LECNOTES - Copy PDF

WEEK 1 The pyramid basically shows the health effects that could be caused by your society, where u live and stuff like that, whereas the top two tiers show you how or with what that can be treated. Epidemiology is the study of often and the reason for the precise or occurrence of a certain diseas...

WEEK 1 The pyramid basically shows the health effects that could be caused by your society, where u live and stuff like that, whereas the top two tiers show you how or with what that can be treated. Epidemiology is the study of often and the reason for the precise or occurrence of a certain disease within different groups of people. An epidemiologist is an individual who keeps the public informed on many different public health issues and provides/purposes solution for the betterment of the public. Some people can abuse their fame and can wrongly promote activities which should not be worth promoting (e.g. Kim’s vampire facial treatment). Epidemiology can be communicated in many different ways including providing diagrams such as histograms, pie graphs or heat maps (many more). Lots of misinformation can be spread and as an epidemiologist it is our job to debunk these myths with evidence and provide your information and solution into society for instance the WHO has a site specifically mysting busts about COVID. Therefore there is many instances where false information is published through the public and has to be myth busted. WHAT IS IT? - Diseases and what their cause or reasoning is behind it, and you can find this course through systematic ways such as experiments or treatments etc. (thru epidemiology you are trying to find the cause of the disease) The study design in place should be structured What is epidemiology? To simply state, it is the study of places and people and patterns of diseases. It is observation and experimental research which can help determine what impacts people and how at different areas affected with the diseases. - Includes: Surveillance Experimentation And look at geographically overtime It is the: - Science of epidemics - Science of occurrence of illness - Study of distribution and determinants of disease in humans ALL OF THIS CAN BE DONE THRU A SYSTEMATIC COLLECTIION AND ANALYSIS OF DATA AND THIS CAN BE USED TO INFORM: - Public health interventions - Prevent, treat or manage diseases KEY FACTS: - Focussed on population and not individuals – for this reason it is easier to compare across different regions and populations (including countries) as we look at a rate and not the number of individuals Target pop, sample pop etc. - Also interregate methods behind data to see whether or not it is useful. these different graphs, as epidemologists should raise questions with a desire to know more, e.g. why does this country have that or this one have this – data is useful for answering questions but also generating new ones. Some graphs have trendlines which display the cumulative or decreasing cases. History of EPI John Graunt recorded bills of mortality and then started to indlcude more details such as birth or gender which gave everyone a pattern and a clearer udntersnat of diseases (1603) James Lind a Scottish naval surgeon used the knowledge that people were getting sick from scury and dying and conducted an experiment on diet revealing that oranges n lemons were effective remedies (1700s) Ignaz Semmelweis “discovered hand washing” through observation studies. He did this by measuring mortality rates of babies born at birth when birthed by midwives and doctors, and because doctors did autoposies before doing this, they carried germs and led to the death of babies (1800s). John Snow as well created a map to determine that chloera was a water transmitting disease. Doll and Hill also suggested that tobacco increased the risk of luing cancer and this data was confirmed with later studies (1900s). PICO REVIEW “Is chewing gum more or less effective for reducing pain in children with ear infections when compared to decongestants?" P – children with ear infections I – chewing gum C – decongestants O – reducing paim AND – used to combine concepts and gets more specific n less results OR – combine similar concepts and more results VIDEO Spanish flu – spiked down meaning that bc government forced activities such as war leads to a decline in wealth and health. Western counties were at the top Once they gained independence they improved – the western cuntried were slowing them down Different Chinese provinces had different statuses – not all of the parts of countries have the same level of wealth as stated. Industrial revolution improved a lot in not only wealth but also health RR 1.2, 95%CI 0.9 to 1.5 CI and the range you look at is the 0.9 – 1.5 Because risk ratio is related 1, since it crosses 1 it is not significant When looking at mean differences the reference value is 0 they can have negative and positive