Skip to NavigationSkip to content

The valuable world of Real World Evidence

Published on 08/12/14 at 07:54am
People image

What do you believe? Are you a doubting Thomas, always wanting to have the pragmatic proof personally presented to you, or do you have the blind faith that if you are told something credible, then you accept it as likely to be true?

In healthcare, doubt has come to be the trump card. Phrases like ‘it is generally accepted’ may not be acceptable to those paying for or prescribing healthcare products, and even our television advertising sometimes resounds with the epithet ‘clinically proven’ (but seldom explaining how). Unbiased proof has become the name of the game, but now the sands are shifting.

I spent a lot of my time in the pharmaceutical industry in clinical research worrying about completing projects that would show that medicines were effective and safe, and worth purchasing. There is no such thing as an easy clinical study, and one spent a lot of time planning for disaster, confounding, or simply the possibility that your product did not work or was not safe.

I remember well a discussion I had with a very senior medical scientist who was labouring on a new inhaled medicine. I was working on a similar project, and I asked him why he had chosen a particular specific group of patients for his pivotal study.

“Well, it was easy,” he said. “I chose those patients who would be most likely to benefit from my product, and excluded those who might not. That way, I was more likely to show that we were better than placebo.”

Accepted study format

What he probably did not know on the day he started the work was that this study would become the gold standard, and forever onwards, only patients like the ones he had chosen would get into like studies and this would be reflected on everyone’s SmPC/pack inserts.

So those patients who subsequently needed treatment who did not fit his criteria were never part of the show and were left out in the cold, or would have to be treated off-license.

In many ways, this shows the weakness of the existing system in that medicines are approved on the basis of limited information on limited types of patients treated only with pre-specified doses and with various other exclusions, sometimes representing only a small fraction of the patient pool.

This is representative of the argument that was put forward by Ben Goldacre in his book Bad Pharma – suggesting that industry sponsored studies may be biased, and are difficult to extrapolate to the population at large.

The shift in thinking is that information from what happens in the real world is of value, as it is more likely to reflect what happens when a medicine is used in a wider, real population that has not been part of a clinical trial selection and allocation process.

What is needed is information about what happens when compliance is not checked, co-medications are not so strictly controlled, visits are not controlled by a timetable and even some bizarre alternative therapy may be going on simultaneously.

Increasingly, according to a report from the ABPI in 2011, there is recognition of the role played by data about patients’ use in normal clinical practice, or in settings better reflecting the reality of healthcare delivery.

World of real evidence

Will your medicine still work in these situations, and is it safe? If you can show that a medicine is a societally improving product, will it make it easier to get reimbursement at the best possible price?

The introduction of value-based pricing needs to be underpinned by robust real world evidence of value for money, which simply cannot be gained from the restrictive arena of limited controlled clinical trials. This is the world of real world evidence (RWE).

Real world data has been defined as data that are collected outside the controlled constraints of conventional randomised clinical trials to evaluate what is happening when a medicine or intervention is used in normal clinical practice.

Real world data includes data on:

• Outcomes (both clinical and patient-reported)

• Resource use (NHS, patient and societal)

• Treatment pathways

• Service models

• Patient preference, experience, and compliance.

This list takes in patient reported outcomes (PRO) and other patient preference scales.

If a PRO is part of a clinical study, then it is prospective, and could not be part of real world data. However, if a clinician applies a PRO or questionnaire routinely as part of patient management, for example an annual review of a chronic illness, then that could become part of the real world data.

In these days of data driven to justify costs, the smart money is on patient registries. In its simplest form, a patient registry is a database that will capture demographic information about the population (national or international) with a particular disease, the interventions used and the important outcomes.

Whilst government health economists can use these data to work out how many patients are being seen by healthcare professionals, and the overall effects of national treatment policies, registries can be interrogated or tweaked to generate answers to specific questions – such as what proportion of patients have received a particular management.

A distinction should be made between prospective use of a registry to collect specific data and after-the-event interrogation. These latter activities are real world data because there is no predetermined intervention to gather the data, except participation in the registry, and the data are all-comers data without specification.

Quality outcome frameworks

Being on a registry may not do a patient any good at all – the benefits are to the users of the registry, and at best, the patient is recognised and has an eye kept on them. Registries can be associated with pay-for-performance (P4P) quality-based contracts for individual doctors, groups of doctors or even all doctors in a country.

For example in July 2006, the New England Journal of medicine reported that the UK rewards clinicians according to 146 quality measures related with 10 chronic diseases that were tracked electronically.

This quality outcome framework is tweaked regularly. One item could be annual review of asthma. This will record patient status in a non-standard way, because you cannot strictly control the way some data are collected, for example how lung function is tested.

The data will be the review data that are required, but within the context of the framework, it is real world data on asthma. Can you see now how real world data might be collected from such a system?

An example of where Registry data provided important input to HTA is the modelling of the cost-effectiveness of TNF-alpha antagonists in the management of rheumatoid arthritis: results from the British Society for Rheumatology Biologics Registry data was used to assess response to treatment over three years, including HRQoL measured using SF-36. This was influential in determining guidance on the use of etanercept and infliximab for treating rheumatoid arthritis.

Some companies and organisation have set up specific registries after the launch of a new product, usually in orphan-like conditions, but not always, or where there are significant restrictions on the patient prescription base.

This may be with the encouragement or agreement of the regulatory agencies who view this as a way to keep an eye on exactly how the product is being used, and creating an accountability between the company, prescriber and patient.

This is understandable if there are risk issues for the treatment.

An example of this is the Fortei (teriparatide (rDNA origin) injection) osteoporosis Patient Registry, a voluntary registry for patients in the US. This is a safety-orientated registry connected to the treatment risk of osteosarcoma. 

Big pharma and real word

Earlier this year in June, Merck/MSD announced the launch of a global registry to evaluate the experiences and outcomes of approximately 20,000 patients with type 2 diabetes in real world settings over three years.

A press release said that the registry is part of Merck’s new global research initiative evaluating outcomes of patients with type 2 diabetes in everyday clinical practice – with the aim of generating real-world evidence to help advance diabetes care.

The registry will be conducted at more than 900 sites across the US, Germany, France and Japan. Led by external experts in each country, it will evaluate blood glucose control, healthcare resource utilisation, medication adherence, quality of life and patient satisfaction.

So it can be seen that patient registries can be a hybrid of information sources, but all aim at collecting population statistics in a way that clinical trials cannot, and reflect real life and a mix of hard and soft endpoints.

Receipt of a particular medication may be conditional upon a patient being in a registry, and an example of this is the olanzapine patient registry which limits the supply of medication to patients enrolled in the registry (and their doctors) and therefore not only catalogues patient outcome, but also doctor training and other support for the medication.

In the current safety environment, these activities can become part of the risk management plan/REMS.

The collection of real world evidence is not a new idea. The real world is full of surprises, and if you set out to collect information of a particular kind, amongst the factors that may come into play are:

• The treatment pathways may change, and what you are collecting at the beginning may not be the same as at the end

• Knowing when the beginning and end of data collection are can be a problem, and you may have to define the magnitude of the study, which, once you have done that, makes it a bit like a prospective clinical study

• Despite your prior research, you may not actually get any patients, or a cohort too small to perform meaningful analyses on

• If you get data, because it has not come from a sponsored study or one with a clear principal investigator, there may be debate about who owns the data.

Why is RWE potentially important?

RWE can provide several kinds of information that randomised controlled trials (RCT) cannot.

The collection of pharmacoeconomic data in RCTs is potentially limited, as the numbers are small, and the treatment of patients may not be representative of normal treatment pathways when the study medication is used in real life – for example the frequency of lab testing, or provision of co-medication for symptom relief.

Whilst RWE may not be the best thing for collecting efficacy data, the methods are better suited to the monitoring of safety information in the marketplace.

Clinical trials, of course, provide an efficacy and safety evaluation in a selected population meeting specific inclusion and exclusion criteria – as well as regulatory requirements, but RWE additionally provides:

• Effectiveness in the real world following market authorisation, both in terms of clinical outcome as well as resource and/or financial aspects

• Information on real world patients in terms of the diversity of patients who will be treated (i.e., outside of inclusion/exclusion criteria), disease progression, social and behavioural aspects, etc.

• Adherence: What is this like in a real world setting – what are the barriers and opportunities, how does a new therapeutic agent help or hinder?

• Real world pharmacovigilance – one of the main drivers currently for collection of RWD of many companies, based on post-authorisation requirements for safety evaluation in real world patients

• Uptake, penetration and barriers to adoption within a market – how does the reimbursement landscape look, what are the resource requirements and usage, clinical workflow?

The future of RWE

Real world evidence is something that has been around for some time, although not always given that specific title. Having said that, when marketing people are presented with this style of data, there may be a sharp intake of breath, whilst they work out how it can be used.

Is it as valid as RCT data for the promotion of medicines? It can fall either way, because whilst not level 1a evidence, it has the almost magical property of being what it is – real.

The caveats that are applied to the claims made from RCT data can be ignored, and providing the context of the collection of this non-interventional data are clearly defined, the two sit alongside each other very comfortably.

I predict that in the future, a lot of thought will go into the ways that real world data can be collected. There are the traditional methods using existing healthcare databases and registries, but there will always be new methods.

The interrelationship between social media and healthcare is an area that cannot be ignored. All of a sudden there is access to users of medical care in unprecedented numbers. The cost of collecting information via web-based platforms is a fraction of RCT data, and the time taken to collect may be much briefer.

There will be a number of issues that need to be consistently resolved. The exact status of ethics and consent in such non-interventional studies will need to be consistently determined, and in the case of patient registries, the ownership of data will need to be clear.

One possible approach is that when patients enter into registries, they give blanket permissions for the use of data, and this is agreed with ethics committees prior to the use of specific data probes.

If the data are not owned by study sponsors, who might normally control the study analysis, how will decisions about the analysis process be made? As there is no GCP style data integrity and paper-trail, there can be no audit – and will this be acceptable all of the time?

Do sponsors have any rights or controls or do they just passively accept that the data are what the data are because once there is a sniff of intervention, it becomes a trial?

Real world evidence is here to stay and represents (in the view of the ABPI) an unprecedented opportunity to engage with and collect big data representative of the way that healthcare is utilised.

Dr Martin Goldman is a freelance consultant to the healthcare industries who has worked in mainstream pharma for many years

Mission Statement
Pharmafile.com is a leading portal for the pharmaceutical industry, providing industry professionals with pharma news, pharma events, pharma service company listings and pharma jobs,
Site content is produced by our editorial team exclusively for Pharmafile.com and our industry newspaper Pharmafocus. Service company profiles and listings are taken from our pharmaceutical industry directory, Pharmafile, and presented in a unique Find and Compare format to ensure the most relevant matches