Too Little Information Too Late
THE TRUTH WAS more difficult to suppress this time than it had been in the atmosphere of fear engendered in the cold war of the McCarthy years, the Cuban missile crisis, and the Vietnam war. Nonetheless, an attempt to keep the facts from the people was clearly being made.
The same Freedom of Information Act that had made it possible for the Washington Post to reveal how the truth about the bomb tests in Nevada had been kept from the American people was used only a few days later to provide the first clue to what really happened at Three Mile Island.
On Monday, April 16, 1979, only two days after Bill Curry's story had gone out over the wires, the Pittsburgh Post-Gazette carried across the top of its front page a special report from the Associated Press with excerpts of tape recordings of the proceedings of the Nuclear Regulatory Commission during the crisis at Three Mile Island. Strangely enough, as in the case of the Watergate affair, it would be the private conversations of top government officials recorded on magnetic tape and unexpectedly released to the news media that would provide the crucial information.
My own awareness of the existence of the tapes had begun with a rather amusing phone call from a Washington Post reporter. He wanted me to comment on a not very complimentary, and very bitter and sarcastic remark made by one of the NRC staff members on what I was likely to say about the doses that would be received by the people near Three Mile Island during the first few days of crisis. When the excerpts of the NRC officials' comments appeared in the Post-Gazette, I saw immediately that my worst fears about the true magnitude of the radiation doses were likely to have been correct. I was sure that there would once again be a large rise in fetal deaths, congenital abnormalities, infant mortality, and childhood leukemia, followed by the delayed rises in infectious diseases, heart disease, and cancer among those exposed in the years to come.
The excerpts opened with the following comment by Lee V. Gossick, Executive Director for Operations of the NRC on Friday, March 30, the day after I had urged the evacuation of pregnant women and young children at the news conference in Harrisburg:
"Bill, we have got a deteriorating situation up there with regard to some releases. The Governor is asking us to confirm what he is getting from the plant, which says that they had an uncontrolled release of stuff."
To this, Harold Denton, Director of the Office of Nuclear Reactor Regulation in charge of all NRC activities at Three Mile Island, answered:
"They are getting 63 curies per second, and I can't explain to you the mathematics, but what they are saying is if that's true, by comparing it with what we know the shutdown rate was and the measurements taken at the north gate, and those were yesterday, they were on the order of three times what they were yesterday, which would put us in the 1200 millirems per hour."
These were truly enormous release rates of fresh fission gases, since in a single hour (consisting of 3600 seconds) there would be 3600 times 63 or some 226,000 curies being released in an uncontrolled manner without detailed analysis or significant hold-up to allow the most dangerous short-lived isotopes to decay. Even Millstone, the worst of all reactors, had not released more than fifteen times that much in a whole year.
The corresponding radiation dose of 1200 millirems per hour, equivalent to the dose of some 50 to 100 chest X-rays, was presumably measured in the narrow plume close to the release point at the plant, since it was ten thousand times the dose-rate that I had recorded the previous day in the airplane and on the ground a few miles away in the pockets of trapped gas. The road into Harrisburg was evidently well away from the narrow, invisible plume of radioactive gases meandering upstream, to the northwest, toward Middletown and Harrisburg with the slowly moving prevailing winds. Wherever it touched down, it would lead to an enormous inhalation dose in a very short time.
The taped conversation continued with the following question by Peter Bradford, one of the NRC commissioners who was not trained as a nuclear engineer:
BRADFORD: What actual measurements do you have?
GOSSICK: I can't give you any at the moment. I don't have anything that is current since this happened here, you understand. The source has been sealed up again and I think this is probably being released for one or two hours. We don't know, however, whether that's good for any period of time.
GILINSKY (NRC Commissioner): Do we have any monitoring equipment --
DENTON: There is a lot up there, Vic, but it takes a while for it to ever get back here.
FOUCHARD (Director of Public Affairs for the NRC): I just had a call from my guy in the Governor's office and he says the Governor says the information he is getting from the plant is ambiguous, that he needs some recommendations from the NRC.
DENTON: It is really difficult to get the data. We seem to get it after the fact. They opened the valves this morning, on the let-down, and were releasing at a six-curie-per-second rate before anyone knew about it. By the time we got fully up to speed, apparently they had stopped, there was a possible release on the order of an hour or an hour and a half --
GILINSKY: And when did this plume -- when was the puff released?
DENTON: Within the last two hours.
HENDRIE (Chairman of the NRC): Presumably it has just terminated recently, then.
DENTON: We don't know how long, but if it was a continuous release over a period of an hour or an hour and a half, which from what I understand which is a kind of lot of puff.
HENDRIE: A couple-of-knots wind and the dammed thing -- the head edge of it is already past the five-mile line.
HENDRIE: There has been a suggestion for a five-mile evacuation in the northeast direction. I take it --
DENTON: A good five miles, I would say from first impression.
BRADFORD: It ought to be made clear that you are not talking about lethal doses.
FOUCHARD: Mr. Chairman, I think you should call Governor Thornburgh and tell him what we know. I don't know whether you are prepared at the present time to make a commission recommendation or not. The Civil Defense people up there say that our state programs people have advised evacuation out to five miles in the direction of the plume. I believe that the commission has to communicate with the Governor and do it very promptly.
GILINSKY: Well, one thing we have got to do is get better data.
FOUCHARD: Don't you think, as a precautionary measure, there should be some evacuation?
HENDRIE: Probably, but I must say, it is operating totally in the blind and I don't have any confidence at all that if we order an evacuation of people from a place, where they have already gotten a piece of the dose they are going to get into an area where they will have had 0.0 of what they were going to get and now they move some place else and get 1.0.
GILINSKY: Does it make sense that they have to continue recurrent releases at this time?
DENTON: I guess I tend to feel that if they really didn't stop the release a half an hour ago -- it's probably best to leave it to the operational people up there. The cloud hasn't had a chance to get down to these low levels.
AHEARNE (NRC commissioner): But Harold, what confidence do you have that they won't embark on the same thing?
DENTON: I don't have any basis for believing that it might not happen -- is not likely to happen again. I don't understand the reason for this one yet.
HENDRIE: It seems to me that I have got to call the Governor . . . to do it immediately. We are operating almost totally in the blind, his information is ambiguous, mine is nonexistent, and -- I don't know, it's like a couple of blind men staggering around making decisions.
This last remark by Chairman Joseph M. Hendrie was to be quoted repeatedly by the news media later, but strangely enough it was always referred to in the context of the danger of a major melt-down. But clearly, either by reason of ignorance or design, the public was not told that the remark was made in the context of the danger from the ongoing large gas releases that were rapidly approaching the critical EPA emergency dose of 25,000 millirems to the whole body or key organs such as the thyroid from internal and external sources combined. This high level had been set by the Federal Radiation Council back in 1965 and later adopted by the EPA as the maximum allowable dose before countermeasures or an evacuation of the population should be ordered. Certainly neither the people living nearby nor the reporters at the site had ever been told of the large radiation doses from the inhaled gases that they were receiving.
The excerpts from the tapes continued to record the enormous concern about the growing radiation doses from the inhalation of the gases to the people in the area:
FOUCHARD: Is there anybody who disagrees that we ought to advise the Governor on what to do?
DENTON: I don't. Just on the basis of what we know. It's a good first step.
HENDRIE: Go ahead with the evacuation?
DENTON: I certainly recommended we do it when we first got the word, commissioner. Since the rains have stopped and the plume is going -- I would still recommend a precautionary evacuation in front and under. And if it turns out we have been too conservative --
GRIMES: My view is that it might have been useful right near the site, but now it is down below the EPA [unintelligible] level, so it probably is the most that should be done, in my view, is to tell people to stay inside this morning.
AHEARNE: I was going to ask, what about pregnant women and children?
GILINSKY: Well, Brian says it is a factor of 10 that can be gained by staying indoors. Anyway, I just think it is worth getting that half hour to find out, first of all, you are alerting people that they are going to have to do something, and they are not going to be able to do something in a half-hour anyway.
HENDRIE (on telephone): Governor Thornburgh, glad to get in touch with you at last. I must say that the state of our information is not much better than I understand yours is. It appears to us that it would be desirable to suggest that people out in that northeast quadrant within five miles of the plant stay indoors for the next half hour.
We have got one of those monitoring aircraft up and seem to have an open line to it and we ought to be able to get some information in the next 10 to 15 minutes. They can tell us whether it would be prudent to go ahead and start an evacuation out in that direction.
THORNBURGH: So your immediate recommendation would be for people to stay indoors?
HENDRIE: Yes, out in that -- out in the northeast direction from the plant.
THORNBURGH: The northeast direction from the plant to a distance of?
HENDRIE: To a distance of about five miles.
I have got a reading. During one of these burst, releases up over the plant several hours ago, up over the plant about 1200 millirems per hour which seems to calculate out, by the time the plume comes to the ground where people would get it, would be about 120 millirems per hour. Now, that is still below the EPA evacuation trigger levels; on the other hand, it certainly is a pretty husky dose rate to be having off-site.
