Article Text

Development and evaluation of the Supportive and Palliative Care Indicators Tool (SPICT): a mixed-methods study
  1. Gill Highet1,
  2. Debbie Crawford1,
  3. Scott A Murray2 and
  4. Kirsty Boyd1
  1. 1Palliative Care Service, Royal Infirmary of Edinburgh, Edinburgh, UK
  2. 2Primary Palliative Care Research Group, Centre for Population Health Sciences, University of Edinburgh, Medical School, Edinburgh, UK
  1. Correspondence to Dr Kirsty Boyd, Palliative Care Service, Royal Infirmary of Edinburgh, 51 Little France Crescent, Edinburgh EH16 4SA, UK; kirsty.boyd{at}ed.ac.uk

Abstract

Objective To refine and evaluate a practical, clinical tool to help multidisciplinary teams in the UK and internationally, to identify patients at risk of deteriorating and dying in all care settings.

Methods We used a participatory research approach to refine the 2010 Supportive and Palliative Care Indicators Tool (SPICT) and evaluate its use in clinical practice. We conducted an ongoing peer review process for 18 months via an open access webpage, and engaged over 30 clinicians from the UK and internationally in developing an effective tool. Secondly, we carried out a prospective case-finding study in an acute hospital in SE Scotland. Four multidisciplinary teams identified 130 patients with advanced kidney, liver, cardiac or lung disease following an unplanned hospital admission.

Results The SPICT was refined and updated to consist of readily identifiable, general indicators relevant to patients with any advanced illness, and disease-specific indicators for common advanced conditions. Hospital clinicians used the SPICT to identify patients at risk of deteriorating and dying. Patients who died had significantly more unplanned admissions, persistent symptoms and increased care needs. By 12 months, 62 (48%) of the identified patients had died. 69% of them died in hospital, having spent 22% of their last 6 months there.

Conclusions The SPICT can support clinical judgment by multidisciplinary teams when identifying patients at risk of deteriorating and dying. It helped identify patients with multiple unmet needs who would benefit from earlier, holistic needs assessment, a review of care goals, and anticipatory care planning.

  • Chronic conditions
  • Clinical decisions
  • Prognosis
  • Supportive care
  • Programme development
  • Tool use behaviour

Statistics from Altmetric.com

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.

Introduction

Timely identification of people who are at risk of deteriorating and dying is a prerequisite for effective end-of-life care and a key element of health policies in the UK and internationally.1–7 Identification prompts clinicians to initiate proactive holistic needs assessment, shared decision making about goals of care and anticipatory care planning. However, patients are often identified too late in their illness. General practitioners and hospital specialists both report difficulties with earlier recognition.8–10 Effective triggers for patient identification are underdeveloped in contrast with other more established tools for end-of-life care planning and service delivery.1 ,2 ,5 The emphasis has been on predicting prognosis more accurately in order to introduce palliative care at an appropriate time. The widely used ‘Surprise question’ (Would you be surprised if this patient died within the next 12 months?) is combined with clinical indicators of advanced conditions in referral criteria for hospice and palliative care in the USA and in the UK Gold Standards Framework Prognostic Indicator Guidance.4 ,11 If used alone, the surprise question may identify too many patients for palliative care; a third of all inpatients in a recent UK hospital prevalence study.12 Mortality risk assessment models for single diagnoses are often used by hospital specialists but are of limited value because multimorbidity is now the norm in those with advanced, long-term conditions.13 ,14 Even in patients with cancer, scores based on biomarkers gave similar prognostic estimates to a multidisciplinary team assessment and were less accurate with longer survival times.15

Linking identification of patients for supportive and palliative care with estimates of prognosis means that clinicians tend to focus on when to start ‘planning for dying’ and may delay a review of care goals and unmet needs until the last weeks of life.16 Earlier identification requires a change from estimating when a patient may die to identifying those who are at sufficient risk of deteriorating and dying for proactive assessment and care planning to be appropriate. This includes people with advanced, progressive, incurable conditions, general frailty and coexisting conditions, and those at risk of dying from a sudden acute crisis in their condition or from life-threatening acute conditions caused by sudden catastrophic events.3 ,17

The Supportive and Palliative Care Indicators Tool (SPICT 2010) was designed to provide practical, evidence-informed guidance to help clinicians working in primary and secondary care recognise when their patients might be at risk of dying and likely to benefit from supportive and palliative care in parallel with appropriate ongoing management of their advanced conditions.18 The SPICT supports multidisciplinary assessments of individual patients based on clinical observations, performance status, symptoms, multimorbidity, illness trajectories and the patient's views and goals.5 This paper reports the development and evaluation of the tool.

