RCTs vs observational designs

Question 1

Advantges of observational studies are

Answer 1

• Inclusion of more heterogenous participants - and therefore better generalizability (unless the inclusion/exclusion criteria are chosen rather strictly, too)
• They better reflect the real world
• Good to assess the safety profile, especially rare and late adverse effects
• Observation of naturally occurring events/exposure
• When RCT are not ethically possible

Question 2

LIMITATIONS of observational studies

Answer 2

The results are highly influenced by the quality of the collected data especially in retrospective studies – if the collected data is incomplete or of bad quality, the results will be too
- Groups might be imbalanced due to confounders - a good method to adjust for confounders has to be chosen
- We need to possibly include all existing confounders - but some might be unknown or not measured
- Prospective studies often run for a long time

Question 3

RCT vs observational studies

Answer 3

RCT: evaluación no sesgada de eficacia y seguridad de intervención / la elegibilidad restringida limita la aplicabilidad (generalización)/ exposición a la intervención es corta / costoso y complejo
OBSERVACIONAL: evaluación de intervención potencialmente sesgada / se incluyen amplio rango de pacientes / exposición a tratamiento puede ser prolongada / rápido y barato

Question 4

STRENGHTS OF RCTs

Answer 4

They are prospective - meaning we get to collect the data as it happens (and can also react if there is some more data we need instead of having to hope that it is already in a databank)
They have strict inclusion/exclusion criteria – meaning we control who is in the trial
That results in a more homogenous population which is good to show efficacy
Randomization to minimize selection bias
We control the intervention - we decide on the what and when and how, which also heightens comparability and reproducibility
They have pre-defined endpoints - meaning we know from the beginning what we are looking for and can make sure we measure it right

Question 5

LIMITATIONS RCTs

Answer 5

Some things are simply unethical to test in RCT – for example if we cannot withhold treatment from patients just to have a placebo control, e.g. life-saving or emergency treatment
The strict inclusion/exclusion criteria might impact the generalizability of the results and therefore the usefulness of the study results
We are testing treatments under controlled, ideal circumstances that do not reflect real life
Complicated statistical analysis
They are still subject to bias
Can often not assess rare or late adverse effects
Not feasible for rare diseases
Expensive