Here it was: the NRC knew that the true doses were not just a few millirems to the people in the area, as had been claimed at the Kennedy hearings, with maximum values of the order of 75 to 100 millirems nearest to the plant.
They knew just as I did that the greatest dose arose not from the external gamma radiation measured by a survey meter or a film badge, but from the internal beta radiation from the inhaled fission gases and particles in the lung, the thyroid, and the other critical organs that concentrate the different substances according to their various chemical properties. So when the external gamma-dose rate on the ground was of the order of 1 to 2 millirems per hour, the true dose rate to the lung and other critical organs could be as much as 50 to 100 times greater, or of the same general magnitude as the 120 millirems per hour Hendrie himself had just mentioned.
But they also knew, as I did, that if they ever were to order the full evacuation that should have been ordered long before, it would not only have caused a panic among the completely confused and unprepared population, it would also have been the end of the nuclear industry, whether or not the core would ever go to a complete melt-down.
And so more precious time in which to save lives was being lost by recommending only that people should stay indoors, as was clear from the taped conversation with Governor Thornburgh. At that very moment Governor Thornburgh was under enormous pressures from those wanting to protect this important Pennsylvania industry on the one hand, and his Secretary of Health, Gordon MacLeod, on the other hand, who was at that very time urging that at least the pregnant women and young children should be evacuated.
But the NRC was clinging to the hope that an evacuation of any kind would not have to be ordered because of the continuing gas releases:
KENNEDY: Don't we know that it has been stopped?
FOUCHARD: Vice-President Herbein of the company reports it has been stopped. Chairman Hendrie has talked to the Governor and recommended that he advise people to stay indoors up to a distance of about five miles for the present time.
GOSSICK: We have just lost telephone contact with the site. I assume that it is telephone problems, but -- [inaudible]. Okay, we have got communications with a trailer up there, but we have lost contact with the control room. . . .
GILINSKY: Let me ask you, what is the status of the reactor, by the way?
CASE: Same as it was an hour ago, I guess.
GILINSKY: What is the state of the core?
CASE: Well, it is about like it was yesterday. The temperature is about 280 degrees, the pressure is up [inaudible] . . .
Here was the proof that up to that very moment the primary consideration in the decision to evacuate the people was not the awareness of a serious chance of a melt-down of the core, but the continuing uncontrolled releases of radioactive gases that were threatening to give internal exposures to the public on the order of tens of thousands of millirems.
In fact, the subsequent excerpts made it clear that only at this point in the desperate deliberations did the knowledge of extensive core damage and a hydrogen explosion reach the NRC commissioners:
HENDRIE (TO DENTON): I have talked to the President and I think you ought to go down to the site. He will be sending down, immediately, a sort of communication system that he takes with him when he travels. He will send somebody with it and he wants to be in a position to pick up the telephone and go right through to the site, and be able to talk to his man down there for information and recommendations on what to do. . . .
MATTSON: [Director of the Division of Systems Safety] . . . Now, B and W [Babcock and Wilcox] and we have both concluded . . . that we have extensive damage to this core. That corroborates with the releases we are seeing. . . . My best guess is that the core uncovered, stayed uncovered for a long period of time, we saw failure modes, the likes of which has never been analyzed. . . . We just learned . . . that on the afternoon of the first day, some 10 hours into the transient, there was a 28-pound containment pressure spike. We are guessing that may have been a hydrogen explosion. They, for some reason, never reported it here until this morning. That would have given us a clue hours ago that the thermocouples were right and we had a partially disassembled core.
HENDRIE: Where abouts is the bubble?
MATTSON: The bubble is in the upper head. The upper head volume at 1128 cubic feet as best we can tell. The estimate of the gas in that volume now is 1000 cubic feet, best that we can tell. That is at 1000 psi. If you take the plant to 200 psi, then --
HENDRIE: Yes, you are going to blow right down and empty the core.
MATTSON: I have got a horse race. I'm putting in high head, and if I get down in pressure, low head and coolant, it is coming in the cold leg, it is going down to the lower plenum, it is coming up through the core, it is splashing and it runs into the noncondensibles, I've got a core partially full or maybe totally full of noncondensibles. . . . We have got every systems engineer we can find, except the ones we put on the helicopter, thinking the problem, how the hell do we get the noncondensibles out of there? . . . Do we win the horse race or do we lose the horse race? And if you are lucky and there is not a lot of -- you have overestimated the noncondensibles, you might win. If you are not lucky and you have got the right number on the noncondensibles you might lose it.
