2008 Travel Reports

The Third Sector and Sustainable Social Change: New Frontiers for Research
Organized by The International Society for Third Sector Research

Universitat de Barcelona
Barcelona, Spain
July 9-12, 2008

Presentation by Constanza Ulriksen
Department of Urban and Environmental Policy and Planning, Graduate School of Arts & Sciences

The Significance of Social Capital as a Concept for Understanding and Analyzing Community-Based Nonprofit Organizations in Boston, USA: Lessons for Chile


Constanza Ulriksen Travel Report Abstract:Some development theorists consider social capital a critical input for economic development. Others criticize it as a narrow way to describe poverty, where higher social capital is associated with better socioeconomic outcomes with no consideration of issues of power and economic capital.

This research examines the role of social capital in neighborhood revitalization through an extensive literature review and a case study of Nuestra Comunidad Development Corporation, a successful community-based organization in Boston, USA.

The main question is if this organization applies the social capital theory, and if not, what methods it uses to succeed. The final goal is to identify lessons for Chile, country currently using social capital to frame public neighborhood rehabilitation interventions.The findings indicate that even a comprehensive social capital model is incomplete to design sustainable neighborhood revitalization initiatives. Experts should employ a community development concept, and create a social policy that promotes organizational capacity, and the formation of an institutional support system in community development.

International Society of Environmental Epidemiology/International Society of Exposure Analysis Joint Conference 2008
Pasadena, California
October 12-16, 2008

Presentations by Jyotsna Jagai
PhD Candidate, Tufts University School of Medicine, Dept. of Public Health & Family Medicine

Symposium Presentation:

Epidemiologic analysis of waterborne diseases typically considers socio-economic, demographic characteristics and parameters of the infectious disease. However, hydrological parameters need to be considered as well. Increases in waterborne diseases have been associated with water quality characteristics such as turbidity and river flow. Each watershed has unique characteristics, such as streamflow, water temperature and turbidity which may predict the rates of waterborne diseases within a particular watershed. It is hypothesized that attributes of the watershed will predict the rates of waterborne disease throughout the watershed.

This preliminary analysis explores seasonal patterns of waterborne diseases along two watersheds in the U.S. using elderly hospitalization records from the Center for Medicare and Medicaid Services (CMS) from 1991 to 2004. We considered diagnoses of various gastrointestinal infections including, cryptosporidiosis (ICD 007.2, 007.4, 007.8), giardiasis (ICD 007.1), and GI Symptoms (ICD 558.9, 787), Ill-defined GI infections (ICD 008.5, 008.8, 009) and all GI infections without Clostridium Difficile (ICD 001-009 W/O 008.45). Cases were aggregated according to diagnosis code, location, and date of admission. Preliminary analysis of disease rates has been conducted for the Upper and Lower Mississippi River watersheds and the Ohio River watershed. Cities and counties within a 10-mile buffer of the rivers were selected using GIS. Annual disease rates by county were calculated using linearly interpolated elderly population for 1997 (midpoint of data timeframe) from 1990 and 2000 census as the denominator.

Seasonal patterns of disease rates were assessed using an annual harmonic regression controlling for hydrological parameters. This is a first attempt at modeling rates of waterborne diseases associated with river characteristics and water quality within a watershed. An understanding of the relationship of watershed characteristics to health outcomes will allow for evaluation of drinking water regulations and policies in the U.S.

Poster Presentation:

Proper understanding of seasonality of enteric infections will improve our knowledge of host-pathogen biology and enhance our capability to predict future outbreaks. Rotavirus and cryptosporidiosis have distinct seasonal patterns depending on geographic location. Rotavirus shows distinct peaks during cooler months in temperate climate but no such seasonality is observed in tropical climates. On the other hand, cases of cryptosporidiosis tend to increase during warm rainy season in the tropics, whereas it peaks during winter and fall in temperate zones. This study assessed seasonality of childhood diarrhea in general, and diarrhea due to rotavirus and Cryptosporidium spp. in particular.

Materials and methods:
We conducted a weekly time-series analysis of data available from a birth cohort study in Vellore, south India. A total of 452 children were recruited at birth and followed up for three years on a twice-weekly basis. Fecal samples were collected from all children fortnightly as part of routine surveillance. Additionally, fecal samples were collected from any child who developed diarrhea, identified either by a routine field worker visit or self-referral by the mother.

