Prognostic models for short-term annual risk of severe complications and mortality in patients living with type 2 diabetes using a national medical claim database.

OBJECTIVE: Prognostic models in patients living with diabetes allow physicians to estimate individual risk based on medical records and biological results. Clinical risk factors are not always all available to evaluate these models so that they may be complemented with models from claims databases. The objective of this study was to develop, validate and compare models predicting the annual risk of severe complications and mortality in patients living with type 2 diabetes (T2D) from a national claims data. RESEARCH DESIGN AND METHODS: Adult patients with T2D were identified in a national medical claims database through their history of treatments or hospitalizations. Prognostic models were developed using logistic regression (LR), random forest (RF) and neural network (NN) to predict annual risk of outcome: severe cardiovascular (CV) complications, other severe T2D-related complications, and all-cause mortality. Risk factors included demographics, comorbidities, the adjusted Diabetes Severity and Comorbidity Index (aDSCI) and diabetes medications. Model performance was assessed using discrimination (C-statistics), balanced accuracy, sensibility and specificity. RESULTS: A total of 22,708 patients with T2D were identified, with mean age of 68 years and average duration of T2D of 9.7 years. Age, aDSCI, disease duration, diabetes medications and chronic cardiovascular disease were the most important predictors for all outcomes. Discrimination with C-statistic ranged from 0.715 to 0.786 for severe CV complications, from 0.670 to 0.847 for other severe complications and from 0.814 to 0.860 for all-cause mortality, with RF having consistently the highest discrimination. CONCLUSION: The proposed models reliably predict severe complications and mortality in patients with T2D, without requiring medical records or biological measures. These predictions could be used by payers to alert primary care providers and high-risk patients living with T2D.

Impact of donor ferritin testing on iron deficiency prevention and blood availability in France: A cohort simulation study.

BACKGROUND AND OBJECTIVES: Implementing a ferritin testing policy for whole blood (WB) donors may prevent iron deficiency (ID, ferritin <26 ng/mL) and anaemia, but may induce donation losses. As part of a national prevention plan in France, we aimed to estimate its impact on ID, anaemias and WB donations among donors at high risk of ID. MATERIALS AND METHODS: A micro-simulation model was developed to evaluate different scenarios compared to the current situation without ferritin testing as a reference scenario. The following scenarios were simulated: a minimum scenario with a 6-month deferral for donors with absent iron store (AIS, ferritinemia <15 ng/ml), a main scenario with additional delayed invitations for donors with ferritinemia 15-25 ng/ml and a supplementation scenario with additional iron supplementation for 50% of the donors with AIS. RESULTS: In the main scenario, 52,699 WB donations per year were estimated to be lost after 1 year (-8%), falling to 27,687 (-4.7%) after 5 years. IDs and anaemias were reduced by 13.6% and 29.3%, respectively, after 1 year. The supplementation scenario increased the number of prevented IDs and anaemias to 24.1% and 35.4%, respectively, after 1 year, and halved the number of anaemias at 5 years. The latter scenario also had the least impact on the number of donations (-3.2% after 5 years). CONCLUSION: A ferritin testing policy resulting in delayed donations for ID donors is effective in reducing IDs and anaemias, but significantly impacts the number of donations, thereby posing a self-sufficiency challenge.

Remdesivir and dexamethasone as tools to relieve hospital care systems stressed by COVID-19: A modelling study on bed resources and budget impact.