HENDRIE: . . . It sounds to me like we ought to stay where we are. I don't like the sound of depressurizing and letting that bubble creep down into the core.
And then came the following most revealing words from Roger Mattson, Director of the Division of Systems Safety, whom I knew to be very knowledgeable with regard to radiation dose calculations.
MATTSON: Not yet. I don't think we want to depressurize yet. The latest burst didn't hurt many people. I'm not sure why you are not moving people. Got to say it. I have been saying it down here. I don't know what we are protecting at this point. I think we ought to be moving people.
HENDRIE: How far out?
MATTSON: I would get them downwind, and unfortunately the wind is still meandering, but at these dose levels that is probably not bad because it is [inaudible].
KENNEDY: But downwind how far?
MATTSON: I might add, you aren't going to kill any people out to 10 miles. There aren't that many people and these people have been -- they have had two days to get ready and prepare.
KENNEDY: Ten miles is Harrisburg.
MATTSON: 40,000 [inaudible] five miles. . . . It's too little information too late, unfortunately, and it is the same way every partial core melt-down has gone. People haven't believed the instrumentation as they went along. It took us until midnight last night to convince anybody that those goddamn temperature measurements meant something. By four o'clock this morning, B and W agreed.
Not until later in the day did Governor Thornburgh finally agree to the compromise plan of ordering evacuation of pregnant women and young children. But as I was to learn in an unexpected manner a few months later, most of the damage had already been done before the evacuation was ordered.
During the latter part of the summer I received a phone call from someone who asked me for a collection of all my articles on the effects of low-level radiation. He said that he was working on a study of the total economic and health impact of the Three Mile Island accident for the Presidential Commission chaired by Dr. George Kemeny of Dartmouth College, and he wanted to include an upper-limit estimate based on my statistical findings around various nuclear plants and after various fallout episodes.
I was surprised by this request, since I had not been asked to testify before the Kemeny Commission, which had been holding extensive hearings all summer long. Also, I wondered how the caller could use my earlier studies if there were no detailed estimates of releases and dose measurements available. The NRC had claimed that all the meters in the stack had gone off scale so that no one knew or would ever know just how much had actually been released.
And so I asked whether he had any detailed information on the quantities of gases released or the amounts of radioactivity in the air, the milk, and the diet on which to base a meaningful estimate of the likely health effects and their costs.
To my surprise, he said that he had such a document prepared for the Metropolitan Edison Company by their own environmental consultants, Pickard, Lowe and Garrick, and that he would be glad to send me a copy for my examination, since it was being widely circulated within the NRC and the Kemeny Commission.
A few days later, the two-inch-thick document entitled "Assessment of Offsite Radiation Doses from the Three Mile Island Unit 2 Accident TDR-TMI-116," dated July 31, 1979, arrived in the mail. And there, in the second paragraph, was the proof that the evacuation ordered on the third day had indeed been too late:
Based on techniques used in this analysis, dose estimates are consistent with the release of seven million curies of noble gases in the first one-and-one-half days of the accident, two million in the next two days and one million in the next three days, and a relatively small amount thereafter.
By Friday afternoon, the third day of the accident, when the evacuation took place, between 7 and 9 million curies out of an estimated 10 million curies of radioactive fission gases had already been released, together with a corresponding fraction of the 14 curies of radioactive iodine 131 as given on page III of the report. And since the report also concluded that most of the thyroid dose was due to inhalation -- and not ingestion of drinking water or milk -- in the first five days of the accident, it was clear that by the time the evacuation of the pregnant women had been ordered, most of the thyroid dose to the developing fetuses had already taken place.
These were indeed very large amounts of radioactivity, comparable to those that arrived from the Chinese bomb tests in October 1976 on the East Coast, and for which I had found a 20 to 60 percent increase in infant mortality in the following three months all the way from Delaware to Maine. Just two years earlier, I had prepared a paper on this incident for the Committee on the Biological Effects of Ionizing Radiation of the National Academy of Sciences for its meeting in Washington on July 17, 1977, in which the levels of radioactive iodine 131 in the milk reported by the EPA were summarized.