Methods

Since April 2011, we have developed and evaluated the SPICT using a mixed-methods, participatory approach.19 The research team included palliative care specialists, primary care clinicians, hospital clinicians, and a senior social scientist as the project researcher. Peer review and consensus building contribute to the development of tools with good face validity and utility in routine clinical practice, so we offered open access, via a designated webpage, to each of the 15 major revisions of the SPICT that were developed over an 18-month period. This is now the SPICT project website (http://www.spict.org.uk). An international, electronic mailing list was built from a wide spectrum of interested clinicians and policy makers who contacted us from the UK, Europe, USA, Canada, New Zealand, Australia and Africa. We also invited primary and secondary care leads for palliative care in Scotland, England and Ireland to participate. After each amendment had been agreed by the project steering group and integrated into the next version of SPICT, email alerts were sent to all our peer reviewers seeking comments and suggestions until no further changes were proposed.

At the same time, we carried out a prospective, case-finding study of patients with advanced kidney, liver, cardiac, or lung disease following an unplanned admission to a tertiary, acute hospital in SE Scotland. The research team worked alongside clinical staff in each of the four participating units in turn, supporting data collection, analysis and interpretation.20 Senior nursing staff and specialty registrars agreed to screen all their patients soon after an unplanned admission using a checklist of SPICT clinical indicators for an 8 week period. The checklist contained the SPICT general indicators, the surprise question and disease-specific indicators relevant to the patients in each unit (see online supplementary appendix 1: Liver unit checklist). The staff recorded demographic and clinical details, admission outcome and service use data (unplanned admissions and hospital bed days) for all patients with the SPICT indicators of advanced conditions. The patients were then followed for 12 months by the research team using the hospital electronic patient record system. These data were analysed using a customised Excel database. χ2 Tests for categorical variables were used to compare the SPICT indicators in patients who died and those still alive at 12 months. Scottish Patients at Risk of Readmission and Admission (SPARRA) scores were calculated by the Scottish Information Services Division.21 SPARRA scores are a proxy measure of deteriorating health because patients in the last year of life have increasingly frequent unplanned hospital admissions.22 We calculated the Charlson renal comorbidity index for patients with kidney disease and the UK end-stage liver disease (UKELD) score for liver unit patients.23 ,24

Results

SPICT redesign

Six parameters underpinned the conceptual design of the SPICT, and our peer reviewers agreed they should be retained (box 1).

Box 1

Supportive and Palliative Care Indicators Tool (SPICT): key elements

  • Simple, one-page format.

  • Readily identifiable general indicators of deteriorating health commonly present in advanced conditions.

  • Evidence-based clinical indicators of all the major advanced, life-limiting conditions and multimorbidity.

  • Good face validity for health and social care professionals working in hospital, community and care home settings in the UK and internationally.

  • Promotes early supportive and palliative care in parallel with optimal management of any underlying conditions as part of routine clinical practice.

  • Contains accessible language and concepts that can be used to initiate discussions with patients and families about goals of care and improve communication between professionals/teams.

The 2010 SPICT indicators were reviewed and compared with similar tools identified from the literature to achieve consensus about the items for each section and their wording.6 ,7 ,11 ,25 General indicators of deteriorating health identified in several other palliative care assessment tools, or found in disease-specific mortality tools, are in the first section of the SPICT. We chose the Malnutrition Universal Screening Tool for the weight loss indicator as it is recommended for use throughout National Health Service (NHS) Scotland.26 Descriptions of symptoms were used instead of disease severity scores, for example, in advanced lung and heart disease, to make the tool more accessible (see online supplementary appendix 2: SPICT 2013). Many of the hospital clinicians in our case study and peer review group, and some primary care professionals, said the surprise question was of limited value if they did not know the patient well, or if the patient was at risk of dying but currently more stable. Our collaborative approach has led to continuing partnerships with projects using the SPICT to identify patients in primary and secondary care settings.7 ,27

