PANDAs in China: Prototype fine particulate matter monitors to study air pollution and obesity

Recently, PhD student, David Holstius and I organized a sensor build, inviting friends from within the EHS program and collaborators from EECS program to make prototypes of a particulate matter air pollution sensor.

The result of this was the creation of 15 monitors that have been shipped to China, where they are being used by Hilary Ong, a UCSF Pediatrics student who I am mentoring to conduct a study of pediatric obesity and its relationship to air pollution exposure.  The study is being conducting in Kunming, where my other PhD student Jenna is doing her dissertation work on food environments.

We’re calling these prototypes PANDAs, befitting their first use in China.  PANDA is an acronym for something I don’t remember.  David has also created a companion software package called BAMBOO that processes data from the PANDAs.  BAMBOO is probably also an acronym for something that I can’t remember.

Stay tuned for the results from our PANDA study.

 

A Proposal: An Environmental Health Community Collaborative at the new Richmond UC Campus

Recently, I was provided the opportunity by my colleagues in EHS to think about how space could be used at the new Richmond UC campus.  Here’s my proposal…

The Environmental Health Community Collaborative is a space for environmental health faculty, students, researchers, and community partners to gather to teach and learn, share, and innovate collaboratively to solve environmental health problems.

The Community Collaborative leverages newly available, and thus flexible space at the former Richmond Field Station, to rethink what an innovative, community-serving, and world-class research-promoting facility could be.

The hallmark of the Community Collaborative is a large reconfigurable Workshop that allows for:

  • Traditional lectures (150-person capacity)
  • Distance learning
  • Community meetings
  • Demonstrations of research innovations

The Workshop opens on one side to The Experiment – a large outdoor area adjacent to the water that allows for transfer of work conducted in the Workshop immediately into the environment.  This space allows for partners to conduct environmental health science in a semi-controlled area with access to all environmental media: soil, water, and air.  Like the Workshop, the space can be reconfigured to fit different research projects or community activities.

On the other side of the Workshop is The Laboratory – a combination of wet and dry labs.  This space allows for the invention of new environmental sensors, biological assays, and processing of samples from various research studies.  The walls of the labs will be in glass to allow the public to see the innovation that is occurring at UC Berkeley.

Adjacent to the Laboratory is The Office – a combination of faculty and researcher offices, open area reconfigurable workspaces for students to interact, learn, and problem-solve, and meeting rooms for brainstorming and private video/tele-conferencing with people from all around the world that regularly collaborate with Environmental Health Sciences faculty.

The Community Collaborative as a whole is a Research Site.  Both indoor and outdoor environmental health research can be conducted at the Richmond site – the kind of research that would be impossible to conduct on the Berkeley main campus due to space constraints, and the lack of natural resources (e.g., the shoreline).  The building and its outdoor spaces can be instrumented so as to collect data on indoor and outdoor environmental quality so as to be a source of rich environmental health research data.  Realize that every proposal to NIH requires a statement about the adequacy of research facilities and space to conduct the proposed research.  The Community Collaborative would provide many advantages over other schools in competing for extramural grant funding.

The Community Collaborative is Revenue Generating. Innovative space like the Workshop and Experiment areas are valuable, and can be rented out to other departments and outside entities for events.  Additionally, an on-site café that serves as public meeting space and encourages productivity can also generate revenue.  The Laboratory is valuable space for conducting fee-for-service laboratory work on contract for Industry partners.  The greatest potential for revenue generation is the combination of Workspace, Experiment, and Laboratory, which are geared towards the innovation of new technologies and approaches to Environmental Health Sciences, which can funnel into patentable intellectual property for the UC.

The Community Collaborative is Environmentally Smart.  The initial design and ongoing reconfigurability of both indoor and outdoor spaces can foster continual thinking for how environmental spaces can be managed to optimized health.   This can move us beyond existing LEED energy initiatives.

A mashup of different photos of innovative spaces, some at Berkeley, Stanford, etc. showing combinations of indoor/outdoor spaces for problem-solving, collaboration, and celebration. Spaces for practical learning and teamwork. New and traditional wet/dry laboratory spaces.  Publically-viewable lab spaces like at the Darwin Center at the London Natural History Museum.

 

Agent-Based Model of Obesity

Dunk the Junk Anti-Soda Mural in Richmond, CA

With funding from NIH NIDDK, we have developed an agent-based model of obesity that considers changes in individual’s diet and physical activity over time.  Using data from the National Heart, Lung, and Blood Institute Growth and Health Study (NGHS) cohort, we have applied both novel causal inference modeling approaches, as well as agent-based modeling.  Along the way, this study has quantified and mapped food environment exposures for the girls in the study, and has assessed the associations between changes in food environment and diet and obesity risk.

