Pulm MDM Flashcards
What is the difference between a risk factor and a prognostic factor?
A risk factor is a patient characteristic that is associated wtih acquiring a disease
A prognostic factor is a patient characteristic associated with an outcome of a disease. Prognostic factors are most relevant to people who are already sick
What is the clinical course of a disease?
The evolution (prognosis) of a disease that has come under medical care
Treatment affects the course of events after the disease is recognized
What does “zero time” refer to in prognostic studies?
A common point in time in the course of a disease
Cohorts in prognostic studies should begin at this common point.
What does an “event” refer to in a prognostic study?
A dichotomous clinical outcome that can occur only once
ex: survival, death, cure, recurrence, freedom from symptoms
What does it mean when a patient is censored from a prognostic study?
A “censored” patient is no longer counted in the demonminator (patients remaining in the study)
Patients are censored when they are lost from the study at any point
This allows us to use the information we got from the patient while they were in the study without skewing the data after they left the study
How might sampling bias occur in prognostic studies?
Sampling bias occurs when the patients in the study are not like other patients with the condition
- When patients in the study are…
- More or less sick
- From a certain population
- Also, when sampling produces groups that are systematically different
What causes migration bias in prognostic studies?
Migration bias occurs when the patients who drop out during follow-up are systematically different from those who remain in the study
=> There may be more bias late in the study than early
- Chance has a higher probability of skewing results near the end of a prognostic study than at the beginning
- 95% confidence intervals are larger at the end of a prognostic study than at the beginning
What causes measurement bias in a prognostic study?
When members of the cohor are not all assessed similarly for an outcome
How are prognostic studies analyzed?
Time-to-Event or Kaplan-Meier analysis
- Time-to-Event when the event is anything other than survival
- Every time somebody dies, calculate the #surviving/#at risk from dying
- Look at 5-year survival, median survival
- Interpret confidence intervals, shape of the curve