Acute Deterioration in Long‐Term Care: A Delphi Study of Indicators, Care Pathways and Implementation Strategies

ABSTRACT

Aim

This study aimed to develop recommendations for acceptable physical and psychological clinical indicators of acute deterioration and strategies to support sustained adoption of an approach to detecting and responding to acute deterioration in long‐term care.

Background

Previous studies have trialled programmes to reduce avoidable hospital transfers through early detection of acute deterioration in long‐term care. However, variability in outcomes and implementation experiences highlights differences in the essential components of these approaches.

Design

Three‐round Delphi process.

Methods

The Delphi process conducted between September 2022 and January 2023 included a 22‐member Australian expert panel to elicit consensus and to generate recommendations. Delphi participants were asked to rate the potential items on a 9‐point scale for level of importance (Part A), level of priority of action (Part B) and level of acceptability (Part C). The mean, standard deviation (SD) and I‐CVI of each item were calculated. Clinical indicators with mean ≥ 7, I‐CVI ≥ 0.78 were considered to have a good degree of agreement, appropriateness and relevance and were deemed acceptable for inclusion.

Results

Participants completed a two‐part online survey in Round 1, and three‐part surveys in Rounds 2 and 3. Round 1, Part A included 44 clinical indicators. On completion of Round 3, 36 items remained. Care pathways in response to acute deterioration (n = 20) were included in Part B. After Round 3, 17 care pathways remained. There was agreement in Part C that communication aides, education and mentor programmes were necessary as viable strategies to support sustained adoption.

Conclusion

Detecting the early warning signs of acute deterioration will have significant outcomes for residents, staff and health care services.

Summary

What is already known about this topic

• Most older adults living in long‐term care (LTC) have complex care needs related to cognitive impairment, multiple chronic conditions, high symptom burden, poor social support and reduced quality of life.

• Detecting the early warning signs of acute deterioration is essential in preventing further deterioration, avoidable hospitalisation, increased disease burden and premature death.

What this paper adds

• Items have been identified on the early signs of acute deterioration in older people living in LTC

• Consensus has occurred on recommendations for elements of a programme to build capacity of the LTC workforce to recognise, report and respond to the early signs of acute deterioration of residents.

The implications of this paper

• The recommendations of early warning signs of acute deterioration in LTC in this study will inform the development of an early warning score system.

• The inclusion of recommendations about pathways in response to acute deterioration, and strategies to support capacity building of the workforce will ensure significant outcomes for residents, staff and health care services.

1. Introduction

In Australia, compared with people with no aged care support, older people in long‐term care (LTC) (residents) have a greater likelihood of using hospital services, including emergency department (ED) presentations (Australian Institute of Health and Welfare (AIHW) ²⁰¹⁹). In 2016–2017, almost a third (32%) of residents had an ED presentation as compared to 14% of individuals aged 50 years and over with no aged care support (Australian Institute of Health and Welfare (AIHW) ²⁰¹⁹). It has been estimated that 13%–40% of hospital admissions involving a resident could be avoided (Laging et al. 2018).

The transfer of a resident from aged care to ED and/or admission to hospital places the person at higher risk of adverse outcomes and events, including delirium, medical errors, falls, pressure injuries and further deterioration (Aminzadeh and Dalziel 2002; Fox et al. ²⁰²⁰). Detecting the early warning signs of acute deterioration that can, if unresolved, result in hospital transfer is essential in preventing further deterioration, hospitalisation, increased disease burden and premature death (Laging et al. 2018).

Identification of the deteriorating patient in acute care has been facilitated by tools, such as Early Warning Score Systems (EWSS), for example, the Modified Early Warning Score system (MEWS) (Downey et al. 2017). Mainly used in the acute care setting, EWSS have been widely adopted globally. Used by registered healthcare workers, the MEWS relies on regular (at least daily) collection of clinical parameters (i.e., blood pressure, pulse, temperature respiratory rate and level of consciousness) to track changes in health status. The MEWS also contains visual prompts for staff to take appropriate action in response to deterioration (Downey et al. 2017). The MEWS is effective in predicting length of stay, in‐hospital outcomes, critical care admissions and mortality (Downey et al. 2017).

Barriers to using EWSS tools in LTC include the LTC workforce mix and the unique presentation of deterioration in older people. The unique skill mix of the LTC workforce arguably contributes to the failure to detect early signs of deterioration and delays initiating appropriate action (Royal Commission into Aged Care Quality and Safety, 2020). A notable difference between LTC and acute care workforces is the disproportionate number of regulated nurses (registered [RN]/enrolled [EN] nurses) and unregulated carers (personal care worker [PCW]/assistants in nursing [AIN]) with variation in workforce composition across and within countries, including the presence of more advanced practice roles in some international settings. In Australian LTC facilities, the number of regulated nurses has declined over the last 13 years to 14.6% (from 21%), while the number of unregulated carers has increased to over 71.5% (from 56.5%) (Eagar et al. 2019). As a result, the workforce providing the majority of care and being in a position to notice acute changes in the resident, unregulated carers do not have the training or medical literacy to conduct the assessments needed to use EWSS tools efficiently. Further, due to age‐related changes, changes in clinical indicators, such as vital signs, can be a very late indication of deterioration in older people, with the resident often being critically unwell when change is measurable (Chambers et al. 2023). A recent literature review revealed that in older people, the early signs of deterioration were identified as being subtle, yet observable changes including changed mood, behaviour, level of alertness, appetite and function (Chambers et al. 2023).

