Specialized Outpatient Palliative Care

Clinical course and predictors for living at home until death

Background: Specialized outpatient palliative care (SOPC) is an important component of the palliative medicine care concept in Germany. Its purpose is to improve the out-of-hospital care of patients who cannot be adequately cared for by their primary care physicians and in the setting of general outpatient palliative care (GOPC).

Methods: In this retrospective analysis of anonymized routine treatment data, we analyzed the characteristics of SOPC patients overall and with specific diseases, and depicted them both numerically and graphically. We also carried out a regression analysis of the factors affecting whether or not patients will be able to die in a home environment.

Results: The analysis included data from 14 460 patients who were treated by 14 different SOPC teams in the North Rhine area of Germany in 2017 and 2018. The majority of patients who died were able to live at home until death (85.9%); only a small percentage died as inpatients (7.7%). The symptom burden shortly before death was less than at the beginning of treatment. The factors displaying a statistically significant association with dying at home were: more advanced age (aOR 0.96; 95% CI: [0.95; 0.96]), female sex (aOR 0.85; 95% CI: [0.74; 0.98]), and house calls at night (aOR 0.60; 95% CI: [0.51; 0.71]).

Conclusion: SOPC met its declared objectives of limiting distressing symptoms and enabling patients to live at home until death.

Specialized outpatient palliative care (SOPC) is a key component of the palliative medicine care concept in Germany. The purpose of SOPC is to provide home care for patients who cannot be adequately cared for by their primary care physicians and in the setting of general outpatient palliative care (GOPC). In this context, at home means at the place of residence. This is often a private home, but the term also covers nursing homes, hospices and other place where people live. SOPC is designed to meet the needs of patients requiring special care, for example due to a particularly complex clinical situation and/or difficult-to-treat symptoms (1).

Extensive analyses of routine care data and epidemiological data over time have so far only been analyzed in Bavaria and in the North Rhine area of Germany. However, these analyses largely rely on data from individual SOPC providers (2, 3, 4). In light of local differences, a detailed analysis of larger, multicenter data sets is warranted for advancing the care of dying or incurable, severely ill patients. In addition, large data sets allow for the study of conditions that are regularly encountered in palliative care but are less common than cancer. Such conditions include, among others, congestive heart failure, renal failure and amyotrophic lateral sclerosis (ALS) (5).

The aim of our analysis is to provide a differentiated description of a large SOPC cohort and to identify factors with an impact on dying at home.

Methods

Setting and sites

Data of 14 of the 21 SOPC providers continuously active in the North Rhine area of Germany—more specifically the administrative area of the North Rhine Association of Statutory Health Insurance Physicians—in 2017 and 2018 were extracted anonymously. Each SOPC provider is authorized for a defined geographical area. Collectively, these 14 providers cover approximately 5.7 million (60.0%) of the 9.5 million inhabitants living in the North Rhine area (6).

Data sources and measurement

The SOPC providers collected these patient data as part of their routine case documentation. Data extraction was coordinated by the Association of the North Rhine SOPC teams (VSTN, Verbund der SAPV-Teams Nordrhein e. V.) of which all participating data providers are members. All quantifiable patient information in the electronic records was extracted in an anonymized fashion. StatConsult, a software developer and contract institute for tasks in clinical research, development and health services research, was responsible for data consolidation and data quality management.

The requirements of the most recent version of the Declaration of Helsinki were observed. The research project was submitted to the ethics committee of the Medical Association North Rhine; no objections were raised (procedure number 57/2017).

