AAAA Therapy Flashcards
(35 cards)
what is the AAAA framework?
Assess: what type of healthcare question –
what type of study?
Access: finding the ‘best’ evidence (validity
and relevance)
Appraise: evaluating the quality of the
evidence & interpreting the results
Act: is this evidence relevant to my clinical
practice. Should this evidence change my
practice?
why is the AAAA framework
important?
- An essential professional and academic skill
- Medical knowledge is continually evolving.
- The medical profession frequently fails to use
effective treatments even when they are
available. - Keeping up to date is a lifelong commitment for
every doctor. - You need to develop and use the skills to find,
appraise and act on research evidence.
explain the assess part of the AAAA framework
Assess: what type of healthcare question do you have - what type of study would best answer your question?
For this session we looking at effectiveness questions:
Will this therapy cause a beneficial effect in my patient?
Therefore, we will be using population based research (RCT) to answer the question:
how do you formulate an effectiveness question?
We formulate an effectiveness question by using a PICO framework:
Population
Intervention (treatment we are evaluating)
Comparator (alternative treatment/no treatment)
Outcomes (what are outcomes of interest) – benefits and harms
Formulating a PICO is important to:
To help with searching for evidence in electronic bibliographic databases such as MEDLINE
To help with assessment of applicability of population-based research to your patient population
how would you formulate a PICO question with this scenario:
A 45-year-old man consults you (his GP) for treatment for his chronic low back pain. You would normally refer him for a course of physiotherapy but the patient has requested a referral to a chiropractor. You want to know how effective chiropractic treatment is, compared to physiotherapy, for the treatment of low back pain before deciding where to refer him.
P –Adults; with chronic low back
pain
I - Chiropractic treatment
C –Physiotherapy
O - Pain at 3 months
- Ability to walk a mile
- Ability to work
- Harms from treatment
explain the access part of the AAAA framework
Access: finding the ‘best’ evidence
evidence that is most appropriate for your type of research question (RCTs)
evidence that is clinically relevant to your question
- Randomised control trials can be found:
Cochrane Library: https://www.cochranelibrary.com/central
Bibliographic databases like MEDLINE, PUBMED, EMBASE
MEDLINE and EMBASE have study design filters (including an RCT filter)
PUBMED has a clinical queries facility - Search Method
Use the components of your question (PICO) to make a search - usually terms for Intervention of interest and if necessary, to narrow search, the Population
Also if necessary restrict the search to RCT
Often more than one term is used to describe Population and Intervention components. These can be taken from:
bibliographic database thesaurus’ (MEdical Subject Headings)
Non MESH terms for the question component (free Text terms) – terms that you know are used to describe intervention of interest (use mixture to find most relevant terms.
Combine terms using Boolean operators:
“AND” gives search results which include both search terms
“OR” gives search results which include either one or other or both search terms
how is appraisal of the RCT design carried out
looking at how well RCT carried out, is it free from bias?
At its simplest, Appraisal asks:
Does the study address a research question which is relevant to my clinical problem?
(is there a match between my PICO & the study question?)
Did they use the study design most likely to provide a valid answer to the clinical question?
(for effectiveness question was it an RCT?)
Was the study done well / is the study trustworthy?
(were steps taken to reduce bias?) A poorly conducted RCT might be less valid that a well conducted cohort study.
If the study was done well (valid), what were the results?
(direction, size, precision and statistical significance of results)
were steps taken to reduce BIAS?
The targets below represent seven different measures of blood pressure from same person.
The true blood pressure is the centre of the target.
Which targets are biased?
Which targets are imprecise?
A + B are biased:
There is systematic deviation - not only are measurement deviated from the true BP, but deviations are all in the same direction i.e., pattern of deviation is systematically in one direction.
C + D are not biased:
Although measurements are not a true reflection of BP, the nature of deviation from the truth is more random
This is a random error.
No systematic pattern of deviation.
B + D are imprecise because measurements are scattered.
what is the method of a RCT?
A RCT is perspective: exposure (treatment) comes before the measuring of outcomes.
In a RCT individuals are allocated to receive either a new treatment or comparator treatment (existing treatment/no treatment)
The allocation to the two group occurs via randomisation
what are the key features of the RCT that needs to be included?
Comparator Group
Randomisation
Why have a control or comparator group?
If the comparator group was no treatment, it allows us to evaluate whether patients would get better without any treatment
Or if the comparator was another an existing treatment it, it allows us to see which treatment creates a better prognosis.
Why shouldn’t the drug manufacturer decide which patients are allocated to the new treatment or the comparator?
In any study the population will have a range of prognoses.
This is because individuals may have more/less severe form of the disease or the may have comorbidities.
Therefore, the ability of the tested treatment to benefit them will be different.
If we allow the manufacture to decide who receives new treatment and who gets existing treatment, there may be a tendency for people with a vested interest to make treatment look good to select people who have a better prognosis for their new drug and to allocate people who will receive competitor treatment into other arm.
You have an imbalance in patient characteristics of two arms that are being compared.
So regardless of how effective new treatment is, we would expect outcome to be different.
This is called selection bias
Why shouldn’t the patient’s Dr decide whether they want the new treatment or the comparator?
The Dr and their patient may feel that if they have more severe disease they have more to benefit from receiving a new treatment.
Therefore, again get an imbalance in the two arms so we have another selection bias.