Previous research has trialled versions of EWSS developed for the LTC setting that consider the altered presentation of acute deterioration in this population. These have included Stop and Watch (Basinska et al. 2022; Ouslander et al. ²⁰¹¹; Ouslander et al. ²⁰⁰⁹; Porter Tamara et al. ²⁰²¹) and Significant 7 (Little et al. 2019). Multimodal programmes aiming to improve LTC workforce capacity to detect and respond to acute deterioration, incorporated these tools along with staff training, care pathways, champion models and communication tools (HELP‐LTC) (Boockvar Kenneth et al. 2020; Boockvar et al. ²⁰¹⁶), EDDIE (Carter et al. 2020), INTERACT (Huckfeldt et al. 2018; Kane et al. ²⁰¹⁷; Ouslander et al. ²⁰¹¹; Ouslander et al. ²⁰⁰⁹; Popejoy et al. ²⁰¹⁹; Rantz et al. ²⁰¹⁷; Tappen et al. ²⁰¹⁷), and the Sub Acute Program (O'Neill et al. ²⁰¹⁷). These programmes have been evaluated across multiple high‐income countries, including the United States, Canada, and Australia. While some of these programmes have demonstrated effectiveness in reducing hospitalisations (Boockvar Kenneth et al. 2020; Boockvar et al. ²⁰¹⁶; Carter et al. ²⁰²⁰; Huckfeldt et al. ²⁰¹⁸; Kane et al. ²⁰¹⁷; Loeb et al. ²⁰⁰⁶; Ouslander et al. ²⁰⁰⁹), as well as improved cost effectiveness (Carter et al. 2020; Loeb et al. ²⁰⁰⁶), others did not find a change in hospitalisation rates (Rantz et al. 2017) and that the programme introduced additional paperwork (Ouslander et al. 2009), were time consuming (Renz et al. 2013), and there was not full uptake of programmes (Ouslander et al. 2009). Evidence regarding sustained use of these programmes beyond the life of research projects is limited. These findings indicate variability in outcomes and implementation experiences, highlighting differences in the essential components of early warning approaches in LTC. Hence, this study builds on previous research in considering the enablers and barriers to detecting and responding to acute deterioration in determining through expert consensus what can be appropriately incorporated in current clinical care.

Therefore, this study aimed to generate items for a tailored EWSS relevant and acceptable for LTC (including residential aged care, nursing homes and old age homes) and to provide recommendations for a sustainable programme to support the LTC workforce to identify and manage acute deterioration in‐house.

2. Methods

2.1. Aims

The specific aims of this study were to seek expert consensus to

1. identify physical and psychological clinical indicators to be included in an EWSS to detect and respond to the early signs of acute deterioration in older people living in LTC and

2. generate recommendations for elements of a programme to build capacity of the LTC workforce to recognise, report and respond to the early signs of acute deterioration of residents.

2.2. Survey Development

Recommendations from two earlier phases of this research (rapid literature review; Chambers et al. ²⁰²³) and a qualitative study of LTC staff, and family of residents in Australia (MacAndrew et al. ²⁰²⁵) helped develop the items for the Delphi survey. The rapid literature review identified and categorised the clinical signs of acute deterioration in residents living in LTC while the focus group and interviews in the qualitative study helped determine (1) how clinical deterioration is currently detected and actioned and (2) the role of the aged care workforce and family members/friends in detecting deterioration in residents. The research team met several times to discuss and categorise generated items for inclusion in the survey.

2.3. Design

An online Delphi process is a well‐established flexible approach utilising ‘rounds’ to identify a consensus view across experts in a particular area (Barrett and Heale 2020). This approach allows for reflection after seeing anonymous results of a previous ‘round’, allowing for potential reconsideration of views (Barrett and Heale 2020). A three‐round Delphi process has been indicated to be sufficient to achieve consensus (Donohoe and Needham 2009).

2.4. Participants

A panel (n = 33) with theoretical and/or practical experience in aged care was identified across Australia through the research team's professional network of clinicians and researchers and were invited to participate by email. The invitation was sent to potential panel members by a research assistant to mitigate the risk of potential participants feeling pressured to participate. These experts included LTC RNs (including clinical nurse specialists and Nurse Practitioners), acute care emergency clinicians (RNs and medical doctors), other doctors (Geriatricians and general practitioners) and nurse academics. Consent to participate was indicated by return email, upon which the link to the Qualtrics survey was forwarded. In recognition of their time, all participants were offered an incentive per completed round. Each round was designed to take up to 1 h to complete.

