In search of the quickest way to disseminate health care innovations

Abstract Research Question Innovations in health care are slowly disseminated in The Netherlands and elsewhere. That's why the researchers defined their research question: What is the quickest way of disseminating health care innovations? Research method The design was a comparative, qualitative case study. The researchers invited a group of 52 authors to describe their 21 health care innovations. All case descriptions were published in a book of 261 pages [2]. Results Six types of innovations were distinguished. Most innovations simultaneously improved quality from the patient's point of view (18 out of 21 cases), professional pride (18/21) and speed of introduction (16/21). Clinical outcomes were better or comparable in 13 of the 21 cases. Brainstorm sessions took place with the innovators and the 22 experts on the quickest way to disseminate the innovations more widely in The Netherlands. These sessions looked for the critical success factors for the dissemination of the 21 projects and identified nine. The following factors were identified: 1. A clear distribution of responsibilities between professionals within the innovation (20/21) 2. Enough educational programs about the innovations for the professionals (18/21) 3. Adequate ICT support for the running of the innovations (15/21) 4. Suitable publicity for the innovations (12/21) 5. An adequate payment system for innovative care providers (7/21) 6. The right size of catchment's area for the innovations (6/21) 7. Enough professional freedom to adopt the innovation (5/21) 8. Fast managerial and public decision-making about the adoption of the innovation (3/21) 9. The embedding of the innovations in quality management assurance policy (1/21). Discussion The results of the study had some influence on the political health agenda in The Netherlands, leading to greater emphasis on innovations and quality of care.


Introduction
At the beginning 2002 the former Dutch Minister of Health dr. E. Borst-Eilers gave a New Year's speech to the Royal Dutch Medical Association. She emphasised that many good practices exist in the health services of The Netherlands. However, the dissemination of the good practices to other colleagues or institutes stagnates for many years. This observation was, for the Minister, the reason to invite the first author to compose a book with descriptions of good examples in the field of preventive services, primary health care and hospital care, to analyse common characteristics of them and to come with advice to accelerate the speed of dissemination of health care innovations. The book had to be produced in close cooperation with the Deputy General of health Nico Oudendijk of the Ministry of Health and within eight months. This article is based on the knowledge in this Dutch written book with 261 pages w2x.

Concepts and methods
In 1962 Rogers published his standard work Diffusions of Innovations w3x. A literature review on dissemination of health care innovations by Van der Linden w4x showed that this book, reprinted in 1971, 1983 and 1995, is still leading later papers and books in this field. That's why in the beginning of our research we tried to use Rogers' concepts and models to explain the slow dissemination of Health Care Innovations in The Netherlands. Rogers distinguishes five critical factors with influence on the speed of diffusion: 1. relative advantage 2. compatibility 3. simplicity 4. trialability and 5. observability. The first factor relates to the relative advantage for the professional who uses the innovation. The second factor is the degree with which the innovation is applicable within the existing organisational structure. The third factor is the simplicity of the innovation: are a few or many actors and processes involved to diffuse the innovation? The triability, the fourth factor, has to do with whether the innovation can be diffused in small steps. Or should it be tried out in one big reorganisation of the system? The fifth factor, the observability, relates to the degree the effects of the innovation are visible for professionals and clients.
Most of the innovations we wanted to investigate ( Table 1) are not compatible with the existing organisational structures and are not simple. Nevertheless they have as we will show in this paper, relative advantages for the professionals, are triable and observable. With the absence of the factors compatibility and simplicity Rogers explains why the diffusion of incompatible and complex innovations is so slow in the USA as well as in The Netherlands. However, Rogers does not provide hypotheses to answer the question ''what is the quickest way to disseminate these type of health care innovations''? So we did not opt for a hypothesis testing study but for a hypothesis generating study. That's why we chose for a comparative qualitative case study of good practices in the mentioned fields. We started to answer the research question, using the following concepts.

Concepts
A health care innovation is a change in the delivery of care, consciously chosen by existing organisations with the object of improving the performance of care delivery. In this study we have distinguished six performance fields: safety, clinical outcome, quality of care from a patient's point of view, costs, speed and professional pride. These fields emerged from studying the individual examples. We divided the 21 examples into six types of innovation ( Table 1). We arrived at this classification after studying the individual cases.
The first three types (new methods, standardisation and transmural programs) are innovations of care process. With process we mean steps in carrying out and co-ordinating the activities of health care professionals. A formal description of a process is a protocol.
The other three innovations are structural changes (new structures for the delivery of primary health care, multidisciplinary care in hospitals and infrastructures). By structure we mean an organisational structure with formalised, longer standing task distribution and coordination mechanisms. This can be distinguished from a project (organisation), which is temporary, and from environment, which is not formalised. A structure is not the same as a system, which also contains informal task distribution and co-ordinating mechanisms. With infrastructure we mean a dormant organisational structure, which can be mobilised in certain circumstances. In this research two infrastructures are discussed. The first is the organisation for a nationwide vaccination program ( Apart from this triangulation, the chapter editing focused on reaching a common language for all innovators, a set of common definitions and as complete as possible information about performance and success factors. In an early phase we dropped the notion of best practice: good examples of each of the described practices were available in The Netherlands. We did not compare them. In Table 1  In all parts of this paper we use the following definitions.

Standardisation of existing care
An integrated cataract care program, joint care for total hip or knee replacement patients, a diagnostic mammary carcinoma outpatient clinic: each of these three (case 5, 6 and 7 in Table 1) are examples of care where better performance is achieved by redesigning the care processes. These three became good examples because of external pressure. There was fear of the management of the supplying hospital of being removed (cataract program), of getting a bad reputation about the quality of care (breast clinic) or about the long waiting lists ( joint care program).
Within these examples there was an ambition to learn from international colleagues and from commercial services sectors, and to work with multidisciplinary protocols. According to the innovators themselves the performance includes better clinical outcomes, safety, shorter admissions, less waiting time for patients, more quality for patients, professional pride and greater cost-effectiveness. However, we noticed a lack of robust scientific studies with control groups and pre test and post test designs to support these examples.
A systematic approach with the treatment of cataracts and the hip and knee replacements appears impersonal, because patients are treated in a process like a car in an automatic wash. However, on the contrary the descriptions show more emphasis on the individual patient. But the standardisation created organisational problems for the rest of the hospitals, which are divided according to medical specialty. A distinction in patient flows was necessary to start these types of innovations.

Transmural care programs
Five transmural or shared care programs (cases 8-12 in Table 1) are described in Table 1. They are focused on treatment and care for patients with arthritis, diabetes, COPD, stroke and with psycho geriatric syndromes. In these innovations many simultaneous changes took place. New protocols were introduced everywhere. New professions occurred (rheumatology program). A new infrastructure for emergencies was created (stroke services). Case managers were introduced (psycho geriatric program) as well as systematic monitoring (diabetes and COPD programs). It was impossible to understand which part of the innovations contributed to what part of the performances. Maybe these are innovations in which each of the components empowers the other ones: then the total is more than the sum of its parts. Anyhow, performances were mostly improved on all the specified fields. One factor was common to the five-transmural programs: the driving force of enthusiastic experienced professionals. Other factors differed. For the diabetes program there was external pressure. The small hospital, which supported the program, survived with its broad cureycare approach of diabetes and other diseases (not described here). With this professional objective the hospital overcame many resistances and competence conflicts between family doctors, internists and between doctors and nurses. Support from one or more national agencies was also an important factor for the stroke service and the diabetes program.

New structures for the delivery of primary health care
A new organisation for out-of-hours services of general practitioners in Nijmegen, a joint venture for integrated home and hospital care in the Twente region and integrated primary and long term care in Almere all led to better performance over a variety of fields, as is shown in cases 13, 14 and 15 in Table 1. Improvements of quality and speed of care as well as cost reductions are mentioned in Twente and Almere, although only scientific evidence is available for the latter. The Nijmegen innovation for GP care during evenings, nights and weekends is one of many, which were implemented across the whole country over a period of five years. They essentially improved GPs' job satisfaction, but not their pride. GPs liked their shorter working hours ( job satisfaction increased) but felt guilty not to deliver continuity of care (their pride diminished). The external pressure to introduce this was and is large: otherwise GP recruitment would suffer too much. Maybe both professional quality and the quality from the patient's view improved. However, no research data were available.
External conditions also played a role in Almere: This was that new town and new primary and long-term healthcare organisations could be designed from scratch. The Twente region has clearly defined borders, mostly with neighbouring Germany. Together with a strong and mission-driven health care insurer the defined borders stimulated the start of the joint venture.
The common feature of all three examples is that completely new structures were developed for large groups of patients with many different health problems. Because of these radical structural changes, there were almost daily bottlenecks with financing and other regulations, more than in the other 21 cases.

Multidisciplinary structures within hospitals
In contrast with the transmural care program innovations within hospitals relate to horizontal rather than vertical chains of care. Multidisciplinary outpatient Coronary Heart Disease (CHD) care, integrated oncological care, comprehensive trauma care and regionally integrated emergency services all belong to the central focus of hospitals. They all provided better performances as is shown in cases 16-19 of Table 1. They have two common features. Firstly there is integration between different medical specialists, e.g. surgeons and internists admitting patients on the same oncological ward. Cardiologists and cardiac surgeons cooperate in the CHD clinic. Secondly, more integration also exists between medical specialists and specialised nurses, for instance in trauma care and in the regional integration of emergency services.
Two of the projects are research led, the CHD clinic and the regional emergency services. During the studies, some care bottlenecks emerged. For the regional emergency services these were wrong use of beds, shortages of nursing and medical staff, high sickness absence and the lack of information about available beds. Because of these problems, professionals and managers took initiatives to improve the regional structures.
The trauma care innovation shows that top-down innovations are only successful when there are local enthusiastic professionals to support them. In The Netherlands, the Ministries of Inland Affairs and of Health, Welfare and Sport Affairs prescribed how trauma care should be organised in a common policy document. But case 19 in Table 1 was the only one, which was really implemented.

New public infrastructures
The last two innovations (cases 20 and 21 in Table  1 Table 2 shows the performances of the 21 innovations in comparison with usual care. The scores were based on analyses of documentation from and on interviews with the innovators. Aggregating the data in sub pages 1 to 21 of Table 1 composes Table 2. Most of the innovations improved quality from the patient's point of view (18 out of 21) and professional self-esteem (18y21). Improvements of speed of care delivery also occurred often (16y21). Less frequent findings were cost reductions (9y21) and improvements of clinical outcome (8y21) and of safety (5y21). Ten innovations were carried out without measuring the possible consequences for clinical outcomes respectively safety. Table 3 shows the critical factors in becoming a good example as mentioned by the innovators. Everywhere (21y21) enthusiastic leading professionals were needed to make the innovation work. Nearly everywhere (20y21) this was the case for professionals working as a team. Factors such as external pressure (13y 21), scientific evaluation (12y21) and support by one or more national agencies (11y21) were less important. The use of modern Information and Communication Technology (ICT) was less often important in making successful innovations (10y21). This is also the case for the creation of new professions (10y21) and the support of an independent local developing agency. The infrequent influence (4y21) of organised, local patient groups as a factor in the success of innovations was surprising both for innovators and researchers.

Opportunities for dissemination
Once the 21 cases were described, we discussed their likelihood of being disseminated and the availability of factors, which could make this more successful. The brainstorming took place with the innovators as well as with 22 experts: five from patient organisations, five from the Ministry of Health, Welfare and Sport Affairs and from National Agencies, four from scientific institutes, three hospital Chief Executives, three from non-health sectors and two from consultancy firms.
Four brainstorm sessions were organised for: 1. Innovators, 2. Representatives of patient groups, 3. Other experts within the health care sector and 4. Experts outside the health sector. We made notes of the discussions. Table 4 is based on these and the remarks in the 21 case descriptions. Nine critical factors for dissemination are mentioned in the first column of this table. We discuss these now.
A vague distribution of responsibilities between doctors, nurses and other professionals can reduce the speed of dissemination: professionals want to know for what they are and are not responsible. This, according to innovators and experts, is not a problem: in 20 of the 21 innovations the responsibilities are well described and uncontroversial. Enough educational programs for professionals to introduce the innovations in their own setting can easily be organised: that is no problem in 18 of the 21 cases. Adequate ICT software and hardware can mostly (15y21) be organised to routinely support the innovations everywhere.
The distribution of scientific and experience knowledge about the innovations is a problem for 12 of the 21 innovations. For 9 of the 21 not enough scientific and professional papers, newsletters and websites exist to show new research data and experiences. There are not enough research funds to produce research data and to make the experiences manifest. An adequate payment system is expected for only 7 of the 21 innovations. During our consultation rounds with innovators and experts the opinions did not go in the direction of fee-for-innovative-services systems to stimulate the dissemination of innovations. The disadvantage of such incentives is the threat of pseudo (fake) innovations without genuine quality and efficiency improvements.
A lack of professional freedom to adopt innovations and slow managerial and public decision-making are two sides of the same coin. The professional freedom respectively policy making processes are thought to be adequate in only 5 respectively 3 of the 21 innovations ( Table 4). New processes and structures cannot be introduced top-down, as is shown in the case of the trauma care. At present, health policy making in The Netherlands is focused on the average institution, not on the extremes: the better ones and the worst cases. To create a better position for professionals who want to modernise their services, the innovators and experts expect much of the introduction of an internal market in the Dutch health services.
In this market health insurance agencies have to purchase care of the care providers. In our consultation rounds the metaphor was a comparison of these agencies with citizens buying a new washing machine. They look after price, quality and delivery conditions, and they want the latest model. With this countervailing power, professionals willing to adopt innovations get more space for their initiatives and their managers have to make quicker decisions.
Quality management assurance policy and dissemination of innovations were only expected to be available in one of the 21 cases studied ( Table 4). In theory, professionals and managers should continuously monitor the quality of care. Sometimes, it may happen that quality management can be improved by introducing an innovation. The introduction gets embedded in the quality assurance policy of the institute. Although innovators and experts agreed with this theory, they did not recognise it in practice. Innovation was one thing and a quality system something else.

Discussion
So far, the innovations and the opportunities for disseminating them have been described. But the researchers point out three limitations of their study to readers.
Firstly, the sample size of 21 cases is small, as is always in qualitative research. However, they are thoroughly described, which cannot be done with a cross-sectional, observational study. Secondly, bias may occur because of the selection of only good practices. The selection was made because of the objectives given by the Minister of Health. The 21 innovations were not randomly selected but chosen after serious consideration of the inclusion criteria as described in the introduction. The ordering of the critical factors in Table 3 could be different if we had included failed as well as successful innovations. Future research should be done on failures to get a complete set of critical factors for the success of innovations. The third limitation is the bias, which may occur because 52 innovators described their own work. Maybe they overemphasised the role of enthusiastic, leading professionals and the need for professional freedom in the dissemination process. However, the triangulation procedures mentioned in the introduction minimised the risk of window dressing.