Article Text

Download PDFPDF

Using a validated case mix tool for use in the telephone-assisted triage of patients in a specialist palliative care community setting: a consecutive case series
  1. Emer Hough,
  2. Michael Lucey,
  3. Martina O'Reilly,
  4. Hannah Featherstone,
  5. Feargal Twomey and
  6. Siobhan Coffey
  1. Department of Palliative Medicine, Milford Care Centre, Limerick, Ireland
  1. Correspondence to Dr Hannah Featherstone, Palliative Medicine, Milford Care Centre, Castletroy, Limerick, Ireland; featherh{at}


Objectives Allocating resources in palliative care is challenging due to the nature of life-limiting illness coupled with the propensity for significant physical symptoms and psychological distress. At present, there is no established system for triaging referrals and prioritising resource allocation.

This study aimed to evaluate the feasibility of using a case mix assessment tool for telephone-assisted triaging of referrals to a specialist palliative care service. This assessed a patient’s phase of illness, Problem Severity Score (PSS) for complexity of symptom burden and psychological distress, and functional status.

Methods Using a prospective consecutive case series approach, 450 referrals to community palliative care over a 6-month period were assessed. Scores for phase of illness, PSS and functional status were assessed at triage, as was the triage category of urgency of response.

Results Analysis demonstrated that phase of illness corresponds with triage category, with terminal or unstable phase patients significantly associated with urgent (category 1) referrals and highest priority for review. Decreased functional status and high PSS were useful predictors for increased urgency of referral.

Conclusions These results demonstrate that this case mix tool could assist in the telephone assessment and triage of referrals to community palliative care.

  • Clinical decisions
  • Home care
  • Supportive care

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Optimising triage of referrals to community palliative care can facilitate timely intervention and improve symptom control. However, there is no established triage system to implement this.


  • A case mix assessment tool assessing phase of illness, Problem Severity Score and functional status can be used systematically to triage the urgency of referrals to community palliative care. Telephone triage of urgency of referral using this system corresponds to urgency of referral on in person review.


  • The case mix assessment tool can be considered for use in telephone triage of urgency of referrals.


Referrals to specialist palliative care (SPC) services are increasing in volume and diversity.1 These referrals must be triaged; this is defined as ‘the analysis of the urgency of a referral and the assignment of a corresponding predefined category of urgency to the referral’.2

Enhancing triage systems can facilitate early palliative home care intervention, improve symptom control and decrease hospital deaths.3

Patient phase of illness, Palliative Care Problem Severity Score (PSS) and functional status has been proposed as a case mix model for palliative care patients.4 To date, there have been no studies using this case mix tool in telephone-assisted triage of patients.


This was a prospective, consecutive case series of 450 patients referred to the SPC community service in Milford Hospice in the Mid-West of Ireland. Six months of data for all patients that had first visits was collated from 1 January to 30 June 2020.

Case mix model

A case mix model was developed by Eagar et al in 20044 and implemented nationally in Australia by the Palliative Care Outcomes Collaboration (PCOC). This uses three different domains unrelated to patient diagnosis, described below.

  1. Phase of illness

    This refers to a distinct clinical period of a patient’s illness5: stable, unstable, deteriorating, terminal and bereaved.

  2. Palliative Care Problem Severity Score (PSS)

The PSS numerically evaluates a patient and family’s problem severity across four domains:

  • Pain.

  • Other symptoms.

  • Psychological distress.

  • Family and caregiver distress.

Each domain is scored from zero to three, with zero indicating absence of distress and three indicating severe distress. The overall score is out of 12.

  • 3. Australian-modified Karnofsky Performance Status (AKPS)

Measured from 0% to 100%, the AKPS (figure 1) is validated for use in palliative care patients.6

Figure 1

The Australian-modified Karnofsky Performance Scale.

Categories of urgency of referral

The Health Service Executive is responsible for provision of health and personal social services in Ireland. Timeliness of response criteria was introduced nationally as a key performance indicator.7 Patients referred to be seen are categorised as:

Category 1: Urgent referral (visit within 2 working days).

Category 2: Early referral (within 1 week).

Category 3: Routine (within 2 weeks).

Data collection

The triage clinical nurse specialist (CNS) assessed the patient for SPC needs. The PCOC case mix scores were documented for every referral at triage, and the referrals urgency categorised, determining the time frame of the community SPC CNS visit. Phase scored by the CNS on first visit was also recorded. Data was subsequently analysed using SPSS v.28.


Association between phase and triage category

Category 1 patients

At triage, a Pearson χ2 test for independence indicated that patients in terminal and unstable phases are more likely to be assigned as category 1, x2 (2, n=450) = 126.96, and p<0.01, Φ=0.53 (large effect). Overall, 75.6% of patients in the unstable or terminal phases were in category 1.

Category 2 patients

At triage, a χ2 test for independence (with Fisher’s exact test) found a significant association between patients in the deteriorating phase and patients categorised as category 2, x2 (2, n=450) = 123.06, and p<0.01, Φ=0.52 (large effect). Overall, 76.8% of deteriorating patients were in category 2.

