Search

Objective: To summarize the available literature on using machine learning (ML) for palliative care practice as well as research and to assess the adherence of the published studies to the most important ML best practices. Methods: The MEDLINE database was searched for the use of ML in palliative care practice or research, and the records were screened according to PRISMA guidelines. Results: In total, 22 publications using machine learning for mortality prediction (n = 15), data annotation (n = 5), predicting morbidity under palliative therapy (n = 1), and predicting response to palliative therapy (n = 1) were included. Publications used a variety of supervised or unsupervised models, but mostly tree-based classifiers and neural networks. Two publications had code uploaded to a public repository, and one publication uploaded the dataset. Conclusions: Machine learning in palliative care is mainly used to predict mortality. Similarly to other applications of ML, external test sets and prospective validations are the exception.

BACKGROUND Quality in health care is a complex framework with many components. The word "quality" is used in different official settings and different contexts (public health, certification, patient safety). On individual and team levels, the perception of quality is heterogenous, and the term is often used beyond the theoretical framework. Therefore, it remains a challenge to describe the perceived quality of care in the clinical setting. The aim of this paper is to present a simple concept that can be used to visually define the perceived quality of care for the individual health care professional. METHODS/CONCEPT An experience-based concept that uses different levels of "quality of care" individually to guide the supervision of health care professionals (residents) and quality goal setting in teams is presented, with the assumption that the ambition of any health care professional is to provide excellence in care. Three perceived levels of quality of care are defined, described, and visualized, namely, a) security, b) comfort, and c) perfection. The "comfort level" defines a sustainable level of care where the optimal balance between good patient care and resource use is achieved. Excellence of care is located between the comfort and the perfection level. The practical application of this proposed concept is described in three settings, namely, 1) the threshold for asking advice from the supervisor (resident physicians), 2) in supervision/coaching discussions between residents and supervisors, and 3) in the analysis of perceived quality of care and goals setting within the team. CONCLUSION A simplified, purpose-built but well-defined concept to visually depict the perception of quality of care by clinicians can be useful in clinical practice, for the supervision of residents and for team dynamics.

Additional file 5. Original version (English) of the “Needs Assessment Tool: Progressive Disease-Heart Failure (NAT: PD-HF)”.

Additional file 6. Changes made for the cultural adaptation.

Additional file 8. Table 2. Frequency of answers from the first application of the tool.

Additional file 7. Table 1. Matrix of the weights used to assess inter-rater reliability and test-retest reliability.

Additional file 9. Table 3. Sensitivity analysis to assess the inter-rater reliability for each one of second evaluators.

Anhand einer schriftlichen Befragung sollten folgende Fragen beantwortet werden: Stand der Integration von Palliative Care als Konzept in Institutionen der Krankenversorgung und der Betreuung von alten Menschen in der Schweiz (Spitäler, Alters- und Pflegeheime, ambulanter Sektor und Spezialeinrichtungen), Identifizierung des Angebotes der Grundversorgung in Palliative Care sowie von Angeboten der spezialisierten Versorgung in Palliative Care. Dies auch im Hinblick auf ein internationales Benchmarking und die Weiterentwicklung eines regelmässigen nationalen Monitoring-Instruments.

Additional file 3. Template of the survey to health care personnel.

Additional file 2. Template of the survey to patients.

Background Few hospitals and heart failure (HF) clinics offer concurrent palliative care (PC) together with life-prolonging therapies. To know the prevalence of patients in HF clinics needing PC and useful tools to recognize them are the first steps to extending PC in those settings. However, it is still unknown whether tools commonly used to identify patients with HF needing PC can correctly distinguish them. Two systematic reviews found that the NECesidades PALiativas (NECPAL) tool was one of the two most commonly used tools to asses PC needs in HF patients. Therefore, we assessed 1) the prevalence of PC needs in HF clinics according to the NECPAL tool, and 2) the characteristics of the patients identified as having PC; mainly, their quality of life (QoL), symptom burden, and psychosocial problems. Methods This cross-sectional study was conducted at two HF clinics in Colombia. We assessed the prevalence of PC in the overall sample and in subgroups according to clinical and demographic variables. We assessed QoL, symptom burden, and psychosocial problems using the 12-Item Short-Form Health Survey (SF-12), the Kansas City Cardiomyopathy Questionnaire (KCCQ) and the Edmonton Symptom Assessment System (ESAS). We compared the results of these tools between patients identified as having PC needs (+NECPAL) and patients identified as not having PC needs (–NECPAL). Results Among the 178 patients, 78 (44%) had PC needs. The prevalence of PC needs was twice as nigh in patients NYHA III/IV as in patients NYHA I/II and almost twice as high in patients older than 70 years as in patients younger than 70 years. Compared to –NECPAL patients, +NECPAL patients had worse QoL, more severe shortness of breath, tiredness, drowsiness, and pain, and more psychosocial problems. Conclusion The prevalence of PC needs in outpatient HF clinics is high and is even higher in older patients and in patients at more advanced NYHA stages. Compared to patients identified as not having PC needs, patients identified as having PC needs have worse QoL, more severe symptoms, and greater psychosocial problems. Including a PC provider in the multidisciplinary team of HF clinics may help to assess and cover these needs. Supplementary Information The online version contains supplementary material available at 10.1186/s12904-021-00850-y.

Additional file 4. Template of the interview to assess face validity, applicability, relevance and acceptability of the tool among health care personnel.

OBJECTIVE The Surprise Question (SQ) is a prognostic screening tool used to identify patients with limited life expectancy. We assessed the SQ's performance predicting 1-year mortality among patients in ambulatory heart failure (HF) clinics. We determined that the SQ's performance changes according to sex and other demographic (age) and clinical characteristics, mainly left ventricular ejection fraction (LVEF) and the New York Heart Association (NYHA) functional classifications. METHODS We conducted a prospective cohort study in two HF clinics. To assess the performance of the SQ in predicting 1-year mortality, we calculated the sensitivity, specificity, positive and negative likelihood ratios, and the positive and negative predictive values. To illustrate if the results of the SQ changes the probability that a patient dies within 1 year, we created Fagan's nomograms. We report the results from the overall sample and for subgroups according to sex, age, LVEF and NYHA functional class. RESULTS We observed that the SQ showed a sensitivity of 85% identifying ambulatory patients with HF who are in the last year of life. We determined that the SQ's performance predicting 1-year mortality was similar among women and men. The SQ performed better for patients aged under 70 years, for patients with reduced or mildly reduced ejection fraction, and for patients NYHA class III/IV. CONCLUSIONS We consider the tool an easy and fast first step to identify patients with HF who might benefit from an advance care planning discussion or a referral to palliative care due to limited life expectancy.

BACKGROUND The Needs Assessment Tool: Progressive Disease-Heart Failure (NAT: PD-HF) is a tool created to assess the needs of people living with heart failure and their informal caregivers to assist delivering care in a more comprehensive way that addresses actual needs that are unmet, and to improve quality of life. In this study, we aimed to (1) Translate the tool into German and culturally adapt it. (2) Assess internal consistency, inter-rater reliability, and test-retest reliability of the German NAT: PD-HF. (3) Evaluate whether and how patients and health care personnel understand the tool and its utility. (4) Assess the tool's face validity, applicability, relevance, and acceptability among health care personnel. METHODS Single-center validation study. The tool was translated from English into German using a forward-backward translation. To assess internal consistency, we used Cronbach´s alpha. To assess inter-rater reliability and test-retest reliability, we used Cohen´s kappa, and to assess validity we used face validity. RESULTS The translated tool showed good internal consistency. Raters were in substantial agreement on a majority of the questions, and agreement was almost perfect for all the questions in the test-retest analysis. Face validity was rated high by health care personnel. CONCLUSION The German NAT: PD-HF is a reliable, valid, and internally consistent tool that is well accepted by both patients and health care personnel. However, it is important to keep in mind that effective use of the tool requires training of health care personnel.

Additional file 2. Template of the survey to patients.

Additional file 3. Template of the survey to health care personnel.

Additional file 4. Template of the interview to assess face validity, applicability, relevance and acceptability of the tool among health care personnel.

