All Flashcards
(269 cards)
1
Q
Best research is
A
○ Clinically relevant
○ Illuminates the accuracy of diagnostic tests
○ Highlight the importance of prognostic markers
○ Establishes efficacy and safety strategies
○ Seeks to understand patient experience
2
Q
EBM
- Three major tenets
A
○ An increasingly sophisticated hierarchy of evidence
○ The need for systematic summaries of the best evidence to guide care
○ The requirement for considering patient values in important clinical decisions
3
Q
Omega 3
A
- Cochrane review challenged validity of claim that omega 3 supplements reduce risk of heart disease, stroke or death
- Systematic review found little to no evidence to support the claim that there is an effect
4
Q
5 A’s
A
- Assess - A questions arises during the care of a patient
- Ask - Completely articulate all the parts of the question
- Acquire - Conduct a thorough focused search and select the highest quality evidence
- Appraise - Evaluate the evidence for validity and clinical applicability
- Apply - Use clinical expertise to integrate applicable evidence with attention to patient values
5
Q
therapy design
A
RCT> cohort>case control>case series
6
Q
prevention design
A
RCT> cohort study > case control > case series
7
Q
diagnosis or screening design
A
Prospective, blind comparison to a gold standard
8
Q
prognosis
A
Cohort study > case control > case series
9
Q
aetiology
A
RCT> cohort > case control > case series
10
Q
meaning
A
qualitative studies
11
Q
One sided
A
Causing an increased OR decreased risk (not either)
12
Q
Two sided
A
Related, but not necessarily in one direction or the other
13
Q
- Experimental study
A
○ Randomised control trial
○ Testing something eg. Treatment
14
Q
- Observational study
A
○ Descriptive § No comparison ○ Analytical § Comparison groups § Cohort § Case control § Cross sectional
15
Q
cons of RCT
A
§ While it has internal validity, it may not be externally valid (this means while it will do what it sets out to do, the outcomes may not hold up outside of the study group
§ The is due to prescribed inclusion and exclusion criteria for study volunteers
§ Sometimes you cant just do RCT - ethical concerns
§ Can get expensive
16
Q
passive or active placebos
A
§ Can be passive or active
Passive placebo - mimics administration only, active placebo - mimics side effects of intervention
17
Q
cohort studies used for
A
○ Determining risk factors/predictors of risk
○ Aetiology - what causes outcomes
○ Prognosis - what happens in this disease over time
○ Diagnosis - if text x is positive, what happens to the patient
18
Q
pros of cohort studies
A
○ Can determine incidence, relative risks and confidence intervals
○ Usually a clear view of exposure-outcome timeline (not affected by chicken-egg question, as are some other trial designs
○ Can investigate multiple outcomes
○ Useful to study rare exposures
19
Q
cons of cohort studies
A
Cons
○ Selection bias - often hard to match exposure and non-exposure group exactly
○ Not a great design for looking at rare diseases
○ Loss to follow up can also lead to bias
○ Exposure status might change over time
20
Q
case control studies uses
A
- What questions can. They answer
○ Risks
○ Prognosis
○ Diagnosis
21
Q
cross sectional studies used for
A
○ Frequency - how common is the outcome
○ Aetiology - what risk factors are associated with the outcome
○ Diagnosis - does the new test perform as well as the current standard test
22
Q
cons cross sectional studies
A
○ Cant determine relationship between outcome and exposure
23
Q
Case study or case series study
A
- A study describing a patients or small group of patients affected by a specific disease or exposed to a factor (eg. An intervention)
24
Q
common design for prevalence
A
cross sectional
25
common design for incidence
cohort
26
common design for prognosis
cohort
27
common design for treatment effect
controlled trial
28
Tigecycline
Tigecycline: is it a viable alternative antibiotic for resistant and complicated infection?
○ On the basis of non inferiority trial (clinical trials comparing a drug to another drug), FDA approval in 2005
§ 4 separate trials, each showed similar cure rates when compared to a particular treatment regimen
§ In the serious adverse events associated with tigecycline use, it was noted that there were more deaths in the tigecycline-treated patients
□ This was not a statistically significant difference
□ No specific safety or efficacy findings could be found to account for this difference
○ In 2010, the FDA warned in a safety communication that tigecycline was associated with an increased risk of death
29
Randomised trial of estrogen plus progestogen
Randomised trial of estrogen plus progestogen for secondary prevention of coronary heart disease in post menopausal women
- Conclusions of observational studies
○ Hormone therapy should probably be recommended for women who have had a hysterectomy and for those with coronary heart disease or at high risk for coronary heart disease
○ Although the findings from the epidemiologic studies are not completely consistent, the preponderance of the evidence strongly suggests that women taking postmenopausal estrogen therapy are at decreased risk for CHD
- Conclusion of RCT
○ No evidence to suggest this is effective
30
Trialists vs epidemiologists
- Trialists most often refer to the Cochrane tool for assessing risk in RCTs, selection bias, performance bias, detection bias, attrition bias, reporting bias
- Epidemiologists tend to use terms: confounding, selection bias, information bias
31
trialists terms
- Trialists most often refer to the Cochrane tool for assessing risk in RCTs, selection bias, performance bias, detection bias, attrition bias, reporting bias
32
epidemiologists terms
- Epidemiologists tend to use terms: confounding, selection bias, information bias
33
Confounding
- The objective of an experiment could be 'does A cause B'
| - A confounder is anything that could cause a change in B, but is not A
34
○ Restriction
§ Exclusion or specification
§ The simplest way of avoiding a confounder, participants are screened and excluded from the study if known to possess the confounder
§ Eg. If smoking is believed to be a confounding factor, the study enrols only non-smokers
35
cons of restriction
□ May slow recruitment
□ Precludes study of people who have that particular confounder (ie. Smokers) - this might exclude a significant proportion of the population
□ Will increase internal validity, but likely to result in poorer external validity
