Difficulties in accessing patient data on effects of heart failure drugs

Lack of access may lead to "erroneous clinical decisions” say researchers

In The BMJ this week, a team of researchers describe difficulties in trying to obtain patient data on the effects of heart failure drugs from previously published clinical trials. As a result of these difficulties they could not undertake a review to assess the outcomes of these drugs on patients.

The UK team, led by Robert Fleetcroft from Norwich Medical School at the University of East Anglia, argues that "difficulty accessing data from clinical trials means that only part of the evidence base is available" and this may lead to "erroneous clinical decisions."

Current NICE guidelines on drug treatments for heart failure are heavily based on evidence from patients who had severe symptoms. However, most patients have only minor symptoms and so these drugs might be less effective. So the research team carried out a systematic review of the effectiveness of heart failure drugs
 for patients with minor symptoms. They found 30 studies that looked at the effect of these drugs -- beta blockers, angiotensin converting enzyme inhibitors, and angiotensin receptor blockers -- on heart failure.
However, none of the studies included enough data to assess outcomes for patients with the varying degrees of minor symptoms required. So Fleetwood and colleagues set out to request the additional data by contacting the authors of each study.

Authors from only 24 of the 30 studies could be contacted because of difficulties finding up to date
 email addresses. Of these 24, one of the authors had passed away, and three had left their institutions, but contact details were found on the internet for two of the latter. In total, only six authors replied from the 24 authors contacted.

Three authors said that data were not available; one said that only one class of heart failure patient had been included in their study; one author refused the data on the grounds that such analyses were not appropriate and may lead to misleading results; and one author recommended getting in touch with a co-author.  

The team said they were “surprised” at the difficulties they faced, even though the majority of the studies were published between the 1990s and 2010. Four had been published since 2010. Many funders and institutions now recommend data sharing. Benefits of access to patient data include that it can enable researchers to answer new questions with existing data, validate findings, and combine the power from individual studies. It may also prevent selective reporting, and research fraud, explain Fleetwood and colleagues.

The next stage is for mechanisms of data sharing to be developed, they add, and they discuss possible solutions that "will require coordinated actions from funders, journals, ethics, committees, and national guideline developing bodies."

"We need central national repositories for trial data based in the country of trial sponsor," they explain.
Furthermore, they argue that bodies involved in national guidelines have a role to play: "It is unacceptable for NICE to make decisions about new drugs based on clinical effectiveness data that are not in the public domain. Such transparency is essential to bolster trust in the process of evaluation of new treatments."

Lastly, they recommend that research ethics and funding committees should make future access to data a mandatory requirement. Funders should cover the costs of archiving data, they add, and journals could require evidence of archived data.