Final Report Summary - IBENC (Identifying best practices for care-dependent elderly by Benchmarking Costs and outcomes of community care)
The need for cost effective health systems in Europe is imperative in the context of an increasing demand for health care by growing numbers of elderly people and a shrinking working force. Our health care systems may benefit from methods to ensure, measure and improve quality of care. However it is unclear which models of care delivery are most cost-effective. To advance both quality and efficiency of community care organisations, policymakers need information on which type of community care delivery provides best outcomes against reasonable costs. The overall aim of the IBenC project was to identify best practices in community care delivery for care-dependent community dwelling elderly, by benchmarking the cost-effectiveness of community care delivery systems, across Europe.
Methods and results
Data on health , functioning and resource use of 2884 community care clients from six countries (Belgium, Finland, Germany, Iceland, Italy, and the Netherlands) was collected longitudinally with the interRAI-HC: at baseline, and after six and twelve months. In addition, a total of 1067 community care staff members from 31 organisations filled out the questionnaire on staff characteristics (Q1). Of the 38 recruited organisations, 36 returned a questionnaire on their organisational characteristics (Q2) (WP3).
We tested two summary measures for benchmarking quality of care outcomes: 1) The Independence Quality Scale (IQS) reflects organisations’ performance on maintaining functional independence and engagement of their clients; 2) the Clinical Balance Quality Scale (CBQS) reflects an organisation’s performance on improving functioning of their clients. Both scales range from 0 to 10 and were formed by respectively 11 and nine case-mix adjusted quality indicators (WP2).
After validation of the interRAI-HC for resource use measurements, costs of care utilisation over a period of six months were estimated per client and aggregated per organisation, as well as per care model. Case mix adjusted costs of care were benchmarked for organisations and care models from both a societal (formal and informal costs) and healthcare perspective (formal costs only) in two European community care samples (IBenC and AdHOC) and revealed substantial and significant cost of care differences between care models (WP5).
We distinguished six care models that differed in practice on three factors: 1) the level of patient-centred care delivery, 2) the availability of specialised care professionals, and 3) the extend of monitoring one’s care performance (WP6).
Organisations were benchmarked on their efficiency by means of two measures that integrated costs and quality of care, the efficiency indexes: The IQS index reflects an organisation’s performance to prevent their clients’ functional decline against the costs of resource use relative to other organisations. The CBQS-index reflects an organisation’s performance in improving their clients’ functional status against costs of resource use, also relative to other organisations (WP7). Care providers achieving index values above 1 are efficient, e.g. delivering good quality care without inducing high costs of care utilisation. Staff factors related to organisations’ efficiency appeared to be predictability, emotional demands, influence, physical workload, and quantitative demands at work, and type of contract (full-time/part-time, temporary/fixed). Also patient-centred care delivery, availability of specialised care professionals and monitoring of care performance were related to organisational efficiency. This novel method of benchmarking provides a valid and evidence based way to identify best practices. Moreover, input for the benchmark can be derived from routine care data if organisations use interRAI-HC assessments. interRAI assessments are currently the most widely disseminated assessment instruments across the world.
Scientific impact: A very unique and international data base was established and the consortium was highly successful in disseminating the first project results and impact by reaching out to stakeholders in a large number of scientific conferences, webinars, press releases a final conference. At the end of the study period already five papers were published in peer reviewed journals and many are to follow.
Societal impact: Stakeholders acknowledged the unique learning potential of the benchmarking method for organisations, policymakers, health care insurers, and clients. A blueprint is available for the benchmarking method which can be fully automated by a web application (WP8). A next step is to engage national or regional health authorities, built the web application, involve care organisations, set up a governance and bench learning model, and test viable business models.
Project Context and Objectives:
Policy makers need to reform health care systems in such a way they function as cost-effective as possible in order to serve all our future care-dependent citizens with appropriate and affordable care. In the context of decreased public spending, the sustainability of high quality public health care is even more challenging.
‘What is value for money in community care for care-dependent elderly?’ is the question we will answer in this project within the context of health and social care provisioning. Care-dependent elderly persons at home are frequently served by medical, nursing and social services that are often subject to poor coordination. It is hypothesised that quality of care and health outcomes can be improved by better coordination of these services. But current standards to benchmark outcomes and costs of care are lacking, as is scientific sound evidence to contributors to either costs or quality of care.
The overall aim of IBenC is to identify best practices in community care delivery for care-dependent community dwelling elderly, by benchmarking the cost-effectiveness of community care delivery systems, across the European Union.
To reach this overall aim, the project addresses the following two main objectives:
1. Benchmark costs-effectiveness of community care: identification of best practice costing models, for care-dependent elderly people at home in Europe.
2. Characterise best practices: describe the most cost-effective practices in terms of staff characteristics, management structures, care processes, and reimbursement systems.
The work to address these objectives into eight WPs, each with its own specific objectives and tasks. The main objectives for each WP are described below.
WP1 Project management and coordination
1.1) Internal and external management to ensure implementation of the consortium in conformity with the EC contract and the consortium agreement, and a successful meeting of the aims.
WP2 Benchmarking quality of care
2.1) Develop a method to benchmark the quality of community care (based on 4010 individual patients available from the AdHOC study (Carpenter et al., 2004)).
2.2) Benchmark the quality of care providers, care models and countries (AdHOC data).
2.3) Describe and perform a method to characterise best practices.
WP3 Recruitment and assessment
3.1) Set up the recruitment protocol for: a) community care organisations; b) users of community care services; c) the subsample of users for the validation of formal and informal service use; d) staff and managers of community care services.
3.2) To set up a standardized procedure for (re)training staff responsible for data collection in participating countries, and to monitor the quality of the recruitment and all assessments.
