EPITalk: Behind the Paper

Spatial Patterns in PrEP Accessibility and the Link between PrEP Access & SDOH

• Annals of Epidemiology

Share episode

Send us a text

Dr. Hui (Henry) Luan joins EPITalk host and co-author, Dr. Patrick Sullivan, for a jam-packed discussion about their recent study on PrEP accessibility measurement. “Spatial accessibility of pre-exposure prophylaxis (PrEP): different measure choices and the implications for detecting shortage areas and examining its association with social determinants of health” can be found in the October 2023 issue (Vol. 86) of Annals of Epidemiology.

Read the full article here:
https://www.sciencedirect.com/science/article/pii/S1047279723001412

Link to the R script for implementing the Gaussian 2SFCA measure for PrEP accessibility:
https://figshare.com/articles/journal_contribution/R_script_on_Gaussian_2SFCA_R/21778493/1

Some resources on GIS/spatial statistics in health can be found below:


Episode Credits:

  • Executive Producer: Sabrina Debas
  • Technical Producer: Paula Burrows
  • Annals of Epidemiology is published by Elsevier.



Patrick Sullivan:

Hello, you're listening to EPITalk: Behind the Paper, a monthly podcast from the Annals of Epidemiology. I'm Patrick Sullivan, Editor-in-Chief of the journal, and in this series we take you behind the scenes of some of the latest epidemiologic research featured in our journal. Today we're here with Dr. Henry Luan to discuss his article "Spatial Accessibility of pre-exposure prophylaxis (PrEP): different measure choices and the implications for detecting shortage areas and examining its association with social determinants of health." You can read the full article online in the October 2023 issue of the journal at www. annalsofepidemiology. org. I'll tell you just a bit about our guest today. Dr. Henry Luan is an assistant professor in spatial data science and spatial epidemiology in the Department of Geography at the University of Oregon. Dr. Luan applies and develops Bayesian spatial temporal statistical models and advanced GIS approaches to investigate how health phenomena, especially health events related to HIV, vary over space, time and race ethnicity, and how socioeconomic, demographic, physical and built environmental factors contribute to these space, time and cross-race ethnicity variations. Dr. Luan, thank you for joining us today.

Henry Luan:

Thank you for having me here. I'm excited to be here.

Patrick Sullivan:

And I'm excited to talk some about your study, which gets to geospatial epi, which I'm excited about, and which gets to pre-exposure prophylaxis, which I'm excited about. So tell me a little bit about the purpose of the study. What were you really aiming to answer with the analysis that you did?

Henry Luan:

Yeah for sure. So the major purpose of this study is to compare different spatial accessibility measures of PrEP providers and then examine how different choices of the measures affect the detection of PrEP provider shortage areas, as well as the association between PrEP accessibility and social determinants of health variables, as suggested by the title of the article. So by something I would like to clarify here is the difference between the accessibility measures for individuals and for areas or for a neighborhood, because they're different. So for an individual, we can derive their activity spaces via techniques such as GPS equipped devices and then derive how a PrEP provider is accessible to an individual. But for an area we cannot do that because we cannot track every individual residing in the same area and instead we can only calculate PrEP accessibility for the entire area or calculate the accessibility for a representative point of that area, for example the geometric centroid or the population weighty centroid of an area. So in this sense they are different.

Henry Luan:

I would like to say that before the widespread use of GPS equipped devices, so these area level accessibility measures are used as proxies of individual level accessibility. So to- in short, this article compares the accessibility measures for areas rather than for individuals and the main questions we aim to answer. There are two questions we want to answer. First is how does the choice of PrEP accessibility measures affect PrEP shortage area detection? And second is how does the choice of PrEP accessibility measures affect the findings regarding the association between PrEP accessibility and social determinants of health variables?

Patrick Sullivan:

That's really interesting. I want to break down a couple pieces of this, just because I think maybe we both think about prep and geospatial stuff. But for people who may not be so familiar this idea of PrEP accessibility in areas in which PrEP is accessible, why is it important to know that? If I'm in Portland and you use this method and you tell me where prep is more or less accessible, how does that translate to public health action? How would that data be used by providers, for example?

Henry Luan:

I think. First because we care about who have sufficient accessibility to PrEP providers, especially for those population who need them most. And if there is a mismatch between the supply and the demand, so we want to make sure those demands are made. And we know that these spatial accessibility is not constant over space. There are always some areas that are in short accessibility and some areas where they have good accessibility. So that's, I think, from this perspective, we need to visualize and to know where the accessibility is now and doesn't meet the need.

