NR 503 EPIDIMOLOGY MIDTERM EXAM

NR 503 EPIDIMOLOGY MIDTERM EXAM Student Consult Chapter 2-4 Which of the following is a condition which may occur during the incubation period? • Transmission of infection The incubation period is defined as the interval from receipt of infection to the time of onset of clinical illness. Accordingly, individuals may transmit infectious agents during the incubation period as they show no signs of disease that would enable the isolation of sick individuals by quarantine. Chicken pox is a highly communicable disease. It may be transmitted by direct contact with a person infected with the varicella-zoster virus (VZV). The typical incubation time is between 10 to 20 days. A boy started school 2 weeks after showing symptoms of chicken pox including mild fever, skin rash, and fluid-filled blisters. One month after the boy returned to school, none of his classmates had been infected by VZV. The main reason was: • Contact was after infectious period • Subclinical infections were not yet detected • Disease was endemic in the class The disease is spread by contact with an infected individual who can transmit the agent (VZV) to immunologically naive persons during the incubation period and for several days after onset of clinical illness. Since the boy started school 14 days after showing signs consistent with chicken pox, it is most likely that he was no longer infectious. The ability of a single person to remain free of clinical illness following exposure to an infectious agent is known as: • Hygiene • Vaccination • Herd immunity • Immunity • Latency Immunity is the capacity of a single individual to avoid disease susceptibility when exposed to an infectious agent. Herd immunity is a population characteristic. For certain diseases, individual immunity can be acquired by vaccination, but this is not true for all infectious diseases. Which of the following is characteristic of a single-exposure, common-vehicle outbreak? • Long latency period before many illnesses develop • There is an exponential increase in secondary cases following initial exposures • Cases include only those who have been exposed to sick persons • The epidemic curve has a normal distribution when plotted against the logarithm of time • Wide range in incubation times for sick individuals Single-exposure, common-vehicle outbreaks involve a sudden, rapid increase in cases of disease that are limited to persons who share a common exposure. Additionally, few secondary cases develop among persons exposed to primary cases. A histogram of the outbreak can plot the number of cases by time of disease onset. In single-exposure, common-vehicle outbreaks, a log transformation of the time of disease onset will often take on the characteristic shape of a normal distribution (i.e., a bell curve) with the median incubation time found at the peak of the curve. What is the diarrhea attack rate in persons who ate both ice cream and pizza? • 39/52 • 21/70 • 39/67 • 51/67 • none of the above The attack rate in this example is defined as the number of persons who develop diarrhea divided by the total number of people at risk. In this example, the at-risk group is those who have eaten both ice cream and pizza. Of these 52 persons, 39 developed diarrhea. What is the overall attack rate in persons who did not eat ice cream? • 30% • 33% • 35% • 44% • 58% (14+9)/(40+30)= 33% The attack rate is the number of persons with diarrhea (14 + 9) divided by the total number of persons who did not eat ice cream (40 + 30). Which of the food items (or combination of items) is most likely to be the infective item(s)? • Pizza only • Ice cream only • Neither pizza nor ice cream • Both pizza and ice cream • Cannot be assumed from the data show Among persons eating ice cream, over 70% developed diarrhea regardless of their pizza consumption (39/52 and 11/15). Among both groups of persons who did not eat ice cream, each attack rate was equal to or less than 35% (14/40 and 9/30). Which of the following reasons can explain why a person who did not consume the infective food item got sick? • They were directly exposed to persons who did eat the infective food item • Diarrhea is a general symptom consistent with a number of illnesses • There may have been an inaccurate recall of which foods were eaten • All of the above • None of the above Without knowledge as to the specific agent in this instance, it is also likely that it can be spread by direct contact with infected persons. Since diarrhea is a general disease symptom, it is possible that several infectious agents may be present at this meal or others eaten during the same time period. Further, information regarding food consumption may have been collected long after the disease episode. This may have led persons to incorrectly remember the foods that they consumed. An outbreak of gastroenteritis occurred at a boarding school with a student enrollment of 846. Fifty-seven students reported symptoms including vomiting, diarrhea, nausea, and low-grade fever between 10 p.m. on September 24 and 8 p.m. on September 25. The ill students lived in dormitories that housed 723 of the students. The table below provides information on the number of students per type of residence and the number reporting illnesses consistent with the described symptoms and onset time. Calculate the attack rate among all students at the boarding school. 57 total cases/ 846 total students=6.7% The answer is found by dividing the total number of cases (57) by the total number of students (846). This equals 6.7%. Calculate the attack rates for boys and girls separately. (40+3)/(380+46)= 10.1% boys (12+2)/(343+77)= 3.3% girls For boys, the attack rate includes all cases (40 + 3) divided by the total number of students who are boys (380 + 46). The attack rate is 10.1%. For girls, the attack rate includes all cases (12 + 2) divided by the total number of students who are girls (343 + 77). The attack rate is 3.3%. What is the proportion of total cases occurring in boys? (40+3)/57=75.4% The proportion of cases occurring in boys is equal to the number of cases in boys divided by the total number of cases (43/57). This equals 75.4%. What is the proportion of total cases occurring in students who live in dormitories? (40+12)/57= 91.2% The proportion of cases occurring in dormitory residents is equal to the number of cases in residents divided by the total number of cases (52/57). This equals 91.2% Which proportion is more informative for the purpose of the outbreak investigation? Both proportions are useful. Dormitory residents account for over 90% of the cases indicating an outbreak of an infectious agent that was transmitted at the school. Furthermore, over 75% of the cases were boys indicating that the responsible agent was more likely to have been transmitted in the boys’ dormitory. A group of researchers are interested in conducting a clinical trial to determine whether a new cholesterol-lowering agent was useful in preventing coronary heart disease (CHD). They identified 12,327 potential participants for the trial. At the initial clinical exam, 309 were discovered to have CHD. The remaining subjects entered the trial and were divided equally into the treatment and placebo groups. Of those in the treatment group, 505 developed CHD after 5 years of follow-up while 477 developed CHD during the same period in the placebo group. What was the prevalence of CHD at the initial exam? 12327 potential participants-309 confirmed CHD=12018 remaining subjects 505 developed CHD on 5 year f/u; 477 CHD during placebo group. 309/12327= 0.x1000=25.1% The prevalence of CHD at the initial exam was 309 cases of CHD divided by 12,327 participants. This equals a prevalence of 25.1 cases of CHD per 1,000 persons. What was the incidence of CHD during the 5-year study? 