Healthcare Redesign of Medication Management for Parkinson’s Inpatients

Abstract

Parkinson’s disease is a progressive neurological disorder reliant on medication regime adherence to alleviate symptomology. When hospitalised, people with Parkinson’s disease have specific medication management needs which are consistently unmet. This study aims to develop, implement and evaluate solutions for improving the medication management of inpatients with Parkinson’s disease. A healthcare redesign approach was utilised, focusing on the final three phases: solutions design, implementation and evaluation. Five solutions were derived: formalise routine patient identification, provide improved staff education, develop and install automated prescriber alerts, review and amend ward PD medication stock, and develop systematic prompts for PD medications. The findings suggest that our solutions sustainably improved systems and processes that contribute to quality and safe medication management for Parkinson’s patients. Correct identification of Parkinson’s patients within an acute care hospital leads to correct prescription of medications, timeliness of medication administration and timely pharmacy review. The length of stay was not positively impacted.

1. Introduction

Parkinson’s disease (PD) is a chronic, progressive, and complex neurodegenerative disorder that primarily affects movement [1]. Patients require individualised medication regimes to provide adequate symptom control [2,3]. Medication management for hospitalised PD patients can be challenging for pharmacists and other members of the treating team. People living with PD generally have differing symptom types and severity with complex medication regimes to manage their disease [4]. PD patients require additional time and consideration from experienced clinicians to ensure that the right medication, at the right dose and at the right time, is delivered. Levodopa, prescribed routinely for people living with PD, is a dopamine precursor associated with improved motor function and fewer side effects than other medications [5]. Dosage timing for levodopa is critical, and it must be administered correctly to ensure adequate symptom control. Current evidence suggests that medication errors are common. Poor prescribing and delayed medication administration for PD inpatients has been linked to increased length of stay in hospital, extending up to 4 additional days [6,7]. It has also been shown to increase adverse events for patients and their caregivers [8], with up to 20% of patients receiving a contraindicated dopamine blocker in one study [6] and 71% of patients missing dosages of medication in another [6,8]. Several factors may contribute to mismanagement of PD medications, including patient status and care coordination [9]. Investigating the systems and processes that contribute to poor PD patient experience and outcomes is critical to improving patient safety.

Healthcare redesign is an approach to change that is utilised to develop and implement meaningful and sustainable solutions to identified, evidence-based issues [10]. This approach includes five phases: scoping, diagnostics, solutions development, implementation, and evaluation. Evidence-based practice is the use of current best evidence, combined with clinical expertise and patient values, to guide healthcare decisions and service provision [11]. To address the ‘know-do-gap’ between best evidence and current practice, knowledge translation requires not only synthesising research evidence but also communicating it to stakeholders [12,13]. This is an interactive process that engages the experience and needs of multiple stakeholders. A previous study [14] explored the scoping and diagnostics phases of healthcare redesign, focused on the Parkinson’s Inpatient Experience One project (PIE1). This study identified five key issues contributing to poor medication management during admissions to the hospital for PD patients. Poor identification of PD patients, untimely pharmacy reviews, prescribing errors, untimely administration of PD medications, and staff’s limited knowledge of PD medications all contributed to a poor experience and poorer clinical outcomes for PD patients [14]. Therefore, the aim of this study was to develop and implement solutions for the PIE1 project and evaluate the impact of these solutions in practice.

2. Materials and Methods

2.1. Setting

Royal North Shore Hospital (RNSH) is a principal tertiary referral hospital in the Northern Sydney Local Health District (NSLHD), New South Wales (NSW), Australia, serving an estimated PD population of 5648 in 2024. The hospital uses two electronic medical record (EMR) systems, PowerChart throughout the hospital and FirstNet in the Emergency Department (ED). There are three main locations that service PD patients: the neurology ward, the elderly care ward, and the ED.

This study utilised the healthcare redesign approach, specifically the solutions design and implementation and evaluation phases (Figure 1). Each phase is outlined below. A pragmatic design approach using mixed methods was utilised to design solutions and evaluate outcomes [15,16]. Mixed methods incorporate elements of both qualitative and quantitative research [16].

