How to prepare a systematic review of economic evaluations for informing evidence-based healthcare decisions: data extraction, risk of bias, and transferability (part 3/3)

INTRODUCTION

This article is part of the series "How to Prepare a Systematic Review (SR) of Economic Evaluations (EE) for Informing Evidence-based Healthcare Decisions" in which a five-step-approach for conducting a SR of EE is proposed. Areas covered: This paper explains the data extraction process, the risk of bias assessment and the transferability of EEs by means of a narrative review and expert opinion. SRs play a critical role in determining the comparative cost-effectiveness of healthcare interventions. It is important to determine the risk of bias and the transferability of an EE. Expert commentary: Over the past decade, several criteria lists have been developed. This article aims to provide recommendations on these criteria lists based on the thoroughness of development, feasibility, overall quality, recommendations of leading organizations, and widespread use.

Which Questionnaire Should Be Used to Measure Quality-of-Life Utilities in Patients with Acute Leukemia? An Evaluation of the Validity and Interpretability of the EQ-5D-5L and Preference-Based Questionnaires Derived from the EORTC QLQ-C30

OBJECTIVES

The aim of this study was to assess the validity and interpretability of different preference-based questionnaires (generic 5-level EuroQol five-dimensional questionnaire [EQ-5D-5L], cancer-specific Quality of Life Questionnaire Preference-Based Measure, and European Organization of Randomized Controlled Trials 8 Dimension [EORTC-8D]) in patients with acute leukemia.

METHODS

Patients who participated in Hemato-Oncologie voor Volwassenen Nederland (HOVON - the Haemato Oncology Foundation for Adults in the Netherlands) clinical trials between 1999 and 2011 at a single hospital were invited to complete the questionnaires. Interpretability was evaluated by the frequency of incomplete data and highest and lowest possible scores. Content validity was evaluated by exploring the health-related quality-of-life domains included in the questionnaires. Construct validity was assessed using correlations with other quality-of-life scales (EQ-visual analogue scale score and global quality-of-life scale of the EORTC Quality of Life Questionnaire) and ability to distinguish between patients with different health statuses.

RESULTS

Questionnaires were returned by 89% (111 of 125) of the patients. Six to seven respondents did not return full questionnaires. Perfect health on the EQ-5D-5L was reported by 32 respondents and many of them (N = 17) did report health problems on other questionnaires. All questionnaires were strongly correlated (range 0.61-0.78) with other quality-of-life scales and yielded substantially different utility values for patients with different health statuses. Nevertheless, the disease-specific preference-based questionnaires showed greater discriminatory power.

CONCLUSIONS

Although the Quality of Life Questionnaire Preference-Based Measure and the EORTC-8D appear to have better validity, this study does not provide any strong evidence against the use of the EQ-5D-5L for measuring quality-of-life utilities in acute leukemia. However, our findings need to be confirmed in larger longitudinal studies.

Variation in use of targeted therapies for metastatic renal cell carcinoma: Results from a Dutch population-based registry

BACKGROUND

For patients with metastatic renal cell carcinoma (mRCC), targeted therapies have entered the market since 2006. The aims of this study were to evaluate the uptake and use of targeted therapies for mRCC in The Netherlands, examine factors associated with the prescription of targeted therapies in daily clinical practice and study their effectiveness in terms of overall survival (OS).

METHODS

Two cohorts from PERCEPTION, a population-based registry of mRCC patients, were used: a 2008-2010 Cohort (n = 645) and a 2011-2013 Cohort (n = 233). Chi-squared tests for trend were used to study time trends in the use of targeted therapy. Patients were grouped based on the eligibility criteria of the SUTENT trial, the trial that led to sunitinib becoming standard of care, to investigate the use of targeted therapies amongst patients fulfilling those criteria. Multi-level logistic regression was used to identify patient subgroups that are less likely to receive targeted therapies.

RESULTS

Approximately one-third of patients fulfilling SUTENT trial eligibility criteria did not receive any targeted therapy (29 % in the 2008-2010 Cohort; 35 % in the 2011-2013 Cohort). Patients aged 65+ years were less likely to receive targeted therapy in both cohorts and different risk groups (odds ratios range between 0.84-0.92); other factors like number of metastatic sites were of influence in some subgroups. Amongst treated patients, there was a decreasing trend in sunitinib use over time (p = 0.0061), and an increasing trend in pazopanib use (p = 0.0005).

