Epidemiology Flashcards
Confidence interval definition
Range of values for variable of interest with specified probability of including true value of variable.
Usually 95% significant if confidence limits does not cross “no effect value” (eg. Relative risk of 1)
P-value definition
Probability that results observed in a study is due to chance, quantified as a value between 0-1.
<0.05 generally accepted as statistically significant
Number needed to treat
- Formula
- Definition
- 1/ARR (absolute risk reduction)
2. Number of patients who received the treatment to produce ADDITIONAL benefit for one individual.
When to use relative risk
In cohort studies
When describing the POPULATION IMPACT
(When informing patients, more useful to use Absolute Risk)
- Attributable risk formula
- Population attributable risk
- What does attributable risk show?
- (Incidence in exposed - incidence in unexposed) / (incidence in exposed)
- expressed as a percentage - The above “attributable risk” x prevalence in population
- Demonstrates what proportion of a specific disease is attributable to the risk factor. In a population, can indicate potential impact of control measures when studying what impact of the disease in a population is attributable to this.
Random error.
- What is sampling error
- What is type I (alpha) error
- What is type II (beta) error
- Sample error is commonest cause of error, from randomness of sample collection. Reduce by increasing power.
- False association made when there is none. More harmful as we will then try to make interventions. Set p-values to minimise alpha errors (eg. <0.05 chance alpha error will occur, can even go lower)
- No association made when there is one. Can reduce by increasing power
Systematic error (SIC)
- What is Selection Bias
- What is Information Bias
- What is Confounding bias
- Especially happens in case-control studies, where case population does not really match control population
- Wrongly classified into a group. Eg. Case control, the subject may not remember having had been exposed to a certain risk factor and therefore not counted.
- When the conclusion is confused because another exposure has not been taken into account.
6 steps of Priority Setting for implementing a new intervention
- Need - do corporate needs assessment, and all the other ways of needs assessment
- Intervention - critically appraise evidence. 2 problems:
1) PH interventions often multifaceted with confounding factors
2) When evidence states current treatment is ineffective - Acceptable (by target population. Do surveys and
questionnaires etc) - Cost-effective. Think about DIRECT, INDIRECT and OPPORUNITY costs. Also the cost-efficacy, effectiveness, utility and benefit etc.
- Ethically justified? (4 principles)
- Overall justified? (Look back at all the above)
What are the 3 main approaches to developing a health needs assessment?
- Cooperate
- Comparative
- Epidemiological
Comparative needs assessment
- Where is data obtained and what is it used for?
- Advantages?
- Disadvantages
- Local hospitals and GP practices, to allow comparison between geographical areas to be made, and compares with previously set standards
- Data already present, there is an easy starting point
- Data may be old, different hospitals have different admission criteria for the same conditions, misdiagnoses
Corporate needs assessment
- Where is the data obtained, and what is it used for?
- Advantages?
- Disadvantages?
1. "CQFI" = corporate Citizens juries Questionnaires Focus groups Interviews in various settings
- Involves public opinion and gives insight into the other factors that affect health
- Time-consuming to carry out, and require a lot of effort to do these qualitative assessments
Epidemiological approach to needs assessment
- What main questions do they address
- Purpose of epidemiological needs assessment
- Examples of things they look at…
- Large-scale analyses, identifying several issues:
- how big is the problem
- what is currently being provided
- are the current services effective, and can something else be done?
Used for WHOLE SERVICES, such as mental health, coronary heart diseases etc to assess priorities of service provision in these circumstances… - Identify health improvements which can be achieved by reallocation resources in order to remedy over-provision and unmet needs.
- Whole specialties (eg. Mental health)
Diseases (eg. Coronary heart disease)
Client groups (eg. Alcoholics)
Vulnerable groups (eg. Ethnic minorities)
What are the 5 main stages of a policy-making/ planning cycle.
- Assess health status
- Identify problem and prioritise
- Identify options for solutions
- Choose one solution
- Implement this service, with consultation by material providers etc.
- Evaluate and monitor
- Repeat
Crude death rate definition
Total deaths / 1000 / YEAR
Directly standardised death rate
- “Definition” and method
- Disadvantages
- Death rate that would have occurred in the REFERENCE population if it had same age- and sex- specific deaths the study population
- Tedious to construct, as would need to standardise all the different sex or age groups in every study population to compare to reference. Easier just to keep ref population as a standard, and compare with all the other populations
What is indirect standardisation and SMR?
What does it used for?
Definition of SMR >/< 100
- SMR = deaths observed in study population group: death rate if it were standardised with same age and sex-standardised as ref population x 100
- Used to compare populations that are differentiated by factors such as geographical, social class eg.
- > 100 suggests study population is doing worse than reference population, and vice versa.
Perinatal mortality rate definition
Stillbirths + within first 28days of life per 1000 LIVE BIRTHS
(Basically is Stillbirth mortality rate + neonatal mortality rate)
Neonatal mortality rate definition
No of deaths within first 28days of life per 1000 LIVE BIRTHS
Early = within first week of life
Late = from 2nd week to 4th week
Post-neonatal mortality rate definition
From 28days to 1y/o per 1000 live births
Maternal mortality rate definition
Death from PUERPERAL causes during pregnancy, and within 42days after birth per 1000 live births
Stillbirth death rate definition
Stillbirths per 1000 total live births
NB. Stillbirth is after 28/40 gestation
Infant mortality rate definition
Deaths at <1y/o per 1000 live births
Basically neonatal and post-neonatal deaths summated
Years of life lost definition
No. of deaths lost before 75y/o, over total population.
Can give some information about causes of death.
Clinical governance definition
FRAMEWORK THROUGH which NHS organisations and staff are ACCOUNTABLE for Quality of patient care.
Have responsibility of maintaining a specified standard of practice.
“PRAE”
Patients and public involvement
Risk management - basically error reporting stuff (DH “an organisation with a memory report”) and SYSTEMS failure rather than individual usually. NPSA too
Clinical Auditing
Education staff and training when necessary.
How National Patient Safety Agency (NPSA) helps in risk management aspect of clinical governance
“ROLA”
Report incidents ensured
Open culture of reporting is promoted
Local authority involvement - supports them in Expressing public concerns
Alleviates fear and assures that changes will be made
Who doctors may be accountable towards
“PCCG” hierarchy
Patients firstly (also their teachers, managers etc and all the people who own doctors) Colleagues (important drive for improvement) ColleGES (royal colleges hold standards that we must maintain) GMC (also has standards, appraise and revalidates practitioners. Assesses knowledge, skills, quality of care etc)
What is the “An organisation with memory”?
Department of Health document
“COUS”
mechanisms to implement Changes in response to the errors
Open culture of reporting
Unified protocols for reporting
System cause of error more appreciated
4 main places to obtain information about Morbidity of a population
“HIPD”
- Hospital data - gets lot of patient details, already available, may not be accurate
- Infectious disease registers - information about infectious diseases, prevalence, surveillance and monitor any more new outbreaks. Vaccination requirements or sexual health service provision requirements
- Primary care - GP details. Decennial GP morbidity surveys. May be incomplete sometimes, but sufficient for local audits etc.
- Disease registers - good information about exact details of certain diseases (eg. Primary, mets, outcome + treatment, demographics), and allows comparison between countries. Best example is Cancer. Expensive to maintain.
Other places to obtain Morbidity Statistics
“LANDS”
- Local health surveys - time-consuming, but obtains info on what the local population wants/feel is important (ie. local needs assessment…)
- Accident surveys
- National statistics: population trends, longitudinal studies, congenital malformations…abortions etc.
- Disability register - self explanatory
- Social security stats - sickness and injury benefit claims
Some more: school health reports…
Self-reporting data for Morbidity statistics
Problems with self-reporting stats
Household survey: acute/chronic diseases, other factors affecting health (eg. Income etc.). Only from 12,000 households per year. Allows CROSS-TABULATE health data with other variables, eg. Housing, employment, income etc.
Health survey: all over the country, annual, similar information but on HEALTH STATUS (and other factors that directly affect health, eg. Smoking and drinking, but not employment and social statusetc)
Decennial Census: unpaid care and general health
Problem = recall bias; info can rapidly OUTDATED
What is cost-minimisation?
- Cost
- Benefit
- When is it used?
- Money
- No known benefit
- When two drugs or something have presumed same efficacy, so pick the cheaper one
What is cost-effectiveness?
- Cost
- Benefit
- When is it used
- Money
- Per unit of benefit
- Usually for comparing new intervention with existing one. Take costs of 2 interventions, and number of benefit per intervention. Directly compares 2 interventions with same outcome.
Cost-utility
- Cost
- Benefit
- When it is used
- Usually money
- QALY
- To compare what services to fund
Problems with QALY
Biases age - discriminates extremes of age. (Decreased LE disadvantages elderly)
Not based on individual assessment, but values determined by others
Disadvantage the already disabled
Lack sensitivity within a disease area, very complex levels within all the diseases
Cost-benefit
- Cost
- Benefit
- Uses
- Money
- Difficult to measure and controversial
- Used even more globally in determining where tax payers money’s get spent - whether in health or education etc.
Definition of EMERGENCY
Event or situation which threatens serious damage to human welfare, the environment or security.
At least 1 of:
>10 people killed
>100 people affected
International assistance called
Definition of an outbreak
Local epidemic, where >2 cases are linked in TIME and SPACE
How to manage an outbreak
- Confirm this is an outbreak using the criteria on another card
- Call outbreak control team: PH, microbiologist, doctors, communication experts(! - to control what the media prints, and also advertise for more cases)
- CASE DEFINITION - time, place, symptoms
- Treat patients, while investigate at same time. CULTURES + identify organism
- SMALL DESCRIPTIVE STUDY, with Epidemiology Curve (?point source outbreak, any further cases from secondary infection by main people infected). Surveillance
- Prevent further infection, using imagination about what could be improved. ?improved training, shut place down etc.
Eradication definition
Disease no longer in environment, but still exists
Eg. Smallpox
Elimination of an infectious disease definition
Still persists in the environment, but control transmission
Eg. Vaccination of chicken to prevent S. enteritidis transmission, milk pasteurisation for Listeria and O157
Definition of containment of an organism
Accept that it cannot be eliminated or eradicated, aim to prevent it from becoming a Significant Public Health problem
Eg. Annual influenza vaccinations
Main reasons why infectious diseases are still such a huge problem, ie. barriers
Government - politics, health inequalities
Individuals - attitudes and behaviours do not change (STIs, HIV etc)
Vaccinations and pharmacology - drug resistance and slowness in developing vaccinations.
Diseases itself - reservoirs still exist, new diseases emerge
Definition of Herd Immunity
Sufficient proportion of the population is immune, such that those non-immune are protected from disease.
Usually >95% of population immune, of Reproductive Rate <1.
(Reproductive rate = mean number of cases generated by each new original case emerged)
Screening: The Condition
- Natural history well known, ie. pathogenesis. Pre-symptomatic phase for secondary prevention
- Significant public health problem
- All practical and cost-effective primary prevention strategies must have been implemented
Screening: The Test
- Good specificity and sensitivity
- Acceptable to the general population
- Further investigations and treatments can follow positive result picked up with the test. Agreed policy.
- Population CUT OFF specified to define what population group would need it.
Screening: the Treatment
- Should be available and effective.
- If unpleasant, should be made clear to the patient at the beginning of the screen (eg. Down’s syndrome “treatment” could involve TOP)
Screening: The Program
- Cost-effective compared to total expenditure on healthcare
- Evidence shows reduce M&M
- Evidence-based that it is generally acceptable by the general public and healthcare professionals, and that the long term benefits of M&M outweighs any psychological harm to patients.
- Adequately resourced
Reasons why Prostate Cancer currently does not have a screening programme
- The condition: natural history not fully understood. Many men die with small degrees of prostate cancer that never go on to cause problems.
- The test: PSA - not specific nor sensitive enough. Many things cause raised PSA, incl. infection, BPH etc.
- also the following Ix is very invasive - prostate biopsy - Treatment: radical prostatectomy, has complications incl. impotence and irritation?
- Little supporting evidence. High NNT…?
Risk factors associated with teenage pregnancy
“MORSE”
Children to teenage Mothers Young Offenders Race - Caribbean and black Social deprivation Education and self esteem - low educational attainment and low self esteem
What are the layers that determine health?
- Constitutional factors (eg. Age, sex, race, genes)
- Lifestyle factors
- Community
- Living and work (WW HH EE A)
- Water and sanitation
- Work environment
- Healthcare
- Housing
- Education
- unEmployment
- Agriculture and food - General socioeconomic conditions
Problems associated with teenage pregnancies and mothers
- Baby: more likely to suffer from illnesses, have conduct disorders, education problems etc. vicious cycle. Breastfeed less
- More likely get ante/perinatal complications (eg. Pre-eclampsia, premature births etc.)
- Socioeconomic: less education, more likely to settle with a partner also of low attainment. Become poorer, Drug abuse.
- Domestic violence more likely
- Maternal health: more likely to suffer Mental illness, Drug abuse problems
Problems associated with domestic violence for woman
Poorer self esteem
Mental health problems
Physical consequences of abuse
Interventions for prevention of Teenage Pregnancy
Targeted youth population?
Sex education classes
Contraception clinics
Sexual health clinics
Relative risk equation
Risk of developing something in Exposed group / risk in unexposed group
Risk = (no. of people who develop the disease/total number of people)
Therefore, exposure has 5 times the risk of developing condition than if not exposed. (Someone is 5x more likely to give develop a disease)
Odds ratio equation
Odds in exposed / odds in unexposed
Odd = no. developed the disease / no. who did not develop the disease
Odds of someone who develops the disease in exposed is 8 times more than someone unexposed
What is a Forest Plot?
A graphical design to illustrate effects of treatments in multiple studies that try to answer the same question - ie. a metanalysis.
Vertical line is no effect line
Diameter of diamond is overall confidence interval
Area of the blobs show representation in the metanalysis
4 things needed in contemplation of behavioural change
- Desire to change
- Belief that they can change (self-efficacy)
- Belief that change will bring about the desired effects (action-efficacy)
- Know how to change
Pachaska and DiClements stages of behaviour change
“PCAMR”
- Pre-contemplate
- Contemplate
- Action
- Maintenance (eg. Special strategies to recognise high-risk situations)
- Relapse (usually when low emotional state. We should realise these as relapses as opposed to weakness..), or remittance
Maxwells dimensions of evaluating a service
"EEEAAR" Efficiency Effectiveness EQUITY Acceptability Accessibility Relevance
