Dr Peter Moore: When a research project goes wrong...

Back in the 1970s, a chest specialist in London started a research project.

He had the idea, which we now know is correct, that chest disease is made worse by air pollution.

Unfortunately, his research was not all plain sailing. The idea was simple. He gave each patient with chest disease a diary. Everyday they were asked to write A, B, C or D ranging from A meaning a wonderful day to D meaning dreadful. He also recorded the level of air pollution in London. At the end of each month he compared the diaries with the level of pollution. Was there a correlation?

What he did not anticipate were the vagaries of his patients. When one died, the family gave him the diary. It was filled in until the end of the month after he had died. The only good news was that the day after he died there was a A.

Another research project which went wrong involved a friend of mine who was a junior doctor at the time.

Medical research usually involves an active group of patients involved in the trial who are compared with a control group who are not treated. In the study where my friend was the junior doctor, patients were split into active treatment and the control group at random by the doctor picking up an unmarked envelope.

Half the envelopes contained all the paperwork for the study, the other was empty for the control group.

Fairly soon he realised that the thick envelopes meant a large workload for the junior doctor, the thin one meant he had no extra work. After a few months the consultant in charge commented that it was amazing. Patients were being selected at random and yet every patient so far had been in the control group.

One research project I was involved in as a junior hospital doctor appeared to have a major flaw. We were trying out a new tablet for heart attack patients. The protocol required any patient who entered the trial to have a particular type of large heart attack, the blood test for the enzymes which are released during damage to the heart had to double but the patient must not go into heart failure, when the heart does not beat hard enough.

The problem we had was that anyone who had such a severe heart attack almost always went into heart failure. In the six months I worked on this ward we did not have a single patient who fitted the protocol.

Some research appears to prove the obvious. I remember reading a paper in the 1970s looking into how patients contacted their GP. It was at a time when many patients did not have a phone. The amazing conclusion was that people are more likely to ring their GP if they have a phone.

Research is not always scientists in white coats in a lab. Every hospital will be involved in several research projects.

Today, research in hospitals is far more sophisticated. Any idea has to be approved by an independent ethics committee and the doctors have to ensure that any patient involved fully understands and has the option to say no.

The system of random allocation is no longer the junior doctor picking thin envelops but computer generated. If we look at how one aspect of a patient’s life affects their health and compare it with other people the group of people may be different. If, for example, we look at people who go to the gym and compare them to people who sit a home, is it likely that there are other differences. Are there the same number of smokers, do both groups take the same amount of exercise at other times and are more of couch potatoes overweight?

The statistics underlying the results are analysed in detail by computer programmes.

Research has moved on from giving patients a diary although I have not come across any other research in which patients scored a A for their health after they had died.