Potential solutions to improve the governance of multicentre health services research

In reply:

There is much on which we agree with Evans and colleagues’ letter on our article.1 We do need streamlined ethics and governance arrangements for health care research; national agreement on requirements; standardised forms and a single portal; widespread understanding of what constitutes negligible, low and high risk research; and a shared mental model for ethics committees — supported by a national training strategy. Research should be accountable, but the current system is excessively cumbersome, with millions of dollars of scarce research funding spent obtaining approvals and clearances. Our case study1 is merely the most recent exemplar.

Our departure point from Evans and colleagues is on the requirement that each participant site should have a local principal investigator who must be an employee (and on the further requirement that external researchers must be employees, albeit honorary). This requirement introduces unnecessary bureaucracy and unwarranted bias if the only people who can perform research on health services are employees. It also impedes the capacity to standardise a research protocol across multiple sites.

Experience and organisational studies tell us that employees, even independently minded ones, are almost always compromised.2 They are under subtle (or often, explicit) organisational pressure to conform to corporate wishes to be seen in a favourable light. We have yet to meet a teaching hospital, for example, that does not describe itself in glowing terms as a “centre of excellence” — even ones that fail accreditation or are under review for poor performance. All organisations impose loyalty demands of some form or another on their staff.3

We do need bridges between researchers and local employees, and it is important that researchers are approved by the local site, but insisting that a local principal investigator is the principal researcher, or mandating that any external researcher must be a deemed staff member, risks jeopardising objectivity. Readers of the research, once published, might worry that it is not research at all, but simply another version of organisational spin. For these powerful reasons, sites hosting research should be facilitators of external researchers, not employers of them.