2.5. Data Collection

The Delphi process was conducted between September 2022 and January 2023. Participants completed a two‐part (A and B) online survey in Round 1, and three‐part (A, B and C) surveys in Rounds 2 and 3. Delphi participants were asked to rate the potential items on a 9‐point scale for level of importance (Part A), level of priority of action (Part B) and level of acceptability (Part C). The items generated under the three topic areas (A, B and C) are explained in the following section.

In Round 1, participants were asked to complete demographic information (role, highest qualification, experience in care of the older person and qualifications in geriatric care) and to rate a series of statements related to the physical and psychological indicators of acute deterioration (Part A) and possible care pathways (Part B). Items in Part A were rated using a 9‐point Likert scale (1 = not at all important, 9 = extremely important) to indicate the level of importance of the items. Part B also used a 9‐point Likert scale (1 = very low, 9 = extremely high) to indicate the priority of the action. There was space provided at the end of each section to add other early signs and/or actions not included in the lists and/or for participants to suggest any word changes or other edits for items in the presented list. New and edited items suggested by participants in Round 1 were then added to Round 2 to be rated.

In Round 2, participants were individually sent a summary of their individual ratings and an aggregate (de‐identified) summary of Round 1 one results indicating which items had been accepted (items with mean score ≥ 7.0 and item‐content validity index (I‐CVI) ≥ 0.78) (Polit et al. 2007) and which items had been excluded (items with mean score ≤ 3.0 and I‐CVI ≤ 0.55). Items with a mean score > 3.0 and < 7.0 and I‐CVI > 0.55 and < 0.78 were included for panel members to re‐rate, and new items or reworded items were also included. There was space at the end of the sections again to add any further thoughts. New items suggested by participants in Round 2 were then added to Round 3 to be rated. In addition, a third part (Part C), focused on strategies to support LTC staff to adopt a new approach to identifying and responding to acute deterioration, was included in Round 2. The panel were asked to rate Part C items using a 9‐point Likert scale (1 = not acceptable at all, 9 = extremely acceptable) to indicate how acceptable each item was for LTC. The same criteria for inclusion, re‐rating or exclusion were applied to all Round 2 items.

In Round 3 (deciding round), the same process was followed as for Round 2. As aforementioned, some items were resent for rating for a third and final round of the Delphi process. Additional suggestions for items were not an option for Round 3; however, the panel could still add comments.

2.6. Ethical Considerations

Ethical approval was obtained from the university Human Research Ethics Committee.

2.7. Data Analysis

An excel spreadsheet was used to organise the data and the open‐ended responses. IBM SPSS Statistics 28 was used for the descriptive statistical analysis. The mean, standard deviation (SD) and I‐CVI of each item were calculated. The CVI can be defined as the degree to which an instrument has an appropriate sample of items for the construct being measured (Polit and Beck 2004) and among nurse researchers has been a widely used method in health and nursing research (Almohanna et al. 2022; Cocchieri et al. ²⁰²¹; Zhang et al. ²⁰²⁴). Further, a study by Franc et al. (2023) of 490 000 simulated Delphi questions concluded that SD and mean were valid methods to analyse ratings from Delphi studies. Polit and Beck (2004) recommend a minimum of I‐CVI = 0.78 for items to have excellent content validity. In Rounds 2 and 3, participants were instructed to consider moving their original item score either closer to nine indicating the item should be included, or closer to one indicating the item should be excluded. Authors CP, DS and MM met after each survey round to discuss the items to be included, resent, reworded or excluded according to the mean and/or I‐CVI and the comments/feedback from free text responses. For the final analysis after Round 3, all items < I‐CVI of 0.78 were excluded. Results from closed‐ended questions are reported in summary form and presented with no identifying information.

3. Results

Twenty‐six of the 33 potential participants invited (78.8%) consented to take part in the Delphi process. Of these, 22 (84.6%) completed Round 1. Rounds 2 and 3 were sent only to those who completed Round 1. Table 1 summarises the disciplines of and education attained by participants and numbers completing each round. Seventy per cent of the overall participants had experience of working as a clinician for over 5 years and 30% had experience as a clinician for less than 5 years in the care of the older person. In their current roles, 45% of the health professionals who participated in the Delphi survey were working in the LTC setting and 32% in the acute care setting (Table 1).

TABLE 1.

Discipline of, and education attained by, participants in each round.

