Applied risk mapping and spatial analysis of address-level decedent data to inform opioid overdose interventions: The Massachusetts HEALing Communities Study

BACKGROUND

Death certificate data provide powerful and sobering records of the opioid overdose crisis. In Massachusetts, where address-level decedent data are publicly available upon request, mapping and spatial analysis of fatal overdoses can provide valuable insights to inform prevention interventions. We describe how we used this approach to support a community-level intervention to reduce opioid-involved overdose mortality.

METHODS

We developed a method to clean and geocode decedent data that substituted injury locations (the likely location of fatal overdoses) for deaths recorded in hospitals. After geomasking for greater privacy protection, we created maps to visualize the spatial distribution of decedent residence addresses, alone and juxtaposed with drive and walk-time distances to opioid treatment programs (OTPs), and place of death by overdose address. We used spatial statistical analyses to identify locations with significant clusters of overdoses.

RESULTS

In the 8 intervention communities, 785 individuals died from opioid-involved overdoses between 2017 and 2020. We found that 19.7% of fatal overdoses were recorded in hospitals, 50.2% occurred at the decedent's residence, and 30.1% at another location. We identified overdose hotspots in study communities. By juxtaposing decedent residence data with drive- and walk-time analyses, we highlighted actionable spatial gaps in access to OTP treatment.

CONCLUSION

To better understand local fatal opioid overdose risk environments and inform the development of community-level prevention interventions, we used publicly available address-level decedent data to conduct nuanced spatial analyses. Our approach can be replicated in other jurisdictions to inform overdose prevention responses.

Spatial access to buprenorphine-waivered prescribers in the HEALing communities study: Enhanced 2-step floating catchment area analyses in Massachusetts, Ohio, and Kentucky

INTRODUCTION

The opioid overdose epidemic continues to impact a large swath of the population in the US. Medications for opioid use disorders (MOUD) are an effective resource to combat the epidemic; however, there is limited research on MOUD treatment access that accounts for both supply of and demand for services. We aimed to examine access to buprenorphine prescribers in the HEALing Communities Study (HCS) Wave 2 communities in Massachusetts, Ohio, and Kentucky during 2021, and the association between buprenorphine access and opioid-related incidents, specifically fatal overdoses and opioid-related responses by emergency medical services (EMS).

METHODS

We calculated Enhanced 2-Step Floating Catchment Area (E2SFCA) accessibility indices for each state, as well as Wave 2 communities in each state, based on the location of providers (buprenorphine-waivered clinicians from the US Drug Enforcement Agency Active Registrants database), population-weighted centroids at the census block group level, and catchment areas defined by the state or community's average commute time. In advance of intervention initiation, we quantified the opioid-related risk environment of communities. We assessed gaps in services by using bivariate Local Moran's I analysis, incorporating accessibility indices and opioid-related incident data.

RESULTS

Massachusetts Wave 2 HCS communities had the highest rates of buprenorphine prescribers per 1000 patients (median: 165.8) compared to Kentucky (38.8) and Ohio (40.1). While urban centers in all three states had higher E2SFCA index scores compared to rural communities, we observed that suburban communities often had limited access. Through bivariate Local Moran's I analysis, we identified numerous locations with low buprenorphine access surrounded by high opioid-related incidents, particularly in communities that surrounded Boston, Massachusetts; Columbus, Ohio; and Louisville, Kentucky.

CONCLUSION

Rural communities demonstrated a great need for additional access to buprenorphine prescribers. However, policymakers should also direct attention toward suburban communities that have experienced significant increases in opioid-related incidents.

Perspectives of opioid prescribers in overdose hotspots and coldspots, Massachusetts, 2019

BACKGROUND

Prescription opioids (POs) have had a devastating effect on people and public health systems in the U.S. Due to the urgency and complexity of the opioid crisis, there is a need to expand qualitative research on the medical community's perspectives on opioid prescribing practices and the role that prescription drug monitoring programs (PDMPs) have played in mitigating this crisis.

METHODS

We conducted qualitative interviews with clinicians (n = 23) across specialties and a range of overdose hotspot and coldspot locations in Massachusetts during 2019. We aimed to capture their perspectives on the opioid crisis, changes in clinical practice, and experiences with opioid prescribing and PDMPs.

RESULTS

Respondents consistently recognized the role clinicians played in the opioid crisis and reported reductions in their opioid prescribing, which were motivated by the crisis itself. The limitations of opioids in pain management were frequently discussed. While clinicians appreciated having greater awareness of their opioid prescribing and increased access to patient prescription histories, they also expressed concerns about surveillance of their prescribing and other unintended consequences. We observed that clinicians in opioid prescribing hotspots had more detailed and specific reflections on their experiences with the Massachusetts PDMP, MassPAT.

CONCLUSION

Clinician perceptions of the severity of the opioid crisis in Massachusetts and thoughts on their role as prescribers were consistent across specialty, prescribing level, and practice location. Many clinicians in our sample cited use of the PDMP as an influence on their prescribing. Those practicing in opioid overdose hotspots had the most nuanced reflections about the system.

