An Alternative Approach to Comparative Effectiveness Research
An Alternative Approach to Comparative Effectiveness Research
A second clinical example demonstrates the unique challenge of conducting CVER. Recently, we employed this CER approach to analyze efficacy, QOL, and cost data for alternate management strategies for acromegaly. Our goal was not to identify a single, "best" strategy, but rather to provide comprehensive, multidimensional information for patients and clinicians faced with the challenge of selecting a therapeutic strategy. Our approach included comparisons of all 3 domains for several treatment options, all of which enter into the conversations between patients with acromegaly and their physicians on a daily basis. With various therapies having approximately equivalent control rates but highly variable physical, emotional, and financial costs to the patient, it is essential for physicians to understand the trade-offs that exist with regard to all of these domains, so that they can work with patients to design individualized strategies that are in accordance with the patients' personal values. It borders on negligence to fail to inform a patient that a strategy of daily injections may be associated with a significant out-of-pocket cost, particularly if the patient's financial situation prohibits compliance with this treatment regimen. Yet without cost analysis, the physician risks remaining uninformed regarding these real-world issues that affect management, and lacks adequate data to discuss objectively the relative risks and benefits of other treatment alternatives with which the patient may better comply.
Although this methodology allowed us to review and rank a series of management strategies with regard to each of the 3 domains, developing an overall ranking of the actual comparative effectiveness of these strategies required application of a semiquantitative model of the relative weights of efficacy, QOL, and cost in the medical decision-making process. This type of model creates an objective, logical, and quantifiable framework that allows clinicians and patients to understand the dynamic relationships between these 3 domains so that they can work in concert to make the best possible decisions given a particular set of values and circumstances. However, because no such model has been adequately developed, we could only adopt an approach based on equal weighting in our analysis, and advise clinicians to individualize the model on an ad hoc basis.
This example demonstrates how the ability of CVER-type investigations to comprehensively inform clinical decisions is limited by the absence of objective, semiquantitative models that reflect societal values and preferences, as well as by adequate data to inform their development. More research regarding cost and QOL analysis is needed to facilitate the technical and social aspects of developing such a model, but this is not the major obstacle to the evolution of a clinically relevant CER enterprise. Rather, the primary impediment is the reluctance of the medical community and the public to recognize that the present health care environment, in which a perpetual torrent of novel but costly therapeutic strategies contributes to unsustainable health care spending, mandates that difficult decisions be made regarding the fundamental values and practical aspects of health care delivery. Comparative, value-based effectiveness research provides a framework for confronting these questions in an objective fashion, whereas PCOR does not.
Opponents of this position argue that building such models and using them to inform comparative effectiveness analyses controverts attempts at "personalized medicine." This is only true in the setting of rigid and unyielding adherence to these models, a condition that most reasonable clinicians and patients would be likely to denounce. Instead, semiquantitative models used in conjunction with a CVER approach to comparative effectiveness analysis provide a concrete framework in which the relative risks and benefits of alternate treatment strategies can be compared objectively, and the dynamics and implications of their interactions can be studied on both the individual and population level. Investigations designed to build these models and to gather data that facilitates their application are the necessary first step in developing a strategy for comparative, value-based effectiveness analysis that can stand in contrast to the already-handicapped PCOR model that is currently overtaking the US CER enterprise.
The Role of Semiquantitative Modeling in CVER
A second clinical example demonstrates the unique challenge of conducting CVER. Recently, we employed this CER approach to analyze efficacy, QOL, and cost data for alternate management strategies for acromegaly. Our goal was not to identify a single, "best" strategy, but rather to provide comprehensive, multidimensional information for patients and clinicians faced with the challenge of selecting a therapeutic strategy. Our approach included comparisons of all 3 domains for several treatment options, all of which enter into the conversations between patients with acromegaly and their physicians on a daily basis. With various therapies having approximately equivalent control rates but highly variable physical, emotional, and financial costs to the patient, it is essential for physicians to understand the trade-offs that exist with regard to all of these domains, so that they can work with patients to design individualized strategies that are in accordance with the patients' personal values. It borders on negligence to fail to inform a patient that a strategy of daily injections may be associated with a significant out-of-pocket cost, particularly if the patient's financial situation prohibits compliance with this treatment regimen. Yet without cost analysis, the physician risks remaining uninformed regarding these real-world issues that affect management, and lacks adequate data to discuss objectively the relative risks and benefits of other treatment alternatives with which the patient may better comply.
Although this methodology allowed us to review and rank a series of management strategies with regard to each of the 3 domains, developing an overall ranking of the actual comparative effectiveness of these strategies required application of a semiquantitative model of the relative weights of efficacy, QOL, and cost in the medical decision-making process. This type of model creates an objective, logical, and quantifiable framework that allows clinicians and patients to understand the dynamic relationships between these 3 domains so that they can work in concert to make the best possible decisions given a particular set of values and circumstances. However, because no such model has been adequately developed, we could only adopt an approach based on equal weighting in our analysis, and advise clinicians to individualize the model on an ad hoc basis.
This example demonstrates how the ability of CVER-type investigations to comprehensively inform clinical decisions is limited by the absence of objective, semiquantitative models that reflect societal values and preferences, as well as by adequate data to inform their development. More research regarding cost and QOL analysis is needed to facilitate the technical and social aspects of developing such a model, but this is not the major obstacle to the evolution of a clinically relevant CER enterprise. Rather, the primary impediment is the reluctance of the medical community and the public to recognize that the present health care environment, in which a perpetual torrent of novel but costly therapeutic strategies contributes to unsustainable health care spending, mandates that difficult decisions be made regarding the fundamental values and practical aspects of health care delivery. Comparative, value-based effectiveness research provides a framework for confronting these questions in an objective fashion, whereas PCOR does not.
Opponents of this position argue that building such models and using them to inform comparative effectiveness analyses controverts attempts at "personalized medicine." This is only true in the setting of rigid and unyielding adherence to these models, a condition that most reasonable clinicians and patients would be likely to denounce. Instead, semiquantitative models used in conjunction with a CVER approach to comparative effectiveness analysis provide a concrete framework in which the relative risks and benefits of alternate treatment strategies can be compared objectively, and the dynamics and implications of their interactions can be studied on both the individual and population level. Investigations designed to build these models and to gather data that facilitates their application are the necessary first step in developing a strategy for comparative, value-based effectiveness analysis that can stand in contrast to the already-handicapped PCOR model that is currently overtaking the US CER enterprise.