Transitioning from Push to Pull: Emerging Trends in Implementation Support and Technical Assistance
By Dr. Julia E. Moore, Executive Director, and Dr. Sobia Khan, Director of Implementation
4-min read
Emerging research and competencies on implementation support and technical assistance
It is an exciting time for implementation support practitioners—the number of us continues to grow, and there are so many new developments and directions in the evidence surrounding this very important role.
For example, over 1,000 registrants attended our core competencies event, which tells us that many people are starting to identify with the role of an implementation support practitioner. On the evidence front, this month’s featured resource highlights new research and examples of technical assistance to help further people’s implementation support practice.
A very important trend we have noticed emerging in both practice and research is a transition away from “push” models to “pull” models of supporting implementation. While the notion of push and pull has been around since the early days of implementation science, what we saw reflected in the field were mostly push models. This was reflected in implementation models and frameworks, and just generally the way we talked about implementation where organizations, researchers, or governments were trying to push out specific evidence-based THINGs (e.g., programs, policies, or practices) to people who were positioned as passive recipients. As we venture towards more end-user and equity-centered approaches to implementation, pull models are getting more attention.
Push models of implementation support
Push models focus on disseminating or implementing the THING from the source (e.g., researchers, policymakers) to the audience (e.g., practitioners, organizations). This is a top-down approach emphasizing delivering evidence-based practices to the intended users, often assuming that outside groups, like researchers and policymakers, know what needs to be implemented. In fact, in push models of implementation, the decision to implement a specific THING has already happened without the input of the people affected by the THING.
We have seen many examples where research teams have an evidence-based program that they are actively trying to spread and scale to new sites and settings. There are also many examples of local, state/provincial, or federal/national governments trying to push out policies and programs across their jurisdictions. Additionally, it has become commonplace to have purveyor organizations support the implementation of specific programs, practices, and policies; therefore, they have a vested interest and specialization in pushing specific THINGs rather than supporting the implementation of other THINGs identified as needs by people in the setting.
A huge pain point that many of the people we have interacted with above is that people in the setting aren’t always motivated or receptive to implementing the THING; we often discuss with them that is that what has effectively been stripped from the people affected by the THING is their ownership of it and agency to implement it. Imagine, for example, you were told that you had to organize your own home in a specific way. Even if that method of home organization were a great idea, you would likely push back because someone came along and told you to do it. But if someone instead asked you what you need and offered you ways that they could support you to organize your home, and you had the agency to decide what best suited your needs, you would be more receptive to this change. Perhaps push models were never developed to be overly prescriptive, but people’s experiences of them now tend to feel like someone is trying to get everyone to organize their homes in the same way, whether they want it or not.
Pull models of implementation support
Meanwhile, pull models are driven by the demand and engagement from the audience looking to implement something. In these models, practitioners or organizations actively seek out information, resources, or interventions to address their specific needs or challenges. The focus is on creating environments and systems that facilitate the search, access, and use of relevant THINGs (e.g., evidence-based programs, practices, and policies).
As the world of implementation science and implementation practice increasingly focuses on community-driven, equity-focused, and/or patient-oriented implementation, the concept of pull models increasingly makes more sense in the implementation space. Additionally, as there are more and more evidence-based THINGs to choose from, it is easier to have pull models, where communities, organizations, or groups identify the challenges they are trying to address and then select a thing that will help them address that challenge. In this case, ownership and agency are back in the hands of the people affected by change.
In pull models, implementation supports look very different because instead of being set up to support people to implement one specific THING, implementation support practitioners need to be equipped to support teams, organizations, and communities through the process of selecting what it is they’re going to implement and then planning and enacting implementation. This requires a more generalized implementation support skill set that doesn’t focus on the ins and outs of implementing a specific THING, but instead focuses on the process of implementing at large.
Advantages of pull models
We have been very happy to see this transition towards pull models. This approach facilitates equity-focused implementation, where those who are implementing and affected by implementation drive the conversation and directions, rather than outsiders deciding what needs to happen.
As there have been more and more questions about fit and adaptations, we see pull models as a much easier way to address fit. In this case, an organization might review a few different options and pick the one that is the best fit for their setting and system. This has huge advantages over models where a specific program is pushed into a site, when it isn’t a great fit, but is the THING being implemented.
Finally, from an implementation support perspective, we believe there are huge advantages of supporting multiple sites to implement different THINGs. We are definitely biased by our own topic-agnostic approach to implementation. Still, we have found that the act of supporting people across settings and topic areas is what has been most helpful in teasing apart what aspects of implementation support are program- or setting-specific and what elements are good for general implementation support.
This article was featured in our monthly Implementation in Action bulletin! Want to receive our next issue? Subscribe here.