Skip to content

Workforce predictive risk modelling: development of a model to identify general practices at risk of a supply-demand imbalance

Research output: Contribution to journalArticle

Original languageEnglish
Article numbere027934
Number of pages12
JournalBMJ Open
Issue number1
DateAccepted/In press - 1 Nov 2019
DatePublished (current) - 23 Jan 2020


Objective. To develop a risk prediction model identifying general practices at risk of workforce supply–demand imbalance

Design. Secondary analysis of routine data on general practice workforce, patient
experience and registered populations (2012 to 2016), combined with a census of general practitioners’ career intentions (2016).

Setting/Participants. A hybrid approach was used to develop a model to predict workforce supply–demand imbalance based on practice factors using historical data (2012–2016) on all general practices in England (with over 1000 registered patients n=6,398). The model was applied to current data (2016) to explore future risk for practices in South-West England (n=368).

Primary outcome measure. The primary outcome was a practice being in a state of
workforce supply–demand imbalance operationally defined as being in the lowest third nationally of access scores according to the General Practice Patient Survey and the highest third nationally according to list size per full time equivalent GP (weighted to the demographic distribution of registered patients and adjusted for deprivation)

Results. Based on historic data, the predictive model had fair to good discriminatory ability to predict which practices faced supply–demand imbalance (area under the ROC curve 0.759). Predictions using current data suggested that, on average, practices at highest risk of future supply–demand imbalance are currently characterised by having larger patient lists, employing more nurses, serving more deprived and younger populations, and having considerably worse patient experience ratings when compared with other practices. Incorporating findings from a survey of GPs career intentions made little difference to predictions of future supply–demand risk status when compared with expected future workforce projections based only on routinely-available data on GPs’ gender and age.

Conclusions. It is possible to make reasonable predictions of an individual general practice’s future risk of undersupply of general practitioner workforce with respect to its patient population. However, the predictions are inherently limited by the data available.

    Research areas

  • primary care, workforce, supply-demand, organisation of care

Download statistics

No data available



  • Full-text PDF (final published version)

    Rights statement: This is the final published version of the article (version of record). It first appeared online via BMJ at Please refer to any applicable terms of use of the publisher.

    Final published version, 1.89 MB, PDF document

    Licence: CC BY


View research connections

Related faculties, schools or groups