36
matching
§ Involves 'matching' the confounder with a participant in the control group
□ Eg. If smoking is a confounder, when enrolling a smoker in the case group, the control group must also enrol a smoker
§ Frequently used for sex and age
37
matching cons
□ If there are several confounder, this can become onerous
| □ By doing this, by definition the study can no longer measure the effect of a matched variable
38
stratification
§ Applied after study completion
§ Essentially uses a restriction approach during analysis
□ Eg. If smoking is the confounder, the then population is stratified into smokers and non-smokers, and the analysis conducted separately for the two groups
§ Comparisons can then be made between the two groups to confirm the effect of the confounder
§ This strategy facilitates recruitment and enables assessment of the influence of the confounder
39
Bias
- Is any systematic error is collecting and interpreting data
- They can result in both the under and over estimation of a study effect
- Bias undermines the internal validity of a study
40
- Selection bias
○ Selection bias occurs when there are systematic group difference when might affect the outcome
41
Performance bias
○ Systematic differences between groups in the care that has been provided or exposure to other factors other than the interventions of interest
42
- Detection bias
○ Systematic differences between groups in how outcomes are determined/measured
When evaluating the effect
43
- Attrition bias
○ Is bias that arises from systematic differences in the ways participants are lost from a study
44
attrition values
○ Attrition of 5% or less is unlikely to involve significant bias
○ However, attrition of >20% would indicate attrition bias
45
how to avoid attrition
§ Over recruitment
| § When there is knowledge of the bias, post hoc analysis strategies can be used eg. Sample weighting
46
the funnel plot
○ Is a scatter plot of the effect estimates from individual studies against some measure of each study's size or precision
○ A triangle centred on a fixed effect summary estimate and extending ~2 standard errors either side will include about 95% of studies if no bias is present and the fixed effect assumption (that the true treatment effect is the same in each study) is valid
47
- What can systematic reviews synthesise
○ Quantitative studies
§ Meta analysis
§ Or narrative analysis
○ Qualitative studies
§ Aggregate or integrate findings from primary qualitative research studies into these or metaphors
○ Quantitative and qualitative studies
§ Mixed methods review
48
- Locating systematic reviews
○ DARE: database of abstracts of reviews of effects
○ Specialist evidence databases
○ Discipline-specific databases
○ Large biomedical databases
§ Most have a systematic revies filter
§ PubMed - systematic reviews filter in clinical queries
§ Medline
§ Embase
49
- How is a systematic review conducted
```
§ Formulate the question
§ Define a eligibility criteria for the systematic review
Conduct the search for studies
§ Screen studies for inclusion
§ Review full texts of eligible studies
§ Asses the risk of bias
§ Extract the data
§ Meta analysis - if being done
```
50
§ Screen studies for inclusion
□ Undertaken by more than or equal to 2 authors to avoid bias
□ Titles and abstracts are screening for relevance and eligibility
51
a forest plot
The length of that horizontal line represents the length of the confidence interval (CI).
} The square is called the point estimate - the result of a particular study
} The size of the point estimate echoes the length of the confidence interval
} The diamond is called the summary estimate. It represents the summary of the weighted results from the combined studies
– The left and right tips of the diamond are the two ends of the confidence interval. With each study that gets added to the plot, those tips will get closer together, and it will move left or right if a study's result tips the scales in one direction
} The vertical line in the centre is the line of no effect -if a result touches or crosses it, then the result is not statistically significant
52
point estaimate on a forest plot
} The square is called the point estimate - the result of a particular study
– Often, it's sized according to how much weight the study has in the meta analysis. The bigger it is, the more confident we can be about the result
53
summary estimate
} The diamond is called the summary estimate. It represents the summary of the weighted results from the combined studies
– The left and right tips of the diamond are the two ends of the confidence interval. With each study that gets added to the plot, those tips will get closer together, and it will move left or right if a study's result tips the scales in one direction
54
line of no effect
} The vertical line in the centre is the line of no effect -if a result touches or crosses it, then the result is not statistically significant
55
Critical appraisal of quantitative systematic reviews
○ Various appraisal checklists can be used to guide the process eg. CASP, AMSTAR
56
○ Reporting statement
§ For quantitative reviews
□ Preferred reporting items for systematic reviews and meta analysis (PRISMA statement)
□ Includes a 27 step checklist for authors, to ensure all important information if required
□ Many journals insist on compliance with the PRISMA statement in order to accept articles for publication
§ For qualitative reviews
□ Enhancing transparency in reporting the synthesis of qualitative research (ENTREQ statement)
□ Similarly includes a 21 step checklist
57
PRISMA
□ Preferred reporting items for systematic reviews and meta analysis (PRISMA statement)
□ Includes a 27 step checklist for authors, to ensure all important information if required
□ Many journals insist on compliance with the PRISMA statement in order to accept articles for publication
58
Basic concepts of medical ethics
1. Respect for autonomy of the patient/subject: autonomy refers to the capacity to think, decide and act on one's own free initiative
2. Beneficence: promoting what is best for the patient
3. Non-maleficence: do no harm
4. Justice
59
The Nuremburg Code (1946)
- A set of research ethics principles for human experimentation set as a result of the subsequent Nuremberg trials at the end of the second world war
- Informed consent and absence of coercion; properly formulated scientific experimentation; and beneficence towards experiment participants
60
principals of the Nuremberg code were not universally followed
○ Eugenics boards 1970's; 31 US states routinely sterilized people ('feeble minded', young rape victims) and conducted medical experiments on humans without informed consent
○ 7500 people sterilized in north Carolina
61
Declaration of Helsinki
- The world medical association drew up the first version of the DoH that broadened the concerns of the Nuremberg code
- Additions include
○ Right to privacy
○ Confidentiality
○ Privacy regarding the use of identifiable human information
○ Seven revisions, with the last in 2013
62
declarartion of helsinki was the broaden the concerns of
the Nuremberg code
63
additions in the declaration of helsinki
○ Right to privacy
○ Confidentiality
○ Privacy regarding the use of identifiable human information
○ Seven revisions, with the last in 2013
64
Henrietta lacks
- African American, 1st child at 14 years old, dies aged 31
- Attended johns hopkins hospital, baltimore (29/01/1951) diagnosed with carcinoma of the cervix and treated with radium
- 2 samples were taken without her permission or knowledge; healthy tissue and cancerous cells
- Cells were cultured by george Otto Gey to create the first human immortal cells line HeLa cells
65
HeLa cells
- Used by Salk to develop the polio vaccine
- Research on cancer, AIDS, the effects of radiation and toxic substances, gene mapping, used to test human sensitivity to tape, glue, cosmetics, and many other products
66
hela cells and medical ethics
- Family medical records published without family consent (1980s)
- Supreme court of California case of Moore v regents of the university of California in 1990
○ A persons discarded tissue and cells are not their property and can be commercialised
- 2013, German researchers published the HeLa genome, without permission from the lacks family
- 2013, NIH agreed to give the family some control over access to the data and a promise of acknowledgment in scientific papers
- Two family members will join a six member committee to regulate access to the genomic data
67
Tuskegee Syphilis Experiment (1932-1972)
○ Initially not inherently unethical in 1932 - no cure available
○ Involved 6-8 months observation and then contemporary treatments such as Salvarsan (organoarsenic), mercurial ointments, and bismuth (possibly mildly effective, but very toxic
○ Men were actively prevented from accessing treatment and exiting the study
68
Office for human research protections (OHRP) and the Belmont report
- Commissioned as a result of the Tuskegee study
○ USA OHPS issued the Belmont report
§ Federal laws and regulations regarding human subjects
§ Requirement for Institutional review Boards (IRB) for the protection of human subjects in studies
○ OHRP manages this responsibility within the US department of health and human services HHS
69
The Belmont report
○ The boundaries between biomedical and behavioural research and the accepted practise of medicine
○ The role of assessment of risk benefit criteria in the determination of the appropriateness of research involving human subjects
○ Guidelines for the selection of human subjects for participation in such research
○ The nature and definition of informed consent in various research settings
70
Fundamental principles for using human subjects
- Respect for persons
- Beneficence
- Justice
71
What is human research
- Data, samples or exhaled breath that can be identified as belonging to a living human
- Use of de identified cell lines, DNA, data etc. is not human research (national institutes of health, USA)
- Use of anonymised samples = human research
72
HRECs Australia
- Human research ethics committee provide ethical oversight of research involving humans (ie. In accordance with relevant standards and guidelines
- >200 HRECs in institutions within Australia
73
types of data and patient identifiability
Individually identifiable data,Re-identifiable data (anonymised), Non- de(identifiable)
74
Individually identifiable data,
where the identity of a specific individual can reasonably be ascertained. Examples of identifying include the individuals name, image, date of birth or address
75
Re-identifiable data (anonymised)
from which identifiers have been removed and replaced by a code, but it remains possible to re identify a specific individual by, for example, using the code or linking different data sets
76
Non- de(identifiable)
data that have never been labelled with the individual identifiers or from which identifiers have been permanently removed, and by means of which no specific individual can be identified. A subset of non-identifiable data are those than can be linked with other data so it can be known that they are about the same data subject, although the person's identity remains unknown
77
deep mind
- Google owned deep mind
| - An algorithm capable of predicting acute kidney injury (AKI) up to 48 hours before it occurs
78
Medical breakthroughs thanks to animal research
- Penicillin - 1940 - treated mice were saved form lethal streptococcus infection
- Blood transfusion - 1915 - citrate first demonstrated to permit transfusion of blood between dogs
- Tuberculosis - 1943 - streptomycin treatment of TB infected guinea pigs showed full suppression of disease
- Macular degeneration - 1988 development of surgical technique
- Asthma
- Meningitis
- Kidney transplants
- Breast cancer
- Parkinson's disease
- Diabetes
79
The Three Rs
- Reduce, replace, refine
- 1959 - Russel and birch first presented this concept in their book 'the principle of human experimental technique'
- These have become widely accepted ethical principals, now embedded in scientific practise virtually worldwide
80
Reduction
○ Minimising the number of animals required to demonstrate a specific outcome
81
Replace
○ Methods that permit a given purpose of an activity or project to be achieved without the use of animals
82
refine
○ Methods that alleviate or minimise potential pain and distress, and enhance animal wellbeing
83
reduction can be achieved by
§ Improving experimental technique
§ Improving data analysis (statistical methods)
§ Information sharing/ensuring the study is novel by evaluating the available literature
§ Minimising variable that might influence experimental outcomes (stress, diet, overcrowding, sex, genetics)
§ Ensuring the species is appropriate
§ Replacement with non-animal models where possible
84
replacement c an be achieved by
§ Replacing higher organisms with lower organisms (eg. Rodents with microorganisms, plants, invertebrates etc.)
§ Substitution with dummies, mechanical or computer models, or in vitro cell cultures
§ Studying human volunteers
§ Consulting on epidemiological studies
85
refinement can be achieved by
§ Identifying sources of pain or stress and removing these
§ Ensuring all research staff are adequately trained to ensure proper animal handling and that procedures are performed using best standards of practice
§ Refining the experiment so methods are less invasive and provide the best levels of acre (eg. Anaesthesia and/or analgesia are required)
Provide best practice animal husbandry to encourage the natural behaviours of animals
86
- Physical measures
○ Measures of bodily activity eg. Pulse or blood pressure
○ Physical measures are much more objective than behavioural measures
○ A physical measure is not simply an observation (which may be subjective)of how a person is acting
87
disadvantages of physical measures
○ Disadvantages: keep in mind that humans are still responsible for running the equipment that takes the measures and ultimately for interpreting the data provided by the measuring instrument, may impact on accuracy of the data
88
advantages of surveys
- Advantages
○ Anonymity
○ Asking sensitive questions
○ Standardised data to monitor trends
○ Quick to administer
○ Simple data analysis can be used (% frequencies)
○ Data can be presented on graphs and charts
○ Simple distribution channels (online, paper, social networking sites)
89
disadvantages of surveys
- Disadvantages
○ Limited explanations of the data
○ Comments open to a range of interpretations
○ Don’t know why questions were skipped or why survey was partially completed
○ No opportunity to explain questions that people don’t understand
○ Cant guarantee the survey was even received or completed by the right person
○ Basic literacy required
90
Reliability
an indication of the consistency of stability of a measuring instrument
91
Validity:
an indication of whether the instrument measures what is claims to measure
92
Rorschach test
- Ink blot test
| - Not reliably measuring what it claims to measure
93
types of reliability
```
○ Test/retest reliability
○ Alternate-forms reliability
○ Split-half reliability
○ Inter-rater reliability
○ Inter-item reliability
```
94
○ Test/retest reliability
§ A reliability coefficient determined by assessing the degree of relationship between scores on the same test administered on two different occasions
§ If the test is reliable, we expect the two scores for each individual to be similar, and thus the resulting correlation coefficient will be high (close to 1)
95
○ Alternate-forms reliability
§ A reliability coefficient determined by assessing the degree of relationship between scores on two equivalent tests
tests need to be truly parallel eg. Equal difficulty, same instructions
96
○ Split-half reliability
§ A reliability coefficient determined by correlating scores on one half of a measure with scores on the other half of the measure
97
○ Inter-rater reliability
§ A reliability coefficient that assesses the agreement of observations made by two or more raters or judges
98
○ Inter-item reliability
§ It is used to determine how much the items (or questions) on a scale are measuring the same underlying dimension
99
acceptable level of test/retest reliability
0.8
100
Practise effects
types of test/retest reliability
| some people get better at the second testing, and this practise lowers the observed correlation
101
Short time between testing
type of test/retest reliability
- individuals may remember how they answered previously, then we may be testing their memories and not the reliability of the testing instrument, and we may observe a spuriously high correlation
102
The kappa statistic
§ The kappa statistic is frequently used to test interrater reliability
§ Problem with kappa:
□ Judgements about what level of kappa should be acceptable for health research are questioned
103
Cronbach's alpha
§ Cronbach's alpha is a common measure of internal consistency (a measure of reliability)
104
types of validity
○ Content validity
○ Face validity
○ Criterion validity
○ Construct validity
105
○ Content validity
§ The extent to which a measuring instrument covers a representative sample of the domain of behaviours to be measured
106
○ Face validity
§ The extent to which a measuring instrument appears valid on its surface
107
○ Criterion validity
§ The extent to which a measuring instrument accurately predicts behaviour or ability in a given area
108
○ Construct validity
§ The degree to which a measuring instrument accurately measures a theoretical construct or trait that is designed to measure
§ If a measure is based on a full and close examination of the underlying concept, along with the related theoretical approaches, then it is said to have high construct validity
109
types of criterion validity
§ Concurrent
□ Used for diagnosis of existing status e.g. hoe well a new test compares to a well established test. You would give both tests and see if the results match
§ Predictive
□ The tests can be used with some degree of accuracy to predict future behaviour e.g. year 12 marks have predictive validity because performance on the tests correlates with later performance in university
110
concurrent validity
□ Used for diagnosis of existing status e.g. hoe well a new test compares to a well established test. You would give both tests and see if the results match
111
predictive validity
□ The tests can be used with some degree of accuracy to predict future behaviour e.g. year 12 marks have predictive validity because performance on the tests correlates with later performance in university
112
types of construct validity
§ Convergent validity
□ If it is related to things which we expect the concept we are trying to measure to be related