3.3) Collect data on a) the status of community care users by interRAI-HC assessments (follow up assessments after six and 12 months); b) utilization of formal/informal health care services; c) staff of community care services; d) management of community care services.
WP4 Data management
4.1) To develop a procedure to aggregate and integrate nationally collected IBenC data.
4.2) To set up a web based infrastructure for collection, storage and issuing of all IBenC data.
4.3) Management, quality control and issuing of IBenC data.
WP5 Benchmarking cost of care
5.1) To validate the interRAI-HC for the assessment of formal and informal care utilization.
5.2) To establish European standards for valuation of formal and informal care use of individual patients.
5.3) To establish European standards for valuation of formal and informal care use of individual patients by conducting a review and a Delphi study.
5.4) Development of a European guideline to value health care utilization from a societal perspective.
5.5) Calculating societal costs based on AdHOC and IBenC data.
5.6) Describing and comparing societal costs across countries and care models.
5.7) To benchmark care models on their costs using IBenC and AdHOC data.
WP6 Identifying care models for community care in health systems in six countries
6.1) To review the care system and structure of community care for care dependent elderly in six countries including benchmarking practices.
6.2) To assess the structure of community care in six countries and perform linguistic validation of this assessment.
6.3) Identification of community care models by benchmarking IBenC data on organisation and staff characteristics across six countries.
WP7 A benchmarking method to identify best practices
7.1) To develop a method to benchmark by means of an integrated of quality of care and costs.
7.2) To rank cost-effective community care practices.
7.3) To identify best practice community care models.
WP8 Dissemination of IBenC results
8.1) To develop and regularly update a dissemination plan.
8.2) Maintenance of public website.
8.3) To set up dissemination activities.
8.4) To perform a feasibility study for the development of a web-based benchmarking tool for European community care.
The main S & T results of the work carried out within each relevant WP during the project are described below:
Benchmarking quality of care (WP2)
Summary measure to capture quality of care
One of the aims of the project was to develop a summery measure for quality of care, that enabled benchmarking of care practices. The instrument on which data collection within the project capitalized, was the comprehensive geriatric assessment instrument interRAI-Home Care (interRAI-HC). Prerequisite for this summary measure under development was that it could use QIs generated from the interRAI-HC.
A systematic review of the literature was performed through electronic searches to identify quality of care indicators for home care. The search reviewed, refined, and selected quality indicators extracted from the interRAI-HC instrument which comprise a) outcome indicators based on the status of individual clients and its change in the course of one year; b) process indicators to measure the production of care and its efficiency; c) indicators for management rations for comparison of structure and explore the feasibility of its’ use. Members from the consortium could also propose papers for inclusion. A search was conducted in MEDLINE for papers published until March 2013. Studies were found to be eligible if the publication described the development process and/or characteristics of Quality Indicators (QIs) specifically for home care, numerators and denominators were described or could be deduced, or a description of performance standards was given; publications in English, or a language that could be translated by consortium members; described QIs that could be extracted from the interRAI-HC.
A total of 539 papers was identified in the search, of which five papers met the inclusion criteria, of which one paper was still in progress. In total 43 QIs were identified, 30 outcome and 13 process QIs. No QIs on structure or management were identified. The main finding of this review was the identification of two interRAI-HC summary measures, at that time still in development (Morris et al, 2013) The two summary measures were composed from 20 interRAI-HC QIs and represent different types of care quality. The Independence Quality Scale (IQS) reflects the absence of decline (defined by functional decline) and focusses on functional independence and engagement. The IQS is generated from 11 QIs; ADL decline, IADL decline, cognitive decline, communication decline, not going out, falls, injuries, hospitalisation/ED visits, mood decline, bladder decline, and uncontrolled pain. The Clinical Balance Quality Scale (CBQS) reflects functional improvement and return to clinical balance, and consists of nine QIs: ADL improvement, IADL improvement, cognitive improvement, communication improvement, bladder improvement, pain improvement, caregiver not distressed, and not alone and distressed. Both scales have good internal consistency and behave very differently, as was reflected by the strong negative correlation between the scales (r = -0.59). This work was very meaningful for the IBenC project and the development of the benchmarking method on costs and quality of care.
Testing and applying the summary measures
A second step was to test the IQS and CBQS in a sample of European community care clients on fitness for the project’s purposes. To this end, the AdHOC study (FP5; 2001-2003) sample was used, since data collection within IBenC was ongoing. A total of 1354 community care clients from six countries were included in the analyses. Quality of care was benchmarked on country level. Scores for the IQS and CBQS varied across countries, and identified areas for quality improvement of services (Foebel et al, 2015). This study showed the suitability of the IQS and CBQS for achieving the project aims.
Recruitment and data collection (WP3)
A study protocol was prepared and medical ethical approval for the study was sought from the relevant legal authorised medical ethical committees in Belgium, Finland, Germany, Iceland, Italy and the Netherlands. Community care organisations in the six countries were selected based on differences in organisational structure, processes and type of regions.
Data on care organisation characteristics (Q2) was collected via the key person of the participating care organisations via online data collection. Each care organisation was provided with a web link, an individual login code and password. Administration of the organisation questionnaire took in most cases approximately up to an hour. If required, assessment support from the local research centre was provided. Care staff characteristics were collected with a self-report questionnaire (Q2) and were collected predominantly online. Completion of the staff characteristics questionnaire took approximately 30 minutes.
Patient outcome data were digitally collected at the homes of care recipients by trained (research) nurses using licenced software. A training protocol for the project is available. Data collection on clients capitalised on the interRAI-HC, a validated comprehensive geriatric assessment instrument providing information on the medical, psychological, social and functional capabilities, care needs, and formal and informal resource utilisation of community dwelling older adults.
