统计问题第178问：什么是风险?

Question

Titrated oxygen was compared with the standard treatment of high flow oxygen for patients with acute exacerbation of chronic obstructive pulmonary disease in the pre-hospital setting. A cluster, randomised controlled, parallel group trial was performed. In total, 405 patients with a presumed acute exacerbation of chronic obstructive pulmonary disease that had been treated by paramedics and who had been transported and admitted to hospital were included. A total of 226 patients were randomised to high flow oxygen and 179 to titrated oxygen. The main outcome measure was mortality, either pre-hospital or in-hospital mortality.

The risk of mortality in patients randomised to titrated oxygen treatment was less than that in the patients randomised to high flow oxygen (seven deaths (3.9% of patients) versus 21 deaths (9.3%)). The researchers concluded that titrated oxygen treatment significantly reduced mortality in comparison with high flow oxygen in acute exacerbations of chronic obstructive pulmonary disease (relative risk 0.42 (95% confidence interval 0.2 to 0.89; P=0.02)).

Which of the following statements, if any, are true?

·a) The risk of mortality is the probability of death.

·b) Each patient in the titrated oxygen treatment group had a risk of death of 0.039.

·c) The risk of death in the titrated oxygen treatment group was reduced by 58% relative to the high flow oxygen treatment group.

提示‍：正确答案不止一个。

Answer

Statements a and c are true, while b is false.

In the above example, the effects of titrated oxygen on mortality were compared with the standard treatment of high flow oxygen. Participants were included if they had had a presumed acute exacerbation of chronic obstructive pulmonary disease in the pre-hospital setting. Mortality was less with titrated oxygen than with high flow oxygen (seven deaths (3.9%) versus 21 (9.3%)). The risk of death (pre-hospital or in-hospital) is equivalent to the probability of death (a is true). Probability is on a numerical scale from zero to one, inclusive, and describes the frequency of an event occurring. In the example above, the event was death, and the probability of death was the proportion of patients who died within the trial period. In the high flow oxygen treatment group 21 of a total of 226 patients died, giving a probability of death of 0.093. In the titrated oxygen group seven of 179 patients died, a probability of death of 0.039. This derivation of probability is termed the frequentist approach, as it based on the observed frequency in the sample.

The word “risk” is typically used in a negative connotation, and when used as a verb it implies the exposure of someone to danger. However, in statistics risk is an alternative word for probability and does not necessarily imply that something harmful will happen. Nevertheless, because healthcare is typically concerned with the prevention of adverse outcomes, risk may have become an acceptable alternative word for probability. In essence, probability measures the uncertainty of an event occurring. The event may be harmful, such as death as in the above example. However, the event could be pregnancy, which is more often than not regarded as a positive outcome.

The probability of death was 0.039 for titrated oxygen treatment and 0.093 for high flow oxygen. By definition, these probabilities apply to the treatment groups and not single patients within the groups. Therefore, patients in the titrated oxygen treatment group did not each have a risk of death of 0.039 (b is false); rather, the treatment group had a risk of death of 0.039.

The relative risk of death for titrated oxygen treatment in comparison with high flow oxygen was 0.42. This is a sample estimate of the population parameter. Relative risks have been described in a previous question. The relative risk describes how exposure to titrated oxygen affected the risk of death relative to treatment with high oxygen flow. Therefore, the risk of death was reduced by 58% with titrated oxygen relative to high flow oxygen (c is true). This reduction in risk applies to the titrated oxygen treatment group and not to a single patient receiving that treatment.

The risk of death was reduced if patients were treated with titrated oxygen rather than high flow oxygen. The sample results estimate the effects of treatment in the population. However, it may be difficult to apply the results to a single patient. The possibility of pre-hospital or in-hospital death of a single patient with a presumed acute exacerbation of chronic obstructive pulmonary disease will not be known in advance. Obviously, a patient may have comorbidity that increases the possibility of their death, but it is not possible to attach a probability to that event. Treatment with titrated oxygen does not guarantee survival; the above study showed only that the treatment was associated with a reduction in the frequency of death in comparison with high flow oxygen treatment. Communicating that concept to patients and clinicians alike may be difficult. Furthermore, what is not known is whether patients who died having been treated with high flow oxygen would have survived with titrated oxygen or whether those patients who survived with titrated oxygen would have died with high flow oxygen.

所以答案是选择 a c

每天学习一点，你会更强大！