Implementing an Advance Care Planning Program in German Nursing Homes
Results of an Inter-Regionally Controlled Intervention Trial
; ; ; ; ;
Background: Advance Care Planning (ACP) is a systematic approach to ensure that effective advance directives (ADs) are developed and respected. We studied the effects of implementing a regional ACP program in Germany.
Methods: In a prospective, inter-regionally controlled trial focusing on nursing homes (n/hs), we compared the number, relevance and validity of new ADs completed in the intervention region versus the control region. Intervention n/h residents and their families were offered professional facilitation including standardized documentation.
Results: Data from 136 residents of three intervention n/hs were compared with data from 439 residents of 10 control n/hs over a study period of 16.5 months. In the intervention region, 49 (36.0%) participating residents completed a new AD over the period of the study, compared to 18 (4.1%) in the control region; these ADs included 30 ADs by proxy in the intervention region versus 10 in the control region. Proxies were designated in 94.7% versus 50.0% of cases, the AD was signed by a physician in 93.9% versus 16.7%, and an emergency order was included in 98.0% versus 44.4%. Resuscitation status was addressed in 95.9% versus 38.9% of cases (p<0.01 for all of the differences mentioned above). In the intervention region, new ADs were preceded by an average of 2.5 facilitated conversations (range, 2–5) with a mean total duration of 100 minutes (range, 60–240 minutes).
Conclusion: The implementation of an ACP program in German nursing homes led, much more frequently than previously reported, to the creation of advance directives with potential relevance to medical decision-making. Future research should assess the effect of such programs on clinical and structural outcomes.
An advance directive (AD) or “living will” is the written record of a person’s wishes with regard to the limits of medical treatment in the hypothetical future event of the person's decisional incapacity in a medical crisis (1). More than four decades after its first description (2), the traditional approach to creating ADs has to be considered a failure conceptually and empirically (3). Advance directives are
- not widely used (4, 5)
- not available when needed (6)
- often not relevant (5)
- of dubious validity (5, 7), and
- frequently not honored by medical staff (4, 5).
- ACP is regarded as a lifelong communication process with updates at regular intervals or when necessary
- Professional facilitation is actively offered to all members of a target group to help them develop and articulate their wishes regarding future treatment in the sense of an informed consent process—involving relatives whereever possible. Their wishes are documented on standardized regional forms, including a plan for emergency situations such as the POLST form (Physician Orders for Life-Sustaining Treatment) (14).
- The relevant regional health care institutions and professionals are involved and receive regular training and updates; they assume responsibility for ensuring that the completed, relevant and valid ADs are available when needed and honored reliably.
So far, no controlled prospective studies have been conducted to investigate the effects of a regional ACP program. Our study was designed to evaluate the feasibility of implementing an ACP program specifically developed for use in German nursing homes (n/hs) and associated health care structures of a given town, and whether it leads to an increase in the number of clearly formulated, valid advance care plans.
This prospective, inter-regionally controlled, nonrandomized study compared an intervention region (four n/hs in one town of the German state of North Rhine–Westphalia) with a control region (five n/hs in each of two other towns in North Rhine–Westphalia) (eTable 2).
The intervention took place at an institutional level. The primary object of analysis was the resulting ADs.
All residents of the participating n/hs were included. The recruitment phase ran from 1 February 2009 to 30 June 2009, the observation phase (data acquisition) from 15 February 2009 to 30 June 2010.
Based on the US ACP program “Respecting Choices®” (11), we developed the ACP program “beizeiten begleiten®” (16), adapted to the specific circumstances in Germany. Implementation of the program involved:
- A 20-hour training course (9–13 February 2009) for the two to four non-physician facilitators from each participating n/h
- Four hours of training for the 20 primary care physicians who cooperated in the program
- Educative sessions (introductory presentations followed by discussions) for each of the following: nursing staff at the n/hs; nursing staff at the regional hospital; medical and paramedic emergency staff; professional guardians.
Furthermore, standardized forms had to be designed: personal ADs for residents capable of decision making; ADs by proxy for the legal proxies of those who were permanently unable to give consent; and a form for documentation of physician orders for life-sustaining treatment (i.e., an adaptation of the US POLST form: the “Hausärztliche Anordnung für den Notfall”, HAnNo®: eFigure, cf. ).
The core of the intervention was the offer of professional ACP facilitation, communicated by flyers, posters and personal contact with the n/h residents. The certified “beizeiten begleiten” facilitators offered to the residents—involving relatives, whereever possible—conversations to support individual advance care planning, with the option to assist in the completion of a written AD after at least two sessions.
Moreover, the facilitation concept included involvement of the certified primary care physicians, who were asked to sign and thereby “validate” the resulting directives (in analogy to the informed consent standard).
End points and definitions
The primary end point was the number of new advance directives completed in the study cohort during the 16.5-month observation period, i.e., from the beginning of the intervention (t0 = the day before the intervention started, 15 February 2009) to the end of the intervention (t1 = 30 June 2010) (incidence). As a secondary end point we looked at the number of existing advance directives at the end of the observation period (t1), regardless of whether they had been written before or after t0 (prevalence).
We counted both the directives signed by the individual him- or herself and those signed by a legal proxy (18), defined as documents according to § 1901 of the German Civil Code, in which a person or (in the case of permanent incapability to give consent) his or her representative (5, 19) had laid down limits for future medical treatment in the hypothetical event of incapability to give consent in future health crises.
Further secondary end points were:
- The relevance of the AD, operationalized by the clarification of the resuscitation status, which is particularly relevant in n/hs
- The suitability of the AD for emergency situations, operationalized by the availability of an emergency plan containing unambiguous instructions to the responsible carers or emergency staff with regard to resuscitation.
- The validity of the AD, i.e., clear documentation of the author's ability to give consent at the time the directive was written, and of correspondence between the directive's text and the author's intention—operationalized by the signature of a physician.
- For patients able to give consent: the designation of a representative (proxy) for the event of acute or permanent inability to give consent.
- The cumulative amount of time required for the professional facilitation.
Evaluation and statistical analysis
The data from the resident samples are expressed as absolute and relative frequencies or means and standard deviations, respectively, after intervention and in the control group.
For every parameter, we report absolute numbers and relative proportions. Furthermore, we present the p-values of the corresponding factor in the model and the intraclass correlation coefficients (ICC) as a measure of sample heterogeneity.
The inferential analysis of the uniformly dichotomous study parameters was performed using logistic regression models.
Participation rate and dropouts
Six hundred forty-five (40.8%) of the potentially recruitable residents of the 14 nursing homes agreed to participate in the study (intention-to-treat [ITT] population).
The proportion of persons with a high level of need for nursing care (care level III, i.e., the highest level) was significantly lower among the nonparticipants (14.9% versus 25.6%, p<0.05), while participants and nonparticipants were similar in terms of age, sex distribution and the proportion of persons with ADs (see  for details of the nonresponder analysis).
Despite agreement to participate, one of the four intervention n/hs was reluctant to implement the agreed measures from the outset (recruitment of only 21.8% of the residents, nonadherence to agreements, no formally completed facilitation) and withdrew from the study in April 2010 (early dropout).
After excluding the 33 residents of this nursing home, 28 residents who died or moved away before data acquisition, and 9 residents with incomplete data, 575 residents of 13 nursing homes remained eligible for analysis (per-protocol [PP] population) (Figure).
Sociodemographic data and advance directives before t0
The demographic data showed a predominantly female and Catholic collective, almost all of them born in Germany, with an average age of about 82 years (Table 1). Approximately half of the residents showed signs of dementia, one quarter were classified as nursing care level III, and one eighth were being fed via a PEG tube. Over one fifth of pre-existing advance directives (26 of 121) were proxy directives. The proportion of residents with advance directives predating t0 did not differ significantly between the intervention region (16.2%) and the control region (20.8%). Significant differences between n/hs were found for the proportion of residents with one or more of the following characteristics: existence of an AD, signs of dementia, placement of a PEG feeding tube, male sex, and average age. However, there were no significant differences between the two regions.
Intervention effect: new advance directives during the observation period from t0 to t1 (incidence)
Over the observation period of 16.5 months, 36.0% of the participants (or their proxies) in the intervention region (IR) completed new ADs, compared to 4.1% in the control region (CR) (primary end point, p<0.001), based on the PP population (corresponding to 27.2% and 3.9% of the ITT population). Both personal ADs and ADs by proxy were much more frequent in the intervention region, while ADs by proxy made up the majority of the newly completed directives in both regions (Table 2).
Analysis of the new ADs (Table 2) showed that all of the following parameters were far more frequent in the IR (p<0.01 in all cases):
- Designation of a proxy (in 94.7% vs. 50.0% of personal ADs)
- Signing of the AD by a third party (95.9% vs. 77.8%), particularly by a physician (93.9% vs. 16.7%)
- Completion of an emergency plan (98% vs. 44.4%)
- Resuscitation status clearly defined (95.9% vs. 38.9%), and presence of a corresponding eye-catcher on the resident’s file (81.6% vs 38.9%).
Within each region, the n/hs did not differ significantly in terms of the proportion of residents with a new AD. Differences were found only in the rate of POLST-like emergency plans and in archiving (see the figures for “Proportion of variance between homes” in Tables 2 and 3).
Existing ADs could be actually found at their supposed repository with similar frequency in both regions (Table 2). Prominent references (eye-catchers) in the file or on its back to the existence of an AD, an emergency plan, or a do-not-resuscitate order were significantly more frequent in the intervention region.
In only one third of the residents (of both regions) who had been transferred to hospital, a copy of the AD could be found in the hospital medical records (Table 2).
Intervention effect: all advance directives at t1 (prevalence)
Table 3 provides an overview of all ADs (i.e., signed before or after t0) at the end of the observation period (t1 = 30 June 2010).
Although the intervention period was short, the intervention effect is clearly discernible (52.2% vs. 24.8%), despite the interference by ADs that were drawn up before the beginning of the intervention (t0 = 15 February 2009).
Time and effort required for the facilitation process in the intervention region
Sixty of the 74 persons who had an initial conversation with a facilitator went on to complete an AD. The median number of conversations was 2.5 (range 2 to 5), and the median total facilitation time amounted to 100 (60 to 240) minutes (Table 4).
This study is the first to show the feasibility and procedural efficacy of the introduction of a regional ACP program in Germany. At the same time, it is the first publication internationally of a prospective, inter-regionally controlled study of the implementation of an ACP program.
Only a few controlled studies of the implementation of ACP programs have been published to date. A Canadian study can be regarded a milestone; however, its randomized assignment to intervention or control was at nursing home level rather than regional level. It showed improvement not only in process quality but also in outcome quality, evidenced by, for example, decreasing hospital admissions and lengths of stay with unchanged satisfaction and mortality, as well as lower total costs in the intervention nursing homes (20).
In a retrospective longitudinal study in the La Crosse region (Wisconsin, USA), where the ACP program Respecting Choices® has been established since the 1990s (11), consecutive analysis of all nonsudden deaths showed that nearly all of the persons in this group had ADs and emergency plans (POLSTs) and that in almost all cases the wishes expressed therein were respected by the institution where the patient was last treated (21). The cost effectiveness of Respecting Choices® has not been systematically investigated, but in a nationwide survey La Crosse was the region with the lowest Medicare hospital costs in the last 2 years of life (12).
An Australian group implemented the ACP program Respecting Patient Choices® in a hospital and showed, in a randomized study, that deceased patients in the intervention group were much more frequently treated according to their wishes and that their relatives suffered far less from posttraumatic stress, anxieties, and depression (22).
In contrast to the Canadian and Australian studies described above, but in agreement with the study from La Crosse, our study chose the regional rather than the institutional level as the focus of the intervention because the systematic changes necessary for improved respect for patients’ preferences stated in an AD cannot be achieved at institutional level alone. The intervention n/hs were therefore recruited from one region and compared with homes from another region to avoid spillover effects. In contrast to the La Crosse study, our study is prospective and included a control group. The observation did not focus on the deaths, but rather on all residents of the participating h/hs.
The most prominent finding of our study, alongside the nine times higher incidence of ADs in the intervention group over the 16.5-month observation period, is the predominantly high standard of the newly completed directives in the intervention region with regard to relevance and validity (Table 2). Almost all of these advance care plans included designation of a proxy, were signed by a physician, and included—documented on our emergency plan adapted from the POLST form—clear instructions regarding resuscitation in the event of cardiac arrest; a glance at the back of the resident’s file usually sufficed to ascertain the resuscitation status. The difference from the control region was considerable, even though the latter—with a smaller number of cases—performed better in this respect than did another collective that we had studied previously (5).
The overall prevalence of ADs at the end of the observation period (including directives predating the intervention) was higher in the intervention group (52%) than in the control group (25%) (Table 3), but lower in both groups than in a study—of similar duration—by Molloy et al. (20) (70% versus 57%); in both studies there were more ADs by proxy than personal ADs (signed by the resident). While we found considerable differences between the intervention and control regions with regard to the relevance and validity of the ADs, Molloy et al. (20) report specific (treatment-relevant) stipulations in the ADs of both intervention and control n/hs; the validity of the ADs was not investigated.
The strengths of our study include minimal selection, because all three of the districts and 13 of the 14 n/hs that we initially approached participated. The study is not randomized, however, so an effect due to the pragmatically determined selection of regions (and thus of n/hs) cannot be excluded, so a future cluster-randomized study with inter-regional comparison seems desirable. The early withdrawal of one institution from the study may be an expression of possible barriers in accepting the intervention at n/h level and thus an indicator of limited generalizability of the results. The power of the study is also limited by the participation rate of 40.8% of residents (or their proxies), although our nonresponder analysis found no signs of systematic selection apart from a higher level of demand for nursing care. On the other hand, the incidence of new ADs, in relation to all potential participants of the intervention region, of 36.0% × 40.8% = 14.7% in 16.5 months of observation is probably underestimated, because our initial recruitment letter to the n/h residents contained no detailed description of the intervention, but requested that the nursing home’s files be made available to the study team for 1.5 years. This barrier would not have been present if the intervention had been implemented as part of standard care. In our experience the offer of a facilitated conversation is rather likely to receive widespread interest; a much higher participation rate could therefore be expected if implementation was not linked with a study.
Further evidence for the feasibility of our ACP program is provided by its integration into the institutional policies of the intervention n/hs and of the local hospital. The median time of 100 minutes required for the facilitation process is higher than the 60 minutes reported in the only comparable study to date (22). One reason for this difference could be that we documented the total time, including preparation and follow-up. We cannot judge whether the outcome “justifies” this effort (23), because we were unable to investigate, in a comparative way, relevant end points such as reduction of unwanted days in hospital, transfers between institutions, or parameters such as avoidance of time-consuming conflicts with relatives about contentious therapeutic interventions, or the overall level of satisfaction on the side of all parties involved. We see an urgent need for studies evaluating the outcome quality of a German ACP program. The results of the groups cited above (20–22) indicate that the costs in time and money of initial implementation of an ACP program, and continuous facilitation, are matched by savings due to such a program. Above all, however, it has to be questioned whether, in ethical terms, society can afford to continue in its failure to systematically elicit in advance the treatment preferences of the elderly and the chronically ill and make sure that these wishes are respected in the event of a future medical crisis.