Participants and case definitions

The data set comprises all patients treated between 2017 and 2018; cases that ended in 2017, but had commenced earlier as well as cases that did not end until after 2018 were also included. Information about the place of death consists of data about the reason for discharge from SOPC during the last treatment period (“died”) and/or the availability of documentation about the place of death after discharge from SOPC. In most cases, SOPC is provided continuously and without interruption between the beginning of treatment and time of death of patient. Over the course of the disease, however, SOPC may be paused for some time or terminated. Examples of possible reasons for treatment interruption include hospital stay, transfer to a hospice and improvement in symptom load as the result of intensified care. In about one fifth of the patients in our data set, one or multiple SOPC interruptions were found. A decision was made to group cases with treatment interruptions of no longer than 13 days together so that these patients could be included in the analysis. Clinically, the rationale behind this approach was that short interruptions usually occurred when patients were hospitalized to receive special treatment, such as palliative chemotherapy or red blood cell transfusions, i.e. due to administrative needs and not because specialized outpatient palliative care was no longer indicated in these patients. A cut-off value of 13 days was chosen based on expert opinions; in addition, this choice was supported by statistical considerations as with a 13-day cut-off value approximately 90% of patients with treatment interruption could remain in the data set.

Quantitative variables and statistical analysis

We decided to take a step-by-step approach to describing the total patient sample. The first step was to divide the data into oncological and non-oncological diseases. Patients with an oncological diagnosis and multiple other treatment diagnoses/main diagnoses were assigned to the group of patients with oncological disease, because we assumed, based on clinical experience, that most patients receiving SOPC were enrolled in the program because of an oncological diagnosis. The next step was to perform in each of the two groups a sensitivity analysis of common conditions with regard to clinical course and duration of treatment to identify clear differences within the groups. Where specific SOPC-relevant conditions differed significantly from the total group with regard to treatment duration, these were analyzed and described separately.

Differences in terms of symptom load at treatment beginning and end of the total sample were evaluated using the McNemar test. Kaplan-Meier estimates were used to visualize the time in the SOPC to death, both for the total patient group and for the subgroups. For the regression analysis, the place of death was dichotomized into “home environment“ (hospice, short-term care, retirement home and nursing home, other place of residence as well as at home) versus “inpatient care in hospital” (acute care hospital, palliative care unit). In this context, we categorized places such as a hospice as “home environment”, as it represents a self-determined place of residence outside the inpatient sector where SOPC can be provided as a medical partial service. Associations between place of death and various variables were found: age (per year), gender (female), days in active SOPC (per 5 days), relatives/trusted person (yes), living will (yes), health care proxy (yes), utilization at night (yes), including/excluding house calls. These variables were evaluated using logistic regression analysis, and adjusted odds ratios (aOR) with 95% confidence intervals were calculated. All analyses were performed using R (version 4.0.2) and SAS 9.4.

Results

Our analysis included data from 14 460 patients who were treated by 14 different SOPC teams in the North Rhine area of Germany in 2017 and 2018. In the defined treatment period, 82.2% (11  884) of these patients died, which corresponds to a share of approximately 9% of the total annual deaths in the region (6). The median patient age was 78 years (Q1–Q3: 68–85 years), the oldest enrolled patient was aged 107 years, the youngest aged 15 years; 51.5% of patients were female. The median length of SOPC enrolment was 19 days (Q1–Q3: 6–53); the majority of patients (80.9%) had only one continuous treatment period before their death; further 8.9% of patients had two or more successive treatment periods with only short interruptions (<13 days) in-between; therefore, these were combined for the purpose of analysis (Table 1).

Table 1

The characteristics of the sample, overall and by disease group

Approximately 85.9% of the patients who died were able to live at home until death; only 7.7% of patients died as inpatients in hospital (acute care hospital, palliative care unit). In 6.4% of patients, the place of death was unknown (missing documentation).

Approximately three quarters of the cases were categorized as oncological diseases (n = 10  631, 73.5%) and approximately one quarter as non-oncological diseases (n = 3  829, 26.5%).

The most common types of cancer were (multiple entries allowed): lung cancer (n = 2  454, 23.1%), breast cancer (n = 1  119, 10.5%), prostate cancer (n = 965, 9.1%), pancreatic cancer (n = 919, 8.6%), and colorectal cancer (n = 800, 7.5%). In the group of oncological diseases, the sensitivity analysis for treatment duration showed a homogeneous patterns; therefore, this group was not further differentiated in the tabular presentation (Figure).

Figure

Clinical course and time in SOPC for various disease groups over a 90-day period as Kaplan-Meier curves

The most common non-oncological diseases were (multiple entries allowed): heart failure (n = 2  954, 77.1%), COPD (n = 1  938, 50.6%), renal failure (n = 1  510, 39.4%), dementia (n = 1  000, 26.1%), and ALS (n = 98, 2.6%). In the group of the non-oncological diseases, the sensitivity analysis showed a rather heterogeneous pattern. Especially after exclusion of patients with multiple treatment/main diagnoses, considerable differences between the main non-oncological conditions were found (Figure).

The type of distressing symptoms present at the time of inclusion in the SOPC program varied between the different diseases. For example, in non-oncological diseases (e.g., dementia or renal failure) nausea was the cardinal symptom and significantly more prominent than in oncological diseases (eFigure).

eFigure

Symptom pattern (moderate and severe combined) at the beginning (blue) and end of treatment (purple), by disease groups. ALS, amyotrophic lateral sclerosis; COPD, chronic obstructive pulmonary disease

The overall symptom burden was rather stable over time in all subgroups and tended to be decreasing. In the total sample, symptoms such as pain (beginning 45.6%; end 37.4%; p<0.001), dyspnea (beginning 35.4%; end 32.4%; p<0.001) and nausea (beginning 15.9%; end 11.0%; p<0.001) were slightly decreasing over the period between inclusion in the program and last assessment before death, while weakness (start 89.9%; end 93.4%; p<0.001) and disorientation (start 29.3%; end 37.4%; p<0,001) were slightly increasing.

The probability to die in hospital was found significantly reduced with more advanced age and female gender. Table 2 provides a detailed overview of the independent variables analyzed in relation to dying in hospital. In the regression model, the heart failure and dementia groups were not further analyzed because the number of patients who died as inpatients was too small (<5).

Table 2

Characteristics with impact on the adverse event “hospital admission before death“

Discussion

This study is the first to characterize a large cohort of SOPC patients in Germany. It comprises 14 SOPC areas in the North Rhine region.

Especially with regard to the main idealistic goal of SOPC—dying in the desired residential environment with symptoms controlled to the greatest extent possible—, our data show that in many cases this goal is achieved. The majority of patients dies at their last place of residence and at the time of the last assessment before their death the symptom burden is slightly reduced compared to the beginning of SAPV, although the natural course of terminal illness would suggest an inverse relationship.

One of the strengths of this study is the size of the data set which enables the differentiation of the symptom burdens of patients with various non-oncological and oncological condition in the SOPC in Germany. This multi-center study involved 14 different service providers (approximately two third of all SOPC services in the North Rhine region at the time of survey) and is thus the most representative SOPC survey conducted to date based on patient records. Since the SOPC contracts in the North Rhine region contain binding requirements for all services (accounting, staffing ratio and documentation), distortions due to the missing services are unlikely. Compared to the national average in Germany, however, slight biases in our data set regarding the frequency of specific conditions are to be expected, because utilization varies widely by region. It is with 9% of all deceased insured persons lower in the North Rhine region in comparison to the German national average of 13.1—a fact that is often explained by the well-established GOPC in the North Rhine region (7). Moreover, it is not possible to determine from our data how many of the patients with proper indication did not receive SOPC at all and whether there is a bias in this respect. With regard to symptom burden, this retrospective patient record analysis is further limited by the fact that the actual SOPC experience of the patients and their relatives was not determined with qualitative instruments, but solely quantitatively.

With regard to age, gender distribution and place of death, the cohort analyzed in this study is in line with other comparison cohorts from the SOPC program—one example of this is the older age of patients with non-oncological diseases (2, 4, 8, 9, 10).

Looking at the specific diseases, patients with ALS differed most clearly from the overall average and the other diseases: Patients with ALS were significantly younger than the average of the total cohort and spend considerably more time in the SOPC program. These findings are consistent with a retrospective analysis of more than 200 000 deceased persons in Canada which also found significantly longer periods of care in ALS patients (11).

