Two medical experts say more research needs to be done on the treatments GPs give patients.

Australian GPs are at risk of treating patients with severely restricted evidence because of federal funding cuts to primary health care research, warn two experts.

In a perspective paper published in the latest Medical Journal of Australia, Professor Tania Winzenberg, from the University of Tasmania, and Professor Gerard Gill, from Deakin University, call for more research to be done on the treatments GPs give patients, which would require more federal funding.

They say studies have shown good primary health care (PHC) systems are linked with better health, fewer deaths and greater efficiency, yet the "plug is being pulled" on federal funding.

"There is a mismatch between the burden of diseases commonly managed in general practice and the number of randomised controlled trials exploring their effective management," they wrote.

An inefficient PHC system leads to overtreatment or less cost-effective treatment, they claim.

Many GP research programs, including the Australian Primary Health Care Research Institute, the Primary Health Care Research and Information Service and the Bettering the Evaluation and Care of Health study, have recently lost or will soon lose funding.

This means there is now less direct funding for PHC research than there has been for decades, say Prof Winzenberg and Prof Gill.

The Royal Australian College of General Practitioners provided a pre-budget submission to the federal government on prioritising GP research, which, if implemented, would go a long way to improving the situation, say the authors.

Among the RACGP recommendations was a $27 million commitment over nine years to establish a general practice research fellowship program.

"In the meantime, we are in grave danger of wasting the investment made to achieve current gains in capacity, leaving our profession and the Australian population to make do with a severely restricted evidence base to support PHC in this country," the authors concluded.