Background

As patients are increasingly asked to pay for all or part of the cost for their medicines, perspectives on the symptoms they experience, how they feel and function, and their quality of life (QOL) associated with their health condition and its treatment have become important in the evaluation of the benefits and the demonstration of the value of new medicines. The patients’ perspectives were not considered in traditional clinical endpoints. Today, patient-reported outcome (PRO) measures are used to evaluate trial outcomes from the patients’ perspective (without interpretation by investigators, physicians or anyone else). PROs cover a range of concepts, from simple symptoms such as pain to more complex multi-domain concepts such as health-related quality of life (HR-QOL) or treatment satisfaction. These can be captured in a variety of formats, such as event logs, rating scales or checklists, using electronic or paper and pencil forms, as well as responses recorded by telephone or in-person interviews. PROs can be used as measures of effects ranging from symptoms of a condition to a variety of impacts, some of them proximal and some of them more distal.

Inclusion of HR-QOL and other PROs in the medicines approval process has been evident at a disease-specific level to varying degrees for a number of years. In the US, a review and analysis of PRO endpoints as reported in clinical study descriptions in approved product labelling of new medical entities (1997–2002) concluded that PROs had a significant role in the development and evaluation of new drugs, but that more formal guidance on the use of these measures was required from the US FDA.[1] Several reviews of disease-specific guidelines from the European Medicines Agency (EMA) during a similar period (1995–2003) concluded that PRO assessments were welcomed, but that more detailed guidance and a more systematic approach to the assessment and interpretation of HR-QOL was required.[2–4] Differences in terminology and recommendations were noted between the FDA and the EMA that highlighted a need for closer collaboration between the two agencies.[2]

In 2006, new draft guidances were issued by both the FDA and the EMA, describing the contribution that PROs can make to support the case for the approval of a new medicine in the respective regulatory environment. In the US, guidance on the use of PROs was issued in draft form in February 2006[5] and in its final form in December 2009;[6] in Europe, the EMA reflection paper on the use of HR-QOL in the evaluation of new products was issued in July 2005 and became effective from January 2006.[7]

Fundamental to these guidance documents, and defined specifically in the FDA guidance, PROs to be measured in any given therapy area should be included within an ‘endpoint model’, which defines in broad terms the hierarchy of endpoints necessary to demonstrate treatment benefit.[6] The endpoint model defines the primary endpoints required for the indication and secondary endpoints for supportive concepts. Depending on the condition, symptom improvement and other function benefits reported by patients can be either primary or secondary evidence of treatment benefit. For example, the endpoints used to evaluate treatment benefit in primary insomnia (i.e. difficulty in sleep initiation and maintenance, or lack of restorative sleep associated with impairments of daytime functioning) range from polysomnography to self-reported time to sleep onset, wake time after sleep onset, total sleep time, daytime symptoms (e.g. tiredness, sleepiness), daytime impact on activities and mood, and interactions with others. Symptoms and, less often, functional impacts are used as primary endpoints in phase III studies. However, other PRO endpoints, such as HR-QOL, may be included as secondary endpoints and may be regarded as additional evidence of treatment benefit that may be considered for inclusion in the labelling.

The relative importance of PROs as factors in the decision to grant approval for a new product can be difficult to evaluate from publicly available documents. As labelling displays “the information the most useful to prescribers in treating patients,”[8] study endpoints (primary clinical or otherwise) are not usually mentioned in the ‘indication’ section of the product labelling,[5,9] except for treatment used for symptom relief. However, data from approved PRO endpoints may appear in the ‘clinical studies’ section of the Full Prescribing Information (FPI) in US product details and in the ‘pharmacodynamic properties’ section of the Summary of Product Characteristics (SmPC) in Europe (section 5.1). These sections both describe the main results of the clinical trials. In the FPI, they summarize the evidence supporting effectiveness in subjects who were studied, including statistically and clinically meaningful effects on prospectively defined endpoints (but not whether the end-point was identified as primary or secondary).[5] In the SmPC, the section includes concise information relevant to prescribers such as the main results (statistically compelling and clinically relevant) regarding pre-specified endpoints or clinical outcomes.[9]

The objective of this research was to review the extent to which HR-QOL and other PROs have played a role in medicines approval and labelling since 2006, when the FDA issued its draft guidance on the use of PRO measures and the EMA issued its reflection paper on HR-QOL.

