Effect of Advance Care Planning to Improve Patient-centered Care of Nursing Home Residents

Advance Care Planning Clinical Trial

— BEVOR  

Official title:

Effect of Advance Care Planning to Improve Patient-centered Care of Nursing Home Residents: Cluster-randomised Controlled Trial

Clinical Trial Details — Status: Completed

Administrative data

• NCT number: NCT04333303
• Other study ID #: 01VSF18004
• Status: Completed
• Phase: N/A
• Start date: April 6, 2020
• Est. completion date: January 31, 2023

Study information

• Verified date: August 2023
• Source: Heinrich-Heine University, Duesseldorf
• Contact: n/a
• Is FDA regulated: No
• Study type: Interventional

Clinical Trial Summary

Advance Care Planning (ACP) is an approach to support individuals in understanding and sharing their values, treatment goals, and preferences regarding future medical care. This study aims to evaluate the clinical effectiveness of implementing a complex regional ACP program focussing on nursing homes and their related health care providers.

For this purpose, a cluster-randomised controlled trial (cRCT) is being conducted in four study centres in Germany: Düsseldorf, Göttingen, Halle (Saale), and Munich. 44 nursing homes with 3,520 residents will be randomised into either an intervention group, receiving the ACP-intervention described below, or into a control group that will receive usual care.

The complex ACP intervention comprises training of professional ACP facilitators (micro level), supporting organisational development in nursing homes and other relevant institutions (meso level), and moderating regional change management among a network of all related health care institutions and actors (macro level).

Clinical outcomes will be measured in order to describe whether the intervention improved care consistency with care preferences.

Data will be collected by two independent approaches:

Firstly, participating nursing homes will provide anonymous data on all their long-term care residents of a defined time period, including the hospitalisation rate (primary outcome). The primary hypothesis will test whether the complex intervention reduces the rate of hospital admissions. A group comparison of all hospitalisations in the past 12 months is made 21 months after randomisation.

Secondly, all residents who have given informed consent in the first three months of the study will be repeatedly surveyed until month 21 (or until they die earlier). Treatment decisions in the face of life-threatening illness that affect any of a list of pre-defined outcomes will be evaluated as to whether they were consistent with the residents' care preferences.

If residents decease, patient-centered care in the last weeks of life will be evaluated by additional interviews with the next bereaved relative, and an involved nurse. Besides, bereaved relatives will be assessed for trauma, depression and anxiety.

In addition, a process evaluation and a health economic evaluation are carried out according to Medical Research Council (MRC) recommendations.

Description:

Aim of the study:

The aim of the study is to evaluate the effect of a complex regional Advance Care Planning (ACP) intervention on care consistency with residents' care preferences in the face of potentially life-threatening disease.

Study design:

Multi-center cluster-randomised controlled trial (cRCT) to evaluate the clinical effects of regional implementation of a complex German ACP program (confer www.div-bvp.de) focussing on nursing homes and related health care providers. 44 nursing homes (clusters) with a mean of 80 residents (observation units) each will be randomised into intervention (ACP implementation) and control (usual care) groups.

Randomization:

A cluster-randomization for control and intervention group is carried out after enrolment of the care facilities using a computer-based code with variable block length generated by the Institute for Medical Biometry and Epidemiology, Hamburg-Eppendorf, Hamburg.

Outcome measures, Data Collection and Time Frame:

In this study, a combination of outcomes, derived from two separate and methodologically distinct data collections that have complementary strengths and limitations, allows to reconstruct a valid perspective on care consistency. The primary outcome, hospitalisation rate, has been chosen because of its relevance for sample size calculation; it is not to be misunderstood to represent the primary intention of the study intervention which is rather a reduction of unwanted hospitalisations (and other procedures). For secondary outcomes see separate section in this registration.

For data collection 1, participating nursing homes provide an anonymous data set of all their residents (full census), including hospitalisation rate (= primary outcome) and days, last place of care, presence of advance care plans, and presence of preferences or orders that forbid hospitalisation. These data will be collected in month 1 (t0) for the 12 months preceding the trial; after month 9 (t1) for the preceding 9 months run-in phase of the intervention, and after month 21 (t2).

Data collection 2, in contrast, is from nursing home residents who gave written informed consent to an in-depth analysis of their nursing and medical files, and comprises three elements:

Firstly, at t0 and t2, a short survey will be performed with the residents (WHOQOL-OLD) or, if incapable, with the responsible nurses (QUALIDEM). Residents' files will be analysed for pre-defined life-sustaining procedures (such as transports to hospital, hospital days, tube feeding insertion and days, intensive care days, mechanical ventilation days, CPR attempts), and evidence of ACP process parameters like advance care plans, and durable powers of attorney.

Secondly, residents' files will be assessed every 3-6 months for potential life-threatening events pointing to any of a list of predefined treatment decisions that correspond with the secondary outcomes (hospital admission, antibiotic treatment etc). Every identified treatment decision will be matched with the corresponding care preferences, derived by analysis of residents' files, and short interviews with residents, relatives, and nursing home staff, so in each case a best possible judgment can be made on care consistency.

For deceased participants, thirdly, interviews will be conducted with a bereaved relative and a nurse who was in charge of this resident, again evaluating the correspondence of the recent care delivered, and the deceased resident's care preferences (After Death Bereavement Interview [domains Shared Decision Making and Advance Care Planning], Hospital Anxiety and Depression Scale / German version [HADS-D] and Impact of Event Scale / revised version [IES-R]).

Sample Size Calculation:

The sample size calculation was carried out with the Power Analysis and Sample Size (PASS) software 16.0.4. The type-one error α is set to 5% two-sided. In total, 40 care facilities with an average of 115 residents each and an intra-cluster correlation of 0.03 are required to achieve a power of 80 % and detect a statistically significant absolute reduction in the annual admission rate of 48% by (absolute) 8.6% to 39.4%. The primary analysis of the hospitalization rate in the 21-month observation period includes all residents who lived in the care facility during this time; with an average occupancy rate of 80 home places and a conservatively assumed death and new occupancy rate of 25% per year, this means 80 + [21/12] x 20 = 115 residents per home and 40 x 115 = 4,600 residents in total in 21 months. With an assumed early dropout rate of the facilities (between baseline survey and randomization) of about 10%, 44 facilities with a total of 5060 residents have to be recruited.

Analysis

All applicable statistical tests are two-sided and performed with a 5% significance level. Analyses of secondary endpoints are performed without adjustment for multiplicity. All confidence intervals presented are 95% two-sided.

Descriptive Analysis

A descriptive analysis is carried out at t0 (baseline before intervention) and t1 (after 9 months) regarding the care facilities and residents. Categorical data are summarized by absolute and relative frequencies. Continuous data are summarized by mean and standard deviation (SD) as well as median, quartiles and/or interquartile range (IQR). The number of available observations and the number of missing observations are reported separately for treatment groups. Tests with statistical significance are not performed for baseline characteristics. All participating care facilities are described regarding the sum of all residents, proportion of male and female residents, mean age of residents, bed days per resident, sum of all residents with at least one hospital stay, sum of all hospital admissions, sum of all hospital days and sum of all deceased residents. All residents are described regarding age, sex and sociodemographic data.

Primary Analysis

The primary outcome is evaluated according to the intention-to-treat principle (ITT) analyzing the change in hospitalization rate per 100 residents from baseline to 21 months in intervention and control group. The ITT population consists of all care facilities randomized and included in the study. The primary analysis is performed at care facility level and includes a rate comparison.

A poisson regression model is calculated taking into account the group (intervention vs control) and realized follow-up time as well as baseline hospitalization rate per 100 residents as independent variables. The resulting statistical test for group comparison of intervention and control group is performed two-sided at the 5% significance level. For sensitivity analyses, per protocol analyses are performed as well as an evaluation with the negative binomial distribution to detect a possible but not expected over dispersion. Per protocol analyses will take into account whether a minimum share of residents have been offered ACP conversations in the respective nursing home.

Secondary Analysis

The evaluation of secondary outcomes is performed in an explorative manner and evaluated in the ITT population. All analyses relating to the residents of care facilities are performed within a framework of multi-level models. For variables including a rate comparison poisson regression models are calculated with events up to t2 as the dependent variable and group (intervention vs. control), study center, region and realized follow-up time as fixed effects, the particular baseline measurement as a covariate, and care facilities as random effect. Similarly, changes in other endpoints are analyzed in accordance to their statistical distribution by using appropriate regression models.

Intervention

The complex ACP intervention comprises an offer to all residents to lead ACP conversations with comprehensively trained ACP facilitators in close cooperation with family physicians (micro level), staff education and organisational development in nursing homes, related hospitals and emergency services (meso / institutional level), and a regional change management process involving all relevant actors and stake holders (macro level).

Process Evaluation:

In accordance with guidelines for complex interventions published by the Medical Research Council (2019), a concomitant process evaluation aims to analyse the acceptance by and impact of the intervention on residents, their relatives, nursing staff, nursing home management, and family physicians. This will identify potential barriers, burdens, and facilitating factors.

Health Economic Analysis

The objective of the health economic evaluation in BEVOR is to assess the economic implications of the intervention by comparing the cost and outcome of the intervention group (treatment according to the ACP approach) to the cost and outcome of the control group (usual care). All costs associated with the intervention as well as costs resulting from the consumption of health-related goods and services are considered from the perspective of the German social insurance.

In order to assess the economic implications of the intervention, a cost-consequence analysis (CCA) is conducted. In addition, a cost-effectiveness analysis (CEA) is performed for selected endpoints by calculating the incremental cost-effectiveness ratio (ICER), for example for an additional hospital admission averted or an additional case of goal concordant care.

Standardised instruments will be used to record the health care consumption of study participants such as outpatient visits, hospital stays, medical aid, contact with therapists and care in nursing homes or retirement homes at baseline, t1 and t2. Due to the short study period, no discounting of the effects and costs is planned.

95% confidence intervals for costs and outcomes will be calculated non-parametrically using bootstrap procedures. Univariate and probabilistic sensitivity analyses will be performed and cost-effectiveness acceptance curves executed to account for uncertainty.

Change of Timeline during the Study

Originally, this study was designed to last 3 years from its beginning in September 2019. After a 6-month preparatory phase, randomization and intervention were to start at month 7 (i.e., March 2020), followed by a 9-month run-in period of the intervention and a 12-months observation period. During 03/2020 to 03/2021, however, the SARS-CoV-2 pandemic restricted the NHs' capacities for participation to times almost zero. Therefore, NH recruitment and randomization were extended until 11/2020 in order to compensate for SARS-CoV-2 related withdrawals. In 05/2021, when NHs had become accessible again 2 to 3 months after complete vaccination of NH residents, it was decided to modify the original timeline:

The start of the observation period was postponed by 9 months in order to allow for a robust run-in period of the intervention. The observation period remains 12 months (now 09/2021 to 08/2022), preceded (in the intervention group) by a variable run-in period of the intervention of 9 to 17 months, depending on the time of randomization.

Recruitment information / eligibility

• Status: Completed
• Enrollment: 5927
• Est. completion date: January 31, 2023
• Est. primary completion date: August 31, 2022
• Accepts healthy volunteers: Accepts Healthy Volunteers
• Gender: All
• Age group: N/A and older

Eligibility

Inclusion Criteria:

• all long-term care residents of the enrolled nursing homes

Exclusion Criteria:

• none

Related Conditions & MeSH terms

• Advance Care Planning
• Nursing Homes

Intervention

Other:
• Implementation of a complex regional Advance Care Planning program
Sustainable implementation of ACP at micro- (qualification of professional ACP facilitators to lead conversations in cooperation with family physicians), meso- (qualification of relevant co-actors, organisational development of participating institutions) and macro-level (regional ACP steering group and network), ensuring that valid ACP is encouraged, facilitated, documented, known and honored. Core of the intervention is a modular 9-day-training for ACP facilitators. These will be preferably located in regional pools, serving more than one nursing home, and reimbursed, if certified, by German sickness funds based on recent legislation (§ 132g Social Code Book [SGB] V: 1 full-time facilitator for 400 residents). ACP conversations will be offered to all residents. Regional intervention at meso- and macro-level comprises various formats of specific ACP training and information for leadership and staff, and aims to support professional change management.

Locations

Country	Name	City	State
Germany	Heinrich-Heine-University	Düsseldorf
Germany	University Medical Center	Göttingen
Germany	Martin Luther University	Halle (Saale)
Germany	University Hospital Großhadern, LMU	Munich

Sponsors (8)

Lead Sponsor	Collaborator
Heinrich-Heine University, Duesseldorf	BARMER, Federal Ministry of Health, Germany, Ludwig-Maximilians - University of Munich, Martin-Luther-Universität Halle-Wittenberg, Universitätsklinikum Hamburg-Eppendorf, University Hospital Goettingen, University Hospital Munich

Country where clinical trial is conducted

Germany, 

References & Publications (1)

In der Schmitten J, Lex K, Mellert C, Rotharmel S, Wegscheider K, Marckmann G. Implementing an advance care planning program in German nursing homes: results of an inter-regionally controlled intervention trial. Dtsch Arztebl Int. 2014 Jan 24;111(4):50-7. doi: 10.3238/arztebl.2014.0050. — View Citation

Outcome

Type	Measure	Description	Time frame	Safety issue
Primary	Change of hospital admissions	Number of hospital admission per 100 residents	21 months
Secondary	Change of hospital days	Number of hospital days	21 months
Secondary	Change in the number of residents whose preferences were (a) known and (b) followed in the face of potentially life-threatening events	Proportion of potentially life-threatening events in which the resident's preferences were (a) known and (b) followed at the time	21 months
Secondary	Change in the number of pre-specified unwanted invasive treatments	Rates of unwanted measures / treatments: feeding tube insertion and tube feeding days, transferral to the hospital, number of hospital stays and days, attempts of cardiopulmonary resuscitation (CPR), mechanical ventilation days, ICU days	21 months
Secondary	Change in the rate of hospital as last place of care	Proportion of residents who died in the hospital	21 months
Secondary	Change in the perception of concordance	Rates of residents, proxies and nurses who judge care in the past 3 months delivered to be consistent with the resident's care preferences, measured by key item F2 selected from the After Death Bereavement Interview	21 months
Secondary	Change in the number of process parameters indicating valid ACP: advance care plans	numbers of advance care plans signed by facilitators, physicians and/or representatives;	21 months
Secondary	Change in the number of process parameters indicating valid ACP: proxy designations	numbers of proxy designations	21 months
Secondary	Change in anxiety, depression and trauma among bereaved relatives	Scoring on Revised Impact of Event Scale (IES-R): min. 0 pts, max. 88 pts., higher score indicates higher impact (trauma) Scoring in German version of Hospital Anxiety and Depression Score (HADS-D): min. 0, max. 14, higher score indicates higher degrees of anxiety and depression	21 months