§ Divergent validity
□ Independent of those things of which the concept should be independent
113
convergent validity
□ If it is related to things which we expect the concept we are trying to measure to be related
114
§ Divergent validity
□ Independent of those things of which the concept should be independent
115
The relationship between reliability and validity
- If valid, will be reliable
116
Sensitively
- The proportion of patients with the disease who also test positive for the disease in this study. It is the probability that a person with the disease will have a positive test result
117
Specificity
- The proportion of patients without the disease who also test negative for the disease in this study. It is the probability that a person without the disease will have a negative test result
118
Predictive value
- PPV - expresses the proportion of those with positive test results who truly have disease
- Another way of saying this is, given that a patient tests positive, what is the probability that they have disease
- Positive predictive value (PPV)
○ If you have a positive test, what is the likelihood you actually have the disease
- Negative predictive value (NPV)
○ If I have a negative test, what is the likelihood I do not have disease X
119
- Positive predictive value (PPV)
○ If you have a positive test, what is the likelihood you actually have the disease
120
- Negative predictive value (NPV)
○ If I have a negative test, what is the likelihood I do not have disease X
121
Effect of prevalence on predictive value
- Sensitivity and specificity are not effected by the prevalence
- PPV/NPV is dependant - less helpful
122
Likelihood ratios (LR)
- LRs express how many times more or less that a test result is to be found in diseased, compared with non diseased people
- Likelihood ratios are independent of disease prevalence
123
PPV formula
prob of + in those with disease/prob of + in those without disease
124
NPV formula
prob of - in those with disease/prob of - test in those without disease
125
positive LR formula
(a/(a+c))/(b/(b+d))
126
negative LR formula
(c/(a+c))/(d/(b+d))
127
- The interpretation of likelihood ratios
○ The larger the positive likelihood ratio, the greater the likelihood of the disease
○ The smaller the negative likelihood ratio, the lesser the likelihood of the disease
○ Diagnostic tests with positive LRs >10 and /or negative LRs<0.1 can be thought of as fairly powerful tests
128
LR values
○ Diagnostic tests with positive LRs >10 and /or negative LRs<0.1 can be thought of as fairly powerful tests
129
Interquartile range
- Difference between the 25th and the 75th percentile values
- Can be useful to report with median
- n = number of people in the sample
- Median = (n+1)/2 th value
130
standard error
std deviation/sqrt n
131
Type 1 error
- The probability of rejecting the null hypothesis when it is actually true
132
Type 2 error
- the probability of failing to reject the null hypothesis when it is actually false
- Happens with small samples
133
power values
- Should aim for a power of 0.8 - 80% chance of finding the population effect
134
power is determined by
- Alpha (a) - usually set to be 0.05 - the probability of type 1 error
135
Power can be increased by
- Increasing the alpha level of you statistical tests.
- Minimising the number of dependant variables.
- Reduce error
- Increasing the size of the treatment effect
- Recruit the correct number of participants for the statistical test you are planning
136
increasing alpha level
○ Increasing the a-level for your statistical test is not a recommended strategy for increasing the power of your study
137
Reducing error increases
effect size
138
If your sample size is too large
the study will be overpowered and you'll run the risk of capturing trivial population effects of no practical or clinical importance
139
- Failure to find a statistically significant effect in sample data is consisted with either of two interpretations
○ The effect does not exist in the underlying population, and the sample data merely confirm this
○ The effect does exist in the underlying population, but the study has insufficient power to detect it
140
Point prevalence
§ Occurrence of disease at a specific point in time
141
○ Period prevalence
§ Occurrence of disease with a defined period of time
142
Incidence
- New cases is a disease that occur during a specified period of time in a population at risk for developing the disease
143
most of the time cumulative incidence and incidence rate are equal unless
§ When they get the disease early and die off quicker
§ Very common - more people drop out of person time
□ Rare condition is more likely to approximate
144
Arithmetic comparisons
- Measures of association are mathematical comparisons
| - Mathematic comparisons can be done in absolute terms or relative terms
145
absolute effect measures
reflects the excess deaths and 'body count'
146
relative effect measures
reflect relative strength of association
147
risk difference
R1-R0
148
relative risk
R1-R0
149
odds ratio
- When doing a case control study
- Approximates a relative risk
ad/bc
150
- Scale variables
○ Interval variables have central characteristic that can be measured along a continuum and they have a numerical value eg. temperature - Celsius
151
- Categorical
○ Nominal variables have two or more categories, but which do not have an intrinsic order e.g. houses, condos, bungalows
152
Ordinal variables
have two or more categories which can be ordered or ranked e.g. not very much, ok, a lot
153
Independent sample t-test
- Two independent categorical groups
- A continuous outcome
- Observations are independent
- Outcome is approximately normally distributed
- No massive outliers
154
Paired t-test
- Two related group (twin, before/after, etc.)
