Ischemic heart diseases remain one of the most frequent causes of morbidity and mortality in Europe and throughout the developed world (Reference Allender, Scarborough, Peto and Network1). In a substantial portion of patients (≥25 percent) with coronary heart disease (CHD), a myocardial infarction or sudden cardiac death without prior symptoms is the first manifestation of the disease (Reference Greenland, Smith and Grundy16). The risk of CHD can be lowered by changing behavior alone or in conjunction with medical therapy. The choice of preventive action depends on an estimation of the long-term risk of suffering a serious cardiovascular event (e.g., death, myocardial infarction). For the prediction of these events, well-known risk factors for CHD such as age, gender, smoking, hypertension, increased blood lipid level, and comorbidities such as diabetes mellitus are used (Reference Graham, Atar and Borch-Johnsen15). Questions have also arisen as to whether additional risk factors should be used to better predict the occurrence of CHD and stratify risk groups. In the pathogenesis of arteriosclerosis, a central role is assigned to inflammatory processes (Reference Libby and Ridker22). The high-sensitivity C-reactive protein (hs-CRP) is a biomarker that indicates systemic inflammation processes. Therefore, hs-CRP can be considered a potential candidate for risk prediction (Reference Myers, Rifai and Tracy25).
When considering a test as a screening instrument for asymptomatic people, it is important to take into account the economic aspects of testing in addition to the medical benefits. The German Agency for Health Technology Assessment (DAHTA@DIMDI), a body of the German Federal Ministry of Health, commissioned a health technology assessment (HTA) report to evaluate the predictive value, clinical effectiveness and cost-effectiveness of the hs-CRP-measurement for risk prediction.
The aim of this HTA was to present evidence that systematically assessed whether additional hs-CRP-measurement (i) improves the risk prediction of cardiovascular events in asymptomatic patients as compared to previously established prediction models, (ii) adds clinically relevant value by changing prevention strategies for a substantial portion of people resulting in a decrease of cardiovascular mortality or morbidity, and (iii) is cost-effective when compared with risk assessment based only on traditional risk factors.
METHODS
Eligibility Criteria
We included all prospective studies of initially asymptomatic subjects evaluating the association of hs-CRP and cardiovascular events comparing a risk prediction model with hs-CRP to a prediction model using only traditional risk factors such as age, sex, smoking, cholesterol, glucose metabolism, and blood pressure. In addition, an effect measure had to be reported for the test accuracy (e.g., sensitivity, specificity, receiver operating characteristic, area under the receiver operator characteristic curve (AUC), and C-statistics, respectively). For details on eligibility criteria of studies analyzing the clinical effectiveness or cost-effectiveness of hs-CRP-screening, see Supplementary Table 1, which is available at www.journals.cambridge.org/thc2010002.
Literature Search
We performed a literature search up to January 2007 of twenty-six electronic databases including Medline, Embase, and the Cochrane Library (see Supplementary Table 2, which is available at www.journals.cambridge.org/thc2010002). Bibliographies of systematic reviews, meta-analyses, and HTA reports were used to identify further studies.
Selection, Validity Assessment, and Data Abstraction
Two authors independently screened the titles and abstracts of all identified studies. On retrieving full text versions, at least two authors independently determined whether the inclusion criteria were fulfilled. Reasons for exclusion of literature obtained in full text were indicated. The evaluation of the enclosed articles took place on the basis of standardized check lists (Supplementary Table 3, which is available at www.journals.cambridge.org/thc2010002), whereas the extraction was performed on the basis of extraction tables and forms that were developed before the evaluation (Reference Altman, Egger, Smith and Altman2;14;Reference Randolph, Guyatt, Calvin, Doig, Richardson and Scott30; Reference Siebert, Behrend, Mühlberger, Wasem, Greiner, v.d.Schulenburg, Welte and Leidl38; Reference Williams, Brunskill, Altman, Briggs, Campbell, Clarke, Glanville, Gray, Harris, Johnston and Lodge42).
Data Synthesis
We systematically described the characteristics and quantitative parameters of the included studies in evidence tables. For cost data used in the studies, currency conversions were performed using purchasing power parities of the Organisation for Economic Co-operation and Development countries and an adjustment for inflation was performed to the year 2006 (26;39) (Supplementary Table 4, which is available at www.journals.cambridge.org/thc2010002).