values. So if the range crosses 0 it is not significant. INTERPRETATIONS OF RISK VALUES AND STUFF ON EXAM! Siginicnce is dedfine by a p value less than 0.05 WEEK 2 What is health? Firstly, we need to define health to maintain and improve it as well as preventing poor health outcomes. WHO says that health is state of complete physical, mental and social well-being not just the absence of disease or infirmity. What is a disease? A pathological process, most often physical, whose quality deviates from the biological norm. (something physical) What is illness? It is a feeling or experience of unhealth which is personal to each patient. It could often accompany disease but the disease may be undeclared such as early stages of cancer or TB etc. EPI defining health and disease focuses on: - Disease presence - Disease absence How to measure death? To measure death, we look at case fatality and there calculate a fatality rate. Numerator: people who die Denominator: everyone that was diagnosed with something in a certain period Examples of mortality data: To establish a population at risk, we take what we know about the certain condition for further study in epidemology. Prevalence - It is the frequency of existing cases (number of defined people in a pop which alr have the condition of interest) - To calculate it you need to know: Number of people who have the condition (regardless of when they got it) The total population who are at risk of having it To calculate: (# of people w disease at a certain time)/(# of people at risk at certain time) - There are two types of prevalence Point prevalence – case that exists at a certain point in time e.g. census, percent of people at GP clinic (AT THIS POINT IN TIME) Period prevalence – cases existing over a period of time (week, month, year) e.g. percent of nursing home paitents who had parkinson’s in 2016 - Factors affecting prevalence: Increased by: Longer disease duration In-migration of cases Out-migration of healthy people Improved diagnostics (better reporting)  Increase in new cases Decreased by: Shorter duration of disease High-case fatality Out-migration of cases Improved curate rate of cases Decrease in incidence (new cases) - Incidence can be defined as the frequence of new cases and is a measure of risk (as people in the denominator can be part of the numerator) Quantifies the # of problems which develop in a population at risk over a period of time Usually given as number of cases To calculate it you need: # of people in defined population # of people who developed the condition in that time period Specified time period Ljnswodwbdi - Proportion is way of expressing counts with recognition that those people are part of a bigger population (population does not stay the same it can fluctuate) Expressed as % or rate - Ratae takes into account the time each person is at risk of an outcome - Crude and specific rate? Commented [SS1]: What is the diff? Burden of Disease - How a disease affects the quality of life of people There are three things which impact the quality of someone’s life: 1. Impairment Any loss or abnormality of psychological or anatomic structure/function e.g. someone loses a leg 2. A restriction (which results from an impairment) e.g. disability of a person who has lost a leg to perform a task which a person with both legs can easily perform 3. Handicap A disadvantage for an indivdiual resulting from impairment or disability e.g. losing a leg doesn’t allow them to play soccer - Years of life lost due to death (YLL) → takes someone’s life expextancy to when they ac die e.g. if someone dies suddenly from illness then they have lost years of life Commented [SS2]: Clearly explain please - Years of life lost due to disability (YLD) → the time between when they got the disability to death - Disability adjusted life year (DALY) → amount of healthy time you have lost due to an illness = YLL + YLD Commented [SS3]: Difference? - Qaulity adjusted life years (QALY) → how good is your quality of life once you acquired a certain illness … use this to determine the number of years which an intervention/treatment can add WEEK 3 : p Generalisliablity – looks at two things to help deicede if study is relevant or not: - Reliability → getting same result everytime (random error – control by increasing sample size or statistic analyses) - Validity → whether or not the results are applicable in other settings with same population – how close to true value (systematic error – to improve have good methodlogical design) SAMPLING: the method in which how we find the people to study or test or run a clinical trial with. Probability sampling – used in quantitative studies, random, we know who our pool of people are coming from - Simple random → take a random sample from known set of people - Stratified → based on strata groups (separate based off age, sex, geo location) - Systematic → accoridng to some order e.g. recruiting every 10 th person - Cluster → recruited from cluster or group (e.g. school, hospital ward) Non-probability – mostly in qualitative but can be used in quant as well - Convenience → people who just were there at the right time e.g. people in a shopping centre, people replying to an add or answer your survey - Snowball → people sharing something (usually a topic that is hard to discuss or something that is hard to get someone to talk about) - Purposive → looking for people with a particular characteristic or maybe they use a particular service These two are methods of sampling and e.g. convenience is just sampling BIASES IN SAMPLE - volunteer bias → bias based on why those people want to go into the study e.g. family history, money involved, they know someone who has this concern etc - The result may not apply to the whole population - Self selection bias – non-reps of target pop, could usually be becausse of self- selection - Healthy volunteer effect → needs to be considered in occupatoinal studies If the randomisation list is not concealed then allocation bias can be introduced. Randomisation is a two step process:  1. Randomisation sequence needs to be established 2. Sequence or allocation needs to be protected (e.g. by someone who is not involved in the study at all) Commented [SS4]: Does this include the surveyers or detectors or the... HAWTHORNE VS PLACEBO Hawthorne → people behave differently knowing theyre in a study Placebo → people have different expectations knowing theyre in a study/or on a treatment Attrition bias → are people staying in the study or dropping out (if they are we need to know why) and it is overcome by an intention to treat analysis If intention to treat has occurred look for this sort of diagram which also says why people have dropped out look at: people who went into the study, and if they leftwhy they left Researcher vs assessor? Commented [SS5]: Do both need to be blinded to avoid detection bias and wat is the difference between them? Researcher – do everything study (recruiting particpaunbts and esseing outvome) Assessing outcome (they don’t know the drugs or people in groups etc) – blinding if they don’t mention Recruiter – just for recruiting Allocation assessor – allocate people e There can be two types of classifications:  1. Nondifferential – errors occur equally in all the groups being compared e.g. if a bp machine is inaccurate, it will be inaccurate for both groups – resulting in random error and instrument bias – but also a deflection from “true” results 2. Differential misclassification – likely to occur in one group more than another due to systematic error e.g. recall, response or observer bias, Interviewer/observer bias – occurs when there is a chance that different people mi8ght see things differently e.g. in an interview all people should be asked the same questions and be asked them in the same way too - It can be corrected by: Standardised questions (same q for everyone and same way of asking) Inter-observer/inter-rater reliability (see if different observers record data in the same or different way) – that’s why it is important to ensure there are many people in the study EXPLAIN WILL ROGERS PHEN! – it is just an effect which could also introduce bias Change in categories and how that influences in results, and it doesn’t have anything to do with the intervention Confounding bias – it can be overcome by stratification (it can be thought of as another variable which could influence the results) Publication bias – more relevant to systematic reviews Sampling – particualr pop of interest and you are sampling that pop to ensure that the participants are representative of the true pop Healthy volunteer effect – people who are healthy will participate and will be more inclined to be a part in those studies Hawethprn effect – acts a bit differently opposed tonot in the study because they are being monitored in a study Placebo – perceived effect bc the person thinks they are taking the actual intervention but there are not Attrition bias – lost to follow up, enter a study and they could just drop out (could be because theyre notresponsive or maybe they are facing adverse effects) – intention to treat analysis (takes into account the people who have dropped the study) Will Rodgers effect – reclassify individual into a different category, Publication bias – having a finding which has a strong association with something else Funnel plots display results of each individual study → everything above 1 = increase in risk and below is not, No pub bias bc even spread of studies throughout graph Pub bias if the spread of the studies is only on one side using the reference line AT2 – whos publishing it bc this influences the actual results published Compare the drug with current best practice which in turn improves the healthcare system Whether it was undustry funded on independently funded (look at context