Medical Hypotheses xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Medical Hypotheses journal homepage: www.elsevier.com/locate/mehy

Is it the time to rethink clinical decision-making strategies? From a single clinical outcome evaluation to a Clinical Multi-criteria Decision Assessment (CMDA) Alberto Migliore a,⇑, Davide Integlia b, Emanuele Bizzi a, Tomaso Piaggio c a b c

UO of Rheumatology, ‘‘Ospedale S.Pietro FBF’’, Rome, Italy ISHEO, Italy University of Genova, Italy

a r t i c l e

i n f o

Article history: Received 27 May 2015 Accepted 30 June 2015 Available online xxxx

a b s t r a c t There are plenty of different clinical, organizational and economic parameters to consider in order having a complete assessment of the total impact of a pharmaceutical treatment. In the attempt to follow, a holistic approach aimed to provide an evaluation embracing all clinical parameters in order to choose the best treatments, it is necessary to compare and weight multiple criteria. Therefore, a change is required: we need to move from a decision-making context based on the assessment of one single criteria towards a transparent and systematic framework enabling decision makers to assess all relevant parameters simultaneously in order to choose the best treatment to use. In order to apply the MCDA methodology to clinical decision making the best pharmaceutical treatment (or medical devices) to use to treat a specific pathology, we suggest a specific application of the Multiple Criteria Decision Analysis for the purpose, like a Clinical Multi-criteria Decision Assessment CMDA. In CMDA, results from both meta-analysis and observational studies are used by a clinical consensus after attributing weights to specific domains and related parameters. The decision will result from a related comparison of all consequences (i.e., efficacy, safety, adherence, administration route) existing behind the choice to use a specific pharmacological treatment. The match will yield a score (in absolute value) that link each parameter with a specific intervention, and then a final score for each treatment. The higher is the final score; the most appropriate is the intervention to treat disease considering all criteria (domain an parameters). The results will allow the physician to evaluate the best clinical treatment for his patients considering at the same time all relevant criteria such as clinical effectiveness for all parameters and administration route. The use of CMDA model will yield a clear and complete indication of the best pharmaceutical treatment to use for patients, helping physicians to choose drugs with a complete set of information, imputed in the model. Ó 2015 Elsevier Ltd. All rights reserved.

Introduction: the limits of randomized controlled trial – RCT The gold standard for clinical research is the randomized controlled trial (RCT). RCTs are employed to test the effect of an intervention in which patients are randomly assigned to receive the interventions that are being compared, thus the decision about which group the participant is allocated to depend completely on the play of chance. Most importantly, this design allows the construct of two or more comparable groups of patients. Done properly, random allocation of participants to intervention or control groups removes the risk of selection bias and confounding. It also facilitates blindness (masking) of assigned treatments from both

⇑ Corresponding author. Mobile: +39 335 8146506.

investigators and patients preventing other bias such as the differential placebo effect. In addition, it enables to use probability’s theory to express the likelihood that any difference in outcome between treatment groups merely indicates – chance [1]. In clinical trials, primary and secondary endpoints seeks to provide evidence about which interventions work best for which type of patients and under what circumstances. Primary end point is a very critical issue in the design of RCTs since it is used to reach a decision on the overall result of the study. Moreover, it also serves as basis to calculate the sample size for a particular RCT. Finally, a RCT must have only one primary outcome, which should be decided at the outset of the study. Overall the estimates of the efficacy or relative efficacy of the treatments from RCTs are likely to be a more accurate reflection of the ‘true’ efficacy than non-randomized experiments and

E-mail address: [email protected] (A. Migliore). http://dx.doi.org/10.1016/j.mehy.2015.06.024 0306-9877/Ó 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Migliore A et al. Is it the time to rethink clinical decision-making strategies? From a single clinical outcome evaluation to a Clinical Multi-criteria Decision Assessment (CMDA). Med Hypotheses (2015), http://dx.doi.org/10.1016/j.mehy.2015.06.024

