University of Sussex
Browse

File(s) not publicly available

Why modelling a complex intervention is an important precursor to trial design: lessons from studying an intervention to reduce falls-related injuries in older people

journal contribution
posted on 2023-06-07, 14:53 authored by Sandra Eldridge, Anne Spencer, Colin Cryer, Suzanne Parsons, Martin Underwood, Gene Feder
Objectives: To develop a cost-effectiveness model of a complex intervention from pilot study data in order to inform the viability and design of a subsequent falls prevention trial. Methods: We used two models; the first estimated the probability of falling over a 12-month period based on a probability tree; the second used Markov simulation to assess the impact of the programme over time. Results: The first model indicated that our intervention would reduce the proportion falling by only 2.8% over a 12-month period. The major reason for this small effect was that less than a quarter of older people at risk of falling were assessed using our screening tool. Even if policy-makers were willing to spend £30,000 per quality-adjusted life-year gained, there is only a 40% chance that the intervention would be cost-effective. Sensitivity analyses showed that the only scenarios that produced a substantial increase in the effect of the intervention were those in which all older people are assessed. Conclusions: The model-building approach described in this paper is vital when designing complex trials and where a trial is not possible. Information from the modelling can be used to re-design the intervention. The effectiveness of our proposed intervention appears very small due to its inability to reach those at risk of falling. It is most likely not to be cost-effective. If inability to reach the target group is a weakness common to other similar interventions, this suggests an area for further research.

History

Publication status

  • Published

Journal

Journal of Health Services Research and Policy

ISSN

1355-8196

Publisher

SAGE Publications

Issue

3

Volume

10

Page range

133-142

Department affiliated with

  • Primary Care and Public Health Publications

Full text available

  • No

Peer reviewed?

  • Yes

Legacy Posted Date

2008-10-21

Usage metrics

    University of Sussex (Publications)

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC