Understanding complexity – the palliative care situation as a complex adaptive system

Background The concept of complexity is used in palliative care (PC) to describe the nature of patients’ situations and the extent of resulting needs and care demands. However, the term or concept is not clearly defined and operationalised with respect to its particular application in PC. As a complex problem, a care situation in PC is characterized by reciprocal, nonlinear relations and uncertainties. Dealing with complex problems necessitates problem-solving methods tailored to specific situations. The theory of complex adaptive systems (CAS) provides a framework for locating problems and solutions. This study aims to describe criteria contributing to complexity of PC situations from the professionals’ view and to develop a conceptual framework to improve understanding of the concept of “complexity” and related elements of a PC situation by locating the complex problem “PC situation” in a CAS. Methods Qualitative interview study with 42 semi-structured expert (clinical/economical/political) interviews. Data was analysed using the framework method. The thematic framework was developed inductively. Categories were reviewed, subsumed and connected considering CAS theory. Results The CAS of a PC situation consists of three subsystems: patient, social system, and team. Agents in the "system patient" are allocated to further subsystems on patient level: physical, psycho-spiritual, and socio-cultural. The "social system" and the "system team" are composed of social agents, who affect the CAS as carriers of characteristics, roles, and relationships. Environmental factors interact with the care situation from outside the system. Agents within subsystems and subsystems themselves interact on all hierarchical system levels and shape the system behaviour of a PC situation. Conclusions This paper provides a conceptual framework and comprehensive understanding of complexity in PC. The systemic view can help to understand and shape situations and dynamics of individual care situations; on higher hierarchical level, it can support an understanding and framework for the development of care structures and concepts. The framework provides a foundation for the development of a model to differentiate PC situations by complexity of patients and care needs. To enable an operationalisation and classification of complexity, relevant outcome measures mirroring the identified system elements should be identified and implemented in clinical practice. Electronic supplementary material The online version of this article (10.1186/s12913-019-3961-0) contains supplementary material, which is available to authorized users.


Interview guide for clinical experts
Background information on care institution -Which are the predominant conditions of patients treated here? • Only patients with malignant diseases? Which non-malignant conditions?
• Do you provide chemotherapy, ventilation, or radiotherapy?
-What is the proportion of patients being discharged every year? • Special symptoms or needs?

Complexity
• In addition to symptoms/needs, which other individual patient and/or family characteristics might influence the level of complexity?
• If not mentioned by the interview partner: in which direction (higher or lower level of complexity) do factors influence complexity?
-Do the factors you mentioned apply to all patients?
-Are certain factors more important than others for determining the degree of complexity? If yes, which ones?

Resources
I'd like to talk to you about resources now. The concept of resources does not refer to the patients' ability to cope with life and the illness, but to the resources that need to be provided by the care institution and the health system.
-In the international discussion it is assumed that a patient with a higher degree of complexity needs more resources than a patient with a lower degree of complexity. • To what extent do the factors that you indicated to determine complexity influence the level of resource needs, respectively, can the same factors be regarded as increasing resource needs? (refer to flashcards) • Which of these factors would you not consider to increase resource needs and why?
• Can you think of any other factors increasing resource needs?
-The following factors have been identified as the main drivers of resource use in palliative care (show cards with Australian factors): problem severity, functional status, age, phase of illness.
Do you think that these factors may be main drivers for resource needs in Germany, too? a) If "No" -factor is not considered to describe resource needs: Why not? b).If "Yes" -factor is considered to describe resource needs: Please explain.
-Which additional/other factors should be taken into account?
If not mentioned: What do you think about the lack of diagnosis in the Australian factors?

Resources -Reimbursement
So far we talked about complexity and resource needs. Now I would like to talk with you about reimbursement.
-How are services reimbursed in your institution, or which reimbursement system is used here?
-What strengths and weaknesses do you see in this reimbursement system? -Do you see any possibility for misplaced incentives with this system? If so, which are these?
-How appropriate do you think this reimbursement system is?
-From your point of view, how cost effective is this reimbursement system and why?
KLINIKUM DER UNIVERSITÄT MÜNCHEN SEITE 3 VON 6 3 -Do you think that reimbursement should be based on individual use of benefits (as in the DRG system) or would you prefer a structural funding (as in the case in per-diem rates)? And why?

Outcome Quality
We think that some of the factors describing complexity of a patient situation may also describe outcomes or outcome quality, if collected at a second point in time.
Explanation: Measured at one point in time, a specific need describes the complexity of a patient situation. If the same need is measured again at a later time point, it can be judged whether things have improved. This would then lead to a statement about outcome quality. A pain score, for example, gives information about the need for pain management at the time of the first measurement. At a second point in time, the pain score provides information whether the therapy was successful.
-Do you think the distribution of resources should be determined by needs or rather by outcome quality? And why do you think so?

Comments
-Is there anything else you would like to add? Any comments?

Opinion on Reimbursement System
-What strengths and weaknesses do you see in the current reimbursement system for palliative care?
-How appropriate do you think is the current reimbursement system for palliative care in Germany?
-Do you consider this reimbursement system to be cost effective and why?
-Do you think that reimbursements should be based on individual use of benefits (as in the DRG system) or should structural funding be available (as in the case in perdiem rates)? And why?
-If you compare reimbursement models in palliative care with other areas of the health care system, do you see any particularities or specific challenges, and if so, which ones?

Complexity
In case of the DRG system, the reimbursement is based on diagnoses and procedures (OPS).
Considering the aim of treatment and care in palliative care, it is discussed whether diagnosis and OPS are actually suitable in this domain.
In other countries, such as United Kingdom and Australia, where palliative has long been established, an alternative system focusses on the complexity of a patient situation and not on diagnosis.
-What would you understand by the complexity of the patient situation?

Outcome Quality
We think that some factors describing complexity of a patient situation may also describe outcomes or outcome quality, if collected at a second point in time.
Explanation: Measured at one point in time, a specific need describes the complexity of a patient situation. If the same need is measured again at a later time point, it can be judged whether things have improved. This would then lead to a statement about outcome quality. A pain score, for example, gives us information about the need for pain management at the time of the first measurement. At a second point in time, the pain score provides information whether the therapy was successful.
-Do you think the distribution of resources should be determined by need or rather by outcome quality? And why do you think so?