Data on all diarrheal episodes experienced by the cohort were extracted and merged with the database containing stool microbiology results. For rotavirus, only those samples positive by RT-PCR was considered positive. Cryptosporidium spp. was identified by microscopy of stool samples. Staggering recruitment was adjusted for by assessing seasonality of prevalence rates per 1000 child weeks. Seasonality was assessed for all diarrheal episodes as well as for the rotavirus and cryptosporidial diarrhea separately. Linear and quadratic trend were adjusted to account for potential changes in symptomatology due to protective immunity associated with re-exposure in the aging cohort.

The cohort experienced a total of 2005 episodes of diarrhea, out of which 289 were due to rotavirus and 57 due to Cryptosporidium spp. Diarrhea overall peaked in the first week of June (week 24). In the last week of August in the second and fourth year of follow-up peaks of diarrhea due to Cryptosporidium spp. were detected. Rate of rotavirus diarrhea peaked in second week of October (week 42) and had high amplitude in the first year that diminished over time. G1 and G2 were the most common genotypes for rotavirus whereas C. hominis was the most common species of Cryptosporidium identified. Strain-specific analysis showed both G1 and G2 rotavirus to have a well-defined seasonality with a significant negative trend. G1 appeared to exhibit a biannual peak with high intensity, whereas G2 strain appeared in an annual basis with less intensity. Coinciding peaks of G1 and G2 amplified the overall rotavirus intensity.

This study uses data from a community-based diarrheal disease study to look at the seasonality of diarrhea overall, and rotavirus and cryptosporidial diarrhea in children in south India. Preliminary findings suggest a strong seasonal pattern in potentially waterborne infection that was pronounced during hot and dry season. Further analysis will control for meteorological parameters including temperature and precipitation, behavioral factors, and water quality.


ISEE-ISEA Joint Annual Conference 2008
Pasadena, CA, USA
Oct 12-16, 2008

Presentation by Ken Chui
Friedman School of Nutrition Science and Policy

Temporal Variation in Salmonella Infection in the US Elderly, 1991-2004

Evaluation of new policies in environmental sciences validates cost-effectiveness of decisions and informs future ones. However, there are caveats in such evaluation when the outcome shows trend and seasonality. Improper modeling of long-term trend may undermine associations between the outcome and the policy, while failing to account for the seasonal pattern of the outcome can lead to biases. We used Salmonella infections (SI) to demonstrate the use of time series analysis as a policy evaluation tool.

In 1997, USDA started the Pathogen Reduction and Hazard Analysis and Critical Control Points Systems (HACCP), which demands regular testing for Salmonella in the chicken broiler industry however a temporally and geographically focused prevention strategy may improve cost-effectiveness. Using nationally representative data, we evaluated the trend and seasonality before and after the start of the HACCP and identified potential areas for preventive service.

Weekly hospitalizations in US elderly involving SI (ICD 9-CM 003) in 1991–2004 (n=27,790) were obtained from the Center of Medicare and Medicaid Services. To assess long-term changes, we fit a broken-stick model with the first week of 1997 as the inflection point, adjusted for annual seasonality. To assess seasonal patterns, we applied annual harmonic regression and annual oscillations were compared. Analysis was conducted for the nine Census divisions in the contiguous US.

Overall, there was a decrease in SI hospitalization rates. Changes before and after HACCP varied, e.g. New England had a decrease before HACCP and a significantly steeper decrease after HACCP; whereas, East South Central had a decrease in the pre-HACCP period, and then remained stable. Regarding seasonality, West South Central, East South Central, and South Atlantic could be potential foci of targeted interventions due to higher rates, and consistent annual oscillation.

The effect of interventions and regulations targeting outcomes with strong seasonality should be examined in three ways: (i) change of overall rate regardless of seasonality, (ii) change of seasonal patterns with respect to time and (iii) with respect to rate of the outcome. We suggested applying time series analysis to assess the effectiveness of policies and provide feedback to policy makers and health professionals.




ISEE-ISEA Joint Annual Conference 2008
Pasadena, California
October 12-16, 2008

Presentation by Christine Rioux
Graduate School of Arts & Sciences and Engineering

Comparison of Four Methods for Characterizing Traffic-Related Air Pollution (TRAP) Exposures in Boston’s Inner Core

Christine Rioux conference

Inner-city traffic burdens are difficult to characterize due to the complexity and heterogeneity of the roadway networks. Our analysis compares four different methods for consistency in estimating relative magnitude of TRAP burden. The study subjects were obtained from a larger study into the effects of TRAP on biomarkers of inflammation in Puerto Rican adults residing in the Boston area. A subset of cases (n=48) subjected to intensive positional confirmation was selected to minimize errors commonly associated with the geocoding of participant addresses, inappropriate extrapolation of traffic volume data and spatial misalignment of data sourses. TRAP exposures were assessed using: (1) representative traffic counts for specific road segments; (2) GIS-based road density analysis within 50m of each residence; (3) estimates of vehicle miles travelled per square mile (VMT/SM); and (4) modeled levels of selected TRAP parameters including carbon monoxide (CO) and oxides of nitrogren (NOx). VMT/SM, CO, and NOx estimates were specific to individual Traffic Analysis Zones in which the residence was located. TAZs are defined by state planning organizations and used in calibrating transportation planning models according to state and federal protocols.