We would expect outcomes to be different in two groups, irrespective to how effective new treatment it.
what needs to be done to ensure comparison of groups?
To ensure a valid comparison we want the groups to be similar on the outset of the trial.
This ensures any characteristics of participants that may affect the outcome will be equal across the groups compared.
This is done by randomisation.
Randomisation mitigates against section bias.
how is randomisation done in RCT?
Allocation to the intervention or the comparator / control arm should be done via randomisation:
Randomisation: Every person has an equal chance of ending up in either the intervention or comparator arm
Therefore, allocation to intervention or control is decided by chance alone
This should:
Minimises the chance of selection bias 🡪 minimises differences between groups being compared at the start of a trial.
Therefore, at the end of the trial any differences observed should be due to differences in the intervention and comparator and not due any difference between groups
In practice randomisation is achieved by computer programs generating a random string of numbers to allocate patients to different arms
what is the first design feature of the RCT to mitigate against bias?
Randomisation alone is not enough to mitigate bias…
Firstly:
It is compulsory that the random allocation sequence is not disclosed to patients and those involved in recruiting trial participants.
Recruiters could select more optimal patient (more likely to benefit) for intervention arm resulting in selection bias.
This can be done by ensuring the computer-generated random sequence is performed distant to where the trial is being conducted.
what is the second design feature of the RCT to mitigate against bias?
Secondly Allocation Concealment to avoid section bias
Also, where possible, treatment allocation should be concealed until its administration (sealed envelopes, similar packaging, placebo etc)
This is called allocation concealment
It helps to ensure that allocation knowledge cannot influence (as per previous slides) who actually receives which treatment
Therefore allocation concealment minimises the chance of selection bias
what is the third design feature of the RCT to mitigate against bias?
Thirdly Blinding to avoid measurement bias
Patients may have preconceived ideas about how effective a treatment is.
This can influence their perception of subjectively measured outcomes such as pain, low mood or quality of life.
We might observe the placebo effect:
‘a beneficial effect produced by a placebo drug or treatment, which cannot be attributed to the properties of the placebo itself, and must therefore be due to the patient’s belief in that treatment’.
Patients may also offer socially desirable responses if they know they are in the intervention arm – regardless of they are feeling better and they know they are in intervention arm they may feel they should say they feel better.
Observers (those measuring outcomes) may have preconceived ideas about how effective a treatment is. This can influence how they measure subjective outcomes.
Therefore, BLINDING is done to avoid measurement bias
what is the fourth design feature of the RCT to mitigate against bias?
Fourthly Blinding to avoid performance bias
Healthcare professionals may have preconceived ideas about how effective a treatment is.
This can influence (consciously or subconsciously) how they deliver treatments, and may lead to participants in either the intervention or comparator arm receiving ‘extra’ (either subconscious or conscious) – less likely to happen if less flexibility around how treatment is delivered.
E.g. physiotherapy more open to manipulation than giving e.g. a tablet
Study participants may be disappointed about which arm they are randomised to which may lead to them behave differently, for example not engaging with treatment.
Therefore, BLINDING also helps us to avoid performance bias.
who can be blinded in a trial?
Blinding can take place at several levels in trial
People who can be blinded:
Participants
Clinicians delivering treatment
Researchers measuring outcomes
Statisticians undertaking analysis
Depending on how many types of participant of a trial are blinded we can have a:
Single Blind trial
Double blind (any 2)
Triple blind (any 3)
In a RCT of chiropractic treatment versus physiotherapy for chronic back pain, pain was measured after 3 months.
Which of the following measures is least likely to undermine the validity of RCT?
The most objective will be least prone to manipulation by those who are measuring outcomes or by patients.
This means it is least likely to undermine the validity of the RCT.
what are the 2 elements of follow up that should be fulfilled?
Follow up should be as complete as possible:
Patients may be lost during the trial so cannot be followed up.
Patients may be available for outcome measurement but are not adhering to the trial protocol (not taking/receiving allocated treatment)
Follow up should be ‘long enough’ for the outcome being measured:
If we are looking for recovery from infection: days / weeks follow up
For improved mobility following hip replacement: several months
For weight loss: several months / a year
For remission free recovery in cancer: follow up of years
In an RCT with 1000 participants, which of the following scenarios is least likely to undermine validity?
85 patients were lost from the intervention group and 80 from the comparator group
70 patients were lost form the intervention group and 10 were lost from the comparator group
5 patients were lost form the intervention group and 30 from the comparator group
Answer: 1
Although in total we have larger number of patients lost there is a similar loss in either arm.
So, it is also about whether loss is similar. Why?
Attrition bias: is a systematic error caused by unequal loss of participants from a randomized controlled trial (RCT).
In our study participants with a mix of prognosis we would expect that imbalances in loss of follow up are likely to occur due to patient characteristics which determine prognosis.
This is called attrition bias.
E.g. patients who tend to drop out of trials might be those with more severe disease who are randomised to receive a comparator treatment or no treatment. As those with most severe disease have most to benefit.
So we have imbalance in loss.
Those who have milder disease have least to benefit from new treatment but most to lose from side effects so also a systematic difference between how people with milder disease drop out of intervention vs comparison arm.
Therefore, a valid RCT has a minimal loss to follow and up and any loss should be similar across groups. I.e. no attrition bias.
