Discussion
Consensus was identified for most elements of the CPWC through the Delphi process. Where consensus could not be achieved, data analysis identified a ‘national best representative’ figure instead. For several activities relating to the delivery of pharmaceutical care the consensus on frequency was that ‘it depends on the patient’. This was explored in more detail by participants in round two of the Delphi study and allowed the generation of average frequencies of activities for the purposes of the CPWC being functional. The ‘experts in the field’ therefore contributed to determining the ‘Activity Standard’ for the WISN algorithm and in determining the ‘unavailable’ time for staff groups. The operator evaluation demonstrated the transferability of the CPWC to other operators, producing reliable and repeatable outputs.
The validity and reliability of a tool requires the consideration of a number of elements which are discussed in the context of the results of this study[13]. Content validity concerns the extent to which any tool addresses the full scope of the phenomenon being measured. This study achieved a national consensus on the tasks that are necessary for the delivery of pharmaceutical care, how long they take and how often they have to be performed, by which staff groups i.e. to establish an ‘activity standard’ for the delivery of pharmaceutical care. Content validity of the CPWC is therefore demonstrated through the consensus study data.
Two types of criterion validity are applicable to the CPWC; ‘concurrent’ and ‘predictive’. Concurrent validity compares a tool with an existing ‘gold standard’. The development of the CPWC followed the WISN[14] ] process and since this is the WHO ‘gold standard’ approach to calculation of workforce requirements, the CPWC could be considered a ‘gold standard’. There are no current ‘gold standard’ calculations for pharmacy workforce resource, since the 1997 ‘Purkiss Model’[25] no longer reflects current workforce requirements. Direct comparison of the CPWC with more recent literature (post-2010)[8,9,26] requires presenting staffing requirement in terms of the number of beds per pharmacist. This comparison (see Table 6) demonstrated that the output of the CPWC matched two of the three reference sources [8,9]. Its advantage over this previous work is the practical development of a simple to use workforce calculator, which can be applied in practice. The outlier in the comparisons the figure identified from National Health Service benchmarking 2015/16[26]. This suggests that many sites are delivering services with far fewer staff than the Activity Standard would suggest. What is unknown is the difference in patient outcomes associated with these different staffing levels and further work is required to determine this.
Construct validity is demonstrated if outputs of some elements of the tool can be correlated with values calculated by different methods[13]. This is particularly important if the tool being validated is theoretically novel, which is directly relevant to the validation of the CPWC and this has been demonstrated in several ways. Many of the timings that drive the algorithm of the CPWC have construct validity in comparison with the literature [27-32]. This is particularly relevant for medicines reconciliation (MR), which is the single longest task that needs to be completed for pharmaceutical care. It also has the greatest influence on the value generated by the tool as it is required for all patients and is associated with reduction in patient harm from medicines[5].
The consistency and reproducibility of the data generated by a tool represents its reliability. The two relevant types of reliability are considered in the context of the CPWC, namely equivalence and stability. The tool has to produce consistent measurements in the hands of two or more investigators to demonstrate equivalence. The ‘operator evaluation’ in this study demonstrated equivalence between users of the CPWC.
The results of this study have validated the CPWC for application in acute general hospital in-patients only and this is acknowledged as a limitation of the study. Application in community or mental health in-patient settings has not been demonstrated due to insufficient participant numbers from these settings. Likewise, the CPWC is also not validated for use in specialties, such as critical care. However, this study does demonstrate the value of applying the WISN approach to pharmacy practice and with sector or speciality-specific adaptions to the Activity Standard, the CPWC could be applied to clinical pharmacy services to sectors such as mental health or specialisms such as critical care.