4 Single Case Studies Flashcards
How are Single case reports and experimental case studies similar?
Provide a detailed account of the therapy
Attempts to measure change
Evidence of patient improvement
Maybe range of measures to capture the change and extent
Early stages of therapy investigation – newly established therapy
Good for studying rare phenomena.
How do experimental case studies differ from single case reports?
🔸 Priori methodology
🔸 Systematic manipulation of variables
🔸 Controlled measures
🔸 Statistical data analysis
🔸 empirical evidence for intervention
Benefits of experimental case studies
🔸 Can be integrated into normal clinical practice
🔸 Provide clinically replicable data
🔸 Answers clinically relevant questions
🔸 Contributes to evidence base – use of statistical tests
BUT conclusions can’t be generalised to the wider population
What are the principles of single case methodology?
Must supply clinical history so the reader can determine the applicability of treatment
Must identify target behaviours – this will form the basis of the primary outcome measure
Must have an established design, including experimental control; this allows us to attribute changes to the therapy
Must do use reliable measures
Repeated baseline measures used; this is so that treatment can be showing to exceed normal variation
Provides raw data
Independent assessors are used.
A statistical analysis is carried out – significance of results
Reduplication
List the three design examples
1️⃣ Time series
2️⃣ Control items
3️⃣ Control behaviours
Time series. Explain.
Use baseline trends and see how much therapy outcomes exceed them.
Measure the target behaviour before therapy (A1), during therapy (B), and after therapy is withdrawn (A2).
A1 -> B
🔸shows change therapy may cause
A2
🔸shows whether effects are maintained
📝use extrapolated baseline to see progress w/o intervention
Problems with time series design thingy
Respondent burden and/or learning effects with repeated testing.
Repeated measures may influence the behaviour of interest – test effects.
Requires a reasonably long no intervention baseline period which raises ethical concerns.
How do we analyse timeseries data?
Use sign test to determine the number of observations that go against the hypothesis.
C statistics.
Mann test
How do Carry out the sign test?
Count the number of observations that go against the hypothesis – how many points are above the trend line.
(S) = number of observations that go against the hypothesis
🔸look up value in table -> p value
🔸must be under 0.05 to be significant
What’s the Mann test and how do you carry out?
Use this one when there is no trend line.
Assessment at time points and score. Need to see how many values are greater (p) than the score assessment point, and how many are lower (q).
S = Σp - Σq
🔸look up on table for p
🔸improvement if Σp>Σq
Control items: how can we adjust variables and design to improve the experiment?
🔸follow up results - maintained?
🔸repeated measures
🔸control sets - generalisation
Repeated measures
🔸 Pre 1 - before therapy
🔸 Pre 2 - Imm before therapy
🔸 Post 1 - imm after therapy
🔸 Post 2 - after period of time
Comparison of times:
🔹pre 2 vs post 1 ➖ imm therapy effects
🔹pre 2 vs post 2 ➖ maintenance
How do we analyse repeated measures data?
Chi - squared.
🔸 how many items improved (a)
🔸 how many items worse (b)
X^2 = (a-b-1)^2 ➗ (a+b)
🔸 look up p on table
For between groups, we want there to be no diff in baseline for groups.
What must be considered when controlling behaviours?
Can be reliably measured several target behaviours?
Are they are compatible difficulty?
Is control more difficult than the therapy behaviour?
Will there be no contamination?
Does one influence the other – there should be no overlap
When controlling behaviours, how can we make sure that there is no contamination?
Use crossover addition – treat a second behaviour after treating the first one???