All assessors followed a standardized training for admission of the interRAI-HC. In care organisations using the interRAI-HC in routine care practice, only new staff was trained (Belgium, Iceland, Italy, and Finland). In Germany, care organisations were new with the use of the interRAI-HC and nurses were trained prior to the assessments. In The Netherlands research nurses were trained. For completion of the interRAI-HC all sources of information are used: patient interviews, care files, observation, and information obtained from informal and formal carers. Duration of an interRAI-HC assessment was approximately 60 minutes. Additional interviews with the RUD-Lite and EQ-5D, with the aim to validate the interRAI-HC for resource utilisation, were conducted within one month of the interRAI-HC assessment (at baseline or at six month follow up). Validation interviews were performed with the care recipient, and if available an informal carer. In case of cognitive decline (CPS ≥ 3), the validation interview was performed with the informal carer.
In total 38 community care organisations participated in the study: 18 in Belgium, three in Finland, 11 in Germany, one in Iceland, two in Italy, and three in the Netherlands. The response on the questionnaires on organisational characteristics was 94.7%. The community care organisations provided predominantly nursing care only (66.0%). Of the 38 community care organisations, 16.7% operated only in rural areas. Organisation size varied considerably, the average number of (second) level nurses ranged between 33.9 up to 4224.0 per organisation (M = 399.0; SD = 1519.5).
A total of 2884 community care clients were enrolled in the study at baseline, divided in relatively equal size samples for all countries: 525 (18%.2) from Belgium, 456 (15.8%) from Finland, 493 (17.1%) from Germany, 420 (14.6%) from Iceland, 499 (17.3%) from Italy and 491 (17.0%) from the Netherlands. Two third of the community care clients were female, the average age was 82.9 (SD = 7.3) almost one third of the participants was married of with partner (31.0%), and the majority of the participants lived alone (57.7%). Overall 414% of the community care clients needed extensive ADL support (ADL ≥ 3) and 17.6% experienced moderate to severe cognitive impairment (CPS ≥ 3). In general, community care clients in Italy had the highest level op dependency and community care clients in the Netherlands had the lowest impairments.
In total 1067 care professionals participated from 31 organisations. The overall response was 60.9%. The majority of the respondents was female (93.4%), nurse (67.5%), over 45 years of age (52.8%) and had on average 11.6 years of experience in home care (SD = 10.0).
The data collection resulted in an unique data set for community care in Europe, containing longitudinal data on community care clients in six different countries, and cross sectional data on care organisations and staff characteristics that can be related to client outcomes.
Data management (WP4)
The data collection within the project and the creation of the project data base formed a challenge for the team: collecting data on different levels (patient, staff, and organisation), the longitudinal design of data collection amongst patients, the need to link all different data sets, the different languages of data collection, and the different software programs used in the countries. This required a dedicated data management team to process, clean, integrate, and store all project data. This work resulted in a secured web-based database consisting of 92 data files, stored on a secured data issuing system. The data base is arranged in such a way that sections of instruments can be issued separately, the same can be done with the assessment waves. All records in the data sets have unique identifiers, enabling linking to other data sets. Downloadable code books are also available from the system. The work in this WP resulted in a unique and deployable data base that in the future will be available for researchers outside of the consortium who are seeking for collaboration. Data are available on request at www.psydata.nl.
Benchmarking costs of care (WP5)
Cost of care assessments for community care from a societal perspective
To determine whether the interRAI-HC could be used for cost of care assessments, a study was conducted to evaluate the convergent validity of the interRAI-HC instrument in comparison to the RUD-Lite instrument for the measurement of formal and informal resource utilisation among care-dependent community dwelling elderly (Van Lier et al, 2016).
A cross-sectional within subject design was used. Data collection took place within a subsample of 790 community care clients participating in IBenC. The interRAI-HC and RUD Lite instrument were administered within a one month period by trained (research) nurses. The RUD Lite measures resource utilisation from a societal perspective among older adults with dementia and has good psychometric properties for costs of care assessments. Resource utilization was valued by using Dutch standard costs. We hypothesised that there was strong agreement (Spearman’s ρ > 0.5) between ten resource utilisation categories, six costs of care categories and total societal costs estimates derived from both instruments.
Strong agreement was found between RUD Lite and interRAI-HC resource utilisation assessments for physician visits, physical therapy, psychological treatment, frequency and duration of hospital admission, homemaking services, meals on wheels, and informal care. The agreement for home health care and home nursing was weak, and the agreement for occupational therapy and emergency room visits was moderate. Agreement for all six cost categories; home care, physician visits, other health care services, hospital admissions, supportive care services, and informal care, was strong. The difference in total societal costs was significantly lower for cost of care estimates calculated with the RUD Lite as compared to the interRAI-HC (€ -804; 95% CI €-1340 - €-269). Agreement between the instruments for total societal costs was strong (Spearman’s ρ = 0.60).
The results suggest that the interRAI-HC instrument can be validly used to measure resource utilisation of formal and informal care services and to estimate societal cost of care utilisation in community dwelling older adults. Since interRAI-HC assessments are part of routine care in many organisations and countries already, this finding increases the feasibility of performing economic evaluations among community dwelling older adults with the interRAI-HC.
Benchmarking care models on costs of care utilisation from a societal perspective
The way in which home care is delivered by care organizations within and across European countries varies considerably. Variation exists in terms of funding, organizational structures, care processes, access and quality of services, reimbursement systems, and public versus private delivery. These variations in home care delivery can be expected to lead to differences in costs of care utilisation. To prepare for a future increase in long term care needs of dependent older adults, it is important to get more insight in community care provision and associated costs for society across different types of home care models within and across countries. When differences in costs between home care models are identified, there may be opportunities for significant cost savings by learning from best practice. A study was conducted to benchmark care models for cost of care utilization in two different data samples (AdHOC and IBenC).
Care models identified by Henrard et al (2006) were used for this purpose, these included a medical model, a medico-socio model and a fragmented model. Costs of care in both samples were estimated for a period of six months by extrapolating resource utilization at baseline and 6-month follow-up assessments and valuing this with Dutch standard costs. Cost of care estimates were adjusted for case-mix.
A total of 1080 participants from AdHOC were included in the analyses. In the AdHOC sample total societal costs in the fragmented model were lowest, but did not statistically differ from the medico-social model (mean difference €1589, 95% CI -1129; 4280. Total societal costs in the medical model were statistically significantly higher than in the fragmented model (mean difference €9501, 95% CI 5839; 13518) and medico-social model (mean difference €11089, 95% CI 7802; 14567). Cost of care estimates were adjusted for age, gender, CPS, DRS, ADL, IADL, and CHESS.
A total of 2030 participants from IBenC were included in the analyses. In this more recent sample of community care clients, different results were found. After adjustment for case mix, the medical model was associated with the lowest costs per client , and the medico-social model was associated with the highest costs per client. The cost-differences between all three care models were statistically significant. The difference between the fragmented and the medico-social model was €-7080 (95% CI -11548; -2874), the difference between the medical and the medico-social model was €-11053 (95% CI-15705; -7269), and the difference between the medical model and the fragmented model was €-3973 (5% CI -7850; -176).
The calculated costs in the benchmark do not reflect actual costs. Adjusting the costs allows for reliable comparisons between care models, and organisations independent of the populations they serve. By adjusting actual costs for case mix, the method provides insight into meaningful differences in costs between care practices.
Another achievement of this WP5 was the execution of a Delphi study to develop cross-European recommendations for the identification, measurement and valuation of resource use and lost productivity in economic evaluations. This Delphi study was based on recommendations from existing guidelines that were identified in a comprehensive literature search. The Delphi study contained of three written rounds in which European health economic experts participated.
Recommendations from 18 guidelines were extracted. Consensus among 26 panel members from 17 countries was reached to adopt a societal perspective when performing a health economic evaluation, to use self-report for measuring resource utilisation and lost productivity, to value both constructs using country-specific standardised/unit costs and to use country-specific discounting rates.
This study provides consensus-based cross-European recommendations on how to measure and value resource utilisation and lost productivity in economic evaluations. Based on the outcomes of this study a costing guideline for use in cross-European health economic evaluations was written.
Identifying care models for community care in health systems in six countries (WP6)
Differences in health care practice and policy in an European context
To understand why some organisations and health care systems provide better quality of care than others, studying the micro (client), meso (organisation) and macro (policy) levels is equally important.
A review was conducted to compare the context, regulations and conditions for community care system delivery to care dependent elderly on the macro (country / regional) level in countries that participate in IBenC (Belgium, Finland, Germany, Iceland, Italy and the Netherlands) (Van Eenoo et al, 2015). The following dimensions of community care were described and benchmarked: governmental vision, governmental regulation, funding, organisations and professionals, eligibility criteria and equity and the involvement of informal caregivers and placed in the context of proportion of aging persons and welfare level. To this end, indicators were used based on two recent literature reviews on community care in Europe and an international comparative study. The descriptions from EurHOMap of the community care system were updated per IBenC country.
In summary, the Netherlands have the most formal community care with low informal care use, while Italy mainly relies on informal care. Community care in Belgium distinguishes itself from community care in the other countries by the definition of nursing care, which also includes personal care such as washing, toileting and dressing. In Germany, more than half of the community care organisations are for profit and are very small. Since municipalities are primarily responsible for community care in Finland, Finland has the most decentralised community care regulation. Finally, the community care system in Iceland is less formal, due to the young population. Nevertheless, Iceland is also preparing the community care system for an ageing population in the near future.
Identifying care models
In order to identify best practices, we first needed to categorise community care organisations into community care models and identify similarities in care practice for this purpose. The aim of this part of the IBenC study was to identify community care models and to classify participating community care organisations accordingly. The way care is provided varies largely in Europe as was identified by the literature review that was conducted on macro indicators for community care. Accordingly it was expected for our sample. The community care organisations recruited for the study offer nursing care (activities of nurses that are of a technical, supportive or rehabilitative nature), personal care (assistance with activities of daily living (ADL) such as dressing, feeding, washing and toileting and getting in or out of bed) and/or domestic care (help with instrumental activities of daily living (IADL), such as shopping, food preparation, housekeeping, transportation, taking medication and financial administration) in the community.
To identify organisational community care models, we used data collected about the care organisations. A questionnaire was designed specifically to capture data on structure and processes of care organisations (Q2). Q2 was structured along four main themes: 1) the structure of the organisations: e.g. type of the home care organisation, ownership of the organisation, location and size of the service region; 2) the caseload of the organisations: e.g. total number of clients, number of clients who are 65 years or older, and total number of client visits during the previous calendar year; 3) the characteristics of care professionals: e.g. disciplines of care professionals, number and FTE per discipline, turn-over rate and education and training of the care professionals; and 4) the organisational processes: e.g. professionals on call, care management, referral and transfer procedures, meetings and accountability. All participating community care organisations were asked to fill out Q2.
To identify community care models, only dichotomous nominal variables were used. Fifty-four variables were included: 14 variables concern structural elements, 29 variables concern care coordination, five variables concern meetings, and six variables concerning accountability. A score of 1 was assigned when the variable is true for a community care organisation and a score of 0 was assigned when the variable is not true for an organisation. If data was missing for a variable, we assumed that the variable was not true for the organisation and a score of 0 was assigned.
