Part 5. It may be difficult to accurately estimate the prevalence of depression
ID: 2439834 • Letter: P
Question
Part 5. It may be difficult to accurately estimate the prevalence of depression in the general population, even in high income countries. Studies face issues of selection bias and information bias in their estimates of depression. Read the following research scenarios. For each, describe one way selection bias may affect the estimate of depression and one way information bias may affect the estimate of depression. Indicate whether the bias you have identified could over- or underestimate the prevalence of depression in the general population.
a. A research study tries to estimate the prevalence of depression in a general population by collecting diagnosis information from medical records at a primary care physician office. [Bias Detection and Impact]
Selection bias:
Information bias:
b. A research study tries to estimate the prevalence of depression in the general population by surveying a member of a household through a telephone interview. [Bias Detection and Impact]
Selection bias:
Information bias:
Explanation / Answer
a. A research study tries to estimate the prevalence of depression in a general population by collecting diagnosis information from medical records at a primary care physician office.
Selection bias: Selection bias may affect the above example where the study takes in a small group of people from a primary care centre. The study has localized its selection sample rather than selecting a sample from a wider area which allows for randomization and a more unbiased group. The bias could over estimate the amount of depression as it is likely that a some amount of people in the sample suffer from tendencies of depression.
Information bias: It affects the study in such a way that the diagnostic records of the sample taken for study does not provide the information needed to determine depression in the general population. The study could over estimate or under estimate the depression in general population the records may not contain information or contain errors with respect to information on depression.
b. A research study tries to estimate the prevalence of depression in general population by surveying a member of a household through a telephone interview.
Selection bias: The sample is too small as surveying one person in a household will not be enough to determine estimates for a large population. The sample may overestimate or underestimate depression in general population depending on the outcome of the interview.
Information bias: It is likely that the person being interviewed does not provide correct information. It is possible that the person may be biased while providing information. The outcome of the interview is once again likely to overestimate or underestimate the depression in general population depending on the kind of information that the interviewee has provided.