Skip to main content
Back to Top


Reflections on the Preventing Overdiagnosis Conference

Content on this page is provided for reference purposes only. It is no longer maintained and may now be outdated.

In September, I was fortunate enough to attend the 4th international conference on Preventing Overdiagnosis in Barcelona, Spain. This was my first time attending the Preventing Overdiagnosis Conference (POD), which has been held at Dartmouth (2013), University of Oxford (2014), and NIH (2015) in the past. This year, the conference was hosted by the Agency for Health Quality and Assessment (AQuAS) of the Ministry of Health of the government of Catalonia as well as many associate partners. 

Briefly, the conference is focused on bringing together researchers, practitioners, policymakers, patients, consumer/citizen advocates, journalists, and other important stakeholder groups to share ideas, report on research findings, discuss ways to advance the field, and debate ways to advance the field and prevent overdiagnosis. It many ways, the overall theme and climate of the conference reminded me of the annual D&I Conference in Health; it was really nice to see a diverse group of stakeholder representatives and important themes that cut across health topics, delivery settings, and content areas. The conference also had a nice mix of presentation formats, including 90-minute workshops (e.g., 2-3 presentations followed by discussion with panelists and attendees) and 3-minute elevator pitches as well as the more traditional formats of panel presentations, plenaries, and posters.     

In collaboration with colleagues, I organized a 90-minute workshop during the conference as well as a 3-hour interactive session during Research Day, held immediately following the conference. Both sessions were focused on advancing the science of de-implementation of ineffective, harmful, inappropriate, and/or unproven health practices, programs, interventions, guidelines, and treatments. The sessions focused mostly on ways to advance the science around reducing or stopping the use of such programs although ways to prevent their introduction and adoption in systems and care settings were also discussed (e.g., replication and reproducibility studies, higher threshold for warranting introduction of practices, etc.).

The conference workshop included three brief presentations followed by a 45-minute open discussion with attendees. Dr. Russ Harris (UNC) led the session discussing potential patient- and provider-level strategies for de-implementation, followed by Dr. Barry Kramer (NCI) on policy-level strategies for de-implementation, and concluded with an overview presentation on de-implementation by myself. Attendees provided great comments, suggestions, examples, and questions about our specific presentations as well as the overall topic of de-implementation. We discussed the importance of studying de-implementation as an important next step in preventing and curbing overscreening, overdiagnosis, and overtreatment and some of the barriers to actual de-implementation.

In collaboration with Drs. Barry Kramer (NCI) and Karsten Juhl Jorgensen (Nordic Cochrane Center), the Research Day session included two presentations, two breakout working groups, summaries, and identification of next steps. Dr. Jorgensen discussed the process for developing 10 criteria for de-implementation and I presented challenges associated with advancing the science of de-implementation (e.g., study designs, measures, outcomes). Attendees split into two groups to further discuss each topic area, both of which were stimulating, informative, and constructive. Each group identified next steps for advancing each separate but related topic and established ways to continue ongoing conversations and discussions within and between the groups.

Overall, it was a wonderful experience and I really appreciate my colleagues’ collaboration on the presentations (thank you Barry, Russ and Karsten!) as well as the incredible feedback from attendees. I look forward to continuing to discuss this topic and work with everyone to try to move the needle forward in terms of advancing research in this area.

Note that the next POD conference will be held in Quebec (August 2017)—details available here.

Congratulations to my NCI colleagues who also presented at the POD conference, including (but not limited) to those particularly relevant to de-implementation and implementation science (i.e., workshops, posters, panel presentations):

·         De-implementation: Individualized cancer screening decisions as a way to achieve optimal care for older adults (Dr. Erica Breslau);

·         The US PROSPR Research Network: A platform for studying overdiagnosis in the community setting (Dr. Paul Doria-Rose);

·         Cancer overdiagnosis: Discussing our disagreements (Dr. Pamela Marcus).