The analyses were performed in three steps. In a first step we conducted a principal component analysis (PCA) with varimax rotation to determine the meaningful number of components to retain for further analyses. In a second step, a hierarchical cluster analysis was carried out on the generated components to group the community care organisations into clusters. In a third step, the clusters of community care organisations are interpreted by using a 3D plot and by calculating the organisations’ scores on each of the generated factors. Analyses were carried out with SPSS version 23.
The six European countries included 38 community care organisations, of which 36 managers filled out a questionnaire on the characteristics of the community care organisations.
The PCA resulted in a rotated matrix with three components, 15 variables, a KMO value of 0.76 and a significant Bartlett’s test (p<0.001). The three rotated components accounted for 75.4% of the total variance. The first rotated component consisted of variables emphasizing care planning and involving clients and informal carers in care planning accounted and for 28.9%, this components is called patient-centred care delivery. The second component (availability of specialized care professionals) contained variables on availability of specialized care professionals in the organization (e.g. dementia care, or palliative care nurse) and accounted for 24.7% of the total variance. The last component for 21.8% of the total variance and was formed by items on whether quality of care was systematically assessed and internal meetings were held within the organization. This component was called monitoring of quality of care.
The cluster analysis on the 36 community care organisations identified six organisational community care models:
Care model (CM) 1: A (very) strong focus on patient-centred care delivery, a high availability of specialized care professionals, and a strong focus on monitoring of care performance. Fifteen home care organizations corresponded to this model.
CM2: A very strong focus on patient-centred care delivery, little or no availability of specialized care professionals, and a very strong focus on monitoring of care performance.
CM3: A strong focus on patient-centred care and a high availability of specialized care professionals, but limited attention for monitoring care performance.
CM4: A very limited focus on patient-centred care delivery, no availability of specialized care professionals and a little focus on monitoring care performance.
CM5: A strong focus on patient-centred care delivery, a low availability of specialized care professionals and a low focus on care delivery outcomes.
CM6: A very limited focus on patient-centred care delivery, little or no availability of specialized care professionals, and a strong focus on monitoring care performance. One Belgian and nine German home care organizations corresponded to this model.
Organisations that could be categorized according to care models 1 and 2 were a mix of organisations from different countries, in care model 6 organisations from two different countries were present, in the ration of 1 against 9. Organisations from one single country were classified in care model 3. Finally in care models 4 and 5 only one organization could be classified. The results for these two models were not considered robust. It was decided that these two models were not considered for further analyses within the project.
Although the identified models might not be totally representative for European care practice, since analyses were solely based on organisations participating in IBenC. It might be possible that we were not able to collect information and practices that differed from these. However, the identified care models, with exception of models 4 and 5, seem robust and explain almost 75% of the total variance in care practice. The care models are considered fit for use in further analyses of the IBenC project. Also the relevance of the three identified core domains to distinguish practices in community care, patient-centred care delivery, specialized care professionals and monitoring of care performance has been reported in literature earlier. The identified care models will be used to benchmark quality and cost of care.
A benchmarking method to identify best practices (WP7)
The need for cost effective health systems is imperative in the context of increasing care demands of a rapidly aging population and a shrinking working force. High quality community care may prevent costly institutionalisation. There is ample evidence on what type of community care provides best outcomes against reasonable costs. In this WP all technology and knowledge gathered within the IBenC project came together in order to develop a solid new benchmarking method. This method is described and applied in this study, and best practices and predictors for best practice outcomes are explored.
IBenC capitalises on a standardised instrument (interRAI-HC) that is widely used across Europe in routine care to assess patient outcomes, care use and quality of care. To determine organisations’ performance and estimate costs of care utilisation, patient data from baseline and six month follow-up assessments were used. Costs of care over a six month period were calculated from a societal perspective, taking health care, welfare, and informal care costs into account. First, care utilisation was inventoried with interRAI-HC items that assess care utilisation from a patient perspective, registering the amount of community care, physician visits, other health care services (e.g. physical therapy), hospital admissions and emergency room visits, supportive care services, institutional care, and informal care. Quality of care was represented by two summary scales, each ranging from 0 to 10. The Independence Quality Scale (IQS) reflects care associated with maintenance of functioning and prevention of functional decline, while the Clinical Balance Quality Scale (CBQS) gives on overview of quality of care that is aimed at improvement of functioning. Costs of formal and informal care utilisation were estimated using Dutch standard costs. Quality of care outcomes and costs of care utilisation estimates were case-mix adjusted. To perform benchmarking of efficiency of community health care practices, the IQS and CBQS were integrated with the societal costs of care into two new costs-quality indexes: the IQS index and the CBQS index. To enable an in-depth interpretation of best practices, the characteristics of community care organisations (Q2) and staff characteristics (Q1) were related to the IQS index and the CBQS index in order to explore predictors for efficiency.
Within the IBenC sample, mean adjusted societal costs were €21,004 for a community care client over a six month period, ranging from €14,300 to €24,209 across organisations. For 14 community care organisations IQS, CBQS, and societal costs could be calculated. Belgian organisations were excluded from the benchmark: due to a missing variable informal care costs could not be calculated. Other organisations were excluded for the benchmark due to small sample sizes. QIs cannot be calculated for less than 20 eligible clients and costs estimates were not considered reliable when calculated for less than 20 clients per organisation. Scores for individual organisations on the IQS ranged from 2 to 7, scores on the CBQS ranged from 4 to 8.
Choosing best practices based on both costs and quality of care is meaningful in the light of sustainable care systems. Where a good balance between costs and quality of care is taken into account, e.g. delivering care that is of a reasonable quality standard, against reasonable cost, and avoiding care provisioning that is of very high quality, but also extremely expensive. In other words, we are looking for efficient organisations. Plotting the summary scales against the costs of care utilisation, enables to distinguish organisations with poor efficiency from those who are highly efficient, by means of the upper left and the lower right quadrant. But this method does not give a decisive answer on how to rank all organisations on efficiency, also the ones that fall in the lower left quadrant (low quality and low costs) or the upper right quadrant (high quality of care and high costs). Therefore the cost-quality of care index for both types of care was developed: the IQS index and the CBQS index. The indexes are formed by combining the estimates on average societal costs and performance of an organisation into one outcome measure. Both reflecting the organisational efficiency on respectively care directed at maintenance or prevention of functional decline, and care focused on improvement of functioning relative to other organisations. Average efficiency is valued by a score of 1 on the index, whereas higher scores indicate better efficiency and a score lower than 1 indicates poor efficiency. The index is a continuous measure and distinguishes between organisations on different efficiency scores. Face validity for the IQS-index and the CBQS-index was high as compared to the plotted costs against the IQS and CBQS summary measure scores.
Five out of the 14 organisations included in the benchmark scored > 1 on the IQS index, indicating better than average efficiency on care related to functional decline. Only the care model in which care practice had a limited focus on patient-centred care delivery and a low availability of specialised care professionals had a IQS-index value > 1 and was considered efficient.
All organisations, except for one, scored > 1 on the CBQS. For the CBQS-index, the only care model not considered efficient and with an CBQS-index value < 1 was the same model that was efficient in care provisioning aimed at maintenance and prevention of functional decline, care practice in this model had a limited focus on patient-centred care delivery and a low availability of specialised care professionals.
Stepwise linear regression analysis was performed to explore predictors for the IQS index. Independent variables included in the analysis related to care professionals were: physical workload, emotional demands, quantitative demands, work pace, influence at work, possibilities for development, predictability at work, having a part-time job, and having a temporarily contract. Furthermore patient-centred care delivery, availability of specialized care professionals and monitoring of performance were included in the analyses for the organisational perspective. The final model was highly significant (F = 74.00 p < 0.000) and explained 47.2% of the total variance of the IQS index. When controlling for patient-centred care delivery, availability of specialised care professionals, and monitoring quality of care, the model became even more significant (F = 101.92 p < 0.000) and explained 64.5% of the total variance in the IQS-index. Predictors for efficiency related to maintenance, and prevention of functional decline were lower levels of predictability, emotional demands, and influence at work, higher levels of physical workload, and working full time, a limited focus on patient-centred care delivery, a focus on monitoring quality of care in the organisation, and availability of specialised care professionals.
A similar predictive model was built for efficiency related to care aimed at functional improvement (CBQS-index). When controlling for patient-centred care delivery, availability of specialised care professionals, and monitoring quality of care, the final model explained 54.8% of the total variance (F = 89.85 p < 0.000). Higher levels of experienced influence and predictability at work, lower levels of monitoring quality of care and availability of specialised care professionals, working part-time, and having a permanent contract were found to be predictors for efficiency related to improvement of functioning.
The benchmarking method developed in this study is a new and powerful method to compare organisational efficiency, taking into account quality and costs of care. The method helps to gain more insight into the functioning of health care systems. The explorative predictive analyses performed in this study are not to be generalise to European community care in general, since the sample it was tested on was relatively small. But the high proportion of explained variance in the models, indicate that the identified variables highly likely are important for organisational efficiency. The IQS and CBQS reflect different types of quality. It is known that quality of care is not a unidimensional concept, and this is reflected in the results, as most of the organisations do not perform equally on both summary scales (thus performing differently on the two types of care). A possible explanation is that organisations perform well on the type care clients they regularly serve require. If these clients in general are sort term clients that need rehabilitation, organisations will in general perform better on the CBQS. If organisations serve mainly clients with chronically illnesses that need support to remain in the community, organisations probably perform better on the IQS. The study does not contain information on the type of referral of clients, or about specific missions and visions of the community care organisations to confirm this hypothesis.
The benchmarking method capitalises on the interRAI-HC instrument, which is used in numerous organisations in many countries in routine care at a client level. Therefore this method does not require additional, burdensome assessments for care organisations. A wide implementation of this method gives a rigorous view of the functioning of the European care system and can help organisations and policy makers identify, and learn from, best practices in terms of efficiency. The method can support health care policy makers in their search for sustainable health care systems.
Feasibility Study on online benchmarking tool
A feasibility study was conducted to explore the need for a benchmarking method as developed within the IBenC project and to explore whether it is technically feasible to apply the benchmarking method in an online tool. The aim of this study is to explore the feasibility of an interactive web based benchmarking tool on community care, which would ‘transfer’ online the benchmarking method developed by the IBenC consortium. To provide a complete overview on the feasibility of the web based tool, the study has been structured in the most comprehensive way, covering general requirements and development processes, definition of a possible tool architecture, an analysis of an already developed comparable web benchmarking tool (Your Health System by CIHI), stakeholder feedbacks and identification of possible challenges and points of attention.
The hypothetical tool architecture, developed following brainstorming sessions and exchanges at both internal level and with the external partners, entails in general terms four key steps, covering the key stages of an effective web-based benchmarking tool, together with the identification of possible challenges and obstacles: data collection and data upload; data cleansing and validation; calculation and metrics; results visualization.
The Feasibility Study results show that the development of an online benchmarking tool based on the IBenC methodology is not only possible, but also advisable and extremely valuable to improve quality in healthcare at European level.
However, the scope of the project and the complexity behind the innovative benchmarking method developed by IBenC, makes it necessary to carefully focus on some stages and aspects of its development and utilisation that might pose particular challenges, as data handling and necessary human and financial resources.