Methods

This study reviewed the role of HR-QOL outcomes and other PROs in the US and European regulatory processes through a systematic manual review of therapy area-specific regulatory guidelines (US and EU) and product labels issued during the period 1 January 2006 to 16 November 2010. In Europe, we focused our analysis on the products approved and guidelines issued by the EMA, which is responsible for the scientific evaluation of medicines developed by pharmaceutical companies for use in the EU through the centralized procedure.

We adopted the definition of PROs as stated in the final FDA guidance: “A PRO is any report of the status of a patient’s health condition that comes directly from the patient, without interpretation of the patient’s response by a clinician or anyone else,”[6] and the definition of HR-QOL given by the EMA in its reflection paper: “In the context of drug approval, HR-QOL is considered to represent a specific type/subset of PROs, distinguished by its multi-dimensionality. Indeed, HR-QOL is a broad concept which can be defined as the patient’s subjective perception of the impact of his disease and its treatment(s) on his daily life, physical, psychological and social functioning and well-being.”[7]

For this review, we excluded products for which evaluation was based only on composite outcomes mixing physicians’ ratings, patients’ ratings and biological markers. Guidelines relying only on observer ratings or composite endpoints (as above) were excluded as well.

All documents were read individually by two independent raters. Discrepancies were discussed until agreement was reached.

Regulatory Guidances

We researched clinical guidance documents released by both agencies during the study period and identified those that included recommendations for patient-reported assessments.

In the US, we analysed the FDA guidances from the Center for Drug Evaluation and Research and the Center for Biologics Evaluation and Research, but not from the Center for Devices and Radiological Health. Guidance documents represent the FDA’s current thinking on a particular subject and do not create or confer any rights for or on any person and do not operate to bind the FDA or the public.

In Europe, we analysed the EMA scientific guidelines, in particular the clinical efficacy and safety guidelines. Only the draft and adopted guidelines were reviewed (concept papers and points to consider were not included). Guidelines are intended to provide a basis for practical harmonization of the manner in which the EU member states and the EMA interpret and apply the detailed requirements for the demonstration of quality, safety and efficacy contained in the community directives. They also help to ensure that applications for marketing authorization are prepared in a manner that will be recognized as valid by the Agency.

Labelling Containing Patient-Reported Endpoints

The number and type of PROs mentioned in the labelling for products approved during the study period were identified from review of the PROLabels database. In addition, we compared the FDA and EMA approvals for the same product: to document patterns in the number and types of PROs included in approved product labels and to determine whether the nature of the approvals reflects recent FDA and EMA guidances on assessment of PRO and HR-QOL endpoints.

PROLabels (www.mapi-prolabels.org) is an online database that tracks PROs included in FDA-and EMA-approved medical product labels, in the FPI and the SmPC, respectively. It covers PRO information available through regulatory agency websites about PRO labelling claims (1995 to present for the EMA; 1998 to present for the FDA). All original new drug applications (NDAs) and biologic labelling applications (BLAs) approved by the FDA, as well as all supplemental NDAs and BLAs, are reviewed as soon as they are published on the FDA website. A similar process is followed for EMA approvals.