Background The Needs Assessment Tool: Progressive Disease-Heart Failure (NAT: PD-HF) is a tool created to assess the needs of people living with heart failure and their informal caregivers to assist delivering care in a more comprehensive way that addresses actual needs that are unmet, and to improve quality of life. In this study, we aimed to (1) Translate the tool into German and culturally adapt it. (2) Assess internal consistency, inter-rater reliability, and test–retest reliability of the German NAT: PD-HF. (3) Evaluate whether and how patients and health care personnel understand the tool and its utility. (4) Assess the tool’s face validity, applicability, relevance, and acceptability among health care personnel. Methods Single-center validation study. The tool was translated from English into German using a forward–backward translation. To assess internal consistency, we used Cronbach´s alpha. To assess inter-rater reliability and test–retest reliability, we used Cohen´s kappa, and to assess validity we used face validity. Results The translated tool showed good internal consistency. Raters were in substantial agreement on a majority of the questions, and agreement was almost perfect for all the questions in the test–retest analysis. Face validity was rated high by health care personnel. Conclusion The German NAT: PD-HF is a reliable, valid, and internally consistent tool that is well accepted by both patients and health care personnel. However, it is important to keep in mind that effective use of the tool requires training of health care personnel. Supplementary Information The online version contains supplementary material available at 10.1186/s12955-021-01817-6.

Background: People living with heart failure (HF) have a wide range of physical and psychological symptoms and comorbidities. These affect their quality of life throughout the HF trajectory. Therefore, patients with HF, especially in advanced stages of this disease, may benefit from palliative care (PC). However, despite PC being widely recommended, few hospitals and HF clinics offer concurrent PC along with life���prolonging therapies. Due to the unpredictability of the HF disease trajectory, prognostication is challenging, as well as identifying patients who might benefit from PC or have unmet needs. Additionally, patients with HF and implantable cardioverter efibrillators (ICD) have specific end���of���life (EoL) device management needs. Aims: In this thesis, I contribute knowledge about the challenges and gaps in the delivery of PC to patients with HF. Through the 4 articles that form my thesis, I provide knowledge about the challenges of identifying patients who might benefit from PC or those who have unmet PC needs. I evaluate the performance of the surprise question (SQ) predicting 1���year mortality in ambulatory HF clinics, in article 1. In article 2, I assess the prevalence of patients with PC needs in outpatient HF clinics with the NECesidades PALiativas (NECPAL) or Palliative Needs tool. Additionally, I use the NECPAL tool to identify patients with HF who might benefit from PC. For article 3, I evaluated the psychometric characteristics of the German Needs Assessment Tool: Progressive Disease ��� Heart Failure (NAT: PD���HF). Finally, I quantify gaps from anticipatory care planning and EoL care of patients with ICD in article 4. Methods: I used a cohort of 178 patients from 2 ambulatory HF clinics in Colombia for articles 1 and 2. To assess the performance of the SQ to predict 1���year mortality (article 1), I consulted Colombia���s national mortality register for participants��� 1���year vital status. To assess the NECPAL tool���s identification of patients with HF who might benefit from PC (article 2), I conducted a cross���sectional analysis that compared health���related quality of life and physical and psychosocial problems, between patients needing (+NECPAL group) and not needing (���NECPAL group) PC. To validate the German NAT: PD���HF, I used a singlecenter study at Inselspital���s Heart Failure Clinic in Bern, Switzerland (article 3). The tool was translated from English into German using a forward���backward translation. I assessed the German NAT: PD���HF���s psychometric characteristics, including internal onsistency, inter���rater reliability, test���retest reliability, and face validity. I conducted a systematic review and meta���analysis for article 4. Results: These studies��� results are presented in 4 articles. Article 1 shows the SQ���s sensitivity to predict 1���year mortality is 85% and its specificity is 57%. The SQ���s positive and negative likelihood ratios were 1.98 and 0.26, respectively. Its performance was similar among women and men, yet performed better in patients younger than 70 years, in patients with reduced or mildly reduced ejection fraction, and in patients at the New York Heart Association class III or IV. Article 2 shows that among patients under optimal medical 8 treatment in outpatient HF clinics, 44% met the NECPAL tool criteria to receive concurrent PC (+NECPAL). Compared to ���NECPAL patients, +NECPAL patients had worse quality of life; more severe shortness of breath, tiredness, drowsiness, and pain; and greater psychosocial problems. Article 3 shows that the German NAT: PD���HF validation had good internal consistency, substantial inter���rater agreement for most of the items, and an almost perfect test���retest reliability. Moreover, patients thought well of the tool, and they agreed that it could help to improve their quality of care. Article 4 shows that nearly 3 out of 4 patients (pooled estimate 28%, 95% CI 22���36%) died with their ICD���s shock function active, despite guidelines recommending deactivation of this function at the EoL. For those with advance directives, few directives mentioned what to do with ICD devices at the EoL; the pooled prevalence estimate was only 1% (95% CI 1���3%). Conclusion: In this thesis, I contribute knowledge about gaps and challenges delivering palliative care to patients with heart failure. I show gaps regarding anticipatory care planning and end���of���life care for patients with implantable cardioverter defibrillators and offer strategies to address these gaps. Additionally, I provide knowledge and suggestions to overcome identifying patients��� palliative care needs. I assessed 3 tools to support the identification of palliative care needs; 2 of them, to support the identification of the needs due to limited life expectancy (the SQ and the NECPAL), and 1 to identify palliative care needs regardless of prognosis (the NAT: PD���HF). The best screening tool depends on the situation, and whatever tool we use, it is better to screen and think about the palliative care needs of the patients using any tool than no screening at all.