e.g. political sources may influence bias or may not be most reliable), different dosage (e.g. giving too much or too less for a drug you want to put down or may it look poorer or with less side effects) → all of this influences publication bias How could publication bias be minimised? - Cochran group? - Alltrials (looks at all truals and summarises it) Look at types of data and Strata – charactertistics or just pregnant people or based off specific characteristics Cluster – based on groups e.g. people in hospitals In the case study – attrition bias addressed by intention to treat analysis, the sample size was smaller Selection bias because no exclusion or inclusion criteria (what they need and what they don’t want in the study for selecting people) -- > for AT2 put it in a table bc it doesn’t acco7unt for word count Week 4 Studies are a scientific process of answering questions using data from a population. Steps in studies: Types of studies: - Case series/study Commented [SS6]: Multiple cases whereas study is a single case Describes characteristics of same disease and exposure Aim: to understand the demographics, presentation, prognosis and other characteristics of people witj particular disease - Cross-sectional study snapshot in time which measures prevalence (measures exposure and outcome) Takes a certain population and measures health information at given point in time It just tells how many people have gotten disease overtime Involves asking people questions from questionnaires, health surveys Inexpensive Assess exposures/outcomes Assess health needs - Case control study Starts off with cases (people with a certain disease) and compares that with a control, in which people do not have the disease – in which both groups are asked about their previous risk factors and then they look at the odds ratio (main measurement)  Usually when you see the graph or plot which says case/control it would be case/control Presence of outcome or disease (case), control (no presence of disease) WHAT DO YOU MEAN BY MATCHING? – selecting Commented [SS7]: ESDS Minimise amount of bias – like same gender, location, to ensure that everyone is in the close.., to eliminate confondounf Past records – finding out exposures such as Observational study which collects data about past exposure, hence why, it is a retrospective study. Commonly used in food or outbreak investigations Advantage – quick and cheap, and bc they start off with cases it is good for identifying or working with common diseases Disadvantages – control selection is difficult, recall may not be as accurate (because they ask about past) You try to see if the exposure matches outcome that is why u] Selection bias – why - Cohort studies Group of people is followed overtime to see what happens to them and information about risk factors is collected People are not randomly assigned Then compare the occurrence of the disease to those exposed and those not exposed Main measurement used is relative risk – riks of disease in exposed group/risk of disesaszse in unexposed group Prospective – go forward in time Start off with exposure ad proceed off with outcome Retrospective – go back and collect data (e.g. looking at medical records to retrieve past info to see if they have exposure and then you would look forward in time to collect data in exposed and not exposed) Advantage: Time sequence of events can be determined (useful for knowing what caused the disease) Sub-analysis can occur from the collection of outcome/risk factors Disadvantage: High cost Not suitable for rare diseasde Risk of attrition bias as people might want to drop out DIFF BETWEEN PROSPECTIVE AND RETROP Commented [SS8]: ! - Interventional study Intervention is done on a group of people and the outcome is studied  Advantage Good evidence of outcome Randomisation ensures that each person has equal chance of receiving intervention and have similar characteristics Disadvantages Expensive Large number of participants Not always possible Study types that summarise: - systematic review Identifies al relevant studies on topic, asses quality, synthesises and interprets the findings, and presents and impartial and unbiased and balanced summary of evidence  Uses data from all the studies which analyse the same3 question with a similar study design → uses this data to do a combined statistical analysis and produces a single summary result - Randomised controlled trial Experimental study design (giving an intervention and placebo/control) Case control – look at outcome first and then go back to see exponses and unexposed e.g. lung cancer and not lung cancer (outcome) and go back in time to see If they wre exposed cohort (in general) – start off with exposure and then look at outcome is case control is alwaysretro proepects cohort – start with exposure (look ar records or stdies to divide into grouos) cancers- registries use data both go forards in time direction – hw they infor about the exposure (start with exposure and f8ind outcome) for case control (start with outcome and go to exposure – has a backwards inquiry but forward time) cohort (retro – forward in time but back inquiry) (prospec – forward in time and forward inquiry) information about exposure is already given to you – retrospective WEEK 5 What is a cause? - A cause is a concept which must always become before the outcome e.g. the cause of a disease or injury is an event or condition (or a comb of these) which plays an important role in yielding the health outcome Types of causes: Component → a variety of separate requirements which contribute to a cause e.g. obesity, insulin resistance, hypertension, high triglycerides, low LDL cholesterol - If you had family history or other confoundcing fators Sufficient → when all components are part of the one sufficient cause that lead to the effect e.g. metabolic syndrome Commented [SS9]: Wtf? You could get the outcome – factors influencing how you could possibly get it how you engage with - Breast cancer stuff you could get the radiation from e.g. sun The way the you gage with stuff Necessary cause → one component you must have in order to determine cause (environmental, biological, social determinants of health) e.g. breast cancer and BRCA gene -- usuallu patjogens Commented [SS10]: Double wtf Breast cancer – radiation ASSOCIATION DOES NOT MEAN CAUSATION - this means that something related to a concern or issue does not always mean that it was the cause of it what are baseline characteristics? – peoples characteristics whedn they jujst enrolled Commented [SS11]: help in the study (were they healthy did they have condutoins, later in the study they progressed and got diseasesd – could be other things usually in table The lecturer also said that the p value was less than 0.005 – what does this mean? Commented [SS12]: helpp What does significance actually mean – like practically Causation There are many risk factors: - predisposing factors → are there any do age, sex, genetics - disabling/enabling factor → do they further effect low income, poor nutrition, bad housing, access to healthcare - precipitating factors exposure to a specific disease agent - reinforcing factors repeated exposure bradford hill criteria is used for determining a casual relationship bradford hill criteria (considerations for causation): 1. temporal relation - the cause MUST happen before outcome - most essential one - sometimes difficult to demonstrate with case control cross-sectional studies e.g. patients with stomach cancer have low levels of vitamin C OR people who smoke develop lung cancer OR does having a vasectomy get you prostate cancer 2. plausibility - is it plausible – does it make biological sense - if theories are similar then it is plausible 3. consistency - have there been other studies which demonstrate similar findings - if results are similar then it is consistency - is it consistet with other peoples findings – refer to results of study 4. strength RR>2 is considered to demonstrate a strong relationship - strength of association (mean difference, RR, odds ratio)  5. dose response relationship - as dose of exposure increases so does the outcome - shown in a cluster bar graph or line graph 6. reversibility - if we remove exposure, does the chance of outcome also decrease? e.g. if someone gets lung cancer from smoking that doesn’t mean that when they stop smoking their lung cancer will go away 7. study design 8. judging the evidence - temporal relationship needs to be demonstrated - look causation – certain and it definitely leads to the outcome association – not certain but it could potentially lead to the outcome RCT is the strongest form of identifying the causation relationship Case control and cohort are more useful for association – this is because you are assessing for multiple things HOW TO INTERPRET P VALUES - p value only tells us significance of result (less than 0.05 it is statistically significant, if greater then it is not) if it is less than 0.05 then it does not mean the results have causation it just means the results are strongly correlated/associated – similar concept to greater than 0.05 WEEK 6 :P Reliable – consistent no matter who is measuring the results – the degree of stability exhibited when a measurement is repeated under identical conditions Intra-rater reliability – reliability of a single investigator Inter-rater reliability – reliability across multi-investigators Valid – correct (an expression of the degree to which a measurement actually measures what it claims to measure) Types of validity Content related - face validity – meets validity criteria makes sense what is being able to asked - construct validity – weather something measures something that it sets out to measure Criterion Related validity - concurrent – comparison to other tests - predictive validity – future outcomes (NAPLAN or ATAR) Primary data – original datra you collecr yourself for a study Secondary data – data you use for a study but has alr been collected by someone else or organisation - a study can use both primary and secondary data primary outcomes – main outcomes a study seeks to measure (often the big things like mortality, cure or number of cases) usually objective stated in the paper secondary outvome – additional outcomes collected to help interpret results (often include adverse and side effects) need to look for within the texts (you don’t need to report all of them for AT2 but at least 1 for the assessment) surrogate outcome – alternative to measures the “big thing”, could be quicker or more ethical e.g. measuring blood pressure when trying to asses to for stroke Collecting Data Two types of data: 1. numerical - continuous (BP) - discrete (number of children) 2. categorical - nominal (blood type, hair colour) - ordinal – has an implied order (measure of pain) OBSERVATION When observing data the setting of the study is important as it can influence the results. e.g. white coat effect – when doctors were in hospitals and wore white coats that increased the BP levels compared to when measured in relaxed settings. INTERVIEWS Ways to conduct this/types of: 1. Scheduled standardized interview – the structure is uniform as everyone is asked the same questions 2. Non-scheduled standardised – still the same qs but a bit more relaxed or the order of questions might change  3. Non-standardised – just having a conversation QUESTIONNAIRES - The people fill this out themselves - Collect demograph questions to be able to dewcribe the sample to be able to generalise the results to diff populations QUESTION TYPES 1. Open or closed - Open (what did you think) – detailed data - Closed (do you have any questions) 2. Double barrelled – you ask two things in one question 3. Leading questions – you ask a question in a way which leads people in a certain direction TYPES OF SCALES  -- opposite feelings on opp side of the scale Week 7 SIGNIFICANCE Statistical – it is statistically sig if it does not include 1 in the stage Clinical – MEAN DIFF – absolute difference between the means of two continuous variables - A mean diff of 0 means no effect (no difference between the two groups) e.g.  P VALUE - Probability that an event is due to chance Week 8 3 types of disease transmission KNOW THE DIFFERENCE AND BE ABLE TO IDENTIFY THEM: 1. direct – transmitted directly e.g. thru touch 2. Indirect Vehicle borne (started from inaminate object) Vector borne (animal mode of transmission e.g. a mosquito bite, ebola as it is carried from bat – can also be mechanical (no reproduction of agent) Vector-borne diseases are human illnesses caused by parasites, viruses and bacteria that are transmitted by vectors 3. Airborne Droplet – larger particles mostly from respiratory tract e.g. if someone coughs or sneezes Dust – smaller particles e.g. Types of outcomes: 1. Endemic → normal cases which we are usually present in an area e.g. the flu is normal/expected in some seasons 2. Sporadic → infrequent, when diseases happen at different places at different times and are not related 3. Cluster → a disease which occurs at different types of locations but there is not link there 4. Outbreak → there needs to be an epidemiological link (a common characteristic e.g. person, place or time) When identifying outbreak you must look at: - Person - Place - Time 5. Epidemic → an entire location or region has a particular part of a disease (confined in a particular country) 6. Pandemic → an entire location or region has a particular part of disease but it has now gone international REPRODUCTION NUMBER  - Reproduction number (R) – number of people that can potentially spread the disease looks at how easily disease spreads Less than 1 – the disease number is decreasing (GOOD) IDEAL RANGE/VALUE Greater than 1 – disease is continuing to spread within the population (BAD) Also considers vaccination rates e.g. 50% has vaccination that means that if the Ro was 2 then the new Ro is 1, meaning that 1 other person will get affected instead of affecting 2 other people - Basic reproduction number (Ro) – doesn’t consider of external influences and looks at disease itself and how likely that is to pass onto someone else (if the number is 3, then every 1 person can transmit the disease to three more people) WATCH LECTURE FOR TYPE OF GRAPH! Comparator – placebo, golden standard Population group – people who have history of hypertension and are current smokers EXAMPLE OF PICO/RESEARCH QUESTIONS: POPULATION, INTERVENTION, OUTCOME, COMPARATOR Mesh heading – particular word catergorised to a particular article Filter – RCT, english based Pop – adults who frequently fly Keyword – words that are included only in title or abstract OR – use either of the terms AND – they have to include both the terms Population and synomyms only