Chapter 4 Notes Epidemiology The branch of medical science that investigated the frequency and distribution of diseases in human populations Page 61 Expressing data using relative numbers such as ratios twice as frequent or percentage 100 increase can be misleading because such statements hide the underlying absolute numbers upon which the ratios or percentages were calculated Page 61 The larger the sample size the less likely it is that differences will arise solely through random fluctuations Page 62 A statistic called the p value which estimates the probability that an observed difference between two measurements would appear by chance even if there were in reality no such difference at all Page 62 In common practice a p value of 0 05 or less is required before scientists will conclude that an observed difference is statistically different that is likely to be genuine numerical difference rather than a random fluctuation Page 62 By definition a p value of 0 05 only means that the likelihood of being fooled is quite small 5 Page 62 Another problem is the possibility of bias which refers to the systematic distortion of results in a certain direction caused by the failure to account for some influencing factor Page 62 Experimenter bias Which arises when the expectations of the scientist carrying out the investigation interfere with objective evaluation of the data For example if an investigator believes that a particular chemical causes cancer this belief may affect his or her assessment of how many people exposed to the chemical have actually developed cancer Page 62 Detection Bias A phenomenon encountered when scientists fail to ensure that equivalent procedures are being used to measure cancer rates in the populations being compared Page 62 Selection bias Is a type of error that arises when people volunteer nonrandomly for research studies Page 62 If a self selected participants without cancer happen to volunteer because they have a healthier life style and are interested in dietary studies it may introduce a bias in which people without cancer appear to have a different diet than those with cancer Page 62 Publication bias Arises from a common practice of scientific journals They rarely publish studies in which investigators have failed to detect some kind of relationship Page 62 63 Confounding Can also lead to the incorrect identification of agents that supposedly cause cancer A confounding variable is a factor that 1 affects the risk of developing and 2 is linked in some way to the factor being investigated Page 63 In reality smoking is a confounding factor in many epidemiology studies because tobacco is a potent cancer causing agent and a variety of other behaviors are commonly associated with smoking Page 63 Another important confounding variable is age Cancer rates increase dramatically as people get older Page 63 Retrospective studies That asses the past exposures of people who have already developed cancer and prospective studies that follow people into the future to see who will develop cancer Page 63 When a person incorrectly concludes that two events are connected by a cause and effect relationship just because they are associated with each other this error in reasoning is referred to as the post hoc fallacy Page 64 Dose response relationship That is by showing that the risk of developing cancer increases in direct relationship to the amount of exposure to the agent being investigated A dose response relationship is considered to provide strong but not absolute evidence for a cause and effect relationship because it is unlikely that common sources of confounding or bias would generate data in which cancer rates progressively increase as exposure to a particular agent increases Page 64 To obtain definitive evidence that something would need to be performed in which an investigator randomly assigns people to receive varying doses or no dose at all of a suspected cancer causing agent and then waits to see who develops cancer This is a randomized trial This is not ethical to perform in humans Animal testing and epidemiology are complementary approaches page 65 Carcinogens Cancer causing agents By law any new chemical introduced for human consumption in the U S must first be tested in animals to check for possible carcinogenic effects Such tests are usually carried out in rats or mice whose life expectancy is short enough to permit the effects of chemical exposures to be monitored throughout their lifetime Page 65 The main advantage of animal testing is that experiments can be carried out under precisely controlled conditions in which difference groups are animals are exposed to different doses of the chemical being examined Page 65 There are numerous physiological differences between humans and laboratory animals means that there is no guarantee that a substance found to cause cancer in animals will always do so in humans Page 65 66 To overcome this problem animal studies often employ high doses of chemicals to obtain a sufficient number of cancer cases If the dose response curve is linear and if animals and humans respond the same way to the chemical in question cancer rates observed in animals exposed to these higher doses can be used to calculate the number of cancer cases expected at lower doses in humans Page 66 Cancer causing agents often act as mutagens that is they cause cancer by triggering DNA mutations changes in DNA base sequence Page 66 Ames Test Utilizes bacteria as a test organism because they can be quickly grown in enormous numbers in culture The bacteria used for the ames test are a special strain that lacks the ability to synthesize the amino acid histidine The bacteria are placed in a culture dish containing a growth medium without histidine along with the chemical being tested for mutagenic activity Normally the bacteria would not grow in the absence of histidine If the chemical being tested is mutagenic however it will trigger random mutations some of which might restore the ability to synthesize histidine page 66 The ames test also gives a significant number of false positive and false negative results so it cannot be relied on as a sole indicator of whether or not a substance causes cancer page 66 Chemical agents called promoters increase risk but do not cause mutations Page 66 67 In contrast cancer is relatively rare in children accounting for less than 1 of the total number of cancer cases diagnosed each year Still cancer is second only to
View Full Document
Unlocking...