Remdesivir and dexamethasone are the only drugs providing reductions in the lengths of hospital stays for COVID-19 patients. We assessed the impacts of remdesivir on hospital-bed resources and budgets affected by the COVID-19 outbreak. A stochastic agent-based model was combined with epidemiological data available on the COVID-19 outbreak in France and data from two randomized control trials. Strategies involving treating with remdesivir only patients with low-flow oxygen and patients with low-flow and high-flow oxygen were examined. Treating all eligible low-flow oxygen patients during the entirety of the second wave would have decreased hospital-bed occupancy in conventional wards by 4% [2%; 7%] and intensive care unit (ICU)-bed occupancy by 9% [6%; 13%]. Extending remdesivir use to high-flow-oxygen patients would have amplified reductions in ICU-bed occupancy by up to 14% [18%; 11%]. A minimum remdesivir uptake of 20% was required to observe decreases in bed occupancy. Dexamethasone had effects of similar amplitude. Depending on the treatment strategy, using remdesivir would, in most cases, generate savings (up to 722€) or at least be cost neutral (an extra cost of 34€). Treating eligible patients could significantly limit the saturation of hospital capacities, particularly in ICUs. The generated savings would exceed the costs of medications.

Machine learning versus regression modelling in predicting individual healthcare costs from a representative sample of the nationwide claims database in France.

BACKGROUND: Innovative provider payment methods that avoid adverse selection and reward performance require accurate prediction of healthcare costs based on individual risk adjustment. Our objective was to compare the performances of a simple neural network (NN) and random forest (RF) to a generalized linear model (GLM) for the prediction of medical cost at the individual level. METHODS: A 1/97 representative sample of the French National Health Data Information System was used. Predictors selected were: demographic information; pre-existing conditions, Charlson comorbidity index; healthcare service use and costs. Predictive performances of each model were compared through individual-level (adjusted R-squared (adj-R2), mean absolute error (MAE) and hit ratio (HiR)), and distribution-level metrics on different sets of covariates in the general population and by pre-existing morbid condition, using a quasi-Monte Carlo design. RESULTS: We included 510,182 subjects alive on 31st December, 2015. Mean annual costs were 1894€ (standard deviation 9326€) (median 393€, IQ range 95€; 1480€), including zero-claim subjects. All models performed similarly after adjustment on demographics. RF model had better performances on other sets of covariates (pre-existing conditions, resource counts and past year costs). On full model, RF reached an adj-R2 of 47.5%, a MAE of 1338€ and a HiR of 67%, while GLM and NN had an adj-R2 of 34.7% and 31.6%, a MAE of 1635€ and 1660€, and a HiR of 58% and 55 M, respectively. RF model outperformed GLM and NN for most conditions and for high-cost subjects. CONCLUSIONS: RF should be preferred when the objective is to best predict medical costs. When the objective is to understand the contribution of predictors, GLM was well suited with demographics, conditions and base year cost.

Evaluation of the PRADO home discharge support programme for orthopaedic surgery in 2018.

BACKGROUND: Between 2015 and 2017, nearly 80,000 hospital stays in orthopaedic surgery were entered into a home discharge support programme (PRADO) offered by the statutory health insurance system. The objective of this study was to assess the impact of the PRADO programme on enrolled stays in orthopaedic surgery over the last three years. HYPOTHESIS: The home discharge support programme used in orthopaedic surgery shortens hospital stays and decreases the rate of readmission within 30days. MATERIAL AND METHODS: The home discharge support programme PRADO was evaluated both quantitatively and qualitatively. The quantitative study used a multicentre retrospective cohort design with matching to controls identified in the national healthcare database. All hospital stays entered into the home discharge support programme between January 2015 and December 2017 were enrolled in the study. Follow-up was 6months after discharge. The main outcome measure was the rate of readmission within 30days after discharge. The secondary outcome measures were emergency department visits, admission to rehabilitation, mean stay duration, visits to recommended healthcare professionals, medication consumption, and total healthcare expenditure at 6months. The statistical analysis used the per protocol approach. The qualitative study involved semi-structured individual and group interviews designed to investigate adhesion of the professionals and their perceptions of programme implementation, funding, and costs. RESULTS: Of 82,202 stays in the programme, 71,761 (87%) were matched and included in the analysis. Characteristics were comparable between the programme stays and the control stays. The programme stays had a significant reduction in the number of all-cause ambulatory and non-ambulatory readmissions (4.5% vs. 4.9%, p<0.0001 and 3.9% vs. 4.2%, p=0.0009, respectively). Emergency department visits and rehabilitation admissions within 30days were significantly less common in the programme group than in the control group (mean values, 2.1% vs. 2.3%, p=0.01 and 3.4% vs. 8.4%, p≤0.0001, respectively). Mean stay length was not significantly different between the two groups. Visits to recommended healthcare professionals occurred significantly more often and earlier in the programme group. The delivery of analgesics and heparin was significantly higher in the programme group, whereas no difference occurred in the delivery of antibiotics. Mean total health expenditures at 6months were lower in the programme group (2248 € vs. 2485 €). The success of the PRADO programme was dependent on leadership from the medical staff within the institution and on assistance provided by the hospital throughout its implementation. The criteria for patient eligibility to the programme were not routinely shared by or clear to the healthcare staff. DISCUSSION: The PRADO programme effectively improves the care of orthopaedic surgery patients and raises the issue of whether some admissions to rehabilitation may be unnecessary. LEVEL OF EVIDENCE: III; comparative retrospective study.