2

A. Migliore et al. / Medical Hypotheses xxx (2015) xxx–xxx

observational studies. In other words, RCTs often have greater internal validity than other designs; their results are less likely to be biased and less subject to the effects of confounding than observational studies. Therefore randomized controlled trials when large, well conducted and correctly analyzed, provide the most reliable estimates (internal validity) of the difference in effect between two or more interventions. RCTs, however, has also their disadvantages [2,3]. They can be expensive and time consuming to carry out and hard to organize well, and most importantly those patients who do volunteer and meet the entry criteria of the Trial may not be representative of all the patients for whom one would like to use the intervention (poor external validity). The issue of external validity or generalizability is important and many trials have been weakened because, for example, they only included younger men in trials of treatments for heart disease whereas potential patients are older, sicker and include women. Another important problem with the assessment of clinical trials is due to their mono-dimensional nature of focusing on the various endpoints separately or on only composite endpoint [4]. For instance, the evaluation of primary or secondary endpoints for a particular drug, in RCTs, may result in different or discordant evidence confounding the decision maker. For example, in rheumatoid arthritis (RA) the agent A may work better in symptomatic outcome and the agent B may work better in radiologic or structural outcome, leaving the decision maker to decide which is the most relevant endpoint. Since the sample size and the statistical power of a RCTs are based only on the primary endpoint selected, neither RCTs and meta-analysis are able to offer simultaneously evidence on all important outcomes needed in order to establish the overall relative clinical efficacy for each treatment. This is of particular importance in the context of decision-making where clinicians, other health care professionals and public and private insurers who aim to assess whether a health care intervention represents good value for money, are empowered to make a decision on which intervention to use or fund. In this particular context the exclusive use of data from RCTs represent a limitation. It is rarely the case that the effectiveness of an intervention can be fully assessed simply by using data from one clinical study. Therefore, as a decision must be made, some form of analytical framework is required which is able to take data from a number of sources (RCTs, observational studies and etc.) and these, together with explicit assumptions, are the basis of an overall estimation of clinical efficacy. The network meta-analysis is an extension of traditional or frequentistic meta-analysis by including multiple different pair-wise comparisons across a series of interventions for a given outcome [5]. This is a new tool to synthesize evidence to compare multiple interventions in the absence of head-to-head trials [6]. The Bayesian approach based Network meta-analysis permits a probabilistic analysis and then leads logically to the decision-making context [7], since an advantage of these probabilistic analyses is the capability to rank treatments in terms of their probability to be the most successful treatment. NMA is a Bayesian MTC method that offers several advantages over frequentist indirect methods, including the ability to produce results for all comparisons of interest in a single analysis. In fact, classic meta-analysis, also called frequentistic meta-analysis, only allows dichotomous results regarding the efficacy of a single intervention, reporting by 95% credibility intervals if the selected intervention is effective or not, while probabilistic meta-analysis also grants the possibility to establish the size of the effectiveness of each intervention. The results of network meta-analyses allow us to rank interventions according to the probability of being effective.

The need for a multi-criteria decision model In clinical research, typically, there are several criteria used to assess the efficacy of clinical intervention. These criteria can differ on several respects and might depend on the target of the evaluation. RCTs focus just on one criteria which is – typically – the target of the evaluation (primary endpoint). By focusing only on the main endpoint, usually efficacy (in terms of hard or surrogate endpoints), the analyst excludes the possibility to get further relevant information on other important outcomes, such as other clinical endpoints, safety and compliance. RCTs allows to gather information on one important aspect but we are not able to consider many other parameters that can help to better understand which is the best treatment for a target patient population [3]. Also from an economic point of view, the assessment of one single endpoint is not sufficient. For a proper cost-effectiveness analysis it is necessary to consider also other clinical parameters, the compliance and administration route, long term-efficacy and all aspect of safety and as well as costs. There are plenty of different clinical, organizational and economic parameters to consider in order having a complete assessment of the total impact of a pharmaceutical treatment. In the attempt to follow a holistic approach aimed to provide an evaluation embracing all clinical parameters in order to choose the best treatments, it is necessary to compare and weight of multiple criteria. Therefore, a change is required: we need to move from a decision-making context based on the assessment of one single criteria towards a transparent and systematic framework enabling decision makers to assess all relevant parameters simultaneously in order to choose the best treatment to use. One commonly used approach for including multiple domains in a unique measure of outcome is the estimation of the quality of life [8]. The quality adjusted life year (or QALY) is the most well-known outcome measured used in economic evaluation of healthcare, accounting for the length of life adjusted for the quality of life. However, there are several limitations and methodological problems with the use of QALYs [9], and often decision makers are faced with evaluating the result of clinical trial/observational studies that do not express the primary outcome in terms of QALY. Therefore, it is crucial to create an alternative multi-criteria decision model able to balance all the possible endpoints of interest and to help the decision maker make the right decision.

An emerging tool in decision-making: MCDA The development of a multi-criteria approach to priority setting has recently been identified as one of the most important issues in the decision-making. In several scientific disciplines and social fields, Multi-criteria Decision Analysis is well developed, has achieved general acceptance and is consistently used [10–13]. There are only a very few applications of MCDA in health system research, even if this trend is increasing [14–16]. Multi-criteria Decision Analysis is a valuable tool to face many complex decisions involving multiple criteria goals or objectives of conflicting nature. It is particularly indicated in the resolution of problems characterized by the need to choose among several alternatives [17]. It enables the DM to identify the relative importance of any parameter assessed, it is a logical, consistent and transparent framework, and is reasonably easy to use [18]. Here we describe the steps to accomplish a Multi Criteria Decision Analysis process. The first step of MCDA is the listing of all possible alternatives of intervention in order to solve a complex problem. The second step is to clarify or identify all relevant criteria for evaluating the available alternatives. This step consists in the production of a table of consequences. This should be a table with

Please cite this article in press as: Migliore A et al. Is it the time to rethink clinical decision-making strategies? From a single clinical outcome evaluation to a Clinical Multi-criteria Decision Assessment (CMDA). Med Hypotheses (2015), http://dx.doi.org/10.1016/j.mehy.2015.06.024

