The collective radiation dose to persons residing within 50 miles of the Three Mile Island nuclear power plant accident was at least twice as high as previously estimated, but the level was still so low that almost no health damage should occur, federal officials said yesterday.

At a Senate Subcommittee hearing yesterday, Health, Education and Welfare Secretary Joseph A. Califano Jr. said that based on the new data "we would expect one additional cancer death" and "one nonfatal cancer" above normal among the two million persons living around the nuclear facility near Harrisburg, Pa

He added that some scientists using the same data would predict as many as 10 additional cancer deaths and 10 additional nonfatal cancers.

At the same time, officials admitted, the data they have is so sketchy that they have no way of telling that how much radiation any single person may have received. Thus, their estimates of health effects are just educated guesses.

The difficulty in setting the extent of radiation exposure, they said, is because of the limited number of radiation monitors working outside the plant site during the first three days after the March 28 accident when most radiation was released. As reported at the by The Washington Post, the 17 thermal luminescent dosimeters (TLDs) maintained by plant operator Metropolitan Edison Co. were, during the initial key days, the only devices within 12 miles of the plant used to determine off-site exposure.

The first of about 200 federal government TLDs were not put in place until most of the radiation releases had ended.

The cumulative radiation exposure figure obtained from TLDs is important in setting dose rates, which in turn are needed to estimate future health risks to the exposed population.

On April 4, before a different Senate subcommittee, Califano predicted that there would be "no additional cancer deaths" as a result of the lowlevel radiation released by the accident.

At that time, within a week of the initial accident, the total population dose for persons within 50 miles of the plant was calcutlated by a Nuclear Regulatory Commission task force at 1,800 person-rems.

The term person-rem represents the estimated average dose within a geographic area multiplied by the number of persons living there. Person-rem measurements are used to estimate the future risk to the population of cancer.

Califano told the Senate subcommittee "as of mid-April," the official estimate was 3,500 person-rems.

At about the same time Califano was giving his figure, the NRC was being told by its staff that the current estimate is 2,000 to 4,000 person-rems.

An official said yesterday afternoon the number could go to 5,000 person-rems by Monday, when the NRC task force is to make public its latest report.

The NRC was told yesterday that the highest cumulative dose now calculated for an individual less than a mile from the plant is "about 90 millirems." A month ago it was 80 millirems.

At that level it still is below the 100 millirems that most Americans receive from natural background radiation sources.

One company dosimeter measured 905 millirems but it was on an uninhabited island adjacent to the facility.

NRC confusion over radiation measurements was illustrated yesterday when the commission's investigative and evaluation staff announced it could not find the source for the 1,200 millirems per hour reading over the plant stack that was publicized on Friday morning March 30.

That reading was instrumental in Pennsylvania Gov. Richard Thornburgh's decision to evacuate pregnant women and preschool-age children from a five-mile radius of the plant.

One NRC official said Metropolitan Edison's helicopter had made the measurement but NRC's investigative staff has not been able to vertify it.

An official said yesterday additional changes in the dose estimate may occur because the dosimeter were not accurate.

The commission was told that dosimeters used at Three Mile Island had been tested by the National Bureau of Standards and found in their readings to range from 20 percent lower to 30 percent higher than the known radiation source they were measuring.

In adition, the TLDs used may not have measured beta radiation. Another Bureau of Standards test on this is pending. It could lead to a sharp increase in the dose estimate, an NRC source said, if the TLDs are found to have measured only gamma, and not beta, radiation.

Nonetheless, he added, the projected health effects would be virtually unchanged.

One thing that may change, however, is the NRC policy toward offsite radiation monitors.

The NRC staff reported there are no requirements for how many dosimeters a nuclear plant operator should have nor are there criteria for the accuracy of such devices.

Currently, the nuclear plant owners must only present an acceptable safety plan that includes an offsite TLD program.