505+477=982 982/12018=0.0817 0.0817x1000=81.71 per 1000 The incidence rate reflects the number of new cases developing in the population at risk. Since prevalent CHD cases were excluded from the study, the population at risk was 12,018 (12,327 persons less 309 cases of CHD). During the 5-year study period, 982 incident cases of CHD developed. This equals an incidence rate of 81.7 cases of CHD per 1,000 persons. Which of the following are examples of a population prevalence rate? • The number of ear infections suffered by 3-year-old children in March, 2006 • The number of persons with hypertension per 100,000 population • The number of cases of skin cancer diagnosed in a dermatology clinic • b and c • All of the above Prevalence is the number of affected persons in a specified population size at a given time. Only answer (b) fits this definition. Example (a) is more consistent with an incident rate while answer (c) is a selected group of persons who may not be representative of a general population. What would be the effect on age-specific incidence rates of uterine cancer if women with hysterectomies were excluded from the denominator of incidence calculations assuming that most women who have had hysterectomies are older than 50 years of age. • The rates in all age groups would remain the same. • Only rates in women older than 50 years of age would tend to decrease. • Rates in women younger than 50 years would increase compared to women older than 50 years of age. • Rates would increase in women older than 50 years of age but may decrease in younger women as they get older. • It cannot be determined whether the rates would increase or decrease. Women who have had hysterectomies (i.e., removal of the uterus) are no longer at risk for uterine cancer. For women older than 50 years of age, this would increase the age-specific incidence rate as there would be the same number of uterine cancers occurring among fewer women at risk. Further, rates may decrease among younger women who have had hysterectomies as they are no longer at risk for uterine cancer and thus may decrease the number of potential cases occurring in their age group over time. A survey was conducted among 1,000 randomly sampled adult males in the United States in 2005. The results from this survey are shown below. The researchers stated that there was a doubling of risk of hypertension in each age group younger than 60 years of age. You conclude that the researchers’ interpretation: • Is correct • Is incorrect because prevalence rates are estimated • Is incorrect because it was based on proportions of the population sample • Is incorrect because incidence rates do not describe risk • Is incorrect because the calculations do not include adult females The survey reports the disease status of a population at a specific point in time. In this case, a random sample of adult males in 2005 provides a reliable estimate of the prevalence of hypertension. Since there is no information on duration of hypertension in these men, incidence cannot be calculated. Therefore, the researchers are not able to make a statement concerning risk of hypertension in the population. The incidence and prevalence rates of a chronic childhood illness for a specific community are given below. Based on the data, which of the following interpretations best describes disease X? • The duration of disease is becoming shorter. • The duration of disease is becoming longer. • The case-fatality rate of this disease is decreasing. • Efforts to prevent new cases of this disease are becoming more successful. • The risk of the disease has decreased over the past 20 years. Prevalence and incidence are related by the duration of disease. If incidence is increasing over time, then duration of illness has to decrease in order to keep the prevalence rate constant. This may occur through better treatments to cure disease or through higher case-fatality rates as a disease becomes more lethal. Since incidence is increasing over time, it is evident that risk is also increasing and that prevention efforts are not successful. A prevalence survey conducted from January 1 through December 31, 2003 identified 580 new cases of tuberculosis in a city of 2 million persons. The incidence rate of tuberculosis in this population has historically been 1 per 4,000 persons each year. What is the incident rate of tuberculosis per 100,000 persons in 2003? (580/) x = 29 The answer is 29 new cases of tuberculosis per 100,000 persons. This is found by dividing the new cases of tuberculosis by the total population at risk (580/2,000,000) and multiplying this rate by 100,000 to standardize the rate. Has the risk of tuberculosis increased or decreased during 2003? /4000= 25x1= 25 per The risk of tuberculosis has increased over the historic incident rate. This comparison can be made by standardizing the historic rate to a rate per 100,000 persons. To do this, multiply the numerator and denominator by 25. Which of the following is an advantage of active surveillance? • Requires less project staff • Is relatively inexpensive to employ • More accurate due to reduced reporting burden for health care providers • Relies on different disease definitions to account for all cases • Reporting systems can be developed quickly Active surveillance entails a concerted effort to collect information about disease occurrence. It typically involves dedicated staff members who have been specifically directed to contact physicians and hospitals in order to collect reports of disease cases in a specified population. This activity requires a large amount of staff and resources in order to accomplish its goals. The population of a city on February 15, 2005, was 36,600. The city has a passive surveillance system that collects hospital and private physician reports of influenza cases every month. During the period between January 1 and April 1, 2005, 2,200 new cases of influenza occurred in the city. Of these cases, 775 persons were ill with influenza according to surveillance reports on April 1, 2005. The monthly incidence rate of active cases of influenza for the 3-month period was: • 4 per 1,000 population • 17 per 1,000 population • 20 per 1,000 population • 39 per 1,000 population • 130 per 1,000 population 2200 new cases of flu/36600 population=0.0601 0.0601/3=0.002 0.002x1000=20 per 1000 The monthly incidence rate is calculated based on the number of new cases of a disease developing during the 3-month time period. In this example, 2,200 cases of influenza developed among an average population of 36,600 persons at risk during the surveillance period. The incidence rate equals 2,200 divided by 36,600. In order to calculate the average monthly rate, the rate should then be divided by 3. Finally, the monthly rate can be multiplied by 1,000 in order to express it per the responses listed. The prevalence rate of active influenza as of April 1, 2005, was: 705/()x1000= 20 per 1000 The prevalence rate as of April 1, 2005, is equal to the number of active influenza cases reported divided by the number of persons at risk in the population at that time. The best estimate of the population size is that from the February 15 count, less who are no longer at risk as they have already recovered from influenza and have developed immunity. Therefore, prevalence equals 705 cases divided by 36,600 less 1,495 recovered cases. This number can be multiplied by 1,000 in order to estimate a prevalence rate of 20 cases per 1,000 persons. What can be inferred about influenza cases occurring in the city? • Active surveillance would enable better prevention of influenza • The incidence rate would decrease if active surveillance were employed • The average duration of influenza is approximately 1 month • The actual number of influenza cases occurring in the population is less since hospitals and private physicians may be reporting the same patients. • The prevalence rate should be higher since it should be calculated based on all cases of influenza occurring from January 1 through March 30, 2005. Since the average monthly incidence rate is 20 per 1,000 and the prevalence rate is also 20 per 1,000, then the duration of disease must equal 1 month. A study found that adults older than age 50 had a higher prevalence of pneumonia than those who were younger than age 50. Which of the following is consistent with this finding? • Younger adults have a higher incidence of pneumonia • Older adults have a higher case-fatality rate from pneumonia • Younger adults with pneumonia are more likely to report being ill than older persons • Incidence rates do not vary by age, but older adults have pneumonia for a longer duration compared to younger adults • None of the above For prevalence to be higher among older adults, either incidence or duration of pneumonia must be increased in this age group. Which of the following statements are true? More than one answer may be correct. • Prevalence rates are always larger than incidence rates • In a steady state, the prevalence of disease is equal to the attack rate • Diagnostic criteria rarely impact estimates of disease prevalence and incidence • Prevalence rates are useful for public health planning • Incidence rates can be used to estimate prevalence when the mean duration of the disease is known A disease has an incidence of 10 per 1,000 persons per year, and 80% of those affected will die within 1 year. Prior to the year 2000, only 50% of cases of the disease were detected by physician diagnosis prior to death. In the year 2000, a lab test was developed that identified 90% of cases an average of 6 months prior to symptom onset; however, the prognosis did not improve after diagnosis. Comparing the epidemiology of the disease prior to 2000 with the epidemiology of the disease after the development of the lab test, which statement is true concerning the disease in 2000? • Incidence is higher and prevalence is higher than in 1999 • Incidence is higher in 2000 but prevalence remains the same • Incidence is the same in 2000 but prevalence is higher than in 1999 • Both incidence and prevalence remain the same as in 1999 • Incidence is the same in 2000 but prevalence is lower than in 1999 With increased ability to detect cases of the disease at earlier times, both the number of incidence and prevalent cases will increase through better detection. Which statement is true concerning the duration of the disease after the development of the lab test? • Mean duration of a case of the disease is shorter in 2000 • Mean duration of a case of the disease is the same in 2000 • Mean duration of a case of the disease is longer in 2000 • No inference about mean duration can be made since the lab test has only been available for 1 year Though the prognosis is similar after the development of the lab test, the duration of new cases identified by the test can be increased by up to 6 months due to earlier detection. Which statement is true concerning the disease-specific mortality rate after the development of the lab test? • The mortality rate for the disease is decreased in 2000 • The mortality rate for the disease is the same in 2000 • The mortality rate for the disease is increased in 2000 • No inference about the mortality rate can be made since the lab test has only been available for 1 year With the implementation of the lab test, the increase in early detection of cases will increase incidence, duration, and prevalence; however, since the prognosis is still the same, at least 80% of patients will die during the year 2000. This should result in a similar mortality rate as the previous year given no change in transmission, prevention, or medical care of the disease. In a coastal area of a country in which a tsunami struck, there were 100,000 deaths in a population of 2.4 million for the year ending December 31, 2005. What was the all-cause crude mortality rate per 1,000 persons during 2005? (100,000 deaths/2,400,000 population)x1000=41.7 per 1000 The answer is 41.7 per 1,000 persons. The rate is calculated by dividing 100,000 deaths by the population of 2,400,000 persons. To express as a rate per 1,000 persons, the rate is multiplied by 1,000. In an industrialized nation, there were 192 deaths due to lung diseases in miners ages 20 to 64 years. The expected number of deaths in this occupational group, based on age-specific death rates for lung diseases in all males ages 20 to 64 years, was 238 during 1990. What was the standardized mortality ratio (SMR) for lung diseases in miners? (192deaths/238expected number of death) x 100= 81% The answer is 81. The ratio is calculated by dividing 192 observed deaths by the 238 expected deaths for this age group. To express it as an SMR, the ratio is often multiplied by 100. In 2001, a state enacted a law that required the use of safety seats for all children under 7 years of age and mandatory seatbelt use for all persons. The table below lists the number of deaths due to motor vehicle accidents (MVAs) and the total population by age in 2000 (before the law) and in 2005 (4 years after the law was enacted). What is the age-specific mortality rate due to MVAs for children ages 0 to 18 years in 2000? • 1.8 per 1,000 • 2.9 per 1,000 • 4.0 per 1,000 • 6.1 per 1,000 • Cannot be calculated from information given (44+105 deaths)/(3500+21000population)x1000=6.1 per 1000 The rate is found by combining the MVA deaths and total population size for the two age groups under 7 years and 7 to 18 years during the year 2000. This equals (44 + 105) divided by (3,500 + 21,000). Multiplying this rate by 1,000 persons gives the answer indicated. Using the pooled total of the 2000 and 2005 populations as the standard rate, calculate the age-adjusted mortality rate due to MVAs in 2005. 350 (total deaths 2000) +640 (total deaths 2005) =990 (population 2000)+ (population 2005)=418,500 990 (total deaths)/ (total population=0.0023 0.0023x1000=2.3 per 1000 The correct answer is 2.3 MVA deaths per 1,000 persons. The key to calculating the age-adjusted rate is to pool the observed numbers for both time periods and to calculate the expected numbers of deaths in the 2005 population assuming that a common rate applied to the population. For example, for those under 7 years, the pooled rate equals (44 + 20) divided by (3,500 + 4,000). The pooled rate for this group is 8.5 per 1,000 persons. When this rate is multiplied by the 4,000 children under 7 years of age in 2005, the expected number of deaths is 34.13. Performing the same calculation for each age group results in 111.7 deaths in those 7 to 18 years of age, 175.8 deaths in those 19 to 49 years, and 237.35 deaths for those 50 years or more. The total number of deaths expected in 2005 based on this pooled rate is 558.98. Therefore, the age- adjusted overall rate for 2005 is 558.98 deaths divided by 240,000 persons. Based on the information in the table, it was reported that there was an increased risk of death due to MVAs in the state after the law was passed. These conclusions are: • Correct, because there were 1.8 times as many MVA deaths in 2005 as in 2000 • Correct, because for each age group, the mortality rates were higher in 2005 than they were in 2000 • Correct, because both the total and the age-adjusted mortality rates are higher in 2005 than in 2000 • Incorrect, because the age-adjusted mortality rate due to MVA is actually lower in 2005 than in 2000 • Incorrect, because the overall mortality rate is the same in both years The overall crude (unadjusted) mortality rate is 2.6 per 1,000 persons in 2005. This is found by dividing 640 deaths by a population of 240,000 persons. This rate is then multiplied by 1,000. The overall adjusted mortality rate is 2.3 per 1,000 persons as calculated in question 34. Both of these rates are higher than the overall crude mortality rate of 2.0 per 1,000 persons for the year 2000. For colorectal cancer diagnosed at an early stage, the disease can have 5-year survival rates of greater than 80%. Which answer best describes early stage colorectal cancer? • Incidence rates and mortality rates will be similar • Mortality rates will be much higher than incidence rates • Incidence rates will be much higher than mortality rates • Incidence rates will be unrelated to mortality rates • None of the above For diseases with a long duration as indicated by high 5-year survival rates for early stage colorectal cancer, the incidence will be much higher than the mortality rate since more persons are being diagnosed with the disease than are dying of it. The following table gives the mean annual age-specific mortality rates from measles during the first 25 years of life in successive 5-year periods. You may assume that the population is in a steady state (i.e., migrations out are equal to migrations in). The age-specific mortality rates for the cohort born in are: • 2.4 2.8 1.7 1.5 0.4 • 2.9 3.7 2.8 2.0 0.6 • 2.9 2.4 1.7 1.3 0.8 • 2.4 3.3 2.0 0.6 0.1 • 1.7 2.8 2.2 1.1 0.2 This is found by tracking the cohort of children born between 1915 and 1919 by each 5-year age group. For example, this group would be 0 to 4 years of age in 1915 to 1919 with a rate of measles mortality of 2.4. In 1920 to 1924, this group of children would be 5 to 9 years of age and have a rate of measles mortality of 3.3. Continuing in a diagonal manner, the remaining three rates can be found in the table. Based on the information above, one may conclude: • Age-specific mortality rates for measles decreased for the period 1910–1914 to 1925– 1929 • Age-specific mortality rates for measles increased for the period 1910–1914 to 1925– 1929 • The case-fatality rate decreased for the period 1910–1914 to 1935–1939 • Children born in 1910–1914 had the highest rate of death in all periods • Children ages 5 to 9 had the highest rate of death in all periods For each 5-year period, the highest mortality rate is reported among those 5 to 9 years of age. This is seen by comparing the rate for this age group to all other age groups in a row. Which of the following characteristics indicate that mortality rates provide a reliable estimate of disease incidence? More than one answer may be correct. • Case-fatality rate is low • The case-fatality rate is high • The duration of disease is short • The prevalence of disease is greater than 5% • The proportionate mortality is high Which of the following statements are true? More than one answer may be correct. • A mortality rate is an example of an incidence rate • Death certificate data are generally valid regardless of the cause of death • Type of disease is the most important predictor of mortality • Changing diagnostic criteria does not affect estimates of prevalence and incidence • The case-fatality rate is calculated based on the entire population at risk A mortality rate can approximate an incidence rate under conditions of a high case-fatality rate and a short duration of disease. Among those who are 25 years of age, those who have been driving less than 5 years had 13,700 motor vehicle accidents in 1 year, while those who had been driving for more than 5 years had 21,680 motor vehicle accidents during the same time period. It was concluded from these data that 25-year-olds with more driving experience have increased accidents compared to those who started driving later. This conclusion is: • Correct based on the data • Incorrect because rates are not reported • Incorrect because prevalence estimates are given when incidence rates should be reported • Incorrect because there are no comparison groups identified • Both b and d are correct The information provided only enumerates motor vehicle accidents in two groups. In order to fully compare these counts, information is needed on the denominator, i.e., the number of persons driving in each group, so that rates can be calculated. For a disease such as liver cancer, which is highly fatal and of short duration, which of the following statements is true? Choose the best answer. • Mortality rates will be much higher than incidence rates • Mortality rates will be much higher than prevalence rates • Incidence rates will be much higher than mortality rates • Case-fatality rates will be equal to mortality rates • Incidence rates will be equal to mortality rates Since the 5-year survival rate for liver cancer is 4%, most incident cases of liver cancer will result in a premature mortality. In this case, the mortality and incidence rates will be approximately equal. The prevalence rate of a disease is two times greater in women than in men, but the incidence rates are the same in men and women. Which of the following statements may explain this situation? • The duration of disease is shorter in women • Men are at greater risk for developing the disease • The case-fatality rate is lower for women • The age-adjusted mortality rate will be higher for women • The proportionate mortality rate for the disease is higher for men Since men and women develop the disease at the same rate, the survival rate in women must be increased in order to increase duration and prevalence. A low case-fatality rate would contribute to an increased duration of the disease. The table below describes the number of illnesses and deaths caused by plague in four communities. The case-fatality rate associated with plague is lowest in which community? • Community A • Community B • Community C • Community D The case-fatality rate equals the number of deaths occurring from plague divided by all persons with the plague. In Community C, the CFR is 300 divided by 400, or 60%. This is lower than A (67%), B (75%), and D (77%). The proportionate mortality ratio associated with plague is lowest in which community? • Community A • Community B • Community C • Community D The proportionate mortality rate equals the number of deaths occurring from plague divided by all persons with the plague. In Community D, the PMR is 500 divided by 5000, or 10%. This is lower than A (50%), B (75%), and D (38%). Student Consult Chapter 5 and 6 In a community-based hypertension testing program called HT-Aware, the detection level for high blood pressure is set at 140 mmHg for systolic blood pressure. A separate testing program called HT-Warning in the same community sets the level at 130 mmHg for high systolic blood pressure. Which statements are likely to be true? • The sensitivity of HT-Warning is greater than that of HT-Aware • The specificity of HT-Warning is greater than that of HT-Aware • The number of false positives is greater with HT-Warning than with HT-Aware • The number of false negatives is greater with HT-Warning than with HT-Aware • The sensitivity and specificity are the same for both tests A school nurse examined a population of 1,000 children in an attempt to detect nearsightedness. The prevalence of myopia in this population is known to be 15%. The sensitivity of the examination is 60% and its specificity is 80%. All children labeled as “positive” (i.e., suspected of having myopia) by the school nurse are sent for examination by an optometrist. The sensitivity of the optometrist’s examination is 98% and its specificity is 90%. How many children are labeled “positive” by the school nurse? School Nurse: 1000 children x 15% prevalence of myopia=150 cases 1000 population – 150 cases (prevalence)= 850 cases of negative for myopia 150 cases x 60% sensitivity= 90 cases true positives (TP) 850 cases of negatives – 80% specificity = 170 false positive (FP) 90TP+170FP=260 cases will be referred to optometrist 90 TP/ (90TP+170 FP) x 100= 34.6% PPV Optometrist: 90 TP from school screening X 98% sensitivity of optometrist: 88 cases TP 170 FP from school screening - 90% specificity of optometrist: 17 cases FP 88TP+17FP= labeled as myopic 88TP/(88TP+17FP)= 84% PPV There are 150 children with myopia in the school population (15% prevalence among 1,000 children). The school nurse will identify 60% of those who truly have the condition, or 90 cases (60% sensitivity multiplied by 150 myopic children). Further, the school nurse will incorrectly identify 170 false positive cases of myopia among those who do not have the condition (80% specificity multiplied by 850 non-myopic children). The sum of the cases labeled as positive by the school nurse equals 260 children (90 true myopic children plus 170 false positive children). What is the positive predictive value (PPV) of the school nurse’s exam? 90 cases of TP/ 260 children labelled as myopic (90 TP+170 FP)= 0.346x 100= 34.6% The PPV of the school nurse’s exam is equal to the number of true positive cases divided by the total number of those that the school nurse labels as positive. In this exam, the PPV is 34.6% (90 true myopic children divided by 260 children labeled as myopic by the school nurse). How many children will be labeled myopic following the optometrist’s exam? 90 TP - 98% sensitivity optometrist= 88 cases + 170 FP cases - 90% specificity optometrist=17 cases =105 Since the optometrist will only test children who have been labeled as myopic by the school nurse, the testing group for this sequential exam is 260 children. The optometrist labels 105 children as myopic. Among the 90 myopic children correctly referred by the school nurse, the optometrist identifies 88 of them as myopic (98% sensitivity multiplied by 90 true cases of myopia). Further, the optometrist will incorrectly identify 17 false positive cases among the 170 children referred by the school nurse who do not have myopia. The sum of the cases labeled as positive by the optometrist equals 105 children (89 true cases plus 17 false positive cases). What is the positive predictive value (PPV) of the optometrist’s exam? 88 true positive/ (88 TP + 17 FP)= 84.8% What is the negative predictive value (NPV) of the optometrist’s exam? 260 total cases referred by school nurse-105 confirmed positive by optometrist=155 true negatives 90 TP School nurse case – 88 TP optometrist case=2 false positives identified 155 true negatives – 2 false positives= 153 NPV= 153/155= 0.987 x 100= 98.7% The NPV of the optometrist’s exam is 98.7%. The NPV equals the number of true negative cases divided by all negative cases indicated by the exam. In this instance, the optometrist correctly identifies 153 children as negative for myopia; however, there are 2 false negative cases following the optometrist’s exam (90 true cases referred by the school nurse less the 88 cases detected by the optometrist). The NPV equals 153 divided by 155, or 98.7%. What is the overall sensitivity of the sequential examinations? 88 confirmed positive cases / 150 prevalence case= 58.7% The overall sensitivity of the sequential exams is 58.7%; 88 true positive cases of myopia are found following the optometrist’s exam among the 150 prevalent cases in the school population. What is the overall specificity of the sequential examinations? 1000 children screened –15% prevalence rate= 150 1000-150= 850 negatives 680 TN from school nurse + 153 TN from optometrist = 833 TN cases 833/850 = 98% specificity The overall specificity of the sequential exams is 98%; 833 children will be correctly labeled as negative for myopia among the 850 true negative cases. This is found by summing the number of true negatives after each exam (680 following that of the school nurse plus 153 following the optometrist) and dividing by the true negative children in the population. This equals 833 divided by 850, or 98%. What would be the positive predictive value (PPV) of the exam for myopia if the optometrist tested all 1,000 children? 150 prevalence 150 x 98% sensitivity= 147 TP 850 TN cases x 90% specificity=85 FP cases PPV: TP/(TP+FP) PPV: 147 TP/(147TP+85FP)= 63.4% The PPV of the optometrist’s exam would be equal to the number of true positive cases divided by all children labeled positive by the optometrist. Applying the sensitivity and specificity of the optometrist’s exam to the 1,000 children would indicate that 147 true positive cases are labeled positive by the optometrist. Additionally, the optometrist would find 85 false positive cases (850 true negative cases multiplied by 90% specificity). The PPV would be 63.4% (147 true positive cases divided by 232 total positives indicated by the optometrist). Which of the following improves the reliability of diabetes screening tests? • Having the same lab analyze all samples • Taking more than one sample for each subject and averaging the results • Insuring that the instrument is standardized before each sample is analyzed • a and c only • All of the above Reliability is improved by consistency of analyses, especially when multiple samples are taken for a subject and the analytic instrument is routinely standardized. A prostate specific antigen (PSA) test is a quick screening test for prostate cancer. A researcher wants to evaluate it using two groups. Group A consists of 1,500 men who had biopsy-proven adenocarcinoma of the prostate while group B consists of 3,000 age- and race-matched men all of whom showed no cancer at biopsy. The results of the PSA screening test in each group is shown in the table. What is the sensitivity of the PSA screening test in the combined groups? 1155 TP /1500 TP= 0.77 x 100=77% The sensitivity equals the number of true positives detected among all true positives. Since a biopsy is the gold standard test for prostate cancer, all 1,500 men in group A are positive for prostate cancer. The PSA test indicated that 1,155 of these men had prostate cancer, a sensitivity of 77%. What is the specificity of the screening test in the combined groups? 3000-240= 2760 true negatives 2760/3000=0.92 x100= 92% The specificity equals the number of true negatives detected among all true negatives. Among the 3,000 men who did not have prostate cancer, the test correctly identified 2,760 men as negative for prostate cancer (3,000 minus 240 false positives). This gives a sensitivity of 92% What is the positive predictive value (PPV) of the screening test in the combined groups? TP/TP+FP 1155/(1155+240)=0.827x100= 83% The PPV is 83%. This value is found by dividing 1,155 true positives by the total number of all positives indicated by the PSA test (1,155 plus 24). The PSA screening test is used in the same way in two equal-sized populations of men living in different areas of the United States, but the proportion of false positives among those who have a positive PSA test in the first population is lower than that among those who have a positive PSA test in the second population. What is the likely explanation for this finding? • It is impossible to determine what caused the difference • The prevalence of disease is higher in the first population • The specificity of the test is lower in the first population • The specificity of the test is higher in the first population • The prevalence of the disease is lower in the first population We can assume that the specificity of the test will be similar in each population. Therefore the proportion of false positives found among the true negatives should be the same in each population. However, the proportion of false positives among all positives on the PSA screening test will be influenced by the number of true positives detected by the test. Since the sensitivity of the test will also be the same, we can assume that more true positives exist in the population of men with a lower proportion of false positive tests due to an increase in the PPV. Test A has a sensitivity of 95% and a specificity of 90%. Test B has a sensitivity of 80% and a specificity of 98%. In a community of 10,000 people with 5% prevalence of the disease, Test A has always been given before Test B. What is the best reason for changing the order of the tests? • The net sensitivity will be increased if Test B is given first • The total number of false positives found by both tests is decreased if Test B is given first • The net specificity will be decreased if Test B is given first • The total number of false negatives found by both tests is decreased if Test B is given first • There is no good reason to change the order of the tests A sequential testing process would only refer those with positive results to the second test. Since Test B has a higher specificity, then fewer false positives will be referred for Test A, thereby decreasing the number of false positives found. This can be shown by calculation if we assume that 500 persons have the disease among the 10,000 in the population. Test B will find only 190 false positives for referral (9,500 true negatives less the number of true negatives multiplied by 98% specificity). Performing Test A first results in 950 false positives referred for the second test (9,500 true negative less the number of true negatives multiplied by 90% specificity). Two neurologists, Drs. J and K, independently examined 70 magnetic resonance images (MRIs) for evidence of brain tumors. As shown in the table below, the neurologists read each MRI as either “positive” or “negative” for brain tumors. Based on the above information, the overall percent agreement between the two doctors including all observations is: • 37.1% • 62.9% • 65.0% • 68.4% • 84.6% 26 positives + 18 negatives= 44 agreed on readings/ 70 MRI=0.628x100= 62.9% The two doctors agree on 44 of the 70 MRI readings. This includes the 26 that they both labeled as positive for brain tumors and the 18 that they both agreed were negative for brain tumors. What is the estimate of kappa for the reliability of the two doctors’ test results? • 10.1% • 24.9% • 50.6% • 57.4% • 68.4% The estimate of kappa expresses the observed agreement of two testers in excess of chance alone. It is found by applying the expected agreement rates for both testers. In this case, Dr. K labeled 38 of the 70 MRIs as positive (54.3% of all MRIs) and 32 as negative (45.7% of all slides). Dr. J labeled 57.1% of the MRIs as positive (40 of 70) and 42.9% as negative. We would expect that if Dr. K had the same rate of positive and negative findings as Dr. J then they would agree by chance on 21.7 of the 38 positive MRIs that were found (38 multiplied by 0.571). Further, they would agree by chance on 13.7 of the 32 negative MRIs that were found (32 multiplied by 0.429). Therefore, we would expect the two doctors to agree by chance on 50.6% of the MRIs (21.7 positive agreements plus 13.7 negative agreements equals 35.4, then divide this by the total of 70 to get an expected overall agreement of 50.6%). Now, kappa can be calculated as the observed agreement less expected divided by 100% less the expected agreement— in this instance, 62.9% minus 50.6% divided by 100% less 50.6%. 12.3% divided by 49.4% results in a kappa of 24.9%. This table represents the results of coronary magnetic resonance (CMR) angiography compared to x-ray angiography (the gold standard in diagnosis of coronary artery disease) in a high-risk population of patients scheduled to undergo x-ray angiography for suspected coronary artery disease. In the general population, the prevalence of coronary artery disease is apporximately 6%. Assuming that this sample of patients is representative of the general population, the sensitivity of the CMR test in the general population would be approximately: • Less than 75% • Between 75% and 85% • Between 85% and 90% • Between 90% and 95% • Greater than 95% 56/ (56+4)=0.9333x 100= 93% If we assume that the prevalence of disease is similar, then we can accept that 60 persons with a positive x-ray will be true cases of coronary artery disease. In this instance, the CMR test positively identifies 56 of the 60 true cases, a sensitivity of 93.3% After reviewing the results of the test comparison, an epidemiologist decides that the specificity of the test is too low. Using the same CMR images, he raises the cutoff value for a positive test to increase the specificity. What is the likely effect on the sensitivity? • Sensitivity will increase • Sensitivity will decrease • There will be no effect because the two characteristics are unrelated • The effect cannot be predicted as it will depend on the prevalence rate • Sensitivity will be higher if the positive predictive value is increased The increase in the cutoff value for a positive test will reduce the sensitivity of the test even though the specificity is increased. This will result in the misidentification of true positive cases as false negatives if their CMR values are below the cutoff value suggested by the epidemiologist. In comparing the mammography readings of two technicians who evaluated the same set of 600 mammograms for presence of breast cancer from a generally representative sample of women from the population, • Agreement regarding negative or normal mammograms is likely to be low • The kappa statistic measures agreement due to chance only • Overall percent agreement calculated for both readers may conceal significant disagreements regarding positive tests • A kappa of 0.9 would be unsatisfactory • A kappa of 0.6 represents poor agreement Since the sample is from the general population, it is likely that very few will have prevalent breast cancer indicating that many readings will be regarded as normal, or negative for the disease. Since a large proportion of the readings will be negative, it is likely that the two technicians will have a high value for overall percent agreement though they may differ significantly in their readings for the few women who are labeled positive for breast cancer. In a country with a population of 16 million people, 175,000 deaths occurred during the year ending December 31, 2005. These included 45,000 deaths from tuberculosis (TB) in 135,000 persons who were sick with TB. Assume that the population remained constant throughout the year. What was the annual mortality rate for the country during 2005? 175,000 total deaths/16,000,000 total population=0.x100,000=1094 per 100,000 mortality rate The annual mortality rate equals the number of deaths divided by the total population. In this example, 175,000 deaths occurred among 16 million persons. Dividing these numbers and multiplying by 100,000 gives a rate of 1,094 deaths per 100,000 persons, approximately 1% of the population. What was the case-fatality rate (CFR) from TB during 2005? 45,000 deaths from TB/135,000 sick from TB=0.3333 x 100= 33% The CFR is the number of cause-specific deaths divided by all cases of the specific disease. In this example, 45,000 TB deaths occurred in 135,000 persons with TB. This equals a CFR of 33%. What is the proportionate mortality ratio (PMR) for TB during 2005? 45,000 deaths from TB/175,000 total deaths =0.26x100=26% The PMR is the number of deaths due to a specific cause divided by all deaths. In this example, the PMR equals 45,000 TB deaths divided by 175,000 deaths, or approximately 26%. Which of the following statements is true? • The case-fatality rate provides a reasonable estimate of incidence • The prevalence of TB for 2005 is equal to the denominator of the case-fatality rate • The duration of TB is brief • All of the above • None of the above Since the duration of TB can be longer than 1 year, neither disease incidence nor prevalence can be validly estimated by mortality indicators. Which of the following statements pertains to relative survival? • Refers to survival of first-degree relatives • Is equal to the case-fatality rate • Is generally closer to observed survival rates in younger age groups • Is generally closer to observed survival rates in older age groups • Provides an estimate of proportionate mortality Relative survival is close to observed survival rate when there are few competing causes of death. This occurs primarily in younger age groups who are less likely to experience mortality events compared to older age groups. What was the probability of surviving the second year given survival to the end of the first year? 950 alive beginning of the year- 30 death at year end= 920 survivors 920 survivors/ 950 alive beginning of the year= 0.968x 100= 97% The probability of surviving the second year given survival to the end of the year indicates that we are concerned with the survival proportion of those alive at the end of year 1. In this example, we have 950 persons alive at the beginning of year 2 (thus, the end of year 1). Of this group, 30 die by the end of the second year. This gives a survival rate of 920 divided by 950, or 97%. What was the cumulative probability of surviving after only 2 years of follow-up? 920 survivors end of year 2/1000 initial number of live people= 0.92x100= 92% The cumulative survival is the total number of those surviving by the end of the second year divided by all persons who were alive at the beginning of follow-up. In this example, there were 920 survivors among the 1,000 persons who were alive at the beginning of observation. This equals a cumulative survival of 92%. Alternatively, this cumulative survival can be calculated by multiplying the survival rates for each period of interest. In this example 95% survival for year 1 multiplied by 97% survival for year 2 equals a cumulative survival of 92%. An important assumption in this type of analysis is that: • No change has occurred in the effectiveness of treatment during the 3-year period • Treatment has improved during the period of the study • Persons lost to follow-up are counted in the table • The data are age-adjusted • Both a and c An important assumption of survival analysis is that separate strata, in this example, years of follow-up, have similar underlying rates of survival. If some external factor were to differentially influence survival during a portion of the follow-up time then we would not be able to assume a cumulative survival that is consistent during the entire study period. Complete the table. What is the probability that a person enrolled in the study will survive to the end of the third year? The answer is 48.6%. Completing the table gives the following results for each column: Column 5 from top to bottom (Column 2 - (Column 4/2): 350, 255, 184 Column 6 from top to bottom (Column 3/ Column 5): 0.229, 0.228, 0.185 Column 7 from top to bottom (Column 1-Column 2)/ (Column 1): 0.771, 0.772, 0.815 Column 8 from top to bottom (Column 1-Column 2)/356: 0.771, 0.596, 0.486 The cumulative survival at the end of the follow-up period equals the probability of survival during each of the years of follow-up. In this example, multiplying 0.771 by 0.772, then multiplying this product by 0.815 equals the cumulative survival rate of 0.486. Before reporting the results of this survival analysis, the investigators compared baseline characteristics of the 38 people who withdrew from the study before its end to those who had complete follow-up. This was done for which of the following reasons: • To test whether randomization was successful • To check for changes in treatment • To check whether those remaining in the study represent the total study population • To check whether the survival estimate of those remaining in the study is the same as the general population • To check that the survival estimate for those lost to follow-up is the same as the general population A key assumption for the use of survival analysis is that persons who are lost to follow-up have the same mortality experience as those remaining in the study. The failure to satisfy this assumption introduces a bias in the survival estimates since the observed population has different attributes that are associated with survival compared to the population that is lost to follow-up. Which of the following is a key assumption involved in the use of life-table analysis? • The risk of disease does not change within each interval over the period of observation • There are no losses to follow-up in the study population • The frequency of exposure is similar in treatment and comparison groups • The disease is common • The study subjects are representative of the population from which they were draw Life-table analysis depends upon a consistent rate of survival during all periods of the study. Changes in the rate of survival may be due to external influences that are operating at later times on only a portion of the initial population. Since those who have died earlier in the study period will not experience these external influences, the comparison between periods is rendered invalid. Which of the following is a measure of disease prognosis? • Prevalence • Median survival time • Age-adjusted mortality rates • Standardized mortality ratio • Proportionate mortality ratio Disease prognosis indicates the likelihood of survival once a disease has become manifest. The median survival time reflects the length of time that the 50th percentile of affected persons has. It differs from the mean survival time in that the mean survival time is an average that may be influenced by extremely low or high survival times. The median survival time consists of an ordering of all survival times with the midpoint of the distribution taken as the duration of survival. In 2003, Sudden Acute Respiratory Syndrome (SARS) appeared in several countries, mainly in Asia. The disease was determined to have been caused by a virus that could be spread from person –to person from the index case occurring in mainland China. This table reflects the total number of reported cases of SARS and deaths among those cases as best as can be determined. What is the overall case-fatality rate for the worldwide epidemic of SARS? • 9.5% • 12.4% • 16.0% • Cannot be determined as the data are not age-adjusted • Cannot be determined as the total number of all deaths is not known 807 total deaths from SARS/8458 total cases of SARS=0.0954x100= 9.5% This can be found by dividing the total number of deaths due to SARS by the total number of cases. This equals a case-fatality rate of 9.5%. Based on the table, we can conclude that the case-fatality rate (CFR) in Vietnam: • Is the same as the case-fatality rate in Singapore • Is twice as great as the case-fatality rate in Singapore • Is almost one half that of the case-fatality rate in Singapore • Cannot be determined because the data are not age-adjusted • Depends on the number of secondary cases 5 deaths from Vietnam/63 cases from Vietnam= 0.079x 100= 7.9% CFR 31 deaths from Singapore/206 cases in Singapore= 0.1504x100=15% The CFR in Vietnam equals 5 divided by 63, or 7.9%, while that of Singapore equals 15%. This is approximately one half the rate. Following a revision in the case definition, more persons were found to have suffered from an infection with the SARS virus. The inclusion of these cases, almost all asymptomatic, did not impact the total number of SARS fatalities. What happened to the case-fatality rate (CFR) following this reclassification? • It remained the same • It was increased • It was decreased • It cannot be determined without knowing the number of new cases • It cannot be determined without age-adjustment The increase in prevalent cases with no change in mortality would decrease the CFR since the numerator, number of deaths due to SARS, would stay the same while the denominator, number of cases, increased. What is the probability of surviving the second year of the study given that a person survived the first year? The independent probability of surviving the second year for all persons who survived the first year is found by dividing the number of survivors at the end of the period by the total number present at the beginning of the period. In addition, for those who withdraw during the interval, only 50% of these persons should be counted as being present during the interval. The table should be completed with the following values: Column (B) from top to bottom: 248, 124(Column B-Column C-Column D), 55(Column A- Column B-Column C) Column (E) from top to bottom (Column C)/(Column B- ½ of Column D): 0.410, 0.470, 0.296 Column (F) from top to bottom (Column B-Column C- ½ of Column D)/ (Column B- ½ of Column D): 0.590, 0.530, 0.704 Column (G) from top to bottom (: 0.590, 0.313, 0.220 Therefore, the second year survival probability among all those surviving in the study past the first year is 53%. The probability of dying during the second year equals the number of deaths during the interval (55) divided by the total number of persons alive at the start of the interval less one half of those withdrawing from the study (117). Subtracting this value from 100% results in a survival rate of 53% for the interval. For all people in the study, what is the probability of surviving to the end of the second year? The cumulative probability of survival through the second year equals the probability of survival for the first year multiplied by the probability for the second year. This equals 59% multiplied by 53%, or 31.3%. What is the probability chance of surviving 3 years after diagnosis? The cumulative survival probability for all 3 years equals the product of the independent interval survival probabilities. In this example, 59% multiplied by 53% multiplied by 70.4% gives a cumulative survival probability of 22%. What is the total number of person-years of follow-up for patients in the study assuming a median survival time of one half of the year for all persons dying during an interval and an observation time of one half of the year for all persons withdrawing from the study? This calculation involves attributing the correct amounts of person-years to each group during an interval. For the first year of the study, 96 deaths occur. Using the median survival time, we can calculate that these persons contributed 48 person-years of observation. Additionally, 28 persons withdraw from the study. Again, allocating one half of the year to each of these patients results in 14 person-years. Of the remaining 124 persons who survive for the full year, they contribute 124 person-years of observation. The total person-time for the first year of the study is 186 person-years. Continuing with this same approach for years 2 and 3 of the study, we arrive at a total of 321.5 person-years of observed study time. Before reporting the results of this survival analysis, the investigators compared baseline characteristics of the 44 people who withdrew from the study before its end to those who had complete follow-up. This was done: • To test whether randomization produced similar groups • To check for changes in prognosis • To check whether those withdrawing from the study are similar to persons remaining in the study • To check whether the outcome of those remaining in the study are the same as the underlying population • To check for confounders in the exposed and nonexposed group A key assumption in life table analysis is to insure that the experience of those lost to follow-up, or withdrawals from the study, is the same as those remaining under observation Chapter 7&8 Question 1 Which of the following statements best describe efficacy? • It is an estimate of the benefit of treatment under ideal conditions • It is an estimate of the benefit of treatment under routine conditions Question 2 A study is conducted for a pharmaceutical agent that has shown promise for reducing heart disease among women. In order to more fully test the agent, an additional study is done restricting the participants to be randomized to those who have a history of hypertension. Which of the following advantages cannot be claimed by the researchers? • Power of the study is increased • Potential benefits in high-risk populations are increased • Validity of the study is increased by focus on a homogenous population • The generalizability of the study is increased Restricting the study to a high-risk group will only fail to increase the generalizability of the findings since the study results will apply only to a more specific group of women who share a history of hypertension. Question 3 A new drug treatment for cardiac thrombus claims to have a higher success rate than the current drug. A strong sign of the potential success is the lack of internal hemorrhaging starting 2 days after treatment. 168 patients who require treatment for cardiac thrombi are randomized after agreeing to participate in a trial of the new drug. The researchers were interested in whether the new drug reduced the need for blood transfusions due to internal hemorrhage compared to the current treatment. The following table summarizes the results of her study: What is the incidence of needing a blood transfusion in the group of persons who were randomized to the new drug treatment? • 31.0% • 41.1% Incidence is the development of the outcome under investigation among those at risk. In this instance, 43 participants “developed” the need for a blood transfusion among the 84 persons in the new drug treatment group. Question 4 A new drug treatment for cardiac thrombus claims to have a higher success rate than the current drug. A strong sign of the potential success is the lack of internal hemorrhaging starting 2 days after treatment. 168 patients who require treatment for cardiac thrombi are randomized after agreeing to participate in a trial of the new drug. The researchers were interested in whether the new drug reduced the need for blood transfusions due to internal hemorrhage compared to the current treatment. The following table summarizes the results of her study: What is the number of persons who died in hospital in the study? • 7 • 17 The death rate in the study for all participants is 0.167. Multiplying this number by 168, the total number of participants, results in 28 deaths in hospital following treatment. Question 5 A new drug treatment for cardiac thrombus claims to have a higher success rate than the current drug. A strong sign of the potential success is the lack of internal hemorrhaging starting 2 days after treatment. 168 patients who require treatment for cardiac thrombi are randomized after agreeing to participate in a trial of the new drug. The researchers were interested in whether the new drug reduced the need for blood transfusions due to internal hemorrhage compared to the current treatment. The following table summarizes the results of her study: What is the main advantage of the randomization of the 168 study participants to one of the two drug treatment groups? Though the study is being conducted in a group of persons with a similar diagnosis, there should be differences in age, gender, race, and severity of conditions in the 168 participants. In order to ensure that those with a healthier profile are not preferentially selected for one of the treatment groups, randomization is used. Question 6 A new drug treatment for cardiac thrombus claims to have a higher success rate than the current drug. A strong sign of the potential success is the lack of internal hemorrhaging starting 2 days after treatment. 168 patients who require treatment for cardiac thrombi are randomized after agreeing to participate in a trial of the new drug. The researchers were interested in whether the new drug reduced the need for blood transfusions due to internal hemorrhage compared to the current treatment. The following table summarizes the results of her study: The researchers interpret the findings to conclude that the new drug treatment is more likely to result in a blood transfusion and subsequent death. This statement is: • Incorrect as the data do not indicate the death rate among only those receiving a blood transfusion • Incorrect as the number of expected deaths is not known • Incorrect as the cause-specific death rate for internal hemorrhaging is not reported • Incorrect as the duration of time from blood transfusion to death is not reported Though the death rate among those receiving a transfusion cannot be directly compared with this table, the high rate