Transplant Registries: Guiding Clinical Decisions and Improving Outcomes

After nearly 30 years of dedication, the International BoneMarrow Transplant Registry (IBMTR) and the Autologous Blood and MarrowTransplant Registry (ABMTR), by nearly any criterion, can be considered asuccess. The Registry contains over 120,000 patient records that are aninvaluable source of information on both autologous and allogeneic bone marrowtransplantations. Through the voluntary efforts of more than 350 institutions,it annually registers over 20% of all transplants occurring in nearly 50countries and, by some estimates, nearly half of all transplants in NorthAmerica.

The IBMTR/ABMTR has evolved to become an invaluable resource tophysicians and researchers around the world. Perhaps the strongest indication ofits value is the vehemence with which Registry data and methods are occasionallyattacked and conclusions from the large data sets dismissed—in science, thisis often a good indication that the traditional paradigms are being successfullychallenged.

Strengths and Weaknessesof Imperfect Data

Clinical trial purists argue that inherent selection biasesseverely compromise the value of Registry data, and that valid treatmentcomparisons come only from prospective, randomized trials. This issue is toobroad to be considered in detail here, but the continuing broad-based,long-term, and voluntary cooperation of large numbers of investigators aroundthe world is a clear indication of the value of the Registry to the field.

Although there clearly are other limitations to the use ofRegistry data, it is fair to say that the major limitation is the lack ofcomparable control data from a registry of patients not treated bytransplantation. However, the Registry has not allowed the good to become theprisoner of the perfect. The clinical world is not always perfect, and we canoften learn much from imperfect information. Indeed, the imperfections ofRegistry data result in a certain reserve and humility, which is sometimeslacking in analyses of randomized trial results. The rest of oncology would dowell to develop similar comprehensive data sets.

As with many areas of clinical research, the major strength andweakness of the Registry is one and the same: the homogenization of data frommany varied sources. The best gets lumped in with the worst—in essence aregression to the mean. The number of patients from any given institution may besmall, the treatments varied, and the skill and depth of the support teamsheterogeneous. In my mind, this limits the use of Registry data in selecting the"best" treatment approach.

On the other hand, the broad-based nature of the Registryprovides an unparalleled opportunity for identifying factors that may contributeto variations in outcome and to understanding the state of the art of the wholefield for policy considerations. In addition, it currently offers the onlyrealistic method of collecting data on transplant treatment of rare diseases.

Selective and Long-Term Benefits

As it becomes even larger, the Registry will serve anincreasingly important role in the study of rare events, particularly forunraveling the etiology of unusual toxicities. Rare events are inherentlydifficult to study in medicine; through its voluntary nature and its broad-basedparticipation, the Registry provides the opportunity to benefit individualpatients with unusual syndromes by allowing their physicians to gain insightsfrom the experience of other transplant institutions with these rare events.

Because data are collected over an extended period, issuesaffecting survivorship can be identified and properly addressed during activetreatment periods, even when such effects do not appear for many years. Thisfact could be one of the Registry’s greatest strengths.

Conclusions

In the end, perhaps the greatest overall value of the Registryis its ability to facilitate the generation of hypotheses. Three factors—thelarge size of the database; the varied nature of the patients, diseases, sites,and physicians; and the intrinsic curiosity of the participating members—combineto provide the proper seed and soil for the intellectual restlessness andintercourse that conceives and ignites new ideas. Finally, the development ofrandomized trials is truly facilitated by Registry data—again, all oncologywould be well served by similar systematically collected data from nontransplantsettings.

As a clinical scientist who for many years has been skeptical ofthe role of the Registry, I must admit that there is great value here. The hopeis that 30 years from now the Registry will no longer be needed because its usehas led to greater therapeutic success.