	Delphi rounds
Participants	Round 1	Round 2	Round 3
	n (%)	n (%)	n (%)
Nurses: Registered nurses, nurse practitioners, clinical nurse specialist
14 (64)	12 (60)	13 (65)
Medical professionals: General practitioners, medical officers with gerontology expertise	3 (14)	3 (15)	3 (15)
Gerontological academics	5 (23)	5 (25)	4 (20)
Total	22 (85)	20 (91)	20 (91)

Highest level of education	n (%)
Higher degree in research	9 (41)
Post Graduate degree	8 (36)
Bachelor's degree	3 (14)
Diploma (advanced)	1 (5)
Other (certificate in nursing)	1 (5)

Current role	n (%)
Health professional (e.g., nurse and doctor) working in LTC	10 (45%)
Health professional (e.g., nurse and doctor) working in acute care	7 (32%)
Academic in gerontology	5 (23%)
Total	22

Tables 2, 3, 4, 5, 6, 7 summarise all items included in Parts A, B and C, and the round the item was included/eliminated based on their mean and I‐CVI values. The final survey included 12 new items in Part A (general appearance n = 2 new items, physiological and neurological state n = 4 new items, and nutrition, hydration and elimination n = 6 new items), two new items in Part B (staff requiring no specialised training n = 1 new item, and staff requiring specialised training n = 1 new item), one new domain (education for family n = 1 item) and two new items in Part C (education for staff n = 1 new item, and programmes involving external health professionals n = 1 new item).

TABLE 2.

Final items in Part A (physical and psychological indicators of acute illness in long‐term care residents).

Items	Mean	SD	I‐CVI	Deciding round
General appearance (changes in how the resident looks)
The resident is finding it harder to breath (e.g., effort of breathing)	8.41	0.89	0.95	Round 1
The resident is short of breath	8.23	1.08	0.91	Round 1
The resident experiences significant blood loss	8.45	1.23	0.91	Round 1
The resident reports their health is worse	7.68	1.29	0.86	Round 1
Family/friends report the residents' health is worse	7.77	1.04	0.86	Round 1
The resident's skin colour has changed (i.e., pale, flushed and cyanosed)	7.68	1.22	0.82	Round 1
The resident's breathing is more noisy	7.55	1.99	0.82	Round 1
The resident's breathing is different (faster, slower)	7.73	1.76	0.82	Round 1
The resident has a new or worsening cough	7.35	1.59	0.80	Round 2
Functional changes (changes in how the resident performs routine tasks)
The resident becomes more breathless during care activities	7.73	1.01	0.91	Round 1
The resident communicates more or less (wording edited in Round 2 survey)	7.86	0.97	0.91	Round 2
The resident had a fall or nearly fell	8.05	1.40	0.91	Round 1
The resident has a new injury (i.e., skin tear, pressure injury, bruise or swelling)	7.86	1.42	0.91	Round 1
The resident has a new or increased difficulty getting out of bed (wording edited in Round 2 survey)	7.45	1.53	0.80	Round 3
Psychological and neurological state (changes in the resident's mood, thoughts or behaviour)
Resident has a new onset of or increased confusion	8.86	0.63	1.00	Round 1
Resident has a new onset of or increased agitation	8.55	0.66	1.00	Round 1
Resident is more difficult to wake or could not be woken	8.68	0.55	1.00	Round 1
Resident has a new onset or increase of delusion or hallucination (new item included in Round 2 survey)	8.70	0.65	1.00	Round 2
Resident has a new onset of or increased aggression	8.50	0.78	0.95	Round 1
Resident is more drowsy	8.45	0.78	0.95	Round 1
Resident has a new onset of or increased disorientation (new item included in Round 2 survey)	8.55	0.82	0.95	Round 2
Resident reported new or increased pain	7.86	1.14	0.86	Round 1
Resident did not want to get out of bed	7.05	0.97	0.85	Round 3
Resident's ability to follow instructions has changed (new item included in Round 2 survey)	7.80	1.31	0.85	Round 2
Resident fell asleep during a meal or activity	7.09	1.38	0.82	Round 1
Resident is lightheaded or fainted	7.82	1.34	0.82	Round 1
Family/friend reported the resident had new or increased pain	7.64	1.33	0.82	Round 1
Nutrition, hydration and elimination
Resident passed no urine (new item included in Round 2)	8.65	0.44	1.00	Round 2
Resident's legs/feet are swollen (new or worsened)	7.82	0.89	0.95	Round 1
Resident had no oral intake (new item included in Round 2)	8.25	0.98	0.95	Round 2
Resident passed less urine	7.95	1.07	0.91	Round 1
Resident has blood in the urine, stool or sputum (new item included in Round 2)	7.95	1.49	0.90	Round 2
Resident has difficulty swallowing (new item included in Round 2)	8.05	1.14	0.85	Round 2
Resident coughed while eating or drinking	7.77	1.13	0.82	Round 1
Resident has new incontinence of urine and/or faeces	7.55	1.24	0.80	Round 3
Resident has had unintended weight gain or loss (new item included in Round 2)	7.55	1.70	0.80	Round 2

Abbreviations: I‐CVI—item‐content validity index; SD—standard deviation.