RESULTS
Included Studies
Our systematic literature search yielded a total of 1,577 references after the exclusion of duplicates; Figure 1 depicts the selection process. Seven studies (eight publications) (Reference Cook, Buring and Ridker7;Reference Danesh, Wheeler and Hirschfield9;Reference Folsom, Chambless and Ballantyne13; Reference Koenig, Lowel and Baumert21; Reference Ridker, Rifai and Rose35;Reference St-Pierre, Cantin and Bergeron40;Reference Van Der Meer, de Maat and Kiliaan41;Reference Wilson, Nam and Pencina43) investigated the incremental prediction of hs-CRP for myocardial infarction and cardiac death, one study assessed the effectiveness of hs-CRP as a screening test for prevention of cardiovascular events (Reference Blake, Ridker and Kuntz3), and three studies examined health-economic aspects of the hs-CRP-test (Reference Blake, Ridker and Kuntz4;Reference Ess and Szucs11;Reference Ess and Szucs12).
Literature search and selection of studies.
Clinical Evaluation
Study Characteristics. Four cohort studies and three nested case-control studies altogether included data from a total of 46,458 people. The average follow-up ranged from 6.6 to 20.6 years. Two studies explored exclusively men (Reference Koenig, Lowel and Baumert21;Reference St-Pierre, Cantin and Bergeron40), one exclusively women (Reference Cook, Buring and Ridker7;Reference Ridker, Rifai and Rose35). In the remaining studies, 44 percent to 72 percent of the participants were male. Supplementary Tables 5 to 7 (available at www.journals.cambridge.org/thc2010002) depict the study characteristics and prediction models included.
Cox proportional hazard models and logistic regression models were used for risk prediction. A decision-analytic Markov model using a life-time horizon to assess both the clinical benefit and the cost-effectiveness is described in the following section on health economic results (Reference Blake, Ridker and Kuntz3;Reference Blake, Ridker and Kuntz4).
Assessment of Study Quality. Outcomes measurement primarily took place in a nonblinded manner and the proportion of the observed subjects was typically below 80 percent due to the long observation period, although in some cases it remained unclear (see Supplementary Table 8, which is available at www.journals.cambridge.org/thc2010002). The applied model-building methods and presentation of results were adequate (Reference Harrell17) as far as could be judged based on the methods descriptions provided. Only one of the models had been validated in another study population. For further details on the study quality, see Supplementary Table 9, which is available at www.journals.cambridge.org/thc2010002.
Incremental Prognostic Value of hs-CRP-Measurement. In six of the seven studies, crude and adjusted predictive effect measures such as odds ratios (OR), or hazards ratios (HR) of the hs-CRP-value and the subsequent occurrence of cardiovascular events were reported. Adjustment for traditional risk factors was performed (see Supplementary Figure 1 and Supplementary Table 10, which are available at www.journals.cambridge.org/thc2010002). The adjusted OR or HR, which typically compared the stratum with the highest hs-CRP-value to that with the lowest, fell between 0.7 and 2.47. In three of the studies, the adjusted effect measures were no longer statistically significant for hs-CRP (Figure 2).
Crude and adjusted measures of association for high-sensitivity C-reactive protein (hs-CRP) and cardiovascular events. HR, hazards ratio; OR, odds ratio; 95 percent CI, 95 percent confidence interval; NR, not reported.
Incremental Accuracy of hs-CRP. In all studies, the gain in prognostic value associated with the risk models that included the hs-CRP-level as a predictor in addition to traditional risk factors was examined by means of the discrimination of the models on the basis of the AUC (Figure 3). The absolute differences between both models were between 0.00 and 0.027. In four (Reference Cook, Buring and Ridker7;Reference Danesh, Wheeler and Hirschfield9;Reference Folsom, Chambless and Ballantyne13;Reference Koenig, Lowel and Baumert21) of the seven studies, the difference in the AUC was statistically significant (p < .05). Because only one of the prediction models (Reference Cook, Buring and Ridker7) was validated in another study population, one cannot exclude the possibility that this gain in prognostic value was overestimated. In summary, the accuracy data, at least on the basis of the AUC, improved marginally on the whole.
Area under the curve without and with high-sensitivity C-reactive protein (hs-CRP) in the prediction model. AUC, area under the curve; sig, statistically significant; ns, not statistically significant. Discrimination according to Harrell (17 l AUC 0.7–0.8: acceptable; AUC 0.8–0.9: excellent; AUC > 0.9: outstanding.
