From: Marc Swiontkowski, MD
Three methodological and logistical concerns about Registries are:
1) - one can never anticipate every data point that one might need to answer all questions in a registry in order to answer a post-hoc constructed research question - the data collection process would be so overwhelming so as to halt the process
2) compulsory data quality in registries is suspect even under the best of circumstances
3) the expense in maintaining and starting up a registry far exceeds the number of high quality research questions that will ever be addressed in this fashion making prospective, focused, hypothesis driven research a much better bargain
From: Augusto Sarmiento, MD
1) A clear definition of a REGISTRY. Reasonable goals.
2) The Scandinavian examples. Reproducible in the U.S.? An objective look at their alleged success.
3) Not all Registries have been successful? Why have some failed?
4) Financial constrains. How to overcome them.
5) End points.
6) Who can, or cannot participate in the project?
7) If not every surgeon will participate, how will the ones left out react to the situation?
8) If not every commercial product can be included in the registry, how will industry react to this?
9) Credibility assurance.
10) Industry's ubiquitous presence. Can its influence be controlled?
11) Could the registry's reports encourage litigation?
12) On balance, do we need Registries? Is there a better way to accomplish their proposed goals?
From: Seth Leopold
Subject: Re: Registry Roundtable Questions
Date: Wed, 27 Feb 2008 19:32:25 -0800
Although I too am a fan of hypothesis-driven research, I actually think that registries provide very important information. So in contrast to some of the others who have posted questions to the group at the URL listed below, I think well-managed registries are real assets to my subspecialty. I think we've learned a great deal from the experience in northern Europe, and gotten some insight into problems that require really large N or real "external validity" (e.g. real-world applicability of certain things outside the hands of designers or high-volume academicians) to figure out -- the ones that come to my mind include problems with certain bone cements that failed early (easily identifiable with N=10k patients, but would've taken much longer to discover with the typical orthopaedic study of N=150), potential benefits of mixing antibiotics into cement (impossible to discern with N 1. For a registry to be worthwhile, participation has to be total or near-total. A registry with 50% of the cases going undocumented leaves an unacceptably severe problem of transfer bias (loss to follow-up). Without a single payor system to provide financial inducements for complete participation, what will incentivize complete or near-complete participation on the part of surgeons?
2. We have learned from northern Europe that registries cannot be run on the cheap. Absent a governmental or rather massive philanthropic source of funding (which is not forthcoming), the remaining option is industry. This is the topic of my powerpoint, but speaking broadly, I have some real questions pertaining to data integrity, as well as actual and perceived conflicts of interest that might be associated with this option.
3. There are some serious implications to gathering and publishing registry data in our medicolegal climate that in large measure do not exist in northern Europe. What steps can be taken to manage this problem if a registry is created?
Bill, thanks again for the invitation to present at HWB; I look forward to sharing the results of our work on publication bias and how this might impact the registry question at your meeting on Thursday.
From: Augusto Sarmiento
Dear Bill,
Responding to your request to summarize my comments at the meeting, I submit to you the following:
My opposition to registries is based on several factors:
1) Bad personal experiences with the Fracture Registry I proposed and for which I received financial and moral support from the AAOS in the 1970's. The failure of the Hip Registry, which as president of the Hip Society, I proposed in the late 1970's. This project also subsequently failed under the leadership of Clem Sledge. A similar disaster occurred with a J&J supported Fracture Registry at USC. Needless to say, the failure of these projects was, to a great extent, due to my lack of talent in these areas.
2) The failure of Maurice Muller Documentation Center, to which Professor Muller faithfully and passionately devoted enormous effort and millions of dollars. It ended up being nothing more than the accumulation of millions of data points, without a resulting new body of "knowledge".
3) Our infatuation with the Scandinavian Registries is based, thus far, on the perception that it has been a success. I am unable to see that success. More time is needed to judge whether the huge expenditures and broad support will produce tangible benefits. "The way to hell is paved with good intentions".
4) Since it is very likely that universal participation in this country is not a realistic scenario, the inclusion or exclusion of surgeons and institutions would create numerous problems. Will registries dealing with specific subjects, such as total joint replacements, or fracture implants and instrumentation, accept the input of people with known conflicts of interest? Will owners of implants be members of the "group"?
The veracity issue, the seminal one in this entire subject, is impossible to ensure. That being the case, the future augurs disaster. Naming "monitors" to oversee the appropriateness of the data will call for recalling Juvenal's criticism of Plato's Republic: "Who will guard the guardians?"
5) Registries are expensive. If Industry is invited to provide financial support, it would be equivalent to the driving of nails into the coffin of stillborn altruistic ventures. Industry would take over the entire project, pretty much the same way they, so successfully, ended up controlling the education of the orthopaedist.
Gus
Further Links:
To: William Burman
Subject: On the San Francsico Meeting.
Date: Fri, 14 Mar 2008 21:36:36 -0400