Spreading Improvement
In order to benefit large populations, improvements in health care cannot be limited to a few pilot sites. Spreading successful improvements to a significant portion of a health system and its catchment population is an imperative for large-scale change in health care.
The USAID Health Care Improvement Project defines “spread” as the practice of taking a local improvement—an intervention, the redesign of a process, or a new system—that has demonstrated better results than the current way of doing things and actively disseminating it across a larger segment of the health system or across the entire system.
In planning to spread a proven intervention, three key questions are considered:
- 1. What are we trying to spread?
- 2. Where do we want to spread to, and by when?
- 3. How will we spread?
What are we trying to spread? This question delves into what it takes to implement the intervention and to ensure that it is optimally packaged. The nature of the intervention influences the choice of spread approach. Some interventions require systemic changes that involve the interaction of many persons in the care delivery process, while others are straightforward and can be easily implemented within existing care delivery systems or may require just a few agents to ensure their implementation. In planning a spread effort, it is important to differentiate between the core elements of the intervention—the components that cannot be changed without compromising the intervention—and those that represent variations around that core—elements that merely enable the implementation of the core elements.
Where do we want to spread to, and by when? This question considers the full scale to which the intervention needs to be spread: geographic scale, the number of facilities or health workers involved, the population or number of patients to be covered, etc. The geographic scope and the timeline for reaching it also influence the choice of spread approach. Several factors influence the size of achievable scale and the time required for an intentional spread effort to reach its goals. Some of these factors relate to the improved process identified for spread—its complexity, cost, alignment with existing organizational processes and values—while others relate to the agents/providers who will adopt the improvements: their level of motivation toward adopting the change, their skills and competence, their number, etc. Other factors relate to the institutional environment—existing resources, perceived legitimacy of the intervention, channels for dissemination, and legal context, and presence of a champion (discussed below). The time, cost, and achievable scale of a spread are a function of these factors combined.
How will we spread? This question considers the nature of the intervention and the scale to which it should be spread to determine a suitable spread approach. There are many possible ways to organize spread. The suitability of a particular approach for a particular spread aim depends on the answers to the first two questions.
A Conceptual Model for Spread
A number of theoretical models from the social and behavioral sciences offer frameworks for characterizing the factors that facilitate or inhibit the spread of health care improvements. Diffusion of innovation theory, proposed by Everett Rogers in 1962, has proven to be a useful theoretical model for understanding how improvements are “diffused” in a health care system by describing how innovations are communicated and adopted by individuals within such a system, who can be thought of as a social group.
Rogers argued that individuals seldom adopt new ideas impulsively. Rather, acceptance of change typically involves passing through five phases: awareness (learning about the innovation), interest (seeking more information), evaluation (forming positive or negative attitudes about the innovation), trial (testing the innovation’s acceptability), and adoption (accepting the innovation).
Rogers observed that individuals vary in their response to innovation. He grouped people into five categories defined by the speed with which they adopt innovation. Listing the fastest adopters first and the slowest last, he ordered them as follows: (a) innovators, (b) early adopters, (c) early majority (relatively early acceptors), (d) late majority (relatively late acceptors who approach innovation with caution and do not adopt until a majority of others in their social group has done so), and (e) late adopters.
Rogers’ model suggests tactics and strategies that can enhance the design of spread. To be effective, spread approaches must incorporate strategies appropriate for each of the adopter categories, to enable them to move through the phases of adoption. He argued that impersonal information sources (such as media campaigns) are most important at the awareness stages and that personal sources (such as friends, relatives, colleagues) are most important at the evaluation phase. Change agents are also important for influencing individual decisions about the adoption of an innovation. Innovators and early adopters may serve as opinion leaders and play a critical role in encouraging others to adopt changes in practice. The early and late majority groups, in turn, are likely to be persuaded by peers and opinion leaders. Extra efforts, incentives, resources, and even regulations and sanctions may be needed to reach late adopters.
Approaches to Spread
Many different approaches may be used to spread improvements, either alone or in combination. Dissemination of guidelines, training, media campaigns, policy decrees, and endorsement by institutions or individuals of prestige are among the common methods that have traditionally been used in health care and other sectors to spread improvements. These approaches focus on one or several factors necessary to achieve spread: dissemination focuses on raising awareness among practitioners about the benefits of best practices; training focuses on developing technical competency; campaigns focus on building commitment and political will; and policy development and endorsement focus on increasing the perceived legitimacy of the improved intervention and alignment with accepted institutional values.
Spread using collaborative improvement approaches has recently shown impressive results in USAID-assisted countries. An improvement collaborative brings together multiple teams, often from interdependent health care facilities, for structured improvement work and shared learning around common aims and interventions. Because a collaborative involves many sites, it itself is a spread approach.
A “wave sequence spread” is another systematic approach to rapidly spread a proven change to a large, nested system in which the individual units are interconnected and arranged in a hierarchical order. The wave sequence approach builds on the improvement collaborative approach and emphasizes developing champions from within the system to lead the charge in furthering the subsequent spread. The full scale of intended spread is mapped first, and a slice of the system representing the different levels of care in each administrative division is selected to participate in the phase I (or “demonstration”) collaborative. Champions from the phase 1 collaborative are identified and equipped to conduct the phase II (or spread) collaborative in their respective subdivisions.
Experiences of the Institute for Healthcare Improvement and other leaders in large-scale change in health care offer some lessons. First, the starting point for any spread effort is identifying a superior model proven to offer better results. This model needs to be developed and packaged for adoption by others in the social system in which the activity occurs. The second key ingredient for successful spread is the commitment of the health system’s leadership to spreading the model. Such leadership includes both senior- or executive-level sponsorship and day-to-day leadership. Executive sponsorship is a crucial component to provide accountability and encouragement for spread. Day-to-day leadership for spread and effective channels of communication are needed to communicate the new model at the operational level. Champions for the new model or change are also important to provide social influence to support spread. Finally, the participation in the spread effort of teams who have already successfully implemented the new model provides both practical knowledge about how the new model can be implemented and voice to the fact that it is feasible.