In contrast to ALS, it is striking that patients with renal failure, dementia and heart failure are included in the SOPC program at a very late point in the course of their respective disease (on average just one to two weeks prior to death). This could indicate that in these disease groups patients experience distressing symptoms requiring SOPC rather late in the course of disease. Likewise, it is conceivable that because of the slower deterioration of the condition of patients with chronic, non-oncological disease, both patients and their families have more time and opportunity to adapt to disease progression. This may result in the development of more resilient coping strategies so that SOPC is needed later in the clinical course.

The aim of SOPC is to maintain the patient‘s independence, foster their quality of life and allow the patient to die with alleviated symptoms at the desired place of death (12). In the cohort analyzed in this study, 85.9% of patients died at their place of residence and only 7.7% died in hospital. A special situation was observed in ALS patients: They died significantly more often in hospital compared to patients on average and patients with other diseases. This may reflect the exceptional burden on the caregiving family members who, confronted with the coincidence of mental clarity and lack of strength to breeze, more frequently initiate hospitalizations in emotional emergency situations. Overall, however, the number of ALS patients in our cohort who died in a home environment was significantly above the numbers for Germany and the UK reported for a period before the introduction of SOPC (1995–1999) (Germany: 55%; UK: 52%) (11).

Our data show that a high percentage of patients were able to die at home (85.9%). This proportion is considerably higher than in the total population where only 27% died at home in 2017—this percentage has continuously increased since 2001 (2001: 21%; 2011: 23%) (13, 14). The increase in the proportion of patients in the total population who died at home could be related to the expansion of SOPC services since 2007 and the increase in SOPC patients over time; however, a causal relationship has not yet been methodically demonstrated (15). With regard to the other important aim of SOPC, symptom alleviation, our data also show positive effects of SOPC in terms of symptom stabilization. For example, symptoms such as pain, dyspnea and nausea which respond well to drug treatment were overall less common at the time of the last assessment compared to the initial assessment in many cases. This indicates that pharmacotherapy is used successfully in SOPC. For the other symptoms described (disorientation and weakness), no symptomatic drug treatment is available and thus their intensity was found increased before death. However, since the overall differences are rather small, we interpret them primarily as signs of system stabilization by SOPC, in keeping with studies from the APVEL project (15).

An analysis of factors influencing a classical aim of SOPC (“dying at home”) reinforces existing assumptions about the importance of relatives and precautionary instruments. The utilization of health services at night (phone consultations or house calls) was associated with a significantly increased probability of dying at home. House calls in particular showed a strong association in this regard (aOR: 0.6; 95% confidence interval [0.51; 0.71]). However, it remains unclear whether the home visit itself or the willingness to utilize this care offering has an impact on the place of death. With regard to the usual precautionary instruments (living will and health care proxy), no statistically significant effects on the place of death were found. As to the direction of the effects, having a trusted person was more conducive to dying at home.

Conclusion

In Germany, SOPC delivers on its core treatment propositions. These include dying at home and alleviation of distressing symptoms. Even though the symptom burden differs slightly between the various conditions, a relatively homogeneous picture emerges with regard to the symptom burden of the dying over the clinical course, regardless of the disease.

Acknowledgement
We would like to thank the data-providing SOPC teams: Home Care Städteregion Aachen gGmbH, HomeCare Linker Niederrhein gGmbH, Palliativ Care Team Kreis Heinsberg, Palliativ Netzwerk Niederrhein GmbH, Palliative Care Duisburg eG, Palliative Care Team Düsseldorf, Palliative-Care-Team Oberhausen, Palliativteam SOPC Köln rechtsrheinisch GmbH, Palliativteam SOPC RheinErft GmbH, Regionales Gesundheitsnetz Leverkusen eG, SOPC Team Solingen GmbH, SOPC Wuppertal GmbH, SAPV-Krefeld GbR and SAPV-Team NoPaiN GmbH.

Conflict of interest statement
Data extraction was funded by VSTN e. V.

The authors declare no conflict of interest.

Manuscript received on 20 December 2021; revised version accepted on 22 March 2022

Translated from the original German by Ralf Thoene, MD.

