Framing a Research Question & Study Designs

Question 1

Outline the possible steps of turning an issue into a question?

Answer 1

1. You could develop a hypothesis to test as a means to offer an explanation for a particular issue or behaviour eg. Junior Drs emigrate because of low pay.

Note: this is not always appropriate, if a topic is exploratory you may just want to ask ‘Why’ to gather enough info to develop a hypothesis.

2. Quantitative Work: Developing a research question from a broad topic.

Step 1. ID concept or phen. that is crucial to the issue you are researching.

Step 2. ID the variable associated with it - which ones are relevant to the research question?

Step 3. Decide on indicators: decide on what you can measure (numerical) and observe.

Question 2

Give some examples of indicators?

Answer 2

MRs
Time taken off work due to injury
Education level
Age
SE Status
Weight etc. 
Self reports of illness

Question 3

What are some limitations of self reported data?

Answer 3

1. Education level
2. the health seeking behaviour of the individual
3. The needs of the person in question (people with more needs may not be the one self reporting)
4. The setting where the survey was filled out - primary hc setting or secondary HC setting.

Question 4

Define Proxy Indicator?

AKA indirect indicator.

Answer 4

An indicator that doesn’t measure the variable directly but has some approximation of the measure.

Example: Indicator = clinical measured weight
proxy indicator = self reported weight.

Question 5

What is an experimental design?

What are they good for?

Answer 5

Generally a trial Ex. RCT.

Good for testing the effectiveness of an intervention Ex. new medicines or therapies.

Question 6

How many groups are generally involved?

What does random allocation allow us to do?

Answer 6

Generally has an experiment/intervention group + Control

RA allows us to control for selection bias. Control for known and unknown confounders by spreading them evenly across 2 groups.

Question 7

What are pragmatic trials?

give an example of a well known one?

Answer 7

• Trials that aim to test an intervention under real world settings so as to increase the generalibility of a study.

Example: Ann Oakley (1990) - pragmatic trial of social support for pregnant women to improve their own and their childs health.

Question 8

What are some issue with PTs? (3)

Answer 8

1. Random allocation to intervention vs no intervention is difficult in real life settings e.g if someone really needs an intervention, but they may not get it by RA.
2. Contamination - patients may find out about the benefits of the intervention and try to get in the group or speak to other patients.
3. Conflict between informed consent and rigorous testing - more patients know about a trial the more chance for bias in outcomes.

Question 9

When is experimental design appropriate?

Answer 9

To answer causal questions

ie. when we need to be confident a intervention has a health benefit in a population.

Question 10

When is experimental design inappropriate? (4)

Answer 10

When answering explanatory questions

Lack of resources - RCTs are expensive and resource intensive

Ethical or practical reasons depending on Research Q

Sometimes difficult to separate factors of a phenomenon to study - may be more appropriate to study multiple at once, rather than trying to control everything.

Question 11

What do we need to take into account when evaluating a complex intervention? (6)

Name the author - Craig, 2008.

Answer 11

1. That simple and complex interventions are not clear cut
2. Complex ones typically have anumber of interacting compnenets.
3. The number of behaviours being studied and the difficulty of behaviours
4. No. of groups being targeting by the intervention
5. No. of variablility of outcomes.
6. The degree of flexibility or tailoring in an intervention.

Question 12

What 2 things are we trying to find out when evaluating experimental designs?

Answer 12

1. Whether they work in everyday practice

2. How do they work in everyday practice