Relevant Actions Regarding hs-CRP-Measurement
The clinical relevance of the aforementioned increase in the AUC was examined only rudimentarily in one of the studies by means of a reclassification analysis (Reference Cook, Buring and Ridker7). The study reported how much of the population was reclassified by the model with hs-CRP using four risk categories based on the 10-year risk of experiencing a cardiovascular event (>20 percent, 10–20 percent, 5–10 percent, <5 percent) as is consistent with the current cardiovascular prevention guidelines. The proportions of the study population in these categories were 0.8 percent, 3.0 percent, 8.4 percent, and 87.9 percent, respectively. A total of 14.4 percent of the high-risk study population as well as 18.7 and 21.3 percent of the two medium-risk categories were reclassified, although only 2.1 percent of the low-risk category was reclassified. The actions suggested by this reclassification were not investigated, and worse reclassified cases were not distinguished from better reclassified cases.
Relevant Benefit to Patients with Respect to Myocardial Infarctions and Cardiac Mortality. We did not identify either randomized or nonrandomized trials that compared the effectiveness of preventive actions based on a risk assessment using only traditional risk factors to the effectiveness of preventive actions based on a risk assessment using traditional risk factors and hs-CRP-level.
In one decision-analytic modeling study (Reference Blake, Ridker and Kuntz3), a Markov model was used to compare the increase in life expectancy due to statin therapy in three groups: individuals without hyperlipidemia but with elevated hs-CRP-levels, those with hyperlipidemia but without elevated hs-CRP-levels, and those with normal cholesterol and hs-CRP-levels. The gain in life expectancy for the base case analysis was similar for 58-year-old people for the first and second group (6.6 months versus 6.7 for men 6.4 versus 6.6 for women), whereas those in the third group did not benefit considerably (0.6 months for men and women). In sensitivity analyses, the assumptions regarding the rate of myocardial infarctions and the efficacy of statin therapy with respect to the prevention of myocardial infarctions influenced the results most. For persons with elevated CRP values but normal LDL values, they varied between 2.5 and 18 months of life gained.
ECONOMIC EVALUATION
Study Characteristics. Three publications addressed the cost-effectiveness of hs-CRP-screening by applying two different decision-analytic models that evaluated primary prevention with statins after stratifying the population based on the results of a CRP test for the US healthcare system (Reference Blake, Ridker and Kuntz4) or different European Countries (Reference Ess and Szucs11;Reference Ess and Szucs12) (Supplementary Table 11, which is available at www.journals.cambridge.org/thc2010002).
Both models compared only the LDL-lipid level as opposed to a comprehensive risk score as an alternative risk assessment to hs-CRP for directing decisions about preventive measures. Therapeutic interventions that were considered included additional therapy with statins or aspirin. For further details on the study design, see Supplementary Table 11.
Assessment of Study Quality. Both studies investigating CRP-screening used data from various sources on the effects and costs of screening and statin therapy as well as on frequency and costs of subsequent cardiovascular events. Blake et al. (Reference Blake, Ridker and Kuntz3;Reference Blake, Ridker and Kuntz4) used data on the efficacy of statin therapy for the prevention of cardiovascular events in asymptomatic patients with elevated CRP levels that were derived from the post hoc analysis of the AFCAPS/TexCAPs study (Reference Ridker, Rifai and Clearfield33); Ess et al. (Reference Ess and Szucs11;Reference Ess and Szucs12), using data from a secondary prevention study of statins called the CARE study (Reference Ridker, Rifai and Pfeffer34), overestimated the efficacy of statin therapy in a primary prevention setting.
Overall, the data on effects and costs included in the models carry high uncertainty. Only the model by Blake addresses this issue with comprehensive one- and multi-way sensitivity analyses (see Supplementary Table 12 for details, which is available at www.journals.cambridge.org/thc2010002).
Cost-Effectiveness Results. The results of different screening and treatment strategies after currency conversion and inflation adjustment for the model by Blake (Reference Blake, Ridker and Kuntz4) are depicted in Figure 4. The results for the models by Ess (Reference Ess and Szucs11;Reference Ess and Szucs12) are presented in Supplementary Figure 1.
Results of the Study of Blake 2003: Incremental cost-utility ratios in Euro 2006 per quality-adjusted life-year gained. Strategy 1: hs-CRP screening with statin therapy for elevated hs-CRP-levels. Strategy 2: hs-CRP-test and statin therapy if hs-CRP > 0.16 mg/dl. Strategy 3: Statin therapy for all patients.
In the model by Blake, a hs-CRP-screening with statin therapy for elevated (≥1.6 mg/l) hs-CRP-levels (strategy 1) was compared with standard therapy (no hs-CRP-screening and no therapy) (Figure 4).
The incremental cost-effectiveness ratio (ICER) and incremental cost-utility ratio (ICUR) were calculated for the base case of 58-year-old men and women as well as for different age groups between 35 and 85 years. With few exceptions, the ICER and ICUR fell above €50,000 per life-year and quality-adjusted life-year (QALY) gained for men in higher age groups. Statin therapy for all (strategy 3) compared with hs-CRP-screening with selective statin therapy (strategy 2) was greater than €500,000 per QALY gained for both genders.
Univariate sensitivity analyses identified the annual risk of myocardial infarction, and the costs and effectiveness of the statin therapy to prevent myocardial infarction (MI) as sensitive variables. A three-way sensitivity analysis was conducted in which the 10-year risk of an MI was allowed to vary between 5 and 25 percent, the costs of the statin therapy were set at US$500 (€540), US$1,000 (€1,081), or US$1,500 (€1,621), and the annual efficacy of the therapy varied between 30 percent and 60 percent.
For the low-cost group, the results ranged from being cost-saving to a maximum ICUR of €40,000 per QALY gained for men; whereas for women in the low-risk group with 5 percent risk of MI and an efficacy of 30 percent, the ICUR was above €43,750 per QALY; all other measures were substantially lower. In the high-cost statin therapy group, the ICER for men only fell below €43,750 per QALY for an efficacy of at least 45 percent and a 10-year risk of MI of at least 10 percent. For women in the high-cost statin therapy group, even at an assumed efficacy of 60 percent, the 10-year MI risk had to be at least 15 percent for the ICUR to remain below US$50,000 per QALY. For high- and medium-cost groups, the ICUR was above €43,750 for those with low cardiovascular risk of both genders and it was as high as €146,000 for men and 243,300 for women.
In conclusion, the sensitivity analysis demonstrated that hs-CRP-screening in individuals without overt hyperlipidemia may be cost-effective under certain conditions when compared with not performing an hs-CRP-screening.
DISCUSSION
We were unable to identify comprehensive intervention studies comparing the use of prediction models with and without hs-CRP as a screening test to guide interventions to prevent cardiovascular mortality and morbidity. The evidence from prospective cohort studies showed that the addition of the hs-CRP-test moderately improves the risk prediction of cardiovascular events. Adjusted relative risk measures ranged from 0.7 to 2.47. A statistically significant increase in the AUC was only reported in four studies, but the extent of the increase was minimal (between 0.00 and 0.027). Prediction studies published after the literature search for this HTA report was conducted showed similar results (Reference Shah, Casas and Cooper36;Reference Wilson, Pencina and Jacques44) or did not find an association of hs-CRP with cardiovascular events (Reference Olsen, Hansen and Christensen27).
A general scientific discussion arose regarding which statistical measures were appropriate for comparing different prediction models and what conclusions could be drawn regarding the clinical relevance of the improvement in statistical summary measures (Reference Cook6;Reference Pencina, D'Agostino and D'Agostino29). The changes in AUC alone cannot indicate whether the risks predicted are sufficiently different enough to alter treatment decisions (Reference McGeechan, Macaskill and Irwig24). Reclassification analyses allow one to do this in the case of cardiovascular disease, because in this field treatments change at defined risk levels according to practice guidelines. After this HTA report was finished, four prediction studies were analyzed or reanalyzed by reclassification analyses (Reference Cook5;Reference Shah, Casas and Cooper36;Reference Wilson, Pencina and Jacques44). Two of these studies reported the proportion of individuals whose prescription of statins was changed by the addition of hs-CRP: 28 of 2,412 individuals and 31 of 919 individuals (Reference Shah, Casas and Cooper36).
