Reducing the number of beds in facilities for psychiatric patients is linked to an increase in prison populations, according to a new study.

Prison population rose as the number of psychiatric beds decreased since 1990 among six South American countries, researchers found.

“This is clearly not a definitive study that can’t provide cause and effect findings,” said the study’s lead author Dr. Adrian Mundt. “Clearly this is a starting point to discuss this more and provide a new angle to this problem. We now should go into more depth and find out more about pathways for individuals.”

Mundt, who is affiliated with Queen Mary University of London, and his colleagues write in JAMA Psychiatry that the theory of a connection between psychiatric bed availability and prison population was proposed in 1939 by English scholar Lionel Sharples Penrose.

Studies examining the so-called Penrose Hypothesis have produced mixed results, however. More recent studies suggested that there is no connection between the number of psychiatric beds available in a community and local prison population.

Mundt told Reuters Health in a phone interview that it’s important to examine whether a connection exists, but governments haven’t done the research.

“There’s not much coordination or knowledge about changes in one system causing changes in another system,” he said. “Also saving money in one system can cause costs in another.”

For the new study, the researchers used data on the availability of psychiatric beds since 1990 in Argentina, Bolivia, Brazil, Chile, Paraguay and Uruguay. Then they compared that data with data on the prison populations of those countries.

The reason the researchers looked at data starting in 1990 is that those countries signed a declaration reducing the number of psychiatric beds and increasing the availability of outpatient mental health services.

Overall, they found the number of psychiatric beds in each country decreased while their prison populations increased.

For example, the number of available psychiatric beds in Argentina dropped by about a third between 1990 and 2012. Its prison population increased by about 137 percent during that time.

In Bolivia, the number of psychiatric beds fell by about 2 percent. Meanwhile, its prison population increased by 16 percent.

On average, the researchers found that prison populations increased by about three people for every psychiatric bed that was eliminated after they accounted for economic growth and income inequality.

Mundt said previous studies in more economically wealthy countries may not show a link between psychiatric bed availability and prison population, because they have better community welfare programs that can treat people in the community.

Dr. H. Richard Lamb, who wrote an editorial accompanying the new study, said decreasing numbers of psychiatric beds may be one reason for the growing prison population.

“I think it’s definitely a part of it,” said Lamb of the University of Southern California in Los Angeles. “It wouldn’t be the only reason. There are a number of reasons there are more people in prisons.” Those include the war on drugs and so-called tough-on-crime initiatives.

“About 350,000 people in the U.S. who have serious mental illness are in our jails and prisons,” Lamb said in a phone interview. “One reason for that is the psychiatric beds that would have met their needs are not there at this point.”

Mundt said the next step would be to see whether people released from psychiatric facilities actually ended up in prisons or jails.