CONCLUSIONS

Targeted therapies have largely replaced interferon-alfa as first-line standard of care. Nevertheless, many eligible patients in Dutch daily practice did not receive targeted therapies despite their ability to improve survival. Reasons for their apparent underutilisation should be examined more carefully.

Individualized Angiotensin-Converting Enzyme (ACE)-Inhibitor Therapy in Stable Coronary Artery Disease Based on Clinical and Pharmacogenetic Determinants: The PERindopril GENEtic (PERGENE) Risk Model

Background

Patients with stable coronary artery disease (CAD) constitute a heterogeneous group in which the treatment benefits by angiotensin‐converting enzyme (ACE)‐inhibitor therapy vary between individuals. Our objective was to integrate clinical and pharmacogenetic determinants in an ultimate combined risk prediction model.

Methods and Results

Clinical, genetic, and outcomes data were used from 8726 stable CAD patients participating in the EUROPA/PERGENE trial of perindopril versus placebo. Multivariable analysis of phenotype data resulted in a clinical risk score (range, 0–21 points). Three single‐nucleotide polymorphisms (rs275651 and rs5182 in the angiotensin‐II type I‐receptor gene and rs12050217 in the bradykinin type I‐receptor gene) were used to construct a pharmacogenetic risk score (PGXscore; range, 0–6 points). Seven hundred eighty‐five patients (9.0%) experienced the primary endpoint of cardiovascular mortality, nonfatal myocardial infarction or resuscitated cardiac arrest, during 4.2 years of follow‐up. Absolute risk reductions ranged from 1.2% to 7.5% in the 73.5% of patients with PGXscore of 0 to 2. As a consequence, estimated annual numbers needed to treat ranged from as low as 29 (clinical risk score ≥10 and PGXscore of 0) to 521 (clinical risk score ≤6 and PGXscore of 2). Furthermore, our data suggest that long‐term perindopril prescription in patients with a PGXscore of 0 to 2 is cost‐effective.

Conclusions

Both baseline clinical phenotype, as well as genotype determine the efficacy of widely prescribed ACE inhibition in stable CAD. Integration of clinical and pharmacogenetic determinants in a combined risk prediction model demonstrated a very wide range of gradients of absolute treatment benefit.

Using meta-regression analyses in addition to conventional systematic review methods to examine the variation in cost-effectiveness results - a case study

Background

Systematic reviews of cost-effectiveness analyses summarize results and describe study characteristics. Variability in the study results is often explained qualitatively or based on sensitivity analyses of individual studies. However, variability due to input parameters and study characteristics (e.g., funding or study quality) is often not statistically explained. As a case study, a systematic review on the cost-effectiveness of drug-eluting stents (DES) versus bare-metal stents (BMS) using meta-regression analyses is performed to explore the usefulness of such methods compared with conventional review methods.

Methods

We attempted to identify and review all modelling studies published until January 2012 that compared costs and consequences of DES versus BMS. We extracted general study information (e.g., funding), modelling methods, values of input parameters, and quality of the model using the Philips et al. checklist. Associations between study characteristics and the incremental costs and effectiveness of individual analyses were explored using regression analyses corrected for study ID.

Results

Sixteen eligible studies were identified, with a combined total of 508 analyses. The overall quality of the models was moderate (59 % ± 15 %). This study showed associations (e.g., type of lesion) that were expected (based on individual studies), however the meta-regression analyses revealed also unpredicted associations: e.g., model quality was negatively associated with repeat revascularizations avoided.

Conclusions

Meta-regressions can be of added value, identifying significant associations that could not be identified using conventional review methods or by sensitivity analyses of individual studies. Furthermore, this study underlines the need to examine input parameters and perform a quality check of studies when interpreting the results.

Electronic supplementary material

The online version of this article (doi:10.1186/s12913-015-1230-4) contains supplementary material, which is available to authorized users.

Treatment variation in stent choice in patients with stable or unstable coronary artery disease

AIM

Variations in treatment are the result of differences in demographic and clinical factors (e.g. anatomy), but physician and hospital factors may also contribute to treatment variation. The choice of treatment is considered important since it could lead to differences in long-term outcomes. This study explores the associations with stent choice: i.e. drug-eluting stent (DES) versus bare-metal stents (BMS) for Dutch patients diagnosed with stable or unstable coronary artery disease (CAD).