- A continuous outcome eg. Age,
- Observations are independent within groups
- Outcome is approximately normally distributed
- No massive outliers
155
Mann-Whitney U test
- Non parametric test - doesn’t rely on normality
- Two independent categorical groups
- A continuous or ordinal outcome
- Observations are independent
156
Wilcoxon signed-rank test
- Two related groups (twins, before/after etc)
- A continuous or ordinal outcome
- Observations are independent within groups
- Does not assume normally distributed
157
Chi square test
- Used to assess the distribution of a categorical variable between to or more groups
- Non parametric test
- Assumes that expected cell frequency is at least 5, in each cell (if this assumption is not valid, we can use fisher's exact test
- The null hypothesis is that the distribution of observation between columns is independent of the rows
158
Pearson's correlation
○ Probably the most widely known correlation
○ Measures if X and Y are linearly related
○ Assumptions
§ Both variables should be continuous
§ Both variables should be normally distributed
Errors are normally distributed about the regression line
159
assumptions of pearsons correlation
○ Assumptions
§ Both variables should be continuous
§ Both variables should be normally distributed
§ Errors are normally distributed about the regression line
160
Kendall and spearman's
```
○ Both rank correlations
○ Do not assume that X and Y are linearly related
○ Often used for ordinal variables
○ Both nonparametric correlations
○ Spearman's probably more common
```
161
Simple linear regression
- Estimating from data points to describe the relationship between two variables
- R2 - measure as the percentage of variability in Y
○ Measures goodness of fit
○ Higher the better
○ Dependant variable = weight
○ Explanatory = height
162
R2
measure as the percentage of variability in Y
163
Survival analysis
- The Kaplan Meier methods is used to estimate the probability of survival past given time points
- The survival distributions of two or more groups can be compared for quality
○ Eg. Measuring the time until death of groups of rats with cancer taking different drugs
○ Time to failure for a knee replacement
164
The Kaplan Meier methods
used to estimate the probability of survival past given time points
165
H0
no difference between populations in the probability of an event occurring at any point
166
H1
there is a difference
167
The effect of intervention
fallacy
- A statistically significant effect at the group level does not mean you should assume every person has benefited
- In order to determine the degree to which a particular person has been changed by outcome measure during the course of the intervention, researchers need to compute a reliable change index (RC) for that person
168
reliable change index (RC)
- The degree to which the person changes on the outcome variable divided by the standard error of difference between the pre-test and post test scores
- When the RC is greater than 1.96, it is likely that the post-test score is reflecting a real change rather that measurement error
169
reliable changes values
- When the RC is greater than 1.96, it is likely that the post-test score is reflecting a real change rather that measurement error
170
Restricted ranges
- Failure to sample across a sufficiently broad range of scores
- If you restrict the range the sample correlation becomes attenuated such that it underestimates true population value
171
To reach a conclusion that one variable causes another, three criteria must be met
- Covariation - the two variables must correlate
- Directionality - the presumed cause must precede the presumed effect
- Causal closure - all other variables that might influence the relationship between the two variables must be eliminated or controlled
172
Causal closure
all other variables that might influence the relationship between the two variables must be eliminated or controlled
173
Independence of observations
- Each score on the dependant variable must be independent of another source
174
Unequal groups
- Group equivalence ensures causal closure and is therefore essential for establishing a causal relationship between the independent variable and the dependant variable
175
- There are two methods for approximating group equivalence
○ Pairwise matching of participants on a confounding variables
○ Statistical matching of participants on the key confounding variables
176
Intra group dependency
- Intragroup dependency occurs whenever the data are collected from intact social groups
177
○ St john's wort
§ Herb supposed to reduce depression
178
○ Aspirin post heart attack
§ May reduce secondary heart attack
179
○ Lumpectomy vs mastectomy
§ Lumpectomy has same/slightly better survival
| § Lead to change in surgical practise
180
Philosophical assumptions underpinning qualitative research
○ Ontological
○ Epistemological
○ Axiological
○ Methodological
181
○ Epistemological
§ Dualism between researchers and research participants refuted: researcher and participants are interactively linked
§ Researcher try to get as close as possible to participants in the study
§ Research findings emerge from the interaction between the research participant and the researcher
182
○ Axiological
§ Facts and values are not separate; researchers and research being undertaken are inseparable
§ Researchers position themselves in a study
183
○ Methodological
§ Assuming multiple realities (ontology) and an epistemological imperative to be 'close' to participants - methodologically, research need to be undertaken using methods to facilitate this
§ Methodology must, therefore, account for dialogue required between investigator and participants
184
3 types of case study
§ Instrumental case study
§ Collective case study
§ Intrinsic case study
185
§ Instrumental case study
□ Researcher focuses on an issue or concern, and then selects one bounded case to illustrate this issue
□ Case is of secondary interest: plays a supportive role and facilitates our understanding of something else
186
§ Collective case study
□ Also called multiple case study - instrumental case study extended to several cases
□ One issue of concern selected, but inquirer selects multiple case studies to illustrate the issue
187
§ Intrinsic case study
□ Undertaken because, first and last, the researcher wants better understanding of this particular case: the case itself of interest
188
○ Grounded theory
□ To move beyond description and to generate or discover a theory, an abstract analytical schema of a process, or action or interaction
189
○ Phenomenology
