Discussion

Overall pharmacists were very confident about using PAST, and agreed that it helped to assign an acuity level and monitor the pharmaceutical needs of inpatients. However, pharmacists appear to rely more on professional judgement than the tool itself to assign a PAL, regardless of experience. The differences in PALs assigned to the case study patients gave a good illustration of the variations between pharmacists and likely use of professional judgement.

In four of the case studies, 10/28 (36%) of pharmacists allocated a higher level than recommended in the guidance tool and in three of the case studies, 9/28 (32%) pharmacists allocated a lower PAL, the research team found the same when they assessed the allocation of PALs in daily practice.7 The departmental pharmaceutical care standards state that a L3 patient should be seen every weekday by a senior pharmacist, a L2 patient should be seen by any pharmacist two or three times a week, and a L1 patient can be managed by a pharmacist or pharmacy technician unless medication or clinical condition of the patient changes.

All case study patients had different comorbidities and in three of the case studies the patients had one decompensated organ and were not on any high-risk medications, so would have been expected to have been classified as L2 using PAST. The different levels allocated for patients may again reflect the professional judgement of the pharmacists and experience on what they feel comfortable managing.

Experienced pharmacists may be allocating a lower PAL to a patient as they feel comfortable managing the patient at a lower level. In comparison, less experienced pharmacists may professionally feel less confident and want to assign a higher PAL to ensure the patient is seen more frequently and by a more experienced pharmacist.

The most notable variation in PAL choice was between the two L3 cases, 26/28 pharmacists (93%) correctly allocated case study 4. The patient had one decompensated organ (the brain), active treatment for prostate cancer (goserelin) and he was also ‘nil by mouth’. However, in case study 6, 21/28 pharmacists (75%) assigned a lower level (L2) than PAST would have recommended (L3), this patient had two decompensated organs (the lung and kidneys). The difference may be due to the perceived severity of the decompensated organs and what pharmacists feel more competent to manage due to experience (eg, community acquired pneumonia and acute kidney injury are pharmaceutically managed more commonly than a subdural bleed and active prostate cancer).

In the other major deviation, 24/28 pharmacists (86%) assigned case study 2 a higher PAL (L2) than recommended in PAST (L1). This patient had been diagnosed with a urinary tract infection and had been prescribed trimethoprim. The medical history of the patient included chronic back pain, angina, epilepsy and dementia, but the medication regimen was stable and included sodium valproate and regular codeine. Codeine is excluded from the high-risk medicines category as it is only a weak opiate but it is possible that the majority of pharmacists identified sodium valproate as a drug requiring therapeutic drug monitoring (even though serum levels are only used to detect non-adherence) thus allocating the patient the L2 status.

Overall the strength of agreement was ‘fair’, indicating that there was not a true agreement with PAST.8 This may reflect the overestimation and underestimation of allocated PALs seen in the case studies and again emphasises the pharmacist’s use of professional clinical judgement when assigning a PAL. Exploration of the agreement between PAL extremes for the different levels of pharmacy experience showed that the strength of agreement was ‘poor’. This suggests that the majority did not agree with the structured PAL guidance allocation for L2 patients. The overall results from this and our previous study7 indicate that the criteria for all the levels of the patients are clearly in need of further review, which should encourage the tool to be used more consistently in practise.

Confidence in an assessment screening tool may also be improved by ensuring the tool is fully validated to detect adverse drug events such as that designed by Urbina and colleagues in a Spanish hospital pharmacy department.9

The open text answers provided further insightful views about the limitations of the PAST. The most prominent limitation is that the PAST is at times used as a system to remind pharmacy team members that they need to follow-up a patient, especially on wards where there is not a regular pharmacy technician and the review may only involve a supply of medication. There was also a lack of confidence about how up-to-date the PAL was on the ward’s patient identification boards after admission. This lack of confidence could mean that patients who do not need to be reviewed regularly are being reviewed, while patients who have become acutely unwell during their hospital admission are being overlooked.

Despite confident about using PAST, one-third of the respondents believe more training is needed to use it more effectively. Cottrell et al found that the usefulness of a pharmacy risk screening tool required substantial involvement by the pharmacy staff3 and Hickson et al 7 suggested that reiterating the true purpose of the tool could improve its reliability.

The strength of the study was that the questionnaire had an excellent response rate, which instils confidence that the results reflected the whole department and not solely either junior or senior pharmacists. A limitation is that the questionnaire results only reflect the attitudes of pharmacists at one hospital and so its application elsewhere may not be valid. Deeper insights into the use and understanding of the tool by pharmacists may also have been possible using more searching qualitative research methods, such as focus groups or interviews.

For future research, it would be beneficial to conduct intensive staff training after redesigning and validating the tool and then reassess adherence to it.