A consumer health group's study claiming to show that a third of heart pacemakers may be implanted needlessly is more than 98 percent wrong, according to a new medical report.

The report, in the American Journal of Cardiology, says the incidence of the implantation of needless or questionable pacemakers is less than 2 percent, not the 36 percent charged by the consumer group.

The original charge was made in a 1981 study by the Public Citizen Health Research Group, which studied pacemaker implants for two years in Maryland. The group found that about 23 percent were implanted needlesssly, and another 13 percent were "questionable."

The study said that the unnecessary surgery and implantation had cost $2.8 million in Maryland. Using that figure to estimate the problem in the rest of the nation, the report said, it is possible that 25,000 pacemakers are being implanted needlessly each year in the United States at a cost exceeding $280 million.

But the numbers in the original report were "flat wrong" by 95 percent, according to Leonard Scherlis, of the University of Maryland and the Maryland General Hospital, one of the authors of the new article.

The original report said that some 817 of 2,222 cases studied showed questionable implantation.

Scherlis and co-author Donald H. Dembo now report that, after a study of patient records by the hospitals involved, it appears that only about 30--not 817--implants actually were needless or questionable.

The other 785 cases were errors, Scherlis said, caused by the method used by the Health Research Group.

The Health Research Group used computer records from the Maryland Health Services Cost Review Commission. They compared the diagnosis of a patient's heart problem listed in the computer with a list of generally acceptable reasons for pacemaker implantation.

Scherlis said that the computer records were wrong so often, or omitted so many crucial details, that the method was "worse than useless."

For example, he said, 16 percent of the pacemakers counted as permanent implants in the computer records actually were only temporary. In 53 percent of the cases, Scherlis and Dembo reported, the significant fact which justified the pacemaker simply was omitted inadvertently from the record.

Allen Greenberg, one of the chief authors of the Health Research Group's study, said it was necessary to use the computer records because hospital records were not available.

He said that while some inaccuracies are bound to be present in the records, the new study's 1.4 percent figure was "wrong" and "meaningless." He said the biases built into the Scherlis study are "incredible."

For example, he said, Scherlis sent a letter to the hospitals expressing doubts about the Health Research Group study, and in effect announced his intention to knock down its results, Greenberg said. Then, Scherlis asked the hospitals themselves to decide in each case whether implantation was justified.

Scherlis and Greenberg said that further study is needed.