New or revised labels are reviewed to identify PRO data reported in the description of treatment efficacy results. Information is extracted from the label and from medical regulatory reviews (if available) and the European Public Assessment Reports (EPARs), including the SmPC and the scientific discussion (when available) to document the methods used to develop the PRO endpoints, details regarding the instrument used (e.g. domains, name, recall period, response scale), therapeutic class of the medical product, mechanism of action, therapeutic area (Medical Subject Headings [MeSH] classification), and general (MeSH term) and specific (reported in label) therapeutic indications. In addition, information about the application is recorded, including regulatory agency, reference number, marketing authorization holder/sponsor, date of drug approval and date of revision of the label. Further information about the review criteria and information collected in the PROLabels database has been published elsewhere.[10–12]

Results

Regulatory Guidances

Between 1 January 2006 and 16 November 2010, 15 and 34 guidance documents were released by the FDA[13–27] and the EMA,[28–61] respectively, containing recommendations for the inclusion of one or more PRO endpoints in clinical trials. Table I presents a summary of the data, and supplementary table I (see the Supplemental Digital Content 1, http://links.adisonline.com/PMZ/A3) lists the guidance documents identified. These guidance documents can be sourced in full from the websites of the FDA[62] and EMA.[63] Additional web links to the individual guidance documents and extracts providing details of specific recommendations regarding PRO or HR-QOL instruments to be used are also included in the references.

Table I
figure Tab1

Patient-reported outcome (PRO) endpoints recommended in US FDA or European Medicines Agency (EMA) guidance documents issued between 1 January 2006 and 16 November 2010

Full size image

These guidance documents reflect 39.5% of the total number of clinical guidance documents published in draft or final form by the FDA and 35.8% of the total draft and adopted guidelines published by the EMA during that time.

Explicit reference to the HR-QOL reflection paper was made in five EMA guidelines (14.7% of all guidelines; 22.7% of the guidance documents referring to HR-QOL).[40,41,46,50,53] Nine (60%) of the FDA guidelines referred to the guidance on PRO measures.[13,17,18,20–23,26]

The clarity of the guidances regarding the way in which PROs should be evaluated varied across therapy areas, ranging from recommendations of specific validated instruments to general acknowledgement that PROs might be useful, but without recommendation of a suitable measure to use (see supplementary table I in the Supplemental Digital Content). Nine of the 15 FDA guidances and 26 of the 34 EMA guidances recommended using a validated instrument, although a smaller number specified a named validated instrument (3/15 FDA; 13/34 EMA).

The FDA referred to HR-QOL specifically (as a secondary endpoint) in 3 of the 15 (20%) guidances mentioning PRO endpoints —for chronic obstructive pulmonary disease (COPD), oncology and weight management.[15,17,19] The EMA recommended use of HR-QOL endpoints in 22 of the 34 (65%) guidances mentioning PRO assessments.[28–31,33,35,37–41,44–46,48–50,52,54,56,58,60]

Indications and disease areas differ slightly between the agencies (see figure 1). In only two cases did the guidelines refer to the same indication: i.e. COPD and weight management. Both agencies advise the use of HR-QOL as a secondary end-point in the latter, but differ regarding the use of HR-QOL endpoints in COPD (see the Comparison of Claims with Guidance Documents section).

Fig. 1
figure 1

Number of US FDA and European Medicines Agency (EMA) guidance documents published between 1 January 2006 and 16 November 2010 that contained recommendations for the inclusion of one or more patient reported outcome endpoints in clinical trials, grouped by disease area.

Full size image

The FDA never recommends HR-QOL as a primary end-point, whereas the EMA recognizes that it can be used as such in three specific and restricted indications: in cystic fibrosis,[50] COPD[58] and haematological malignancies.[56]

In cystic fibrosis, it is clearly stated that: “A claim of improvement in QoL (or lack of deterioration) would be acceptable only in section 5.1 of the SmPC. Such a claim should be supported by studies specifically designed to demonstrate a HRQL benefit, with HRQL assessment as primary endpoint.”[50]

In haematological malignancies, HR-QOL is considered as a co-primary endpoint in the case of a treatment administered without curative intent where the prognosis is poor and where only short-term disease control is expected (palliative therapy): “In a study conducted with BSC [best supportive care] as reference therapy, the objective should be to demonstrate prolonged OS [overall survival] and/or improved symptom control or quality of life (QoL). The latter requires that the study is conducted under proper double-blind conditions.”[56]

As for COPD, HR-QOL is only envisaged as a co-primary endpoint with lung function.[58]

In these three cases, HR-QOL is never envisaged as a sole primary endpoint, and the claim will only be acceptable in section 5.1 of the SmPC.

