How can we measure the complexity of general practice consultations?
We have previously demonstrated that GP consultation rates are increasing. However, the complexity of consultations may also be increasing due to the rising prevalence of multimorbidity, transfer of work from secondary to primary care, the requirements of the QOF and increasing policy and public expectations. If so, this will further compound workload pressures in general practice. In order to assess the complexity of GP consultations we need a valid and reliable measure which can be applied to routine medical records, but no suitable measure currently exists. The aim of this study was to develop such a measure.
We conducted a Delphi study over two rounds involving 32 experienced GPs to identify potential indicators of consultation complexity. We created initial Read code sets for each of the endorsed complexity indicators; the code sets were assessed independently by two GP academics with a third involved if disagreement. We excluded indicators which could not be clearly defined or with very low prevalence. We are currently assessing the face validity of each indicator by exploring its relationship with consultation duration, using a 10% age-sex stratified random sample of 304,937 people and 1.7 million consultations from the Clinical Practice Research Datalink (CPRD). We will then use factor analysis to combine the indicators into separate measures of patient and consultation complexity. Finally, using CPRD we will explore how complexity varies in different populations and how it changed between 2007 and 2018.
We found it necessary to distinguish between indicators of complex consultations (consultations involving complex problems) and complex patients (enduring patient characteristics which make most of their consultations complex irrespective of the presenting problem). Between rounds 1 and 2 of the Delphi several potential indicators were re-worded in the light of panel members’ comments. After two rounds the panel endorsed 17 of a possible 19 indicators of consultation complexity (e.g. consultation raises child protection concerns) and 17 of 26 indicators of patient complexity (e.g. polypharmacy). The panel sometimes found it difficult to define what makes a patient or consultation more complex than others when the norm was already complex. Findings from the work to combine indicators and to validate the patient and consultation complexity scores will be presented, along with how complexity has changed over time.
Developing measures of complexity is challenging both conceptually and practically, but necessary given the changing demands on general practice in the UK and internationally. Valid and reliable measures of patient and consultation complexity are important for research (e.g. as a case-mix measure) and for policy, with implications for how general practice is organised and resourced, for example to inform decisions about appointment lengths and how funds are distributed between practices.