What are my chances, doctor?

Critically appraising prognosis

J Irlam
Primary Health Care Directorate
UCT Faculty of Health Sciences
jirlam@cormack.uct.ac.za

Clinical scenario:

Does obesity increase the risk of dementia?

An obese 41-year-old woman patient with a body mass index (BMI) of 32 presents in your clinic. She remembers reading a newspaper report about the link between obesity and dementia in later life, and is concerned about her risk. She asks you about her husband’s risk of dementia too, as he is also middle-aged and overweight.

Prognostic factors:

Characteristics of a particular patient that can be used to more accurately predict that patient’s clinical outcome.

• Prognostic factors are associated with outcomes, but are not necessarily causes.
• Risk factors are not necessarily prognostic factors (e.g. smoking vs. tumour stage in lung cancer).

1. Ask the right question:

2. Access the evidence:

Whitmer RA et al. Obesity in middle age and future risk of dementia: a 27 year longitudinal population based study. BMJ. 2005 Jun 11;330(7504):1360.

3. Appraise the evidence - 1:

Is the study design appropriate to the question?

• Cohort studies with a well-defined representative sample, objective outcome criteria and good follow-up are best.
• Case-control studies are very efficient, especially for rare diseases, but are more prone to bias and confounding.
• RCTs are usually not feasible for assessing prognosis in the long term.

3. Appraise the evidence - 2:

Are the results of the study valid?

1. Was there a representative and well-defined sample of participants?
2. Were the groups similar at baseline w.r.t to potential confounders (i.e. important prognostic factors)?
3. Were measurements of exposure valid?
4. Were sufficient (>80%) and similar numbers of exposed / nonexposed participants followed up to completion of the study?
5. Was duration of follow-up sufficient to detect outcomes?

6. Were outcome measurements accurate?
   • Were those assessing outcomes blinded?
   • Were measurement procedures standardised?

What are the results?

7. How likely are the outcomes over time?

   • Examine hazard ratios (HRs)

8. How precise are the estimates of likelihood?

   • Examine 95% confidence interval.<.li>

4. Apply the evidence:

• Are the results relevant to my patients?
• Are the results useful in my management of patients?