On the other hand, there are indications where the EMA clearly specifies that HR-QOL measures are not fully established (migraine[36] ), or validated (insomnia[51] ), or that the use of HR-QOL as a primary efficacy variable in pivotal studies is discouraged (Parkinson’s disease[42] ).

Label Claims Containing Patient-Reported Outcome Measures

During the study period, the FDA approved 93 products (of 432 total approvals) with label claims that included PRO end-points (table II). Of those, eight products (8.6% of all products with a PRO claim) documented treatment benefits characterized as HR-QOL. The vast majority of the PRO claims reflected treatment benefit measured by patient-reported symptoms (104 claims representing 79 products, corresponding to 85% of all products with a PRO claim approved during this period).

Table II
figure Tab2

Patient-reported outcomes (PROs) included in the Full Prescribing Information or Summary of Product Characteristics (SmPC) for products approved by the US FDA or the European Medicines Agency (EMA) between 1 January 2006 and 16 November 2010

Full size image

During the same period, the EMA approved 54 products (of 248 total approvals) that included PRO endpoints, of which 16 products (29.6% of all products with a PRO claim) reflected HR-QOL data (table II). Like the FDA, the vast majority of the PRO claims reflected treatment benefit measured by patient-reported symptoms (i.e. 64 claims representing 48 products, corresponding to almost 89% of all products with a PRO claim approved during this period).

Thus, the inclusion of HR-QOL data in approved product documents was >3 times more common in EMA than FDA approvals, although both agencies reported PRO data in roughly 21.5% of labels for new products approved.

Since 2006, the number of products approved by the FDA with an HR-QOL claim has steadily decreased. As for the EMA, the number of products with HR-QOL cited in section 5.1 of the SmPC showed a rise in 2007 (12%) but has remained stable since then (around 6%).

Figures 2 and 3 summarize PRO and HR-QOL claims by broad therapeutic areas. In the US, HR-QOL claims were dominated by respiratory conditions (four out of eight; however, PRO claims in general are more widely distributed across therapeutic areas, as seen in figure 2).

Fig. 2
figure 2

Summary of US FDA patient-reported outcome (PRO) and health-related quality of life (HR-QOL) claims in product labelling for agents approved between 1 January 2006 and 16 November 2010, grouped by broad therapeutic areas.

Full size image
Fig. 3
figure 3

Summary of European Medicines Agency patient-reported outcome (PRO) and health-related quality of life (HR-QOL) claims in product labelling for agents approved between 1 January 2006 and 16 November 2010, grouped by broad therapeutic areas.

Full size image

In Europe, HR-QOL claims represent a broader range of conditions and indications, dominated by six HR-QOL claims for anaemia treatments. In contrast, PRO claims in general are more widely distributed across therapeutic areas, as seen in figure 3.

Comparison of the FDA-and the EMA-Approved Labels

Although each agency granted a number of claims that included PROs, there were few products that had PRO claims from both agencies. During the time period studied here, the FDA and the EMA granted 12 compounds (24 products) approval for PRO content in the product label (see supplementary table II in the Supplemental Digital Content).

Out of these 12 compounds, three present labelling claims that differ according to the agency: fosaprepitant (Emend® or Ivemend®) for chemotherapy-induced nausea and vomiting, golimumab (Simponi®) for ankylosing spondylitis, and varenicline (Chantix™ or Champix®) for smoking cessation.

