more_vert
A RECENT case of research fraud highlights the need for independent, publicly funded research and enhanced access to raw datasets, according to experts.
An editorial in this week’s MJA, by Professor John Myburgh of the George Institute for Global Health, describes how a German anaesthetist, Joachim Boldt, fabricated a published study on colloids. (1)
After Boldt’s research was published in the journal Anesthesia and Analgesia, letters of concern sent to the journal’s editor prompted an inquiry that found there were no original patient or laboratory data to support the findings and the report was retracted. Subsequently a further 87 of Boldt’s reports in 18 journals were also retracted.
Eleven of the Boldt reports relating to a specific colloid used for fluid resuscitation had been used in submissions to regulatory bodies including the Therapeutic Goods Administration in Australia.
“The magnitude of the Boldt case is unprecedented in intensive care medicine. It ranks with the fraudulent research conducted by Wakefield (measles–mumps–rubella vaccine) …”, Professor Myburgh wrote.
Professor Jamie Cooper, director of the Australian and New Zealand Intensive Care Research Centre at Monash University, said he believed research fraud was rare and the Boldt case was “exceptional”.
However, the case indicated the “enormous pressures” sometimes involved in industry-funded research, Professor Cooper said. Boldt had reportedly received research funding from several colloid manufacturers.
Professor Cooper said that having more publicly funded research might help protect against research misconduct. “I think one of the answers is that we need to put more government money into independent research, which is not subject to these pressures from industry”, he said.
Ethics committees at universities or research institutes could also take more steps to protect against fraud or misconduct, such as ensuring that all trials were subject to independent monitoring, Professor Cooper said.
Journals could also impose more rigorous standards such as requiring receipt of all versions of the research protocol submitted to the ethics committee.
The Australian Government recently took steps towards protecting against research misconduct with the establishment in January of the Australian Research Integrity Committee. (2)
The committee can review institutional investigations into research misconduct upon request, however it has not undertaken any reviews to date.
There is also a move within Australia and internationally to enhance sharing of raw datasets from clinical trials.
MJA editor Dr Annette Katelaris writes in the latest journal that sharing data could help deter fraud, as well as facilitate research by allowing testing of secondary hypotheses and simplify data collection for meta-analyses. (3)
In January, Professor Warwick Anderson, chief executive officer of the NHMRC, signed a joint statement with 16 other international health research funders, committing to work towards sharing data to “unlock the full potential of research”. (4)
Dr Tim Dyke, executive director of governance and strategy at the NHMRC, said sharing raw data would also help prevent research misconduct, which undermined public trust in science and research and had implications for patient safety.
“The enhanced transparency would certainly help because people could check results against the raw data and determine whether they’re fabricated,” he said.
Dr Dyke said the NHMRC would work with the research community to determine the best mechanisms for data sharing.
Dr Suzanne Morris, research integrity officer at the University of Queensland, said universities around Australia were discussing the possibilities and practicalities of data sharing.
“I think there’s a lot of merit to making data more accessible, subject to confidentiality and intellectual property concerns”.
Dr Morris said institutions needed to support researchers to share their data, such as by developing data-sharing tools.
– Sophie McNamara
1. MJA 2011; 194: 621-622
2. Australian Research Council. New independent committee to assure research integrity.
3. MJA 2011; 194: 619
4. Wellcome Trust. Global funding agencies commit to enhancing access to research data to improve public health.
Posted 20 June 2011
I agree with Professor Loblay. Persons willing to publish fraudulent studies would easily be able to fabricate data to support them. Many of the fraudulent studies that I recall have been undertaken by research fellows who are under tremendous pressure to make a mark in the year or so allotted to them. Their supervisors are very willing to share the credit for “breakthroughs” but then plead ignorance of the fraud. Supervision needs to be tighter.
Even public funded (Grants) research is under incredible pressure to produce results, usually in the form of LPU* and the stress can have predictable results. So while academia is kept under such stress I suspect we will have more and more fraud.
Dr Phil (PhD, BMBS)
* LPU= least publishable units, which is the absolute minimum required to get a paper published. It is worth more academically to publish 20 mediocre papers than one really good one. So it is even harder to read the literature for good quality research.
If Tim Dyke at NHMRC thinks sharing ‘raw data’ will help prevent research misconduct, he is seriously misguided. People who fabricate data are generally very clever about covering their tracks. In most cases that have come to light, even those working in the same department or laboratory, with full access to ‘raw data’, are usually not aware of the fabrication until suspicions are raised by a student or research assistant who is intimately familiar with the study details. Painstaking investigation by careful examination of lab books or patient records is usually required to uncover the fraud. By contrast, the current interest in ‘sharing of raw data’ from clinical trials does not involve examination of patient records or signatures on consent forms; it involves access to electronically recorded ‘raw data’ on which others can perform independent statistical analysis. This may certainly reveal important information about reporting bias, but is very unlikely to uncover fraud or outright fabrication.
Isn’t the issue about the conduct of the research rather than than the source of funding? Presumably the industry sponsor of this research will be just as upset as everyone else about this.
More useful may be a process to confirm that the many companies undertaking conduct and monitoring of research are indeed qualified. Rather than increase the workload of our publicly funded committees, it may be better to require that the research be monitored/conducted by a “certified” study company, ie. clinical research cannot be conducted in house.