2.2. Solutions Design Phase

One of the major components of a successful solutions phase of healthcare redesign is engagement with key stakeholders [10]. Stakeholder mapping identified key people to be engaged and empowered throughout this phase (Appendix A), including medical, nursing, and allied health professionals, administrators, and leaders. Solutions design involved clarifying the key issues and the root causes and prioritising these issues based on impact and feasibility through a series of workshops and meetings. Prioritised issues were then confirmed with key stakeholders, and a range of engagement and collaboration strategies were utilised to develop evidence-based solutions.

The prioritised issues and their root causes are presented in

Table 1. Key issues and root causes for the solution phase.

Key Issue	Root Cause
PD patients are not readily identifiable to hospital staff	Staff are unaware that the PD alert icon in FirstNet (eMR system used in ED) exists and what it achieves
	Staff are unaware of how to set up the PD alert icon in FirstNet
	No alert system in PowerChart for identification of PD patients on the ward
	No bedside (non-electronic) means of identification
Pharmacist reviews of PD patients are not conducted in a timely fashion	Insufficient funding to increase pharmacist-to-patient ratios
	Pharmacists have competing priorities
	PD patients are not readily identifiable to pharmacists in the ED due to poor use of the electronic PD alert icon
	No alert system exists in the eMR for the identification of PD patients on wards
Staff lack knowledge about PD medications	Lack of education on PD medications
PD medications are often not administered on time	Competing nursing priorities prevent timely administration from being prioritised
	Lack of education regarding the importance of administering medications on time
	eMR system does not advocate for the timely administration of PD medications
	High-risk or low-use medications may not be available on the ward
	Delays in obtaining medication, which is not available on the ward—there is no alert system for pharmacists and nurses to action ordering medications available only in the dispensary
PD medications are often incorrectly prescribed	Medication regimens are complex, and the history is poorly taken due to heavy workload and lack of education
	The eMR function to chart custom medication administration times is not user-friendly
	There is a lack of education on how to chart custom medication administration times

.

2.2.1. Data Collection and Analysis

A mixed methods approach was used to collect data on solution development.

Quantitative Data

Organisational reports were requested and provided by the RNSH pertaining to the root causes. This information was collated and analysed descriptively in Microsoft Excel spreadsheets.

Qualitative Data

Focus groups consisting of emergency nurses (n = 10), neurology nurses and allied health (n = 15), aged care nurses and allied health (n = 18), pharmacists (n = 31), geriatricians (n = 8), and neurologists (n = 5) were consulted for process mapping and brainstorming. Patients and carers were not included in the solution design work, as the information gathered during the diagnostic phase indicated their perception of the care received was not consistent with hospital performance, although one patient and one carer were included in the steering committee for the duration of the project [14].

Participants were presented with the five key issues and their root causes, previously identified during the diagnostic phase of this project [14]. Participants were asked for their ideas, asking questions such as ‘In an ideal world what would make this easier for you?’, ‘How would improve [root cause]? Details of the ideas were elicited asking probing questions such as ‘What would that look like? How would that work? Who would we need to involve?’ Focus group discussions were transcribed verbatim, and data was deductively analysed according to each key issue. Themes were derived through content analysis and collated to develop solutions.

2.3. Implementation Phase

The implementation phase consisted of executing activities developed as part of each solution. These activities are detailed in

Table 2. Final Solutions and Implementation Plan.