The impact of nausea and vomiting on patients’ daily lives as measured by the Functional Living Index —Emesis (FLIE) is clearly indicated in the labelling claim of Emend® (FDA), while there is no mention of such impact in the section 5.1. of the SmPC of Ivemend® (EMA), although the FLIE was administered to patients in the studies submitted to the EMA.

Results of this evaluation are fully described in the scientific discussion of the EPAR dated 5 February 2008 and its variation published on 24 September 2010.[64,65]

In the scientific discussion of 2008, the assessment is clearly in favour of Ivemend®: the results of the combined analysis of studies P052 and P054 show that a higher proportion of patients receiving the aprepitant regimen reported minimal or no impact of nausea and vomiting on daily life (74.4% vs 63.9% —table V of the report[64,65] ). In the Ivemend® —H-743-X-06 : EPAR —Assessment Report —Variation it is clearly stated that “Also with respect to this outcome measure, it appears reasonable to conclude that overall non-inferiority has been shown, even though trends towards worse results in the vomiting specific domain have been noted.”[64,65] These outcomes are quite similar to those reported in the studies submitted to the FDA. Unfortunately, no reasons are given for the omissions in the EMA SmPC.

Simponi® is another example of differences between the EMA and the FDA. The EMA SmPC for the indication of ankylosing spondylitis describes assessments of physical function and HR-QOL. Those PROs are not even considered in the studies submitted to the FDA. It is interesting to note that the EMA SmPC stipulates that HR-QOL is measured by the physical component of the Short Form 36 (SF-36) Health Survey Questionnaire. This seems to be in contradiction with the statement of the EMA reflection paper on the use of HR-QOL, which indicates: “ … the notion of multidimensionality is a key component of definition of HR-QOL. A single domain, e.g. physical functioning or fatigue, is not considered as a HR-QOL (i.e. it cannot be the basis for a claim for a global HR-QOL improvement), even though it is a patient-reported.”[7]

It could be argued that the physical component of the SF-36 is determined by four dimensions. However, these four dimensions only cover physical aspects of health.

As for the FDA Chantix™ labelling, the major difference with the EMA lies in the use of a patient’s self-report to measure abstinence from smoking as a primary endpoint verified by measurement of exhaled carbon monoxide (CO <10 parts per million) at weekly visits. In addition, unlike the EMA, craving is not indicated in the FDA label, although it is assessed as a secondary endpoint by the same PRO instrument (i.e. the Brief Questionnaire of Smoking Urges [QSU-Brief]). Reasons for this decision are found in the FDA Medical Review: “Varenicline’s effects on « craving » and « urge to smoke » both appear to be measuring the same construct, probably described, per Dr. Scott, as “urge to smoke.” However, a consistent effect on the instruments and questions measuring this concept appears to be demonstrated.”

Comparison of Claims with Guidance Documents

Comparing claims with guidance documents, our analysis of the PRO claims issued shows that for the FDA, not a single HR-QOL claim was in a condition having an HR-QOL recommendation in the guidelines issued between 1 January 2006 and 16 November 2010 (table III). Thus, we could not see a pattern between HR-QOL recommendations in guidances and actual label content reflecting these concepts. For the EMA, we found four conditions in which there was a match between the products approved and the guidances issued: psoriatic arthritis, ankylosing spondylitis, insomnia and multiple sclerosis, corresponding to three different products: Circadin®, Extavia® and Simponi® (table III). One product (Circadin®) was approved before the publication of the corresponding guidance. The analysis of the SmPC revealed that chosen endpoints were similar to those advised in the insomnia guidance; this is consistent with the guidance as a description of current practices. Extavia® and Simponi® (for the psoriatic arthritis indication) were approved after the publication of their corresponding guidance, and the endpoints chosen are in line with those recommended in their respective guidance documents. As for ankylosing spondylitis, the almost simultaneous dates of guidance publication and product approval (Simponi®), plus the concordance of endpoints, suggest the recognition of existing practices within the guidance.