Patrick Sullivan:

And it's very actionable, right? Because if I'm responsible for HIV prevention in this city and I can look and say, oh, I know the areas of the city where new HIV diagnoses are occurring, and then you are able to use these kind of data to show where there may be low service areas, and if those two things line up like high risk and low service, that's the most impactful place, exactly. So you sort of mentioned the methods a little bit. So you must have inputs of where the providers are and then what are the methods that sort of lead to identifying these areas. How does that work?

Henry Luan:

Yeah, actually like a great question, because we compared four different categories of measures in our study.

Henry Luan:

So the density-based, the proximity-based, the more complicated two-step floating catchment area-based approach and finally, the most complex one, the Gaussian version of the two-step floating catchment area approach. So I would like to briefly introduce the principles behind these four categories of measures. First is the density-based measures, so this is the simplest one. So we use the count of PrEP providers within an area, so in our context, the zip code, the zip code compilation areas. So we use the count in that area over the total demand. So this demand, in our study we use two proxies One is the total population, the other is the total account of new HIV diagnosis and we also compared how these different proxies of population demand affect the identification of PrEP shortage areas. So this is the first one, first category. The second category is proximity, so we simply calculate the distance along the road network, so it's not like straight-line distance. We do account for the network, like in structure in reality, and then we calculate the distance from the zip code centroid to the nearest PrEP provider, or the average distance to the three or five closed PrEP providers, because we want to account for more than one provider, because usually they have been researched, that reviewed, that the residents probably do not necessarily go to the nearest PrEP provider. So that's the second type. And for the third type is more complicated, that's the two-step floating catchment area approach.

Henry Luan:

So, as the name suggests, there are two steps in the calculation of this measure. So first we create the catchment area for each PrEP provider, so that's from the supply side. So based on, for example, the travel distance along the road network, based on a specific cut-off value, for example 10 minutes driving, 15 minutes driving or 30 minutes driving, and then we calculate the supply-demand ratio. So for example, let's assume a prep provider can serve five zip codes when we serve, that is the centroid of the zip codes for inside that catchment area. And then, if we assume that each zip code has one resident only, and then the supply-demand ratio will be 0.2. That means not the full capacity will be used by each zip code.

Henry Luan:

There is a computation along the supply side but of course the actual population can be accounted for to calculate the supply-demand ratio. So that's the first step in this method. The second step is that for each zip code, so that is at the demand side, we identify which prep providers are accessible to that zip code. And then we add up the supply-demand ratios associated with different accessible prep providers so that the final results are the accessibility measures for each zip code. So you can see that this measure accounts for the computation from the demand side as well as the spatial variations in the demand population. So that's the third type of accessibility measures.

Henry Luan:

The last one, which is the most complicated one, the Gaussian-based two-step floating catchment area, goes one step further compared with the standard two-step floating catchment area approach. That is, we add Gaussian function to account for the distance decay effect. In other words, we account for the fact that a prep provider further away from the zip code's centroid will have lower accessibility. So this is not the case for the standard version of that method. So because as long as the prep providers are within the cutoff value in the standard method, they are assumed to be equally accessible, but that's not the case in the Gaussian version. So that's the four categories of the measure that we used and compared in our study.

Patrick Sullivan:

Thanks for that explanation. From a methodologic point of view, how did those measures compare, because the bottom line here is sort of trying to identify these PrEP shortage areas. So were they pretty much aligned, or was there one that you felt like ended up being more informative than the other?

Henry Luan:

I would say the most complicated one is more informative than the other ones because, as based on the introduction to the principles of these methods, it accounts for different factors into the calculation. But here I wouldn't conclude in terms of which measures can more accurately detect the PrEP shortage areas, because we don't know the actual ground. Choose which area has a higher accessibility, which doesn't. But in state I would compare them based on different factors, for example if they account for the variations in the demands, if they account for the computations among those demand populations and if they are insensitive to, for example, the spatial scales, which is also one factor we explored in our study. So because based on these comparison criteria, like the most complicated Gaussian-based two-step floating catchment area, is the most robust or most consistent. So that's why I say that this one is most robust measure to quantify accessibility to PrEP providers.

Patrick Sullivan:

Great. Was there anything that surprised you by the findings of this study, or were things sort of as you expected when you went in?

Henry Luan:

To be honest, I am not surprised by the performance of the most complicated Gaussian-based approach, but what surprises me is the simple proximity measures. So if I go over our article, we see that regression coefficients identified from this measure is consistent, is in the same direction, with the regression coefficients obtained by using the most complicated PrEP-accessibility measures. So my interpretation is that probably is because this measure also accounts for the distance decay effect. Again, further away that the zip code is from the PrEP provider, the lower accessibility. So then, like this, distance decay effect is accounted for in that measure, and so that's why I'm thinking that it is challenging to implement those most complicated measures. I would recommend to use the proximity measures as a starting point and, in contrast, I wouldn't recommend using the density measures, especially for those areas that are denated by very arbitrarily defined boundaries for administrative purposes.

Patrick Sullivan:

Yeah, because sometimes the responsible entity, the public health entity, may be responsible within those boundaries, but the clinic that might serve people is just across the board.

Patrick Sullivan:

I mean, it's not all about just the distance because of administrative roles and where the public health response is yeah. So thank you so much for that explanation and, on a really important issue, I want to pivot a little bit and talk about just your process and we call this Behind the Paper is sort of trying to understand, because we do work in these very technical areas, but we're also, you know, people who have colleagues and who have interests, and so I like to a little sense of just how your way of work and how this work comes about. So one of the things I noticed is that you and your co-authors really have diverse academic backgrounds and disciplines, and so I wonder whether you think like having that kind of multi-disciplinary team influences the work, or is it that people have been trained in different areas but also to gravitate towards a very similar interest in the end? So does it make a difference that there's this sort of diversity of backgrounds on the team?

Henry Luan:

Yeah, this is a great question. Yeah, I have to admit that it does make like difference and make impact on the final presentation or the findings of our study. So first here, like I would like to explain this from different perspectives. From the statistical modeling perspective, for example, like including the second author, Dr. Li from the UK, who is a statistician. I believe that makes our analysis statistical rigorous. In fact, I had some questions where I was fitting the marginalized two-part log-node models. It very sophisticated like a set of model. So because that's the first time I fit the model, I had some questions and that's why I reached out to Dr. Li, who I have known for a long time, and he answered my questions and make me more confident regarding the statistical rigor of our study. So that's from the statistical modeling perspective. And, on the other hand, other co-authors of this study, including Dr. Duncan, Dr. Ransome and, of course, Patrick, are experts in like social epidemiology, spatial epidemiology, and it's actually research. And you also significantly helped the presentation of the final article, including the writing style.

Henry Luan:

So how to make the article more accessible to the public, health professionals and researchers who might be interested in implementing the methods for their own data sets and of course also what kind of information was the public?

Henry Luan:

Health professionals and researchers might be interested in knowing, so we should present them in the article and so on. So one specific example I would like to mention here is one suggestion from Dr. Ransome, from Yale so, who suggested to include an abstract video to show how the complex Gaussian two-step for the catchment area approach is implemented in GIS and then posted, for example, via YouTube. So this could help, like folks who are not trained in GIS but interested in implementing the method but don't know where to start. So I really like the idea. But because the method was implemented in R rather than in GIS, so I decided to post the R script online rather than making a video. But even for that we didn't make a video. I really hope that sharing the R code that implemented the method could encourage and inspire the practical usability of our study. So I'm really hoping that these script sharing could encourage people to implement this method for their own data sets.

Patrick Sullivan:

Yeah, that's great, and we'll be sure to put the link to that R code in the show notes as well, so people are listening, that can find the code. That's great. I think just sharing your code is great all the way around. It's a little bit vulnerable because I feel like when I write code it's not always the most elegant, it's not always like some people write and it's like the absolute minimum number of characters and steps and mine. It more reflects how I'm thinking. So it's kind of clunky sometimes, but it's a generous thing to share your code.

Henry Luan:

Yeah, but I think that's totally fine, because we are not trained as programmers. As long as the function works, the job can be done in different ways, right?

Patrick Sullivan:

and I think that's another good point is, I mean, I love that you talk about the collaboration with your colleague who's like more on the statistical side, because to do this kind of work, where we're really trying to like take data and make knowledge that improves health, you really have to understand every piece of it and I'm not sure any of us by ourselves could understand you know, like, why people get PrEP where they do and why somebody might want to go to the closest place and why somebody else might want to go to a place that's not closest to their house.

