Background
Patients approaching the ends of their lives, their relatives and clinicians all value accurate prognostic information [
1‐
5]. This information is routinely provided by clinicians using their clinical intuition. However, clinicians’ estimates of survival (CES) are often inaccurate and over-optimistic [
6]. The need for more accurate methods of prognostication was highlighted by the Neuberger report [
7] into the short-comings of the implementation of the Liverpool Care Pathway [
8].
Determining more accurately how long patients with advanced cancer have left to live would enable both patients and their relatives to make plans for their future [
3]. It would also aid clinicians to target treatments to those patients most likely to benefit and it would safeguard other patients from receiving treatments that they are unlikely to benefit from [
9]. Prognostic information may help clinicians to plan services and to ensure that patients are cared for in the most appropriate environment. More reliable prognostic estimates may also facilitate the identification of patients for inclusion on palliative care registers [
10] and the prioritisation of patients who are referred to palliative care services.
The Prognosis in Palliative care Scales (PiPS) are predictive models of survival and were previously developed by members of our research team in order to provide an objective aid to clinicians’ intuition [
9]. The original study recruited prospectively a cohort of 1018 patients with advanced cancer, who were no longer undergoing disease-modifying treatment. This was a multi-centre study involving 18 specialist palliative care services across England. Separate prognostic models were created for patients without (PiPS-A) or with (PiPS-B) available blood test results. The PiPS scores were able to predict whether a patient was likely to live for “days” (less than 14 days), “weeks” (between 2 and 7 weeks) or “months” (2 months or longer). These categories were chosen as they seemed to have the greatest face validity among clinicians. Both PiPS models were shown to perform as well as CES. The PiPS-B prognostic estimate was found to be significantly better than doctors’ prognostic estimates. Before recommending PiPS for use in routine clinical practice, it is important to check that the models provide accurate and reliable estimates of survival in a new group of patients.
More recently, two research studies have published validation data about the PiPS models [
11,
12] in different clinical settings. Baba and colleagues (2015) [
11] undertook an independent validation of the PiPS models in Japanese cancer patients and reported that the PiPS instrument performed as well as in Gwilliam’s original study [
9]. However, the Japanese study did not compare the accuracy of PiPS to the accuracy of CES. Differences in cancer epidemiology and oncology practice between Japan and UK may also be a relevant consideration. The study by Kim and co-workers (2014) [
12] reported that the PiPS instruments performed approximately as well as in the original study. However, the Kim study included a relatively small number of participants and the study population was restricted to palliative cancer patients in a specialist cancer hospital in South Korea. No large scale validation has yet been undertaken in the UK.
Taken together, these two studies strengthen the case for undertaking a large-scale validation study in the UK using CES as a comparator. Based on systematic reviews [
13,
14], four other prognostic models have been identified that might also be useful in clinical practice and which are in need of further evaluation. These are the Palliative Prognostic Index (PPI) [
15], the Palliative Performance Scale (PPS) [
16], the Palliative Prognostic (PaP) [
17] score and the Feliu Prognostic Nomogram (FPN) [
18].
The PPI and the PPS can both be calculated without the need for a blood test (like PiPS-A). The PPI model stratifies patients into three groups; survival shorter than 3 weeks; shorter than 6 weeks; or more than 6 weeks [
15]. The PPI has shown a high accuracy level in patients with short estimates of survival [
19]. The PPS is a measure of functional status and is one of the variables included in the PPI score. Although not specifically designed as a prognostic instrument, and therefore lacking some face validity as a stand-alone prognostic tool, the PPS has been found to have prognostic significance in patients with advanced disease [
20,
21]. Using data from large observational studies, the PPS can be used to predict the probability of dying across a range of survival times [
20].
The PaP and the FPN, both require blood test results (like PiPS-B). The PaP score classifies patients into three risk groups based on a 30-day survival probability of less than 30%; between 30 and 70%; and more than 70% [
17]. There is increasing evidence to support its validity in a variety of settings [
22‐
26]. The total PaP score was shown in one study to be more accurate than a simple clinician prediction of survival [
27]. One practical concern with the PaP score is that it relies on CES. This can make the PaP challenging to use when clinicians are unsure about survival times or when an “objective” estimate is required that is free from the influence of CES. The FPN predicts survival at 15, 30 and 60 days [
18]. In one study the FPN was found to be more accurate than the PaP [
18] and does not rely on subjective CES.
The PiPS2 study was designed with the overall aim of validating the accuracy of the PiPS-A and PiPS-B [
9]. The primary objectives of the study are: to validate the PiPS models and; to compare the performance of PiPS-B against CES. The secondary objectives are to validate other selected prognostic models (PPI, PPS, PaP and FPN).
A nested qualitative sub-study has also been embedded in the PiPS2 study to assess the acceptability of the prognostic models to patients, carers and clinicians and to identify barriers and facilitators to clinical use. This is particularly relevant because in clinical practice it is often the relatives and carers of non-competent or semi-conscious patients who most wish to have access to accurate prognostic information.
Nested qualitative sub-study
Semi-structured, face to face interviews were conducted with a purposive sample of approximately 15 patients, 15 carers and 15 clinicians to determine the acceptability of using prognostic indicators, and barriers and facilitators to clinicians’ use. Data saturation will determine the final sample sizes, which may be larger or smaller than anticipated. To date, 28 patients, 19 carers and 32 clinicians were enrolled in the qualitative sub-study.
Interviews were audio recorded, and used iterative topic guides based on reviews of the literature, themes arising from preceding interviews, and the MORECare recommendations for conducting research at the end of life [
41] The patient and carer interviews explored whether participants wish to know the patient’s prognosis, and if so whether they would have preferred clinicians’ estimates of survival or prognostic modelling. Those wishing to know their prognosis were asked how this information should have been presented to them. Interviews were of 30–60 min duration to ensure that participants were not overburdened.
The clinician interviews were interactive and explored the acceptability of PiPS, PPI, PPS, PAP and FPN. Participants were shown the prognostic models, and commented on their utility. Topic guides included questions about potential barriers and facilitators to using the models, and to discussing prognostic information with patients and carers. Interviews were of 60 min duration to allow enough time for in-depth discussion.
Interview data will be entered into NVivo 10.0 (
https://www.qsrinternational.com/nvivo/) and analysed using the five stages of Framework Analysis [
42]: familiarisation, developing a thematic framework, indexing, charting, and mapping and interpretation.
Acknowledgements
We would like to thank the UCL PRIMENT Clinical Trials Unit for their support, Karolina Christodoulides and Jane Harrington for their help with administrative tasks and data monitoring, and Florence Todd-Fordham for contributing to data quality control procedures. We would like to thank Peter Buckle and Dori-Anne Finlay for acting as Patient and Public Representatives for the study. We would finally like to thank all the patients who participated in this study and our collaborators across participating sites.