TABLE 3.

Items that were excluded in Part A (physical and psychological indicators of acute illness in residential aged care residents).

Items	Mean	SD	I‐CVI	Deciding round
General appearance (changes in how the resident looks)
Resident's posture or gait has changed (new item included in Round 2 survey)	7.30	1.52	0.75	Round 3
Resident's skin has become sweaty (new item included in Round 2 survey)	7.00	1.82	0.70	Round 3
The resident's skin feels hot	6.35	2.10	0.60	Round 3
The resident has dryer lips	6.35	1.82	0.55	Round 2
The resident's skin feels cold	5.80	2.18	0.45	Round 3
Functional changes (changes in how the resident performs routine tasks)
The resident needs more assistance to complete daily activities	7.40	1.02	0.75	Round 3
The resident walks less	6.75	1.44	0.65	Round 3
The resident has not participated in their usual planned activity	6.75	1.18	0.55	Round 3
The resident slept less at night	6.35	1.97	0.50	Round 2
The resident walks more	5.40	2.28	0.35	Round 2
Psychological and neurological state (changes in the resident's mood, thoughts or behaviour)
Resident's mood has changed (new item included in Round 2 survey)	6.65	1.88	0.60	Round 3
Nutrition, hydration and elimination
Resident had one or more episodes of vomiting/nausea	7.15	1.59	0.75	Round 3
Resident had one or more episodes of diarrhoea	6.65	2.37	0.60	Round 3
Resident has not had bowels open for 2 days	6.50	2.04	0.60	Round 3
Resident did not agree to take medications (new item included in Round 2)	6.80	1.25	0.60	Round 3
Resident drank less	6.85	1.45	0.55	Round 2
Resident ate less	6.65	1.46	0.55	Round 3
Resident asks for more drinks than usual	5.85	1.83	0.40	Round 2
Resident passed more urine	6.15	1.88	0.40	Round 2
Resident ate more	4.70	1.91	0.15	Round 2

Abbreviations: I‐CVI—item‐content validity index; SD—standard deviation.

TABLE 4.

Final items in Part B (care pathways).

Items	Mean	SD	I‐CVI	Deciding round
Care pathway not requiring specialised training
Escalate concerns to your immediate supervisor/shift coordinator	8.77	0.42	1.00	Round 1
Review resident's pain	7.73	1.32	0.82	Round 1
Raise emergency alert	7.75	1.13	0.80	Round 3
Initiate basic first aid (new item included in Round 2 survey)	7.20	1.82	0.79	Round 2
Care pathway requiring specialised training
Assess vital signs (e.g., BP, pulse, RR, temperature, SpO2 and pain)	8.59	0.72	1.00	Round 1
Conduct a delirium assessment	8.41	0.51	1.00	Round 1
Notify the resident's family and inform them that the changes could indicate the person is unwell or becoming unwell and may need additional treatment	8.23	0.95	0.95	Round 1
Notify the resident that the changes noted could indicate they are unwell or becoming unwell and may need additional treatment	8.45	0.89	0.95	Round 1
Discuss care options with the resident/family	8.27	1.09	0.91	Round 1
Notify the GP or other health professional, for example, call local ED	8.36	0.93	0.91	Round 1
Send assessment data to GP/ED/external assessment team and await instruction	8.23	1.00	0.91	Round 1
Conduct a comprehensive health assessment	8.00	1.45	0.90	Round 3
Review medication (new item included in Round 2)	7.80	1.54	0.90	Round 3
Conduct a neurological assessment	8.14	1.18	0.86	Round 1
Follow the resident's advance care plan	8.14	1.87	0.86	Round 1
Increase frequency of monitoring	8.00	1.28	0.82	Round 1
Transfer the resident to ED	7.55	1.83	0.80	Round 3

Abbreviations: BP—blood pressure; ED—emergency department; GP—general practitioner; I‐CVI—item‐content validity index; RAC—residential aged care; RR—respiratory rate; SD—standard deviation; SpO₂—saturation of peripheral oxygen.

TABLE 5.

Items that were excluded in Part B (care pathways).

Items	Mean	SD	I‐CVI	Deciding round
Care pathway not requiring specialised training
Stay with the resident and monitor their health	6.85	2.06	0.70	Round 3
Review resident's bowel movements	6.00	2.30	0.40	Round 3
Care pathway requiring specialised training
Follow service/organisation care pathway	6.70	2.51	0.70	Round 3
Follow external care pathway (i.e., hospital in the home programme)	6.75	2.38	0.70	Round 3
Call an ambulance to transfer the resident to an emergency department	6.75	2.36	0.70	Round 3
Check the resident's urine for abnormalities	6.80	1.50	0.68	Round 3

Abbreviations: I‐CVI—item‐content validity index; SD—standard deviation.

