I’m now in Washington for the ACR conference (see explanation here). We’re half way through the Clinical Research Conference and I’m jotting some notes in the break.

The theme of this part of the conference (actually a pre-conference) is ‘reducing disparities in prevalence and impact of rheumatic diseases’. We’ve heard a nice introductory talk by Dr. Agustin Escalante who presented some nice background on the differences between race (genetic) and ethnicity (social). Also about measuring ‘admixture’ of various ‘racial’ characteristics, looked at on a genetic level. It’s good stuff. I wish he had some slides I could refer to or that I’d taken better notes.

We then heard a series of talks on evidence for disparities.  There is plenty of this. Some of this is accounted for by true differences between men and women (e.g. 10 times as many women as men have lupus – is this related to endogenous or exogenous hormones?). Others appear to reflect differences due to perceived race. There can, of course, be true physiologic differences between ‘races’ as well as sexes. But one of the issues that has been pointed out repeatedly is that there are a lot of variables that contribute to health care disparities and it can be very difficult to tease them all out and figure out what’s really due to race or sex or whatever. But in general, there is evidence that disparities do exist, for whatever reason.

The most interesting talk of this first part was by Dr. Matthew Liang who basically challenged researchers to stop (or at least do less of) describing disparities and to start conducting randomized trials in an attempt to do something about it. As part of this argument, he noted that general health seems to be related less to technology than to the social and economic environment. That seems counter-intuitive at first, but he presented a nice study of tuberculosis that shows that death rates were falling long before there were direct treatments for the disease. The conclusion is that adjusting the social environment may be more important than developing new technologies. , of course, includes socially-attributable disparities in health care.

The second part was a series of talks about ‘social and behavioral determinants of differential outcomes in rheumatic diseases’. For some reason I found this part not to be very interesting, so I’m going to ignore it. Maybe I was just tired.

We were then presented with a very interesting abstract by Cornelia Borkhoff in
Toronto. I thought this was so interesting that I’m going to present the whole abstract below:

Author(s): Cornelia M. Borkhoff1, Gillian A. Hawker2, Hans J. Kreder3, Richard H. Glazier4, Nizar N. Mohamed5, James G. Wright6. 1Population Health Sciences,The Hospital for Sick Children, University of Toronto, Toronto, ON, Canada; 2Division of Rheumatology, Department of Medicine, Women’s College Hospital, University of Toronto, Toronto, Ontario, OH, Canada; 3Division of Orthopaedic Surgery, Department of Surgery, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, Ontario, OH, Canada; 4Department of Family and Community Medicine, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, OH, Canada; 5Division of Orthopaedic Surgery, Department of Surgery, Toronto Western Hospital, University of Toronto, Toronto, Ontario, OH, Canada; 6Department of Surgery and Population Health Sciences, Research Institute, The Hospital for Sick Children, Toronto, Ontario, OH, Canada
Presentation Number: CRC1

 PURPOSE: Unmet need for total joint arthroplasty is more than three times as great in women as in men. Physicians indicate that patient gender has no effect on their decision making. The purpose of this study was to determine the effect of patients’ gender on physicians’ recommendations for total knee arthroplasty (TKA).
METHODS: Standardized patients (SPs) (one man and one woman) with moderate knee osteoarthritis visited 38 family physicians and 33 orthopaedic surgeons located in Ontario, Canada. Physicians were blind to the patient. SPs were trained to present with chronic knee pain as their chief complaint and submit the same essential clinical features of a scenario with the level of function and pain and prior treatment appropriate for moderate knee osteoarthritis. SPs recorded physicians’ final recommendations (yes/no) to refer for, or perform, TKA. Binary logistic regression analysis using a generalized estimating equation (GEE) approach was used. Estimation of sample size was based on a one-sided alternative hypothesis because the focus of this research was investigate the potential for physician gender bias. Given that a one-sided test of significance at α=0.05 was used to estimate sample size, for the results we present one-sided P values and 95 percent confidence intervals.
RESULTS: Sixty physicians were male. Of the 11 female physicians, 8 were family physicians. The overall odds that men were recommended for TKA by a physician was 4 times higher than the odds of recommendation for women (odds ratio, 4.4; 95% CI 2.5-7.5; p<0.001). After detecting an overall main effect for gender, we tested for an interaction for the pre-specified subgroup: physician type. The odds that men were recommended for TKA by an orthopaedic surgeon was 24 times higher than the odds of recommendation for women (odds ratio, 23.9; 95% CI 7.0-81.0; p<0.001). The odds that men were referred for TKA by a family physician was 2 times higher than the odds of recommendation for women (odds ratio, 2.21; 95% CI 1.04-4.71; p=0.043).
CONCLUSIONS: Physicians are more likely to recommend TKA surgery to men than to women, suggesting that physician bias may contribute to the gender disparity in the utilization of TKA.

I thought this was very interesting. Often our whole point as scientists is eliminate all extraneous differences between two groups (or people, as above) and look at the difference in outcome of only the variable of interest. This is what Dr. Borkhoff has done and in doing so she appears to have shown that there is indeed a difference in how physicians offer therapy to the two sexes. The reasons for this may be complex and hard to tease out, but the demonstration is lovely.

Now the study isn’t perfect. There were only 2 standardized patients and there may have been subtle differences in how they presented themselves that influenced the outcome; but in the presentation Dr. B. outlined the methods they used to minimize this. It would be nice to see the effect with a few more standardized patients.

Part 3 of the day was a series of presentations which the organizers failed to title but that mostly concerned differences in access to joint replacement. The gist of this session was that white men get joint replacements more often than everyone else, even though the burden of the disease may be higher in other groups.

But it’s more complicated that just assuming racial bias. There are also differences in the preferences of men vs women, or blacks vs whites. For instance, it turns out that blacks minimize their disability when compared to whites or Hispanics which means that they are less likely to undergo surgery. And often they are more skeptical of undergoing a procedure, or more skeptical that the results will be positive. Interestingly enough Dr. Suarez-Almozar found that American blacks were more likely to be offered joint replacement, yet lesslikely to undergo the procedure. She did not find the same apparent bias in race or gender on the part of what physicians offered as Dr. Borkhoff did.

In the same vein, Dr. Losina found in her studies that patient factors seemed to play a major role in outcomes and complications of total joint replacement. She found that minorities, those in rural areas, the relatively poor and undereducated, tended to be seen and operated on in lower-volume centers (i.e. places that do relatively few total joint replacements), and thus suffered more adverse outcomes. Interestingly, if I interpret her right, she suggest that if some of these patients can be directed to higher volume hospitals, they will get better care and cost society less money.

And at this point, I had to go to another session that I might write about tomorrow.

Oh, one more thing.  I’m very happy to say that about half the people in the room were women and that more than half the speakers were. I’m happy to say that women are active in my fields (both rheumatology and epidemiology) and doing work that it at least on a par with men. I hope their pay is reflecting that.