use OR For more than one word add quotation marks If they are all related to the same concept (e.g. all terms from pop concept you use OR but from different concepts use AND) OR – only looking at POP first or synomyms associated with the pop TAKE SS OF FULL SEARCH Only apply RCT limit if there are less than 100 Over 100 like 133 is too broad WK9 Systematic Reviews – right at the top of the evidence pyramid because they collect all information from different studies and then result in a whole review Funnel plots can be used to determine if publication bias is present: The blue box one or the first lopsided one indicated presence of publication bias Interpreting Forest Plots ◼ Square can be used to determine the weighting JUST MAKE SURE TO READ THE SIDES BECAUSE SOMETIMES THE AXIS ARE SWAPPED AROUND – E,G, THE CONTROL MAY BE ON THE LEFT SIDE CONTINOUS OUTCOME  ◼ You still look at the diamond in the same way (significant or not) Chi – look at p value = if it is below 0.05 then there is significant heterogeneity WK10 Screening vs Diagnostic Diagnostic tests – performed when a patient who has a clinical problem is more likely to have the disease Screening – performed on healthy people before they display symptoms Diagnosis - Provides closure for a patient - Can prevent the severity of a disease or can be helpful in curing it - In infectious disease it is important to diagnose to prevent spread and complete a disease control Implication of Misdiagnosis - You might not be able to correctly treat the condition - The patient might feel angry or frustrated – a degree of mistrust - Adverse effects of if a strong medicine is taken Sensitivity – proportion of patients with the disease who have a positive test – how accurate is the test at identifying the true positives ◼ True positive (people who tested positive and had the disease)/total amount of people who have the disease (total number of people who have the disease including the people who tested negative for this disease) Specificity – proportion of patients without the disease who have a negative test – part of the diagnostic journey to rule out a condition ◼ True negative (the number of people who tested negative and don’t have the disease)/total no disease (those who tested negative and positive who both do not have the disease)  = 75% Positive Predicted Value – if someone tests positive test, what is the probability that they have the disease ◼ True positive / total positive (all the people that tested positive) = 64% confident that they have the disease Negative predictive value – for a patient with a negative result, what is the probability that they do not have the disease ◼ True negative / total negative  = 94% confident that their result is not accurate or true. Predictive value relate to the patient but there is a limitation because they rely on alr published values. Likelihood ratios - Uses the characteristics of a test and defines how much a positive or negative test result modifies the probability of a disease and is expressed as a ratio - Come as a positive (how likely is it someone will have the disease if they test positive) and negative (if someone tests negative how likely is it that someone does not have the disease) 1 – reference value Lower than 1 – decrease in association Greater than 1 – increase in association EXAM – strong or weak association (just know how large or small the number is) REVIEW THE GRAPHS!!!! – NANOGRAMS>> Prevention and Screening PREVENTION Primary prevention – things to do to try and stop the disease from occurring (washing your hands properly, educatuoin and awareness of safety habits, immunisations) Secondary prevention – aims to reduce the impact of disace/injury that has alr occurred (e.g. taking aspirin in people who are at risk of cardiovas or screening) Teritialy prevention – aims to alleviate the umpact of ongoing illness/injury (e..g. diet, rehabilityion to prevent a cardio event occurring again) SCREENING BIASES LENGTH-TIME BIAS - Occurs when screening is more likely to detect slow-growing disease that has a long phase without symptoms - Occurs due to the heterogeneity of diseases LEAD TIME BIAS - Patient is diagnosed earlier and appear to live longer because they know they have the disease for longer.  - Awareness of disease may make it falsely seem like early diagnosed patients live longer – when it doesn’t change it at all but allows for early detection VOLUNTEER BIAS - People who are selected are healthier - Same thing as healthy volunteer bias EVALUATION AND SCREENING PROGRAMS RE LOOK AT THE TYPES OF BIASES! WK12 Efficacy – does the intervention actually work Effectiveness – if we administer it in real-life will it work e.g. if you had one pill that reduces cholrestroel and bp compared to 7 pills which reduce chloresterole, BP and - Barriers affecting: Efficiency – if the intervention is effective what is cost/benefit ratio Just review w12 again!