I have attended the

I have attended the international Preventing Overdiagnosis conference since its inception four years ago.  Working primarily in the realm of cancer research, it was an eye-opener at my first meting to see that the problem of overdiagnosis in not restricted to cancer, but pervades many areas of clinical medicine and public health.  Some of the drivers and potential solutions are common across specialties and disease types, but some are specific to a given area.  A value of the conference is the discussions about both the commonalities and differences.  Even more importantly, the conferences have been moving the field from simple description of the prevalence and impact of overdiagnosis in various areas to actual intervention to mitigate or avoid the problems.

The strategies will not be simple.  The very concept of overdiagnosis is highly counterintuitive to the public and healthcare professionals, especially in the realm of cancer screening and diagnosis.  Strategies include de-implementation of entrenched belief systems and practice.  They will cut across areas of expertise and involve "medical reversal" interventions on multiple levels: patient, health professional, policy, social.  That is why I found it so useful to plan the workshop and interactive session outlined in Dr. Norton's note.  We were able to combine our individual areas of expertise, particularly in her area of research--de-implementation.  The field of de-implemention to decrease overdiagnosis has plenty of room for more research.  That's why I look forward to next year's meeting in Quebec City.

Hi Wynne—Fantastic post!

Hi Wynne—Fantastic post! Research is sorely needed to illuminate successful deimplementation strategies in cancer care and other areas. I agree with Barry Kramer’s post that overtreatment and overdiagnosis are difficult public health problems that many may be reluctant to engage with. These issues are especially acute in cancer care, where the understandable panic of a diagnosis and anxiety about recurrence may be associated with “more is better” thinking, contributing to overdiagnosis and overtreatment. Clinicians may use non-recommended health care services (e.g., serum tumor markers for early stage breast cancer surveillance) to address issues like patient anxiety, while recognizing that these tests are not clinically useful to detect cancer recurrence.  Thus use of this test addresses a perceived need (managing patient anxiety); however, there are other approaches that may better serve these patients including distress screening and referral or survivorship clinics with symptom education classes.

In addition, how can the cancer community best learn from other areas of research that have been trying to address issues of overtreatment and overdiagnosis? Antibiotic stewardship efforts are a good example where many approaches have been tested, with some approaches resulting in reduction of overutilization. The conference sounds like an incredible opportunity to bring together researchers and clinicians from diverse areas of medicine and public health to collaborate and learn about these cross-cutting issues. I hope to attend in the future. 

Hi Wynne, thanks for raising

Hi Wynne, thanks for raising awareness about this meeting and the work going on regarding over diagnosis. I think one of the interesting intersections at the provider and patient levels is in research on different types of cognition and cognitive biases in the diagnosing process and overuse. Geoffrey Norman and colleagues have a great piece just out in Academic Medicine on this, addressing how different types of cognition are associated with different types of bias in making diagnoses depending on physician experience and the patients' conditions I suspect that their findings reflect a fundamental challenge about how much we can expect physicians to develop and maintain accurate diagnostic skills across a range of conditions.
On the patient level, we now know that people systematically exhibit what we would term biases (in comparison with the purely rational, utility-maximizing agent of classical economic models), including responding differently to equivalent outcomes depending on whether framed negatively or positively (e.g., patients will choose an option with a 10% chance of mortality less often than one with a 90% change of survival). I think one of the interesting questions to tackle is how much, if at all, these types of biases play a role in over diagnosis and over treatment, and to what extent these "biases" reflect what all of us would choose for ourselves. For example, people hate uncertainty; are there situations where a test with no utility for informing a treatment plan is still valued by patients because it helps the clinician and by extension the patient understand what's going on?
Finally, this is such an interesting area of research because it's such a minefield. I can remember the days of the Health Maintenance Organization (HMO) in the '90s when primary care doctors were asked to serve as "gatekeepers" and HMO administrators required physicians to obtain prior authorization for some referrals or treatments, in some cases to improve quality and in some cases to reduce costs. Patients and clinicians alike came to revile managed care, and they lumped the good in with the bad. Over diagnosis and over treatment are serious issues and perhaps represent one of the few opportunities could be true win-win-wins (more patient-centered, better outcomes, less costly). But we have to tread lightly, listen and make sure we understand before acting. There's so much we don't know, and the stakes are high. And for a researcher, that makes for heady work.
I'm looking forward to continuing the conversation with you and all of our colleagues around the world working on this issue.