TORONTO -- A new Canadian study suggests nearly one in three new prescriptions goes unfilled, with expensive drugs and medications used to control some chronic conditions more likely not to be taken as directed.

The lead author of the work admitted the magnitude of the gap between prescriptions written and drugs purchased was greater than she and her co-authors expected, even though the figure was in line with a previous American study looking at the issue.

"We were startled," said Robyn Tamblyn, an epidemiologist who teaches in McGill University's school of medicine.

"In fact, we kept on lengthening the window that we looked at to make sure that was the case, actually."

The study was published Monday in the journal Annals of Internal Medicine. It was funded by the Canadian Institutes of Health Research (CIHR).

The findings are based on data from Quebec. Everyone in that province has drug coverage, either through their employment or through a provincial program that covers people like seniors, students, welfare recipients, people who are self-employed and employees of small businesses who wouldn't otherwise have drug plans.

The authors analyzed data on nearly 16,000 patients who were prescribed 37,506 new prescriptions from the beginning of 2006 to the end of 2009. Of those, 31.3 per cent were not filled nine months after the prescription was written.

Researchers who were not involved in the paper called the findings important, though one pointed out that they may not be completely generalizable to the entire Canadian population.

Steve Morgan, director of the University of British Columbia's Centre for Health Services and Policy Research, suggested the people studied represent a "somewhat vulnerable" population -- meaning they might have lower incomes and less capacity to cover copayments if they are required by the drug plan.

(Quebec's public plan has three tiers which are linked to income. People on the lowest rung are not required to make any copayment; they get medication for free. The next tier involves paying for 25 per cent per medication to a limit of $45 a month. Top income earners in the plan pay 25 per cent per prescription to a maximum of $80 a month.)

Tamblyn acknowledged the findings might have been somewhat different if the study also encompassed people with workplace drug plans. Those people might be more affluent, and in some cases their drug plans might be more generous. But Tamblyn said some findings would likely remain the same, noting other research has shown that prescriptions for expensive drugs are less likely to be filled than more moderately priced ones.

The study looked at prescription records only. The authors did not talk to the patients about why they did or didn't fill a prescription, so they cannot say what is behind the behaviour. Tamblyn said she and her colleagues plan to do a followup study to mine that issue with patients.

But there are clues in the types of prescriptions that are filled and those which go unfilled.

For instance, the percentage of unfilled prescriptions for treatment of urinary tract infections -- which are painful and difficult to ignore -- was comparatively low, 21 per cent. Likewise, antibiotics, which are used (and abused) to tackle infections, are generally filled.

But prescriptions for episodic conditions -- things that come and go such as migraine headaches or rashes -- had comparatively high rates of non-adherence, meaning they were more often left unfilled. Prescriptions for higher priced drugs were more likely to go unfilled. And patients who had higher drug co-payments were more likely not to fill prescriptions.

Prescriptions for antidepressants, which many experts believe are over-prescribed, were not filled 38 per cent of the time.

But drugs for coronary artery disease, high blood pressure, high cholesterol and diabetes were also among those for which prescriptions were less likely to be filled -- which in many cases probably isn't the right approach. "Some drugs, when appropriately prescribed, will keep people out of the hospital," Morgan noted.

Tamblyn said the finding related to drugs for chronic conditions may be due to the fact that these types of medications are generally not short-term.

"My suspicion would be it's not an easy thing for people to start therapy that presumably there had been some discussion that this would be forever. So maybe there's a reluctance to do that," said Tamblyn, who is also director of the CIHR Institute of Health Services and Policy Research.

"Maybe it's 'If I can get my weight down' maybe I don't have to take this -- that kind of thing."

Dr. David Juurlink, a specialist in internal medicine and clinical pharmacology, said that while the one-in-three number looks surprising at first, part of the issue relates to the fact that doctors write too many prescriptions.

Juurlink, who practises at Sunnybrook Health Sciences Centre in Toronto, notes some prescriptions are written with clear instructions that patients should only fill them if necessary. So in some cases not filling a prescription is the rational approach, he said -- though he acknowledged a portion of these prescriptions should have been filled and were not.

"You can't from a paper like this know which unfilled prescriptions should not have been (unfilled) and which filled prescriptions should not have been filled," he noted.

"As a general pronouncement, I wouldn't say that it's a good thing. But in some instances, I think it might not be such a bad thing."

Still, Juurlink said the preferable course of action would be for doctors to write prescriptions only when drugs are needed and are likely to be effective. "Patients shouldn't be the ones making that decision," he said.

Tamblyn said given these findings, it's important for doctors to keep in mind that patients may not be taking all the medication that's been prescribed for them. She guessed that as much as half the time doctors would be unaware of this fact.

"We have never really fully recognized that this is an issue. And when people come back and they're worse off than they were before, they'll get a new therapy thinking that 'Well, I guess that first therapy didn't work' when in fact the prescription was never filled."

Morgan said based on earlier studies, as many as 10 to 15 per cent of unfilled prescriptions might be due to cost constraints. He noted that many patients may be embarrassed to reveal they can't afford prescribed medication.

Other times there may be a lack of communication between doctors and patients, he said, with the latter not having the chance or the gumption to say that they would rather pursue other options before resorting to medication.