3

A. Migliore et al. / Medical Hypotheses xxx (2015) xxx–xxx

criteria listed down the left hand side of the table and alternatives on the right side. Then in the third step the consequences of choosing each alternative from the point of view of each listed criterion are reported in the appropriate cell. The reported data may be notes, or answers to specific questions (yes/no/do not know), or a numerical scale; in this last case the meaning or the value should be clearly explained. The fourth step is to rank each option against each criterion. It will produce a series of ranks based on the consequences table. The ranks, expressing the decision maker’s preferences or the summary of objective evaluations, may identify which options are preferred from the point of view of each selected criterion. In the fifth step the rankings table can be used to eliminate alternatives that are dominated by another alternative, or that fail to reach a minimum acceptable level on one or more of the objectives, as well as any criteria, which rank the remaining options equally. An alternative ‘‘A’’ may be considered as dominant over alternative ‘‘B’’ if: (1) alternative A is better than the other on all criteria; (2) alternative A is better than the other only for some criteria and equal in others. Clearly, this means the elimination of the dominated alternative. In the case that more than one favorable option emerges from this procedure it will be necessary to make a choice, since there will be no option which is best on all criteria (if not one of these alternative would dominate the others). At this point a trade-off between the different criteria should be performed. Since the remaining available alternatives will be worse on some of the criteria than other options and all criteria do not have the same relevance, it is necessary a method of identifying and ranking the performance on each criterion (this is the sixth step). There are several ways by which this can be done. The most effective and transparent method may be converting consequences to value scores. This consists of estimating the value of each consequence from the perspective of each criterion. A similar scale for each criterion is essential to pool these values for a correct final computation. Generally, a scale of 100–0 is used, where 100 (or 10 or 1) represents the best performance and 0 the worst; this way the best and worst performances are clearly defined. After identifying the best and the worst, you need to derive the intermediate scores. It is very easy if you have a measurable criterion, the measures correspond to value scores. In the absence of the measurability of the criteria, the DM will have to rate each alternative directly. The seventh step consists in assessing weights for value scores. Since some criteria are more important than others are, the weights should represent the relative prominence of each criterion taking account of the value scoring system. It is crucial to consult the decision maker or all stakeholders at this point. Although this may be not a problem for individual decision-making, it deserves extreme care in the case of business, clinical or social decision making since different stakeholders may indicate different preferences. It is important to ensure that the weights accurately reflect the DM’s or stakeholders’ values. The weights must be expressed as a percentage in order to add up to 100%. After normalization, the final step is to multiply value scores by weights and add up the total for each alternative, working out aggregate (total) weighted scores for each alternative. This will produce the final rank of the available alternatives. In the case of a group of decision makers, MCDA helps people to discuss the problem to be solved in a way that reflects the contributions of all stakeholders. It gives stakeholders the opportunity of thinking, re-thinking, questioning, weighting, evaluating, correcting and finally deciding. MCDA can be applied also in the process of evaluation of the best pharmaceutical treatment for a specific pathology in a given territory. The hypothesis we suggest is to apply MCDA model in the clinical setting in order to help clinicians

to achieve the best decision based on the Evidence Based Medicine, and on a transparent and systematic methodology enabling clinicians to asses all relevant criteria simultaneously. The MCDA model applied in this field could be called ‘‘Clinical Multiple Criteria Decision Assessment – (CMDA)’’ and could be integrated successively in a context of a wider MCDA framework for further economic or social evaluation.

Clinical Multiple Criteria Decision Assessment (CMDA) – a hypothetical model In order to apply the MCDA methodology to clinical decision making for the best pharmaceutical treatment (or medical devices) to use to treat a specific pathology, we can define a specific Multiple Criteria Decision Analysis for the purpose, like a Clinical Multi-criteria Decision Analysis (CMDA). Like for the general framework of MCDA, in CMDA there is not one predominant criterion. But, in addition to the logic of MCDA, in CMDA the innovative approach is to consider a ‘‘clinical consensus’’ for the selection of all relevant clinical criteria (domains) and related parameters, and then for the attribution of specific weights to each domain and parameter of evaluation. This means that clinical outcome measured on a specific endpoint (parameters) will be weighted according to the importance that a clinical consensus, composed by a scientific board (at regional or national level) will give to the specific domain. Using CMDA approach, the clinical consensus becomes the cardinal point of the evaluation procedure that aims to choose the best clinical treatments, guarantying a transparent and objective process. In CMDA, results from both meta-analysis and observational studies are used by a clinical consensus after attributing weights to specific domains and related parameters. The decision will result from a related comparison of all consequences (i.e., efficacy, safety, adherence, administration route) existing behind the choice to use a specific pharmacological treatment. The Clinical Multiple Criteria Decision Assessment is able to consider many domains and parameters with different weights – attributed by the clinical consensus – in order to evaluate the most efficient clinical pathway for patient population (Table 1 and Fig. 1). CMDA matches meta-analysis results, literature review evidences and results from observational studies with clinical consensus weights on domain and parameters. The match will yield a score (in absolute value) that link each parameter with a specific

Table 1 Description of items taking part in the CMDA process and their evaluation process. Items

Description

Performer

Domain

Main criterion selected in order to assess each intervention

Weight of domain (%) Parameter

Relevance attributed to the domain

Weight of parameter (%) Score from clinical consensus Probability to reach outcome Score

Relevance attributed to the parameter

Listed by clinical consensus Attributed by consensus Listed by clinical consensus Consensus

Global weight attributed to the domain for different parameters

Automatically calculated

Probability of the treatment to reach the desired result

Network Meta-analyses

Overall rating, alias the sum of the scores of all parameter for a given intervention

Automatically calculated

Outcome or measuring instruments for the evaluation of each domain

Please cite this article in press as: Migliore A et al. Is it the time to rethink clinical decision-making strategies? From a single clinical outcome evaluation to a Clinical Multi-criteria Decision Assessment (CMDA). Med Hypotheses (2015), http://dx.doi.org/10.1016/j.mehy.2015.06.024

4

A. Migliore et al. / Medical Hypotheses xxx (2015) xxx–xxx

before. In the Fig. 2 we can see that the sixth column describes the results from meta-analysis, literatures and results from observational studies of treatment 1, and the seventh column reports the final score for each parameter/clinical target. Column eighth and ninth do the same for treatment 2, and so on. These columns contain two different colors: blue, that considers probability to reach parameter (comes from literature research and meta-analysis), and yellow that considers the score obtained from CMDA model. At the bottom of the yellow columns there will be the sum up of all scores of each parameters. This is the final score attributed to each pharmaceutical treatments. The highest score indicate the best treatment to use according to the CMDA model.

1st example: hypothesis of CMDA application to antiosteoporotic drugs

Fig. 1. Flow chart of the evaluation process in CMDA.

intervention, and then a final score for each treatment. The higher is the final score; the most appropriate is the intervention to treat disease considering all criteria (domain and parameters). The results will allow the physician to evaluate the best clinical treatment for his patients considering at the same time all relevant criteria such as clinical effectiveness for all parameters and administration route. The use of CMDA model will yield a clear and complete indication of the best pharmaceutical treatment to use for patients, helping physicians to choose drugs with a complete set of information, imputed in the model (see Fig. 2). Starting from left side, the first column indicates the name of possible domains to be considered for the specific pathology. In the second column there is the weight (in percentage) attributed by the clinical consensus (the sum of weights of each domain will be 100%). In the third column there is the name of several parameters that refers to each domain, and in the fourth column there is the weight (in percentage) of the parameters attributed from clinical consensus (the sum of weights of the parameters related to 1 domain will yield 100%). The fifth column indicates the weight of each parameter compared to the relative weight of the domain. This percentage indicates the weight of each parameters compared with all other parameters. The sum of the weights of all parameters will be 100%. This first 5 columns indicate the process of attribution of weights by the clinical consensus. Going towards right side, we find a series of columns that consider the pharmaceutical treatments that we want to compare. For each of them there are two columns: one column indicates how well performs the pharmaceutical treatment for the specific parameter/clinical target according to meta-analysis, literature review and results from observational studies, and de second column indicate the match between the score formulated from the clinical consensus and the performance of treatment as described

Osteoporosis is characterized by reduced bone mass and disruption of bone architecture, resulting in increased risk of fragility fractures. In nine industrialized countries in North America, Europe, Japan, and Australia, country-specific osteoporosis prevalence (estimated from published data) at the total hip or hip/spine ranged from 9% to 38% for women and 1% to 8% for men. In these countries, osteoporosis affects up to 49 million individuals [19]. Piscitelli et al. estimated that the number of postmenopausal osteoporotic women would increase from 3.3 million to 3.7 million between 2010 and 2020 (+14.3%) in Italy, due to demographic changes, the burden of fractures is expected to increase markedly by 2020 [20]. The economic burden of fragility fractures was estimated at € 37 billion. The costs are expected to increase by 25% in 2025 [21]. Epidemiological data have consistently demonstrates that the annual incidence of fragility fracture increases with age [22]. The burden of fracture is expected to increase with an ageing population, especially in western countries. Therefore, the management of fracture prevention is an important purpose for decision-making. After correcting modifiable risk factors and life style, pharmacological intervention is needed in fracture high risk patients affected by OP. Actually bisphosphonates are considered the first-line therapy for the prevention and treatment of osteoporotic fractures, both vertebral and non-vertebral [23]. Reduction in incidence of new vertebral fractures is the most common and relevant endpoint to achieve registration of any new drug for the primary or secondary prevention of PO. The majority of randomized controlled trials (RCT) on prevention of osteoporosis considers prevention of non-vertebral fractures as a secondary end point even though mortality and morbidity of non-vertebral fractures and specially hip fractures are more relevant than vertebral fractures. Reductions in relative risk (RR) to develop new vertebral and non-vertebral fracture ranged respectively from 41% to 70% and

Fig. 2. The steps to apply CMDA.

Please cite this article in press as: Migliore A et al. Is it the time to rethink clinical decision-making strategies? From a single clinical outcome evaluation to a Clinical Multi-criteria Decision Assessment (CMDA). Med Hypotheses (2015), http://dx.doi.org/10.1016/j.mehy.2015.06.024