METHODS & RESULTS

Associations with treatment decisions were based on a prospective cohort of 692 patients with stable or unstable CAD. Of those patients, 442 patients were treated with BMS or DES. Multiple logistic regression analyses were performed to identify variables associated with stent choice. Bivariate analyses showed that NYHA class, number of diseased vessels, previous percutaneous coronary intervention, smoking, diabetes, and the treating hospital were associated with stent type. After correcting for other associations the treating hospital remained significantly associated with stent type in the stable CAD population.

CONCLUSIONS

This study showed that several factors were associated with stent choice. While patients generally appear to receive the most optimal stent given their clinical characteristics, stent choice seems partially determined by the treating hospital, which may lead to differences in long-term outcomes.

A conceptual disease model for adult Pompe disease

Background

Studies in orphan diseases are, by nature, confronted with small patient populations, meaning that randomized controlled trials will have limited statistical power. In order to estimate the effectiveness of treatments in orphan diseases and extrapolate effects into the future, alternative models might be needed. The purpose of this study is to develop a conceptual disease model for Pompe disease in adults (an orphan disease). This conceptual model describes the associations between the most important levels of health concepts for Pompe disease in adults, from biological parameters via physiological parameters, symptoms and functional indicators to health perceptions and final health outcomes as measured in terms of health-related quality of life.

Methods

The structure of the Wilson-Cleary health outcomes model was used as a blueprint, and filled with clinically relevant aspects for Pompe disease based on literature and expert opinion. Multiple observations per patient from a Dutch cohort study in untreated patients were used to quantify the relationships between the different levels of health concepts in the model by means of regression analyses.

Results

Enzyme activity, muscle strength, respiratory function, fatigue, level of handicap, general health perceptions, mental and physical component scales and utility described the different levels of health concepts in the Wilson-Cleary model for Pompe disease. Regression analyses showed that functional status was affected by fatigue, muscle strength and respiratory function. Health perceptions were affected by handicap. In turn, self-reported quality of life was affected by health perceptions.

Conclusions

We conceptualized a disease model that incorporated the mechanisms believed to be responsible for impaired quality of life in Pompe disease. The model provides a comprehensive overview of various aspects of Pompe disease in adults, which can be useful for both clinicians and policymakers to support their multi-faceted decision making.

Electronic supplementary material

The online version of this article (doi:10.1186/s13023-015-0334-6) contains supplementary material, which is available to authorized users.

Which factors may determine the necessary and feasible type of effectiveness evidence? A mixed methods approach to develop an instrument to help coverage decision-makers

OBJECTIVES

Reimbursement decisions require evidence of effectiveness and, in general, a blinded randomised controlled trial (RCT) is the preferred study design to provide it. However, there are situations where a cohort study, or even patient series, can be deemed acceptable. The aim of this study was to develop an instrument that first examines which study characteristics of a blinded RCT are necessary, and then, if particular characteristics are considered necessary, examines whether these characteristics are feasible.

DESIGN

We retrospectively studied 22 interventions from 20 reimbursement reports concerning medical specialist care made by the Dutch National Health Care Institute (ZIN) to identify any factors that influenced the necessity and feasibility of blinded RCTs, and their constituent study characteristics, that is, blinding, randomisation and a control group. A literature review was performed to identify additional factors. Additional expertise was included by interviewing eight experts in epidemiology, medicine and ethics. The resulting instrument was called the FIT instrument (Feasible Information Trajectory), and was prospectively validated using three consecutive reimbursement reports.

RESULTS

(Blinded) RCT evidence was lacking in 5 of 11 positive reimbursement decisions and 3 of 11 negative decisions. In the reimbursement reports, we found no empirical evidence supporting situations where a blinded RCT is unnecessary. The literature also revealed few arguments against the necessity of a blinded RCT. In contrast, many factors influencing the feasibility of randomisation, a control group and blinding, were found in the reimbursement reports and the literature; for example, when a patient population is too small or when an intervention is common practice, randomisation will be hindered.

CONCLUSIONS

Policy regarding the necessity and feasibility of different types of evidence of effectiveness would benefit from systematic guidance. The FIT instrument has the potential to support transparent, reproducible and well-founded decisions on appropriate evidence of effectiveness in medical specialist care.

Access to orphan drugs in western Europe: can more systematic policymaking really help to avoid different decisions about the same drug?

