensity score to account for confounding. Prognostics factors in first line population were identified with Cox model selected by a LASSO procedure. 391,106 patients with lung cancer were identified, of whom 43,359 received up-front pembrolizumab for an advanced disease. There were 67% (29,040/43,359) of male and the median age at diagnosis was 65 years old. After a median follow-up time of 25.9 months (min-max, [0-97.6]), the median overall survival (OS) after pembrolizumab initiation in first line was 15.7 [CI 95, 15.3-16.0] months. In multivariable analysis, several covariables were independently associated with worse OS, including male sex with chemo-immunotherapy, age, hospital category, high deprivation index, inpatient hospitalization for first pembrolizumab, and history of diabetes, diuretic, beta blocker, painkiller prescription. At landmark time of 29 months after pembrolizumab initiation, continuation beyond 2 years was not associated with better OS than a fixed 2-year treatment, HR=0.97 [0.75-1.26] p=0.95. This study supports the notion that stopping pembrolizumab after 2 years could be safe for patients with advanced NSCLC. However, because observational studies are prone to confounding and selection bias, causality cannot be affirmed. This study did not receive any specific grant.

HIV status and lay bystander cardiopulmonary resuscitation initiation for witnessed cardiac arrest

IntroductionEarly initiation of cardiopulmonary resuscitation (CPR) by bystanders of out-of-hospital cardiac arrest (OHCA) significantly improves survival and neurological outcomes. However, misconceptions about human immunodeficiency virus (HIV) transmission risk during CPR can deter lay bystanders from performing resuscitation. The aim of this study was to compare the rate of CPR initiation by lay bystanders who witnessed OHCA in subjects with and without HIV infection. MethodsWe analysed data from the two French cardiac arrest registries (SDEC and RéAC) from 2012 to 2020. We identified HIV-positive individuals from the French National Health Insurance database for the SDEC registry, and directly from the RéAC registry data. We used logistic regression models to assess the association between CPR initiation by lay bystanders and the victim’s HIV status. ResultsOf 58,177 witnessed OHCA cases, 192 (0.3%) occurred in HIV-positive subjects. These individuals were younger, more often male, and presented more shockable initial rhythms compared with subjects without HIV. Overall, there was no difference in the CPR initiation rate according to the HIV status (57.3% vs 47.6%, adjusted odds ratio 1.11, 95% confidence interval 0.83–1.48). The CPR initiation rate also did not differ by location between victims with or without HIV (home: 57.7% vs 45.4%; public places: 56.0% vs 53.6%; p for interaction = 0.46). Survival and neurological outcomes at hospital discharge did not differ based on the HIV status. ConclusionsThis study revealed that the rate of CPR initiation by lay bystanders did not differ between HIV and non-HIV subjects during OHCA.

#581 Therapeutic management of patients with GPA and MPA: a nationwide observational study in France

Abstract Background and Aims Granulomatosis with polyangiitis (GPA) and microscopic polyangiitis (MPA) are rare systemic necrotizing vasculitis. Treatment per rituximab was approved in 2013, and recent guidelines recommend corticosteroids sparing. A real-life pharmacoepidemiologic retrospective study was set up to describe the demographic characteristics of hospitalized GPA and MPA patients and compare the treatments of patients included before rituximab approval (over 2010–2012, Group 1) and after rituximab approval in France (over 2014–2017, Group 2). Method All adult patients hospitalized for GPA or MPA between January 1, 2010 and December 31, 2017 were included from the French National Health Insurance database (SNDS) which covers &amp;gt;97.5% of French population. Patients were separated in 2 groups according to their date of inclusion: (i) Over 2010-2012 for group 1 and (ii) over 2014-2017 for group 2. Descriptive analyses were performed. Kaplan-Meier method was used to model the duration of treatment induction and maintenance. Fine and Gray tests were used to compare treatment phases durations. Results The study involved 4445 prevalent GPA patients (including 1578 incident patients) and 1833 prevalent MPA patients (878 incident patients). Group 1 consisted of 694 and 283 GPA and MPA patients respectively; group 2 included 668 and 463 GPA and MPA patients respectively. Mean age and the sex ratio were stable over time. During the first year after inclusion, the proportion of patients with first-time hospitalized infections, diabetes or renal failure increased between 2010–2012 and 2014–2017. Between the two inclusion periods, the proportions of patients treated with rituximab increased both in the induction and maintenance phases whereas treatment with azathioprine declined (Table 1). As for methotrexate, cyclophosphamide and glucocorticoids, the proportions remained stable. Conclusion According to real-life, pharmacoepidemiologic data in GPA and MPA patients, the introduction of rituximab reduced the use of azathioprine without affecting the use of glucocorticoids, although corticosteroids sparing should become a standard of care with the emerging treatment regimens. No reductions in adverse events were observed with the increased use of rituximab.