The topic of 2012 Health 2012.3.2-1 called for improvement of the organisation of health care delivery in a pan-European manner. In specific, the main objective of this topic was to benchmark best practices regarding the structure, care processes and performance of health care organisations in Europe. This was exactly the main objective for the IBenC project. By combining the expertise of members of the interRAI network in Europe, the consortium succeeded to provide an evidence based, integrated, and quantitative approach to benchmark community health care organisations with respect to quality and costs of care. The method reliably identifies the most efficient practices on regional, national, and international levels. When relating organisational efficiency with organisational processes and structures, and staff characteristics, the methodology provides evidence based directions for care organisations and health care policy makers on how to reform our current health care system to a future sustainable health care system.
The project consortium partners have published the project’s findings in a range of media:
- in scientific papers published in peer reviewed journals (five papers to date, including three open access journals), and national and international scientific conferences. Up to now three dedicated symposium on the IBenC project were held, and oral presentations in 15 scientific conferences;
- in newsletters, press releases, and social media, including Twitter, LinkedIn, ResearchGate, and three webinars that are available on YouTube;
- through the dedicated IBenC website;
- in meetings for a wider audience, such as the Public Health Research Day workshop held in Brussels in February 2017, addressing policy makers; the Amsterdam Center on Aging Conference held in Amsterdam in March 2016, addressing policy makers, health care providers, patients, and scientists; and a meeting in Lausanne in March 2017 for policy makers organized by the Department of Public Health of the Canton of Vaud (Switzerland). The Department of Public Health is in the process of implementing interRAI on a regional level in order to set up a regional benchmark on quality of care.
IBenC’s potential impact was discussed in the Description of Work in four categories: 1) academic impact; 2) economic impact; 3) impact on industry; and 4) societal impact. The achievements are discussed below.
1. Academic impact.
One of the unique features of the IBenC project is that the project capitalises on one single assessment instrument to obtain cost of care and quality of care information. This instrument, the interRAI-HC, is a validated comprehensive geriatric assessment instrument, and often used in routine care to support the direct care of clients. This makes the IBenC approach very patient-centred.
The IBenC project hopes to push the field of Health Technology Assessment (HTA) further. The validation of the interRAI-HC for cost of care assessments from a societal perspective in community care already provides new opportunities for HTA researchers to obtain reliable cost of care information from routine care data. When case-mix adjustments are applied to these data, the data support reliable benchmarking of costs of care utilisation across organisations, regions, or countries. Information for such comparisons is already potentially available, since every country where the interRAI-HC is being used has its central database. The field of HTA is pushed even further, by the integration of the cost of care estimates with the summary scales for quality of care. The IQS and CBQS are composed of reliable, case-mix adjusted quality indicators, extracted from the interRAI-HC assessments and discriminate organisational performance with respect to two different types of care. The summary scales are benchmark scales and reflect the relative care performance of an organisation in relation to other organisations in the benchmark. The integrated measures (IQS-index and CBQS-index) on cost and quality of care provide a innovative perspective on the relationship between costs and quality of care to HTA research. Usually this relationship is described by means of the incremental ratio, which expresses the costs of quality gains, whereas the IQS- and CBQS index reflect the relative balance between costs of care utilisation and provided quality of care. In other words, are the costs of care in proportion with respect to the organisational performance, or is an organisation efficient.
The comprehensive approach adopted by the IBenC project also brings the field of health care research further. By collecting data on organisational characteristics on structure and processes, and on staff characteristics from the participating community care organisations, in addition to longitudinal patient data, a unique international database on community care is created. By our knowledge this is a first attempt to create a database on community care including different stakeholder data, and this database will provide health care researchers more insight in the current status of European community care clients and the general functioning European community care.
2. Economic impact
The current organisation of our health care system is not sustainable for the future. Costs of caring for elderly people will put a large economic strain on the European society. The costs of care will rise exponential since the number of people over 65 years of age will be doubled by the year 2060 in comparison to 2010. As people are expected to grow older and healthier, the relative proportion of people of 80 years and older will increase even more. Higher age is associated with a higher risk of co-morbidity and therefore relatively more people are expected to be dependent on professional care services in the coming decades. Concomitant with the rising number of elderly people in the future, the proportion of younger people will decrease. This demographic shift will lead to a projected shortage of professional and informal carers and puts a strain on the labour market. Currently main cost drivers for health care in elderly people are hospital admission and admission to long term care facilities, the key for cost reductions in elderly health care spending lies in community care. To move forward to a sustainable health care system, a shift from institutionalised care to community care has already been initiated in the past years in many countries. But adequate and good quality community care is essential to avoid high expenses and adverse effects. In order to monitor care performance and costs, reliable and comparable costs and quality information on care as collected within the IBenC project, is essential. This type of information is currently unavailable on a large scale. Since the data used for the benchmark are case-mix adjusted, comparisons are not influenced by the severity or dependency level of the population organisations serve. The benchmark therefore reflects realistic differences caused by practice.
The benchmark provides informed directions for stakeholders on organisations that are in need of changing their practice. Either because the organisational performance is substandard, or due to disproportional high costs of care utilisation. The indexes provide specific information on whether to choose for an organisation delivering very good quality of care, but against very high costs, or for an organisation performing less in terms of quality, still meeting reasonable quality standards, but against less care costs. These are considerations that will become more important in the future, where the costs of care (as cheap as possible) can prevail over quality of delivered care. In taking these choices, quality of a reasonable standard should be guaranteed, and the efficiency indexes deliver exactly this information.
When relating the efficiency indexes to characteristics of organisations, and care staff, insight can be gained into what actually contributes to a better organisational efficiency. Explorations on this were performed in the project, but these results are not to be generalised, since the (number of) participating care organisations are not representative for community care in Europe. When applied on a large scale, generalizable information can be obtained on factors within organisations that lead to better efficiency. This is valuable information for care providers and policy makers in order to reform care practices evidence based. Changes that can be initiated can be relatively small, such as changing the level of influence, or predictability of work for staff, but also larger as starting actively to monitor quality of care performance in the own organisation.
