Geographic Variation in the Use of Doublet Therapy for Metastatic Prostate Cancer - Samuel Washington

December 19, 2022

Samuel Washington shares insights from a study analyzing treatment practices for metastatic prostate cancer across the US. Utilizing a broad dataset provided by ConcertAI, a data warehouse company focusing on AI and big data, the team investigated patterns of doublet therapy administration among patients diagnosed with M1 disease. Dr. Washington and his team identified over 1,700 men across the country, aiming to understand what type of treatment men were getting and the variables influencing it. A significant geographical variation was discovered in doublet therapy use, with factors such as patient location, state population, prior treatment, and urologist and radiation oncologist density playing a part. The study, however, did not reveal causation, underscoring a need for more granular and targeted research to better understand these variations.


Samuel L. Washington III, MD, MAS, Assistant Professor of Urology, Goldberg-Benioff Endowed Professorship in Cancer Biology, University of California, San Francisco

Alicia Morgans, MD, MPH, Genitourinary Medical Oncologist, Medical Director of Survivorship Program at Dana-Farber Cancer Institute, Boston, Massachusetts

Read the Full Video Transcript

Alicia Morgans: Hi. I'm so excited to be here with Dr. Samuel Washington, who is an assistant professor of urology at the University of California San Francisco. Thank you so much for being here with me today.

Samuel Washington: Thanks for having me. Happy to chat about this project.

Alicia Morgans: Great. Thank you so much. So I think this is so interesting, this project that you and your team have performed where we're really looking at the utilization of doublet therapy for metastatic prostate cancer. Can you tell me a little bit about the population and the dataset you used to investigate this?

Samuel Washington: Yeah, so this was an exciting project that came from a data research award from ConcertAI, data warehouse company, primarily focused on artificial intelligence and using large data sets to further our understanding. This provided a data set of medical oncology practices from across the country. We're able to identify patients who had M1 disease at diagnosis and start to cluster and look at which medications they're getting within roughly 90 days of that diagnosis. The great thing about this dataset was that it was not limited to one insurer or one healthcare system. So it really provided us a kind of novel and unique look at these practice patterns that was different than the regularly accessible data sets that we have currently available.

Alicia Morgans: And can you tell me a little bit about the geography that this data set covers? Kaiser is really just as an example, California. Other data sets are really looking at certain portions of the country. What kind of geographic diversity did you have within this data set?

Samuel Washington: Yeah, it was actually quite broad because we weren't limited to one say, SEER Registry for example, or United Healthcare Insurance providers or healthcare system like Kaiser. It was really multiple settings or multiple clinical settings from across the country. And these different healthcare systems would pay into ConcertAI to kind of curate and clean that data.

Alicia Morgans: Really exciting. So tell me, what did you investigate and what did you find?

Samuel Washington: So as part of this award, I was really trying to understand how who you are and where you get your care, impacts the type of care that you're going to get. So we're looking at men with M1 disease, a diagnosis. We identified over 1,700 of those men across the country and we are trying to focus and really understand what type of treatment are men getting, are they getting doublet therapy, particularly in the context of us talking about triplet therapy now. What are we seeing? We knew that this dataset didn't provide all of the granular geographic information, so what else is happening around this area that they're getting care? So we merged data from the area health resource file. So it gave us granular kind of state and county level information that we could combine with what we knew from ConcertAI granular clinical information to get a more comprehensive picture.

Alicia Morgans: And just to confirm, when you say doublet, do you mean AR targeting agents or do you mean ADT plus potentially docetaxel? What is a doublet in this study?

Samuel Washington: Yeah, we actually kept this fairly broad just to get a general understanding of what people are doing across the board. So we combined doublet in this sense was very general, very generous, androgen deprivation therapy plus some other agent. And from that we get a very optimistic view on what's happening across country.

Alicia Morgans: Very good. And so what is happening? Tell me.

Samuel Washington: Yeah, so what we saw was that overall we saw a pretty significant geographic variation by kind of the four census regions with the majority of patients living in the southern area of the US census region. Roughly a third of the cohort received ADT by itself. So no other me medications combined with it. What we did was further drive home. Okay, what factors at a geographic level are associated with doublet therapy? Narrowing our cohort a little bit to about 1600 patients. In this cohort, about 85% were receiving doublet therapy. So those for whom we had complete information. And we found that based on geographic region we saw differences in care. So region was associated with differences in treatment, state population, prior treatment if you had surgery or radiation, as well as interesting specialist density factors. So greater urologist density and radiation oncologist density, were associated with differences in doublet therapy use.

Alicia Morgans: So can you tell me a little bit about which direction that went? So when you say state population or density of urologists or radiation oncologists, how did it actually affect it?

Samuel Washington: Yeah, so we saw that when we in... Each state, we were able to get kind of a total count of the number of urologists or radiation oncologists, that density for that geographic region. When we did this, we saw that increasing urologist density was associated with lower odds of doublet therapy. Conversely, we saw increasing radiation oncologist density was associated with greater odds of doublet therapy compared to ADT alone.

Alicia Morgans: So that's really interesting and I think is a little bit counter to what I would necessarily think. Obviously you don't have causation or really a reason behind that. It's an association, but really interesting, especially I'm sure when you present to a urology or urologic oncology community. What are your thoughts around this particular finding?

Samuel Washington: Yeah, so I think two main thoughts that I have about this. One, it drives home kind of the black spot in our thinking, how do we practice? And I think more studies should focus on that because that's in our actionable domain of things that we can manage and change. Two, I would say with this dataset, as with any other, we can't drill down to the individual. So again, we don't know causation, we don't know what was happening on the ground, but this does give us targets or areas that we can hone in further with mixed methodologies, more granular data to understand the why. But this is giving us an area to hone in on, which is I think an important part when we look at practice variations or disparities, we need more targeted focused research.

Alicia Morgans: That is a great conclusion and really I think to your point, raises questions that we can investigate further because these are opportunities for us, I think as a field. Now the other thing that you mentioned when you talked about association was prior therapy and type of prior therapy also seemed to be associated perhaps with the use of these couplet therapies or doublet approaches. How did prior surgery and how did prior radiation affect the use of these therapies?

Samuel Washington: Yeah, so for this, we noticed that if they had surgery or radiation, we kept those two together just from a number standpoint, increase the odds of having doublet therapy, which makes sense to us in a clinical setting. These are people who are plugged in already, surveilled by oncology teams already. So it makes sense that they are kind of in a better situation to get doublet therapy compared to monotherapy alone, potentially.

Alicia Morgans: Very interesting. So any other take home points that you want to characterize related to your findings? And then if you wouldn't mind just kind of acknowledging some of the limitations, some of which you've already sort of gone through.

Samuel Washington: Yeah, so I think the main thing here is just this provides a great opportunity with kind of a novel data set to drill down more of that where you live and how it impacts care than I think we've been able to get at before. Obviously this does have limitations. We're unable to drill down more to smaller geographic regions, for example, because of numbers. Causation inherent to the dataset, we're unable to address that. I think there are obviously other variables that we could start to capture prospectively. So we can drill into these questions further. But I think it's all wide open, which is very exciting.

Alicia Morgans: Well, I really appreciate this and just want to thank you and your team for taking such an approach and really trying to answer this question. I think that in general, in the field, we talk a lot about under utilization, but we don't really know why. And to your point, this is helping us understand rather than just sort of pointing a finger blindly in this direction or that where some areas may be that could act as levers for change. And I really think that this is important work and sincerely appreciate your efforts. Thank you so much for your time today.

Samuel Washington: Thank you for the opportunity.