Table III
figure Tab3

The US FDA-and European Medicines Agency (EMA)-approved products including a health-related quality of life mention in their label and corresponding guidance (1 January 2006 to 16 November 2010)

Full size image

Overall concordance would not be expected yet, since trials take years to complete and the earliest guidance we reviewed was dated 2006. Trials that specifically considered each guidance may still be continuing and the data not yet available.

Discussion

Our review of guidance documents and products with an approved PRO label indicates that representing the patient’s perspective in clinical research is important for both the EMA and the FDA. However, HR-QOL endpoints play a minor role in the approval of drugs. Although the percentage of products with an HR-QOL mention in the label is >3 times higher for the EMA (29.6%) than for the FDA (8.6%) among all products with a PRO label, when compared with all drug approvals, the percentages are quite low (FDA 1.9%, EMA 6.4%). Among guidances that included any PRO, the FDA recognized the relevance of HR-QOL endpoints in only 20%. In contrast, the EMA acknowledged the importance of HR-QOL endpoints in 65% of guidances that included any PRO endpoints. Those findings suggest that receptivity to HR-QOL concepts is higher in the EMA, while the FDA seems to prefer symptom-based data.

There may be several reasons for the differences observed between the percentage of the FDA and EMA guidelines re-commending HR-QOL endpoints and the number of labels including HR-QOL for the corresponding therapeutic area: (i) the reluctance of regulators, particularly in the US, to allow labels including HR-QOL wording; (ii) the lack of documentation and sometimes quality of the HR-QOL data, including the non-compliance with current standards for evaluating PROs in clinical trials; and (iii) the decision of sponsors to limit the end-points to critical ones to get the product approved.

Willke et al.[1] reviewed and analysed PRO endpoints as reported in clinical study descriptions in product labelling of new molecular entities approved in the US from 1997 through 2002. They found that PRO endpoints appear in 30% of all labels (64 products) and that 22 of the 23 drugs approved only on PRO endpoints were approved on direct patient symptom reports. The latter is in concordance with our finding about FDA preference for symptom-based data. The percentage decrease in the use of PRO endpoints (in our review, PRO endpoints represent 21.5% of all FDA approvals) might be explained by the approval of products belonging to different therapeutic classes in Willke’s review (e.g. anti-inflammatory, urological, ophthalmical, gastrointestinal or allergic).

It is interesting to note that Gondek et al.[66] came to a similar conclusion about EMA receptivity to HR-QOL, although their study was focused mainly on oncology. In another study in oncology (a review of PRO supporting FDA anticancer approval from 1995 to the end of 2006), Rock et al.[67] showed that nine treatment indications have been approved for seven anticancer products based either on symptom palliation or improvement in a functional endpoint. These approvals represented only 10% of all treatment indications approved. HR-QOL was used only once (gemcitabine for non-small cell lung cancer).

In the research conducted by Szende et al.,[4] all published EMA guidance documents and regulatory information for products authorized at the EMA and appearing in the EPAR database between 1995 and 2003 were examined for reference to HR-QOL and other PROs. They found that 53% of the guidances retrieved included recommendations on PROs. However, they included ‘concept’ papers and ‘points to consider’ in their review, which we did not. When those are deleted, the percentage rises to 58% (21 out of a total of 36). Our review identified 34 guidelines out of a total of 95 (35.8%). When we examined the list of guidelines already including PROs and compared it with the indications retrieved in our review, we discovered six matches: HIV infection, migraine, multiple sclerosis, Parkinson’s disease, stable angina pectoris, and weight control. Interestingly, at that time, ankylosing spondylitis was not seen as a potential candidate for PRO recommendation (2003 concept paper). Fifteen of the 21 draft and adopted guidelines retrieved included an HR-QOL recommendation (71.4%). Our review found that up to 65% of the guidances included any PRO endpoints. All these difference might be explained by different disease areas covered by the 1995–2003 guidelines (only six matches were found to those covered in our review).