Corresponding author
Dr. med. Johannes Just
Institut für Allgemeinmedizin und Interprofessionelle Versorgung
Universität Witten/Herdecke
Alfred-Herrhausen-Straße 50
58453 Witten, Germany

Cite this as:
Just J, Schmitz MT, Grabenhorst U, Joist T, Horn K, Weckbecker K:
Specialized outpatient palliative care—clinical course and predictors for living at home until death. Dtsch Arztebl Int 2022; 119: 327–32.
DOI: 10.3238/arztebl.m2022.0172

►Supplementary material

eFigure:
www.aerzteblatt-international.de/m2022.0172

References

1. Nauck F, Jansky M: [Palliative Home Care Teams in Germany]. Dtsch Med Wochenschr 2018; 143: 558–65

2. Dengler R, Kirchgessner C, Albrecht E, Rockmann K, Cassens M: Versorgungscharakteristika und Ergebnisqualität in der Spezialisierten Ambulanten Palliativversorgung (SAPV): Analyse und Verlauf von Patienten- und Versorgungsdaten am Beispiel eines SAPV-Teams. Gesundheitswesen 2020; 82: 770–6

3. Heckel M, Stiel S, Frauendorf T, Hanke R, Ostgathe C: Retrospektive Datenanalyse von Patienten in der Spezialisierten Ambulanten Palliativversorgung (SAPV) – Vergleich zwischen Stadt und Landkreis. Gesundheitswesen 2016; 78: 431–7

4. Weckbecker K, Puth MT, Just J, Horn K, Grabenhorst U, Porz J: Spezialisierte ambulante Palliativversorgung (SAPV)–Querschnittsanalyse von Symptomlast, Diagnosen und Sterbeort über 7 Jahre an einem Standort in NRW. Gesundheitswesen 2021; 83: 993–7

5. Jansky M, Lindena G, Nauck F: Stand der spezialisierten ambulanten Palliativversorgung (SAPV) in Deutschland – Verträge und Erfahrungen. Z Palliativmed 2011; 12: 164–174

6. www.it.nrw/statistik/eckdaten/lebendgeborene-und-gestorbene-956 (last accessed on 7 March 2022).

7. Ditscheid B, Krause M, Lehmann T, et al.: Palliativversorgung am Lebensende in Deutschland: Inanspruchnahme und regionale Verteilung. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2020; 63: 1502

8. Kratel U: Fortschritte in der ambulanten Palliative Care-Versorgung in Deutschland – Eine Bestandsaufnahme. Deutsche Zeitschrift für Onkologie 2014; 46: 8–15

9. Geist MJP, Keßler J, Bardenheuer HJ: Analyse der medizinischen Versorgung onkologischer und nichtonkologischer Palliativpatienten in der spezialisierten ambulanten Palliativversorgung (SAPV). Z Palliativmed 2018; 19: 149–55

10. Just J, Schmitz MT, Grabenhorst U, et al.: Factors influencing length of survival in ambulatory palliative care—a cross sectional study based on secondary data. BMC Palliat Care 2021; 20: 69

11. Neudert C, Oliver D, Wasner M, Borasio GD: The course of the terminal phase in patients with amyotrophic lateral sclerosis. J Neurol 2001; 248: 612–6

12. Alt-Epping B, Nauck F: [Specialized palliative home care: an interprofessional network]. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2015; 58: 430–5

13. Dasch B, Blum K, Gude P, Bausewein C: Place of death: trends over the course of a decade—a population-based study of death certificates from the years 2001 and 2011. Dtsch Arztebl Int 2015; 112: 496–504

14. Dasch B, Zahn PK: Place of death trends and utilization of outpatient palliative care at the end of life—analysis of death certificates (2001, 2011, 2017) and pseudonymized data from selected palliative medicine consultation services (2017) in Westphalia, Germany. Dtsch Arztebl Int 2021; 118: 331–8

15. Dasch B, Zahn PK: Place of death trends and utilization of outpatient palliative care at the end of life—analysis of death certificates (2001, 2011, 2017) and pseudonymized data from selected palliative medicine. https://innovationsfonds.g-ba.de/downloads/beschluss-dokumente/2/2020–04–03_APVEL_Ergebnisbericht.pdfonsultation services (last accessed on 7 March 2022).