The study was financed by the German Federal Ministry of Education and Research (Fkz. 01 GX 0753), with an additional donation from the B. Braun Foundation. We are grateful to the directors of the US advance care planning program Respecting Choices®, Dr. Bud Hammes and Linda Briggs, RN, MSN, MA, for their invaluable assistance in designing the intervention and for their advice during its implementation.
We thank Prof. S. Rixen (Bayreuth) for instructive legal advice on project development and implementation and Prof. H.-H. Abholz (Düsseldorf) for critical revision of the manuscript with numerous valuable suggestions.
Thanks are also due to the directors and staff of the participating nursing homes in the intervention and control regions, the participating primary care physicians, and the directors and staff of the hospital, emergency service, outpatient hospice, and other cooperating facilities in the intervention region.
We thank the North Rhine Medical Association and the administrations of the two control districts for their support. Special thanks are due to the administrative staff of the district Neuss, without whose advice and practical assistance it would not have been possible to conduct the study.
Finally, we are particularly grateful to the certified “beizeiten begleiten” facilitators, Dipl. Soz.-Päd. Inga Lücke and Barbara Mandt, for their outstanding commitment and their contribution of numerous crucial ideas and suggestions.
Conflict of interest statement
“beizeiten begleiten®” is a registered wordmark and logo of the University of Düsseldorf and the University of Augsburg, “HAnNo®” a registered wordmark and logo of the University of Düsseldorf. The authors developed these wordmarks and logos, but possess no commercial rights thereto.
Beyond that, the authors declare that no other conflicts of interest exist.
Manuscript received on 18 June 2013, revised version accepted
on 7 October 2013.
Translated from the original German by David Roseveare.
Dr. med. Jürgen in der Schmitten, MPH
Institut für Allgemeinmedizin
Medizinische Fakultät der Heinrich-Heine-Universität
40225 Düsseldorf, Germany
@eSupplement, eTables and eFigure:
Dr. in der Schmitten MPH, Dipl. Psych. Mellert
Institute for Patient Safety, Rheinische Friedrich-Wilhelms-Universität Bonn: Frau Lex MScN
Faculty of Social Work, Catholic University of Eichstätt-Ingolstadt: Dr. jur. Rothärmel
Department of Medical Biometry and Epidemiology, University Medical Center Hamburg-Eppendorf:
Prof. Dr. rer. pol. Wegscheider
Institute of Ethics, History and Theory of Medicine at the Ludwig Maximilians University Munich:
Prof. Dr. med. Marckmann MPH
|1.||Emanuel LL, Barry MJ, Stoeckle JD, Ettelson LM, Emanuel EJ: Advance directives for medical care—a case for greater use. N Engl J Med 1991; 324: 889–95. CrossRef MEDLINE|
|2.||Kutner L: Euthanasia: Due process of euthanasia: the living will, a proposal. Indiana Law J 1969; 44.|
|3.||Fagerlin A, Schneider CE: Enough. The failure of the living will. Hastings Cent Rep 2004; 34: 30–42. CrossRef MEDLINE|
|4.||Evans N, Bausewein C, Menaca A, et al.: A critical review of advance directives in Germany: attitudes, use and healthcare professionals' compliance. Patient Educ Couns 2012; 87: 277–88. CrossRef MEDLINE|
|5.||Sommer S, Marckmann G, Pentzek M, Wegscheider K, Abholz HH, in der Schmitten J: Advance directives in nursing homes: prevalence, validity, significance, and nursing staff adherence. Dtsch Arztebl Int 2012; 109(37): 577–83. VOLLTEXT|
