In order to determine whether a new treatment or technology will benefit patients, it must be tested. Usually this is done by clinical trials involving groups of patients with the same problem who volunteer to be randomly selected either to try the new treatment, or to be treated according to accepted practice. But what measures are used during these trials to gather the information on the new therapy? And how is that information analyzed and presented? These questions are decided by the study designer who, in collaboration with an expert on the study's content (i.e., the medical questions asked), designs the study, creates the statistical methodology, and coordinates the analysis.

As a biostatistician and epidemiologist, Dr. George Wells frequently takes the role of study designer. Often, his collaborators are fellow Loeb researchers. For example, Dr. Wells and Dr. Peter Tugwell have just published the results of a randomized control trial of combination therapy in rheumatoid arthritis. Dr. Wells also works with Dr. Ian Stiell in his development of decision rules for pre-hospital diagnoses, and in his studies of cardiac arrest. In an investigation that has just been funded by the Medical Research Council, Dr. Wells will be collaborate with obstetrician/gynecologist Dr. Erica Eason of the Ottawa Civic Hospital in an investigation of the antisepsis of vaginal gels for preventing infection in hysterectomies. This is an important problem because it occurs frequently and causes significant discomfort. In this multicentre study, which will involve research centres from Quebec City to Toronto and London, specific gels will be evaluated, and their efficacy compared.

In another study that is just concluding, Dr. Wells investigated blood transfusion requirements in the ICU with Dr. Paul Hébert. The two investigators compared the standard practice of transfusing blood with high levels of hemoglobin (100-120 x per y) with transfusions of blood with low levels (70-90), in order to show that the safety and effectiveness is the same.

Dr. Wells also teaches an advanced course on the design of clinical trials in the Department of Epidemiology and Community Medicine at the University of Ottawa.

In a related area, Dr. Wells collaborates on meta analysis projects, such as those conducted with Loeb scientist Dr. Graham Nichol in the field of decision analysis. In this case, Dr. Wells is responsible for how the information used in the analyses is gathered and presented. That means first conducting a systematic review of the literature to find the various publications on a single subject, then creating the mathematical models by which they will be combined, and finally conducting the meta analysis to arrive at a single overall estimate of the data yielded by all the different studies.

Dr. Wells also works on specific projects in the field he refers to as "big 'E' epidemiology." Often, this means looking at large studies where people are exposed to particular diseases, and then either have or have not developed them. Many of these studies use surveying methods that are themselves an area of interest for Dr. Wells. He also explores occupational health issues, often with police forces like the RCMP. One recent study investigated whether radar units cause cancer. This question arose because policemen, who often set the radar guns between their legs when they drive off after reading the speed of passing cars, appeared to have a high incidence of testicular cancer. Dr. Wells determined that this was not the case, that in fact the observed level of testicular cancer was normal for any similar group of men.

Dr. Wells has moved on to other occupational health studies, such as one he is currently conducting on eyewear. The issue here is that the element of danger for all concerned increases when police officers' eyeglasses are knocked off in a scuffle--particularly when guns are involved. Contact lenses, which don't fall out, might be a safer alternative. In another investigation, Dr. Wells is investigating back pain among those in occupations that require them to sit for long periods of time.

Dr. Wells has also just embarked on a national survey evaluating a Canadian perinatal program. This benefitted specific groups of disadvantaged women, such as teenagers or aboriginal people, to help them ensure healthy pregnancies. The program was implemented in different spots across the country, and must now be assessed to determine if any changes were effected. Dr. Wells and his team will survey different populations of participants selected by Statistics Canada, to compare baseline characteristics and to evaluate the program.

As part of Dr. Wells's study of how information is synthesized, he is researching the way data is gathered for studies of hormone replacement therapy and osteoporosis. Along with Dr. Tugwell, Dr. Wells is also a member of the editorial committee of the Musculoskeletal Group of the Cochrane Collaboration, a worldwide effort to collect and synthesize all the data from every clinical trial undertaken around the world. As part of this effort, he conducts systematic reviews, particularly concentrating on continuous outcomes, which represent a shift of focus from success-or-failure approaches. In osteoporosis, for example, the most clear-cut outcome--known as a CIO, or "clinically important outcome"--is a fracture: Either the patient has one, or she doesn't. Therefore, in osteoporosis research, a surrogate outcome, such as bone-marrow density, is often used. This would be measured by how much bone the patient has had, and how much their care has helped prevent them from losing. Any sampling of papers on this subject will measure and relay this information in different ways, which is what makes meta analysis of continuous outcomes both difficult and important. Similarly, the severity of rheumatoid arthritis can be measured simply by counting the number of swollen or tender joints a patient has, which again is expressed in different ways from one study to another.

But other outcomes must be considered, too--such as the quality of patients' lives. As part of an international study, Dr. Wells is now assessing different methods of measuring quality of life and functional outcomes. In this particular study, five questionnaires are being evaluated for how sensitive they are to changes in people with rheumatoid arthritis. Once a patient is enrolled in the study, a physician conducts a clinical profile. Then, the patient is put on methotrexate, a drug that has a strong clinical impact, and a joint-count is done. Patients are assessed at three and six months, and asked to fill out questionnaires, including functional scales like the Health Assessment Questionnaire, and general quality-of-life forms like the Nottingham Health Profile. Thus, not only is the clinical outcome measured--and expected to be improved, with this particular treatment--but quality of life is also assessed, to see if it's sensitive to clinical outcomes. This will also reveal if the quality of life assessment questionnaires pick up changes in quality of life due to the drug treatments.

With so many efforts in so many fields, Dr. Wells's three main pursuits--how to assess outcomes, how to gather and present evidence, and how to design and present clinical trials--are all directed toward one aim: improving health care for all Canadians.