TABLE 6.

Items that were included in Part C (strategies to support residential aged care staff to adopt an approach to identifying and responding to changes in health that indicate acute illness).

Items	Mean	SD	I‐CVI	Deciding round
Communication approaches
Could be used by the RAC community, including family
An electronic care pathway to direct actions taken when early signs of acute illness are identified	8.20	1.10	0.90	Round 2
A paper‐based tool to support all RAC staff and families to report changes in a resident's baseline health status indicating acute illness	7.65	1.63	0.85	Round 2
An electronic tool to support all RAC staff and families to report changes in a resident's baseline health status indicating acute illness	7.90	1.21	0.85	Round 2
A paper‐based assessment tool that includes indicators of acute illness, supporting the recognition of the early changes in the resident's health	7.75	1.74	0.80	Round 2
A paper‐based care pathway to direct actions taken when early signs of acute illness are identified	7.70	1.69	0.80	Round 2
Could be used by RAC staff only
Introduce a handover procedure including the current health status of residents and the presence of new changes in the indicators of acute illness	8.00	1.15	0.90	Round 2
A digital application (app) to prompt staff and families to recognise and report early signs of acute illness and guide optimal actions	7.40	1.71	0.80	Round 3
Education programmes
For staff
Face‐to‐face education about the early signs of acute illness and escalation care pathways	7.75	1.44	0.85	Round 3
Self‐paced online education about the early signs of acute illness and escalation care pathways	7.25	2.38	0.80	Round 3
Simulation training	7.85	1.24	0.80	Round 3
Immersion training for new grad RN in aged care, such as having access to some supernumerary time in the acute environment with an experienced RN (new item included in Round 3)	8.10	1.37	0.80	Round 3
For family
Education about the signs of acute deterioration and actions to take for family (new item included from feedback in Round 3 which was the final round‐no subsequent rounds for rerating)				After Delphi a
Mentor programmes
A champion programme, for example, selected staff receive specialised training to then lead others in practice change	8.37	1.14	0.90	Round 2
A mentoring programme, for example, a RAC RN would work with a nurse practitioner to enhance assessment skills	8.16	1.47	0.85	Round 2
A targeted campaign to increase awareness such as acute deterioration awareness week (new item included in Round 3)	7.40	1.77	0.80	Round 3
Programmes involving external health professionals
Residential outreach service with clinical nurse specialist linked to geriatrician	8.45	0.86	0.95	Round 3
Assessment information that is electronically linked to the treating GP who is alerted when there is an important change in the resident health status (excluded after Round 3 based on open text responses)	7.85	1.85	0.85	After Delphi a

Abbreviations: ED—emergency department; GP—general practitioner; I‐CVI—item‐content validity index; RAC—residential aged care; RN—registered nurse; SD—standard deviation.

^a Decision to include/exclude made from feedback given by expert after Round 3; these will need to be further examined in future research.

TABLE 7.

Items that were excluded in Part C (strategies to support residential aged care staff to adopt an approach to identifying and responding to changes in health that indicate acute illness).

Items	Mean	SD	I‐CVI	Deciding round
Communication approaches
Addition of the assessment tool and care pathway in the organisation's electronic medical records to automate data collection with prompts for an appropriate care pathway	7.20	1.86	0.75	Round 3
Introduction of care routines that includes daily documentation of all residents' health status using an assessment tool	6.65	2.26	0.60	Round 3
Visual stimulus
Poster listing the early signs of acute illness	7.15	1.53	0.75	Round 3
Flip chart to show early signs of acute illness and escalation care pathways	6.95	1.86	0.70	Round 3
Keyring card with early signs of acute illness and escalation care pathways	5.90	2.68	0.50	Round 3
Over the bed chart displaying the resident's baseline health status	6.15	2.67	0.50	Round 3
Education programmes
For staff
A website where staff and families could access more information about how to detect acute illness	7.35	1.39	0.70	Round 3
Real time online education about the early signs of acute illness and escalation care pathways	6.45	2.67	0.60	Round 3
A desktop manual containing information about how to use the new resources, what the early signs of acute illness are, and options for a response to an identified change	6.20	2.58	0.60	Round 3
Mentor programmes
Community of practice for aged care staff (new item included in Round 3)	7.50	1.86	0.75	Round 3
Programmes involving external health professionals
The RN contacts the local ED to get advice regarding best treatment	6.20	2.79	0.60	Round 3
ED staff from local hospital visit unwell resident to conduct a comprehensive assessment and develop a management plan	6.15	2.52	0.50	Round 3

Abbreviations: I‐CVI—item‐content validity index; SD—standard deviation.

3.1. Part A: Physical and Psychological Indicators

A total of 36 items remained in Part A after Round 3, including (1) n = 9 general appearance items, (2) n = 5 functional changes items, (3) n = 13 psychological and neurological state items and (4) n = 9 nutrition, hydration and elimination items (see Table 2).