Epidemiology and disease burden of sickle cell disease in France: A descriptive study based on a French nationwide claim database.

CONTEXT: Sickle cell disease (SCD) is a severe hematological disorder. The most common acute complication of SCD is vaso-occlusive crisis (VOC), but SCD is a systemic disease potentially involving all organs. SCD prevalence estimates rely mostly on extrapolations from incidence-based newborn screening programs, although recent improvements in survival may have led to an increase in prevalence, and immigration could account for a substantial number of prevalent patients in Europe. The primary objective of this study was to estimate SCD prevalence in France. METHODS: A cross-sectional observational study was conducted using a representative sample of national health insurance data. SCD patients followed up in France between 2006 and 2011 were captured through hydroxyurea reimbursement and with the International Classification of Diseases (ICD-10) SCD specific code D570.1.2, excluding code D573 (which corresponds to sickle cell trait (SCT)). Nevertheless, we assumed that ICD-10 diagnosis coding for inpatient stays could be imperfect, with the possibility of SCT being miscoded as SCD. Therefore, prevalence was analyzed in two groups of patients [with at least one (G1) or two (G2) inpatient stay] based on the number of SCD-related inpatient stays in the six-year study period, assuming that SCT patients are rarely rehospitalized compared to SCD. The prevalence of SCD in the sample, which was considered to be representative of the French population, was then extrapolated to the general population. The rate of vaso-occlusive crisis (VOC) events was estimated based on hospitalizations, emergencies, opioid reimbursements, transfusions, and sick leave. RESULTS: Based on the number of patients identified for G1 and G2, the 2016 French prevalence was estimated to be between 48.6 per 100,000 (G1) or 32,400 patients and 29.7 per 100,000 (G2) or 19,800 patients. An average of 1.51 VOC events per year were identified, with an increase frequency of 15 to 24 years of age. The average annual number of hospitalizations was between 0.70 (G1) and 1.11 (G2) per patient. Intensive care was observed in 7.6% of VOC-related hospitalizations. Fewer than 34% of SCD patients in our sample received hydroxyurea at any point in their follow-up. The annual average cost of SCD care is €5,528.70 (G1) to €6,643.80 (G2), with most costs arising from hospitalization and lab testing. CONCLUSION: Our study estimates SCD prevalence in France at between 19,800 and 32,400 patients in 2016, higher than previously published. This study highlights the significant disease burden associated with vaso-occlusive events.

Cost-effectiveness analysis of alternative colorectal cancer screening strategies in high-risk individuals.