Solution	Implementation Plan (Activities)	Additional Tasks	Outcome Measures
Solution 1. Formalise routine patient identification	Emergency Ward—PD patients entered can be identified within the ED by the Green PD icon on the FirstNet board in the ED. Neurology Ward—PD patients are included in the safety huddle, identifying issues and risks for patients on each shift. Neurology and Aged Care—Patients are highlighted on the nursing handover.	Training nurses in ED to check for and enter the code at triage and the bedside nursing assessment. Training pharmacists to scan FirstNet and the Patient Flow Portal to identify PD patients to Ongoing education sessions, quick reference instructions, and emails for nurses and pharmacists. Open feedback loop for staff. Periodic report of outcome measures to managers	Data collection: Report from EMR. Measure: proportion of PD patients who have the green icon activated in FirstNet. Measure: Proportion of PD patients who receive a pharmacist medication reconciliation within 48 h and during admission
Solution 2. Provide improved staff Education	Face-to-face/virtual educational sessions: PD medication overview Importance of routine identification of PD patients Email—Educational emails to reinforce the learnings from the educational sessions. Educational Packages Quick reference instructions—Provide staff easy access to reminders of how to enter the PD ‘problem’ into FirstNet and PowerChart.	Open a feedback loop for staff to provide feedback on the solution to the team. Monitoring via monthly audits to track the measurable benefits. PD education embedded into annual education programmes and orientation programs for JMO’s, nursing new graduates, departmental educational meetings, including Elderly Care, Neurology, Pharmacy, and Emergency.	Data collection: staff survey Measure: knowledge that PD medication on time means within 15 min of the prescribed time Data Collection: Report from EMR Measure: proportion of pharmacy medication reconciliations completed in 24 h.
Solution 3. Develop and install an Automated Prescriber alert	A pop-up prescriber alert in the medication administration record when any levodopa medication or entacapone is prescribed	Education sessions and quick reference instruction cards for JMO ID tags Periodic report of outcome measures	Data collection: Manual review of Pharmacy medication reconciliations. Measure: prescriber errors.
Solution 4. Review and amend ward PD medication stock	Review and amendment of ward stock medication available on 7F, 9E, and ED occurred in consultation with Nursing Unit Managers (NUMs) and Senior Pharmacists, improved access to PD medications on the wards. Poster listing what PD medications are available on the ward imprest and instructions for nurses to administer. Posters identifying PD medications included image of the medications and formulation details.	Ongoing education to nurses and pharmacists on what changes are made to the imprest supply and why timely administration benefits the patient. Posters and emails about the solution. Communication feedback loops. Annual review of imprest stock to confirm information on the posters remains current.	Data collection: Focus groups and feedback sought during education sessions embedded into business as usual to confirm resources and ward stock remain up to date and relevant Outcome: Resources and ward stock are reviewed annually.
Solution 5. Develop systematic prompts for nurses regarding the time-critical nature of PD medications	EMR Electronic component 1 ‘Order Comment’ for all levodopa and/or entacapone containing medications on the medication administration record. Time-critical must be administered within 15 min of prescribed time EMR Electronic component 2 Tile indicating administration is due, turns red after 15 min for all levodopa and/or entacapone—containing medications Pharmacy Electronic component 3 and Physical component Time-critical prompt automatically included on medication labels when dispensing levodopa-containing products.	Ongoing education provided to nurses and pharmacists on the need to prioritise the timely administration of PD medications. Posters and emails about solutions. Open feedback loop Periodic report of outcome measures to Nurse Unit Managers and Clinical Nurse Educators to provide feedback on performance.	Data Collection: Report from EMR. Measure: timeliness of PD medication administration.

.

2.4. Evaluation Phase

The evaluation phase in healthcare redesign involves measuring resource use, process management, staff satisfaction, well-being, and productivity [17]. An evaluation is about making a comparison to discover if the solutions that were implemented have delivered the expected benefits and whether these have been sustained. A mixed-methods, pre-post approach was utilised to evaluate the impact of the solutions implemented in the previous phase. The pre-intervention period was January to June 2022. The post-intervention period was July to December 2022. Data collection continued until June 2024 to demonstrate the sustainability of solution implementation, a vital component of healthcare redesign [9].

The success of the project was defined by the following project objectives:

• Increase proportion of PD patients clearly identified in ED—flagged with a ‘PD problem code’ in the electronic medical record (EMR) from 31% to 80%;
• Increase the proportion of PD patients who receive a pharmacist medication reconciliation within 48 h of admission from 16% to 60%;
• Increase the proportion of levodopa doses administered within 15 min of the prescribed time from 52% to 80%;
• Increase the proportion of pharmacist medication reconciliations free of PD medication prescribing errors from 38% to 80%;
• Improve service efficiency, experience of care, and health outcomes for PD patients by reducing their median length of stay by 10%.

We expect that the outcomes of this study will inform and contribute to improved inpatient care for people hospitalised with PD and improved systems and processes within the local setting for staff working with PD patients, PD patients, and their families and carers.

Data Collection and Analysis

Data was collected through organisational data reports and staff focus groups. A pre-post analysis of organisational data pertaining to PD medication management at RNSH and experiential data from staff (Table 2) was undertaken. All quantitative data were collated and descriptively analysed in Microsoft Excel. All qualitative data were themed using content analysis. The administration of PD medication was analysed in two ways. Firstly, by “on time within 15 min of the prescribed time” and secondly, by “on time within 30 min of the prescribed time”. Whilst this study focused on a 15 min window as best practice, previous studies consider a 30 min window to be standard practice. Therefore, the 30 min window is comparable with other previous research outcome measures.

3. Results

3.1. Solutions Design and Implementation Phase

Five solutions were developed (Table 2). These solutions focused on various physical and behavioural processes across and within the wards. They were both staff-focused and patient-focused and included a range of achievable outputs for implementation.

3.2. Evaluation Phase

There were a total of 1358 PD patients admitted between January 2022 and June 2024 (mean: 272 per 6-month period) (

Table 3. Number of PD patients admitted.

Admissions	RNSH	ED	Neurology	Elderly Care
January–June 2022 (Pre-intervention)	279	225	40	37
July–December 2022 (Post-intervention)	252	205	34	25
January–June 2023 (Sustain phase)	266	225	44	21
July–December 2023 (Sustain phase)	287	239	46	28
January–June 2024 (Sustain phase)	274	234	38	27

). PD patients were identified using ICD10-AM 12th edition [18] codes G20—indicating PD was considered active during the admission, or U80.1—indicating the PD was not considered active during the admission (

Table 4. Demographic data.

	RNSH Total Admissions	Male (%)	Average Age PD pt (Years)	Average Length of Stay (Days)	PD Primary Diagnosis, n (%)	G20, n (%)	U80.1, n (%)
January–June 22 (Pre-intervention)	279	61	79	11	26 (9)	111 (40)	168 (60)
July–December 22 (Post-intervention)	252	65	78	12	11 (4)	82 (33)	170 (67)
January–June 23 (Sustain phase)	266	58	78	17	31 (12)	101 (38)	165 (62)
July–December 23 (Sustain phase)	287	62	78	13	23 (8)	96 (33)	191 (67)
January–June 24 (Sustain phase)	274	62	78	12	24 (9)	95 (35)	179 (65)

).

3.2.1. Solution 1: Patient Identification

Following the implementation of solution 1, the proportion of PD patients identified in ED using the PD Icon on the FirstNet tracking board improved on average, from 39% (n = 74) pre-intervention to 45% (n = 91) post-intervention (Figure 2). This improvement was sustained with results continuing to trend towards the 80% target. December results are consistently poorer than usual performance.

The proportion of PD patients who received pharmacy medication reconciliation within 48 h of admission improved from 16% (n = 35) pre-intervention to 27% (n = 84) 6 months post-intervention and was sustained between January and December 2023 at 27% (n = 69). The proportion of PD patients receiving a pharmacy medication reconciliation was not sustained as it reduced to pre-intervention levels, 17% (n = 50), during July to December 2023 and remained poor at 18% (n = 51) during January to June 2024 (Figure 3).

The proportion of PD patients who received a pharmacy medication reconciliation during admission followed a similar trajectory of initial but not sustained improvement, with 27% (n = 59) pre-intervention, 43% (n = 135) post intervention June to December 2022, 41% (n = 135) January to June 2023, and 29% (n = 80) July to December 2023, 33% (n = 91) January to June 2024 (Figure 3).

3.2.2. Solution 2: Education

Post-intervention, 71 staff were surveyed (25 doctors, 34 nurses, and 6 pharmacists) about their knowledge of the importance of administering PD medication on time, where on time is defined as within 15 min of the prescribed time. Most respondents (82%, n = 52) demonstrated an improvement in knowledge with an average of 23 percentage points per respondent. Most doctors (88%, n = 22) showed an improvement in knowledge with an average of 28 percentage points improvement per respondent. Most nurses (75%, n = 30) showed an improvement in knowledge with an average of 9 percentage points improvement per respondent. All pharmacists (100%, n = 6) showed an improvement in knowledge with an average of 52 percentage points improvement per respondent (Figure 4).

3.2.3. Solution 3: Medication Prescribing Errors

There has been modest improvement towards the target for this objective of 80% of medications reconciliations free of prescribing errors (Figure 5). The education provided to JMOs when they start each rotation in 2023 has contributed to sustained improvement: Medication reconciliations free of prescribing errors, February–April 2022, were 38% compared to the same period in 2023 of 61%.

There are also improvements in the number of pharmacist medication reconciliations that detected multiple prescribing errors for a single patient. Particularly when comparing January 2022 to April 2022 with January 2023 to April 2023 (Figure 6). While multiple prescribing errors still occur, they are less frequent.

