统计问题第177问：估计风险人群

Question

Risk factors for cancers of the upper digestive tract have been investigated using different study designs. These include a population based case-control study in Golestan province, northern Iran, which investigated the association between tea drinking habits and the risk of oesophageal squamous cell carcinoma. In total, 300 cases of histologically confirmed oesophageal squamous cell carcinoma and 571 matched neighbourhood controls were identified. Participants were asked about their tea drinking habits in face to face interviews. Drinking hot tea, compared with drinking lukewarm or warm tea, was found to be strongly associated with increased risk of oesophageal cancer.

In a study in Denmark, researchers used a population based prospective cohort design to examine the association between different types of alcoholic drink and cancers of the upper digestive tract (oropharyngeal and oesophageal). Participants were assessed at baseline for alcohol intake and smoking habits. Cohort members were 15 117 men and 13 063 women aged 20-98 years, who were followed for a mean of 13.5 years (range 2-19 years). The researchers reported that when compared with non-drinkers a moderate intake of wine was unlikely to increase the risk of upper digestive tract cancer, whereas a moderate intake of beer or spirits was associated with a considerably increased risk.

Which of the study designs, used above, facilitates the estimation of the population at risk?

·a) Case-control study

·b) Prospective cohort study

提示：正确答案只有一个。

Answer

The population at risk can be estimated from a prospective cohort study (answer b) but not from a case-control study (answer a).

To be able to estimate the population at risk entails estimating, from the sample, the incidence or prevalence of cancer of the upper digestive tract in the population. These estimates need to be made not only for the entire population but also for people in the population who participated in the risk behaviour and for those who did not. It was possible to estimate the population at risk from the cohort study because it was prospective in design (answer b) but not for the case-control study (answer a) because it was retrospective in design.

The study of the association between different types of alcoholic drink and upper digestive tract cancers used a prospective cohort study design. The cohort was a sample of 15 117 men and 13 063 women aged 20-98 years, selected as a representative sample of the adult population in Denmark. The cohort was followed prospectively for a mean of 13.5 years, and all cohort members were at risk of developing upper digestive tract cancers. Therefore, the risk of developing upper digestive tract cancers in the population could have been estimated from the cohort—not only for the population as a whole, but also for those members of the population who participated in the risk behaviour of alcohol drinking and for those who did not. Hence, the prospective cohort study was said to estimate the population at risk of upper digestive tract cancers.

Being able to estimate the population at risk from the prospective cohort study enabled the incidence or prevalence of upper digestive tract cancers in the population to be estimated. It also allowed a relative risk to be derived, so that the risk of upper digestive tract cancers could be compared for alcohol drinkers relative to non-drinkers. Relative risks, described in a previous endgame, should be calculated only if the population at risk can be estimated.

The case-control study undertaken in northern Iran investigated the association between tea drinking habits and risk of oesophageal squamous cell carcinoma. It was not possible to estimate the population at risk of oesophageal cancer from this study. The study design involved selecting two groups on the basis of their disease status—those with oesophageal cancer (the cases) and otherwise healthy people (the controls). Cases and controls were then interviewed about their past tea drinking habits. The numbers of cases and controls, and the ratio of one group to the other, were decided in advance by the researchers. Therefore, the proportion of participants with oesophageal cancer overall would not have estimated the proportion in the population. Furthermore, of those study participants who self reported the risk behaviour—drinking hot tea—the proportion with oesophageal cancer depended on the number of cases and controls selected. This also applied to those study participants who drank warm or lukewarm tea.

Because the population at risk of oesophageal cancer could not be estimated from the case-control study, it was not possible to derive the relative risk of oesophageal cancer for drinking hot tea relative to lukewarm or warm tea. Instead, an odds ratio—which is an estimate of the population relative risk—would have to have been calculated from the sample. Odds ratios have been described in a previous endgame. When the prevalence of the disease is rare—typically thought to be less than 10%—the odds ratio is regarded as a good estimate of the population relative risk. This can be demonstrated by simple mathematics.

所以答案是选择  b

每天学习一点，你会更强大！