BACKGROUND AND AIMS: Current guidelines recommend colonoscopy every 3-5 years for colorectal cancer (CRC) screening of individuals with a familial history of CRC. The objective of this study was to compare the cost effectiveness of screening alternatives in this population. METHODS: Eight screening strategies were compared with no screening: fecal immunochemical test (FIT), Stool DNA and blood-based screening every 2 years, colonoscopy, computed tomography colonography, colon capsules, and sigmoidoscopy every 5 years, and colonoscopy at 45 years followed, if negative, by FIT every 2 years. Screening test and procedures performance were obtained from the literature. A microsimulation model reproducing the natural history of CRC was used to estimate the cost (€2018) and effectiveness [quality-adjusted life-years (QALYs)] of each strategy. A lifetime horizon was used. Costs and effectiveness were discounted at 3.5% annually. RESULTS: Compared with no screening, colonoscopy and sigmoidoscopy at a 30% uptake were the most effective strategy (46.3 and 43.9 QALY/1000). FIT at a 30 µg/g threshold with 30% uptake was only half as effective (25.7 QALY). Colonoscopy was associated with a cost of €484,000 per 1000 individuals whereas sigmoidoscopy and FIT were associated with much lower costs (€123,610 and €66,860). Incremental cost-effectiveness rate for FIT and sigmoidoscopy were €2600/QALY (versus no screening) and €3100/QALY (versus FIT), respectively, whereas it was €150,000/QALY for colonoscopy (versus sigmoidoscopy). With a lower threshold (10 µg/g) and a higher uptake of 45%, FIT was more effective and less costly than colonoscopy at a 30% uptake and was associated with an incremental cost-effectiveness ratio (ICER) of €4240/QALY versus no screening. CONCLUSION: At 30% uptake, current screening is the most effective screening strategy for high-risk individuals but is associated with a high ICER. Sigmoidoscopy and FIT at lower thresholds (10 µg/g) and a higher uptake should be given consideration as cost-effective alternatives. PLAIN LANGUAGE SUMMARY: Cost-effectiveness analysis of colorectal cancer screening strategies in high-risk individuals Fecal occult blood testing with an immunochemical test (FIT) is generally considered as the most cost-effective alternative in colorectal cancer screening programs for average risk individuals without family history.Current screening guidelines for high-risk individuals with familial history recommend colonoscopy every 3-5 years.Colonoscopy every 3-5 years for individuals with familial history is the most effective strategy but is associated with a high incremental cost-effectiveness ratio.Compared with colonoscopy, if screening based on FIT is associated with a higher participation rate, it can achieve a similar effectiveness at a lower cost.

Cost-effectiveness analysis of alternative colon cancer screening strategies in the context of the French national screening program.

BACKGROUND: A nationwide colorectal cancer (CRC) screening program was set up in France from 2009 for average-risk, asymptomatic people aged 50-74 years based on an immunochemical fecal occult blood test [faecal immunochemical test (FIT)] every 2 years, followed by colonoscopy if positive. The European standard recommends a participation rate of 45% for the program to be cost-effective, yet the latest published rate in France was 34%. The objective of this study was to compare the cost effectiveness of screening alternatives taking real-world participation rates into account. METHODS: Eight screening strategies were compared, based either on a screening test (Guaiac or FIT testing, blood-based, stool DNA, computed tomography colonography, colon capsules, and sigmoidoscopy) followed by full colonoscopy if positive or direct colonoscopy. A microsimulation model was used to estimate the cost effectiveness associated with each strategy. RESULTS: Compared with no screening, FIT was associated with a 14.0 quality-adjusted life year (QALY) increase of €50,520 per 1000 individuals, giving an incremental cost-effectiveness ratio (ICER) of €3600/QALY. Only stool DNA and blood-based testing were associated with a QALY increase compared with FIT, with stool DNA weakly dominated by blood-based testing, and the latter associated with an ICER of €154,600/QALY compared with FIT. All other strategies were dominated by FIT. CONCLUSION: FIT every 2 years appears to be the most cost-effective CRC screening strategy when taking into account a real-world participation rate of 34%.

Author Correction: A stochastic agent-based model of the SARS-CoV-2 epidemic in France.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

A stochastic agent-based model of the SARS-CoV-2 epidemic in France.

Many European countries have responded to the COVID-19 pandemic by implementing nationwide protection measures and lockdowns1. However, the epidemic could rebound when such measures are relaxed, possibly leading to a requirement for a second or more, repeated lockdowns2. Here, we present results of a stochastic agent-based microsimulation model of the COVID-19 epidemic in France. We examined the potential impact of post-lockdown measures, including physical distancing, mask-wearing and shielding individuals who are the most vulnerable to severe COVID-19 infection, on cumulative disease incidence and mortality, and on intensive care unit (ICU)-bed occupancy. While lockdown is effective in containing the viral spread, once lifted, regardless of duration, it would be unlikely to prevent a rebound. Both physical distancing and mask-wearing, although effective in slowing the epidemic and in reducing mortality, would also be ineffective in ultimately preventing ICUs from becoming overwhelmed and a subsequent second lockdown. However, these measures coupled with the shielding of vulnerable people would be associated with better outcomes, including lower mortality and maintaining an adequate ICU capacity to prevent a second lockdown. Benefits would nonetheless be markedly reduced if most people do not adhere to these measures, or if they are not maintained for a sufficiently long period.

Facing the COVID-19 epidemic in NYC: a stochastic agent-based model of various intervention strategies.

Global spread of coronavirus disease 2019 (COVID-19) has created an unprecedented infectious disease crisis worldwide. Despite uncertainties about COVID-19, model-based forecasting of competing mitigation measures on its course is urgently needed to inform mitigation policy. We used a stochastic agent-based microsimulation model of the COVID-19 epidemic in New York City and evaluated the potential impact of quarantine duration (from 4 to 16 weeks), quarantine lifting type (1-step lifting for all individuals versus a 2-step lifting according to age), post-quarantine screening, and use of a hypothetical effective treatment against COVID-19 on the disease's cumulative incidence and mortality, and on ICU-bed occupancy. The source code of the model has been deposited in a public source code repository (GitHub®). The model calibrated well and variation of model parameter values had little impact on outcome estimates. While quarantine is efficient to contain the viral spread, it is unlikely to prevent a rebound of the epidemic once lifted. We projected that lifting quarantine in a single step for the full population would be unlikely to substantially lower the cumulative mortality, regardless of quarantine duration. By contrast, a two-step quarantine lifting according to age was associated with a substantially lower cumulative mortality and incidence, up to 71% and 23%, respectively, as well as lower ICU-bed occupancy. Although post-quarantine screening was associated with diminished epidemic rebound, this strategy may not prevent ICUs from being overcrowded. It may even become deleterious after a 2-step quarantine lifting according to age if the herd immunity effect does not had sufficient time to become established in the younger population when the quarantine is lifted for the older population. An effective treatment against COVID-19 would considerably reduce the consequences of the epidemic, even more so if ICU capacity is not exceeded.

Lockdown exit strategies and risk of a second epidemic peak: a stochastic agent-based model of SARS-CoV-2 epidemic in France.

Most European countries have responded to the COVID-19 threat by nationwide implementation of barrier measures and lockdown. However, assuming that population immunity will build up through the epidemic, it is likely to rebound once these measures are relaxed, possibly leading to a second or multiple repeated lockdowns. In this report, we present results of epidemiological modelling that has helped inform policy making in France. We used a stochastic agent-based microsimulation model of the COVID-19 epidemic in France, and examined the potential impact of post-quarantine measures, including social distancing, mask-wearing, and shielding of the population the most vulnerable to severe COVID-19 infection, on the disease's cumulative incidence and mortality, and on ICU-bed occupancy. The model calibrated well and variation of model parameter values had little impact on outcome estimates. While quarantine is effective in containing the viral spread, it would be unlikely to prevent a rebound of the epidemic once lifted, regardless of its duration. Both social distancing and mask-wearing, although effective in slowing the epidemic and in reducing mortality, would also be ineffective in ultimately preventing the overwhelming of ICUs and a second lockdown. However, these measures coupled with shielding of vulnerable people would be associated with better outcomes, including lower cumulative incidence, mortality, and maintaining an adequate number of ICU beds to prevent a second lockdown. Benefits would nonetheless be markedly reduced if these measures were not applied by most people or not maintained for a sufficiently long period, as herd immunity progressively establishes in the less vulnerable population.