1. General appearance

   Nine items remained in the final survey (see Table 2) with the items receiving the highest level of agreement being: resident finding it harder to breathe, resident short of breath, resident experiencing significant blood loss and resident or family reporting worsening health status (see Table 3).

2. Functional changes

   Two items were reworded as a result of feedback in Round 1: resident communicates differently was changed to resident communicates more or less, and resident has difficulty getting out of bed was changed to the resident has a new or increased difficulty getting out of bed. Five items remained in the final survey (see Table 2) with the items receiving the highest level of agreement being: being more breathless during care activities, the resident communicates more or less, the resident had a fall or nearly fell and the resident has a new injury (see Table 3).

3. Psychological and neurological state

   Thirteen items remained in the final survey (see Table 2) with resident has new onset of increased confusion or agitation, difficult to wake or could not be woken and new onset or increase of delusions or hallucinations (new item added in Round 2) having the highest rating (see Table 2).

4. Nutrition, hydration and elimination

   Nine items remained in the final survey (see Table 2), with resident passed no urine (new item included in Round 2), legs/feet are swollen (new or worsened) and no oral intake (new item included in Round 2) rated highest. This category had the greatest number of items removed with nine items removed (see Table 3).

3.2. Part B: Care Pathways

In Part B, 17 items were included (Table 4), and six items were excluded (Table 5) after Round 3. There were four items related to pathways not requiring specialised training (i.e., regulated and unregulated care staff could initiate the action) and n = 13 items related to pathways that would require staff to have specialised training (i.e., could only be initiated by health professionals, e.g., regulated care staff).

All participants agreed that the priority actions that should be initiated when acute deterioration is to escalate concerns to your immediate supervisor/shift coordinator. All participants agreed that the priority actions of assessing vital signs and conducting a delirium assessment were important.

3.3. Part C: Strategies to Support LTC Staff to Adopt an Approach to Identifying and Responding to Changes in Health That Indicate Acute Illness

Seventeen items related to support needed to implement a new change in LTC were included in Part C and were categorised under four subheadings: (1) communication approaches, (2) visual stimulus, (3) education programmes and (4) mentor programmes. Table 6 summarises the items included in Part C, and Table 7 the items excluded. Items receiving the highest level of agreement included face‐to‐face training about the signs of acute deterioration for staff and families, champion programme and an outreach programme involving clinical nurses and doctors.

4. Discussion

This study described the consensus identification of items to be included in an EWSS to detect and respond to the early signs of acute deterioration in older people living in LTC and has generated recommendations for elements of a programme to build the capacity of the LTC workforce to recognise, report and respond to the early signs of acute deterioration of residents. Round 1, Part A included 44 clinical indicators. On completion of Round 3, 36 items remained: general appearance n = 9; functional changes n = 5; psychological and neurological state n = 13; and nutrition, hydration and elimination n = 9. Care pathways in response to acute deterioration (n = 20) were included in Part B. After Round 3, four care pathways remained. There was agreement in part C that communication aides, education and mentor programmes were necessary as viable strategies to support sustained adoption. These recommendations will improve care for people in LTC and for staff and health care services.

Clinical early warning scores are intended to alert healthcare staff to possible early indicators of acute deterioration and are a feature of the National Safety and Quality Health Service standards in Australia: Recognising and Responding to Acute Deterioration (Australian Commission on Safety and Quality in Health Care ²⁰²³). Due to the intricacies of the health needs of older people living in LTC facilities, it is imperative that specific tools be designed specifically for RACs and that sustained use of these is achieved. While there is evidence of the potential for substantial impacts to be made with an EWSS in LTC, current programmes have been developed but are not routinely used in practice.

While not specifically in LTC, a quality evidence synthesis indicated that there are a number of barriers to the escalation of concerns with an EWS protocol (O'Neill et al. ²⁰²¹). These included: lack of standardisation, resources, lack of accountability, rapid response team behaviours, fear, hierarchy, increased conflict, over confidence, lack of confidence and patient variability (O'Neill et al. ²⁰²¹). Through consultation with aged care experts, this project ensured that consensus of recommendations for the development of an EWSS for LTCs to be trialled in the future will be acceptable to the RAC community. This will promote the early adoption and sustained adherence to implemented practice changes in the future (NHMRC ²⁰¹⁸). Further, it is anticipated that the EWSS informed through this consultation is likely to empower LTC health professionals to work to their scope of practice and enhance their competence in the assessment and management of acutely unwell residents.

The agreed items in this study related to the physical and psychological indicators of acute deterioration and were noted to require little measurement, and in most cases, can be observed. The observation of indicators is consistent with findings of a recent rapid review to identify the clinical indicators of acute deterioration in LTC (Chambers et al. 2023). This review found that the ‘soft indicators’, or those that can be observed by all staff such as level of consciousness, mood, oral intake and function were among the most reliable indicators (Chambers et al. 2023; Stocker et al. ²⁰²¹).