Context: Prior research widely reaches the consensus that Palliative Care (PC) has proven clinical benefits. However, far less is known about the economic impact of PC. This thesis contributes to the existing literature by providing a comprehensive overview of PC costs oc-curring in individual organizational units and cost types from the perspective of a European mixed funded health care system. The objective of this thesis is to identify cost drivers of PC and to quantify their effects on hospital costs. Research Question: Does PC reduce hospital costs for adult patients with a non-accidental death? Methodology: The research design of the retrospective, observational cost analysis is based on administrative and medical patient data for all inhospitalised deaths in 2015 in a large aca-demic University hospital in Switzerland. The thesis consists of three separate cost analyses, of which the first compares hospital costs for patients receiving PC to costs for usual care (UC) patients using the propensity score, i.e. inverse probability weighting. The second and third analyses focus on PC patients only. These sub-studies compare the costs of a patient’s pre- and post-intervention hospital stay as well as investigate the role of the timing of PC in-terventions. Hereby, the costs for patients who receive a PC intervention during the first three days of their hospital stay (early PC patients) and patients who receive a PC intervention after three days of their hospital stay (late PC patients) are compared using the bootstrap method. Results: The first cost analysis provides heterogeneous results regarding the hospital costs for PC and UC patients. Average daily costs are lower for PC patients compared to UC pa-tients. However, due to the on average significantly longer hospital stays of PC patients, total costs are on a similar level for the two groups. Considering total ward costs only, costs are significantly higher for PC than for UC patients. Therefore, no unambiguous statement can be made, whether PC patients indeed have lower hospital costs as compared to UC patients. The second analysis reveals that hospital costs increase after a patient receives a PC inter-vention. The results of the third analysis show that late PC patients have higher hospital costs than early PC patients. Therefore, the timing of a PC intervention is of major importance pur-suing a cost avoidance strategy. Conclusion: The thesis provides empirical evidence supporting decision-makers and management accountants of the cost avoidance potential of PC interventions from different cost perspectives. The results contribute to the literature comprehensive information on hospi-tal cost drivers by shedding light on costs from different organizational units as well as individ-ual cost types of a hospital. This enhances transparency for internal and external stakeholders and can serve as a potential controlling instrument.:Content Index of Figures Index of Tables Index of Abbreviations Abstract 1 Introduction 1.1 Motivation 1.2 Research Objective 1.3 Outline of the Study 2 Conceptual Foundation 2.1 Palliative Care 2.2 Economic Evaluation Approach 2.3 Stationary Accounting 2.4 Ambulatory Accounting 3 Literature Review 3.1 Review Method 3.2 Results 3.3 Implications for this Study 4 Hypotheses Development 5 Research Design 5.1 Sample Selection 5.2 Clinical and Financial Data 5.3 Methods of Analysis 5.3.1 Cluster Code 5.3.2 Statistical Analysis 6 Empirical Results 6.1 Sample Characteristics 6.1.1 Baseline Table 6.1.2 Activity Records 6.2 Financial Analysis 6.2.1 Cost Allocation 6.2.2 Palliative Care vs. Usual Care 6.2.2.1 Total Costs 6.2.2.2 Costs per Day 6.2.2.3 Interim Conclusion 6.2.3 Palliative Care Intervention 6.2.3.1 Total Costs 6.2.3.2 Costs per Day 6.2.3.3 Interim Conclusion 6.2.4 Timing of PC Interventions 6.2.4.1 Total Costs 6.2.4.2 Costs per Day 6.2.4.3 Interim Conclusion 7 Conclusion 7.1 Main Findings and Discussion 7.2 Implications 7.3 Limitations and Outlook Bibliography Appendix Appendix I: Literature Review – Search String Appendix II: Baseline Table – Early vs. Late PC Patients Appendix III: Crude Data – PC vs. UC Patients Acknowledgment Affirmation on Oath Kontext: Bisherige Forschungsergebnisse stimmen weitgehend überein, dass «Palliative Care» (PC) einen erheblichen klinischen Nutzen bieten kann. Über die wirtschaftlichen Auswirkungen von PC ist indes wenig bekannt. Diese Dissertation trägt zur bestehenden Literatur bei, indem sie eine umfassende Analyse der