A. Migliore et al. / Medical Hypotheses xxx (2015) xxx–xxx

10% to 39% over 3 years in the RCT conducted on bisphosphonates [24–29]. These data are referred to the effect of any single bisphosphonate compared to placebo, no data from head to head RCT focused on reduction of incidence of both new vertebral and non-vertebral fracture among merchandised bisphosphonates are available, since the sample size of study population and the length of follow up should be much larger and longer respectively and the consequent costs will be unaffordable. Another method commonly used to compare bisphosphonates is based on evaluation of change in BMD [30], since an increase in BMD correlates to a reduction in fracture risk, even if it is only one of the determinants of anti-fracturative activity of bisphosphonates [31,32]. Some examples of network meta-analysis are reported in literature comparing vertebral antifracturative efficacy of antiresorptive drugs [33–35]. RCTs analyzed in such meta-analysis usually are registrative trials with a three years follow up. Even though the term of probability of antifracturative efficacy is a key item influencing the choice on OP treatment, other factors are also relevant, such as, safety, compliance [36], long term efficacy (more than 3 years) and administration route especially in a long standing chronic disease. The most clinically effective drug in an experimental setting may be the least effective in clinical practice if the treatment adherence is very scarce or long term safety is poor or if there is a loss of efficacy over time. On the other hand the best way to investigate safety and adherence is from registry data and cohort studies. In addition data from RCTs and both classic and network meta-analyses as well as cohorts studies are unavoidably weak, since they analyze only one outcome at a time. This leads to uncertainty about selection of therapeutic interventions both in clinical practice and in a decision making setting. A concise data synthesis of all relevant parameters/outcomes is needed to facilitate evidence based decision making. The heterogeneity of the evidence on bisphosphonates, the lack of head to head RCTs and the restriction to one endpoint also in networks meta-analysis ask for a more complete approach encompassing all available endpoints to accomplish a synthesis better able to suggest the right decision. In Fig. 3 we show an example of a model for Clinical Multi-criteria Decision Assessment table to match the multiple criteria in order to choose the most effective pharmaceutical treatment to treat osteoporosis. In order to understand the meaning of each column and row we refer to the paragraph 5 after the Fig. 2 where we describe in details the value in the cells. In this example we have postulated that a clinical consensus might ascertain as domains: clinical efficacy at 3 years, adherence, change in BMD, safety and administration route; the same consensus must define all parameters for each domain (as postulated in the 3rd column of the table) and then attribute the related weight to each one. At the end of the evaluation process, policy makers and payers can understand clearly which is the best treatment to treat osteoporosis or RA for a specific patient population.

2nd example: hypothesis of CMDA application to biologic agents for rheumatoid arthritis Rheumatoid arthritis (RA) is a chronic inflammatory disease characterized by inflammation of the synovial lining of joints, tendons, and periarticular structures [37]. This chronic, systemic, inflammatory disorder affects up to 1.0% of the adult population [38]. The bone and cartilage damage of the affected joints can significantly reduce physical function and the chronic inflammation of RA, due to dysregulation of several cytokines, such as tumor necrosis factor (TNF), IL-1 and IL-6, is associated with increased

5

morbidity and mortality. For this reason RA is a significant economic burden on health-care systems and society in general [38]. Conventional disease-modifying antirheumatic drugs (cDMARDs) are generally employed as first-line treatments, usually methotrexate [MTX] alone, or in combination with another chemical DMARD [39]. In the last decade biological compounds (bDMARDs) that specifically target one of the key cytokines involved in the pathogenesis of RA have been developed. According to worldwide guidelines they are recommended for patients with inadequate response or intolerance to cDMARDs. In addition, patients who fail treatment with a first biologic can be switched to another biologic [40]. Several bDMARDs have been licensed for such use in the EU and US. These drugs have radically changed the treatment of RA, with dramatic anti-inflammatory effects, reducing symptoms and increasing the quality of life. The registrative RCTs reported as primary endpoint the percentage of study participants who achieve after six months treatments ACR20, ACR50, and ACR70 clinical response according to the America College of Rheumatology [41]. It means percentage of patients in the study population that achieved respectively a 20%, 50%, 70% improvement in tender or swollen joint counts as well as 20%, 50%, 70% improvement in three of the other five criteria. Other studies reported as primary endpoint ‘‘The European League Against Rheumatism (EULAR)’’ response criteria, based on the disease activity score (DAS) measure [42]. DAS is a scoring system developed in Europe for assessing the severity of rheumatoid arthritis calculated using a formula that includes counts for tender and swollen joints, an evaluation of general health by the patient, and a measure of circulating inflammatory markers. An assessment DAS28 score greater than 5.1 is considered to be indicative of high disease activity, between 5.1 and 3.2 of moderate disease activity and less than 3.2 of low disease activity [43]. A patient scoring less than 2.6 is defined as being in remission. It is also used to monitor the treatment response; a decrease in DAS28 score of 0.6 or less is considered a poor response, while decreases greater than 1.2 points indicate a moderate or good response, depending on whether an individual’s DAS28 score at the end point is above or below 3.2, respectively [43]. Further RCTs have shown also the efficacy of biological DMARDs not only on clinical outcomes, but also on structural outcomes demonstrating to halt radiological disease progression. Unfortunately, there are few trials comparing bDMARDs head-to-head. Even though data from clinical trials and from the main registries established in various countries indicate that all these drugs are effective both on clinical and structural parameters of RA, displaying similar safety profiles, there is no scientific and common assessment to guide the choice of the most appropriate biologic as first or second line treatment. Some network meta-analysis (NMA) have been performed to compare the clinical efficacy of EU licensed-dose bDMARDs for the treatment of RA patients after failure on cDMARDs [44–47]. Efficacy was measured using American College of Rheumatology (ACR) response end points from RCTs. Moreover, although anti-TNF agents are commonly used in the long-term treatment of RA, few randomized double blind controlled studies have evaluated continued therapy with these agents for a period of up to 2 years. Effective management of chronic progressive diseases such as RA requires adherence to the therapeutic regimen for a long time, and increased experience from long-term data registries or cohort studies must be evaluated. Because the disease is chronic, the long-term efficacy and tolerability of the available therapies should be assessed. In the hypothesis to apply CMDA in RA we can apply the model referred in the Fig. 2 with domains, parameters and pharmaceutical treatments specific for rheumatoid arthritis. In this

Please cite this article in press as: Migliore A et al. Is it the time to rethink clinical decision-making strategies? From a single clinical outcome evaluation to a Clinical Multi-criteria Decision Assessment (CMDA). Med Hypotheses (2015), http://dx.doi.org/10.1016/j.mehy.2015.06.024

6

A. Migliore et al. / Medical Hypotheses xxx (2015) xxx–xxx

Fig. 3. CMDA parameters and weights for OP.

case we have hypothesized that a clinical consensus might identify as domains at least: efficacy at 6 month, adherence, the rate of non-radiological progressors, safety and administration route; the same consensus must determine all parameters for each domain as hypothesized in the 3rd column in the table and then attribute the relative weight to each one. Also for RA the evaluation process is the same (see Fig. 4). In the yellow line the final score of each products is reported allowing a comparison between them. The clinical efficacy domain could be

repeated with 3 or 5 years observation. Also for Fig. 4, in order to understand the meaning of each column and row we refer to the paragraph 5 after the Fig. 2 where we describe in details the value in the cells. At the end of the evaluation process, policy makers and payers can understand clearly which is the best treatment to treat osteoporosis or RA for a specific patient population.

Fig. 4. CMDA parameters and weights for RA.

Please cite this article in press as: Migliore A et al. Is it the time to rethink clinical decision-making strategies? From a single clinical outcome evaluation to a Clinical Multi-criteria Decision Assessment (CMDA). Med Hypotheses (2015), http://dx.doi.org/10.1016/j.mehy.2015.06.024

A. Migliore et al. / Medical Hypotheses xxx (2015) xxx–xxx

Discussion The use of MCDA is increasing progressively, it seems the most accurate method for facing and solving complex problems. Many often DM need to evaluate several criteria to make the right decision [48]. Also in the healthy field, both clinicians and DM have to consider many factors influencing therapeutic chooses. The use of MCDA within the health care framework has also been previously discussed but mainly within the QALY framework [49]. The Clinical Multi-criteria Decision Analysis –CMDA aims to utilize the MCDA framework within a broader framework. Even though network meta-analysis are able to overcome the incomparability of RCTs, they can compare only a single outcome at once. NMA is not able to compare simultaneously many outcomes and to evaluate modification of results over time [5]. CMDA may be a valid attempt in order to overcome these difficulties since it takes into account simultaneously all possible criteria considered suitable by clinical DM. It allows clinicians and then DM to better identify the best available intervention. CMDA may be useful also in a context of a pharmaco-economic evaluation. Results from CMDA may be inserted in a wider MCDA performed by Health DM in order to evaluate national or regional health interventions. Moreover CMDA may easily applied in a closer pharmaco-economic point of view, by applying in the general table costs of interventions and cost/values of each parameter taken into account. In this way, clinicians may give the most scientific possible evidence to DM for further economic or social evaluations. One of the most relevant limit is the relativity of domains and parameters established by the clinical consensus that may vary according to its composition. The variability of the consensus agreement on parameters weight due to different nationality from one hand may cause different final scores for the same interventions in different countries, but fixes better adherence at the local level. Another limit is identifying the value of each intervention for each criterion or parameter considering literature. The best way should be to put the results from head to head studies or probabilistic meta-analysis, but unfortunately, they are not always disposable. In this case, available data from literature should be evaluated by the consensus. Moreover, the quality of reported evidence is not taken into account and the strength of the evidence should be also reported. In conclusion we propose to apply the methodology of MCDA in the clinical setting through a new modified model, CMDA. Decision assessment drawn by this model can help both clinicians and DM to make the best available intervention for a given disease. In addition, the results can run in a further pharmaco-economic setting. However further applicative studies are necessary to overcome some limits and improve the model of CMDA.

Conflict of interest All Authors declare they have no conflict of interest to declare regarding the drafting of the present Article. References [1] Kenneth F Schulz, David A. Grimes Generation of allocation sequences in randomised trials: chance, not choice THE LANCET February 9, 2002;359:515– 19. [2] Sanson-Fisher RW, Bonevski B, Green LW, D’Este C. Limitations of the randomized controlled trial in evaluating population-based health interventions. Am J Prev Med 2007;33(2):155–61. [3] Rothwell PM. External validity of randomised controlled trials: ‘‘to whom do the results of this trial apply?’’. Lancet 2005;365(9453):82–93. [4] Lauer MS, Topol EJ. Clinical trials—multiple treatments, multiple end points, and multiple lessons. J Am Med Assoc 2003;289:2575–7.

7

[5] Ades AE, Welton N, Lu G. Introduction to Mixed Treatment Comparisons. Bristol, United Kingdom: University of Bristol; 2007. [6] Jansen JP, Crawford B, Bergman G, et al. Bayesian meta-analysis of multiple treatment comparisons: an introduction to mixed treatment comparisons. Value Health 2008;11:956–64. [7] Ades AE, Sculpher M, Sutton A, et al. Bayesian methods for evidence synthesis in cost-effectiveness analysis. Pharmacoeconomics 2006;24(1):1–19. [8] Mortimer D, Segal L. Comparing the incomparable? A systematic review of competing techniques for converting descriptive measures of health status into QALY-weights. Med Decis Making 2007;28(1):66–89. [9] Brower Amanda. Is it time to take a harder look at the QALY. Biotechnol Healthc 2008;5(3):47–8. [10] Karvetski CW, Lambert JH, Linkov I. Scenario and multiple criteria decision analysis for energy and environmental security of military and industrial installations. Integr Environ Assess Manage 2011;7(2):228–36. [11] Lienert J, Koller M, Konrad J, McArdell CS, Schuwirth N. Multiple-criteria decision analysis reveals high stakeholder preference to remove pharmaceuticals from hospital wastewater. Environ Sci Technol 2011;45(9):3848–57. [12] Mavrotas G, Ziomas IC, Diakouaki D. A combined MOIP-MCDA approach to building and screening atmospheric pollution control strategies in urban regions. Environ Manage 2006 Jul;38(1):149–60. [13] Cheng S, Chan CW, Huang GH. Using multiple criteria decision analysis for supporting decisions of solid waste management. J Environ Sci Health A Tox Hazard Subst Environ Eng 2002;37(6):975–90. [14] Thokala P, Duenas A. Multiple criteria decision analysis for health technology assessment. Value Health 2012;15(8):1172–81. [15] Littlejohns P, Sharma T, Jeong K. Social values and health priority setting in England: ‘‘values’’ based decision making. J Health Organ Manage 2012;26(3):363–73. [16] Postmus Douwe, Tervonen Tommi, van Valkenhoef Gert, Hillege Hans L. Erik Buskens A multi-criteria decision analysis perspective on the health economic evaluation of medical interventions. Eur J Health Econ 2014;15:709–16. [17] Figueira J, Greco S, Ehrgott M, editors. Multiple Criteria Decision Analysis: State of the Art Surveys Series. New York: Springer; 2005. [18] Baltussen Rob. Louis Niessen Priority setting of health interventions: the need for multi-criteria decision analysis. Cost Effect Res Alloc 2006;4:14. [19] Wade SW, Strader C, Fitzpatrick LA, Anthony MS, O’Malley CD. Estimating prevalence of osteoporosis: examples from industrialized countries. Arch Osteoporos 2014;9:182. [20] Piscitelli P, Brandi M, Cawston H, Gauthier A, Kanis JA, Compston J, et al. Epidemiological burden of postmenopausal osteoporosis in Italy from 2010 to 2020: estimations from a disease model. Calcif Tissue Int 2014 Nov;95(5):419–27. [21] Hernlund E, Svedbom A, Ivergård M, Compston J, Cooper C, Stenmark J, et al. Osteoporosis in the European Union: medical management, epidemiology and economic burden. A report prepared in collaboration with the International Osteoporosis Foundation (IOF) and the European Federation of Pharmaceutical Industry Associations (EFPIA). Arch Osteoporos 2013;8(1):136. [22] THE EUROPEAN PROSPECTIVE OSTEOPOROSIS STUDY (EPOS) GROUP. Incidence of vertebral fracture in Europe: results from the European Prospective Osteoporosis Study. J Bone Miner Res 2002;17:716–24. [23] Kling JM, Clarke BL, Sandhu NP. Osteoporosis prevention, screening, and treatment a review. J Womens Health (Larchmt) 2014;23(7):563–72. [24] Cummings SR, Black DM, Thompson De, Applegate WB, Barrett-Connor E, Musliner TA, et al. Effect of alendronate on risk of fracture in women with low bone density but without vertebral fractures: results from the Fracture Intervention Trial. JAMA 1998;280:2077–82. [25] Harris ST, Watts NB, Genant HK, Mckeever CD, Hangartner T, Keller M, et al. Effects of risedronate treatment on vertebral and nonvertebral fractures in women with postmenopausalosteoporosis: a randomized controlled trial.Vertebral Efficacy With Risedronate Therapy (VERT) Study Group. JAMA 1999;282:1344–52. [26] Reginster J, Minne HW, Sorensen OH, Hooper M, Roux C, Brandi ML, et al. Randomized trial of the effects of risedronate on vertebral fractures in women with established postmenopausal osteoporosis. Vertebral Efficacy with Risedronate Therapy (VERT) Study Group. Osteoporos Int 2000;11:83–91. [27] Mcclung MR, Wasnich RD, Recker R, Cauley JA, Chesnut CH, Ensrud KE, et al. Oral ibandronate study group. Oral daily ibandronate prevents bone loss in early postmenopausal women without osteoporosis. J Bone Miner Res 2004;19:11–8. [28] Black DM, Delmas PD, Eastell R, Reid IR, Boonen S, Cauley JA, et al. Horizon pivotal fracture trial. Once-yearly zoledronic acid for treatment of postmenopausal osteoporosis. N Engl J Med 2007;356:1809–22. [29] Lyles KW, Colón-Emeric CS, Magaziner JS, Adachi JD, Pieper CF, Mautalen C, et al. HORIZON Recurrent Fracture Trial. Zoledronic acid and clinical fractures and mortality after hip fracture. N Engl J Med 2007;357(18):1799–809. [30] Wasnich RD, Miller PD. Antifracture efficacy of antiresorptiveagents are related to changes in bone density. J Clin Endocrinol Metab 2000;85:231–6. Clinical fractures and mortality after hip fracture. N Engl J Med 2007;357:1799–1809. [31] Cummings SR, Karpf DB, Harris F, Genant HK, Ensrud K, Lacroix AZ, et al. Improvement in spine bone density and reduction in risk of vertebral fractures during treatment with antiresorptive drugs. Am J Med 2002;112:281–9. [32] Silverman SL, Watts NB, Delmas PD, Lange JL, Lindsay R. Effectiveness of bisphosphonates on nonvertebral and hip fractures in the first year of therapy:

Please cite this article in press as: Migliore A et al. Is it the time to rethink clinical decision-making strategies? From a single clinical outcome evaluation to a Clinical Multi-criteria Decision Assessment (CMDA). Med Hypotheses (2015), http://dx.doi.org/10.1016/j.mehy.2015.06.024

8

[33]

[34]

[35]

[36]

[37] [38]

[39]

[40]

A. Migliore et al. / Medical Hypotheses xxx (2015) xxx–xxx the risedronate and alendronate (REAL) cohort study. Osteoporos Int 2007;18:25–34. Jansen JP, Bergman GJ, Huels J, Olson M. Prevention of vertebral fractures in osteoporosis: mixed comparision of bisphosphonates therapies. Curr Med Res Opin 2009;25(8):1861–8. Jansen JP, Bergman GJ, Huels J, Olson M. The efficacy of bisphosphonates in the prevention of vertebral, hip, and nonvertebral-nonhip fractures in osteoporosis: a network meta-analysis. Semin Arthritis Rheum 2011;40:275–84. Migliore A, Broccoli S, Massafra U, Cassol M, Frediani B. Ranking antireabsorptive agents to prevent vertebral fractures in postmenopausal osteoporosis by mixed treatment comparison meta-analysis. Eur Rev Med Pharmacol Sci 2013;17(5):658–67. Shiligsmann M, Rabenda V, Gathon HJ, Ethgen O, Reginster JY. Potential clinical and economic impact of nonadherence with osteoporosis medications. Calcif Tissue Int 2010;86(3):202–10. Brooks PM. The burden of musculoskeletal disease–a global perspective. Clin Rheumatol 2006 Nov;25(6):778–81. Helmick CG, Felson DT, Lawrence RC, et al. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States. Part I. Arthritis Rheum 2008;58:15–25. Gaujoux-Viala Cécile, Smolen Josef S, Landewé Robert, Dougados Maxime, Kvien Tore K, Mola Emilio Martin, et al. Current evidence for the management of rheumatoid arthritis with synthetic disease-modifying antirheumatic drugs a systematic literature review informing the EULAR recommendations for the management of rheumatoid arthritis. Ann Rheum Dis 2010. Smolen Josef S, Landewé Robert, Breedveld Ferdinand C, Buch Maya, Burmester Gerd, Dougados Maxime, et al. EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological diseasemodifying antirheumatic drugs 2013 update. Ann Rheum Dis 2014;73(3):492–509.

[41] Felson David T, Anderson Jennifer J, Boers Maarten, Bombardier Claire, Furst Daniel, Goldsmith Charles, et al. ACR preliminary definition of improvement in rheumatoid arthritis. Arthritis Rheum 1995;38(6). [42] Wells G, Becker J-C, Teng J, Dougados M, Schiff M, Smolen J, et al. Validation of the 28-joint Disease Activity Score (DAS28) and European League Against Rheumatism response criteria based on C-reactive protein against disease progression in patients with rheumatoid arthritis, and comparison with the DAS28 based on erythrocyte sedimentation rate. Ann Rheum Dis 2009;68(6):954–60. [43] Fransen J, Creemers MC, Van Riel PL. Remission in rheumatoid arthritis: agreement of the disease activity score (DAS28) with the ARA preliminary remission criteria. Rheumatology (Oxford) 2004;43:1252–5. [44] Singh JA, Christensen R, Wells GA, Suarez-Almazor ME, Buchbinder R, LopezOlivo MA, et al. A network meta-analysis of randomized controlled trials of biologics for rheumatoid arthritis: a Cochrane overview. CMAJ 2009;181(11):787–96. [45] Singh JA, Wells GA, Christensen R, Ghogomu E, Maxwell L, Macdonald JK, et al. Adverse effects of biologics: a network meta-analysis and Cochrane overview. Cochrane Database Syst Rev 2011;2. [46] Orme ME, Macgilchrist KS, Mitchell S, Spurden D, Bird A. Systematic review and network meta-analysis of combination and monotherapy treatments in disease-modifying antirheumatic drug-experienced patients with rheumatoid arthritis analysis of American College of Rheumatology criteria scores 20, 50, and 70. Biologicals 2012;6:429–64. [47] Jansen JP, Buckley F, Dejonckheere F, Ogale S. Comparative efficacy of biologics as monotherapy and in combination with methotrexate on patient reported outcomes (PROs) in rheumatoid arthritis patients with an inadequate response to conventional DMARDs – a systematic review and network meta-analysis. Health Qual Life Outcomes 2014;12(1):102. [48] Drummond et al. Methods for the Economic Evaluation of Health Care Programmes. Oxford University Press; 1987. [49] Thokala P, Duenas A. Multiple criteria decision analysis for Health Technology Assessment. Elsevier; 2012. policy perspective.

Please cite this article in press as: Migliore A et al. Is it the time to rethink clinical decision-making strategies? From a single clinical outcome evaluation to a Clinical Multi-criteria Decision Assessment (CMDA). Med Hypotheses (2015), http://dx.doi.org/10.1016/j.mehy.2015.06.024