Christina Donatti PhD, Oxford Outcomes Ltd., Oxford, UK; Diane Wild MSc, Oxford Outcomes Ltd., Oxford, UK; and Asha Hareendran PhD, Pfizer Ltd., Sandwich, UK

The Use of Conceptual Models, Conceptual Frameworks and Endpoint Models to Support Label Claims of Treatment Benefit Using Patient Reported Outcomes

A range of new, though poorly-defined terminology has been used when proposing label claims about patient reported outcome (PRO) endpoints to the US Food & Drug Administration (FDA). The terms 'conceptual model', 'endpoint model', and 'conceptual framework' are now requirements from some regulators when considering PRO labelling claims.

Graphic models can be used to provide a visual picture of interrelationships between concepts and endpoints. They can be used as frameworks for discussion, to select and refine research questions and to demonstrate the context for a treatment benefit claim. While conceptual models, frameworks, and endpoint models have been discussed in the context of research to support PRO label claims, the nuances among them have not been made clear.

What is a 'Conceptual Model'?
Conceptual models in health education research have been defined as 'diagrams of proposed causal linkages among a set of concepts believed to be related to a particular public health problem [/disease] … the conceptual models…show…the small part of the causal web selected for study.…They often are informed by more than one theory, as well as by empirical findings' [1]. A conceptual model provides a visual picture that represents the research question which is to be investigated. The best models convey complex information in a way that is not only parsimonious but also permits a quick grasp of complicated relationships. For example, a conceptual model can be used to delineate the various concepts that are relevant in a particular disease or condition and the hypothesised links between the concepts, or to provide a framework from which potential targets for treatment can be selected. A conceptual model can guide our choice of what to measure and how to measure it, and provide a context for interpreting the findings. [2]

In 1995 Wilson and Cleary published a conceptual model for health-related quality of life (HRQoL) which integrated both biological and psychological aspects of health outcomes. They proposed five different levels in their model, namely physiological factors, symptom status, functional health, general health perceptions and overall quality of life. They defined their conceptual model as ' …a taxonomy of patient outcomes… according to the underlying health concepts they represent and proposes specific causal relationships between different health concepts' [3]. The Wilson-Cleary Conceptual Model of HRQoL is captured in figure 1. As the reader moves from left to right of the model, the nature of the concepts shifts from the cell toward the individual, and then to the interaction of the individual as a member of society.

A conceptual model is valuable in terms of its ability to facilitate the exploration of a disease area, identify potential treatment benefits, and can also help with the selection of endpoints to evaluate treatment benefit.

Material that can be used to help identify the interrelationships between concepts and outcomes for a conceptual model could include a review of scientific literature, key opinion leader interviews and interviews with patients in the disease area of interest. These should ideally be developed very early on in drug development (phase I and II) to facilitate the development of measurement strategies for endpoints used to support a treatment benefit claim.

A conceptual model can be used to explore the patient perspective of a disease and can help identify/target the aspects that need to be addressed. Such conceptual models could be used to demonstrate the relevance and meaningfulness of patient centred outcomes for measuring treatment benefits in a specific disease or condition. They would be especially useful for discussions with regulatory agencies about treatment benefit claims based on patient reported outcomes.

An example of a conceptual model for Atrial Fibrillation (AF), developed based on a review of literature, is proposed in Figure 2. AF is “characterised by uncoordinated atrial activation with consequent deterioration of atrial mechanical functioning” [4]. A review of scientific literature revealed that the physical symptoms and implications of AF have two primary effects on health-related quality of life. AF patients may experience significant limitations in their daily activities as a result of some or all of the physical AF symptoms. Symptoms such as persistent light-headedness, dizziness, pain, and breathing difficulties make employment, travel, and operating a vehicle extremely difficult for AF patients. Moreover, mobility is also impaired in AF patients (beyond the limitations caused by the observable symptoms) because of the exercise intolerance caused by reduced cardiac output. These limitations also have an impact on the patient's social network and may create a feeling of isolation, while the physical symptoms may limit sexual relations. A further implication of AF on HRQoL relates to patients hearing and feeling the heart palpitations or arrhythmias themselves. While objectively physically harmless, the feeling of an irregular, thumping heartbeat may cause insomnia, mental disturbance, anxiety and fear [5]. Based on these findings a model can be devised that covers some of these HRQoL concepts as demonstrated in Figure 2. This model may need to be modified or elaborated based on additional input from patients and clinicians to further map out or validate the impacts of AF on HRQoL.

The next step would be to identify which aspects of the conceptual model are likely to be impacted as a result of a specific treatment or product, i.e. the specific 'treatment benefit claims' that are being pursued. In drug development terms, decision making about treatment benefit claims are usually guided by both the potential attributes of the product and an understanding of market needs. It is important to consider who the key customers will be when considering 'messages' about treatment benefit. For example, if the product is for use in a paediatric asthma population then the key customers are parents of children with asthma; if the indication is in the treatment of Traumatic Brain Injuries (TBI), then this would be health care professionals treating TBI; if the indication is the treatment of hay-fever, then the patients themselves are the key customer. In essence, the claim should drive the development of the measurement strategy for evaluating treatments.

Regulatory agencies require substantial evidence to support a label claim. To support a treatment benefit claim, more than one endpoint may need to be measured in clinical trials. For example, the EMEA CHMP guidance [6] for the evaluation of products for maintenance treatments for asthma requires both lung function and patient reported symptom endpoints to be assessed. The endpoints and relationships between them can be graphically depicted using an endpoint model.

What is an 'Endpoint Model'?
An endpoint model is a model of the relationships between all measures that may be defined as endpoints (primary or supportive) in a clinical trial. It helps to clarify the appropriate endpoint which can be used to describe particular treatment benefit claims. Early in product development, a preliminary endpoint model that specifies the hierarchy and hypothesised relationships between the end-points will facilitate communication with the FDA [7].

The endpoint model is developed based on a systematic/comprehensive review of disease literature, where this is available. The conceptual model, clinical experts and PRO researchers are useful and valid reference points The product development team who have a scientific understanding of the drug under development and understand the proposed benefits to the patient would be key players for the selection of endpoints to evaluate in a clinical trial. These decisions are ideally made in the context of the target product profile for the product. The draft PRO Guidance Document states that “sponsors [need] to identify all endpoint measurement goals early in product development before studies are initiated” [8]. We recommend that the development of the endpoint model occurs at pre-phase II and be revised through the life-cycle of product development.

An endpoint model should include all the concepts that contribute to the evaluation of endpoints that are used to support a treatment benefit claim. The endpoints can be measured using physiological, patient and/or clinician reported outcome measures. The endpoint model should also specify the expected relationships between the endpoints. These linkages would in turn inform the statistical testing strategy for the clinical trial. In Figure 3 we propose a simplified endpoint model for a hypothetical drug for the treatment of AF. It links the desired claim, with the concept being identified and the instrument needed to properly establish if the claim is being effectively measured.

Once the endpoint model has been agreed, the next step is to determine the strategy to measure each endpoint in the clinical trial. It is useful to understand the concepts that make up an endpoint to ensure that the tools selected provide comprehensive and supportive data.

What is an 'Endpoint Model'?
A conceptual framework has sometimes been referred to as a content map/ measurement model. Its purpose is to demonstrate the expected relationships between items within a domain, and of domains within a PRO tool. A conceptual framework is developed by reviewing the literature that has previously been used to justify the tool and the domains being measured. If the literature does not provide sufficient information, the conceptual framework needs to be developed and validated.

Where existing measures are used or revised, there is usually some evidence which sets out the objective of the instrument and information regarding whether it contains distinct domains measuring distinct concepts, and whether a total score can be derived. From the information available, the conceptual framework should be drafted and validated using a combination of qualitative and quantitative methods. Focus groups with patients, item reduction interviews, factor analysis and other psychometric techniques could aid the revision and validation of a conceptual framework.

