Course Notes: Part I MFPHM Revision

EGJ

Part A MFPH

revision COURSE handout

“I couldn’t wait to get in there and tell them what I’d learned”

Stanley Hauerwas

The Hauerwas reader p25

For maximum benefit -

USE this handout:

Annotate, scribble, write examples

‘Read and forget; write and remember’

EPIDEMIOLOGY......

Epidemiological studies: design......

Expressing the main result......

Concepts and measures of risk......

Interpreting the result......

Chance......

Bias......

Confounding......

EPIDEMIOLOGY

: Epidemiology for the uninitiated

NOTE – Throughout the handout anything in this typeface (Arial 10) is a direct cut-and-paste from the syllabus

a) Epidemiology: use of routine vital and health statistics to describe the distribution of disease in time and place and by person; numerators, denominators and populations at risk; time at risk; methods for summarising data; incidence and prevalence including direct and indirect standardisation, years of life lost; measures of disease burden (event-based and time-based) and population attributable risks including identification of comparison groups appropriate to Public Health; sources of variation, its measurement and control; common errors in epidemiological measurement, their effect on numerator and denominator data and their avoidance; concepts and measures of risk; the odds ratio; rate ratio and risk ratio (relative risk); association and causation; biases; confounding, interactions, methods for assessment of effect modification; strategies to allow / adjust for confounding in design and analysis; the design, applications, strengths and weaknesses of descriptive studies and ecological studies; analysis of health and disease in small areas; design, applications, strengths and weaknesses of cross-sectional, analytical studies, and intervention studies (including randomised controlled trials); clustered data - effects on sample size and approaches to analysis; Numbers Needed to Treat (NNTs) - calculation, interpretation, advantages and disadvantages; time-trend analysis, time series designs; nested case-control studies; methods of allocation in intervention studies; studies of disease prognosis.

Appropriate use of statistical methods in the analysis and interpretation of epidemiological studies, including life-table analysis; electronic bibliographical databases and their limitations; grey literature; evidence based medicine and policy; the hierarchy of research evidence - from well conducted meta-analysis down to small case series, publication bias; the Cochrane Collaboration

Epidemiological studies: design

Descriptive studies: “How much of this stuff have we got?”

Case control studies: “What caused these cases?”

Cohort studies: “What effect does this have?”

Interventions incl. RCTs

PICO

(Modelling studies)

(Systematic reviews)

Retrospective vs prospective studies: ‘Five a day’

Interventional vs observational studies: beta carotene and lung cancer, HRT and CHD

Expressing the main result

Intention to treat analysis

Concepts and measures of risk

Relative risk

Absolute risk

Deaths per 100,000 male doctors per year from lung cancer:

smokers (>25 per day): 327

non-smokers: 14

Ratio of incidence (incidence rate ratio) =

Absolute risk difference =

(Excess rate/ risk attributable to smoking = )

Population attributable “risk”(aetiologic fraction)

Odds ratio

Number needed to treat (NNT)

Interpreting the result

Could the result be due to

Chance?
Bias?
Confounding?
REAL effect?

Chance

P values, CIs etc – but remember Type I and Type II errors

Bias

Systematic differences in

Sample / subjects
Measuring instrument
Observer

Confounding

The ‘other explanation’

Control of confounding:

Design

Analysis

Standardisation

Residual confounding

Over-adjustment

Effect modifiers [interaction]

This is a type of REALITY

Age related macular degeneration (de Jong PTVM NEJM 2006; 355: 1474 – 85)

Smokers (vs non): Odds ratio = 2.4

Homozygous for CFH Y 402H polymorphism Odds ratio = 7.6

Smoker AND homozygousOdds ratio = ?

Graphically:

fluoridation of water supply more beneficial to poor than to rich.

Riley JL et al Int J Epidemiol 1999; 28: 300 –5. Jones CM et al BMJ 1997; 315: 514 – 7

Causation

Bradford Hill criteria for causality (in order of importance):

[AB Hill. The environment and disease: association or causation? Proc R Soc Med 1965; 58: 295 - 300]

1.strength of association

2.consistent in different studies

3.specific

4.temporality

5.biological gradient e.g. more drinks / day -> higher RR

6.biologically plausible

7.coherence

8.experimental evidence

9.analogy (if thalidomide and rubella cause foetal malformation so may other drugs / viruses)

Mnemonic courtesy of Martin Bull:

A Statistical Cohort of Surgeons with TB Postulated the Cause to be an Environmental Agent!

NB if picture muddy may need to think about different types of cause:

Necessary / Sufficient

Underlying / Trigger

Etc

Uses of epidemiology (Jerry Morris):

Morris JN Uses of epidemiology Br Med J. 1955 August 13; 2(4936): 395–401

Historical trend
Community diagnosis
“Individual chances”
Operational research - how well services are working
Completing the clinical picture – study ALL cases
Identification of syndromes – ‘peptic ulcer’, ‘frailty’
Clues to causes

Putting it together – guidelines and recommendations

GRADE – strong and weak recommendations

Importance – prevents death

Size of effect – 30% reduction in risk

Precision – narrow CI

Certainty – many high quality RCTs

Risks and Burdens of therapy – no adverse effect but fortnightly iv infusion

Risk of event

Costs

Values (e.g. life or comfort?)

AGREE – quality of guidelines

Hierarchy of evidence

NB ‘Cultures of evidence’

Surveys

Constructing the survey instrument

Construction of valid questionnaires

Validity

content
face
criterion

concurrent

predictive

construct validity

Convergent/ discriminant

Reliability

‘degree to which measurement is free from measurement error’

Test - retest

Multiple form
Split half

Scales:

should be uni-dimensional
some instruments have domains
[e.g. “total SF36 score” is wrong]

Doing the fieldwork

methods of sampling from a population

The sample

Methods of sampling and allocation

random, quasi-random,

stratified

cluster

quota

convenience

nomination / snowball

the design of documentation for recording survey data

The instrument

Typography: font size, layout, tick boxes etc

Items: ambiguous questions / double questions / leading qq

Whole thing: running order (e.g. sensitive last)

Mode: paper - computer – telephone - internet

The interview

Interviewers

Select

Train

Monitor

Respondents

Introduction – gaining consent etc

Attempts to contact (how many? Time of day?)

Use of proxy allowed?

Methods for validating observational techniques

Validation of observational techniques:

inter-observer

interviewer training

videotaping

Observer variation

HEALTH INFORMATION

- Capture: how accurate? How complete?

– Coding – how fine grained?

-Output: how detailed? how often? How aggregate?

Routine data sources

Populations: conduct of censuses; collection of routine and ad hoc data; demography; important regional and international differences in populations, in respect of age, sex, occupation, social class, ethnicity and other characteristics; methods of population estimation and projection; life-tables and their demographic applications; population projections; the effect on population structure of fertility, mortality and migration; historical changes in population size and structure and factors underlying them; the significance of demographic changes for the health of the population and its need for health and related services; policies to address population growth nationally and globally

Sickness and health: sources of routine mortality and morbidity data, including primary care data, and how they are collected and published at international, national, regional and district levels; biases and artifacts in population data; the International Classification of Diseases and other methods of classification of disease and medical care; rates and ratios used to measure health status including geographical, occupational, social class and other socio-demographic variations; routine notification and registration systems for births, deaths and specific diseases, including cancer and other morbidity registers; pharmacoepidemiology, including use of prescribing and Pharmacy sales data; pharmacovigilance; data linkage within and across datasets

Population

UK Census

Census 2011

Health question

2011: How is your health in general? Very good / good / fair/ bad/ very bad

2001: Over the last 12 months would you say your health has on the whole been good / fairly good / not good ?

Disability question

2011: Are your day-to-day activities limited because of a health problem or disability which has lasted, or is expected to last, at least 12 months?

2001: Do you have any long-term illness, health problem or disability which limits your daily activities or the work you can do?

Include problems which are due to old age

Income question – there isn’t one in the UK census!

Ad hoc censuses

Census based measures

Deprivation scores

Jarman/ Townsend score

Index of multiple deprivation IMD2000 - NOT census based: 7 domains / 33 indicators: Income, Employment, Health and disability, Education skills and training, Barriers to housing and services, Living environment and Crime. See

Population

Estimates and projections

Historical change in population structure

1946 baby boom plus second wave

effect of economic downturn

Routine statistics

Mortality

Hospital

Inpatient

Ambulatory – A&E, outpatient

Diagnostics – lab, radiology

Primary care

Medical

Dental

Pharmacy

Registers

Surveys

Measurement surveys

Self report surveys

Non-health service: fire, police, social services department

[NB – poor definitions in non-health sources]

Research: synthetic estimates

Classifications:

ICD10

[OPCS4 coding for operations]

Read codes - a nomenclature not a classification

Epidemiology: how much do I need to know?

Clinical features (don’t overdo this)

Time (secular trend - last 50 years, more recent)
Place
Person

age, sex, socio-economic

ethnic, occupation, familial

lifestyle

Causes & determinants

THINK ABOUT SOURCES of knowledge (e.g. ‘CHD is declining’)

======

[Infections: covered later]

Neoplasms:

*Breast

*cervix,

*colon,

*lung,

*skin (melanoma and SCC)

Metabolic, endocrine:

*Diabetes mellitus

Psychiatric:

*Schizophrenia,

*dementia

*suicide,

deliberate self harm

Nervous system:

CVD

*CHD

*stroke

Abdominal aortic aneurysm

Respiratory

*asthma,

*chronic bronchitis

Trend summary E&W deaths 1990 – 1999 :

CHD, stroke, asthma, bronchitis: down [smoking]

Digestive:

Caries

peptic ulcer -> Helicobacter

Perinatal

SIDS

Congenital and hereditary

Down syndrome

Injury & poisoning

Falls

Epidemiology of lifestyle

*smoking,

*alcohol,

*sexual behaviour

*diet (obesity)

*exercise

Syllabus: ‘the effects on health of different diets (e.g. the ‘Western diet’), obesity, physical activity, alcohol, drugs, smoking, sexual behaviour and sun exposure

HEALTH ECONOMICS

Health economics: principles of health economics (including the notions of scarcity, supply and demand, marginal analysis, distinctions between need and demand, opportunity cost, discounting, time horizons, margins, efficiency and equity); assessing performance; financial resource allocation; systems of health and social care and the role of incentives to achieve desired end-points; techniques of economic appraisal (including cost-effectiveness analysis and modelling, cost-utility analysis, option appraisal and cost-benefit analysis, the measurement of health benefits in terms of QALYs and related measures); marginal analysis; decision analysis; the role of economic evaluation and priority setting in health care decision making including the cost effectiveness of Public Health, and Public Health interventions and involvement.

Perfect market

Elective surgery / Specialist psychiatry
Many sellers (and buyers)
Free entry (and exit)
Perfect information
Homogeneous product

No externality: I pay, someone else benefits (e.g. host purchaser / infrastructure costs)

Risk pools (Insurance systems)

1.Rare event

2.High cost

3.Population demand predictable

4.Individual's probability of demand independent

adverse selection

moral hazard

Economic appraisal

Measurement of COST

Marginal vs unit costs (and benefits):

e.g screening interval, change in admissions

Incremental cost

Opportunity cost

Direct vs indirect

Tangible (can invoice / bill for this) vs intangible (pain, suffering etc)

Discounting

future costs
?discount future health benefits
NICE recommends 3.5% annual discount for costs and health benefits

Cost effectiveness:

Cost minimisation – (e.g. to achieve no Hep B in drug users)

Sensitivity analysis

Cost utility:

Step 1: Assess health state after treatment using Quality of Life scale e.g. EQ5D

Step 2: Place a value (utility) on that health state

(e.g. on a rating scale score of 0 – 100)

Could use time trade off or standard gamble instead of rating scale

Disability weighting: see

Stouthard MEA et al. Disability weights for disease. Eur JPH 2000; 10: 24 – 30

Cost benefit:

Used by government to decide whether or not to go with a programme: overall cost to society

Air pollution clean up:

cost £785m - £1100m estimate for UK

12,000 - 24,000 deaths in 1996 (COMEAP)

Do costs outweigh benefits?

May need to value life:

“Gross output”

Willingness to pay

Pay to reduce road deaths

Pay for risky occupations

Willingness to spend (e.g. for a smoke alarm)

Decision analysis

Economic appraisal plus sensitivity analysis

May also involve decision tree

Also Richardson WS et al JAMA 1995; 273: 1292 - 5

Option appraisal

Where should paediatric cardiac surgery take place?

Efficiency: generally about getting the most out of your resources

Cost efficiency – no money wasted

Technical efficiency – no inputs wasted i.e. no kit, staff, standing idle

[Doesn’t work when you’re comparing different mixes of inputs and outputs

cf option appraisals]

Allocative efficiency – can’t give A more without taking from B i.e. no surplus

OR Technical – do CABG as cheaply as possible

Allocative – allocate funds for anti-smoking (achieves more / more efficient CHD reduction)

Equity

vertical: greater resource for greater need

horizontal: equal resource for equal need

Equality - of what?

Equal spend per person
Equal spend for equal need
Equal spend for equal benefit

SOCIAL SCIENCES

Sociology

‘study of individuals in groups and social formations’ (Lawson and Garrod) includes institutions

Organisations and management

Social identity – age class gender race

Family and friendship

Power and class (Marxism?)

Work including professions and status

Norms and deviance, discrimination

Social welfare, education etc

Qualitative methods

The principles of qualitative methods including semi-structured and in-depth interviewing, focus groups, action research, participant observation, and their contribution to public health research and policy; their appropriate use, analysis and presentation; the ethical issues which may arise; validity, reliability and generalisability; common errors and their avoidance; strengths and weaknesses.

Capturing qualitative data

ethnography

long interview

diary

analysis of documents and images

Qualitative analysis:

grounded approaches

semiotics (symbolism)

discourse analysis / repertoires

Rigour in qualitative studies:

Researchers' perspective (e.g. feminist)

Full description of fieldwork method

Subject selection

Recording (e.g. tape plus transcription)

Main results

Verbatim quotes

Exceptions noted (e.g. help seeking and masculinity)

Concepts of health, wellbeing and illness and aetiology of illness: the theoretical perspectives and methods of enquiry of the sciences concerned with human behaviour; illness as a social role; concepts of health and wellbeing; concepts of primary and secondary deviance; stigma and how to tackle it; impairment, disability and handicap; social and structural iatrogenesis; role of medicine in society; explanations for various social patterns and experiences of illness (including differences of gender, ethnicity, employment status, age and social stratification); the role of social, cultural, psychological and family relationship factors in the aetiology of illness and disease; social capital and social epidemiology.

Health care: different approaches to health care (including self-care, family care, community care, self-help groups); hospitals as social institutions; professions, professionalisation and professional conflicts; the role of clinical autonomy in the provision of health care; behaviour in response to illness and treatments; psychology of decision-making in health behaviour.

Epilepsy ‘from a sociological perspective’ ?

Concepts of health and illness

Culture and health beliefs: (your culture = your rules on how to eat / drink / etc)

Cultural beliefs about the body

Shape: beautiful baby competitions

Course Notes: Part I MFPHM Revision

EPIDEMIOLOGY

Epidemiological studies: design

Expressing the main result

Concepts and measures of risk

Interpreting the result

Chance

Bias

Confounding

Other problems

Effect modifiers [interaction]

Causation

Putting it together – guidelines and recommendations

Hierarchy of evidence

Surveys

HEALTH INFORMATION

Routine data sources

Population

Ad hoc censuses

Census based measures

Routine statistics

Epidemiology: how much do I need to know?

HEALTH ECONOMICS

Economic appraisal

Decision analysis

SOCIAL SCIENCES

Sociology

Qualitative methods

Concepts of health and illness