Public Heath Flashcards
Causes of associations
Bias Chance Confounding Reverse causality True association
Define bias?
Selection bias?
Types of information bias? (4)
Publication bias?
Definition: A systematic error that results in a deviation from the true effect of an exposure on an outcome
Selection bias- Selection of study population eg. Non response of certain groups, loss to follow up of certain groups
Allocation bias e.g. different participants in the groups
Information bias (MORR) -
Measurement bias (different equipment measures things differently)
Observer bias- Observer’s expectations influence reporting
Recall bias- Past events not remembered or recalled correctly
Reporting bias- Respondent doesn’t report the truth because they feel ashamed/judged
Publication bias- Trials with negative results less likely to be published
Drug trials more likely to be published if sponsored by drug company
What are the brad ford hill criteria for causation ?
STD R Crap
Strength - very high relative risk (Relative Risk of 21)
Temporality - most important - exposure occurs before outcome (people smoke before developing lung cancer
Dose-response - more risk of outcome with more exposure (the more you smoke the higher the risk of lung cancer)
Reversibility - if you take away the exposure then the risk of disease decreases or is eliminated (stop smoking and you have a decreased risk of lung cancer after 10 years or so)
Consistency - the association is seen in different geographical areas, using different study designs, in different subjects (smoking is associated with lung cancer in dogs, mice and people, all over the world)
2 pros and 2 cons of RCT ? What is it/>
😍 Low risk of bias and confounding
😍 Can infer causality
😒 Time consuming and expensive
😒 Unrepresentative study population
Similar participant are randomly assigned to an intervention or control group
Case control basics? 2 pros 2 cons?
Observational study looking at cause of a disease. Compares similar participants with disease and controls without.
Looks retrospectively for exposure/cause
😍 Quick (as outcome has already happened)
😍 Good for rare outcomes (e.g. cancer)
😒 Difficult finding appropriately matched controls
😒 Prone to selection and information bias
Cross sectional basics? 2 pros / cons?
Observational study collecting data from a population and a specific point in time.
A snapshot of a group e.g. census or patient survey
😍 Large sample size
😍 Provides data on prevalence
😒 Risk of reverse causality- which came first?
What is reverse causality?
Outcome may have caused the exposure
Eg survey on whether depressed people are obese -which caused which
Cohort basics? Pros? Cons?
Longitudinal study in similar groups but with different risk factors/treatments. Follows them up over time
E.g. Framingham Heart Study or smoking doctors vs non-smoking doctors
😍 Can follow up rare exposure e.g. natural disaster
😍 Allow identification of risk factors (exposure causing outcome)
😒 Takes a long time, people drop out
§
Over a ten year period, there were 50 cases of lung cancer in Crookes (a population of 1,000 people).
What is the incidence (risk) of lung cancer per year?
(50/1000)/10 = 0.5%/ year
In Crookes (a population of 1000 people), 300 of them smoke. Of those who do smoke, 45 of them developed lung cancer. 5 of the non-smokers developed lung cancer. What is the relative risk of lung cancer in smokers?
Risk in exposed = 45/300 = 15%
Risk in unexposed = 5/700 = 0.7%
Relative risk (risk ratio) = 15/0.7 = 21.4 times more likely to develop lung cancer if you smoke compared to non smokers
In Crookes (a population of 1000 people), 300 of them smoke. Of those who do smoke, 45 of them developed lung cancer. 5 of the non-smokers developed lung cancer. What is the risk of lung cancer that is attributable to smoking?
Risk of lung ca in exposed = 45/300 = 15%
Risk of lung ca in unexposed = 5/700 = 0.7%
Attributable risk = (15/100)- (0.7/100) = 14.3
Attributable risk is the number of cases that are due to smoking so you take away the naturally occurring cases (“that would happen anyway”)
In Crookes (a population of 1000 people), 300 of them smoke. Of those who do smoke, 45 of them developed lung cancer. 5 of the non-smokers developed lung cancer. How many people would have to give up smoking to prevent one death from lung cancer?
Attributable risk = (15/100)-(0.7/100) = 0.143
NNT = 1/AR
= 1/0.143 = 6.99
So if 7 people stopped smoking you would prevent 1 death due to lung cancer in this population
Define sensitivity? Specificity ? PPV? NPV? And how to work out?
Sens - % correctly identified with disease (associated with increased false positives) a/(a+c)
Spec - %correctly excluded as not having the disease (low causes unnecessary follow ups) d/(b+d)
Positive predictive value= % of those with a positive test that actually have the disease a/(a+b)
Negative predictive value= % of those with a negative test who are actually disease free d/(c+d)
What is the Wilson screening criteria
INASEP
Important disease
Natural history of the disease must be understood e.g. detectable risk factors, disease marker
Simple, safe, precise and validated test
Acceptable to the population
Effective treatment from early detection with better outcomes than late detection
Policy of who should receive treatment
Disadvantages of screening
Over detection of sub clinical disease
Needless worry if the screening result comes back positive and/or exposure to harmful diagnostic tests e.g. colonoscopy after faecal occult blood.
Is this higher in low sensitivity of low specificity? Answer= low specificity!
Over-identification and treatment of sub-clinical and harmless disease.
Harm from the screening test or subsequent treatment e.g. antibiotic resistance in GBS pregnant women
lead time vs length time bias ?
Lead time bias= Early identification doesn’t alter outcome but appears to increase survival
e.g. patient knows they have the disease for longer
Length time bias= Disease that progress more slowly is more likely to be picked up by screening (i.e. symptom free and around for longer), which makes it appear that screening prolongs life.
Mnemonic for health inequality ?
PROGRESS Place of Residence (rural, urban, etc.) Race or ethnicity Occupation Gender Religion Education Socioeconomic status Social capital or resources
Definition of health
Key articles of human rights act of 1998
Definition of health: a state of complete physical, mental and social wellbeing. No merely the absence of disease or infirmity
Article 2 – right to life
Article 3- right to freedom from inhuman and degrading treatment
Article 8- right to respect for privacy and family life
Article 12- right to marry and found a family
Article 14- right to freedom from discrimination
Causes of errors
Sloth error – being lazy, not bothering to check results/ information for accuracy. There is incomplete evaluation.- improved by paying attention to detail, and checking information
Lack of skill – lack of appropirate skills teaching or practice- improved by being properly trained in your role and PPD!
Communication breakdown – unclear instructions or plans and not listening to others- improved by being approachable and listening. System failure Human factors Judgement failure Neglect Poor performance Misconduct
4 parts of negligence ? What is bolam ? Bolitho?
Was there a duty of care?
Was there a breach in that duty?
Was the patient harmed?
Was the harm due to the breach in care?
Bolam - Would a group of reasonable doctors do the same?
Bolitho - Would that be reasonable?
What are the 3 domains of public health?
Health improvement
Social interventions and preventing disease, promoting health and reducing inequalities
Health protection
Measures to control infectious disease risk and environmental hazards
Improving services
Organization and delivery of safe high quality services
What is the health needs assessment?
Is a way of systematically reviewing health issues to find a list of agreed priorities and resource allocations that will improve health and reduce inequalities
Needs assessment -> planning -> implementation -> evaluation -> repeat
What is the Bradshaw taxonomy of need? (4 types of need)
Felt need- individual perceptions
Feeling back pain
Expressed need- individual seeks help (demand)
Going to the Dr with back pain
Normative need- professional defines intervention
GP decides you need a physio
Comparative need- comparison
Compare all the referrals from GP X to physio to GP Y
What are the 3 approaches to HNA?
Epidemiological
Defines the problem and looks at the size of it using routine and primary data
Comparative
Looks at services received and compares with a similar area – may compare health outcomes
Corporate
Asks the local population and health profs what the needs are using focus groups etc
Pros / cons of epidemiological approach to HNA
pros
Uses existing data
Provides incidence mortality and morbidity data
Can evaluate services by trends over time
Cons
Quality of data is vsriable
Data collected might not be the data required
Doesn’t consider the felt needs or opinions of those its looking at
Pros / cons of comparative approach to HNA
Comparative
Pros
Quick and cheap if data available
Indicates whether services are better or worse in comparable areas
Cons
Difficult to find a comparable
Pros / cons of corporate approach to HNA
Corporate
Pros
Based on felt and expressed needs
Recognises the knowledge and experiances of those working in the population
Takes into account a wide range of views
Cons
Difficult to establish need from demand
Groups may have vested interests
May be influences by political agendas
3 types of health behaviour
Health behaviour
Prevent disease
E.g go for a run, wear sun screen etc
Illness behaviour
Going to the DR
Sick role behaviour
Taking medications
3 types of behaviour change
Transtheorectical model- Pre contemplation- not ready yet Contemplation Preparation Action Maintenance / relapse
Nudge – nudge the environment for positive change – fruit near the checkouts
Theory of planned behaviours – best predictor of change is intention
Persons attitude to the behaviour (smoking is bad)
Subjective norms- percieved social pressures/norms (GP wants me to give up)
Perceived behaviours – how well the person thinks they can do it (I can do it!!)
What 3 principles is resource allocation based on?
Egalitarian
Based on the idea that the NHS was founded on the requirement to provide all care that is necessary and required to everyone
Equal for everyone
Maximising
Maximise public utlity/ profits? – an act is evaluated soley in terms of its consequences – will it be beneficial?
Give to those who it is most likely to benefit the most
Libertarian
Each is responsible for their own health
The german health incentive scheme – change health behaviours and get bonuses for partipating
What are never events? What do they lead to? Egs?
Serious, largely preventable patient safety incidents that should not occur if the available preventative measures have been implemented
Financial penalties, affect trust reputation, CQC visits
Surgical - wrong site / retained foreign object
Psych - escape of a transfer pt
Medical - wrong route of chemo administration / opioid overdose
A group of patients with lung cancer and a group of age matched controls are compared by looking at pack life history to assess the impact of smoking on lung cancer.
What type of study is this?
Give an example of a covariate?
What type of bias may interfere with results?
Retrospective case-control
E.g. work exposure, gender, type of tobacco, level of passive smoking
Selection bias e.g allocation bias (inappropriately matched controls) and information bias e.g. recall and reporting bias
3 benefits of cohort studies
Identify causation
Ethical assessment of RF (if exposure was thought to be dangerous would not be fair to do a RCT )
Allows assessment of multiple risk factors
A new diagnostic test is being developed: 100 people known to have the disease are tested as are 200 disease-free controls. Ninety of the cases yield positive tests, as do 30 of the controls.
What is the specificity of the test?
What is the positive predictive value?
What do these mean?
85%
75%
85% specific means 85% of people without the disease were correctly excluded. 75% PPV means of those who had a positive result, 75% actually had the disease.
Why do you notify public health
So HPA can take control measures
You may be the only one to tell them
Duty of registered medical practitioners
Role of the consultant in communicable disease control
Surveillance - Notification / lab data to monitor diseases
Prevention - Eg vaccination
Control
Basics of outbreak management
Clarify problem - make a diagnosis
Decide if its an outbreak - 2 or more related cases
Gain help - microbiologist, consultant in infectious disease, health visitors
Call an outbreak meeting
Identify cause
Control measures
4 causes of homelessness
Relationship breakdown
Domestic abuse
Dispute with parents
Bereavement
Health problems faced by homeless
Infectious disease Foot and teeth health Resp problems Violence / rape Sexual health Mental illness Nutrition Addiction
Barriers to healthcare for travellers
Reluctance of GPs to register them Illiteracy Communication difficulties Lack of permanent site Mistrust of professionals
Barriers to healthcare with homeless
Access
-Location / discrimination
Lack of integration between primary care and other agencies
-housing, social services, criminal justice
Other things on mind
-don’t prioritise health
May not know where to find help