Looking at the Metropolitan Edison Company's own measurements in some of the nearby towns, I saw it was evident that levels of iodine 131 had been produced by the Three Mile Island accident comparable to those that had been produced by the heavy rainout of the drifting Chinese fallout as it met a severe rainstorm moving up the East Coast from Delaware to Maine. For the Hardison Farm, to the north of Three Mile Island, the report listed 110 picocuries per liter on April 25, almost a month after the accident. By that time, the activity would have decayed away to only about one-eighth to one-sixteenth of its maximum value, so that the peak values in local farms, if there had been adequate monitoring at the time, would have been in the range of 800 to 1600 picocuries per liter. The amount one worried about was measured in picocuries, which was only a millionth of a millionth, or a trillionth, of one curie. Yet 14 curies or 14 trillion of these units had been discharged into the air, breathed by the pregnant women in the area, and added to what they ingested with the milk. The report showed concentrations in the air as high as 20 picocuries per cubic meter. Since 1 cubic meter was roughly the volume of air inhaled by an adult every hour, 20 picocuries of iodine 131 entered a pregnant woman's lungs each hour at these concentrations.
The milk concentrations compared with the highest values listed in the EPA's December 1976 tabulation published two months after the Chinese fallout had arrived. These had ranged from 36 picocuries per liter in Rhode Island to 123 in Connecticut. At that time, infant mortality rose 60 percent in the first quarter of 1977 compared to the same period in 1976 in Delaware, 41 percent in New Hampshire, 17 percent in Maine, and 13 percent in Connecticut. But what was even more significant was that in Massachusetts, where the health department had ordered the cows to be fed stored hay which was relatively free from fresh fallout, infant mortality continued its rapid decline. Furthermore, infant mortality kept decreasing in the United States as a whole by 7 percent, as it did in Rhode Island, which receives most of its milk from Massachusetts.
Thus, if the measurements reported by Picker, Lowe and Garrick were accurate, there simply would have to be a sharp rise in infant mortality in the Harrisburg area and those parts of Pennsylvania and other nearby states over which the radioactive gases released in the first three days had drifted. On the other hand, infant mortality should again continue its normal decline in areas that happened to be spared by the invisible clouds of radioactive gases from the Three Mile Island nuclear plant.
But in which direction was the wind blowing during the period of highest release rates? Was there any way to find out how much was coming out at any given moment? After all, according to a story sent out by the Washington Post News Service and published on April 22, the Nuclear Regulatory Commission had been told by one of its staff people, Albert Gibson, that the radiation monitors in the stack went off scale on the morning of the accident. Thus, in answer to Commissioner Victor Gilinsky's question, "So we don't really know what went up," Gibson replied, "That's correct."
The story went on to say that as much as 365 millirems per hour of beta and gamma radiation were recorded on the ground some 1000 feet from the stack, and a helicopter had recorded three times this level in the air over the vent, confirming once again that the dose rates were far higher than the public had been told at the Senate hearings when the beta radiation that accounts for most of the internal dose from inhaled gases is taken into account. But Gibson went on to say that "those measurements were very inconclusive," and that "without knowing the precise weather patterns, we don't know if they were made at the appropriate locations."
However, leafing through the report of Met Ed's environmental consultants, I found that these were all completely misleading statements. Contrary to what the NRC commissioners and the public were being told, there were radiation monitoring instruments in the plant that never went off scale, namely in the auxiliary building, whose readings were directly related to the amount of radioactive gas being released. Here was the way the utility's consultants described how it was possible to know how much gas was being released every moment:
Strip chart records from all noble gas radiation monitors in the plant ventilation exhaust show no significant radiation levels during the first three hours of the accident. Since these monitors are in the most probable pathway for release, it is concluded that no significant releases occurred before 0700 March 28. Shortly after 0700, however, these monitors, which are designed to read normal low levels, indicated rapidly increasing radiation concentrations. Within a few minutes, they went off scale on the high side. At about the same time, the in-plant building area monitors which measure radiation levels inside the fuel handling and auxiliary buildings began to record increasing levels from about 1 milliroentgen to 100 milliroentgen per hour at 0740. At about 0900 the readings began to increase again to reach about 100 milliroentgen per hour at 1000 hours. They continued to fluctuate at high levels for about four days. One or more of these area monitors continued to read on scale during the course of the accident.
The report went on to explain in detail that by means of these measuring instruments it was possible to know what went out the stack because:
. . . radiation levels measured by area monitors in the auxiliary and fuel handling buildings are proportional to the rate at which airborne gamma activity was released to the environment . . .
In table after table and chart after chart, the releases and gamma radiation doses in different directions were worked out in detail. For every hour of the accident from 4 A.M. on March 28 until midnight of the fourth day, the readings of the area monitors were given together with the hourly wind direction and wind speeds. It showed that during the period of highest releases, from 10 A.M. on Wednesday the 28th to 7 A.M. on Thursday the 29th, the winds were blowing north, northwest, and west at 6 to 9 miles per hour, sending the radioactive gas toward upstate New York and western Pennsylvania. Only later, when the rate of release had decreased tenfold, did the winds shift briefly to the south, becoming more variable thereafter.
By the time the winds were blowing toward the northeast on Saturday, the fourth day of the accident, the intensity had dropped to less than one-twentieth of its peak value, thus largely sparing the most densely populated areas of Philadelphia, New Jersey, and New York City.
No wonder the NRC staff did not want to let the public know that they knew exactly in which direction the most radioactive clouds had moved, since this information could then be used to tie any later localized rises in fetal deaths, infant mortality, and cancer to the radioactive gas clouds from Three Mile Island. In fact, I remembered only too well the attempts of some of the same individuals formerly working for the AEC to discredit my findings on the rises of infant mortality across the southeastern United States following the first nuclear-bomb test at Alamogordo by claiming that the winds were not blowing in that direction. And they certainly did not want any of this to become known before the Kemeny Commission was scheduled to complete its report in early November.
Once again, as in the case of the Nevada tests, it was essential to keep such knowledge from the public and the scientific community at large. The NRC, the EPA, and all the other federal and state agencies knew full well that the doses were comparable with those experienced by the people of Utah, Montana, Wyoming, and the other states across the northern United States as far as New York and New England during the period of the Nevada tests, or for releases from some of the largest and most heavily emitting reactors, such as Millstone in New London, Connecticut, over a period of a year or two.
If, indeed, there should once again be sudden rises in infant mortality in areas where the radioactive clouds had drifted and the public should learn of them when the televised nightmare of Three Mile Island was still fresh on everyone's mind, this public knowledge would threaten the government's and the nuclear industry's vast program to build a thousand of these giant reactors by the end of the century near all the major cities, and would result in costly damage suits, exactly as in the case of the Nevada tests.
I had tried to obtain the preliminary data on monthly infant mortality rates by county from the Health Department in Harrisburg without success. I was told that this data had not yet been sorted out and processed, and that it would be many months before it could be properly assembled, even though such data had already been sent to me by the Maryland Health Department. All research on the health effects of Three Mile Island in Pennsylvania were under the direction of Dr. George Tokuhata, the same man who prepared the statistical portion of the Shapp Report exonerating Shippingport back in 1973. There was clearly no hope for any help from that direction.
All I could do was to wait for the state-by-state data on monthly infant deaths and births published by the Center for Health Statistics in Washington, which was usually three or four months behind. Thus, if significant effects would first show up for infants two to three months after the accident, or for the months of May and June, the earliest numbers indicating an effect would not become publicly available until August or September, too late for any presentation to the Kemeny Commission.
Nevertheless, I decided to gather whatever data I could as soon as possible, for an enormous media campaign had been launched by the nuclear industry to convince the public that there were no serious health effects due to the accident at Three Mile Island. No one had died, and no member of the public had been injured; the safety systems had worked, and there was no reason to abandon this important source of energy at a time when the United States depended so heavily on imported oil from the unstable Middle East.
In mid-August, the latest monthly report from the U.S. Center for Health Statistics for the month of May arrived in the library. Calculating the rates of infant deaths per 1000 live births, I found what I had expected. Instead of declining from the winter high, infant mortality in Pennsylvania had gone up following the accident at the end of March. Compared to 147 deaths in February and 141 in March, there had been 166 in April and 198 in May, an unprecedented rise of 40 percent. Yet, the number of births had actually declined from 13,589 in March to 13,201 in May. Thus the rate of infant deaths per 1000 live births had increased even more, namely by 44 percent, from 10.4 in March to 15.0 in May.
Yet, at the same time, the rate for the United States as a whole between March and May had declined 11 percent as it normally did, dropping from 14.1 to 12.6 per 1000 live births.
These were highly significant changes, the Pennsylvania figures for March and May representing an increase of 57 deaths, which was more than three times the statistically expected normal fluctuation of about +/- 16, and thus unlikely to occur purely by chance in less than one in a thousand instances. But how else could I test the hypothesis that these increased rates were likely to be due to the releases from Three Mile Island without having the county-by-county and the Harrisburg figures available to me?
Having learned from the utility's own report that the heaviest releases had occurred when the wind was blowing north, northwest, and west, and having seen reports in the papers of high levels of radioactivity being measured in Syracuse, New York, some 150 miles to the north, at the time of the accident, I decided to examine the figures for New York State. By a fortunate coincidence, the U.S. Vital Statistics gave separate figures for New York City and the rest of the state, most of whose population was located in upstate New York -- north, northwest, and northeast of Harrisburg some 100 to 200 miles away. Here, then, was a clear prediction of the hypothesis that could be tested: The figures for the rest of the state outside of New York City should have gone up, while New York City should either have shown no change or an actual decline.
And this is exactly what the numbers showed: Between March and May, infant deaths outside New York City climbed an amazing 52 percent, by 63 deaths, from 121 to 184. For New York City during the same period the number declined from 166 to 129. Again, these changes were many times as large as normal fluctuations, and the number of births changed relatively little, or by less than 10 percent, so that there could be no doubt about the significance of these changes in infant deaths.
So far, the hypothesis had passed its first major tests, but would it hold up for the other nearby states? What about Maryland to the south, where some of the gases had drifted in the morning and afternoon of the second day, according to Met Ed's report. The numbers were smaller than for the more populous states, but the changes continued to support the hypothesis: Infant mortality rose 26 percent, from 39 to 49 deaths, while the number of births remained essentially unchanged -- 4013 in March and 4076 in May.
What about New Jersey, to the east and northeast of Harrisburg? If the hypothesis was correct, there should not be any significant increase, since by the fourth day of the accident, when the winds shifted toward New Jersey, the rate of release had already sharply declined. Again, the hypothesis held up under the test. Between March and May, New Jersey infant mortality rose by only 8 deaths, from 87 to 95. This is not considered a significant increase, with the spontaneous statistical fluctuation of about +/- 13 normally expected.
What about Ohio to the west, which had for decades closely paralleled Pennsylvania in its declining infant mortality figures? Did it show the same 40 to 50 percent rises of Pennsylvania and of New York State outside New York City? It was not to be expected that the gases would have drifted more than 200 to 300 miles west of Harrisburg, counter to the generally prevailing west-to-east movement of air masses across the United States, and so there should really be very little change in the Ohio figures. Again the numbers bore this out: Between March and May, infant deaths declined in Ohio from 177 to 160, the rate remaining constant at 11.5 per 1000 live births.
There simply could be no other explanation for such a localized pattern of sharply increased infant deaths in the areas where confined winds had blown the radioactive gases, while infant mortality rates were steady or declined in all the surrounding states that were not in the direction of the winds during the first two days of highest releases.
And yet, it would be important to have some figures for the area that was most heavily exposed close to the plant. If the figures around Harrisburg should indeed show much higher rises in infant mortality than the 44 percent for Pennsylvania and the 50 percent for New York State outside New York City, then it would be difficult to reject the hypothesis that it was indeed the radioactive gases from the stricken plant that were responsible for the unusual increase in newborn deaths.
But only Tokuhata had the data for the 5-mile and 10-mile zones around the plant, and there was no way that I would be able to obtain them. From everything I had been able to piece together, the numbers in Harrisburg had to show very large increases if areas as far away as 150 miles in upstate New York showed 50 percent rises in infant deaths. From my earlier studies on fallout clouds, I knew that the effect would roughly decrease in inverse proportion to the distance from the point of release. Thus, for Harrisburg, only 10 miles to the north of the plant, the rise in newborn infant mortality would have to be as high as 300 percent to 600 percent, corresponding to a four- to sevenfold increase above normal to be consistent with the rises in upstate New York.
How could I get at least an indication of whether this was the case? Just at this very moment, a way opened up to obtain this crucial information without the need to obtain access to the Pennsylvania Health Department's carefully guarded data.
Earlier in the summer I had been invited to address a public meeting in Harrisburg on the likely health effects of the accident at Three Mile Island. At the end of the meeting, someone introduced himself to me, and asked me whether he could be of any help. His name was Warren L. Prelesnik, and he told me that he was deeply concerned, since he had just moved to the area with his family shortly before the accident, and that he was working in the Harrisburg Hospital as executive vice-president in charge of administration.
I asked him whether it might be possible to obtain information on the monthly number of births and infant deaths together with their cause over the past few years, and he said that he would try to see what he could do.
More than a month later, when I had already given up hope of receiving any information, there arrived in the mail a letter with a list of the monthly infant deaths, fetal deaths, stillbirths, and live births in the Harrisburg Hospital for the previous two years.