It is interesting to compare the items with the highest level of agreement from this study with items included in existing early warning tools for LTC such as Stop and Watch (Basinska et al. 2022; Ouslander et al. ²⁰¹¹; Ouslander et al. ²⁰⁰⁹; Porter Tamara et al. ²⁰²¹). Stop and Watch includes 12 clinical indicators, most of which are observable changes and compares a current state to a previous baseline state. When the items with the highest level of agreement from this study are compared with the items in Stop and Watch, the common items include changes in mood and communication, oral intake and condition of the skin. However, our panel did not rate functional ability such as needing more help, for example, when transferring, change in weight and being weak, as highly as other indicators. Our panel rated new psychotic symptoms (hallucinations and delusions), no urine output, swollen extremities, significant blood loss and falls or near falls more highly; these are not included in the comparison tool. Another item of difference is residents communicate more or less. While Stop and Watch includes communicates less, this ignores the fact that communicating more is an early indicator of hyper‐delirium which can accompany acute illness in older people (Inouye et al. 2014).

However, health professionals do require a baseline knowledge of a resident's normal state to successfully detect these subtle changes, for example, levels of alertness and mood. This highlights the importance of thorough documentation, that is often not done, so that all involved in the care of the individual could identify change. This is an important consideration when developing future programmes or tools (Andrews et al. 2019; Parker et al. ²⁰²⁰).

Another interesting finding was the care pathways that were excluded during the Delphi process. Of particular note was the exclusion of calling an ambulance to transfer the resident to the ED (although noting that transfer to the ED was included) and following external pathways. The items included and excluded suggest some faith in the skills of the aged care workforce to manage acute deterioration in‐house. This was again reinforced when looking at the supports recommended which again excluded having external advisors to manage the situation. Previous studies have found that a lack of confidence in the aged care workforce is a factor contributing to acute deterioration not being detected and managed in LTC (Chambers et al. 2023). Our findings suggest there is an appetite to focus on changing this prevailing rhetoric and to build the capacity of the LTC workforce to work to their scope of practice and manage the treatment of acute deterioration in LTC.

The transferability of these findings should be considered given differences in care systems, workforce structures, and cultural practices across settings. The clinical indicators identified are likely to have broad applicability, given their basis in physiological changes associated with acute deterioration (Ogawara et al. 2025); however, recognition and response may vary across contexts due to differences in workforce roles, communication processes and decision‐making structures in LTC.

These findings have important implications for clinical practice in LTC. The identified clinical indicators, alongside structured care pathways and workforce support strategies, provide a foundation for strengthening early recognition, communication and response to acute deterioration.

4.1. Strengths and Limitations

The strength of this study is found in the consideration of more than just the physiological indicators; rather, the potential psychological indicators are also contemplated (Chambers et al. 2023). Furthermore, while recommendations were initially based on a systematic review, the literature was limited and hence the consensus of experts in the area was essential. Interestingly, new items brought up across the surveys were also then deemed to be important, and most were included in the final outcome. Recommendations were consistent with items that would likely be identified by both regulated and unregulated staff, which is consistent with previous tools developed. While previous studies have indicated the need for a multicomponent aspect to the implementation of an EWSS, this project has provided recommendations on care pathways and strategies, with most of these being able to occur in‐house.

While it would be usual to see a decrease in participant involvement in Delphi studies (Niederberger et al. 2025), a strength of this study was the high participation rate and continued engagement across three rounds of the survey, ensuring good representation of health professionals. There were, however, noted to be inconsistencies in the detailing of items (the tense, in particular, and the wording of items) which will require further refining. The use of the 9‐point Likert scale was problematic for some participants, as some opted for a ‘neutral’ score, and it was difficult for some respondents to rescore in one direction or the other.

5. Conclusion

The consensus of recommendations for acceptable physical and psychological clinical indicators to be included in an EWSS for acute deterioration in LTC will provide the basis of an EWSS that will be appropriate for LTC. This includes care pathways in response to acute deterioration and supporting capacity building of the LTC workforce to recognise, report and respond to the early signs of acute deterioration of residents that will lead to improvements in care for people in LTC and for staff and health care services.

Author Contributions

All authors contributed substantially to this study and take full responsibility for this article.

Funding

This work was supported by Queensland Government Nursing and Midwifery Research Fellowship.

Conflicts of Interest

The authors declare no conflicts of interest.

Clinical Resources

Recognising and responding to acute deterioration Standard Recognising and Responding to Acute Deterioration Standard|Australian Commission on Safety and Quality in Health Care

National Safety and Quality Health Service Stands (2nd Ed) National Safety and Quality Health Service Standards (second edition)|Australian Commission on Safety and Quality in Health Care

Data Availability Statement

The authors have nothing to report.