When the Scottish government recently launched a £21 million fund to provide patients with rare diseases access to treatment, Alex Neill, the Scottish Health Secretary, stated that 'it was only right that Scottish patients with rare conditions had access to innovative medicines which were clinically justified, and that they were not disadvantaged due to the very high cost of these treatments'. This statement raises questions about the criteria that play a role in reimbursement decisions on orphan drugs. This editorial examines the criteria that were used in the decisions about the reimbursement of an orphan drug for Pompe disease and explores methods to improve the transparency and consistency of reimbursement decisions for orphan drugs in general.

Beliefs about medicines in Dutch acenocoumarol and phenprocoumon users

AIMS

Adherence to the generally complex regimen of coumarin derivatives is vital in order to keep patients in the adequate International Normalized Ratio range. Patients' beliefs about medicines are associated with the level of therapy adherence. Our first aim was to assess beliefs about coumarins. Secondly, we compared the beliefs about coumarins with the beliefs about other cardiovascular drugs.

METHODS

The Beliefs about Medicines Questionnaire was used to assess medication beliefs. The questionnaire was completed by new users of coumarins indicated for venous thromboembolism or atrial fibrillation. A necessity score and a concerns score were calculated for all patients. The analyses were repeated for users of antihypertensive drugs or statins (not using coumarins).

RESULTS

Three hundred and twenty patients were included in the analysis of the beliefs about coumarins. The mean necessity score was 15.3, the concerns score 12.3 and the necessity-concerns differential 3.0. Patients with venous thromboembolism (n = 71) had higher necessity scores than patients with atrial fibrillation (n = 249; 16.8 vs. 14.9, P < 0.001). The mean necessity score in 493 users of other cardiovascular drugs was 16.1, the concerns score 13.5 and the necessity-concerns differential 2.6. The necessity score was higher in chronic cardiovascular drug users (n = 192) than in new users (n = 301; 17.9 vs. 14.9, P < 0.001).

CONCLUSIONS

Coumarin users score higher on the necessity scale than on the concerns scale, which is also the case in users of other cardiovascular drugs. Patients with atrial fibrillation have a less positive attitude towards these drugs than patients with venous thromboembolism, and could therefore benefit more from specific attention.

Patient perspective on remote monitoring of cardiovascular implantable electronic devices: rationale and design of the REMOTE-CIED study

Background

Remote patient monitoring is a safe and effective alternative for the in-clinic follow-up of patients with cardiovascular implantable electronic devices (CIEDs). However, evidence on the patient perspective on remote monitoring is scarce and inconsistent.

Objectives

The primary objective of the REMOTE-CIED study is to evaluate the influence of remote patient monitoring versus in-clinic follow-up on patient-reported outcomes. Secondary objectives are to: 1) identify subgroups of patients who may not be satisfied with remote monitoring; and 2) investigate the cost-effectiveness of remote monitoring.

Methods

The REMOTE-CIED study is an international randomised controlled study that will include 900 consecutive heart failure patients implanted with an implantable cardioverter defibrillator (ICD) compatible with the Boston Scientific LATITUDE® Remote Patient Management system at participating centres in five European countries. Patients will be randomised to remote monitoring or in-clinic follow-up. The In-Clinic group will visit the outpatient clinic every 3–6 months, according to standard practice. The Remote Monitoring group only visits the outpatient clinic at 12 and 24 months post-implantation, other check-ups are performed remotely. Patients are asked to complete questionnaires at five time points during the 2-year follow-up.

Conclusion

The REMOTE-CIED study will provide insight into the patient perspective on remote monitoring in ICD patients, which could help to support patient-centred care in the future.

Is it cost-effective to use a test to decide which individuals with an intermediate cardiovascular disease risk would benefit from statin treatment?

BACKGROUND

The 2012 European guidelines recommend statins for intermediate-risk individuals with elevated cholesterol levels. Improved discrimination of intermediate-risk individuals is needed to prevent both cardiovascular disease (CVD) and statin side-effects (e.g. myopathy) efficiently since only 3-15 in every 100 individuals actually experience a cardiovascular event in the next 10 years. We estimated the potential cost-effectiveness of a hypothetical test which helps to determine which individuals will benefit from statins.

