WIC meetings seek to develop a community of researchers, healthcare professionals, physicians, and policymakers. We welcome your contributions on this year’s 5 themes and a high priority topic; we also welcome papers outside these topics on evaluating health care and unwarranted variations. Papers will be presented orally or by poster. Additional sessions will be developed to appropriately group meritorious papers.
Please complete the abstract form by May 31, 2026 and mail to wennberg.collaborative@dartmouth.edu. The WIC will screen all submissions and will send relevant abstracts to be considered by the organizers of the different themes and priority topics. The format of the conference and of each session will depend on responses to this call for papers.
Download the Abstract submission form here.
Major themes
1) Low value care and overuse / Eva Stensland
Low-value care has no added value for the patient and may be harmful. To increase quality of care, healthcare systems aim to reduce low-value care, through de-implementation strategies. De-implementation also supports healthcare systems sustainability reducing wasteful spending and carbon emissions. De-implementation is a challenge because of the many causes of low-value care with responsibility at different levels: patients, professional, organization, and system. In some de-implementation initiatives, healthcare may be replaced by high value care, but other times there is no substitute. Behavioral change is often essential for both healthcare providers and patients. De-implementation requires a tailored strategy based on an inventory of barriers and facilitators. Contributions on this topic include:
- How can we identify low-value care and overuse? What role can practice variation analysis play?
- Why do we still provide low-value care?
- What strategies are effective to reduce low-value care? And which are sustainable?
- What are the financial consequences of reducing low-value care?
- How can we spread and sustain effective strategies?
2) Drawing causal inference from observational data / Jostein Grytten
The gold standard for cause-effect studies is randomized controlled experiments. For several reasons (ethical, financial and political), such experiments are difficult to carry out in studies of practice variation and can lack generalizability to the general population. Hence many studies rely on observational data (i.e., data in which exposure is determined in a different way than random assignment) but it can be difficult to establish causal inference with the most commonly used study designs. Causal inference is required to answer questions such as: What is the impact of hospital payment on practice variation? What is the effect of a specific intervention (such as care protocols) in reducing variation in medical practice? To what extent does an increase in the number of beds increase variation? Econometricians have led the development of methods to examine causality using observational data. Examples are difference-in-difference, regression discontinuity or instrumental variables methods. These methods can be challenging to use.
The aim of the session is to discuss the potential benefits and limitations of these methods, their application in specific research. We welcome all types of work, both in progress and nearly completed. If you have a draft of your work, we may be able to get somebody to give prepared comments.
3) Novel data and data linkage methods for health care evaluation / Therese Stukel
Population-based health care measurement across regions and providers commonly begins with the analysis of health administrative data that has been collected for billing purposes, but the scope and scale of these data differs across regions even within the same country. Some countries have access to rich population-based data that is linkable across health care sectors and time as well as across different domains such as health, education and justice. In many places, researchers and policy analysts have access to other rich data sources that can be linked to patients such as clinical registries, patient and provider surveys, and electronic medical data (EMR). In this session, we will explore the use of novel data sources and data linkage methods to enhance our understanding of the causes and consequences of unwarranted variation.
4) Redesigning systems to reduce unwarranted variation: the role of incentives and choice architecture / Gwyn Bevan
There is so much truth in Paul Batalden’s haunting observation ‘Every system is perfectly designed to get the results it gets’. Hence the paradox: despite fundamental differences between countries in their systems of organizing the finance and delivery of health care, one troubling outcome that these different systems have in common is that they are perfectly designed to produce unwarranted variations. This theme has two objectives.
- First, to identify the causes of those outcomes from systemic characteristics of incentives (of providers of care) and the default options in the available choice architecture.
For example, in England the default option for a general practitioner who wearies of having to see the same patient every week with low back pain is to make a referral for an MRI scan. This is costly and of no benefit to the patient but means the general practitioner won't have to see the patient for weeks, and patient is happier because something is being done.
- Second, working out the obstacles to systemic change, and where it is possible to overcome these by changing the incentives and choice architectures to reduce unwarranted variations.
For example, for low back pain develop online help and local exercise classes.
5) Artificial Intelligence and health care performance evaluation / Philipp Storz-Pfennig
The belief that artificial intelligence-based products and services will be “transformative” for health care is widely voiced – following the availability of large (language) models showing seemingly impressive performance in different medical and healthcare related tasks, well beyond assisting in image interpretation (which was the dominant usage until quite recently).
The use of certain AI tools (e.g. in clinical decision making) might, in theory, contribute to more “rational” - at least in the sense of more uniform, less variable – decisions and may increase efficiency by substituting scarcity in qualified health professionals (time). Contrary to this, there is also concern that decisions made using AI will be biased and prejudiced, may de-qualify health professionals and “hallucinate” (potentially dangerous) treatment needs. As AI use is widely expected to increase tremendously in the near future, it may have a profound impact on service provision – and may itself be a case of supplier driven care. There seems to be a lack of data on the extent of AI use and its consequences already taking place. The following questions might be addressed:
- What is already known about actual AI adoption and uptake, it’s heterogeneity and variation (e.g. regarding health care fields, regions, types of AI technology)? How may data sources on the use of such technologies be developed?
- How could the (variations of) use of certain AI-products/services predominantly be characterized: Supply sensitive, preference based, evidence based – do these categories fit?
- What changes in service provision are observed or can be foreseen? Will AI use constitute or contribute to overdiagnosis, over- or misuse of treatments – or will it’s “rationality” help to preclude this?
Due to the dynamic and evolving nature of the field, conceptual papers and planned research on the topic will also be welcome.
High priority topic
Graduate student and early investigator training in investigation of population-based health care performance – curricular frameworks and syllabus examples / David Goodman
Despite the high interest in evaluating health care and identifying the causes and consequences of unwarranted variation, very few university graduate programs have related courses or curricular content. This is a problem in several respects. Aspiring and early investigators are not exposed to these ideas and methods, and this deprives the field of novel ideas and energy. Established investigators often have no local source of expertise in the systematic evaluation of health care. There are no textbooks which could serve as starting points for syllabus development at universities and as handbooks for scholars in related disciplines. The current informal transmission of theories and methods from WIC participants to the next generation often falls short in advancing the awareness of what is known and what areas are still developing.
This session will present recent efforts to address this training gap, including source materials and collections that could be useful in developing new courses. At this session we strongly encourage members of the WIC community to share any courses or training programs that advance the theories, methods, and knowledge related to our field.