Understanding mortality in multiple myeloma: Findings of a European retrospective chart review.

OBJECTIVES: This study aimed to provide real-world data on the characteristics and treatment of patients with multiple myeloma (MM) at the time of death. METHODS: The study was a retrospective patient chart review across France, Germany, Italy, Spain and the UK during 2016, and included patients who had died in the 3 months before the index date. RESULTS: Data from 786 patients were reviewed. At the time of death, 37% of patients were receiving active treatment, 12% were in a treatment-free interval and 51% were receiving only supportive care. Death before and during active first-line treatment was not uncommon (6% and 24% of patients, respectively) but these deaths were often not solely due to disease progression; factors such as renal failure and infection frequently played a role (in 30% and 20% of patients at first-line, respectively). Most deaths at later lines were due to progressive disease. Cox model results suggested that early deaths were associated with advanced disease stage, high-risk cytogenetics and poor response and relapse profiles. CONCLUSIONS: These real-world data could be used to help develop strategies for improving survival in patients with MM and to support management tailored to the stage of disease.

A Comprehensive Model of Predictors of Suicide Attempt in Depressed Individuals and Effect of Treatment-Seeking Behavior: Results From a National 3-Year Prospective Study.

OBJECTIVE: Multiple factors have an impact on the risk of attempting suicide in depressed individuals. The elevated number of contributing factors and their frequent co-occurrence suggest the need to combine them in integrative models to develop more effective suicide prevention strategies. This report presents a comprehensive model of the 3-year risk of suicide attempt in individuals with major depressive episode (MDE) using a longitudinal nationally representative study, the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC; Wave 1, 2001-2002; Wave 2, 2004-2005). METHODS: Structural equation modeling was used to simultaneously examine effects of 4 broad groups of clinical factors previously identified as potential predictors of suicide attempts: (1) severity of depressive illness, (2) severity of psychiatric and other physical comorbidity, (3) sociodemographic characteristics, and (4) treatment-seeking behavior. RESULTS: About 3.5% of the 2,587 participants with a DSM-IV diagnosis of MDE attempted suicide during the 3-year follow-up period. Several factors predicted attempted suicide independently of each other: the absence of treatment-seeking behavior for MDE, the severity of the depressive illness (ie, recurrent thoughts of death, prior suicide attempts, and severity of the general depressive symptom dimension representing the joint effect of most depressive symptoms), and the severity of comorbidities (ie, severity of the general psychopathology factor representing the mechanisms shared across all comorbid psychiatric disorders and lower mental health-related quality of life). No sociodemographic characteristics independently contributed to this risk. CONCLUSIONS: This model may help identify high-risk individuals with MDE and inform future research on risk of suicide.

Healthcare resource utilization among patients with relapsed multiple myeloma in the UK, France, and Italy.

AIMS: To assess the real-world healthcare resource utilization (HRU) and costs associated with different treatment regimens used in the management of patients with relapsed multiple myeloma in the UK, France, and Italy. METHODS: Retrospective medical chart review of characteristics, time to progression, level of response, HRU during treatment, and adverse events (AEs). Data collection started on June 1, 2015 and was completed on July 15, 2015. In the 3 months before record abstraction, eligible patients had either disease progression after receiving one of their country's most commonly prescribed regimens or had received the best supportive care and died. Costs were calculated based on HRU and country-specific diagnosis-related group and/or unit reference costs, amongst other standard resources. RESULTS: Physicians provided data for 1,282 patients (387 in the UK, 502 in France, 393 in Italy) who met the inclusion criteria. Mean [median] total healthcare costs associated with a single line of treatment were €51,717 [35,951] in the UK, €37,009 [32,538] for France, and €34,496 [42,342] for Italy, driven largely by anti-myeloma medications costs (contributing 95.0%, 90.0%, and 94.2% of total cost, respectively). During active treatment, the highest costs were associated with lenalidomide- and pomalidomide-based regimens. Mean cost per month was lowest for patients achieving a very good partial response or better. Unscheduled events (i.e. not considered part of routine management, whether or not related to multiple myeloma, such as unscheduled hospitalization, AEs, fractures) accounted for 1-9% of total costs and were highest for bendamustine. LIMITATIONS: The use of retrospective data means that clinical practice (e.g. use of medical procedures, evaluation of treatment response) is not standardized across participating countries/centers, and some data (e.g. low-grade AEs) may be incomplete or differently adjudicated/reported. The centers involved may not be fully representative of national practice. CONCLUSIONS: Drug costs are the main contributor to total HRU costs associated with multiple myeloma. The duration of active treatment may influence the average total costs, as well as response, associated with a single line of therapy. Improved treatment outcomes, and reductions in unscheduled events and concomitant medication use may, therefore, reduce the overall HRU and related costs of care in multiple myeloma.

Organised screening for cervical cancer in France: a cost-effectiveness assessment.

OBJECTIVE: According to the third cancer plan, organised screening (OS) of cervical cancer (CC) among women aged 25-65 years should be implemented in France in the forthcoming years. The most efficient way to implement OS in the French healthcare system is yet to be determined. METHODS: A microsimulation model was developed adopting a collective 'all payers' perspective. A closed cohort of women eligible for CC screening and representative in terms of age and participation in individual screening (IndScr) by annual Papanicolaou (Pap) testing every 3 years was modelled on a lifetime horizon. Different OS strategies, additive to IndScr with a 61.9% participation rate based on mailed invitations to non-participant women to perform OS were assessed. Similar modalities were applied to OS and IndScr participants. Strategies implied different screening tests (Papanicolaou (Pap) test, human papillomavirus (HPV) test and p16/Ki67 double staining) and OS periodicity. RESULTS: Compared with IndScr only, all OS strategies were associated with decreased cancer incidence/mortality (from 14.2%/13.5% to 22.9%/25.8%). Most strategies generated extra costs ranging from €37.9 to €1607 per eligible woman. HPV testing every 10 and 5 years were cost saving. HPV tests every 10 and 5 years were the most efficient strategies, generating more survival at lower costs than Pap-based strategies. Compared to IndScr only, an HPV test every 10 years was cost saving. The most effective strategies were p16/Ki67 as primary or HPV positive confirmation tests, with respective incremental cost-effectiveness ratios of €6 541 250 and €101 391 per life year. Pap-based strategies generated intermediary results. CONCLUSION: OS strategies based on the HPV test appear highly efficient. However, our results rely on the assumption that women and practitioners comply with the recommended OS periodicities (3, 5, 10 years). Implementing these OS modalities will require major adaptations to the current CC screening organisation. Pap test-based strategies might be simpler to setup while preparing an appropriate implementation of more efficient OS screening modalities.

Factors associated with the length of stay of patients discharged from emergency department in France.

