In the early 1960s, an abnormally high number of children were born with severe deformities: limbs were missing, or only half there, fingers and toes were found at the end of what looked like flippers.

It turned out that a tranquillizer, marketed as the safest ever, was causing the problems. It was commonly prescribed to pregnant women for morning sickness. The drug was thalidomide.

Thalidomide: Bitter Pills, Broken Promises, from CBC Archives

Devastating for the families affected, the discovery set political wheels in motion around the world. Regulators realized with tragic clarity that it was insufficient to establish the safety and effectiveness of a drug before it went on the market. A system was needed to track medications once the general population had access to them.

In 1962, the year thalidomide was pulled from the market, the Ministry of National Health and Welfare, as Health Canada was then called, asked the Royal College of Physicians and Surgeons to study Canada's drug approval process. The college suggested some form of surveillance to track drugs already on the market as a way of detecting potentially dangerous side effects.

The U.S. Food and Drug Administration, which had started work on a reporting system in the late 1950s, implemented a voluntary reporting system in 1963. (The U.S. had refused to market thalidomide.)

The following year a Canadian parliamentary committee recommended a similar system for collecting reports of suspected adverse drug reactions (ADRs). In 1965, National Health and Welfare established the Drug Adverse Reaction Reporting Program.

Three years later the World Health Organization started the Collaborating Centre for International Drug Monitoring, with 10 founding countries, including Canada, the U.S. and the United Kingdom.

Canada's system of ADR reporting was set up by Dr. Ed Napke, a man who liked to see what was going on. Although the reporting system was computerized, Napke developed a system of "pigeon holes" into which reports were filed. Coloured tabs were attached to reports of severe or unusual reactions as they were filed.

Dr. Ed Napke filing reports in his system of "pigeon holes"

When clusters of coloured tabs began to emerge, Napke and his small team had a visual cue that something might be wrong. In the world of post-market drug surveillance, such cues are called signals, and signals prompt investigations that can lead to warnings, or even to a drug being pulled from the market.

Now, nearly 40 years later, the surviving thalidomide children are middle-aged, and Canada's post-market surveillance is run by the Canadian Adverse Drug Reaction Monitoring Program of Health Canada.

Napke's pigeon holes are long gone, and the signals are now contained in a database called the Canadian Adverse Drug Reaction Information System, or CADRIS.

But critics complain that all is not well, that Health Canada is reactive and not picking up on trends quickly enough.

There are some parents who say Canada's system of post-market surveillance has failed them; notable among them are Susan and Chris Atkinson. Their six-year-old daughter Ashley died in February 2001.

Ashley went to the hospital with croup. She woke up one night having difficulty with her breathing. Four days later, she was dead. She had been sedated using a drug called propofol.

A coroner's jury didn't blame the drug, but in testimony two medical experts linked Ashley's death to propofol toxicity.

Health Canada's CADRIS database, acquired by CBC News through federal Access to Information laws, lists 13 reports of suspected adverse reactions to propofol in children between 1993 and 2003. Three of those patients died. Ashley was the third.

"Why is it taking them so long to do something about these drugs that are reported and flagged over and over and over again?" asked Susan Atkinson after seeing the listings from the database.

Dr. Supriya Sharma, Health Canada

According to Health Canada, that's a very complicated question.

When asked about how good the country's adverse drug reaction reporting is, Health Canada spokesperson Dr. Supriya Sharma split it into two considerations: quality and quantity.

Sharma and Health Canada concede that the ADR system receives only a fraction of all suspected adverse drug reactions in Canada, somewhere between one and 10 per cent.

Currently, 10,000 reports are entered into the CADRIS database annually. In 2002, 450 of those involved a patient who died as the result of a suspected adverse drug reaction.

But Sharma counters that it's not the number that counts, it's the quality of the reports. Those reports are made on a voluntary basis by health professionals, patients and consumers. Drug companies, on the other hand, are required by law to report any known serious or unexpected adverse drug reaction.

• RELATED: Should you file an adverse drug reaction report?

Health Canada says it determines the quality of the reports by using the database in combination with other information, such as medical literature, media reports, international databases of suspected adverse drug reactions and statistics on drug usage in Canada.

Health Canada also cautions that, due to underreporting, it's impossible to project real rates of adverse drug reactions based on the numbers found in their database. So for example, just because there are database listings for three children who reportedly died from a suspected drug reaction to propofol, you can't definitively conclude that 30 children, or 300, have died in Canada as a result of suspected drug reactions to the same drug.

The database is used to find signals and warning flags. It’s also used to corroborate trends found in other countries or in medical literature.

While the quality of the data in the database may not be perfect, it is an integral part of Canada's drug safety program.

Investigations started by those tracking Canada's ADR reports can lead to labelling changes, a manufacturer issuing a Dear Doctor warning letter, or the decision to pull a drug from the market. (Health Canada can step in and do this, but the usual practice is for the manufacturer to take action in consultation with Health Canada.)

Critics complain that something needs to be done to improve the system. Dr. Gordon Guyatt, a specialist working at McMaster University, suggests mandatory reporting.

Others, such as Dr. Eric Wooltorton, say doctors need an incentive to improve the quality of the reports. Faced with a full waiting room, physicians will choose treating patients before filling out a Health Canada form.

And still more critics, including the Atkinsons, say Health Canada needs to track trends more aggressively in order to pick up the warning signals. That's because, as the case of Ashley Atkinson illustrates, the database of ADR reports, blemishes and all, often contains the signals necessary to take action.