Design and validation of algorithms to identify venous thromboembolism in the French National Healthcare Database.

PURPOSE: This paper aims to introduce an algorithm designed to identify Venous Thromboembolism (VTE) in the French National Healthcare Database (SNDS) and to estimate its positive predictive value. METHODS: A case-identifying algorithm was designed using SNDS inpatient and outpatient encounters, including hospital stays with discharge diagnoses, imaging procedures and drugs dispensed, of French patients aged at least 18 years old to whom baricitinib or Tumor Necrosis Factor Inhibitors (TNFi) were dispensed between September 1, 2017, and December 31, 2018. An intra-database validation study was then conducted, drawing 150 cases identified as VTE by the algorithm and requesting four vascular specialists to assess them. Patient profiles used to conduct the case adjudication were reconstituted from de-identified pooled and formatted SNDS data (i.e., reconstituted electronic health records-rEHR) with a 6-month look-back period prior to the supposed VTE onset and a 12-month follow-up period after. The positive predictive value (PPV) with its 95% confidence interval (95% CI) was calculated as the number of expert-confirmed VTE divided by the number of algorithm-identified VTE. The PPV and its 95% CI were then recomputed among the same patient set initially drawn, once the VTE-identifying algorithm was updated based on expert recommendation. RESULTS: For the 150 patients identified with the first VTE-identifying algorithm, the adjudication committee confirmed 92 cases, resulting in a PPV of 61% (95% CI = [54-69]). The final VTE-identifying algorithm including expert suggestions showed a PPV of 92% (95% CI = [86-98]) with a total of 87 algorithm-identified cases, including 80 retrieved from the 92 confirmed by experts. CONCLUSION: The identification of VTE in the SNDS is possible with a good PPV.

Cardiovascular and renal diseases in type 2 diabetes patients: 5-year cumulative incidence of the first occurred manifestation and hospitalization cost: a cohort within the French SNDS nationwide claims database.

BACKGROUND: Myocardial infarction (MI), stroke, peripheral arterial disease (PAD), heart failure (HF) and chronic kidney disease (CKD) are common cardiovascular renal diseases (CVRD) manifestations for type 2 diabetes. The objective was to estimate the incidence of the first occurring CVRD manifestation and cumulative hospitalization costs of each CVRD manifestation for type 2 diabetes without CVRD history. METHODS: A cohort study of all type 2 diabetes free of CVRD as of January 1st 2014, was identified and followed-up for 5 years within the French SNDS nationwide claims database. The cumulative incidence of the first occurring CVRD manifestation was estimated using the cumulative incidence function, with death as a competing risk. Cumulative hospitalization costs of each CVRD manifestations were estimated from the perspective of all payers. RESULTS: From 2,079,089 type 2 diabetes without cancer or transplantation, 76.5% were free of CVRD at baseline with a mean age of 65 years, 52% of women and 7% with microvascular complications history. The cumulative incidence of a first CVRD manifestation was 15.3% after 5 years of follow-up with a constant linear increase over time for all CVRD manifestations: The most frequent was CKD representing 40.6% of first occurred CVRD manifestation, followed by HF (23.0%), then PAD (13.5%), stroke (13.2%) and MI (9.7%). HF and CKD together reached about one patient out of ten after 5 years and represented 63.6% of first CVRD manifestations. The 5-year global cost of all CVRD hospitalizations was 3.9 billion euros (B€), i.e. 2,450€ per patient of the whole cohort, with an exponential increase over time for each specific CVRD manifestation. The costliest was CKD (2.0 B€), followed by HF (1.2 B€), then PAD (0.7 B€), stroke (0.6 B€) and MI (0.3 B€). CONCLUSIONS/INTERPRETATION: While MI, stroke and PAD remain classic major risks of complications for CVRD-free type 2 diabetes, HF and CKD nowadays represent individually a higher risk and cost than each of these classic manifestations, and jointly represents a risk and a cost twice as high as these three classic manifestations all together. This should encourage the development of specific HF and CKD preventive strategies.

Clustering of prostate cancer healthcare pathways in the French National Healthcare database.

Background: Healthcare pathways of patients with prostate cancer are heterogeneous and complex to apprehend using traditional descriptive statistics. Clustering and visualization methods can enhance their characterization. Methods: Patients with prostate cancer in 2014 were identified in the French National Healthcare database (Système National des Données de Santé-SNDS) and their data were extracted with up to 5 years of history and 4 years of follow-up. Fifty-one-specific encounters constitutive of prostate cancer management were synthesized into four macro-variables using a clustering approach. Their values over patient follow-ups constituted healthcare pathways. Optimal matching was applied to calculate distances between pathways. Partitioning around medoids was then used to define consistent groups across four exclusive cohorts of incident prostate cancer patients: Hormone-sensitive (HSPC), metastatic hormone-sensitive (mHSPC), castration-resistant (CRPC), and metastatic castration-resistant (mCRPC). Index plots were used to represent pathways clusters. Results: The repartition of macro-variables values-surveillance, local treatment, androgenic deprivation, and advanced treatment-appeared to be consistent with prostate cancer status. Two to five clusters of healthcare pathways were observed in each of the different cohorts, corresponding for most of them to relevant clinical patterns, although some heterogeneity remained. For instance, clustering allowed to distinguish patients undergoing active surveillance, or treated according to cancer progression risk in HSPC, and patients receiving treatment for potentially curative or palliative purposes in mHSPC and mCRPC. Conclusion: Visualization methods combined with a clustering approach enabled the identification of clinically relevant patterns of prostate cancer management. Characterization of these care pathways is an essential element for the comprehension and the robust assessment of healthcare technology effectiveness.

Real-World Effectiveness of Osteoporosis Medications in France: A Nationwide Cohort Study.

Although drugs for osteoporosis have been demonstrated to be effective in reducing fracture risk in placebo-controlled clinical trials, data on effectiveness in real-world practice is limited. Data from the French national health insurance claims database (SNDS) were used to follow five cohorts of women aged ≥55 years after initiating treatment for ≥6 months with either denosumab, zoledronic acid, oral bisphosphonates, raloxifene, or teriparatide in 2014-2016. Fracture incidence was compared within each cohort between the 3 months following initiation (baseline fracture risk) and the 12month, 18month, and 24 month postinitiation periods. Data are presented as incidence rate ratios (IRRs) with their 95% confidence intervals (CIs)s. Overall, 67,046 women were included in the denosumab cohort, 52,914 in the oral bisphosphonate cohort, 41,700 in the zoledronic acid cohort, 11,600 in the raloxifene cohort, and 7510 in the teriparatide cohort. The baseline vertebral fracture rate ranged from 1.74 per 1000 person years (‰PY) in the raloxifene cohort to 34.75‰PY in the teriparatide cohort, and the baseline hip fracture rate from 0.70‰PY in the raloxifene cohort to 10.52‰PY in the zoledronic acid cohort. Compared with the baseline fracture rate, vertebral fractures involving hospitalization were significantly reduced in the 3-24-month postinitiation period with denosumab (IRR 0.6; 95% CI, 0.5-0.7), zoledronic acid (IRR 0.4; 95% CI, 0.3-0.4), teriparatide (IRR 0.3; 95% CI, 0.2-0.5), and oral bisphosphonates (IRR 0.6; 95% CI, 0.4-0.8). Hip fracture incidence was reduced with denosumab (IRR 0.8; 95% CI, 0.6-0.9), but higher for oral bisphosphonates (IRR 1.7; 95% CI, 1.2-2.3); no significant change in hip fracture rate was observed for zoledronic acid, teriparatide, or raloxifene. A reduction in nonvertebral, non-hip fracture incidence was observed only in the denosumab cohort (IRR 0.8; 95% CI, 0.7-0.9). These findings indicate that treatment with osteoporosis drugs is effective in the real-world setting. © 2023 The Authors. JBMR Plus published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research.

Risk of cancer with immunosuppressants compared to immunomodulators in multiple sclerosis: A nested case-control study within the French nationwide claims database.

PURPOSE: The objective was to compare the risk of malignancies in real-world settings between exclusive immunosuppressant (IS) and immunomodulator (IM) use in multiple sclerosis (MS). METHODS: A nested case-control study was designed within a new-user cohort of all patients with MS who initiated a first IM or IS between 2008 and 2014, and without cancer history, using the information of the SNDS nationwide French claims database. Incident cancer cases were matched with up to six controls on year of birth, sex, initiation date, and disease risk score of cancer. A conditional logistic regression (odds ratio [95% confidence interval]) was used to compare exclusive IS versus IM use during follow-up and according to three use durations. RESULTS: From 28 720 newly treated patients with MS, 407 incident cancers were observed during the follow-up with 2324 matched controls. A significant increase in cancer risk was observed for IS compared with IM (1.36 [1.05, 1.77]), with similar increases for the first 2 years of use but not for ≥2 years (1.06 [0.65, 1.75]). Similar increase was also observed for IS with indications other than MS (1.37 [1.04, 1.81]) but not for IS indicated only in MS (1.03 [0.45, 2.34]). CONCLUSIONS: Compared with IM, a 37% increase in cancer risk was observed for IS with indications other than MS and used for a short duration (≤2 years) but not for IS indicated only in MS. The absence of risk for prolonged exposure of IS with indications other than MS is not in favor of a causal relation with these drugs.

Strong instrumental variables biased propensity scores in comparative effectiveness research: A case study in oncology.

BACKGROUND AND OBJECTIVES: Some medications require specific medical procedures in the weeks before their start. Such procedures may meet the definition of instrumental variables (IVs). We examined how they may influence treatment effect estimation in propensity score (PS)-adjusted comparative studies, and how to remedy. STUDY DESIGN AND SETTING: Different covariate assessment periods (CAPs) did and did not include the month preceding treatment start were used to compute PS in the French claims database (Sytème National des Données de Santé-SNDS), and 1:1 match patients with metastatic castration resistant prostate cancer initiating abiraterone acetate or docetaxel. The 36-month survival was assessed. RESULTS: Among 1, 213 docetaxel and 2, 442 abiraterone initiators, the PS distribution resulting from the CAP [-12; 0 months] distinctly separated populations (c = 0.93; 273 matched pairs). The CAPs [-12;-1 months] identified 765 pairs (c = 0.81). Strong docetaxel treatment predictors during the month before treatment start were implantable delivery systems (1% vs. 59%), which fulfilled IV conditions. The 36-month survival was not meaningfully different under the [-12; 0 months] CAP but differed by 10% points (38% vs. 28%) after excluding month -1. CONCLUSION: In the setting of highly predictive pretreatment procedures, excluding the immediate pre-exposure time from the CAP will reduce the risk of including potential IVs in PS models and may reduce bias.

Real-world data and evidence in health technology assessment: When are they complementary, substitutes, or the only sources of data compared to clinical trials?

Within the life-cycle assessment of health technologies, real-world data (RWD) have until now been of secondary importance to clinical trial data. The availability of massive, better quality RWD, particularly with the emergence of connected devices, the improvement of methods for characterizing populations, make it possible to have a better insight into the effects of treatment, sometimes on a national scale the importance of RWD is likely to progress in the eyes of health technology assessors, going from being traditionally complementary to possibly replacing clinical trial data. This is the fundamental question that the round table, involving experts from the academic and/or hospital, institutional, and industrial worlds, set out to answer. This work served first to establish the current role of RWD in health technology assessment, by distinguishing the main purposes of RWD, the timing of the evaluation in relation to the life cycle of the technology, and then according to the party commissioning or receiving the outcomes of RWD-based studies. Secondly, the round table proposed six general recommendations for more intensive and decisive use of RWD in the assessment and decision-making process.

High-dimensional propensity scores for empirical covariate selection in secondary database studies: Planning, implementation, and reporting.

Real-world evidence used for regulatory, payer, and clinical decision-making requires principled epidemiology in design and analysis, applying methods to minimize confounding given the lack of randomization. One technique to deal with potential confounding is propensity score (PS) analysis, which allows for the adjustment for measured preexposure covariates. Since its first publication in 2009, the high-dimensional propensity score (hdPS) method has emerged as an approach that extends traditional PS covariate selection to include large numbers of covariates that may reduce confounding bias in the analysis of healthcare databases. hdPS is an automated, data-driven analytic approach for covariate selection that empirically identifies preexposure variables and proxies to include in the PS model. This article provides an overview of the hdPS approach and recommendations on the planning, implementation, and reporting of hdPS used for causal treatment-effect estimations in longitudinal healthcare databases. We supply a checklist with key considerations as a supportive decision tool to aid investigators in the implementation and transparent reporting of hdPS techniques, and to aid decision-makers unfamiliar with hdPS in the understanding and interpretation of studies employing this approach. This article is endorsed by the International Society for Pharmacoepidemiology.

Abiraterone acetate versus docetaxel for metastatic castration-resistant prostate cancer: a cohort study within the French nationwide claims database.

OBJECTIVES: To conduct the direct comparison of abiraterone acetate and docetaxel for first-line treatment of metastatic castration-resistant prostate cancer (mCRPC) in real-life settings. METHODS: Data were extracted from the French nationwide claims database (SNDS) on all men aged ≥40 years starting first-line treatment with abiraterone acetate or docetaxel for mCRPC in 2014. A high-dimensional propensity score including 100 baseline characteristics was used to match patients of both groups and form two comparative cohorts. Three-year overall survival and treatment discontinuation-free survival were determined using Kaplan-Meier analysis. RESULTS: In 2014, 2,444 patients started abiraterone for treatment of mCRPC and 1,214 started docetaxel. After trimming and matching, 716 patients were available in each group. Median overall survival tended to be longer in the abiraterone acetate cohort (23.8 months, 95% confidence interval = [21.5; 26.0]) than in the docetaxel cohort (20.3 [18.4; 21.6] months). Survival at 36 months was 34.6% for abiraterone acetate and 27.9% for docetaxel (p = 0.0027). Treatment discontinuation-free median was longer in the abiraterone acetate cohort compared to the docetaxel cohort (10.8 [10.1; 11.7] versus 7.4 [7.0; 8.0] months). CONCLUSION: The findings underline the interest of oral abiraterone acetate over intravenous docetaxel as the first-line treatment option in mCRPC.

Incidence and risk of cancer among multiple sclerosis patients: A matched population-based cohort study.

BACKGROUND AND PURPOSE: Studies have not yet found conclusive results on the risk of cancer in patients with multiple sclerosis (MS). This study aimed to compare the incidence of all cancers and of specific types of cancer between MS patients and the general population by age and by sex. METHODS: All prevalent MS patients identified between 2008 and 2014 in the nationwide French health care database (Système National des Données de Santé) and without history of malignancy were included in a cohort study and followed up until cancer occurrence, date of death, or 31 December 2015, whichever came first. MS patients were matched based on sex and year of birth to non-MS controls from the general population without cancer before index date. Incidence rate was reported per 100,000 person-years (PY), and risk of cancer was estimated by type of cancer, age, and sex using a Cox model (hazard ratio [HR] and its 95% confidence interval [CI]). RESULTS: Overall, 576 cancers per 100,000 PY were observed in MS patients versus 424 per 100,000 PY in the control population. The risk of cancer was higher among MS patients than among population controls whether considered overall (HR = 1.36, 95% CI = 1.29-1.43) or for prostate (HR = 2.08, 95% CI = 1.68-2.58), colorectal and anal (HR = 1.35, 95% CI = 1.16-1.58), trachea, bronchus, and lung (HR = 2.36, 95% CI = 1.96-2.84), and to a lesser extent, breast cancer (HR = 1.12, 95% CI = 1.03-1.23). CONCLUSIONS: MS patients were associated with increased risk of cancer compared to population controls.

Comparative effectiveness of dimethyl fumarate in multiple sclerosis.

AIMS: To assess the effectiveness of dimethyl fumarate (DMF) on annual rate of relapse subject to treatment (ARRt) and disability progression in multiple sclerosis (MS) compared to injectable immunomodulators (IMM), teriflunomide (TERI) and fingolimob (FTY), in real-life setting. METHODS: A population-based cohort study was conducted using data of the French nationwide claims database, SNDS. All patients initiating IMM, TERI, FTY or DMF between 1 July 2015 and 12 December 2017, with 4.5 years of database history and 1-3.5 years of follow-up were included in this study. DMF patients were 1:1 matched to IMM, TERI or FTY using a high dimensional propensity score. Negative binomial regression and a logistic regression model were used to estimate the relative risk (RR ± [95% CI]) of ARRt and the odds ratio (OR ± [95% CI]) of disability progression, respectively. RESULTS: Overall, 9304 subjects were identified: 29.0% initiated DMF, 33.2% TERI, 5.6% FTY and 32.2% an IMM. The matched cohorts consisted of 1779 DMF-IMM patients, 1679 DMF-TERI patients, and 376 DMF-FTY patients. DMF significantly reduced ARRt compared to IMM (RR 0.72 [0.61-0.86]) and TERI (0.81 [0.68-0.96]) and did not show any significant difference when compared with FTY. The risk of the progression of MS-specific disability was not significantly different for any matched cohorts. CONCLUSION: DMF is associated with lower risk of treated relapse for patients with RRMS than other first-line RRMS agents (TERI and IIM).

Patients with stable coronary artery disease and type 2 diabetes but without prior myocardial infarction or stroke and THEMIS-like patients: real-world prevalence and risk of major outcomes from the SNDS French nationwide claims database.

AIM AND HYPOTHESES: The THEMIS randomized trial compared ticagrelor plus aspirin versus placebo plus aspirin for patients with stable coronary artery disease and type 2 diabetes mellitus (CAD-T2DM), and without prior myocardial infarction (MI) or stroke. The aim of the study was to quantify the size of the CAD-T2DM population without prior MI or stroke population in a real-world setting, and more specifically populations with similar THEMIS selection criteria (THEMIS-like and THEMIS-PCI-like populations), as well as their risk of major outcomes in current practice. METHODS: A 2-year follow-up cohort study included all CAD-T2DM without MI/stroke prevalent patients on January 1st, 2014 in the SNDS French nationwide claims database. The THEMIS-like population concerned those ≥ 50 years of age with similar THEMIS inclusion and exclusion criteria. Prevalence was standardized to the European population. The cumulative incidence function was used to estimate the incidence of clinical outcomes (MI, ischemic stroke, and major bleeding according to the TIMI classification) with death as competing risk, and the Kaplan-Meier estimate for all-cause death and a composite outcome of MI, stroke and all-cause death. RESULTS: From a population of about 50 million adults, the prevalence of CAD-T2DM without MI/stroke, THEMIS-like and THEMIS-PCI-like populations was respectively at 6.04, 1.50 and 0.27 per 1000 adults, with a mean age of 72.7, 72.3 and 70.9 years and less comorbidities and diabetic complications for the THEMIS-like and THEMIS-PCI-like population. The 2-year cumulative incidence was respectively 1.7%, 1.3% and 1.6% for MI, 1.7%, 1.5% and 1.4% for stroke, 4.8%, 3.1% and 2.9% for major bleeding, 13.6%, 9.7% and 6.8% for all-cause death, and 16.2%, 12.0% and 9.5% for the composite outcome. CONCLUSION: THEMIS-like prevalence was estimated at 1.50 per 1,000 adults, representing about a quarter of CAD-T2DM without MI/stroke patients, and 0.27 per 1000 adults for the THEMIS-PCI-like populations. In current French practice, the median age of both these populations was about 5-6 years older than in the THEMIS trial, with a 2-year incidence of major outcomes between two or four time above the ones of the placebo arm of the THEMIS trial using very close definitions. Registration No. EUPAS27402 ( http://www.ENCEPP.eu ).

NSAIDs and COVID-19: A Systematic Review and Meta-analysis.

BACKGROUND: Nonsteroidal anti-inflammatory drugs (NSAIDs) have been discouraged for the treatment of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections, fearing that they could increase the risk of infection or the severity of SARS-CoV-2. METHODS: Original studies providing information on exposure to NSAIDs and coronavirus disease 2019 (COVID-19) outcomes were retrieved and were included in a descriptive analysis and a meta-analysis with Cochrane Revue Manager (REVMAN 5.4), using inverse variance odds ratio (OR) with random- or fixed-effects models. RESULTS: Of 92,853 papers mentioning COVID-19, 266 mentioned NSAIDs and 61 mentioned ibuprofen; 19 papers had analysable data. Three papers described NSAID exposure and the risk of SARS-CoV-2 positivity, five papers described the risk of hospital admission in positive patients, 10 papers described death, and six papers described severe composite outcomes. Five papers studied exposure to ibuprofen and death. Using random-effects models, there was no excess risk of SARS-CoV-2 positivity (OR 0.86, 95% confidence interval [CI] 0.71-1.05). In SARS-CoV-2-positive patients, exposure to NSAIDs was not associated with excess risk of hospital admission (OR 0.90, 95% CI 0.80-1.17), death (OR 0.88, 95% CI 0.80-0.98), or severe outcomes (OR 1.14, 95% CI 0.90-1.44). With ibuprofen, there was no increased risk of death (OR 0.94, 95% CI 0.78-1.13). Using a fixed-effect model did not modify the results, nor did the sensitivity analyses. CONCLUSION: The theoretical risks of NSAIDs or ibuprofen in SARS-CoV-2 infection are not confirmed by observational data.

Use of intravenous iron and risk of anaphylaxis: A multinational observational post-authorisation safety study in Europe.

PURPOSE: This post-authorisation safety study estimated the risk of anaphylaxis in patients receiving intravenous (IV) iron in Europe, with interest in iron dextran and iron non-dextrans. Studies conducted in the United States have reported risk of anaphylaxis to IV iron ranging from 2.0 to 6.8 per 10 000 first treatments. METHODS: Cohort study of IV iron new users, captured mostly through pharmacy ambulatory dispensing, from populations covered by health and administrative data sources in five European countries from 1999 to 2017. Anaphylaxis events were identified through an algorithm that used parenteral penicillin as a positive control. RESULTS: A total of 304 210 patients with a first IV iron treatment (6367 iron dextran), among whom 13-16 anaphylaxis cases were identified and reported as a range to comply with data protection regulations. The pooled unadjusted incidence proportion (IP) ranged from 0.4 (95% confidence interval [CI], 0.2-0.9) to 0.5 (95% CI, 0.3-1.0) per 10 000 first treatments. No events were identified at first dextran treatments. There were 231 294 first penicillin treatments with 30 potential cases of anaphylaxis (IP = 1.2; 95% CI, 0.8-1.7 per 10 000 treatments). CONCLUSION: We found an IP of anaphylaxis from 0.4 to 0.5 per 10 000 first IV iron treatments. The study captured only a fraction of IV iron treatments administered in hospitals, where most first treatments are likely to happen. Due to this limitation, the study could not exclude a differential risk of anaphylaxis between iron dextran and iron non-dextrans. The IP of anaphylaxis in users of penicillin was consistent with incidences reported in the literature.

Use of Nalmefene in Routine Practice: Results from a French Prospective Cohort Study and a National Database Analysis.