The draft PRO Guidance sets out the template of the conceptual framework [8] and is clearly explained in the draft Guidance. It discusses in detail what is expected in terms of developing and validating a conceptual framework for a new or existing PRO measure. Thus, if we were to select an endpoint (e.g. psychological well-being) for evaluation from the conceptual model proposed in Figure 2 relating to atrial fibrillation (AF), it would be necessary to carry out the steps outlined in the draft guidance for developing conceptual frameworks. These include using expert and patient input, drafting items, cognitive debriefing of the item list and psychometric validation.

Following on with our previous example using AF, an article on the development of an AF HRQoL questionnaire has been published. The article outlines the steps used by the authors to develop the measure and concludes with an 18 item questionnaire [9]. Figure 4 presents the conceptual framework of the AF-QoL Questionnaire.

Conclusion
In conclusion, endpoint models and conceptual frameworks can be used to support treatment benefit claims.

Ideally the claims should be considered when the 'product concept' is conceived. To develop a measurement strategy to support a label claim, it is useful to review the conceptual model(s) of the disease and its treatments, to help identify the endpoints that support the treatment benefit claim. Seeking input from patients and clinicians helps to ensure the content validity. The conceptual model can be used to support discussions with the internal commercial and clinical teams (as part of generating the target product profile (TPP) [10]). Determining the treatment benefit claims can help to address any unmet needs and present claims which are meaningful for key customers.

When PRO endpoints are proposed to evaluate a particular treatment benefit, a conceptual framework for the PRO instrument helps to describe the concepts/domains/items and the links between them that contribute to the evaluation of the endpoint. The endpoint model and conceptual framework(s) will inform the study design, analysis plans, and hierarchy of testing which are used to support the claim. Importantly, the strategy which is embarked on will need to be reviewed by regulatory agencies as appropriate.

References
[1] Earp JA, Ennett ST. Conceptual models for health education research and practice. Health
Education Research: Theory & Practice 1991;6;163-71.
[2] Rothman ML, Beltran P, Cappelleri JC, Lipscomb J, Tescendorf B, Patient-Reported Outcomes:
Conceptual Issues. Value Health 2007;10(Suppl.)S66-75.
[3] Wilson IB, Cleary PD. Linking clinical variables with health related quality of life. JAMA
1995;273;59-65.
[4] Fuster V, Ryden LE, Asinger RW, et al. ACC/AHA/ESC guidelines for the management of patients
with atrial fibrillation: executive summary. A report of the american college of cardiology/ american
heart association task force on practice guidelines and the european society of cardiology committee
for practice guidelines and policy conferences (Committee to Develop Guidelines for the
Management of Patients With Atrial Fibrillation). J Am Coll Cardiol 2001;38:1231-66.
[5] Hamer ME, Blumenthal JA, McCarthy EA, et al. Quality-of-life assessment in patients with paroxysmal
atrial fibrillation or paroxysmal supraventricular tachycardia. Am J Cardiol 1994;74:826-9.
[6] Notes for Guidance on the Clinical investigation of medicinal products in the treatment of asthma.
The EMEA Committee for Proprietary Medicinal Products (CPMP), 21 November, 2002, page 7.
[7] Patrick DL, Burke LB, Powers JH, et al. Patient-reported outcomes to support medical product
labelling claims: FDA perspective. Value Health 2007;10(Suppl.)S125-37.
[8] Guidance for Industry: Patient-Reported Outcome Measures: Use in Medical Product
Development to Support Labeling Claims. DRAFT GUIDANCE. U.S. Department of Health and
Human Services, Food and Drug Administration Center for Drug Evaluation and Research (CDER).
February 2006. At: http://www.fda.gov/Cder/guidance/5460dft.pdf.
[9] Badia X, Arribas F, Ormaetxe, et al. Development of a questionnaire to measure health-related
quality of life (HRQoL) in patients with atrial fibrillation (AF-QoL). HQLO; 2007;5:37.
[10] See FDA website at: http://www.fda.gov/CDER/GUIDANCE/6910dft.htm
[11] Wild D, Donatti C Hareendran A. Making Treatment Benefit Claims to the FDA: The Role of, and
Interrelationships between, conceptual models, endpoint models and conceptual frameworks.
ISPOR Poster presented at Dublin, 20-23 October, 2007.

  Issues Index | 2008 Issues Index