Setting
The Rive-Neuve specialized palliative care hospice, founded in 1988, is located in the French-speaking Canton of Vaud, in Western Switzerland, and can accommodate 20 inpatients. Admissions aredocumented in a file reviewed by a nurse, who is specialized in palliative care. Referrals are made by neighbouring hospitals or by the physicians of the mobile palliative home care team. Indications for hospitalization are the palliative situation of the patient, especially when symptoms are present, which are difficult to managesuch as pain, anxiety or dyspnea. In 2019, 222 patients were admitted; all of them were included in this study, since the documentation of all charts allowed the scoring with the INTERMED. The study period corresponds to the habitual situation, prior to the Covid pandemic. The staff consists of a multidisciplinary team of nurses, physiotherapists, an occupational therapist, an art therapist, a chaplain, about fifty volunteers, physicians, psychologists, psychiatrists, and administrative personal. The study was accepted by the ethics committee of the Canton of Vaud (CER-VD 2023 − 01200) - based on a request for authorization under Article 34 LRH (no need for consent) - under the condition that no patient had explicitly stated his or her disagreement for using his or her routine medical data for research. Given the aim of the study, no medical data were recorded. Comparison was intended only on a populational level or with regard to setting (e.g., palliative care patients versus patients affected by obesity). Since this is a descriptive study, we decided to include a whole year of admissions.
Assessment instruments
The study was based on a retrospective medical chart review. Sociodemographic and medical data, as well as patient symptom assessments with the ESAS (Edmonton Symptom Assessment System) were retrieved from the charts (28).
Biopsychosocial case complexity was assessed by means of the INTERMED. The INTERMED is a reliable and valid instrument to evaluate patients’ case complexity by taking into account their biopsychosocial situation and the complexity of their interactions with the care system (29, 30). The 20 items rated by the INTERMED are shown in the Appendix. Developed in 1995 (31), numerous studies have been conducted with the INTERMED in different countries, and different health care settings and patient populations. These studies have consistently shown - for example in low back pain (32, 33), chronic shoulder pain (34), diabetes (35), or internal medicine (36) - that the INTERMED identifies complex patients who have a less favorable response to medical treatments (37, 38). Moreover, early and targeted psychosocial interventions in complex patients identified by means of the INTERMED have demonstrated beneficial effects with regard to medical and psychological outcomes, as well as health care utilization (39, 40).
The INTERMED can be scored by means of a semi-structured interview, a self-assessment or retrospectively based on patient charts (41). Retrospective chart-based rating has been proved to be reliable: in a previous study: interrater reliability of retrospective ratings reached an intra class correlation (ICC) of 0.91 (41).
The INTERMED is composed of three columns for each domain (biological, psychological, social and healthcare system domain) (see Fig. 1). The first column refers to the past (history), the second to the present (current state) and the third to the future (prognosis). Two items in the past and present domains and one items in the future domain are rated with a score ranging from 0 (no complexity) to 3 (indicating the highest level of complexity). Each domain can thus reach a score of 15, with a total INTERMED score of 60. A score ≥ 21 is considered as indicating biopsychosocial complexity (42).
The amount of interdisciplinary palliative care interventions – without the care provided by medical doctors and nurses – was documented by the routine registration of minutes per patient spent by the different care professionals for 211 of the 222 patients (for 11 patients data were missing). All patients in Rive-Neuve are equally cared for by physicians and nurses, as in other medical settings. Therefore, care minutes calculated in this study only relates to the additional support provided by psychologists and psychiatrists, the chaplain, physiotherapists and occupational therapists, dietitians, social workers, hypnotherapy nurses, and others. The time spent with patients by volunteers was not collected and is therefore not taken into account in the calculations.
Statistical Analysis
Initially, descriptive statistics for the INTERMED total and domain scores were conducted for admission and the end of the stay (death or end of hospitalization). Differences between groups were analyzed using t-tests for independent or dependent groups. Percentages were calculated to estimate the prevalence of patients with complex health care needs. Correlations between the INTERMED change scores calculated by INTERMED T2-T1 and amount of palliative care provision were estimated by calculating Pearson’s correlation coefficients. Hierarchical cluster analysis was conducted, to investigate if the sample might consist of clearly distinct subgroups. It was worked out thoroughly using several criteria to check for the existence of more than one main cluster. However, cluster analysis - as a relatively rough method to determine clusters - did not result in easily separable groups. To further investigate the main sources of variance with respect to the INTERMED items, we conducted a principal component analysis (PCA). Results identified two factors with an Eigenvalue > 1 with one main factor (1) explaining 35.5% and the second factor (2) explaining 13.5% of the variance. For all 20 items of the INTERMED, the factor loadings were calculated. Items with a high positive factor loading on the main Factor 1 were related to psychological or social health care needs (see Appendix).