RESS Flashcards
A research Study
Generate new knowledge when none is available
Audit
Quality improvement process to improve patient care
2 aims of an audit
1) Aspects of structure, process and outcomes are selected and evaluated
2) Is what ought to be happening actually happening
Service evaluation
Evaluates a current service/practice to inform local decision making
Difference between an audit and a service evaluation
Audit = designed to answer certain questions (Does this service reach a certain standard?) - Only interested in the proportion of patients who receive the care recommended (Outcome)
Service evaluation = What standard does this service achieve - Interested in variation in the characteristics of healthcare delivery (Exposure)
What is the Commision for Patient and Public Involvement in Health (CPPIG)?
An organisation set up to make sure the public is involved in decision making about health and services - over 400
What can/does a CPPIG do?
Have statutory powers to ensure they are listened to
Provided with training and development opportunities
Obtain views of the community
Influence design and access to NHS services
What is the engagement cycle?
A strategic tool to help identify who needs what in order to engage communities at each stage of commisioning
What are the burdens and risks of a study?
Anything that may harm the patient
The discussion of unsettling issues
Physical or mental discomfort,
Disclosure or loss of personal information
3 situations when ethical approval + consent may not be required
1) Changes in practice are routine and not part of the study’s design and either;
2) All the data used have been collected during/ as an established part of routine practice
3) The collection of any additional data used do not impose excessive burdens or risks on participants
4 Ethical responsibilities in the design and conduct of a study
1) Well designed and with potential benefits that justify the risks, burdens and resources involved
2) Seek to minimise the burden and risk
3) Ensure participants provide informed consent
4) Wary of expected and unanticipated burdens/risks during the study
What do NICE do?
Independent organisation responsible for providing national guidance and advice to improve health and social care
3 roles of NICE
1) Reduce variation in availability and quality of treatment
2) Help resolve uncertainty about which medicines and treatments work best and which represent best value for money
3) Set national standards on how people with certain conditions should be treated
Define a Guideline
A comprehensive set of recommendations for a particular disease or condition
Define a Quality Standard
A prioritised, concise set of statements with associated measurable indicators, chosen and adapted from the clinical guideline recommendations.
What do National Collaborating Centres (NCC’s) do?
Provide technical input and draft the guideline
What do Guideline development groups (CDG’s) do?
Review evidence base and make recommendations
Who is involved in a CDG?
Chair - highly experienced and respected clinical leader
Clinical and academic experts
Patients and lay members
NCC technical team
4 checks when choosing outcomes and exposures
1) Exposure must precede outcome
2) Exposure must be different to the outcome
3) Both must be something specific
4) Must be formatted as binary variables - A + B
Define a Confounder
Can cause the outcome and the exposure
Must adjust for them as can create a pseudo-causal path which will generate a statistical relationship between the 2 events when non-exists
Define a Mediator
Can cause the outcome and be caused by the exposure
Don’t need to adjust for them as not part of the causal pathway and they are an important pathway
Define a competing exposure
Can only cause the outcome
Might adjust for these if they cause a substantial amount of variation in the exposure as it may make the association easier to detect
3 criteria for a causal relationship
1) Only possible if the variable precedes the effect
2) Variables must be temporally distributed (Past to present)
3) Should only be ommited when there is very firm evidence that no causal relationship exists
What does a Directed Acyclic Graph do?
1) Summarises the theoretical and speculative causal relationships between variables pertinent to our question
2) Identifies any measurable covariates acting as cofounders, competing exposures and mediators
3 Pros of a Prospective study
1) Fewer sources of bias
2) Less chance of unmeasured confounding variable
3) Some variable can only be measured prospectively
4 Cons of a Prospective study
1) Time and resource intensive
2) Tendency to collect data on more variables than you need or can use in your analysis
3) Participants can drop out during the study - bias
4) Usually infeasible for rare outcome
2 Pros of a Retrospective study
1) Data collection is less time and resource intensive
2) Allows oversampling for rare outcome
3 Cons of a Retrospective study
1) More susceptible to bias
2) Some variables cannot be measured directly
3) If data is from records, little control over how they have been measured
Prospective or Retrospective?
Case Control
Cohort
Cross-sectional
Retrospective
Prospective
Mixture
Define
1) Defining variables conceptually
2) Defining variables operationally
1) Establishing what they should mean
2) Establishing what they actually mean
What does a data collection pro-forma do?
Ensures that project researchers measure/record the variable consistently, accurately and with minimal missingness
What should a questionnaire be?
Series of clearly written questions that are unambiguous, definitive and instructive
Preparing data for analysis - What is in the rows and collumns
Rows = Cases/Participants Columns = Variables
Define the target population
The total, finite population of people/context
Define the study sample
The people/contexts from within the target population for which data is collected
What is a complete sample - give pro and con
The entire study population
Pro - No bias
Con - Resource intense/expensive
What is an unstratified/random sample - give pro and con
Every member of the target population have the same chance of being sampled
Pro - Easy to design and conduct
Con - Small groups may be misrepresented by chance
What is a stratified random/probability sample - give pro and con
Random sample from the target pop within each stata - every member within each strata has the same chance of being sampled and the number of each stratum can be oversampled to strengthen analysis
Pro - Representative and improved power for rare strata
Con - Population may not be be easily divisible into strata
Define the Odds Ratio
Represents the odds that an outcome will occur given a particular exposure, compared to the odds that the outcome will occur in the absence of that exposure
Define the confidence interaval
A range of values so defined that there is a specified probability that the parameter lies within it.
What does a 95% CI give us?
2 values between which the true value lies 95% of the time
What does it mean if the CI round an odds ratio does not include 1?
We are confident the effect is genuine
How to calculate an Odds Ratio
Look at notes