Implementation and evaluation of SPICT

Ward staff completed checklists containing demographic details and SPICT clinical indicators for 130 patients with advanced organ failure following an emergency hospital admission (table 1). The renal team identified 41% of their unplanned admissions (55, n=133) and the liver unit 83% (35, n=42). The cardiology unit screened 570 admissions but only identified 2.8% (16 patients) due to a change in the types of patients being admitted to the ward after the study was planned. Workload pressures meant the respiratory unit struggled to screen all admissions, but the participating ward staff identified 24 patients using the SPICT checklist by asking senior ward staff or the respiratory specialist nurses to direct to them newly admitted patients with advanced lung disease. The major diagnoses were vascular disease, alcoholic liver disease, ischaemic heart disease, and chronic obstructive pulmonary disease. Multimorbidity, defined as two or more active long-term conditions, was common (77%). Patients were usually admitted with acute medical problems relating to their underlying advanced conditions. Almost all the patients were discharged home but 35% (45) died within 6 months. By 12 months, 48% (62) had died; 69% (43) of these patients in hospital (table 2). Most patients had a short final admission (median 10 days, range 1–60 days) and half the deaths were within 88 days of patients first being identified. Five patients had a sudden terminal event. However, the majority followed an archetypical organ failure trajectory and died of multiple complications of their advanced conditions, usually after an acute deterioration on a background of progressive decline.

Table 1

Demography of Supportive and Palliative Care Indicators Tool (SPICT)-identified patients (130)

Table 2

Outcomes for Supportive and Palliative Care Indicators Tool (SPICT)-identified patients

The SPICT indicators identified in patients who died and those still alive at 12 months are shown in figure 1. A significantly greater number of patients who died, compared with those who survived for over 12 months, had: two or more unplanned admissions in the 6 months before they were screened (69% vs 37%, p<0.001, 95% CI on the difference=33±16), persistent symptoms despite optimal treatment of their underlying conditions (60% vs 24%, p<0.001, 95% CI 36±16), and increased care needs (40% vs 13%, p<0.001; 95% CI 27±15). There was some variation between patient groups. A poor or deteriorating WHO performance status affected 60% in total, but only a quarter of the liver unit cases. Weight loss occurred in a third of renal unit patients but in less than 5% of the other groups. The SPICT performed as well as the standard mortality tools. The renal unit patients who died had a mean Charlson renal morbidity index of 7.9; a score of 6+ predicts a 1-year survival probability of about 0.6. In the liver unit, SPICT identified patients had a mean UKELD score of 57; a score of 60 predicts a 50% 1-year survival in advanced liver disease. The ward teams identified many of the patients who had died by 12 months (79%) with the surprise question. They thought that 71% of the group as a whole was at risk of dying within 12 months. Despite this perception that many of the identified patients might die, relatively few had a ‘Not for cardiopulmonary resuscitation’ form in place at the time of their initial SPICT screening; 24% overall and 39% of those who died.

Figure 1

Indicators of deteriorating health in Supportive and Palliative Care Indicators Tool (SPICT)-identified patients.

In addition to documenting how many of the patients had two or more admissions when screened with the SPICT, we recorded admissions and bed days during the last 6 months of life for the 62 patients who died. These patients had a mean of 2.4 unplanned admissions (range 1–7), a median number of hospital bed days of 34.5 (range 3–130 days), and spent 22% of their last 6 months of life in hospital. We also recorded these service use outcomes for the 68 patients who survived during the first 6 months after they were identified. They had a mean of 2.2 admissions (range 1–7), a median number of hospital bed days of 21.5 (range 3–167 days) and spent 17% of that 6 months in hospital. This high use of hospital services by all the patients identified was reflected in the SPARRA scores at discharge following the initial admission. These showed a 65% median readmission risk for those who died and a 60% risk in those who were still alive.

Discussion

The SPICT is a simple tool consisting of easily recognised, clinical indicators of advanced illness. It can support the clinical judgment of multidisciplinary teams in primary and secondary care as they seek to identify patients who might benefit from supportive and palliative care assessment and care planning. The SPICT has been further developed and evaluated in this study using a mixed-methods approach that combined peer review with a detailed case study and web-based dissemination. The tool is now being used in clinical practice by SPICT partner projects caring for patients with a wide range of advanced illnesses in hospitals, the community and care homes in the UK and internationally.7 ,26 In our hospital-based, case study of patients with advanced non-malignant illnesses, those who died had a higher frequency of SPICT general indicators than those who survived for 12 months, particularly unplanned hospital admissions, persistent symptoms and increasing care needs.