Kosten von PC unter den Rahmenbedingungen eines europäischen gemischt finanzierten Gesundheitssystems liefert. Dabei werden die Kosten einzelner Organisationseinheiten sowie die Kostenarten eines Spitals untersucht. Die Arbeit zielt darauf ab, Kostentreiber von PC zu identifizieren und deren Auswirkungen auf die Spitalkosten zu quantifizieren. Forschungsfrage: Senkt PC die Spitalkosten für erwachsene Patienten, welche an einem nicht-unfallbedingten Tod verstorben sind? Methodik: Das Forschungsdesign der retrospektiven Kostenanalyse basiert auf administrativen und medizinischen Patientendaten für das Jahr 2015 bezüglich inhospitalisierter Todesfälle eines grossen akademischen Universitätsspitals in der Schweiz. Die Dissertation besteht aus drei separaten Kostenanalysen. Mithilfe der Propensity Score-Methode (inverse probability weighting) vergleicht die erste Analyse die Spitalkosten von PC Patienten mit den Kosten von «usual care» (UC) Patienten. In der zweiten und dritten Analyse werden ausschliesslich PC Patienten unter Anwendung der Bootstrap-Methode untersucht. Dabei werden einerseits die Kosten eines Spitalaufenthalts vor und nach einer PC Intervention verglichen. Zum anderen wird die Bedeutung des Zeitpunkts einer PC Intervention untersucht, indem die Spitalkos-ten von Patienten mit einer frühen PC Intervention den Kosten von Patienten mit einer späten PC Intervention gegenübergestellt werden. Ergebnisse: Die erste Kostenanalyse liefert heterogene Ergebnisse bezüglich der Spitalkosten für PC und UC Patienten. Die durchschnittlichen Kosten pro Tag sind geringer für PC Patienten als für UC Patienten. Aufgrund des im Durchschnitt deutlich längeren Spitalaufenthaltes von PC Patienten sind die Gesamtkosten beider Gruppen jedoch auf ähnlichem Niveau. Der Teil der Gesamtkosten, welcher auf der Normalstation entsteht, ist für PC Patienten signifikant höher als für UC Patienten. Daher kann keine allgemeingültige Aussage getroffen werden, ob PC Patienten im Vergleich zu UC Patienten tatsächlich geringere Spitalkosten verursachen. Die zweite Analyse zeigt, dass die Spitalkosten steigen, nachdem ein Patient eine PC Intervention erhalten hat. Die Ergebnisse der dritten Analyse verdeutlichen, dass Pa-tienten, welche nach drei Tagen Spitalaufenthalt eine PC Intervention empfangen («late PC-Patienten») höhere Kosten verursachen als «early PC-Patienten», die während der ersten drei Tage ihres Spitalaufenthalts eine PC Intervention erhalten. Demzufolge ist der Zeitpunkt einer PC Intervention von grosser Bedeutung für die Implementierung einer Kostenvermei-dungsstrategie. Schlussfolgerung: Die empirischen Ergebnisse der Arbeit sind bedeutsam für Entscheidungsträger und Controller im Spital, da das Kostenvermeidungspotenzial von PC Interven-tionen aus verschiedenen Kostenperspektiven diskutiert wird. Der Beitrag der Dissertation zur bestehenden Literatur besteht darin, dass die Studie auf Basis detaillierter Kostendaten um-fassende Erkenntnisse über die Kostentreiber eines Spitals liefert, indem sie verschiedene Organisationseinheiten sowie einzelne Spitalkostenarten berücksichtigt. Die Arbeit erhöht die Transparenz für interne und externe Stakeholder und kann als potenzielles Steuerungsinstrument dienen.:Content Index of Figures Index of Tables Index of Abbreviations Abstract 1 Introduction 1.1 Motivation 1.2 Research Objective 1.3 Outline of the Study 2 Conceptual Foundation 2.1 Palliative Care 2.2 Economic Evaluation Approach 2.3 Stationary Accounting 2.4 Ambulatory Accounting 3 Literature Review 3.1 Review Method 3.2 Results 3.3 Implications for this Study 4 Hypotheses Development 5 Research Design 5.1 Sample Selection 5.2 Clinical and Financial Data 5.3 Methods of Analysis 5.3.1 Cluster Code 5.3.2 Statistical Analysis 6 Empirical Results 6.1 Sample Characteristics 6.1.1 Baseline Table 6.1.2 Activity Records 6.2 Financial Analysis 6.2.1 Cost Allocation 6.2.2 Palliative Care vs. Usual Care 6.2.2.1 Total Costs 6.2.2.2 Costs per Day 6.2.2.3 Interim Conclusion 6.2.3 Palliative Care Intervention 6.2.3.1 Total Costs 6.2.3.2 Costs per Day 6.2.3.3 Interim Conclusion 6.2.4 Timing of PC Interventions 6.2.4.1 Total Costs 6.2.4.2 Costs per Day 6.2.4.3 Interim Conclusion 7 Conclusion 7.1 Main Findings and Discussion 7.2 Implications 7.3 Limitations and Outlook Bibliography Appendix Appendix I: Literature Review – Search String Appendix II: Baseline Table – Early vs. Late PC Patients Appendix III: Crude Data – PC vs. UC Patients Acknowledgment Affirmation on Oath

Background The Surprise Question (SQ) “would I be surprised if this patient were to die in the next 12 months?” has been suggested to help clinicians, and especially General Practitioners (GPs), identify people who might benefit from palliative care. The prognostic accuracy of this approach is unclear and little is known about how GPs use this tool in practice. Are GPs consistent, individually and as a group? Are there international differences in the use of the tool? Does including the alternative Surprise Question (“Would I be surprised if the patient were still alive after 12 months?”) alter the response? What is the impact on the treatment plan in response to the SQ? This study aims to address these questions. Methods An online study will be completed by 600 (100 per country) registered GPs. They will be asked to review 20 hypothetical patient vignettes. For each vignette they will be asked to provide a response to the following four questions: (1) the SQ [Yes/No]; (2) the alternative SQ [Yes/No]; (3) the percentage probability of dying [0% no chance – 100% certain death]; and (4) the proposed treatment plan [multiple choice]. A “surprise threshold” for each participant will be calculated by comparing the responses to the SQ with the probability estimates of death. We will use linear regression to explore any differences in thresholds between countries and other clinician-related factors, such as years of experience. We will describe the actions taken by the clinicians and explore the differences between groups. We will also investigate the relationship between the alternative SQ and the other responses. Participants will receive a certificate of completion and the option to receive feedback on their performance. Discussion This study explores the extent to which the SQ is consistently used at an individual, group, and national level. The findings of this study will help to understand the clinical value of using the SQ in routine practice. Trial registration Clinicaltrials.gov NCT03697213 (05/10/2018). Prospectively registered. Electronic supplementary material The online version of this article (10.1186/s12904-019-0413-x) contains supplementary material, which is available to authorized users.