AIMS: Two complementary studies were used to assess the real-life use of nalmefene in alcohol-dependent patients and its impact on alcohol use health status. METHODS: USE-PACT was a prospective cohort study designed to evaluate the real-life effectiveness of nalmefene in the management of alcohol dependence, as assessed by total alcohol consumption (TAC) and number of heavy drinking days (HDD) at 1 year. USE-AM was a cohort study using data from the French nationwide claims database and was used to evaluate the external validity of the population in the prospective study. RESULTS: Overall, 256 of 700 new nalmefene users enrolled in the USE-PACT study had valid data at 1 year. After 1 year, patients treated with nalmefene showed a mean ± SD reduction from baseline in TAC (-41.5 ± 57.4 g/day) and number of HDD (-10.7 ± 11.7 days/4 weeks). Patients took a mean ± SD of 20.0 ± 12.0 tablets/4 weeks (median of 1 tablet/day) for the first 3 months and then reduced the dose. The proportion of patients who no longer took nalmefene gradually increased from 5% at 1 month to 52% at 1 year. The USE-AM study identified 486 patients with a first reimbursement for nalmefene in 2016; baseline characteristics confirmed external validity of the USE-PACT study. Overall, 46.3% of initial USE-AM prescriptions were made by GPs; most (91.8%) patients stopped treatment during follow-up. However, 15.2% of patients resumed treatment after stopping. CONCLUSIONS: In this analysis of French routine practice, patients with alcohol dependence treated with nalmefene showed reduced alcohol consumption, and nalmefene was generally well tolerated.

Intra-database validation of case-identifying algorithms using reconstituted electronic health records from healthcare claims data.

BACKGROUND: Diagnosis performances of case-identifying algorithms developed in healthcare database are usually assessed by comparing identified cases with an external data source. When this is not feasible, intra-database validation can present an appropriate alternative. OBJECTIVES: To illustrate through two practical examples how to perform intra-database validations of case-identifying algorithms using reconstituted Electronic Health Records (rEHRs). METHODS: Patients with 1) multiple sclerosis (MS) relapses and 2) metastatic castration-resistant prostate cancer (mCRPC) were identified in the French nationwide healthcare database (SNDS) using two case-identifying algorithms. A validation study was then conducted to estimate diagnostic performances of these algorithms through the calculation of their positive predictive value (PPV) and negative predictive value (NPV). To that end, anonymized rEHRs were generated based on the overall information captured in the SNDS over time (e.g. procedure, hospital stays, drug dispensing, medical visits) for a random selection of patients identified as cases or non-cases according to the predefined algorithms. For each disease, an independent validation committee reviewed the rEHRs of 100 cases and 100 non-cases in order to adjudicate on the status of the selected patients (true case/ true non-case), blinded with respect to the result of the corresponding algorithm. RESULTS: Algorithm for relapses identification in MS showed a 95% PPV and 100% NPV. Algorithm for mCRPC identification showed a 97% PPV and 99% NPV. CONCLUSION: The use of rEHRs to conduct an intra-database validation appears to be a valuable tool to estimate the performances of a case-identifying algorithm and assess its validity, in the absence of alternative.

Treatment resistant depression incidence and prevalence using the French nationwide claims database.

PURPOSE: To estimate annual incidence and prevalence of Treatment-Resistant Depression (TRD) in France. METHODS: We identified all adult patients (≥ 18 years) with a TRD episode between 1 January 2012 and 31 December 2014 in the EGB (Échantillon généraliste des bénéficiaires), a permanent random sample of the French nationwide claims database. After a 6-month washout period without hospitalization for depression or any antidepressants (AD), and after exclusion of psychotic or bipolar affective disorders, Parkinson's disease and dementia, a TRD episode was defined by three successive sequences of different AD over a 3-month treatment period (6 months for a sensitive analysis), with at least 3 weeks before each sequence change and a Medication Possession Ratio ≥ 80%; or by the dispensing of >two different AD together; or of an AD with a potentiator (lithium, antiepileptic drugs, antipsychotic drugs, thyroid hormones) over the same treatment period. The annual incidence rate was estimated from 2012 to 2014 and the prevalence using a Gamma parametric function based on treatment duration and a 30-year prediction. RESULTS: Between 2012 and 2014, 700 patients were identified in EGB with a TRD episode. The mean age was 47.4 years (±15.3); 52.7% were women. Annual incidence and prevalence of TRD were estimated at 5.8 and 25.8 per 10 000 patients, respectively and at 7.8 and 37.6 per 10 000 patients, respectively in the sensitivity analysis. CONCLUSION: This study provides the first population-based estimates for incidence and prevalence of TRD in France.