Scientific research and the public good

Randomized clinical trials are the state-of-the-art method to study comparative effectiveness of healthcare interventions. However, sometimes a discrepancy is observed between the effects of an intervention in routine clinical practice as compared with the effects that were demonstrated in randomized controlled clinical trials.¹ Real-world data, e.g. from registries or claims databases, could help to unravel the effectiveness of interventions in clinical practice. The hope is that the use of these data will allow healthcare decision makers to be more certain when it is (cost-)effective to provide access to new treatments. Real-world data could therefore contribute to the public good.

Recently, the Dutch National Health Care Institute hosted an international workshop on the IMI-GetReal project. GetReal is a 3-year project of the Innovative Medicines Initiative (IMI), an EU public-private consortium consisting of pharmaceutical companies, academia, Health Technology Assessment (HTA) agencies and regulators, patient organizations and small and medium enterprises (SMEs).² Representatives of these parties got together to discuss the use of real world data to determine the comparative effectiveness of drugs and other medical interventions.

Two participants in the workshop, who represented public HTA agencies, presented the findings of a case study in which they had attempted to combine data from registries in three different European countries. Or rather, they presented the absence of findings, because despite putting in a lot of effort, following the right procedures and demonstrating their willingness to cooperate with the researchers working on the registries, they had not succeeded in gaining access to the data. In the case of one registry, the main argument used to decline access to data was that this would interfere with ongoing PhD-studies of junior researchers affiliated with the registry. The discussion that followed in the workshop revealed that this was not a stand-alone incident, but a response that several participants recognized and had encountered themselves when trying to get access to real world data from registries. In fact, gaining access to individual patient level data is such a common problem that epidemiologists have developed creative methods to ‘re-create’ individual data from published survival tables or curves, or other summary statistics.³

On 19 November 2015 in a letter to The Times, EUPHA-president Martin McKee along with many other stakeholders urged the European decision-makers to abstain from disproportionate limits on the use of personal data in health research, that have been proposed to the EU Data Protection Regulation: ‘It’s critical these restrictions are not included in the final version of the law. Protecting privacy is vital. Strong safeguards and governance structures already exist to help ensure that personal information is used safely, ethically and securely in research. We urge the European institutions to ensure that the Regulation works with these safeguards to strike an appropriate balance that protects the interests of individuals while enabling research that benefits us all.’⁴ However, if we, the collective research community, urge politicians to ensure that we can access individual level data to conduct research that benefits the public, we should be careful not to impose unnecessary barriers to this kind of research ourselves.

Of course, a certain degree of professional competition enhances the quality of research, which is why researchers usually have to compete in order to get funding for their research. They invest personal time and knowledge into the development of grant applications and in order to be successful they have to publish scientific papers. After all, a scientific track record is an important requirement to get access to research funding. And if they are successful and get the grant, they invest time and expertise in the collection of data for one or several research questions. Therefore, it could be argued that they deserve some first hand rights to capitalize on this investment by being the first to publish their findings in scientific journals. Also, secondary use of data must comply with basic principles of confidentiality of personal data, so while the aim for ‘open access to data’ is laudable, it may be not always be easy to implement in health research using individual records.

Obviously, if multiple research teams utilize data from one and the same source, they will have to come to agreements about who is going to publish what, when and with whom. Agreeing on certain embargoes could be a solution. However, access to data for public health purposes—especially to data that have been collected with public money and for public goals—has to be negotiable in principle. Surely, no one wants to spoil epidemiologists' fun in clever and complicated modelling in order re-create individual data, and undoubtedly these methods can be applied to a number of other problems as well. However, there seems to be something fundamentally wrong if public agencies pursuing public goals have to resort to such methods, because Ph.D. theses are prioritized over the interests of patients and the public.