Patrick Sullivan:

And and then, layering on all these methods, it really has to be done in teams, yeah for sure. So I just want to sort of note, in terms of your own career, that you've really I mean, we're talking about HIV today, but you've also been involved with other really important public health challenges, like food access, urban environmental health, crime, like how do you? So? You have a set of methods here, but is there a common theme for you about, like the areas that you've chosen to apply these methods to to try to improve public health? Because it does? I mean, I see them unifying, like geography, GIS, expertise, but what inspires you about some issues to tackle them?

Henry Luan:

Yeah, that's a great question. Yeah, so I was also- sometimes I was reflecting what kind of research I am, you know. So what is like a narrowed theme of my research.

Henry Luan:

So here, in short, I would say I'm a data-driven and system-modeling driven researcher in different health phenomena, because for different health data sets I address different issues, for example, like in our article. So when the density-based measures are used to quantify proper accessibility, we have the skewed data sets with a lot of zeros and we see that if these issues are not properly addressed, they could lead to biased results which could be have like negative impacts on like subsequent interventions. Yeah, in terms of where and to what population the HIV interventions should be, like allocated. Yeah, so in this case, like we, that's why we propose these complicated to mixture models, two-part mixture models, to address this issue.

Henry Luan:

So another example is that so it is common that in HIV and in other some other like similar sensitive data sets, if they are released at a small area levels, we have this data spatial issue because those values that is smaller than a specific like five below five, will be suppressed from releasing to the general public. So there are a lot of uncertainties for to analyze these incomplete data sets. So that's why, like in my article, like I propose to use, for example, bas approaches to address these data spatial issue. So to sum up, I would say again a data driven and a set score, like quantitative method driven, like researcher, who are interested in addressing different data analysis problems from different data sets. That's why, like my application, like span from like food, hiv, crime, urban environments, etc. But I see that they are all closely related to each other.

Patrick Sullivan:

So yeah, and then this element of space and services right, I mean you can see how that applies. For example, the food access, for sure, the crime you know is a place based. There's a place based concentration of crime, for example, of resources. So see how it's set up at this. But then there really are these like sort of access and health equity issues. Yeah, that arise out of place and I think that's an interesting theme.

Patrick Sullivan:

So do you have any advice? For I think that the map based and sort of place based research is so of such interest to people because it's accessible and because the pathway between what you identify with your analyses and how you make health better, to me that's a short jump. Like once you see that app of where the people who need the thing is and where the service locations are, the fix is sort of in the story, like in the map, right. So if people are interested in this but maybe a little intimidated by the methods, what advice would you have for students or early career professionals who say like I'm really interested in this kind of work but I'm not sure how to get started? How might somebody get an introduction to this?

Henry Luan:

Yeah, so my, I do have some advices, like because it's similar to where I started to, for B is the system only because I'm not a trend statistician, so that I see that there's like similarity between my experience and this to this question yeah, I do have some piece of advice is, first, to use those freely accessible resources in this field, in GIS, public health and their applications GIS and spatial analysis applications in public health, because today there are so many like free resources out there, including, for example, the textbooks, the journal articles, youtube videos, training programs, like at conferences or at different institutions, so start from there. So here I would like to provide some examples in terms of these like resources, for example, like for those training programs at conferences and institutions. As far as I know, conferences such as SER, the Society for Epidemiologic Research, and international conference on environmental epidemiology they are now having those pre conference workshops on GIS, special analysis and the applications in public health, and some institutions they also have these like summer workshops like on this field, from down from Columbia University and Drexel University and so on, and some textbooks. They are also a freely accessible on the website because of the widespread use of like R and Python programming, there are also a lot of sample codes, like in those textbooks, so which students, anyone can follow and to reproduce the results or replicate the study.

Henry Luan:

And for some journals they also have like freely accessible, like articles on this topic, for example, spatial and social epidemiology, international journal of health, geographics and even like Annals of Epidemiology, like they also have like articles on this subject. So start from there, like I read these articles and using some simple data sets from these resources. So, and I believe that and there is no doubt that with more practice, like the students and anyone who would like to start this topic will get to do more about this topic and get to be more comfortable and confident about this topic. And finally, I think another piece of advice is that do not be shy to reach out to the experts in this field. So my experience is that most scholars would be happy if you read their article, their work and reach out to them for questions. This is also my own experience. For example, whenever I have a question on sesquimolony, I reach out to those experts. That's how I got connected with the second author of this study in this paper. So that's my suggestions.

