Data collection
For the first part of the study, consensus was electronically sought on the reference data in the CPWC algorithm. In Round one the Expert Panel was asked to identify from the locally-generated list of suggested ‘clinical pharmacy tasks’ those which they believed to be necessary for individual patient care, how long they perceived each task typically took (this could be provided as existing local data, data collected at the time to answer the question or ‘best guess’) and how often it would be necessary to complete each task for each patient admission. In addition, they were asked to identify for a range of staff groups the time in their employed hours that was typically unavailable for patient care. This included activities such as annual leave, sick leave, training, and travel. Round one responses were collated and anonymised and sent to each participant together with their own response in Round two for reconsideration in light of the Panel’s collated responses. The Panel were asked additional questions in Round two to gain clarity on elements where consensus was not achieved in Round one, or where responses required conversion from narrative into numerical values to allow application in the algorithm. This was achieved by asking participants to respond to exemplar patient scenarios in terms of activities undertaken for patients by day of admission.
The second part of the study, the ‘operator evaluation’, was issued to the Delphi Panel participants. They were given instructions and asked to use the CPWC to calculate answers to the three staffing scenarios shown in table 1 and return them for analysis.