Lifestyle data

1
Lifestyle data
2

• What we will cover in this session
• what we mean by lifestyle
• the need for local lifestyle data
• sources of lifestyle data and their comparative
  strengths and weaknesses

3
Lifestyle?

• Smoking
• Physical activity
• Diet
• Alcohol consumption
• Use of drugs
• Sexual behaviour
• Obesity
• Sexually transmitted infections
• Teenage pregnancy
• Breastfeeding
• Blood pressure
• Cholesterol levels

4
Question

• Whats the best way to estimate the percentage
  of people in your local population who smoke?

5

• Potential data sources
• national surveys
• synthetic estimates based on national surveys
• local surveys
• data from primary care
• data from commercial/market research
  organisations

6
different purposes...different aspects of
lifestyle...different sources...

• Which lifestyle factor?
• Smoking
• Physical Activity
• Diet
• Obesity
• Alcohol consumption
• What do we really want to know?
• Between area comparisons
• Within-area comparisons
• Trends over time
• Evaluating local services/initiatives

7
The need for local lifestyle data - general

• prioritising, planning, targeting and evaluating
  local services and lifestyle change initiatives
• data at LA/PCT-level AND below for ...
• comparison with other LAs/PCTs
• within-area comparisons by age, gender,
  ethnicity, ward, practice, etc
• analysis of trends over time and progress towards
  local targets
• measuring the outcomes/impacts of services

8
The need for local lifestyle data - specific

• Local Operational Plans (LOPs)
• based on the Vital Signs indicators
• Local Area Agreement (LAAs)
• based on the National Indicator Set (NIS)
• Joint Strategic Needs Assessment (JSNA)
• Vital Signs and NIS lifestyle indicators include
• child obesity
• smoking quitters
• hospital admissions for alcohol-related harm
• adult physical activity levels
• school PE and sport provision
• breastfeeding

9
The need for local lifestyle data (cont.)

• helps to communicate important public health
  messages ...
• ... to the public/local communities
• ... to decision makers/commissioners/funders

10
Some criteria for assessing and comparing
different sources

• Validity
• Reliability
• Accuracy
• Bias
• Precision
• Timeliness
• Cost

• Does it measure what it aims to measure?
• Do questions produce reproducible answers?
• Measurement or data entry?
• Is the sample representative?
• Adequate sample size?

11
Data from primary care

• records of consultations held on practice
  computers
• incentives - the QOF system
• GP research databases

12
The Quality and Outcomes Framework (QOF) is a
system of financial rewards to general practices
for the provision of high quality care
Obesity The practice can produce a register of
patients aged 16 years and over with a BMI
greater than or equal to 30 in the last 15
months.
Smoking 1 The percentage of patients with any or
any combination of the following conditions
coronary heart disease, stroke or TIA,
hypertension, diabetes, COPD or asthma whose
notes record smoking status in the previous 15
months (except those who have never smoked where
smoking status need only be recorded once since
diagnosis).
Smoking 2 Ditto ... but where the notes contain
a record that smoking cessation advice or
referral to a specialist service, where
available, has been offered within the previous
15 months.
13
Data from primary care

• Advantages
• Standardised national systems.
• Ongoing data collection
• Established coding systems.
• Disadvantages
• Selective focus on particular patient groups
• Variability between practices in the completeness
  and quality of data recording.
• Lack of a geographical focus

14
Data from commercial organisations

• e.g. CACI, Claritas, Experian, Acxiom, Dr Foster
• large volumes of household survey and consumer
  data modelled to provide estimates for all areas
  of the country of e.g.
• expenditure on tobacco, food and drink
• prevalence of smoking and obesity

15
Acxiom smoking prevalence estimates
Data from commercial organisations e.g.

• based on the National Shoppers Survey
• large national sample - some coverage in your
  area
• updates available annually
• adjusted for known biases in the sample e.g.
  undersampling of young people.
• relatively cheap
• 2005 data gave a national prevalence of 17
  compared to 23-24 from 2004-5 GHS and 2004 HSE

16
Data from commercial organisations

• Advantages
• Extensive geographical coverage
• Disadvantages
• the main problem is that detailed methodologies
  are often not available
• Potential for bias (lack of random sampling /
  poor response rates)
• Cost expensive?

17
National surveys

• Health Survey for England
• General Household Survey / Integrated Household
  Survey
• Smoking prevalence available at PCT level from
  2010
• Smoking, drinking and drug use
• Tellus

National surveys are often not large enough to
provide data below regional level data
aggregation / local boost samples may be
options....
18
National surveys local boosts

• Advantages
• Flexibility.
• Robust methodology.
• Comparability with national, regional and other
  benchmarks.
• Disadvantages
• Lack of local historical trend data.
• Costs

19
Synthetic estimates from national surveys (public
domain)
20
Model-based estimates of healthy lifestyle

• IC-funded project
• HSfE 2003-2005 data on smoking, obesity, binge
  drinking, fruit and veg consumption
• statistical modelling to identify social and
  demographic predictors of these aspects of
  lifestyle
• LA- and MSOA-level estimates (and CIs) based on
  the social and demographic characteristics of
  their populations.
• validated against other survey data, e.g. the
  Merseyside boost of the HSfE

21
Example LA-level smoking model
Model-based estimates of healthy lifestyle

• Area characteristics statistically associated
  with prevalence of smoking
• high households with no car
• low of people 16-74, professional managerial
  occupations
• high males, 16-34, White ethnic origin
• low life expectancy (females)
• low with limiting long-term illness

22
Model-based estimates of healthy lifestyle

• Advantages
• Comprehensive geographical coverage - allows
  comparative analysis between areas. 
• Ease of access/cost. The estimates are in the
  public domain, are immediately accessible to
  users and are free of charge.
• Disadvantages 
• Lack of flexibility particular local
  requirements cannot be specified.
• Lack of sensitivity to local lifestyle
  interventions. - represent expected prevalence
  so additional local factors that may impact true
  prevalence are not taken into account, so should
  not be used to measure change over time.

23
Local surveys (including regional surveys)

• Advantages
• generates real local data
• flexibility and control over e.g.
• population to be surveyed - area, age, sex,
  ethnicity, etc
• sample size - trading statistical power v cost
• survey design - cross-sectional/longitudinal,
  census/sample, etc
• method - phone, internet, interview, postal
• subject matter - can be anything!!

24
Local surveys (including regional surveys)

• Disadvantages
• lack of standardisation of questions and derived
  indicators
• lack of comparability with other areas/benchmarks
• securing permissions
• labour intensive
• cost

25
smoking prevalence in your local authority
area rate the different sources for different
applications