Transformation Directorate

Predictive analytics to assess risk and trigger care interventions


Problem to be solved

The model of visiting and supporting individuals at regular times in the day is long-established in domiciliary care. But identifying and acting on care-related issues and possible signs of deterioration in individuals is highly dependent on an individual care worker’s observations, and their ability to promptly escalate those observations.

This can be challenging when more than one care worker is responsible for an individual. And in the context of workforce shortages across social care, there can be additional challenges with inexperienced carers and carers with heavy caseloads struggling for time. Regardless of an organisation’s workforce position, the potential benefits to care outcomes of more immediate and consistent risk assessments are clear - early, proactive and targeted intervention works.

The project

Cera Care is a CQC registered technology-enabled home care company, providing approximately 10,000 care visits per day to individuals across 50 local authorities in England. The company’s AI-powered Concern Predictor tool analyses data collected from care workers’ visit reports to produce a stratified risk assessment of individuals across a region, predicting the likelihood of falls and hospital admission.

Where an alert is triggered, regional service managers compare the automated assessment with the care worker’s written report to fully understand the situation and make appropriate decisions about individuals’ care needs, intervention measures and escalation to other agencies. 

Trained on 68,000 annotated care records, the tool analyses both:
  • structured data e.g. specific health observations
  • free text from care workers’ visit logs, using sentiment analysis 

to predict a level of risk for a given individual.

On any given day, the Concern Predictor analyses data on between 450 and 500 individuals. Between November 2019 and June 2020, the Concern Predictor has highlighted 715 risks of adverse instances, which managers and care workers have then proactively addressed.

To find out more about this project, contact Mahiben Maruthappu using email address:


Explore Further