3.2.4. Solution 4: Ward Medication Stock

Feedback from education sessions suggests that the posters utilised in this solution improve confidence when administering PD medications and also make it easier to access the medication sooner. Focus group data were consistent with this, with two overarching themes.

• Visual cues are beneficial—staff reported that displaying the posters next to where the medications are stored helped them to check the medications they administered easily and efficiently.
• Stock availability improved workflow—having the stock available was crucial for workflow and ensured that staff were able to access medications when needed, minimising delays in administration.

3.2.5. Solution 5: Systematic Prompting

The implementation of systematic prompting (the order sentence and the tile going red at 15 min) in EMR has resulted in sustained improvements in the administration of PD medications on time, even extending beyond the wards where education and resources were provided (Figure 7). Pre-intervention, 52% (n = 4284) of levodopa doses were administered on time within 15 min across the whole hospital. Post-intervention, this improved to 67% (n = 4473) of doses. This improvement has been sustained for a further year post-implementation across the hospital, achieving 70% (n = 6036) for the January to June 2023 period and 75% (7307) for the July to December 2023 period, and 74% (n = 5262) in the January to June 2024 period. The wards of focus have maintained improvements for the same time periods, each achieving the 80% target in individual monthly reports. The emergency department continues to sustain modest improvements of 58% and 56%.

Pre-intervention, 73% (n =5937) of levodopa doses were administered on time within 30 min across the whole hospital. Post-intervention, this improved to 84% (n = 5637) of doses. This improvement has been sustained for a further year post-implementation across the hospital, achieving 87% (n = 7481) for the January to June 2023 period and 89% (n = 8629) for the July to December 2023 period, and 74% (n = 6267) for the January to June 2024 period. The wards of focus have maintained improvements, each achieving over 90% across both post-intervention periods. The ED continues to sustain improvement with 76% and 68% of doses administered within 30 min (Figure 8).

4. Discussion

This study aimed to address previously identified issues and their root causes for medication management in PD patients admitted to Royal North Shore Hospital (Sydney, Australia). The results suggest that the identified solutions improved systems and processes that contribute to quality and safe care for PD patients and that these were generally sustainable. Consistent with previous studies, when PD patients are correctly identified within an acute care hospital, the correct prescription of medications, timeliness of medication administration, and pharmacy review improve [1,4,6,8]. Further, we demonstrated that the identification process can occur through electronic triage, education, and good communication. Previous research suggests that patients can be identified using posters or electronic alerts on patient tracking boards in the emergency or bed management boards [1,9]. Various strategies have been shown to improve the timely administration of PD medications. These include alerts, for example, ‘get it on time’ stickers or posters [1,19,20,21,22,23], and electronic alerts in electronic medical records [21,24], to prompt nurses to administer medications on time, and to improve the availability of stock on the ward [20,25]. This is consistent with our findings, highlighting the use of medical records prompts and consistent information to ensure timely medication administration.

Underpinning all issues related to poor medication management for PD patients is the limited knowledge doctors, pharmacists, and nurses have about PD medications. Education in the form of posters, ‘grand rounds’ presentations, or unit-based education programmes has been shown to improve medication management [3,19,20,21,22,23,25], consistent with our findings. However, unless the education programmes are of high quality, translation of knowledge into practice may not be effective or sustained [26].

One of the main strengths of this study, and the redesign approach to change, is the engagement and empowerment of stakeholders. Solutions must align with the patient’s knowledge, skills, and experience, [27] as well as their values and preferences [28]. Clinicians, in particular local opinion leaders, need to be engaged and supported through the process of implementation [19] until successful adoption has taken place [20]. An implementation process that is not person-centred (both patients and staff), clinically effective, and continually improving in response to a changing context [29] will likely fail. Our study demonstrated that engagement during the solutions phase created a strong sense of ownership of the solutions and investment in successful outcomes.

The objectives of this redesign study were partially met. The outcomes for each objective are described below.

Objective 1: Patient identification. The target of 80% has not been reached. This solution required the largest change in behaviour by the Nurses in the ED and required the most follow-up, ongoing education, and support. ED nursing staff were the stakeholders routinely targeted to enter the problem code that triggers the ‘PD Icon’ on the first net board, so that from the point of triage, patients would be clearly identified as having PD. However, reaching all nurses remains difficult as there are over 150 nursing staff, a high rate of staff turnover, and many critical care topics that compete for space in the ED education programme. As a result of the poor performance in early 2023 project team changed strategy. ED nurses at triage are no longer expected to enter the problem code. However, elderly care CNC in Emergency continues to do so as part of their assessment of PD patients in ED. The PD CNC now enters the PD problem code manually and retrospectively when auditing the data each month. This change in focus has resulted in the modest ongoing improvement achieved through the second half of 2023 and 2024.

Objective 2: Pharmacist medication reconciliation. The improvements in PD patient identification in the ED make it easier for pharmacists to identify PD patients in the ED. However, initial improvements in the number of patients who received a medication reconciliation within 48 h of admission are potentially attributable to increased pharmacy staff over the winter of 2022. These increased staffing levels were temporary and may account for why the improvements were not sustained into 2023 and 2024. Pharmacists are often under-resourced, making timely medication review difficult; however, their work is vital to the detection and prevention of medication errors [14,22]. The Pharmacy department implemented a concurrent change to how patients are referred for medication reconciliation in 2023 within the EMR system. Their improved knowledge of the importance of administering PD medication within 15 min of the prescribed time informs how they organise their workflow to prioritise referrals for PD patients, but has not resulted in any improvements in the number of PD patients receiving a medication reconciliation within the 48 h time frame.

Objective 3: Levodopa administration within 15 min of the prescribed time. Moderate improvements in nurses’ knowledge do not account for the improvements in the administration of PD medication on time across the hospital. Our findings support the implementation of time-critical alerts within EMR to improve the administration of PD medication on time [21,30]. Prompts built into the EMR system are automated, efficient, and immediately scalable to other hospitals using the same EMR system. Furthermore, the alert is easily replicable by other services using a different EMR system.

Objective 4: PD medication prescribing errors. Medical staff often have limited expertise in PD [30]. They often lack the knowledge and confidence to prescribe accurately using electronic systems that are often not intuitive, causing medication timing errors to be the most common type of error [14]. The combination of education and quick reference tags summarising instructions to prescribe patient-specific custom times and to check formulation and dosages is correct did result in a reduction in prescribing errors. Despite embedding this education into business-as-usual annual junior medical officer education orientation programmes at the hospital, and orientation sessions at each rotation to neurology and aged care, the number of pharmacy medication reconciliations free of prescribing errors was not sustained. However, the improvements in the number of pharmacy medication reconciliation reporting multiple errors have been sustained. The low number of pharmacy medication reconciliations conducted for PD patients must be noted as a limitation that may contribute to these findings. It must also be noted that manual auditing of pharmacy medication reconciliations, while providing insightful and valuable information, is time-consuming and not sustainable as a long-term evaluation outcome measure.

Objective 5: Impact on LOS. Despite the improvements made to medication management for PD patients admitted to the hospital, there was no reduction in the average LOS. This is inconsistent with the literature that suggests improved management of PD medication demonstrates a reduction in hospital LOS [3,19,22,31,32]. It is not clear why this work did not impact LOS and raises questions as to what else may be contributing to the PD patient LOS. Further work is required to identify a possible range of issues beyond medication mismanagement that contribute to a PD patient’s LOS.

This study has several limitations. Firstly, the study was undertaken at one hospital and the findings may not be able to be generalised across different health services. Further, whilst patients and their families were invited to participate in the process through focus groups and interviews, only limited data was collected, and this was not able to be included in any analysis of findings. Future work would benefit from a more comprehensive inclusion of patients and their families or supporters.

5. Conclusions

PD medication regimens are complex and specific to individual patients who are often admitted to the hospital for something other than their PD. Improved identification of PD patients through the implementation of EMR alerts and ward-based systems, coupled with education embedded into hospital business as usual, equips staff from multiple clinical backgrounds to prioritise accurate prescribing and timely administration of PD medications. Furthermore, improvements in PD medication management will extend beyond the education programme when EMR and hospital systems have alerts with feedback loops that prompt staff to prescribe accurately and administer medication on time. These changes to hospital systems are sustainable and scalable, improving PD medication management and PD symptom control for PD patients admitted to the hospital. However, further work is required to understand what impacts the PD patients’ LOS beyond quality PD medication management.