Randomised Controlled Trials are a great idea in theory but most are hampered with a near-fatal design flaw.

The controlled trial is an attempt to remove chance, bias and confounding from an experimental observation. By controlling or leaving unaltered as many of the “extenuating circumstances” as possible, it is hoped the the effects of the therapy can be ascertained.

Good randomisation is supposed to ensure that variables in patient selection are eliminated so that the results are not biased by selecting healthier, younger or better-suited people in one group or the other.

However, most RCTs presume that all practitioners are the same and the patient experience will have nothing to do with the outcome.

As an example, in 1995 the MRC Meade Trial compared randomised patients attending chiropractic or hospital outpatient clinics to receive either Chiropractic treatment at one of the private clinics that had volunteered for the study or at hospital outpatient physiotherapy clinics. Three years later, those treated in chiropractic clinics still had more benefit than those treated in the hospital clinics. But what actually caused the benefit? Was it that chiropractic treatment really is better or are chiropractors more confident, or more personable? Did the chiropractic clinics play better music than the outpatient clinics, were the reception staff better trained? And so on. Far from being controlled, randomisation across centres introduced hundreds of confounding factors. Add into this uncertainty, the fact that patients had clearly expressed a preference for the type of therapy they would receive by attending either physiotherapy or chiropractic in the first place, and you have a very unsatisfactory study.

If you want to know whether physiotherapists get better results than chiropractors, or whether McTimmoney chiropractors get better results than ordinary chiropractors, you could just measure their outcomes.

BUT let's say you wanted to know something really interesting like whether low back pain patients did better with stretching exercises or strengthening exercises. Using Record Your Results, this would be a breeze. You could recruit a group of chiropractors using the system already, preferably individuals who were genuinely interested in finding the answer to the question. Researchers who already “know” which is better are not well qualified to investigate an unknown. Each practitioner could then randomise their next 30 low back pain patients into stretch or strength exercise groups and then the software would collect the results without any bias or favour. While the results of one chiropractor would not be generalisable, pooling the results of 100 chiropractors doing the same thing would have tremendous significance because each RCT introduces only one variable, the one under investigation. All other factors in the care of each patient group would remain unaltered.

Patients are happy, they received a widely used and accepted treatment. Chiropractors are happy, they were doing practice-based research on paying customers. The profession is happy because we advance our knowledge and the researchers are happy because they come out with a very well controlled trial with broad validity.

Imagine the questions we could finally answer.

With a standardised and independent measuring system, quality RCT's become reliable and cheap to implement.

Simon King