
                         Roanoke Times
                 Copyright (c) 1995, Landmark Communications, Inc.

DATE: SUNDAY, March 27, 1994                   TAG: 9403270033
SECTION: NATIONAL/INTERNATIONAL                    PAGE: A-6   EDITION: METRO 
SOURCE: New York Times
DATELINE:                                 LENGTH: Medium

U.S. TO AUDIT DOCUMENTS IN BREAST CANCER REPORT

To help restore public confidence in a breast cancer treatment that has been shaken by news of falsification of research data in a major study, federal health officials now plan to audit all documents at the University of Pittsburgh, which coordinated the study.

Auditors will also check the original records at many of the 89 hospitals in the United States and Canada where 1,843 women took part in a surgical study that helped change the way breast cancer is treated by declaring that a partial mastectomy and radiation are just as effective as the more disfiguring full mastectomy.

The audit is expected to take several months, and it may later be extended to 13 other cancer studies.

In a sharply worded letter Tuesday, Dr. Bruce Chabner, a senior official of the National Cancer Institute, informed Dr. Bernard Fisher, the head of the Pittsburgh center, that an auditing team would arrive early Thursday.

An investigation by the federal Office of Research Integrity found that a Canadian scientist, Dr. Roger Poisson of St. Luc's Hospital in Montreal, falsified data he submitted for the 14 studies.

The investigation results were published last year in a paragraph in the Federal Register but did not come to widespread attention until they were discussed in a Chicago Tribune article March 13.

Officials of the National Cancer Institute were alerted three years ago that data may have been falsified. Last week, they stood by their assurances that there is no reason to doubt the major conclusions of the Pittsburgh study.

The University of Pittsburgh has also reanalyzed its data - excluding that from Poisson - and said that the study's original conclusion remains valid. The cancer institute has reviewed this reanalysis and so far found its methodology correct but, in a further effort to reassure the public, will audit some raw data.

The Pittsburgh study, published in 1985, found many women with early breast cancer can be treated with a partial mastectomy, which is also called a lumpectomy, together with radiation.

Lumpectomies were just as effective in preventing recurrences as were full mastectomies, the study found. Several independent studies have confirmed the value of lumpectomies.

But federal officials are facing mounting criticism from Congress, angry women and skeptical doctors, both for the delay and for not independently checking the medical records on which the University of Pittsburgh analysis was based.

The House subcommittee on oversight and investigations plans to hold a hearing on the matter April 13.

by CNB