More emphasis on identifying patients with advanced conditions who are at risk of deteriorating and dying will encourage clinicians to consider a review of care goals and interventions to improve quality of life earlier in the illness trajectory. Simple identification tools based on accepted clinical indicators, such as SPICT, are needed to support this change in practice. Better identification, holistic assessment and anticipatory care planning for future deterioration, could help to reduce the substantial risk of unplanned hospital admissions and prolonged inpatient stays experienced by all the patients we identified.21 ,27 ,28 Primary care teams have previously indicated that they would welcome guidance from hospital specialists about when to put patients on their palliative care registers or open discussions about future care goals.5 ,25 ,29 The SPICT indicators offer guidance on which patients are at risk of deteriorating, and describe changes in health status and care needs that can be used as a starting point for these discussions. The SPICT has been developed and refined in partnership with clinicians working in hospitals, general practice and other community settings with the specific aim of producing a common set of clinical indicators that can support effective communication, shared care and a range of palliative care service developments for patients who transition frequently between care settings.27 Additionally, it identifies patients with advanced liver disease, a rapidly rising cause of mortality among people under 70 years in the UK,30 as well as the many patients with multiple advanced conditions including frailty and dementia. Future work could address the value of combining the SPICT with established needs assessment tools designed to help clinicians who provide generalist palliative care to identify patients who have more complex needs.31 ,32 Some of these patients and families will benefit from referral for specialist palliative care, as indicated in the SPICT section on assessment and planning.33

Although the hospital ward staff told the research team that they were aware of a group of patients with advanced illness who were often readmitted with an acute or chronic deterioration, and who might well die in the next year, a substantial number of the SPICT-identified patients were still for cardiopulmonary resuscitation despite having a very low chance of a medically successful outcome. Identification is only the first step in addressing the multiple, complex barriers that hinder successful introduction of earlier palliative care in acute hospital settings.12 ,16 Primary care clinicians have said they prefer an individualised, needs-driven approach to introducing palliative care, and often use triggers such as poorly controlled pain or a requirement for increased support from the multidisciplinary team.29 ,34 Combining population and service-based screening with individual patient assessments by clinicians, undertaken as part of routine care, could be the most effective way to integrate public health and patient-focused approaches to earlier identification.

Strengths and limitations

The SPICT has been developed from the available evidence base and subject to thorough peer review by a wide range of clinicians who see the need for a pragmatic tool to improve clinical practice. We conducted a case study in one urban area in SE Scotland with relatively small numbers of patients and few from black or minority ethnic groups. We faced the challenge of clinical research in busy care settings, and this limited our ability to survey all admissions in some wards. We did not collect data on outcomes for the patients who were screened but not identified as this was not feasible. Future studies could evaluate the sensitivity and specificity of the SPICT in larger populations. However, it is the identification of individual patients at sufficient risk of deteriorating and dying to prompt a review of their care that is important. The effectiveness of the SPICT in clinical practice needs to be evaluated more extensively in different settings and with a wider range of patients. Although our peer reviewers came from diverse professions, settings and countries, they were a self-selected sample of interested professionals and policy makers. Ongoing peer review and comment via the website should ensure further refinements are made to the tool over time by a wider group of participants.

Conclusions

Key components of effective palliative and end-of-life care are patient identification, assessment and care planning. Unless screening patients with advanced conditions becomes routine practice in all care settings, many patients will not benefit from the improvements being made to the care of people who are at risk of deteriorating and dying. A simple tool to support better identification of patients with unmet general palliative care needs is essential, and the SPICT addresses this gap. Shifting the emphasis to a focus on quality of life as the main goal of care is challenging. The SPICT provides clear indicators which clinicians can use to initiate conversations with patients and families about their preferences for different treatment and care options. Additionally, adopting a common, shared tool promotes effective communication and coordination between primary and secondary care teams.

Acknowledgments

We would like to thank the multidisciplinary teams in our four partner units who collaborated enthusiastically with the research team in data collection and in interpreting the findings, and all those colleagues who contributed to reviewing each version of the SPICT, particularly those from University Hospitals Coventry and Warwickshire NHS Trust and the ‘Coordinate My Care’ group.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • Contributors KB and SM designed the study and KB was the principal investigator; GH was responsible for the data collection supported by KB and DC; KB designed the SPICT project website; KB, GH, DC and SM were involved in data analysis and interpretation. The final manuscript was written by KB with contributions from the other authors. All the authors read and approved the submitted version.

  • Funding The study was funded by a small grant from the NHS Lothian Health Services Research Unit (HSRU004).

  • Competing interests None.

  • Ethics approval The study was approved by the NHS Lothian research ethics committee and the NHS Lothian Caldicott Guardian.

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

  • Data sharing statement The project Excel database has been retained by the principal investigator and contains no patient identifiable data. Further information can be obtained from the corresponding author.