RESS 3 Flashcards
what is a case-control study
selects participants on the basis of their outcome and works back to their exposure
what is a cohort study
longitudinal study following. group of people over time, recording subsequent events
what is randomised controlled trial
participants randomised to treatment groups - minimises bias
what is meta-analysis
combines the quantitive results from multiple studies
define research
generates new knowledge where there is no / limited research evidence available, and which has the potential to be generalisable and transferable
define audit
a quality improvement process which seeks to improve patient care and outcome through systematic review of care against explicit criteria and the implementation of change
define clinical audit
A study exploring whether healthcare practice achieves a particular standard/guideline
clinical audit vs service evaluation
audit: does this service reach a certain standard?
service evaluation: what standard might this service achieve?
define audit cum service evaluations
we can interested in whether variation in patient characteristics or the characteristics of healthcare delivery might influence whether a patient receives standard/recommended care
define ethics
what is morally correct
define governance
what has received permission
define wellbeing
studies improve burdens and risks - we can’t to not let the risks outweigh the benefits and protect the patient’s wellbeing
define rights
studies require informed consent - participation should be informed, voluntary and reversible
the role of ethics and governance approval
Ethics and governance procedures aim to ensure that no human (and many animal-based) studies take place without ensuring appropriate measures are in place
Ethics approval is particularly important/useful when the study’s participants are unable to consent
Ethical review ensures that the burdens and risks have been assessed, and that the consent proposed is considered appropriate to protect the rights of participants
what is engagement
where information about research is provided and disseminated
what is participation
where people take part in a research study
what is involvement
where members of the public are actively engaged in shaping research projects
- ‘Research carried out ‘with’ and ‘by’ members of the public rather than ‘to’ ‘about’ or ‘for’ them.’
examples of engagement
- social media
- newsletters
- websites and blogs
why do PPIE (patient and public involvement / engagement) research?
- ensures a patient focussed outcome
- involving patients helps together funding
- benefits to participants
- lets the patient have a voice
- patients offer a different yet important viewpoint
- sense of community to patients who share experiences
how and when to involve people
- research design
- development of the grant application
- undertaking/management of research
- analysis of data
- dissemination of research findings
role of NICE
- To reduce variation in the availability and quality of treatments and care
- Help resolve uncertainty about which medicines and treatments work best and which represent best value for money for the NHS
- Set national standards on how people with certain conditions should be treated
NICE’s core principles
- scientific rigour
- inclusiveness
- transparency
- independence
- challenge
- review
- support for implementation
- timeliness
(SITICRST) Smart Irresponsible Thoughtful Islands Control Retro Swollen Twigs
what are NICE clinical guidelines
- broad guidance covering management of a particular condition (the care pathway)
- considers clinical and cost-effectiveness and patient/care perspective
- recommendations are advisory, not mandatory
what are evidence based clinical NICE guidelines
a comprehensive set of recommendations for a particular disease or condition
what are guideline based quality standards - NICE
a prioritised, concise set of statements (usually 6-8) with associated measurable indicators, chosen and adapted from the clinical guideline recommendations
- benchmarks of best practice
what are covariates
what causes the outcome?
what causes the exposure?
what are potentially caused by the exposure?
what are cofounders
cause both the outcome and the exposure
why do we need to adjust cofounders?
They created pseudo-causal pathways between the outcome and the exposure, which will generate a statistical relationship between the two, even when none exists
what are mediators?
they cause the outcome and are caused by the exposure
do we adjust for mediators?
no - they are part of the casual path between the outcome and the exposure
what are competing exposures
cause the outcome, but have no relationship with the exposure
do we adjust for competing exposures?
yes because they cause a substantial amount of variation in the exposure - adjusting this out can make any association between the remainder and the exposure easier to detect
primary data
data which has been generated by the researcher e.g., surveys, interviews, experiments
secondary data
data that was formerly collected for other purposes than that of the current research study
define exposure
modifiable variation in clinical practice
defie outcome
clinical standard / guideline
3 steps of data collection
- decide which variables you need to measure
- consider where to source your variables
- choose between different sources of the same variable
prospective studies
record/measure variables during the study period, with the outcome subsequently measured
examples of prospective studies
cohort studies
cross-sectional studies
retrospective studies
record/measure the outcome and then look back to record/measure the exposure and co-variants
examples of retrospective studies
case-control studies
cross-sectional studies
pros of prospective data
- more accurate data regarding exposures, confounders, outcomes and covariates
- less bias
cons of prospective data
- time and resource intensive
- expensive
- usually infeasible for rare outcomes
pros of retrospective studies
- less time and resource intensive
- allows oversampling for rare outcomes
cons for retrospective studies
- more susceptible to bias
- existing data may be of poor quality
- if data from records, little control over how these are measured
importance of recognising sources of measurement errors
helps to…
- minimise errors and improve accuracy / validity when measuring variables
- select the most accurate variable available
what is missingness
the data value that is not stored for a variable in the observation of interest
how does missingness affect data
it reduces the number of participants on whom you have complete data
define sample
comprises data drawl from a specified context or population
what is a target population
the total, finale population of people/contexts
what is simple random sampling
When a sample of size n from a population of size N is drawn so that every possible sampling unit has an equal chance of selection
what is stratified sampling
- population is divided into homogenous strata according to a demographic factor
- randomly sample from the target population within the strata
what is opportunistic sampling
taking the most accessible or willing persons - e.g., people walking sown the street
what is the multivariable regression model
helps to indicate if there is an association between an outcome and exposure adjusting for cofounders
what are nuisance variables
can undermine the interpretation of the associations, but we can adjust for these to eliminate or reduce their influence
examples of nuisance variables
confounders and competing exposures
odds ratio interpretation: OR = 1
The odds in exposed group is the same as the odds in unexposed group
odds ratio interpretation: OR < 1
The odds in exposed group is less than the odds in unexposed group
Odds Ratios (OR) > 1
The odds in exposed group is higher than the odds in unexposed group
misspecification
incorrectly specified multivariable regression model
model misspecification in terms of non-linear relationships
The relationship between the outcome variable and exposure variable maybe non-linear and we fit a linear regression line
model misspecification in terms of omitted variable bias
Occurs when relevant variables e.g., confounders are left out in a multivariable regression model resulting in confounding bias
- associations are made without adjustment for confounders
- confounders may generate sprier associations
model misspecification in terms of omitting competing exposures
competing exposures may cause the outcome, but have no relationship with the exposure
- we must therefore adjust for these as they cause a substantial amount of variation in the exposure and so adjusting out can make any association between the remainder and the exposure easier to detect
model misspecification in terms of adjustment bias
controlling for an intermediate variable or mediator in multivariable regression analysis
- we can investigate the total effect of exposure on outcome using regression analysis without adjusting for the mediator
what is multi-collinearity / collinearity
occurs when there is strong associations or correlations among covariates in a multivariable regression model
what is heterogeneity bias
occurs when there are natural group structures in the data and there are differences in the groups which are correlated with the study variable
- overall regression analysis shows a positive association but in stratified analyses, there is a negative association
how to categorise covariates
convert continuous data to two groups (dichotomising - binary split at the median)
- leads to a comparison of groups of individuals with high or low values of the measurement
advantages of dichotomising
analyses or presentation of results will be simplified
disadvantages of dichotomising
- information loss
- conceals any non-linearity in the relation between the exposure and outcome
