Epidemiological surveillance of frailty
A population-wide platform of frailty indicators for public health decision-making
Turning big data on frailty into actionable intelligence for public health
A public health agency with large-scale longitudinal data on frailty in people over 65, but without the analytical or visualisation capacity to turn them into evidence for public policy.
The client had a systematic frailty screening system in primary care, with five years of data on more than one million people over 65. However, it did not have the analytical and visualisation capacity needed to turn that volume of data into actionable epidemiological intelligence for policy decision-makers, clinicians and health service managers.
The challenge was to process large-scale population data with multiple levels of stratification (sex, age group, health district, primary care area), apply rigorous epidemiological standardisation to allow comparison across areas and with other countries, and translate methodological complexity into intuitive visualisations for non-technical audiences. All of this with reproducible pipelines that would allow automated annual updates.
>1M
people over 65 with longitudinal follow-up
5 years
of data with systematic classification in primary care
4 indicators
key epidemiological measures, with multiple stratification
An interactive web platform for epidemiological surveillance
Automated processing of large-scale population data
Handling of data on more than one million people with multiple longitudinal measurements. Classification into four frailty categories using an electronic frailty index derived from health records (for example, e-SIF or eFI). Automated calculation of prevalence, incidence, mortality and reversibility, with multiple stratification and direct age standardisation to the European Standard Population, to make comparison easier between areas and with European reference figures.
Interactive visualisation for different user profiles
A hierarchical structure by year with intuitive navigation. Epidemiological tables with 95% confidence intervals, charts stratified by sex and age, time series of year-on-year trends and geographical heat maps to spot hotspots visually at health-district level.
Reproducible pipelines for annual updating
Built in R with the Quarto framework to generate a responsive static website. Reproducible pipelines that allow automated annual updates with new surveillance data. Automatic item-by-item text descriptions of the screening, with distributions by sex, age and area.
Epidemiological intelligence for policies on care for older people
Identification of key epidemiological patterns
A marked age gradient, with prevalence above 25% in people over 85. Consistently higher prevalence in women across all age groups. Marked geographical differences, with health districts and primary care areas above the average that required closer monitoring and evaluation.
Geographical prioritisation of resources through maps
The maps make it possible to identify areas with a higher burden of frailty and provide evidence to support geographical prioritisation and the planning of home care services, prevention programmes and care resources.
Epidemiological evidence within reach of decision-makers
A web platform that puts epidemiological evidence within reach of policy decision-makers, primary care clinicians, managers and legislators, without the need for technical intermediaries. Rigorous age standardisation allows fair comparison between areas with differing population pyramids.
Project carried out under a confidentiality agreement. The client's identity has been withheld at their request.
What this case shows about the way we work
Epidemiological big data with complex stratification
Processing of large-scale population data across multiple simultaneous levels of stratification, without loss of statistical robustness.
Rigorous epidemiological standardisation
Implementation of direct age standardisation methods that allow fair comparison between heterogeneous areas and with international reference figures.
Visualisation for non-technical audiences
Translating epidemiological complexity into intuitive visualisations that policy decision-makers can act on.
Reproducible web platforms for surveillance
Public health intelligence systems that put epidemiological evidence within reach of several user profiles, sustainably over time.
Shall we talk?
Tell us what you are working on and we will give you a straight answer on how we can help.