The National Heart, Lung, and Blood Institute Growth and Health Study (NGHS)

The NGHS was a 10-year prospective cohort study of African American and White girls initially aged 9 and 10 years old recruited from three main sites in the United States beginning in 1987. The study aimed to collect longitudinal data to better understand both physiological and behavioral factors that may contribute to obesity development in Black females.  Our current study concerns the cohort of 887 girls from one of the sites, University of California, Berkeley, which was recruited from public and private schools in the Richmond Unified School District.

Food Environment Assessment

West Contra Costa food environment from NETS

An important component of our study was to evaluate different options for characterizing  food environment exposures for the NGHS cohort.  We collected, cleaned, and compared various databases including: Yellow Pages, Internet listings on Yelp, Parcel data, local restaurant inspection data, and Dun and Bradstreet data.  Comparisons of these databases in relationship to nieghborhood social demographic characteristics was presented at the American Public Health Association (Hua, et al., 2011 presentation).  We ultimately licensed and used the National Establishment Timeseries (NETS) Dun and Bradstreet data because it could be used to reconstruct changes in food environment exposures by year for each NGHS participant.  Because NETS includes all business establishments, our research also developed SIC code definitions specific to food environments.

Causal Inference Framework

Our study’s hypothesis is that there are upstream determinants of diet and physical activity that affect growth trends in girls.  Using longitudinal data from the NGHS cohort allowed us to explore this conceptual framework by applying both structured equation model and G-computation causal inference methods, as well as agent-based modeling.

To explore this framework, we have applied longitudinal G-computation to a set of structured equations, and have used both generalized linear modeling as well as machine learning approaches to assess the impact of the built environment on mediators of BMI change.  The machine learning approach is based on ensemble prediction using a method called Superlearner, available here as a package in R.

For those interested in how we use Superlearner for modeling, here is a basic example from our modeling scripts that illustrates how a set of longitudinal structured equations are estimated using the method in R.  Details can be found in our forthcoming papers. The vignette that comes with the Superlearner R package provides more basic use examples.

Study findings

Manuscripts documenting our study’s findings are currently in review.  Until peer-reviewed, these are preliminary findings:
  • Controlling for race, stress, and other covariates, individual diet and physical activity had small, but significant effects on BMI trends.
  • Individual BMI trends were sensitive to the timing of changes in diet and physical activity.
  • Increased exposure to fast food increases caloric intake, but not BMI.  Those more exposed to fast food had:
    • 3 kcal’s/day more intake at baseline than those less exposed;
    • 31 kcals/day more intake at year 5; and
    • 63 kcals/day more intake at year 10.
  • Agent-based modeled trajectories of the impact of fast food exposure on different races (shown below), illustrate the disparate effects of fast food exposure on white versus black girls.  Black NGHS girls had higher BMI z-scores throughout the study period compared to the white NGHS girls. Exposure to fast food had opposing effects on Black girls versus White girls.  These effects are highly modified by income, and to a lesser extent self-reported perceived stress.
Efficient Simulation of the Agent-Based Model

As commonly done for ABM simulation studies, our study required many repeated simulations to observe differences between various scenarios. We ran separate scenarios considering hypothetical changes to the community food environment, neighborhood income levels, and for white versus black girls.  Running these many scenarios required  exploration into parallel computing. Because our ABM was implemented in R, instead of developing our own parallel computing algorithms from scratch, we leveraged existing libraries developed for both multi-CPU and GPU processing libraries.  For example:

Ultimately, both our statistical modeling using Superlearner and ABM models utilized the multi-CPU approach, which scaled quite well with additional CPUs, and which allowed us to use computer clusters on the Berkeley campus.

Next Steps

Our group continues research into the food environment and its relationship to diet and metabolic syndrome risk.  These studies and activities are related our ABM modeling study:

  • APHA 2012 presentation validating the use of 3D street view data to map the food environment.
  • Our Foodscoremap.com website allows researchers to map and download food environment data from around the world.
  • Our CalFit smartphone system can be used to assess individual-level exposures to food and other built environment factors, physical activity, diet, and emotional state.

Contact

Interested in our study, please contact May Wang <maywang@ucla.edu> and Edmund Seto <seto@berkeley.edu>.

Acknowledgements

NIH/NIDDK RC1DK086038  (P.I. May Wang, UCLA;  subaward P.I. Edmund Seto, UCB)

Other key personnel on this grant included:

  • Kate Crespi, UCLA
  • Rob Mare, UCLA
  • Gilbert Gee, UCLA
  • Alan Hubbard, UCB
  • Pat Crawford, UCB
Additionally, this project provided research experience to students at both UCLA and UCB.