to reduce individual experiences with a phenomenon to a description of the universal essence
190
○ Generic qualitative approach
□ Combine several approaches
| □ Claim no particular methodological viewpoint at all
191
○ Common sampling strategies
§ Maximum variation - people with different experience - breapth
§ Homogenous - people with similar experience, more depth
§ Stratified - select based on subgroups
§ Snowball - get people to refer there friends - need to have a broad initial sample
§ Triangulated - using multiple methods combined
§ NB: convenience sampling is not purposive sampling
192
Maximum variation
people with different experience - breapth
193
Homogenous
people with similar experience, more depth
194
Stratified
select based on subgroups
195
Snowbal
get people to refer there friends - need to have a broad initial sample
196
Triangulated
- using multiple methods combined
197
3 requirements for credibility
□ Prolonged engagement in the field
□ Triangulation
□ Member checking
198
□ Prolonged engagement in the field
® Extended presence in the field
® Sufficient time so you know the context in which you are researching
® Engagement and time in the field
® More likely to develop trust and report
® Better able to understand context
199
□ Triangulation
® Data investigator
◊ More than one researcher in research study
◊ Analysis done as a group
◊ Based on data not on personal background
® Methodological identification
◊ Collection of data from different data collection methods
® Different courses of data
◊ Patients, carers, health professionals
200
□ Member checking
® Taking back emerging findings to those participating in research for verification
® Participants need to realise that it is based on data from a range of experience not a singular experience
® Happens once you are into data analysis
201
3 types of triangulation
® Data investigator
® Methodological identification
® Different courses of data
202
® Data investigator
◊ More than one researcher in research study
◊ Analysis done as a group
◊ Based on data not on personal background
203
® Methodological identification
◊ Collection of data from different data collection methods
204
® Different courses of data
◊ Patients, carers, health professionals
205
○ Dependability 'transferability'
§ Refers to the stability and trackability of changes over time
§ A person interviewed one week apart may give different information
§ Data collection must track changes
§ Audit trail
□ Journal tracking decisions made across the research process
□ Analysis
□ Someone will a similar repour/disciplinary background could come to the same conclusion
§ Transferability
□ Sufficient information and context about the research and participants
□ Can these findings be transferred to a different context/setting
□ Eg. Rural/city
206
§ Audit trail
□ Journal tracking decisions made across the research process
□ Analysis
□ Someone will a similar repour/disciplinary background could come to the same conclusion
207
§ Transferability
□ Sufficient information and context about the research and participants
□ Can these findings be transferred to a different context/setting
□ Eg. Rural/city
208
○ Confirmability
§ Interpretations must be plausible
| § To independent researchers must agree about meanings
209
- Qualitative data analysis (thematic analysis)
```
§ Immersion in the data
§ Constant comparison approach
§ Open coding
§ Clustering of concepts - categories
§ Abstracting and linking categories
§ Refinement
§ Theoretical coding
```
210
§ Constant comparison approach
□ Segments of text such as words or phrases
□ Experiences related within an interview
□ Whole interviews
□ The data and knowledge
□ The data and other studies of similar issues
□ Aim - to develop external heterogeneity and internal homogeneity within categories of data
211
Open coding
□ Reading transcripts line by line
| □ Identify and code concepts (an idea) in the data - closely examining or interrogating it
212
§ Clustering of concepts - categories
□ Focussed coding
® Organising the concepts (into categories), examining and refining each category, determining how it relates to other categories
213
§ Abstracting and linking categories
□ Characteristics of categories
® A label/name
® A key idea about which the category is concerned
® Segments of data
® A set of criteria that demarcate the category in relation to other categories being used
214
§ Refinement
□ Purpose - to ensure that categories are fully developed conceptually and we know as much as we can about each category
® Intensive analysis focusing on one category at a time
® Identify within each category what is important to the study and examine the way It is related to other categories
215
§ Theoretical coding
□ Focussing on the main ideas at an abstract level (themes) and examining their interrelation
216
philosophical assumptions of mixed methods research
○ Purist stance
§ Paradigms are incompatible, paradigms relate to methods, therefore mixing is impossible
○ Dialectic stance
§ Multiple paradigms are possible and can be used, but honour and make explicit the paradigms
○ Design stance
§ Multiple paradigms can be used - relate them to your design
○ Single paradigm
§ There is a single paradigm for mixed methods research: pragmatism
217
Purist stance
§ Paradigms are incompatible, paradigms relate to methods, therefore mixing is impossible
218
○ Dialectic stance
Multiple paradigms are possible and can be used, but honour and make explicit the paradigms
219
○ Design stance
§ Multiple paradigms can be used - relate them to your design
220
○ Single paradigm
§ There is a single paradigm for mixed methods research: pragmatism
221
3 types of mixing
○ Merge the data
§ Collect qualitative and quantitative data and present both
○ Connect data
§ Collect qualitative and then quantitative
○ Embed the data
§ Large quantitative study with qualitative studies imbedded
222
triangulation/concurrent design
QUAL+QUAN
223
convergence model
merge results and interpret QUAL and QUAN
224
data transformation model
convert QUAL into QUAN
225
validating quantitative data model
validate QUAN with qual
226
multilevel model
QUAN and qual on levels
227
○ Exploratory design (sequential)
interpretation based on
QUAN->qual
or
QUAL->quan
228
embedded (nested) design
QUAN(qual)
or
QUAL(quan)
229
embedded experimental model
QUAN with qual before during and after intervention
230
Richard Peto
- smoking causing lung cancer
231
Ignaz Semmelweis
- Two clinics used almost the same techniques, and Semmelweis excluded other factors such as age, religious practises, climate differences, overcrowding, poverty level