METHODS AND RESULTS

Prognosis of different age- and gender-specific cohorts with an intermediate risk was simulated with a Markov model to estimate the potential costs and quality-adjusted life-years for four strategies: treat all with statins, treat none with statins, treat according to the European guidelines, or use a test to select individuals for statin treatment. The test-first strategy dominated the other strategies if the hypothetical test was 100% accurate and cost no more than €237. This strategy and the treat-all strategy were equally effective but the test generated lower costs by reducing statin usage and side-effects. The treat-none strategy was the least effective strategy. Threshold analyses show that the test must be highly accurate (especially sensitive) and inexpensive to be the most cost-effective strategy, since myopathy has a negligible impact on cost-effectiveness and statin costs are low.

CONCLUSION

Use of a highly accurate prognostic test could reduce overall CVD risk, frequency of drug side-effects and lifetime costs. However, no additional test would add usefully to risk prediction over SCORE when it does not satisfy the costs and accuracy requirements.

Cost-effectiveness of enzyme replacement therapy with alglucosidase alfa in classic-infantile patients with Pompe disease

BACKGROUND

Infantile Pompe disease is a rare metabolic disease. Patients generally do not survive the first year of life. Enzyme replacement therapy (ERT) has proven to have substantial effects on survival in infantile Pompe disease. However, the costs of therapy are very high. In this paper, we assess the cost-effectiveness of enzyme replacement therapy in infantile Pompe disease.

METHODS

A patient simulation model was used to compare costs and effects of ERT with costs of effects of supportive therapy (ST). The model was filled with data on survival, quality of life and costs. For both arms of the model, data on survival were obtained from international literature. In addition, survival as observed among 20 classic-infantile Dutch patients, who all received ERT, was used. Quality of life was measured using the EQ-5D and assumed to be the same in both treatment groups. Costs included the costs of ERT (which depend on a child's weight), infusions, costs of other health care utilization, and informal care. A lifetime time horizon was used, with 6-month time cycles.

RESULTS

Life expectancy was significantly longer in the ERT group than in the ST group. On average, ST receiving patients were modelled not to survive the first half year of life; whereas the life expectancy in the ERT patients was modelled to be almost 14 years. Lifetime incremental QALYs were 6.8. Incremental costs were estimated to be € 7.0 million, which primarily consisted of treatment costs (95%). The incremental costs per QALY were estimated to be € 1.0 million (range sensitivity analyses: € 0.3 million - € 1.3 million). The incremental cost per life year gained was estimated to be € 0.5 million.

CONCLUSIONS

The incremental costs per QALY ratio is far above the conventional threshold values. Results from univariate and probabilistic sensitivity analyses showed the robustness of the results.

Explaining length of stay variation of episodes of care in the Netherlands

OBJECTIVES

Diagnosis Related Group (DRG) systems aim to classify patients into mutually exclusive groups of patients, with the patients in each group having the same expected length of stay (LOS). We examined the ability of current classification variables to explain LOS variation between DRG-like Diagnosis Treatment Combination (DBC)s for ten episodes of care in the Netherlands, including breast cancer, stroke and inguinal hernia repair. Additionally, we assessed the predictive ability of some other classification variables.

METHODS

For each episode of care, the relevant DBC codes of all hospitalizations in 2008 were identified and all available determinants that may serve as classification variables were acquired from the national database. Ordinary least squares regression was used to examine the predictive ability of these classification variables.

RESULTS

The current classification variables are not sufficiently distinct to classify patients into mutually exclusive groups of patients. ICU admissions and hospital type may serve as valuable classification variables. Additionally, episode-specific variables may improve the Dutch grouping algorithm.

CONCLUSIONS

Although it may not be feasible in the short term, grouping algorithms would benefit greatly from the introduction of classification variables tailored to the needs of specific episodes of care. A first step would be to focus on 'general' classification variables meaningful for specific episodes of care.

The faces of personalized medicine: a framework for understanding its meaning and scope

The objective of this article was to provide a framework for understanding the different definitions of the term "personalized medicine." The term personalized medicine is used regularly but interpreted in different ways. This article approaches the term by starting with a broad view of clinical medicine, where three components can be distinguished: the questions (e.g., what is the diagnosis?), the methods used to answer them (e.g., a test), and the available actions (e.g., to give or not give a particular drug). Existing definitions of personalized medicine disagree about which questions, methods, and actions fall within its domain. Some define the term narrowly, referring to the use of a diagnostic test to predict drug response, thereby clarifying whether or not a patient will benefit from that drug. An example of this combination is the HER2/neu test to predict the effectiveness of trastuzumab in breast cancer. Many who adopt this definition associate the concept of personalized medicine with fields such as genetics, genomics, and other types of "-omics." In contrast, others view personalized medicine as a concept that has always existed, because medicine has always considered the needs of the individual. One definition of personalized medicine that accommodates both interpretations is "the use of combined knowledge (genetic or otherwise) about a person to predict disease susceptibility, disease prognosis, or treatment response and thereby improve that person's health." This predictive ability can increase over time through innovations in various technologies, resulting in further improvements in health outcomes. Moreover, these developments can lead to a better understanding of the underlying causes of disease, which can eventually lead to breakthroughs in the treatment of individual patients. In that sense, a truly personalized form of medicine can also be seen as an ideal, a goal that will be achieved only after multiple advances in science. Although the term personalized medicine was rechristened somewhat recently, our ability to personalize medicine will continue to advance in unimaginable ways as we come to learn more about the heterogeneity that exists among individuals and diseases.

Systematic review of available evidence on 11 high-priced inpatient orphan drugs

Background

Attention for Evidence Based Medicine (EBM) is growing, but evidence for orphan drugs is argued to be limited and inferior. This study systematically reviews the available evidence on clinical effectiveness, cost-effectiveness and budget impact for orphan drugs.

Methods

A systematic review was performed in PubMed, Embase, NHS EED and HTA databases for 11 inpatient orphan drugs listed on the Dutch policy rule on orphan drugs. For included studies, we determined the type of study and various study characteristics.

Results

A total of 338 studies met all inclusion criteria. Almost all studies (96%) focused on clinical effectiveness of the drug. Of these studies, most studies were case studies (41%) or observational studies (39%). However, for all orphan diseases at least one experimental or quasi-experimental study was found, and a randomized clinical trial was available for 60% of the orphan drugs. Eight studies described the cost-effectiveness of an orphan drug; an equal number described an orphan drug’s budget impact.

Conclusions

Despite the often heard claim that RCTs are not feasible for orphan drugs, we found that an RCT was available in 60% of orphan drugs investigated. Cost-effectiveness and budget impact analyses for orphan drugs are seldom published.

Long-term outcomes of continuous intrathecal baclofen infusion for treatment of spasticity: a prospective multicenter follow-up study

Long-term outcomes of 115 patients treated with continuous intrathecal baclofen infusion are reported. A prospective follow-up study was conducted in eight centers. Patients were followed up over a 12-month period. The follow-up scores on the three spasticity scales (Ashworth, spasm, and clonus scales) were significantly lower at every follow-up visit in comparison to the intake score, except for the clonus scale scores at 12 months. Improvements in health-related quality of life (EQ-5D) and functionality (SIP-68, functional independence measure) were small and nonsignificant. A significant reduction in severity of self-reported personal problems rating scale was observed. Sixty-six patients had no adverse events. Types of adverse events reported were wound complications (22%), catheter problems (36%), cerebrospinal fluid leakage (25%), and other complications (17%). Intrathecal baclofen reduces spasticity and severity of patient-reported problems but its effect on quality of life and functionality is less apparent. Improvements are desired in selection criteria, design of spinal catheters, and outcome scales.

The development and validation of a decision-analytic model representing the full disease course of acute myeloid leukemia

BACKGROUND

The treatment of acute myeloid leukemia (AML) is moving towards personalized medicine. However, due to the low incidence of AML, it is not always feasible to evaluate the cost-effectiveness of personalized medicine using clinical trials. Decision analytic models provide an alternative data source.

OBJECTIVE

The aim of this study was to develop and validate a decision analytic model that represents the full disease course of AML.

METHODS

We used a micro simulation with discrete event components to incorporate both patient and disease heterogeneity. Input parameters were calculated from patient-level data. Two hematologists critically evaluated the model to ensure face validity. Internal and external validity was tested by comparing complete remission (CR) rates and survival outcomes of the model with original data, other clinical trials and a population-based study.

RESULTS

No significant differences in patient and treatment characteristics, CR rate, 5-year overall and disease-free survival were found between the simulated and original data. External validation showed no significant differences in survival between simulated data and other clinical trials. However, differences existed between the simulated data and a population-based study.

CONCLUSIONS

The model developed in this study is proved to be valid for analysis of an AML population participating in a clinical trial. The generalizability of the model to a broader patient population has not been proven yet. Further research is needed to identify differences between the clinical trial population and other AML patients and to incorporate these differences in the model.