All 48 cases lived directly on one of Boston’s most highly trafficked roads (>25,000 average weekday traffic volume). Complicating the positional confirmation process in Boston’s inner core, there are 18 distinct neighborhoods with over 200 instances of the same road name being used for two or more different non-contiguous locations. Twenty four cases lived directly on street segments where traffic count stations were located (as matched by neighborhood name or zip code). Seventeen cases lived on the same street segment characterized by traffic counts, yet they were located in 3 different TAZs. VMT/SM estimates ranged from 65,000 to over 197,000. CO and NOx emissions estimates varied by factors of 2.9 and 2.8, respectively. Road densities within a 50m radius differed by nearly 6-fold among the 17 cases and for cases within the same TAZ road densities varied as much as a factor of 2.8. Another eight of these cases lived on another street characterized by the same traffic counts, though within 2 different TAZs. Estimated VMT/SM, CO, and NOx for this set of locations varied by factors of 3.6, 3.5, and 3.0, respectively. Road density varied by a factor of nearly 2.5 within one of the TAZs but was relatively consistent for all cases in the other TAZ.

This study illustrates that the relative magnitude of estimated TRAP burden can vary depending on the exposure characterization method that is used. Our findings suggest a strong potential for exposure misclassification when traffic counts and proximity measures are the primary methods of characterization. For health outcomes research, two relatively accessible and cost-effective tools, a combination of TAZ data and weighted road density analysis, could provide a more comprehensive picture of the intra-urban, and much localized, variations in TRAP burden.


American Association of Zoo Veterinarians
Los Angeles, California
October 10-17, 2008

Presentation by Tierra Wilson
Cummings School of Veterinary Medicine

Development of a PCR Diagnostic Technique for Differentiation of Mycobacterium Species in Elephant Trunk Wash Samples in Nepal

Mycobacterium tuberculosis (M. tb) is a known threat to captive elephants in the USA and elephant range countries. In Nepal, it is unknown whether M. tb and/or Mycobacterium bovis (M. bovis) are responsible for pulmonary mycobacteriosis in elephants. M. bovis is endemic within Nepali cattle and water buffalo populations and M. tb is endemic in the human population. Captive elephants are in close association with humans and share grazing pastures with cattle, water buffalo, and wild elephants providing opportunity for bidirectional transmission of both M. tb and M. bovis. Identification of the Mycobacterium sp. responsible for elephant mycobacteriosis is essential to understanding the mode of transmission in Nepal.

Diagnosis of elephant mycobacteriosis relies on trunk wash (TW) culture and serology. TW culture requires BL3 facilities, takes 8-12 weeks, and lacks sensitivity. Serology while sensitive, does not differentiate the various Mycobacteria sp. We have developed a highly sensitive gyrB-based PCR-RFLP assay capable of differentiating M. tb, M. bovis, and M. avium. When applied to TW samples from 22 suspected positive Nepali elephants, PCR inhibitors were localized to debris present in elephant trunks. Following inhibitor removal, PCR-RFLP was able to successfully differentiate M. tb, M. bovis, and M. avium in TW samples spiked with DNA but was unable to detect Mycobacteria in the 22 suspected positive elephants. Current studies are directed towards improving removal of PCR inhibitors and increasing recovery of Mycobacterium DNA from TW samples.

Terra Madre 2008
Turin, Italy
October 23-27, 2008

Attended by Asta Schuette
Friedman School of Nutrition Science and Policy

Outcome of the conference:
Held concurrently with Salone del Gusto in Torino from October 23 to 27, the third edition of the biennial international meeting of the Terra Madre Network brings together food communities, cooks, academics and youth delegates for four days to work towards increasing small-scale, traditional, and sustainable food production.

Asta plans to use the contacts she met at Terra Madre to organize an Eat-in here in Boston to bring attention to the need for a sustainable food system.

View additional travel reports by year: