Based on the above case, select the one response that best represents the corresponding threat to validity.
Question 1
Because there are relatively few patients in this study, it may give a misleading impression of the actual effectiveness of surgery.
Samples can give a misleading impression of the situation in the parent population, especially if the sample is small.
Question 2
The results of this study may not apply to your patient, a woman, because all the patients in the study were men.
Generalizing the results of a study of men to the care of a woman assumes that the effectiveness of surgery for low back pain is the same for men and women.
Question 3
Fewer patients, who did not have surgery, remained under care at the clinic two months after surgery.
The two treatment groups did not have an equal chance of having pain measured.
Question 4
The patients who were referred for surgery were younger and fitter than those who remained under medical care.
The difference in recovery between patients who received surgery versus medical care may be the result of some other factor, such as age, that is different between the two treated groups and not the result of surgery itself.
Question 5
Compared with patients who had medical care alone, patients who had surgery might have been less likely to report whatever pain they had, and the treating physicians might have been less inclined to record pain in the medical record.
These are biases related to measurement of the outcome (recovery from pain).
Question 6
Patients without other medical conditions were both more likely to recover and more likely to be referred for surgery.
The other medical conditions confound the relationship between treatment and outcome; that is, they are related to both treatment and recovery and might be the reason for the observed differences.
For the following questions, select the best answer.
Question 7
Histamine is a mediator of inflammation in patients with allergic rhinitis (“hay fever”). Based on this fact, which of the following is true?
The observation that histamines mediate inflammation in hay fever leads to a promising hypothesis that blocking histamines will relieve symptoms, but the hypothesis needs to be tested in people with hay fever. The other answers all assume more about the causes of symptoms than is actually stated. For example, histamine is only one of many mediators of inflammation in hay fever.
Question 8
Which of the following statements about samples of populations is incorrect?
Samples may misrepresent populations by chance, especially when the samples are small.
Question 9
You are making a treatment decision with a 72-year-old man with colon cancer. You are aware of several good studies that have shown that a certain drug combination prolongs the life of patients with colon cancer. However, all the patients in these studies were much younger. Which of the statements below is correct?
Generalizing from younger to older patients is a matter of personal judgment based on whatever facts there are that bear on whether older patients respond to treatment the same as younger patients. Internal validity is about whether the results are correct for the patients in the study, not about whether they are correct for other kinds of patients. Both bias and chance affect internal validity.
Question 10
A study was done to determine whether regular exercise lowers the risk of coronary heart disease (CHD). An exercise program was offered to employees of a factory, and the rates of subsequent coronary events were compared between employees who volunteered for the program and those who did not volunteer. The development of CHD was determined by means of regular voluntary checkups, including a careful history, an electrocardiogram, and a review of routine health records. Surprisingly, the members of the exercise group developed higher rates of CHD even though fewer of them smoked cigarettes. This result is least likely to be explained by which of the following?
If volunteers had the same amount of exercise as the nonvolunteers, a difference in the rate of CHD could not be explained by this variable. Differences in the groups (selection bias) or in the methods used to determine CHD (measurement bias) could account for the finding.
Question 11
Ventricular premature depolarizations are associated with an increased risk of sudden death from a fatal arrhythmia, especially in people with other evidence of heart disease. You have read that there is a new drug for ventricular premature depolarizations. What is the most important thing you would like to know about the drug before prescribing it to a patient?
The drug’s effect on mortality is the most important thing to determine. It is possible that some arrhythmia-suppressing drugs can increase the rate of sudden death (in fact, this has happened). In such situations, the intermediate biologic outcome of decreased arrhythmias is an unreliable marker of the clinical outcome—sudden death.
For each of the following statements, select the one response that best represents the corresponding threat to validity based on the above report.
Question 12
Women with thrombophlebitis may have reported the use of contraceptives more completely than women without thrombophlebitis because they remembered hearing of the association.
Measurement bias is frequently an issue when patients are asked to recall something that may be related to the illness, because those with illness may have heightened recall for preceding events that they think might be related to their illness.
Question 13
Doctors may have questioned women with thrombophlebitis more carefully about contraceptive use than they did those without thrombophlebitis (and recorded the information more carefully in the medical record) because they were aware that estrogen could cause clotting.
The questioning (measurement) about contraceptive use was not the same in the two groups.
Question 14
The number of women in the study was small.
Small study numbers increase the possibility that chance accounts for differences between groups.
Question 15
The women with thrombophlebitis were admitted to the hospital by doctors working in different neighborhoods than the physicians of those who did not have thrombophlebitis.
Different neighborhoods often are surrogates for socioeconomic variables that are related to numerous health outcomes, which may mean the two groups are different in terms of important covariates.
Image credits
Unless otherwise noted, images are from Adobe Stock.