Meanwhile, there is good evidence from the randomized trial that statin therapy in individuals with an elevated (≥2 mg/l) hs-CRP-level and a normal (<130 mg/dl) LDL level can prevent cardiac events to a relevant degree. After a median follow-up of 1.9 years, the rate of the combined primary endpoint of cardiovascular disease was reduced from 1.36 (placebo) to 0.77 (rosuvastatin) per 100 person-years (hazards ratio, 0.56; 95 percent confidence interval, 0.46 – 0.69; p < .0001). A higher incidence of diabetes was reported in the rosuvastatin group 3.0 percent versus 2.4 percent p = .01 (Reference Ridker, Danielson and Fonseca32). Long-term safety remains unclear in this low-risk population (Reference Hlatky18). However, this does not automatically mean that only a general hs-CRP-testing in asymptomatic individuals could additionally detect individuals with low LDL levels who could benefit from statin therapy. It is unclear whether hs-CRP is a causal risk factor or not and the known cardiovascular risk factors might also identify an essential portion of this subgroup. To address whether the introduction of hs-CRP-screening in combination with a statin therapy is effective, the JUPITER trial results could be combined with the results of the prediction studies in a decision-analytic model. (Reference Siebert37). Unlike in existing decision-analytic studies, additional hs-CRP (Reference Blake, Ridker and Kuntz3;Reference Blake, Ridker and Kuntz4;Reference Ess and Szucs11;Reference Ess and Szucs12) -screening should be compared with screening with traditional risk factors alone, because the initiation of treatment is not only dependent on the LDL level but also on the global risk of cardiovascular disease. This approach should also be taken in cost-effectiveness analyses.
Our systematic review of cost-effectiveness of hs-CRP-screening was unable to identify health economic studies that used this comparator. In principle, it appears that it is possible to use hs-CRP-testing in a cost-effective manner. Sensitivity analyses demonstrated that the ICER was most sensitive to the baseline risk of myocardial infarction, the effectiveness, and the costs of statin therapy. The ICER reached cost-effective values in some situations. However, the data sources on medical efficacy available in the modeling studies of the review are outdated, and new modeling studies should be done to inform policy. The data of the prediction studies could be combined with the efficacy results of the JUPITER trial, and country specific data on the costs of hs-screening and treatment of the detected individuals to determine the cost-effectiveness of hs-CRP-screening.
Previous HTA reports, systematic reviews and meta-analyses have already investigated the predictive value and to some extent the clinical efficacy of additional hs-CRP-testing for risk prediction in asymptomatic people. Two older HTA reports (Reference Kaul19;Reference Klee, Jaffa and Lakatua20), a more recent systematic review (Reference Lloyd-Jones, Liu and Tian23), three meta-analyses by Danesh et al. (Reference Danesh, Collins and Appleby8–Reference Danesh, Whincup and Walker10) and the 2003 scientific statements by the Center of Disease Control and Prevention (CDC) and the American Heart Association (AHA) (Reference Pearson, Mensah and Alexander28) concluded that hs-CRP did not sufficiently improve the risk discrimination when used as a routine screening test for the general population, but could be used specifically for selective decision making about the type of primary prevention used in moderate risk patients. In his narrative review, Ridker (Reference Ridker31) proposed to extend this recommendation of CRP-screening to all people with a 10-year MI risk between 5 percent and 20 percent based on the prediction model of the adult treatment panel III guidelines, whereas the most recent and most comprehensive systematic review by Shah et al. (Reference Shah, Casas and Cooper36) concluded that improvement in risk stratification due to the addition of hs-CRP to risk models is small and guidance on the clinical use of hs-CRP questionable.
Limitations
In this systematic review, we decided against quantitative pooling of the effect measures for hs-CRP and CAD using statistical meta-analysis, because the studies varied in their cutoff values for hs-CRP-level and because the events used as clinical endpoints and the risk factors used in the adjustment of analyses differed.
CONCLUSIONS
In conclusion, the HTA report indicates there is not sufficient evidence available to demonstrate that measurement of hs-CRP-values in addition to traditional risk factors as part of a global risk assessment for CAD or cardiovascular disease improves patient outcomes.
The additional measurement of the hs-CRP-level increases the incremental predictive value of the risk prediction. It has not yet been clarified whether this increase is clinically relevant and results in a reduction of cardiovascular morbidity and mortality. An altered assessment of the cardiovascular risk by hs-CRP-testing would result in a different decision as to whether additional statin therapy should be initiated for primary prevention, most likely affecting those who fall in an intermediate cardiovascular risk group (5–20 percent in 10 years).
The JUPITER trial provided evidence that statin therapy can reduce the occurrence of cardiovascular events for asymptomatic individuals with normal lipid and elevated hs-CRP-levels. However, a direct assessment of the clinical value and the cost-effectiveness of universal hs-CRP-testing is still lacking.
There is a general demand for further research aimed at developing and validating statistical measures to evaluate prediction models with biomarkers and risk factors and that can also be applied easily by clinicians and interpreted in a meaningful way. Furthermore, it seems important that these statistical measures be easily integrated into decision-analytic approaches, because clinical consequences and costs may vary among biomarkers with identical predictive values that could be addressed in a decision-analytic model.
SUPPLEMENTARY MATERIALS
Supplementary Table 1
Supplementary Table 2
Supplementary Table 3
Supplementary Table 4
Supplementary Table 5
Supplementary Table 6
Supplementary Table 7
Supplementary Table 8
Supplementary Table 9
Supplementary Figure 1
Supplementary Table 10
Supplementary Table 11
Supplementary Table 12 www.journals.cambridge.org/thc2010002
CONTACT INFORMATION
Petra Schnell-Inderst, MPH, PhD (petra.schnell-inderst@umit.at), Senior Scientist, Institute of Public Health, Medical Decision Making and Health Technology Assessment, UMIT—University for Health Sciences, Medical Informatics and Technology, Eduard Wallnoefer Center I, A-6060 Hall i.T., Austria; Institute for Health Care Management, University of Duisburg-Essen, Schützenbahn 70, 45127 Essen, Germany
Ruth Schwarzer, MA, MPH, ScD (ruth.schwarzer@umit.at), Senior Scientist, Institute of Public Health, Medical Decision Making and Health Technology Assessment, UMIT — University for Health Sciences, Medical Informatics and Technology, Eduard Wallnoefer Center I, A-6060 Hall i.T., Austria
Alexander Göhler, MD, MPH, MSc (public-health@umit.at), Assistant Professor, Institute of Public Health, Medical Decision Making and Health Technology Assessment, UMIT—University for Health Sciences, Medical Informatics and Technology, Eduard Wallnoefer Center I, A-6060 Hall i.T., Austria; Senior Scientist, Cardiovascular Research Program, Institute for Technology Assessment and Department of Radiology, Massachusetts General Hospital, Harvard Medical School, 101 Merrimac Street, Boston, Massachusetts 02114
Norma Grandi, MPH (public-health@umit.at), PhD Student, Institute of Public Health, Medical Decision Making and Health Technology Assessment, Department of Public Health, Information Systems and Health Technology Assessment, UMIT—University for Health Sciences, Medical Informatics and Technology, Eduard Wallnoefer Center I, A-6060 Hall i.T., Austria
Kristin Grabein (Kristin.grabein@uni-due.de), freelancer, Institute for Health Care Management, University of Duisburg-Essen, Schützenbahn 70, 45127 Essen, Germany
Björn Stollenwerk, PhD (bjoern.stollenwerk@helmholtz-muenchen.de), Senior Scientist, Helmholtz Zentrum München (GmbH), Institute of Health Economics and Health Care Management, Ingolstädter Landstraße 1, 85764 Neuherberg, Germany; Institute of Public Health, Medical Decision Making and Health Technology Assessment, Department of Public Health, Information Systems and Health Technology Assessment, UMIT—University for Health Sciences, Medical Informatics and Technology, Eduard Wallnoefer Center I, A-6060 Hall i.T., Austria
Jennifer Manne, MSc (jmanne@hsph.harvard.edu), PhD Student, Department of Global Health and Population, Harvard School of Public Health, 677 Huntington Avenue, Boston, Massachusetts 02115
Volker Klauss, MD (klauss@med.uni-muenchen.de), Professor of Internal Medicine, Chair, Department of Cardiology, Medical Clinic, Campus Innenstadt, University Hospital of Munich, Ludwig-Maximilians-University, Munich, Germany
Uwe Siebert, MD, MPH, MSc, ScD (public-health@umit.at), Professor of Public Health (UMIT), Chair, Department of Public Health, Information Systems and Health Technology Assessment, UMIT–University for Health Sciences, Medical Informatics and Technology, Eduard Wallnoefer Center 1, Hall i.T., Austria, A-6060; Adjunct Professor of Health Policy and Management, Center for Health Decision Science, Department of Health Policy and Management, Harvard School of Public Health, 718 Huntington Avenue, Boston, Massachusetts 02115; Director of Cardiovascular Research Program, Institute for Technology Assessment and Department of Radiology, Massachusetts General Hospital, Harvard Medical School, 101 Merrimac Street, Boston, Massachusetts 02114
Jürgen Wasem, PhD (juergen.wasem@uni-due.de), Professor, Chair, Institute for Health Care Management, University of Duisburg-Essen, Schützenbahn 70, 45127 Essen, Germany
