Overview
Initiative type
Service Improvement
Status
Deliver
Published
June 2025
Summary
The project team built a triage tool in Microsoft Power BI to flag high risk patients requiring pharmacist clinical review that changed the entire department workflow to a “roaming” service from the traditional ward-based model.
Dates: 19/02/2024 -
Implementation sites: Mackay Hospital
Partnerships: Pharmacy Department
This project was presented as a Poster at CEQ Showcase 2025 (PDF 435KB).
Aim
Identify medical patients that are high value for pharmacists to clinically review.
Outcomes
- Service provision changed from ward-based to an objective triage system, where multiple pharmacists may be at the same ward to review patients in parallel
- High-risk patients seen earlier in their episode on average (22% reduction from 43.41hrs to 33.71), with 11% increase in count (279 to 311) post project
- Medium-risk patients seen earlier by median (29% reduction from 41 to 29) but no different on average (1.29% reduction). More medium-risk patients seen by 16% (325 to 376)
- Low-risk patients seen later in the episode on average (37% increase from 49 hours to 68), reducing in count by 51% (457 to 225)
- Safe: No change in reported risk incidents between periods
- Removal of patient geographical disadvantage
Background
Staff Deficit: The Pharmacy department faced a 33% staff deficit and needed to review the current ward-based patient care model. The department will not be able to provide service delivery to high-risk medical patients on such a large staff deficit with this model.
Management was facing a challenging decision of which medical ward over the others would be serviced. Prioritisation is time-consuming: Prioritisation for pharmacists is a very time-consuming process, taking around 1.5 hours per day per pharmacist to complete1,2. This tool assists with identifying patients where pharmacists can add significant value, and those less likely to require input without opening the patient chart or flipping through numerous Patient Flow Manager handover sheets. Pharmacist involvement is subjective: Guidelines are available to assist pharmacists with which patient cohorts to review and triage, but lack of process standardization results in a range of approaches.
A tool developed by senior pharmacists from varying medical backgrounds created a consistent triage that is simple to use. Ward-based model limitations: Mackay Base Hospital did not service all wards. As an example, patients at Paediatrics, Day Surgery Unit would not receive a pharmacist clinical review as these were non-serviced wards. Limited resourcing for the rural sites where there is no pharmacy service resulted in no clinical oversight of high-risk patients by a pharmacist and were “flying blind”. Staff burnout/fatigue: Culture that all patients on the ward need to be clinically reviewed to “clear the list”.
Methods
A literature review was conducted that identified several journal articles that have reviewed or developed pharmacist triage tools utilising electronic medical record or automated dispense cabinet data. A paper that could be quickly adapted locally was selected and several workshops with a panel of pharmacists consisting of three HP5s, one HP6 and one HP3 developed the RAT’s 31 indicators2. The data source used is Central Business Intelligence’s Operation Data Store using SQL to extract ieMR data. This includes data fed to it by AUSLAB and HBCIS. Each indicator is assigned a score with a higher score indicating a higher clinical priority. All indicators per patient are summed to provide a final score for the current datetime. The score categorises patients into high, medium or low risk based off the score percentile for each MHHS facility. High Risk is a percentile >=75%, Medium >=50% and <75% and Low risk < 50%.
The project team decided to use a percentile to identify the patient categories instead of a static score (eg. patients over 30 are high risk) as this would move the cohorts up and down depending on hospital medical acuity, and it removed the need for the panel to choose score thresholds for each risk category. The 30 minutely update of the patient’s score allowed it to change over the inpatient episode so low risk patients who changed to high risk would be easily identified. Appendix 1 has more information about each indicator.
A three-month pilot period was established to continually review and validate the data compared to the manual process. The tool was reviewed by pharmacists on the weekends for feedback and prototype usage. Weekends were consequently removed from the pre and post period comparison. User interface design principles were used to allocate a traffic light system with green for “go” for high-risk patients and red for “stop” for low-risk patients.
Continual feedback process to remove unnecessary information that overwhelmed the user assisted in simplifying the tool. Filters and page views can be saved for simple navigation assisted with usage and quality of life. The process of service model change
commenced on 19/02/2024 once the department was comfortable using the tool and it was at a refined state from continual improvement.
The service model involved a team huddle in the morning and after midday where a team leader with the team would review the
high and medium risk patients and the work split accordingly. The team members were requested to review their allocated patients in order of triage, and more high-risk patients reviewed when their list completed. Feedback from staff showed promise that the team-based case discussion and allocation instead of siloed ward work improved morale, comradery and workload distribution. The project team wanted to determine whether the RAT assisted with reducing the time spent reviewing patients in the triage process, so the “sentinel” reporting of users opening patient charts and not documenting activity within an hour of opening the chart was reviewed.
Discussion
This project required a strong project lead in a position of authority, who also works on the floor to drive the change. The Director of Pharmacy is very interested in the success of the RAT and was able to provide on the floor adoption and assist with change management to the department. Feedback from the department about the process showed that the tool was well received because it was evidence based, and local workflows were considered in the development of the tool. The indicators were developed collaboratively where wider feedback was sought from the department where the members felt heard. This engagement with the entire team and setting expectations that the tool will not be perfect initially but will be developed from the team’s feedback assisted with the adoption process.
One of the most significant limitations is heavy reliance on the availability and accuracy of data in our ieMR system. If patient data is not fully captured in these systems, some patients with chronic or complex histories might be under-prioritised. The design of “row level security” excluding other HHS data from our dataset proved to be a massive barrier to obtaining recent patient data. Other source systems such as MOSAIQ, METAVISION, CIMHA, SCRIPTRAKER were not included due to performance limitations and may be opportunities for further development. The RAT could experience refresh failures leading to data delays, data hosting issues resulting in tool downtime and required a business continuity plan to be developed. The tool is designed based on a panel of pharmacists’ input. Its subjective nature could bias how patients are prioritised.
Further validation is needed to improve inter-rater reliability. Other parallel system changes, such as alterations in ieMR documentation processes, code upgrade could also influence the tool's efficiency gains. Clinical risk was reviewed using RiskMan incident comparison during the periods, and did not show any difference in reported incidents. There is an opportunity to develop specialty versions of the tool, particularly for mental health and palliative care settings, which would require customised indicators to serve these patient groups better. Ongoing evaluation and data collection will help refine the tool further, ensuring it continues improving patient outcomes while maintaining safety and efficiency standards. The recent completion of ieMR rural roll out shows opportunities for a pharmacist at MBH to provide pharmacy reviews remotely to the facilities without a pharmacy service. Other Queensland hospitals, including Townsville, have expressed interest in adopting the RAT, while The Sunshine Coast University Hospital and West Morton Health have already implemented it for trial.
The RAT has sparked innovative ideas where similar triaging tools could be developed and implemented in other departments such as Dietetics. There is also interest in embedding the RAT state-wide into the ieMR system in the Pharmacy Care Organiser. The RAT is “free and open”, and the tool has had national interest from New South Wales, South Australia and across Queensland. The project team used the Process Improvement cycle for this project.
References
1. Society of Hospital Pharmacists of Australia. (2013). Chapter 9: Staffing levels and structure for the provision of clinical pharmacy services. Journal of Pharmacy Practice and Research, 43(suppl), S32-S34.
2. Falconer, N., Nand, S., Liow, D., Jackson, A., & Seddon, M. (2014). Development of an electronic patient prioritization tool for clinical pharmacist interventions. American Journal of Health-System Pharmacy, 71(4), 311-320. https://doi.org/10.2146/ajhp130247
Key contact
Robert Knights, / Ron Nightingale / Neve Munro
Clinical Technical Informatician / Director of Pharmacy/ Acting Clinical Educator Pharmacist
Mackay Hospital and Health Service