At first there seemed to be no obvious change, if one looked at the total numbers of all types of fetal and infant deaths combined. But then I examined separately the category of newborn or neonatal infant deaths -- those that were born alive but died within the first year, but mainly in the first few hours after birth. Here was the evidence I needed. In February, March, and April of 1979, there had only been 1 infant death per month. But for each of the two months of May and June, there were 4. Effectively, since the number of births had not only remained nearly the same but had actually declined slightly, this was more than a fourfold increase in the mortality rate, or of the right magnitude required to fit the observed 50 percent rise in the more distant area of upstate New York.
From an average of 5.7 per 1000 live births in the three months of February, March, and April -- before the releases could have had an appreciable effect -- the newborn mortality rate had risen to 24.1 for May and 26.0 for June, an unprecedented summer peak that did not occur the previous year. In fact, for May and June of 1978, there had been a total of only 3 infant deaths, while for the same period in 1979 after the accident, there had been 8.
As some of my colleagues with whom I discussed these findings agreed, by themselves the Harrisburg Hospital numbers were of course small, and only marginally significant, representing only about one-third of all the births and deaths in Harrisburg. But taken together with the vastly more significant and independent numbers for all of Pennsylvania, upstate New York, New York City, New Jersey, Maryland, and Ohio, there was now a much greater degree of certainty: It would have been much too much of a coincidence -- perhaps less than one in a million -- for all these different numbers to show the pattern they did.
At this very moment, there arrived an invitation to address an international meeting of engineers and scientists in Israel. The meeting was devoted to studying the environmental problems of industrialization of Third World nations. It was to be held in December, and I was asked if I would be willing to address the meeting on the environmental health problems connected with nuclear energy.
I could hardly believe that such an opportunity to bring these facts into the open should come at precisely the time when the data from the Harrisburg Hospital had convinced me more strongly than ever of the great danger of nuclear reactors. I decided to accept the invitation and to devote myself in the remaining few months to preparing a detailed paper that would begin with a review of the previous evidence on the effects of low-level radiation from fallout and normal nuclear plant releases and end with the evidence for the rise of infant deaths after Three Mile Island.
One of the remaining important questions that had to be checked, however, was the time and cause of death. Clearly, if the excess deaths were connected with the radioactive iodine released from the plant, then they should be associated with underweight births or immaturity, since damage to the fetal thyroid would slow down the normal rapid growth and development of the baby in the last few months before birth. The development of the lungs, which have to be ready to begin breathing at the moment of birth, is one of the most critical phases of late fetal development. Any developmental slowdown would be most life-threatening if it led to the inability of the tiny air sacs in the lungs to inflate and start supplying the blood with oxygen. Failure of the lungs to function properly would therefore lead to immediate symptoms of respiratory distress, and if efforts to treat the baby should not succeed, it would die in a matter of minutes, hours, or days of respiratory insufficiency or hyaline membrane disease.
Thus, one would not expect to find as large an increase in spontaneous miscarriages well before birth as newborn deaths within a short time after birth, since the lungs did not need to start functioning until the baby was born. Also, there should be no significant increase in gross congenital malformations a few months after the accident, since by the time the baby in the mother's womb had reached the sixth or seventh month of development, all the major organs had already fully developed. Thus, only some six to seven months after the accident would one expect some increase in serious physical malformations, since these infants would have been exposed to radiation in the first three months of development of critical-organ formation.
The data from the Harrisburg Hospital supported these expectations. There was much less of an increase in the number of spontaneous miscarriages and stillbirths than in the number of newborn babies that died shortly after birth because of immaturity and respiratory distress, indicating the strong likelihood that it was the effect of iodine 131 and other shorter-lived iodines such as iodine 133 that had damaged the ability of the thyroid to produce the necessary hormones needed for normal growth and development.
In fact, it was for this reason that I had publicly urged widespread screening for hypothyroidism at the time of the news conference in Harrisburg on the second day of the accident, the kind of simple test that could prevent permanent mental retardation if detected and treated early. This test was already being used routinely for every newborn baby born in hospitals of a number of states in New England, the Northwest Coast, and Pennsylvania, but not yet in New York or Maryland. There would have to be a rise in the incidence of this condition if my past findings on the increase in underweight births and subtle forms of mental retardation during the period of heavy nuclear-bomb testing were indeed related to the action of radioactive fallout. But not until many months later, long after the Kemeny Commission hearing had been completed, would I learn that a rise in hypothyroidism had already been discovered by the Health Department of the State of Pennsylvania among the newborn babies in areas where the invisible radioactive gases from Three Mile Island had been carried by the winds.
back to Secret Fallout | radiation | rat haus | Index | Search