- Maternal death caused by cadavers
232
Eradication of disease
- Edward Jenner (1776) and small pox
- Observed the protective effect of prior cow pox infection for small pox
- Paved the way for the concept of vaccination
- His theory was right and he saved millions of lives and led to the eradication of small pox
233
Goldberger and pellagra (1917)
- Goldberger figured it was as a result of diet and not infectious
- Deficiency of niacin
- Exposed volunteers including his wife to blood urine faeces and epidermal scales of pellagra lesions to prove it was not infectious
234
Breast cancer diagnosis
- ABCs Queensland headquarters in Brisbane
- 10 people diagnosed with breast cancer
- Younger women 18 times more likely to develop breast cancer
- Demolished the building
235
DVT on long haul flights
- Deep vein thrombosis
- After a three hour flight, 5 times more likely within 0-7 days
- More likely to die from DVT than airplane crash
236
Endophthalmitis in WA
- Infection of the eye after cataract surgery
- Nest case control study of 4 controls per case
- Antiseptic preparation and subconjunctival antibiotics
- Subconjunctival antibiotics were used 2% of the time, now used 96% of the time
237
Policy influences
- Public opinion, media, economic climate, legislative/policy infrastructure, political ideology and priorities, stakeholder interests, expert advice, resources, research
238
post project knowledge transmission
advocate, adviser
239
integrated knowledge transmission
commisioned work, partner, co-producer
240
What is effective
| - Little to no effect
○ Educational materials
| ○ Didactic sessions
241
- Sometimes effective
○ Audit feedback
○ Local opinion leaders
○ Local consensus project
○ Consumer mediated interventions
242
- Consistantly effective
○ Reminders
○ Interactive education (with discussion of practise)
○ Social marketing
○ Knowledge brokers
243
Translational medicine
The process of turning observations in the laboritory, clinic and community into interventions that improve the health of individuals and the public
244
T0
basic biomedical research, including preclinical and animal studies, not including interventions with human studies
- Fundamental bench research
- Uses models - animal or cultured models
- Examines samples - from patients and animals
- Innovations - technological advances
245
T1
translation to humans, including proof of concept studies, phase 1 clinical trials, and focus on new methods of diagnosis, treatment, and prevention in highly controlled settings
246
T2
translation to patients, including phase 2 and 3 clinical trials, and controlled studies leading to clinical application and evidence based guidelines
- Translation to patients
- Movement of discoveries into patients; early phase clinical testing
- Clinical research includes
○ Studies to better understand a disease in humans, and relate this to finding observed in animal or cell models of disease
○ Testing of interventions for safety and effectiveness in those with and without disease
○ Behavioural and observational studies
○ Testing new treatments for toxicity and efficacy on healthy people and ill people for whom no other treatment exists
- The goal of many clinical trials is to obtain data to support regulatory approval for an intervention
247
T3
translation to practise, including comparative effectiveness research, post marketing studies, clinical outcomes research, as well as health services, and dissemination and implementation research, phase 4 clinical trials
- Adoption of interventions that have been demonstrated to be useful in a clinical setting into routine clinical care for the general population
- This stage also includes the evaluation of the results of clinical trials and to identify new clinical questions and gaps in care
248
T4
translation to communities, including population level outcomes research, monitoring morbidity, mortality, benefits, and risks, and impacts of policy and change
- Public health implementation
249
T0-T4
```
T0 - basic biomedical research
T1 - translation to humans
T2 - translation to patients
T3 - translation to practise
T4 - translation to communities
```
250
- Collaborations and partnerships
○ Different skills requires
○ Made difficult by: compartmentalized environment of institutions, cultural divide between scientists and clinicians, university system that rewards individual achievement
251
- Data transparency and release - effective dissemination
○ Data needs to be made publicly available to spur innovation and discovery
252
- De-risking therapeutic development
○ There are risks associated with moving through each stage of the translational pathway
○ This risk is a significant barrier to the progression of the translational pathways
○ High-risk = low probability of funding success
253
- Predictive efficacy and toxicology
○ Uncertainty and lack of predictive efficacy is a significant barrier to translation
254
where to find evidence based summaries
up to date, clinical key, best practise, ovid, national heart foudation, astham australia
255
where to find synopsis of synthesis
ACP journal club, databse of abstracts of reveiws of effects (DARE)
256
where to find systematic reviews
cochrane, medline, pubmed
257
where to find synopsis of single studies
ACP journal club
258
where to find single stuides
medline, pubmed, embase
259
Medline - Ovid Medline
- Accessible through UWA library
- Subscription only
- PubMed contains Medline in it - PubMed is the NLM's portal for accessing Medline
- Ovid licences Medline from NLM
- Uses
○ Searching for key terms separately - applying boolean operators
○ Saving literature searches
○ High level filtering functions
260
Up to date
- Evidence based summaries and guidelines
- Tailored for clinicians
- Subscriber based
- Uses
○ For medical practitioners needing prescriptive info
○ For allied health personnel looking for specific summaries to provide to patients
261
point prevalence eq
- Everyone who has it/everyone (who's alive)
262
period prevalence eq
- Everyone who gets it or already had it/average population
263
cumulative incidence eq
- Everyone who gets it/everyone who's capable of getting it
264
Incidence rate eq
- Everyone who gets it/total life years
265
case fatality eq
- Everyone who dies/everyone who gets it or already had it
266
mortality rate eq
- Everyone who died/everyone
267
- Effect size calculated by
```
○ Cohen's d
○ Measures effect size
○ 0.2 - trivial, small
○ 0.5 - medium
○ 0.8 large
○ P tells you if it works, d tells you by how much
```
268
cohens d measures
effect size
269
cohens d values
○ 0.2 - trivial, small
○ 0.5 - medium
○ 0.8 large