Patrick Sullivan:

Yeah, and what's amazing to me is when I do that, people always respond. I feel like in our world you can reach out to someone who you've never met before and say I'm working on this analysis and I saw you had worked on these methods and I have a few questions and people are incredibly responsive.

Henry Luan:

Exactly, yeah, that's why, yeah, don't be shy, just reach out, you'll get some responses, great.

Patrick Sullivan:

Well, this has been a great conversation, both about the substance of your article and some ideas for these sort of research collaborations and how people might get started. Do you have any last thoughts that you want to share with our podcast listeners of, either about this analysis or advice about doing these kinds of analyses?

Henry Luan:

Yeah, just do it like, go ahead. We're simple because of the R&R movement in spatial analysis, like that's reproducibility and replicability. So a lot of work, they have, shared the code they have and some even share the data set. If that's not, the data set is not that sensitive. So just go ahead and do some practice and they will be like at least like be more familiar with this topic. And yeah, that's my last piece of suggestion.

Patrick Sullivan:

Yeah, and then once you have that level of comfort, you can start keeping some pieces of the code but swapping in a different kind of service type, that's a great way to start, I think the idea of just jumping in and seeing how the code works, yeah, and then and a lot of what makes research interesting is bringing good questions.

Patrick Sullivan:

Exactly, the methods are some methods, but people who listen, who may pick up. You know, look at some of this code or play around with it, then we'll bring that to another important question that they understand, that I wouldn't know about. You know, and that's that's how we propagate this information.

Henry Luan:

Yeah, and something like I would like to say that those interdisciplinary collaboration and discussions are also important, because that's also the way. From these like brainstorming, talking with, like experts in different fields, then we can think of some innovative solutions to a specific research question, and sometimes even the research questions arise from these interdisciplinary discussions, yeah. So yeah, go out to talk to people. Yeah, different topics, yeah.

Patrick Sullivan:

I do. This does just make me think that, like so much of my own, like professional history was around going to certain conferences mostly HIV and public health conferences for me and meeting people like I had read, you know, and whose names I knew, and introducing myself at posters, presentations, and, and I think, in some ways, that the pandemic has made some of the conferences and things more accessible because you can do them online and you're not traveling around. But there is that piece of just like making connections with people, yeah, conferences and building your network, that I think that this, you know, the generations of researchers who are really coming into this, yes, are going to have to develop different tools in ways, because for me it was very much driven by and that's why I went to the conferences, yes, to hear the talks, but it was mostly to meet the people giving the talks.

Henry Luan:

Yes, I can't agree more, to be honest. Yeah, and I feel like only through those in person conversations we would be able to know if I can get along with this like person to do, to collaborate with him or her, you know. So, yeah, to me this is like personal connections always matters more than those on like virtual conversations.

Patrick Sullivan:

Right. Although in full disclosure we're recording this today over a zoom call, so it's hopefully this sounds like we're sitting in a room having a conversation. Yeah, ups and downsides of technology.

Henry Luan:

Exactly yeah.

Patrick Sullivan:

Well, that brings us to the end of the episode. Thank you again, Dr. Luan, for joining us today. It was so nice to talk to you just about your thoughts about this field, and obviously I love and I'm interested in the application that you've used to bring about this knowledge and, you know, I hope we have opportunities to work together again. A lot of shared interests, so, but thanks for sharing this with us.

Henry Luan:

Yeah, thanks for having me. Yeah, I'm glad I, yeah, I have this opportunity to share this work, introduce the principles behind those methods and hopefully this like episode would inspire more researchers to adopt this method in their own work. And don't hesitate to reach out to me if you need any suggestions or like need, like advice is all implementing those methods.

Patrick Sullivan:

So two things. One, I'd encourage people to check the show notes. We're going to put the link to the video that you mentioned and some of the other resources, and I'm going to follow up with you in a month or two and just see how many people have reached out. I think I have a goal of five people who listened to this and reach out and say help, get me started

Henry Luan:

So, yeah, great, let's see how that goes.

Patrick Sullivan:

All right, I'm your host, Patrick Sullivan. Thanks again for tuning into this episode and see you next time on EPITalk: Behind the Paper brought to you by Annals of Epidemiology, the official journal of the American College of Epidemiology.

People on this episode

Head Shot

Patrick S. Sullivan, DVM, PhD

Host