Evolution of therapeutic management of patients with ANCA associated vasculitis in France after licensing Rituximab use

IntroductionIn 2013, rituximab was approved in France for the treatment of ANCA-associated vasculitis (AAV). The aim of the study was to compare the treatment and health events of adult incident patients with granulomatosis with polyangiitis (GPA) and microscopic polyangiitis (MPA), included before rituximab approval (over 2010–2012, Group 1) and those included after rituximab approval (over 2014–2017, Group 2).MethodData were extracted from the French National Health Insurance database (SNDS) including outpatient health care consumption and hospital discharge forms. Comparisons between inclusion periods were performed using Wilcoxon and χ² tests. Kaplan-Meier method was used to model the duration of treatment induction, maintenance, and off-drug periods. Fine and Gray tests were used to compare treatment phase durations.ResultsA total of 694 GPA and 283 MPA patients were included in Group 1, while 668 GPA and 463 MPA patients were included in Group 2. Between the two inclusion periods, the proportions of patients treated with rituximab increased in the induction and maintenance phases whereas treatment with azathioprine declined. These proportions remained stable in the case of methotrexate, cyclophosphamide, and glucocorticoid-treated patients. Frequency of first-time hospitalized infections, diabetes and renal failure during the first year after inclusion increased for both groups.Limitations of the studyThis is a retrospective study based on claims data including only 76% of people covered by health insurance in France. The period studied includes the learning phase of using rituximab. This study lacks biological data and precise quantitative analysis for the use of steroids, therefore the criteria for establishing diagnosis and therapeutic choice were unknown.ConclusionsIntroduction of rituximab reduced the use of azathioprine without affecting the use of glucocorticoids or cyclophosphamide.

Triptan use in elderly over 65 years and the risk of hospitalization for serious vascular events

BackgroundSeveral studies have focused on the use of triptan and the risk of acute vascular events but the existence of such association is still debated and has never been quantified in patients over 65 years. To assess whether triptan use among older is associated with an increased risk of hospitalization for acute vascular events.MethodsA propensity score-matched cohort study was designed using the French national health insurance database linked to hospital stays. Patients aged ≥ 65 years, newly treated by triptans between 2011 and 2014, were included… The primary event was hospitalization for an acute ischemic vascular event within de 90 days following triptan initiation. Association with triptan exposure was investigated through cox regression model, considering exposure at inclusion, and with exposure as a time-varying variable A case-crossover (CCO) and a self-controlled case series (SCCS) analyses were also conducted to address potential residual confounding.ResultsThe cohort included 24, 774 triptan users and 99 096 propensity matched controls (mean (SD) age: 71 years (5.9), 74% of women). Within 90 days after cohort entry, 163 events were observed in the triptan group, and 523 in the control group (0.66% vs. 0.53%, adjusted hazard ratio (aHR) exposed/not exposed 1.25 95%CI [1.05–1.49]; aHR time−varying 8.74 [5.21–14.66]). The association was significant (CCO) for all events (adjusted odds ratio (aOR1.63 [1.22–2.19]) with a more consistent association with cerebral events (aOR 2.14 [1.26–3.63]). The relative incidence (RI) for all events was 2.13 [1.76–2.58] in the SCCS, for cardiac (RI: 1.67 [1.23–2.27]) and for cerebral events (RI: 3.20, [2.30–4.45]).ConclusionThe incidence of acute vascular events was low among triptan users. We found that triptan use among older may be associated with a low increased risk for acute vascular events, which may be more marked for cerebral events such as stroke, than for cardiac events.

Cost-Effectiveness of Sleeve Gastrectomy and Gastric Bypass as Revisional Surgery on Antidiabetic Reimbursement: A Nationwide Cohort Study.

This study compared the effectiveness of 4 main revisional bariatric surgery (RBS) sequences after sleeve gastrectomy (SG) and adjustable gastric banding (AGB), on the reimbursement of antidiabetic treatments in France. Few large-scale prospective cohort studies have assessed the changes in antidiabetic treatments after RBS. This nationwide observational population-based cohort study analyzed data from the French National Health Insurance Database. All patients who underwent primary SG and AGB in France between January 2012 and December 2014 were included and followed up until December 31, 2020. The changes in categories and costs of reimbursed antidiabetic treatments across different RBS sequences were assessed (presented as follows: bariatric surgery (BS)-RBS). Among the 107,088 patients who underwent BS, 6396 underwent RBS, 2400 SG-GBP (SG converted to gastric bypass [GBP] during follow-up), 2277 AGB-SG, 1173 AGB-GBP, and 546 SG-SG. Pre-RBS insulin was used in 10 (2.9%), 4 (0.9%), 8 (2.4%), and 10 (2.6%) patients, respectively. Two years after RBS, the treatment discontinuation or decrease (the change of treatment to a lighter one category rates [eg, insulin to bi/tritherapy]) was 47%, 47%, 49%, and 34%, respectively. Four years after RBS, the median annual cost per patient compared with baseline was lower (P < 0.01) for all sequences, except SG-SG (P = 0.24). The most notable effect concerned AGB-GBP (median of more than 220 euros to 0). This study demonstrated the positive impact of RBS over a 4-year follow-up period on antidiabetic treatments reimbursement, through the reduction or discontinuation of treatments and a significant decrease in costs per patient.

Nationwide Audit of Postoperative Mortality and Complications After Digestive Cancer Surgery: Will New Legal Thresholds be Sufficient?

French policymakers recently chose to regulate high-risk digestive cancer surgery (DCS). A minimum of five cases per year should be performed for each of the following types of curative cancer surgery: esophagus/esogastric junction (ECS), stomach (GCS), liver (LCS, metastasis included), pancreas (PCS), and rectum (RCS). This study aimed to evaluate the hypothetical beneficial effects of the new legal minimal volume thresholds on the rates of 90-day postoperative mortality (90POM) for each high-risk DCS. This nationwide observational population-based cohort study used data extracted from the French National Health Insurance Database from 1 January 2015-31 December 2017. Mixed-effects logistic regression models were performed to estimate the independent effect of hospital volume. During the study period, 61,169 patients (57.1 % male, age 69.7 ±12.2 years) underwent high-risk DCS including ECS (n = 4060), GCS (n = 5572), PCS (n = 8598), LCS (n = 10,988), and RCS (n = 31,951), with 90POM of 6.6 %, 6.9 %, 6.0 %, 5.2 %, and 2.9 %, respectively. For hospitals fulfilling the new criteria, 90POM was lower after adjustment only for LCS (odds ratio [OR],15.2; 95 % confidence interval [CI], 9.5-23.2) vs OR, 7.6; 95 % CI, 5.2-11.0; p < 0.0001) and PCS (OR, 3.6; 95 % CI, 1.7-7.6 vs OR, 2.1; 95 % CI, 1.0-4.4; p<0.0001). With higher thresholds, all DCSs showed a lower adjusted risk of 90POM (e.g., OR, 0.38; 95 % CI, 0.28-0.51) for PCS of 40 or higher. Based on retrospective data, thresholds higher than those promulgated would better improve the safety of high-risk DCS. New policies aiming to further centralize high-risk DCS should be considered, associated with a clear clinical pathway of care for patients to improve accessibility to complex health care in France.

Medication abortion during the COVID-19 pandemic in France: A research based on the French national health insurance database.

During the COVID-19 pandemic in France, abortion was recognized as an essential service that cannot be delayed, and such care was therefore presumed to be maintained. The aim is to analyze the changes in the practice of abortion in 2020 to identify the consequences of the two lockdowns and the effects of the extension of the legal time limit. We analyzed the data collected by the French national health insurance system, which covers 99% of the population. All women who had an elective abortion, either surgical or with medication, in France in 2019 and 2020 were included in the study. Trend changes in abortions were analyzed by comparing the ratio of the weekly number of abortions in 2020 with the weekly number in 2019. Both 2020 lockdowns were followed by a drop in abortions, particularly after the first and stricter lockdown. This may be explained not by an abrupt shutdown of access to abortion services, but rather by a decrease in conceptions during the lockdown weeks. The decrease was more marked for surgical abortions than for medication abortions in a hospital setting, and less so for medication abortions in non-hospital settings. Moreover, the proportion of the latter type of abortions continued to increase, showing the reinforcement of a previous trend. Our findings indicate that expanding the legal time limit for abortion, diversifying the settings where abortions can be performed and the range of abortion providers help to facilitate access to this fundamental reproductive care.

P434 Incidence of chronic end-stage renal disease in patients with inflammatory bowel disease: a population-based cohort study from the French National Health Insurance Database

Abstract Background Renal disorders are among the extra-intestinal manifestations of inflammatory bowel disease (IBD), and treatments may also lead to renal failure. However, the incidence of end-stage renal disease (ESRD), defined as the need for renal replacement therapy (dialysis and/or renal transplantation(RT)), is poorly known. The primary objective of our study was to assess the incidence of ESRD in IBD patients, and to compare it with the non-IBD population. Methods We conducted a prognostic cohort study based on data from the French National Health Insurance Database(SNDS). Prevalent cases of IBD affiliated to the general health insurance scheme during the period 2011-2020 were identified on the basis of declarations long-term diseases ALD 24 or the presence of 2 hospital stays for IBD with ICD-10 codes or a single stay with reimbursement of specific treatments. Incident cases of ESRD were identified on the basis of a stay with a RT or a period with dialysis sessions for at least 45 days. Crude incidence rates of ESRD were calculated over the period for the IBD and the non-IBD population, with their 95% confidence intervals. Results A total of 280,283 IBD patients were included in the study over the period 2011-2020, 134,567 with ulcerative colitis(UC) and 145,716 with Crohn’s disease(CD). The mean age at study inclusion was 39.8 (± 17.1) years. The M/F sex ratio was 0.9. The mean age at incidence of ESRD in the non-IBD population was 65.0 (±17.0) years, the M/F sex ratio cases was 1.7. Among IBD patients, 534 patients began replacement therapy by dialysis or RT, they were younger than non-IBD patients (60.3(±16.7)years, P&amp;lt;0.001).The mean incidence rate of ESRD over the period was 26.4 per 100,000 person-years (IC95% 23.1-29.6), 480 (89.9%) patients entered ESRD by dialysis, and 54 (10.1%) by preemptive RT. In the non-IBD population the mean rate of ESRD was 15.9 per 100,000 person-years (IC95% 14.8- 17.1; P&amp;lt;0.0001 compared with IBD): 79,606 (94.5%) patients entered ESRD by dialysis, and 4,634 (5.5%) by RT (as compared with IBD,P&amp;lt;0.001). In subgroup analysis, the mean incidence rate of ESRD was 22.8 per 100,000 person-years (IC95% 18.9-26.7) in UC patients and 29.5 (IC95% 25.3-33.6) in CD patients. The rates of preemptive RT were 12.4% in UC patients (87.6% dialysis) and 8.6% in CD patients (91.4% dialysis). Over a median follow-up of 6.8 years, the survival rate after ESRD was 40.8% in ESRD-IBD patients VS 34.2% in ESRD non-IBD patients (P&amp;lt;0.001). Conclusion This study shows that the incidence of ESRD is significantly higher in IBD than in the non-IBD population, confirming the need for regular monitoring of renal function in our patients. The high incidence of RT confirms that access to transplantation is not restricted in this population.

