CHANCE News 2.11 
               (2 June to 17 June 1993)

Prepared by Laurie Snell

 Please send suggestions for  articles or format to 

Previous issues of this chance news and other 
materials for a chance course are on the chance 
gopher.  Just point your gopher to

"First get the facts straight, then you're free to 
distort them at your leisure.

                                     Mark Twain



>>>>>==========>> Maya Bar-Hillel from the Hebrew University sends the following contribution: One of Israel's large dailies has at the moment promotion campaign. They advertise it as follows: GIGANTIC PRIZES -- NO LOTTERY. They then explain: You are holding a "half of the prize" in the form of half a photo (of a car, say), with a 7-digit number printed at the bottom. You are then instructed to look for the forms that come everyday with every copy of the paper. And if the form you get has the other half of the photo, and a matching number -- then you get the item photographed "without any lottery". <<<========<<

>>>>>==========>> For many fathers, roles are shifting. The New York Times, 20 June, 1993,A20. Douglas Martin This article was inspired by a national study being prepared by the Families and Work Institute. It is not really a statistical discussion of the problem but has a couple of interesting statistics. "The institute's study has found that in households where only the husband works the wife does 94 percent of the cooking and 93 percent of the child care. Where both parents work, the wife's share dips to 80 percent of the cooking and 70 percent of the child care." <<<========<<

>>>>>==========>> Implications of the Copernican principle for our future prospects. Nature, 27 May, 1993. Richard Gott III This is the source of the New York Times article reported in the last chance news about a method for obtaining 95% confidence limits on such things as how much longer a species (for example us) will survive I remind you how it works. Let PAST be the time from the beginning to the present and FUTURE the time from the present to the end of the species. Then with 95% confidence: 1/39*PAST < FUTURE < 39*PAST. Based on PAST = 200,000 years this leads to 95% confidence that we will last .2 million more years but less than 8 million years. Since the journal Nature has been published for 123 years its future publication will last more than 3.15 years and less then 4,800 years with 95% confidence Here Gott gives some indication of how he justifies his confidence intervals. For example, he assumes that intelligent species are being formed in the universe at a uniform rate and are subject to a constant but unknown extinction rate. Then PAST and FUTURE are independent random variables with common exponential density and so their ratio has distribution F(x) = x/(1+x). Thus, for example, P(FUTURE/PAST) > 39 = 1/40 = .025. The author connects his results with other results in cosmology such as an argument of Carter that" there must be of the order of 1 improbable event required for the formation of intelligent life"(Phil. Trans. R. Soc. A310, 347-363 (1983)). You will find lots more things like this discussed in this article. <<<========<<

>>>>>==========>> Survey finds major gains in cutting blood cholesterol. The New York Times, June 16, 1993, A18. Warren E. Leary Results of a survey being published in the current Journal of American Medical Association show that there has been a 4 percent decline in the average cholesterol level, to 205 in 1990 from 213 in 1978. The proportion of adults with very high levels fell to 20 percent from 26 percent. The article includes new guidelines from the National Heart, Lung and Blood Institute which stress continued efforts to lower cholesterol levels with an emphasis on adding exercise and weight loss to dietary changes to reduce the intake of fat and cholesterol. They also recommended delaying the use of cholesterol-lowering drugs or younger adult men and pre menopausal women with a low risk for developing heart disease. There has been a 54 percent decline in deaths from coronary heart disease in the last 30 years which has coincided with drop in cholesterol levels, declines in high blood pressure and cigarette smoking rates. Of course, the question remains open as to how much each of these actually contribute to this decline. <<<========<<

>>>>>==========>> Conference ends with little hope for AIDS cure. The New York Times, 15 June, 1993, C1. Lawrence K. Altman There was lots of news about the ninth international AIDS meeting in Berlin and this article pretty well summarizes this news. There seems little hope that new drugs will be available soon to save the lives of the 14 million people now infected with the HIV virus or that a vaccine will be developed in time to prevent the World Health Organization prediction of 30 million infected by the year 2000 coming true. On the other hand, experts stressed that steps can be taken to very much reduce the risk of transmission of the virus through education programs that lead to safer sex. Preventing HIV infection in women was given special priority because they are a fast-growing group in the AIDS epidemic and they can pass on the virus through their children. There was concern that teen-agers might represent the next wave in the AIDS epidemic. More information was provided about the European study that showed that early use of AZT was not effective. This topic will be the subject of a meeting at the National Institute of Health which is expected to make new recommendations about the optimal time to start AZT treatment. <<<========<<

>>>>>==========>> Board backs continuing quake prediction experiment. Gannett News Service, 3 June 1993. Earle Eldridge A report on the preliminary findings of a board to review the Parkfield experiment designed to improve earthquake forecasting. In 1985 it was predicted that a 6.0 or more earthquake would occur in the Parkfield area within four years of 1988. This was based upon major quakes in 1857, 1881, 1901, 1934, and 1966--an average of every 22 years. With no such quake by Dec. 31, 1992 this prediction was deemed a failure. Then, after a 4.7 quake near the area on Oct. 19, 1992, Parkfield scientists predicted a 6.0 quake would occur in a seven-county region of Central California within 72 hours--another failed prediction. The report concludes that "It is important to educate the public that there is great value to this monitoring effort even if the prediction effort is unsuccessful." <<<========<<

>>>>>==========>> Lyme disease questions; many who test positive aren't infected. Newsday, 16 June, 1993, news section page 37. Mitchell Freedman A Boston study of 800 patients referred to a Lyme disease clinic showed that 45 percent of the patients did not have Lyme disease but tested positive suggesting that the current blood testis not useful for screening purposes. Lyme disease has been reported in every state except Alaska and Montana with 50,000 cases reported since 1982. Shelter Island in Long Island has been the place of highest exposure for Lyme disease. In about 60 percent of the cases of Lyme disease a bull's eye rash develops after a bite from an infected deer tick. Additional symptoms are heartbeat irregularities and swollen joints. Lyme disease is treatable with antibiotics once detected but can cause more serious problems in untreated. Federal officials are providing increased funds for research on Lyme disease. <<<========<<

>>>>>==========>> FDA seeks more data on side effects. Washington Post, 2 June, 1993, A3 John Schwartz A new program called MEDWatch has been introduced to make it easy for physicians to report serious side effects or defects in a broad range of medications and devices. Current clinical trials are well designed but an event that occurs only once in 5,000 patients can be missed. This program replaces an older system that was notoriously difficult to use. Even under this system such reports have played a major role in removing drugs and devices that were approved but later found in fact to have serious risks. A recent example being the restriction on silicone breast implants after 30 years on the market. The program is described in more detail in 2 June issue of the Journal of the American Medical Association. <<<========<<

>>>>>==========>> The ozone backlash. Science, 11 June 1993 Gary Taubs This article describes the current bitter debate between scientists who believe that chloroflouorocarbons (CFCs) provide a serious threat to the ozone layer and their critics who think this is nonsense. These critics argue that chlorine could not be depleting the ozone level because natural sources such as sea water, volcanoes etc. have been around forever and CFC's provide a negligible amount of chlorine compared to these. Their opponents argue that these natural sources do not reach the stratosphere. The opinions of the critics are having a significant influence through their appearance on talk shows, in their books and in the press. They claim that scientists concerned about CFCs are part of a conspiracy to gain funds and keep their own ideas dominant in public policy. This is an excellent discussion of the responsibilities of scientists and the role of the press and media when the full truth is not known and yet important political decisions must be made. <<<========<<

>>>>>==========>> Should unpublished data be included in meta-analysis? Journal of American Medical Association, 2 June, 1993 Cook and others. It is well known that studies with negative results are less apt to be submitted for publication and if they are less apt to be accepted. This suggests that meta-analyses that include only published data will be misleading. On the other hand, if un-published articles are less strong than published articles they might cause a problem if included. The authors sampled meta-analyses experts to see if they did include unpublished data. They also asked meta analysts and editors if they felt that unpublished results should be included in meta analyses and how this should effect future publication of the individual study. Of 150 meta analyses, 30.7% included published data in their primary analysis. 77.7% of the meta- analysts and methodologists sampled thought they should or probably should be included as compared to only 53.2% of the editors sampled who agreed with this. They also asked if inclusion of data in a prior overview should effect publication of the original research. 86.4 l% of the meta-analysts and methodologists felt that prior inclusion in a overview should have no effect on publication while only 53.2% of the editors agreed with this. <<<========<<

>>>>>==========>> Stalking HIV in the red light district. New Scientist, 12 June, 1993. Neil McKeganey The author describes the use of the capture-recapture technique for estimating a populations size in a study of prostitutes in Glasgow's city center. The project is part of a broader, six year study of prostitution, drug use and HIV infection. The authors describe the many practical problems and possible errors in the use of this technique in this setting. <<<========<<

>>>>>==========>> Does prediction of outcome alter patient management?. The Lancet, 12 June, 1993 Murray and others A study that shows that provision of a computer-based prediction of outcome after severe head injury resulted in a measurable appropriate change in patient management without adversely effecting the outcome. The authors state that "Our results, which indicate that clinicians respond appropriately to predictive information, should allay concern that an adverse prediction might results in unreasonable pessimism." <<<========<<

>>>>>==========>> Advances in detection create dilemma on prostate cancer. The New York Times, 17 June 1993, A1 Gina Kolata In recent years there has been a remarkable increase in the number of men diagnosed as having prostate cancer and treated for this cancer. This has resulted from the introduction of the PSA test recommended by the American Cancer Association for all men over 50. Unfortunately, the death rates from prostate cancer have remained virtually the same for decades rather than decrease as would be expected if the early diagnosis and treatment were effective. Worse yet, Medicare data shows a death rate of 1% among men from 65 to 74 and 2 percent for men over 75. In addition, 63 percent of the men who have their prostates removed are incontinent afterward and over 60 percent are impotent as result of the operation. Economists estimate that the cost of carrying out the PSA tests on all men over 50 and subsequent tests and treatment on men cost $28 billion year. We end on this gloomy note. <<<========<<

>>>>>==========>> !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! CHANCE News 2.11 (2 June to 17 June 1993) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Please send suggestions to: jlsnell@dartmouth.edu >>>==========>>|<<==========<<<