OBJECTIVES: The length of stay in the emergency department (ED) has been proposed as an indicator of performance in many countries. We conducted a survey of length of stay in two large areas in France and tested the hypothesis that patient and ED-related variables may influence it. PATIENTS AND METHODS: During 2007, we examined lengths of stay in ambulatory patients, that is, excluding admitted patients. The following variables were considered: (a) at the patient level, age, sex, the day and month of the visit, and the French clinical classification of emergency patients (CCEP) class; (b) at the ED level, annual ED total number of visits, mean age, the proportions of patients less than 15 and more than 75 years, and the proportions of admitted and clinically stable patients with CCEP class 1 and 2. A multilevel hierarchical analysis was carried out. RESULTS: We analyzed 988 591 visits in 53 EDs. The ED-specific median length of stay was 98 (IQR: 62-137) min and the ED-specific median proportion of patients with length of stay of more than 4 h was 15 (5-24) %. There was a strong correlation between the ED-specific median length of stay and the ED-specific proportion of patients with a length of stay of more than 4 h (R=0.96, P<0.001). Using a multilevel analysis, only three variables were associated significantly with the length of stay: the age and the CCEP class of the patient, and the ED census. CONCLUSION: We observed that the length of stay in the ED needs to be stratified by case mix and the total number of visits of the ED.

Main barriers to effective implementation of stroke care pathways in France: a qualitative study.

BACKGROUND: Stroke Care Pathways (SCPs) aim to improve quality of care by providing better access to stroke units, rehabilitation centres, and home care for dependent patients. The objective of this study was to identify the main barriers to effective implementation of SCPs in France. METHODS: We selected 4 types of SCPs currently implemented in France that differed in terms of geographical location, population size, socio-economic conditions, and available health care facilities. We carried out 51 semi-structured interviews of 44 key health professionals involved in these SCPs and used the interview data to (i) create a typology of the organisational barriers to effective SCP implementation by axial coding, (ii) define barrier contents by vertical coding. The typology was validated by a panel of 15 stroke care professionals. RESULTS: Four main barriers to effective SCP implementation were identified: lack of resources (31/44 interviewees), coordination problems among staff (14/44) and among facilities (27/44), suboptimal professional and organisational practices (16/44), and inadequate public education about stroke (13/44). Transposition of the findings onto a generic SCP highlighted alternative care options and identified 10 to 17 barriers that could disrupt continuity of care. CONCLUSION: Lack of resources was considered to be the chief obstacle to effective SCP implementation. However, the main weakness of existing SCPs was poor communication and cooperation among health professionals and among facilities. We intend to use this knowledge to construct a robust set of quality indicators for use in SCP quality improvement initiatives, in comparisons between SCPs, and in the assessment of the effective implementation of clinical practice guidelines.

Hospital performance based on treatment delays: comparison of ranking methods.

BACKGROUND: Reducing time-to-care is crucial in many acute and chronic diseases. Quality indicators based on target delays derived from guidelines are used to compare hospital performance but there is no accepted methodology for comparing performance when no target delay has been established. AIM: To explore by different statistical methods the uncertainty in hospital comparisons that are based on delay indicators, when no target delay is available. METHODS: Data for hospital door-to-needle time were extracted from a 2010 study of 1699 patients in 57 hospitals with ST-elevated myocardial infarction. We determined whether the times in each hospital were statistically different from the overall mean time or the median time for all hospitals by (i) one-way analysis of variance (ANOVA), (ii) non-parametric ANOVA with Nelson-Hsu adjustment (ANOVA R) and (iii) the proportional hazard model (PHM). We also tested for the assumptions underlying the methods: normal distribution for ANOVA, homogeneity of variances (homoscedasticity) for ANOVA and ANOVA R, and proportionality for PHM. RESULTS: Door-to-needle times were available for 889 patients in 44 hospitals. Data distribution was not Gaussian. Test assumptions were verified for ANOVA R (homoscedasticity) for one data subset (>48-h times (48H) excluded) and for PHM (proportionality) for two data subsets (48H or >95th percentile (P95) times excluded). The same five significantly better performers were identified in each case (although ANOVA R missed one). ANOVA R (48H) identified two significantly poorer performers, PHM (48H) identified three and PHM (P95) just one. Poorer performers differed according to method. CONCLUSIONS: The tested statistical methods yielded broadly similar results but no method was truly satisfactory. A transparency statement should therefore always specify the ranking method used to compare hospital performance.