Identification and Economic Evaluation of Differentiated Thyroid Cancer Care Consumption Patterns Using Sequence Analysis.

Objectives: This study aims to assess the impact of care consumption patterns and individual characteristics on the cost of treating differentiated thyroid carcinoma (DTC), in France, with a specific emphasis on socioeconomic position. Methods: The methodology involved a net cost approach utilising cases from the EVATHYR cohort and controls from the French National Health Insurance database. Care consumption patterns were created using Optimal Matching and clustering techniques. The individual characteristics influence on patterns was assessed using multinomial logistic regression. The individual characteristics and patterns influence on care costs was assessed using generalised estimating equations. Results: The findings revealed an average cost of €13,753 per patient during the initial 3years. Regression models suggested the main predictors of high DTC specific care consumption tended to include having a high risk of cancer recurrence (OR = 4.97), being a woman (OR = 2.00), and experiencing socio-economic deprivation (OR = 1.26), though not reaching statistical significance. Finally, high DTC-specific care consumers also incurred higher general care costs (RR = 1.35). Conclusion: The study underscores the increased costs of managing DTC, shaped by consumption habits and socioeconomic position, emphasising the need for more nuanced DTC management strategies.

Strong social disparities in access to IVF/ICSI despite free cost of treatment: a French population-based nationwide cohort study

BackgroundAccess to IVF/ICSI is facilitated when the financial barrier is removed. In a national context where in vitro fertilisation (IVF)/intracytoplasmic sperm Injection (ICSI) treatment is cost-free, how many women do not access IVF/ICSI and what are the factors associated with non-access?MethodsUsing French national health insurance databases, the cohort included 20,240 women aged 18–43 years living in France who underwent unsuccessful treatment (no pregnancy) with clomiphene citrate (CC) and/or gonadotropins with treatment started between January and August 2016. The outcome measure was non-access to IVF/ICSI during the 24-month following start of infertility care. Factors associated with non-access to IVF/ICSI were explored using mixed effects logistic regression.ResultsIn the cohort, 65.4% of women did not access IVF/ICSI. In multivariable analysis, non-access to IVF/ICSI was higher in younger women (18–25 years: (OR 2.17, 95% CI: 1.85–2.54) and in older women (40–43 years: (OR=3.60, 95% CI: 3.25–3.98)). Non-access was higher among women below the poverty line (OR=3.76, 95% CI: 3.34–4.23) and showed a significant upward trend with increasing deprivation of place of residence. Distance to the nearest fertility centre was not significantly associated with non-access to IVF/ICSI.ConclusionsIn a national context of cost-free ART treatment, a large proportion of women did not access treatment, with a strong social gradient that raises important issues. We need to understand the underlying social mechanisms to develop an efficient and equitable health policy regarding infertility care.

Abstract 14337: Identification of New Clinical Phenotypes of Sudden Cardiac Death Using Artificial Intelligence

Introduction: Current classifications of Sudden Cardiac Death (SCD) are based on cardiovascular (CV) conditions and risk factors, and frequently fail to explain numerous SCD cases. Hypothesis: a data-driven machine learning approach could identify new clinical phenotypes of SCD in the general population, using a wide range of CV and non-CV variables. Methods: All case of SCD over 18y that occurred between 2011-15 in Paris and suburbs were included in the Paris Sudden Death Expertise Center. Data from the French National Health Insurance database that occurred up to 10y before SCD were collected (drugs, examinations,diagnosis). We performed a non-supervised statistical approach to identify relevant clinical phenotypes of SCD, using clustering and dimension reduction methods. A validation was performed between 2016-20. Results: 12,651 SCD subjects were included. Mean age was 68y, and 61% were men. The clustering analysis identified 8 distinct clusters of SCD, with homogeneous characteristics. 3 clusters were already well known: a group mainly composed of aged women (3,307, 38% men, 81y), one with mainly CV diseases (3,638, 74% men, 74y) and one of young subjects with very few comorbidities (2,645, 65%men, 54 y). 4 non-CV clusters were identified: a group with psychiatric and/or neurologic disorders (862), a group with chronic or acute pulmonary diseases (688) and a group with malignant neoplasms of different sites and related drugs (518), and a group with kidney diseases (279). The last group was composed of subjects with characteristics related to social deprivation (alcohol related disorder, hepatitis, HIV, etc) (252).The same 8 clusters were identified in the validation cohort (11,485). Conclusions: By extending far beyond CV pathology, our approach provides for the first time a global picture of SCD, revealing the involvement of extra CV medical fields that might eventually lead to discover new pathways and help identifying high risk subjects.

Abstract 347: Artificial Intelligence for Targeted and Personalized Prediction of Sudden Cardiac Death in the General Population: A Machine Learning Study Based on Electronic Health Records

Background: Identification of subjects from the general population with an increased risk of Sudden Cardiac Death (SCD) remains highly challenging, particularly at an individual level. Methods: Every case of SCD + 18 y.o. that occurred since 2011 in Paris and its inner suburbs was prospectively included in the Paris Sudden Death Expertise Center (SDEC) registry (6.7M inhabitants). SCD cases were matched by age, sex, and residence area with a control group (70000) sampled from the general population. All data from the French National Health Insurance database (SNDS) were collected: prescribed drugs, examinations and diagnoses that occurred up to 10 y. before the event, including cardiovascular and non-cardiovascular variables (10000). We provide a detailed explanation of the prediction at the individual level. For each subject, we computed the Shapley scores from the supervised learning model (CatBoost) to evaluate the contribution of variables to its predicted risk. The model used was derived from 12338 SCD cases and matched controls cohort (2011-15) and validated on 11620 cases cohort (2016-20). Results: A total of 23958 equations have been built with personalized variables and coefficients, achieving an AUC of 0.80 in the derivation cohort (with a PPV=77%, sensitivity=68%). One example is shown below, with an individual predicted risk of SCD at 3 months (89.3%), and the most important factors that increase (red) or decrease (blue) the risk of SCD. Conclusions: For the first time, we provide a personalized prediction model of SCD, which explains the predicted risk at the individual level. These findings offer a new step toward personalized prevention and global public health interventions

The burden of very early dropout in infertility care: a nationwide population-based cohort study.

What is the frequency and the associated factors of very early dropout following unsuccessful clomiphene citrate (CC)/gonadotropin treatment in the context of full coverage of treatment cost. Despite free treatment, almost one in four women had a very early dropout following unsuccessful CC/gonadotropin treatment, with patients below the poverty line being more likely to drop out early. Success of infertility care is tarnished by very high dropout rates. Infertility care dropout has been considered as resulting principally from financial barriers because of the high cost of treatment. Nearly all previous work addressed dropout following IVF/ICSI. Factors associated with dropout following CC/gonadotropins may be different and also need to be investigated. Nationwide population-based cohort study. Using the French national health insurance and hospital databases, we included in the cohort 27 416 women aged 18-49 years unsuccessfully treated with CC/gonadotropins in 2017. The main outcome was very early dropout, defined as discontinuation of all infertility treatment after unsuccessful treatment for 1-3 months. Very early treatment dropout was analysed by multivariate logistic regression. Among women unsuccessfully treated with CC/gonadotropins, 22% dropped out of infertility care within 3 months. In multivariate analysis, higher early dropout following unsuccessful CC/gonadotropin treatment was associated with older and younger ages (≥35 and <25 years), being below the poverty line, being treated with CC prescribed by a general practitioner and lack of infertility tests or monitoring. This study is based on health administrative data that do not include reasons for dropout and record only a limited amount of information. It is thus not possible to analyse the reason for early dropout. Despite full coverage of all infertility treatment, women under the poverty line have a higher risk of very early dropout following unsuccessful CC/gonadotropin treatment. Better understanding is needed of the non-financial barriers and difficulties faced by these patients. To address disparities in infertility treatment, practitioner training could be reinforced to adapt to patients from different social and cultural backgrounds. This work was supported by the ANR StimHo project, grant ANR-17-CE36-0011-01 from the French Agence Nationale de la Recherche. The authors have no conflict of interest to declare. N/A.

Survival outcomes for patients with multiple myeloma in France: A retrospective cohort study using the Système National des Données de Santé national healthcare database.

Evaluate the overall survival (OS) of patients with multiple myeloma (MM) at different treatment stages in France. This retrospective observational cohort study used data from the French National Health Insurance database to study patients with MM (diagnosis 2013-2019). Patient outcomes included OS (all-cause mortality), time-to-next treatment (TTNT), and duration of therapy (DoT) from initial diagnosis, the start of different lines of therapy (LOTs), triple-class exposure (TCE), and subsequent treatment following TCE. The Kaplan-Meier method analyzed "time-to-event" data. From diagnosis, death rates increased from 1% at 1 month to 24% at 2 years; median OS was 63.8 months (N = 14 309). Median OS from the start of LOTs declined from 61.0 months (LOT1) to 14.8 months (LOT4). Median OS from TCE start was 14.7 months. There was a large variation in TTNT within LOTs (e.g., LOT1: bortezomib + lenalidomide: TTNT = 26.4 months, OS = 61.7 months; lenalidomide: TTNT = 20.0 months, OS = 39.6 months); DoT was similar for LOT1 and LOT2, then progressively declined at LOT4. Patients with stem cell transplant, younger age, and less comorbidity had better survival outcomes. Patients with MM face a poor prognosis after relapse to multiple LOTs and TCE, demonstrating a worsening of survival outcomes. Access to novel therapies may improve outcomes.

CO4.2 - Cost-effectiveness of an adaptive radiotherapy strategy to decrease xerostomia in locally advanced oropharynx carcinoma: An economic evaluation alongside the ARTIX Randomized Clinical Trial matched to the French National Health Insurance Database

CO4.2 - Cost-effectiveness of an adaptive radiotherapy strategy to decrease xerostomia in locally advanced oropharynx carcinoma: An economic evaluation alongside the ARTIX Randomized Clinical Trial matched to the French National Health Insurance Database

Drug-Drug Interactions and the Risk of Emergency Hospitalizations: A Nationwide Population-Based Study.

Several studies suggest a significant risk of hospitalization because of drug-drug interactions in the general population. However, to our knowledge, this risk has never been measured precisely in a large population. We aimed to estimate the risk of emergency hospitalization associated with exposure to the contraindicated concomitant use of interacting drugs in the general population. A self-controlled case-series analysis was carried out on a cohort of 150,000 subjects randomly selected from the French national health insurance database, between 01/01/2016 and 31/12/2016. Exposure to the contraindicated concomitant use of interacting drugs was defined as the overlapping period of dispensings of drugs contraindicated because of clinically meaningful drug-drug interactions. The main outcome, incidence rate ratios, comparing the incidence rate of emergency hospitalizations during each category of exposure time periods with that during the reference period, was estimated using the conditional Poisson regression model. Over the study period, 967 subjects were exposed to at least one contraindicated concomitant use of interacting drug and 177 had been exposed and presented at least one emergency hospitalization. Compared to the unexposed follow-up time, the risk of emergency hospitalization increased during exposure to contraindicated concomitant use of interacting drug periods (incidence rate ratio: 2.41; 95% confidence interval 1.55-3.76). This could translate into 7200 (4500-8900) potentially preventable emergency hospitalizations yearly in France. We evidenced an almost 2.5-fold increase in the risk of emergency hospitalizations during periods of exposure to contraindicated concomitant use of interacting drugs, with a potential public health impact exceeding 7000 preventable hospitalizations yearly in France. These results confirm the need to reinforce training in prescription practices and tools for prevention concerning contraindicated concomitant use of interacting drugs. These would especially concern drugs involved in an increase in long QT syndrome when associated such as citalopram, and highly prescribed drugs with a risk of overdose if co-prescribed with cytochrome P450 inhibitors, such as antigout and lipid-lowering drugs.