Related Experiment Video
Updated: Jun 3, 2026
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Using routinely collected data to evaluate a leaflet campaign to increase the presentation of people with memory
Tom Chan1, Jeremy van Vlymen, Neil Dhoul
1Division of Population Health Sciences and Education, St George's-University of London, London, UK.
Background:
The Alzheimer's Society wished to raise awareness that people with memory problems may benefit from early assessment and diagnosis, so that appropriate measures could be put in place and management improved.
Objective:
To use routinely collected data to determine whether a leaflet campaign to raise awareness of memory problems would result in increased presentation of people with memory problems to their GPs.
Method:
A locality was identified which met the criteria for locating the pilot intervention. A neighbouring locality was identified which used the same secondary care service and could serve as a comparator. Anonymised routinely collected computer data were gathered before and after the intervention.
Results:
The intervention locality had a much greater proportion of elderly patients and a higher proportion had memory problems recorded at baseline (OR 1.67; 95% CI 1.47-1.91; P<0.001). In both localities just under 40% of people with memory problems had blood tests. Approximately 80% would be referred to secondary care, and this was more likely for those in the intervention group (OR 1.29; 95% CI 0.99-1.93; P=0.044). However, the use of antidepressants was greater in the control locality; 34% vs 9% (OR 0.19; 95% CI 0.13-0.27; P<0.001). Whilst the absolute number of people prescribed cholinesterase inhibitors was greater and increased more in the intervention practices, the proportion of people with memory problems prescribed was not significantly greater (OR 1.21; 95% CI 0.77-1.89; P=0.38). The increased prescribing in the intervention practices was due to people restarting therapy. From a lower baseline there was a greater increase in the control locality for all variables for which we had a before and after measure.
Conclusions:
During a leaflet campaign the recording and management of memory problems increased. However, there was greater improvement in the control locality. This study demonstrates the importance of including a control group and the strengths of routine primary care data.
