1. global determinants of health Flashcards

Question

evaluating evidence

Answer 1

--> Reference resource where experts pool together evidence for a particular topic * Peer-reviewed material * Conclusion statement * Position statement

Answer 2

Get a good question by looking at Popualtion – who Intervention - what Control group – comparison outcome Study design (RCT) Do the study How well was the study done What do the results mean Could they be due to chance

Answer 3

• Title – does it adequately describe the work, participants, (place), intervention/ exposure and outcome?

Answer 4

• Abstract – does it give a clear overview of the work, that helps answer your PICOS (other other question)?

Answer 5

• Introduction – remembering this is mostly the opinion of the authors. Only key point from scanning is does this go from the big picture to position the research question and justify why they did what they did?

Answer 6

• Methods – Key part, does it make sense, can it answer their question and does it appear transparent and justify the number of participants etc.

Answer 7

• Results – The other key part, these should be logically laid out, key information in a figure, secondary data in a table. It should only be what was found and no opinions

Answer 8

• Discussion – As with introduction, mostly authors opinion, can be useful for first paragraph which states what the authors think the research found, and then check the conclusion is not overstated and links to the primary outcome, methods and data

Answer 9

* Can I now read a paper or a review article and understand why this topic is being investigated? * Is this an observation or an experimental study (and is it the best approach to ask this question)? * How are the authors testing their hypotheses? * Has the trial been registered, did the primary outcomes change? Have the number of participants been justified * Do conclusions follow from the data presented in tables and figures? * Are the results generalisable?

Answer 10

observational | experimental

Answer 11

* Establish links/associations * Investigator takes measures but does not intervene. * Retrospective or prospective

Answer 12

• E.g. cross-sectional studies/surveys, case-control studies, cohort studies Longitudinal period of time – suggest causality

Answer 13

* Establish cause & effect | * An intervention is performed by the researcher, all else held constant

Answer 14

• E.g. intervention trials, randomised controlled trials (RCTs)reduce confounding and bias

Answer 15

the explanatory or predictor variable – the variable hypothesised to explain the outcome variable. The cause in cause and effect

Answer 16

the variable that is observed as the effect of manipulation of the independent variable. The variable we expect to change

Answer 17

---> can't change, residual things in population e.g. amount of fat people eat and amount of energy, as people eat more fat = more energy Gender Age Education Control for confounding variable • Occurs when it is not possible to disentangle the effects of two or more variables or events

Answer 18

• Accuracy Closeness to the true value, also called specificity – it measures what is supposed to measure

Answer 19

• Precision Repeatability, also called reliability – gives same result over repeated trials

Answer 20

The likelihood that a test will turn out positive when the result is positive Example: % of people with clinical depression who score as “depressed” on a scale measuring depression

Answer 21

The likelihood that a test will turn out negative when the result is negative Example: % of people who do not have clinical depression who score as “not depressed” on a scale measuring depression

Answer 22

Can sacrifice a high sensitivity But don't sacrifice a high specifity as that means that you get false negatives

Answer 23

Distortion, due to: a. Methods used to select subjects b. Factors that influence study participation Therefore – findings represent factors determining disease AND factors determining participation

Answer 24

---> make sure there are enough people in the study

Answer 25

The statistical approach to determining the sample size is a “power calculation” or “power analysis” Typically involves accepting no more than 5 % level of error, i.e. event occurring by chance * Sample size calculations are based on * Level of error you will accept * Magnitude of the effect under investigation * Prevalence of factor you want to measure * Relative sizes of the groups being compared

Answer 26

A statistical value indicating the probability that your findings are due to a real effect, and not likely to be by chance

Answer 27

A “95% C.I. (A, B)” means that if we were to repeat our experiment a very large number of times, in 95% of cases, our result would fall always between A and B. The difference with the p-value is that the C.I. gives you a range or limits, of the lowest and highest values your result can possibly take.

Answer 28

A measure of how strong an effect is. • The chance (“odds”) of an event occurring in the “control” group vs. the same event occurring in the “cases group”.

Answer 29

OR = 1 means no difference between groups OR > 1 the event is more likely in the cases = “induces” disease OR < 1 the event is more likely in the control group = “protects

Answer 30

Think of “surveys” * Measure disease & exposure at the same time in a population * Useful for assessing prevalence, exploring relationships & generating hypotheses * Do not give evidence of cause & effect or change over time * Key issues in critical appraisal include sampling, confounding & validity/reliability of methods/data

Answer 31

Think of “rare diseases” * Select “subjects” and “patients” (i.e. on basis of outcome variable, with or without disease). * CCS look back to behaviour in the past * Case-control studies are best for rare disease, assessment of multiple exposures * When evaluating a case-control study consider suitability of control group and sources of bias

Answer 32

Think of “large groups” being followed over time * Group 1 - people exposed to variable of interest * Group 2 - people not exposed to variable of interest * All subjects are healthy at the start of the trial except in “inception” cohorts (Greenhalgh, ch.3)

Answer 33

Prospective cohort – exposure measured now and in the future

Answer 34

Retrospective cohort - based on measurements of exposure taken in the past

Answer 35

All measure the risk of future outcome on the exposed group vs. on the non-exposed group

Answer 36

risk if exposed/ risk if not exposed – RR = 1 no effect of exposure – RR > 1 the exposure increases risk – RR <1 the exposure decreases risk

Answer 37

• Note that the OR is a related measure to RR, it approximates to the RR – The OR gives an estimate based on number of existing events (“cases”) now; the RR esimates a risk based on diseaese development over time

Answer 38

RR is not absolute risk, needs changing to absolute risk, e.g. recent stories about processed meat and cancer risk (RR 17% absolute risk 1 extra case per 1000) Effect absolute risk is different

Answer 39

randomisation, blinding, - do researchers know who is getting what intervention placebo ITT (intention to treat, do people that start the study end the study – problem with intervention may cause dropout) • Experimental studies provide the best evidence of cause and effect – results can be used to perform meta-analyses

Answer 40

Study population Define selction criteria for participsnts Select and invite partticipants Randomisation of participants into control and treatment groups Complete or drop out

Answer 41

• Limitations include ethical considerations, cost and factors leading to bias

Answer 42

• Systematic reviews are a type of Secondary Research – Overviews primary (original) research – Conducted following of a pre-defined, reproducible, and non-biased method (follows agreed standards) • Systematic reviews allow a quick evaluation of the cumulative evidence behind a particular intervention

Answer 43

– Was the research question relevant and specific? – Was the literature search conducted thorough and accurate? – Was the quality of the studies assessed (and weighed)? – Are conclusions reasonably deduced from the numerical results?

Answer 44

Why Interviews, observations Get opinions, views Non numerical, qualitative, descriptive ? Unrepresentative Non statistical

Answer 45

How many, what, how Quantified observations; interventions; laboratory Associations; causeeffect relationships Testing hypotheses; numerical or quantifiable Representative Statistical

Answer 46

* A process of review necessary to be able to judge the evidence * Involves identifying both positive and negative aspects of research or evidence * Allows to identify relevant information to enhance practice

Answer 47

Identify problem Review literature Develop testable hypothesis Design intervention Alect recruit subjects Obtain and analyse data Interpret results Diseeminate results

Answer 48

* Relevance of topic chosen when research will be finished * Practical issues (resources, funding, staff) * Ethical considerations * Study design and methodology * The play of chance (statistical analysis) * Relevance to practice

Answer 49

“To weigh up the evidence critically to assess its validity (closeness to the truth) and usefulness (applicability)”. (Sackett and Haynes, 1995) “A technique for increasing the effectiveness of your reading”

Answer 50

* A description * Making assumptions * Making unsupported generalisation * Not questioning the information you have * Giving inaccurate or misleading information * Quoting what the author’s say without giving your own views and stating the difference between the author’s opinions

Answer 51

* Information overload * “ Science is just one of many truths”. Professor Robert Winston * Variable quality of research * Need for evidence-based practice * Supports the decision-making process in clinical practice

Answer 52

* What was the research question / hypothesis? * Why is the research needed, is it original? * What was the study design? * Was the study design appropriate for the question?

Answer 53

• The sample – Sample size, power of the study – Inclusion/exclusion criteria – Randomised, differences between groups – Recruitment * Intervention * Outcome measures and measurement tools * Duration / follow up (loss to follow up) * Bias

Answer 54

* Comprehensive, precision * Bias in representation * Presentation * Descriptive * Analysis – Quantitative statistics – Qualitative

Answer 55

Analyse and evaluate – Question the information you have. • Argue – What is in favour of authors arguments, what is against? * Conclude and summarise your understanding of the main points * Develop your own view with justification

1. global determinants of health Flashcards

(79 cards)