|6.||Morrison RS, Olson E, Mertz KR, Meier DE: The inaccessibility of advance directives on transfer from ambulatory to acute care settings. JAMA 1995; 274: 478–82. CrossRef MEDLINE|
|7.||Brett AS: Limitations of listing specific medical interventions in advance directives. JAMA 1991; 266: 825–8. CrossRef MEDLINE|
|8.||Wiesing U, Jox RJ, Hessler HJ, Borasio GD: A new law on advance directives in Germany. J Med Ethics 2010; 36: 779–83. CrossRef MEDLINE|
|9.||Höfling W: Das neue Patientenverfügungsgesetz. NJW 2009; 62: 3.|
|10.||Yates JL, Glick HR: The failed Patient Self-Determination Act and policy alternatives for the right to die. Journal of Aging & Social Policy. 1997; 9: 29–50. CrossRef MEDLINE|
|11.||Hammes B: Update on Respecting Choices four years on. Innovations in End-of-Life Care 2003; 5: 18.|
|12.||Marckmann G, in der Schmitten J: Patientenverfügungen und Advance Care Planning: Internationale Erfahrungen. [Advance Directives and Advance Care Planning: International Experiences]. Zeitschrift für Medizinische Ethik 2013; 59: 213–27.|
|13.||in der Schmitten J: Advance care planning: putting an end to the agonizing perpetuation of a pointless debate. Onkologie 2013; 36: 395–6. CrossRef MEDLINE|
|14.||Hickman SE, Nelson CA, Perrin NA, Moss AH, Hammes BJ, Tolle SW: A comparison of methods to communicate treatment preferences in nursing facilities: traditional practices versus the physician orders for life-sustaining treatment program. J Am Geriatr Soc 2010; 58: 1241–8. CrossRef MEDLINE PubMed Central|
|15.||in der Schmitten J, Rotharmel S, Mellert C, et al.: A complex regional intervention to implement advance care planning in one town's nursing homes: Protocol of a controlled inter-regional study. BMC Health Serv Res 2011; 11: 14.|
|16.||in der Schmitten J, Marckmann G: Sackgasse Patientenverfügung. Neue Wege mit Advance Care Planning am Beispiel beizeiten begleiten®. [Traditional advance directives—a dead end. Exploring new ways with the advance care planning program beizeiten begleiten®]. Zeitschrift für Medizinische Ethik 2013; 59: 229–43.|
|17.||in der Schmitten J, Rothärmel S, Rixen S, Marckmann G: Patientenverfügung im Rettungsdienst (Teil 2). Neue Perspektiven durch Advance Care Planning und die Hausärztliche Anordnung für den Notfall. Notfall Rettungsmed 2011: 10.|
|18.||Volicer L, Cantor MD, Derse AR, et al.: Advance care planning by proxy for residents of long-term care facilities who lack decision-making capacity. J Am Geriatr Soc 2002; 50: 761–7. CrossRef MEDLINE|
|19.||Bühler E, Stolz K: Das neue Gesetz zu Patientenverfügungen in der Praxis. BtPrax 2009: 6.|
|20.||Molloy DW, Guyatt GH, Russo R, et al.: Systematic implementation of an advance directive program in nursing homes: a randomized controlled trial. JAMA 2000; 283: 1437–44. CrossRef MEDLINE|
|21.||Hammes BJ, Rooney BL, Gundrum JD: A comparative, retrospective, observational study of the prevalence, availability, and specificity of advance care plans in a county that implemented an advance care planning microsystem. J Am Geriatr Soc 2010; 58: 1249–55. CrossRef MEDLINE|
|22.||Detering KM, Hancock AD, Reade MC, Silvester W: The impact of advance care planning on end of life care in elderly patients: randomised controlled trial. BMJ 2010; 340: c1345. CrossRef MEDLINE PubMed Central|
|23.||Marckmann G, Sanktjohanser AM, in der Schmitten J: Sterben im Spannungsfeld zwischen Ethik und Ökonomie. In: Bormann FJ, Borasio G (eds): Sterben Dimensionen eines anthropologischen Grundphänomens. Berlin: Walter de Gruyter Verlag; 2012: 351–67.|