As for the product-level information, Szende et al.[4] identified 81 products (34% of those submitted) that included HR-QOL or other PRO information. Most of them were antineoplastic agents (32%).[4] In our review, we identified 54 products with PRO information representing 21.8% of all drug approvals. Most of them had an indication in respiratory, neurological, haemic and rheumatological diseases. In the Szende et al. review, HR-QOL information was included in 55 products (66.9%).[4] Our review identified 16 products, representing 29.6% of all products with a PRO claim. Again, the difference might be explained by the differences in therapeutic classes of approved products.

Papanicolaou et al.[3] reviewed the European guidelines on how HR-QOL research should be conducted in clinical trials. Published product-level information through EPARs of all approved drugs was also reviewed to investigate the actual role of HR-QOL data in the European regulatory process. From 1995 to 23 March 2003 they identified 20 of 50 guidances including reference to, or recommendations for, HR-QOL. When we deleted all ‘points to consider’, we found reference to 13 draft or adopted guidelines (15 in the Szende et al. review[4] ), with two draft guidelines not included: panic disorder and psoriasis. The main difference from the Szende et al. review is a severe criticism of the terminology used: “In general, most of the recommendations that referred to the measurement of HR-QOL were generic and vague, using nonstandard terminology that was also inconsistent across the documents.”

Based on our review, our judgement is less severe, and we believe that the publication of the HR-QOL reflection paper as well as training on HR-QOL issues among experts involved in the design of the EMA guidance had an impact on the clarity of the guidance documents issued between 2006 and 2010. The guidelines about osteoarthritis, irritable bowel syndrome, COPD and Crohn’s disease, which were only points to consider in 2003, were considerably revised and clarified when their status changed to draft or adopted. On the other hand, new guidelines in gastroesophageal reflux disease, premenstrual dysphoric disorder and ankylosing spondylitis are very detailed, clear and precise.

Differences between the EMA’s and FDA’s appreciation and preference of PRO endpoints, especially HR-QOL, might be explained by their differences in structure and organization. As the world’s largest medical regulatory agency, the FDA employs an ‘army’ of medical and methodological experts to review submissions and develop guidance. This internal expertise enables the FDA to take a very active role in defining endpoints that are needed to evaluate treatment effects, developing guidance and regulations that clarify the FDA expectations, and enforcing regulatory policy. The statutes that the FDA enforces require that product claims be supported with well defined and reliable clinical trial endpoints. The FDA’s most recent position is that PRO endpoints must be informed by an understanding of the condition that is grounded in systematic qualitative research with patients who have the condition that will be studied in the trial. This helps ensure that PRO assessments used to measure endpoints provide accurate data because they are easy to understand and relevant to the patients asked to complete them. Using their internal expertise in measurement and clinical matters, the FDA conducts detailed reviews of existing and newly developed PROs to ensure they have content validity and that claims based on the end-points are neither false nor misleading. Finally, as part of its Critical Path Initiative, the FDA has recently released a draft guidance on the Qualification Process for Drug Development Tools (DDTs).[68] The guidance provides a framework for how drug developers and manufacturers may submit and seek qualification approval for tools that may have only been used in one instance, but could potentially serve as platforms for accelerating development activities. Biomarkers and PRO instruments are the main tools considered in this guidance. Knowing that a certain PRO is qualified in advance with the FDA as a DDT can accelerate trial and approval work.

By contrast, the EMA, working with a network of over 4500 European experts, relies on clinicians and the scientific community for the definition of validity and the acceptance of an endpoint. In addition, experts involved in the development of guidelines or review of approval submissions are often clinicians still active in their countries with, for some of them, daily contact with patients. Historically, the EMA has been more likely to accept existing measures, including global assessment and diaries, provided that the assessments are supported by peer-reviewed publications of the development and validity of the instruments. Thus, the peer-review process and the acceptance by the scientific and medical communities greatly influence the likelihood of PRO claims in the EMA product approvals.

The FDA’s apparently greater interest and detailed review of PRO endpoints in clinical trials and product labels may be driven, in part, by the potential use of such information in direct-to-consumer advertising. Clearly, direct-to-consumer messages based on PRO data can be very compelling to the healthcare consumer, offering a promise of treatment benefits in terms that mean the most to patients themselves: i.e. how they feel and what they are able to do. Anticipating such promotion based on the content of the package insert or drug labelling, the FDA undertakes an extreme level of care in determining whether and which PROs may serve as clinical trial endpoints that are meaningful and interpretable by patients.

What of the representation of HR-QOL claims in proportion to other patient reports, such as symptom assessments? We believe that two forces are at work here, particularly in the US. First, the FDA has set a very high hurdle with its definition of HR-QOL and requirements for proving an improvement —acknowledging the concept is multidimensional and potentially abstract. It is these qualities that may make the HR-QOL claim of limited usefulness in communicating treatment benefit. Second, clinicians recommend interventions to improve specific signs or symptoms of a condition.

Bottomley et al.,[69] in their review of the FDA PRO guideline and the EMA HR-QOL reflection paper, considered that these documents were “important steps towards the acceptation and appreciation of patient viewpoint, and the creation of significant evidence in the drug approval process.” However, they deplored the stringency of the FDA for well established PRO or HR-QOL measures (e.g. the European Organisation for the Research and Treatment of Cancer Quality of Life Questionnaire [EORTC QLQ] C30) with a long history of effective use and significant evidence of real-world validity.

There are several limitations to our study. First, it takes time to see the effect of guidance in the results of clinical trials and thus claims, for the simple reason that it takes years to run a trial. Therefore, effects of the 2006–10 published guidances are not fully seen. Second, the differences cited between the EMA and the FDA may simply be because of differences between drugs that were approved during the limited time period of the study and may not reflect any tendency for the EMA to approve more HR-QOL endpoints.

As for the relatively low percentage of products with PRO claims (around 21.5% for both agencies), it is difficult to draw conclusions. We cannot determine whether this percentage is just right or too low. We may wonder whether PRO evaluation for all the drugs without PRO claims (78.5%) was judged not relevant or not even considered. However, we agree with Doward et al.[70] who, in a recent paper, aptly pointed out that PROs should be considered beyond the label claim and advocated regulators to consider patients’ views with more dedicated attention: “The FDA may be unwilling to consider PRO data beyond first order impacts (signs and symptoms). However, it is clear when talking to patients and patient groups that such concerns are often of minor concern to their determination of the impact of disease and the effectiveness of treatments. Patients have very real ideas about what states of physical and emotional well-being (and ultimately QoL) are acceptable and may not always agree with clinicians and regulators on whether treatments are beneficial.” Doward et al. also recognized the importance of health payers and agencies involved in reimbursement and pricing decisions, who should be considered as key players in the future.[70] They may have an influence in the extension of PRO evaluation to new therapeutic areas or in the identification of broader PRO outcomes.

Conclusions

Our study shows that representing the patients’ perspective in clinical research is important for both the EMA and the FDA, with broad HR-QOL endpoints playing a minor role in product claims. Our analysis suggests that the receptivity of the EMA to HR-QOL endpoints is greater than the FDA’s, and that both agencies value patient-reported symptom data. Our results are comparable to previous analyses within an acceptable margin, with the main differences observed due to products approved in different therapeutic classes or guidances published in different therapeutic areas. We showed that the generic guidances published on PRO measures (FDA) and HR-QOL (EMA) are acknowledged in disease-specific guidelines, and that guidelines issued by the EMA during our period of analysis have gained in precision and clarity compared with previous periods.

As for the future of PRO evaluation, we believe that the shift toward patient-centric healthcare and the empowerment of patients in healthcare decision-making through an increase in access to information and active advocacy groups will foster the development of specific PRO endpoints relevant to them for evaluating treatment benefits, in addition to traditional clinical endpoints.