Category 3 patients

There was a significant association between patients in the stable phase and patients categorised in category 3, x2 (2, n=450) = 241.12, and p<0.01, Φ=0.73 (large effect). Overall, 75% of stable patients were categorised as being in category 3.

Congruency of phase between triage and first visit

The ability to accurately assess patients clinical need over the phone is an important aspect of triage. To look at this, we assessed congruency between the phase allocated by the community CNS on their initial visit and the phase allocated at triage using Fisher’s exact test.

Unstable phase

There is an association between patients in the unstable phase at triage and the unstable phase at first visit, x2 (1, n=427) = 14.59, p<0.01, Φ=0.18 (small effect).

Stable phase

Of those patients who were stable at triage, 63.5% were stable at first visit, x2 (1, n=385) = 78.04, p<0.01, Φ=0.44 (medium effect).

Deteriorating phase

There is an association between deteriorating phase at triage and scoring in the deteriorating phase at first visit, x2 (1, n=427) = 42.05, p<0.01, Φ=0.32 (medium effect).

Association between PSS and urgency of visit

A Kruskal-Wallis test revealed a statistically significant difference in PSS total score values across patients in three different categories (category 1, n=73; category 2, n=230; category 3, n=120), x2(2, n=423) = 157.12, p<0.01. Patients in category 1 had higher median score PSS Total Score (Md=7) than patients in category 2 or 3 (Md=5, 4).

A Mann-Whitney U test revealed significant difference in the total PSS scores of patients in category 1 (Md=7, n=73) in comparison to category 2 (Md=5, n=230), U=4047, z=−6.83, p<0.01, r=0.39 (medium effect) and in comparison to category 3 (Md=4, n=120), U=638, z=−10.180, p<0.01, r=0.73 (large effect).

Functional status

A Kruskal-Wallis test revealed a statistically significant difference in AKPS scores across three different categories (category 1, n=65; category 2, n=235; category 3, n=139), x2(2, n=439) = 153.23, p<0.01. Patients in category 1 had lower median scores (Md=20) than patients in category 2 or 3 (Md=60, 60).

A Mann-Whitney U test revealed significant difference in the median AKPS scores of patients in category 1 (Md=20, n=65) in comparison to category 2 (Md=60, n=235), U=1034.5, z=−10.87, p<0.01, r=0.63 (large effect) or category 3 (Md=60, n=139), U=191, z=−11.17, p<0.01, r=0.78 (large effect).


Previous attempts to develop a suitable prioritisation tool for triaging palliative care referrals8–12 have to date not resulted in a widely accepted evidence-based tool.1

In this study, we used an established case mix tool in the telephone-assisted triage of patients referred to an SPC community service. Urgent referrals (category 1) were most likely to be in the terminal phase of illness. Category 2 referrals were strongly associated with patients in the deteriorating phase of illness, reflecting non-urgent needs. Category 3 referrals were strongly associated with stable patients.

Another aspect of the tool that we wished to evaluate was the consistency of phase categorisation between telephone triage and first visit. There was a small association between unstable phase at the time of triage and first visit. This can be explained by the activity of the triage nurse, such as contacting the patients General Practitioner (primary care physician) and public health nurse to ensure there are urgent interventions, which often results in a change in phase of illness. There was a strong correlation between stable phase and deteriorating phase at the time of triage and first visit, showing the tool performed well in assessing phase by phone.

The patient’s functional status was significantly different across referral categories of urgency—lower AKPS scores were associated with increased urgency of referral.

There was also a significant difference in the PSS across all three categories of urgency of referral. The higher the PSS, the more urgent the referral.

We can see, therefore, that urgency of referral was significantly associated with terminal phase (and to a lesser extent unstable phase), high PSS and low functional status. Deteriorating and stable phases, lower PSS scores and better functional status were associated with lower urgency categories.

This study builds on existing international evidence relating to the key elements of effective triage in palliative care—identifying physical suffering, imminent dying, psychological suffering and caregiver distress as key drivers of urgency.13–15 These parameters are captured in the PCOC tool—PSS focuses on physical, psychological and caregiver distress, whereas terminal phase indicates proximity to death. Our study also showed that declining functional status can be added to the drivers of urgency of referral.

The single-centre nature of this study is a possible limitation—further studies at a national level could be conducted to confirm our findings. The COVID-19 pandemic during the time of data collection may or may not have also affected results. Future studies could be conducted to evaluate the effectiveness of the CNS initial visit by assessing change in case mix scores over that time period.


Categorisation of specialist community palliative care referrals for urgency is essential to optimise delivery of services. This study demonstrates that the PCOC case mix tool is feasible and acceptable to aid in this process as part of a comprehensive assessment, but further validation studies in the triage setting should be undertaken.

Ethics statements

Patient consent for publication

Ethics approval

Not applicable.


This work has not previously been presented or published.



  • Contributors EH: Investigation, formal analysis, writing—original draft. ML: Conceptualisation, methodology, formal analysis, writing—original draft, supervision. MO'R: Conceptualisation, methodology, formal analysis, writing—original draft. HJF: Writing—original draft. FT: Conceptualisation, methodology, supervision. SC: Investigation, methodology.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; internally peer reviewed.