Since this benchmark is driven by the interRAI-HC, which is used in routine care and directly supports patient care, the data collection does not put any extra burden on care professionals in addition to their standard practice. And in contrast to many other (legally required) benchmarks, data collection for this benchmark does not require any additional financial efforts from care providers. In conclusion, the IBenC benchmarking method is a powerful and practical tool to push the reform of the our current health care system towards more cost-effective health care provisioning.
3. Impact on industry
The interRAI assessments are done digitally by health care workers, and in many countries SMEs have developed information technology for this. In the IBenC project data was collected with software solutions from six different SMEs. A widespread adoption of the benchmark method will offer them new opportunities in terms of rising demands for assessment software, since more care organisations will be starting to use interRAI-HC in their care practice. It also offers opportunities to ICT industry in general. To use the full potential of the benchmark method, information should be easy accessible, up to date, securely stored, and as current as possible, which could be realised by a web-based benchmarking tool. Within the project the development of such a tool was not foreseen, however the executed feasibility study on such a tool resulted in a potential architectural blue-print of the benchmarking tool. The study encompasses both user and technical requirements, opportunities and threats, needs and acceptability of such a tool, therewith providing a ‘controlled’ way of introducing innovation in the sector without the blind investment in a costly and under-researched instrument and provides an excellent starting point for further development of a web-based tool.
4. Societal impact
The aforementioned increase of the proportion elderly people, including the number of frail elderly people with multiple chronic diseases, will lead to higher demands on care services and higher economic burden for society in the future. Despite being care dependent, most people prefer to remain in their own environment and in their community. Community care is much cheaper in comparison to institutionalised care, and therefore seen as the solution for the future sustainability of the health care system by policy makers. Policy reforms that support a shift from institutionalised care to community care have been initiated and implemented in various countries in Europe in the past years. Such a shift can only benefit care-dependent elderly if community care is adequate, sufficient, and cost-effective. The results of IBenC can contribute to this in multiple ways. The use of the interRAI-HC in care practice has proven its merits for direct patient care, monitoring of quality of care, and care improvement trajectories. Nurses are empowered by using the instrument, the regular and systematic way of assessing clients provides them a comprehensive understanding of the cognitive, functional, clinical, social, and psychological status of their clients. The output of the assessments provides specific information for individual care plans, gives insight into changes over time in terms of improvement or decline in patient status, and the interRAI-HC generates signals for patients at risk of decline. The use of the interRAI-HC instrument leads to more adequate care provisioning, tailored to individual care needs. Ultimately this will lead to a higher quality of life of community care clients, since their care needs are met, institutionalisation is avoided or delayed, and people can remain in their own environment.
With regard to monitoring of care, the IBenC benchmark method provides an overview on the strengths and weaknesses of care providers. The benchmark capitalises on two kinds of information: costs of care utilisation, and two summary measures representing quality of two different types of care. The first type is care aimed at maintenance of functioning and prevention of functional decline (IQS), and the second type of care is aimed at improvement of functioning (CBQS). Both types of care are important in community care, since the target group has diverse needs. The community care setting differs from institutional settings in which the aim of care is mainly focused on prevention of decline, instead of on improvement, or hospital settings where the focus is on recuperation of patients. Community care providers serve both types of patients, the ones that need support in managing chronic diseases, and those in need of (short term) rehabilitation care. By publishing information on the care performance and efficiency of organisations, clients can make an informed choice on the provider that will most likely meet their care needs. Furthermore, the benchmark provides other stakeholders direct and relevant information to push the efficiency of care providers to a higher level. Care providers themselves get a better insight into their own functioning, as compared to other organisations and offers the opportunity to learn from each other. By having insight into each other’s practices, organisations can change their practice according to identified ‘best practices’ in order to become more efficient.
The individual quality indicators, that form the two quality of care summary measures are of importance for organisations as well. These QIs provide a profound insight on specific care domains and identify areas on which specific care quality improvement trajectories are necessary.
Finally, the benchmark provides valuable information for health care insurers, policymakers, ministries, and inspectorates. All are powerful stakeholders with regard to national and regional care provisioning and policy. The first, the care insurers can base their care purchasing on organisational efficiency, using the benchmark. By reimbursing efficient organisations higher than less efficient organisations, they provide organisations with strong financial incentives as an extrinsic motivation for organisations to improve and strife to deliver good quality, cost-effective care. Also the second group can provide strong extrinsic motivations to organisations in order to push towards sustainable health care systems. The benchmark provides information to change policies and legislation, and helps to ensures that quality of care does not fall below a critical level, but will be sufficient to support elderly in the community. Overall, good data, good standardised and rigorous monitoring will not only result in better informed medical decision making, but also in better informed policy making, and consequently in better care for individual patients.
The impact of the IBenC project can reach further than community care alone. Since the method capitalises on the interRAI-HC, that is part of a suite of instruments for dedicated to settings for vulnerable patient groups, it is to be expected that this method can be developed and applied to other care settings as well. The interRAI suites covers settings as long term care, palliative care, acute care, mental health, etc. All interRAI instruments have the same core items, with additional setting specific items. The instruments generate quality indicators and in potential quality of care summary measures, as calculated with the interRAI-HC, can be generated from other interRAI instruments too. A similar benchmark as developed for, and performed in community care in current project can potentially be executed in other health care settings. Therewith supporting improvement of organisational efficiency. By widely implementing the interRAI instruments and adopting the benchmark method, society in general, and in specific patients from a large number of health care settings can profit from well-organised care, receiving good quality care against reasonable costs.
List of Websites: