AIDS TREATMENT NEWS Issue #301, August 21, 1998
John S. James

CHORUS (Collaborations in HIV Outcomes Research--U.S.) began in fall 1997. We are covering it now because of the growing interest in side effects of antiretroviral drugs, especially the body-shape changes and metabolic effects which have been seen in some patients using protease inhibitors. The CHORUS database is one tool for studying such problems; many people are thinking about how to research these side effects, and they should know that this database is available, and that persons outside the program can suggest questions for research.

CHORUS is funded by Glaxo Wellcome Inc., but an independent scientific board of physicians and other HIV experts, including community advocates, will decide what research is conducted and published. Database housing and management is being provided by Research Triangle Institute, of Research Triangle Park, North Carolina.

Two reports from the Chorus database have already been presented.(1,2)

How It Works

CHORUS collects data at only four sites (listed below; a fifth site may be added). Therefore, only patients at those practices can enter the program. Their medical records are maintained by computer. At the end of each day, new information about the patients who have consented to enter the program is copied in an "anonymized" form, using a unique identifier without the patient's name or other identifying information. All CHORUS data, identified by this number, is transmitted securely to Research Triangle Institute and added to the database. The research data can be analyzed by site, but individuals cannot be identified.

The four medical practices where patients can volunteer for the CHORUS database (and the principal investigators on the project there) are:

* Pacific Horizon Medical Group, San Francisco (Stephen Becker, M.D.);

* Pacific Oaks Medical Group, Los Angeles (Anthony Scarsella, M.D.);

* Comprehensive Care Center, Nashville (Stephen Raffanti, M.D.); and

* Liberty Medical, New York (Douglas Dieterich, M.D.).

Anyone can submit questions through a member of the scientific advisory board. The current co-chairs are Stephen Becker, M.D., Pacific Horizons Medical Group in San Francisco, and Richard Moore, M.D., Johns Hopkins University in Baltimore. Other board members include the principal investigators at the four CHORUS sites (listed above), community representatives Robert Frost at the American Foundation for AIDS Research (AmFAR) and Dawn Averitt, outside consultants, and scientists from Glaxo-Wellcome.

Interview with Stephen Becker, M.D.

We asked co-chair Dr. Stephen Becker why the CHORUS database is important.

Dr. Becker: I believe that observational databases will become increasingly useful in our understanding of HIV disease, and in answering many clinical questions. Historically they have played a major role, for example the MACS data [Multicenter AIDS Cohort Study, which has contributed to hundreds of publications], and data from Johns Hopkins finding no different clinical course of HIV disease by race or by gender.

Observational data will become more important because of the limitations of doing randomized controlled trials. These trials for HIV now need to be much longer, because clinical events are pushed out further. Adverse events are not completely seen during the duration of clinical trials. And these studies are increasingly expensive to run, particularly if they enroll large numbers of people.

Randomized trials will certainly remain important, but our understanding will be complemented by the larger number of patients, the longer duration, and perhaps the more real- world experience that observational databases will give us. Observational databases follow patients for years. The CHORUS database will have 6,000 patients, after we add a fifth site.

And the issue of real-world experience is important. There is a selection bias in many trials done by referral of patients to a study site, because highly motivated patients are the ones that tend to enroll. They may behave differently: they may be more compliant, they may be willing to tolerate more side effects, they may be a sicker patient population. A community-based observational database such as CHORUS includes almost all the patients in the practices, and therefore reflects prevailing medical practice.

AIDS TREATMENT NEWS: There is much interest today in adverse effects being seen in some patients--the lipodystrophy or the body-shape changes, abnormal cholesterol and triglyceride values, and problems in glucose metabolism. To what extent is the CHORUS database tracking these?

Dr. Becker: We are certainly following adverse events. Lipodystrophy is difficult, because until we have a working definition that we can all agree upon, it will be hard for doctors to code for it and have it be retrievable in large numbers through any automated system. It would be difficult to get this information from any database until we have uniformity of definition.

The CHORUS database includes clinical outcomes, various blood counts, viral loads, diagnoses, CDC demographic criteria, epidemiological information like the risk factor and how the patient likely acquired HIV; these are entered daily from the clinical records into the database. But free-text entries are not recorded; they would need to be coded in a standard way.

ATN: What about the triglycerides, cholesterol, and blood glucose?

Dr. Becker: All of those lab data are in the system, and we are beginning to look at them. For example, we counted patients with cholesterol above 240, and those with triglyceride levels above 350. We can ask about trends--how many patients had high values at certain dates? And we can often match this information with dates when particular therapies were started. We can ask how many patients with these side effects have never been on protease inhibitors or other drugs or classes of drugs.

With observational databases we can look at these problems in a more systematic way than with anecdotal case reports. Large observational databases can help advance our understanding by showing phenomena that are happening; however, they do not answer the question of why. Randomized clinical trials are still needed to prove cause and effect.

ATN: Could these results help guide studies of the mechanisms--of adverse drug effects for example?

Dr. Becker: Absolutely. They should be seen as hypothesis- generating studies, as well as to help our understanding. For example, we could look at a large number of patients who started therapy with protease-inhibitor or protease- inhibitor-sparing regimens, and look at the duration of response--for much longer times and with larger numbers of patients than in clinical trials. And we can match the durability of virological response to treatment with quality of life--to give people more information about balancing the pros and cons of particular regimens.

ATN: One question is, since we have observed that many people on regimens with protease inhibitors are still doing well clinically even after viral load returns, will that be true for other classes of drugs as well?

Dr. Becker: We have set up the means to do just that analysis--distinguishing between the virologic (viral load), the immunologic (CD4 count), and clinical duration of therapies.

Also, when we start or stop therapy, we have to provide a reason for doing so. We are going to code a new reason, "immune reconstitution," to help study stopping treatment or prophylaxis for opportunistic infections. Instead of reports from just a handful of patients, there could be analysis of several hundred.

The database includes a validated questionnaire on compliance (adherence), and two questionnaires on quality of life. We will also be studying economic aspects: frequency of visits, frequency of laboratory use, ordering of various tests, the need for consultants, to try to understand better what the costs of HIV care actually are.

ATN: How is CHORUS different from other observational databases?

Dr. Becker: What distinguishes this database from some of the others is that CHORUS is community-based [meaning that it is based in physicians' offices] and therefore closer to real life medical practice. Also, the data from the charts are entered daily--rather than charts being abstracted every six months or so--with the database linked to automated medical records that the doctors use in their offices. Also, CHORUS includes data on quality of life, economics, and epidemiology.

ATN: What steps are taken to make sure that the data are entered correctly?

Dr. Becker: Quality control is essential, and CHORUS has a robust and overlapping quality control program to ensure that all data is accurate. I cannot comment on other observational studies, but the CHORUS quality control program is very inclusive.

ATN: How can people submit research ideas to the CHORUS team?

Dr. Becker: Through a member of the advisory board. The board prioritizes the questions, based on our view of their importance, because we only have so many analysts and biostatisticians. We welcome ideas for database queries, including those from outside industry or academia.

*****

Comment

In a better world, everybody could choose whether or not to make data from their medical record available anonymously for research. The cost to the healthcare system would be small, since there are many reasons to computerize charts anyway for better patient management--and development costs could be spread among millions of patients. Privacy concerns are greatly reduced if the data is anonymous and even the anonymous individual patient records are never released from the research center. We would all benefit, since such databases would make possible many research projects to improve the quality of care. For example, physicians could look for predictors of who is likely to do well with certain treatments or strategies, and who should avoid them.

With 6,000 patients when fully enrolled, the CHORUS database is large enough to be useful. It can also be a prototype for larger systems which computerize medical charts, extract anonymous information for research, and keep the research database current through daily updates.

References

1. Bigley JW, Becker S, Raffanti S, and others. Patterns of protease inhibitor use by healthcare coverage in the U.S.: Results from the CHORUS observational database. 12th World AIDS Conference, Geneva, June 28 - July 3, 1998 [abstract #42310].

2. Bigley JW, Raffanti S, Becker S and others. Viral detectability & CD4 levels in patients receiving protease inhibitors in clinical practice: Results from the CHORUS observational database. 12th World AIDS Conference, Geneva, June 28 - July 3, 1998 [abstract #42403].

Published twice monthly: Subscription and Editorial Office: P.O. Box 411256 San Francisco, CA 94141 800/TREAT-1-2 toll-free U.S. and Canada 415/255-0588 regular office number fax: 415/255-4659 Internet: aidsnews@aidsnews.org

980821
ATN30101

AEGiS presents published material, reprinted with permission and neither endorses nor opposes any material. All information contained on this website, including information relating to health conditions, products, and treatments, is for informational purposes only. It is often presented in summary or aggregate form. It is not meant to be a substitute for the advice provided by your own physician or other medical professionals. Always discuss treatment options with a doctor who specializes in treating HIV.

Copyright ©1980, 1998. AEGiS. All materials appearing on AEGiS are protected by copyright as a collective work or compilation under U.S. copyright and other laws and are the property of AEGiS, or the party credited as the provider of the content. .