Methods for jurisdictional vulnerability assessment of opioid-related outcomes

In 2020, an estimated 2.7 million people in the US had opioid use disorder, increasing their risk of opioid-related morbidity and mortality. While jurisdictional vulnerability assessments (JVA) of opioid-related outcomes have been conducted previously in the US, there has been no unifying methodological framework. Between 2019 and 2021, we prepared ten JVAs, in collaboration with the Council of State and Territorial Epidemiologists, the Centers for Disease Control and Prevention, and state public health agencies, to evaluate the risk for opioid-involved overdose (OOD) fatalities and related consequences. Our aim is to share the framework we developed for these ten JVAs, based on our study of the work of Van Handel et al. from 2016, as well as a summary of 18 publicly available assessments of OOD or associated hepatitis C virus infection vulnerability. We developed a three-tiered framework that can be applied by jurisdictions based on the number of units of analysis (e.g., counties, ZIP Codes, census tracts): under 10 (Tier 1), 10 to <50 (Tier 2), and 50 or more (Tier 3). We calculated OOD vulnerability indices based on variable ranks, weighted variable ranks, or multivariable regressions, respectively, for the three tiers. We developed thematic maps, conducted spatial analyses, and visualized service provider locations, drive-time service areas, and service accessibility relative to OOD risk. The methodological framework and examples of our findings from several jurisdictions can be used as a foundation for future assessments and help inform policies to mitigate the impact of the opioid overdose crisis.

Relationships between places of residence, injury, and death: Spatial and statistical analysis of fatal opioid overdoses across Massachusetts

Understanding the factors associated with where people who use opioids live, where their fatal overdoses occur, and where deaths are recorded can improve our knowledge of local risk environments and inform intervention planning. Through geospatial analyses of death certificate data between 2015 and 2017, we found that a majority of opioid-involved fatal overdoses in Massachusetts occurred at home. Age (adjusted odds ratio [AOR], 1.03; 95% confidence interval [CI], 1.02-1.04), living in a census tract with a higher percentage of crowded households (AOR, 1.04; 95% CI, 1.01-1.08), households without vehicles (AOR, 1.01; 95% CI, 1.00-1.02), and Hispanic ethnicity (AOR, 0.56; 95% CI, 0.42-0.74) were independently associated with fatal overdose at home. Using geographically weighted regression, we identified locations where these associations were stronger and could benefit most from home-based and culturally sensitive overdose prevention efforts, including expanded overdose education and naloxone distribution.

Improving Healthy Food Access Spatial Analyses With Community Input and Retailer and Food Pantry Surveys

Objectives

Healthy food access can be analyzed spatially, but availability, affordability, desirability, and transport mode are among many potential impediments that have not been adequately measured in previous geographic information system (GIS) based food access studies. We aimed to identify and measure non-spatial factors to inform and improve the GIS/spatial analysis of healthy food access in three Health Equity Zone designated municipalities in Rhode Island.

Methods

Here, healthy food was defined as produce. Three instruments were deployed: a community questionnaire; a survey of fresh and canned produce availability, general pricing, and produce quality at food retailers; and a survey of produce availability at food pantries. Retailer affordability was ranked using the collected pricing data and verified against community perceptions of affordable stores. Community input defined travel time parameters for mapping provider service areas for car, bus, and walking transportation. A novel bus access analysis incorporated walk, wait, and weighted drive times. Accessibility was mapped separately for regular-cost supermarkets, low-cost supermarkets, and food pantries. Results were compared to poverty, car ownership, and older households at the block group level to evaluate how neighborhoods most in need were being served.

Results

Pricing data corresponded closely to community perceptions of store affordability, allowing us to categorize two of 16 supermarkets serving our study communities as expensive and to be excluded from spatial analyses. Five were categorized low-cost. Four out of 10 food pantries were excluded because they did not offer fresh produce. We found no convenience, corner, or dollar store met a priori criteria for a healthy food provider, with the threshold for fresh produce met in only one out of 48 of these stores. Two communities had no low-cost supermarkets accessible by car, bus, or foot. A limited number of people had bus or foot access to healthy food.

Conclusions

Healthy food access analyses can be more accurate and useful when (1) travel time parameters are set by the community and the providers assessed for affordability, availability of fresh and canned produce, and quality, and (2), refined GIS techniques are used, including multi-step transit analysis and multiple layers of retailer and pantry service areas.

Funding Sources

None.

Opioid-Involved Overdose Vulnerability in Wyoming: Measuring Risk in a Rural Environment

BACKGROUND

Between 2009 and 2019 opioid-involved fatal overdose rates increased by 45% and the average opioid dispensing rate in Wyoming was higher than the national average. The opioid crisis is shaped by a complex set of socioeconomic, geopolitical, and health-related variables. We conducted a vulnerability assessment to identify Wyoming counties at higher risk of opioid-related harm, factors associated with this risk, and areas in need of overdose treatment access to inform priority responses.

METHODS

We compiled 2016 to 2018 county-level aggregated and de-identified data. We created risk maps and ran spatial analyses in a geographic information system to depict the spatial distribution of overdose-related measures. We used addresses of opioid treatment programs and buprenorphine providers to develop drive-time maps and ran 2-step floating catchment area analyses to measure accessibility to treatment. We used a straightforward and replicable weighted ranks approach to calculate final county vulnerability scores and rankings from most to least vulnerable.

FINDINGS

We found Hot Springs, Carbon, Natrona, Fremont, and Sweetwater Counties to be most vulnerable to opioid-involved overdose fatalities. Opioid prescribing rates were highest in Hot Springs County (97 per 100 persons), almost two times the national average (51 per 100 persons). Statewide, there were over 90 buprenorphine-waivered providers, however accessibility to these clinicians was limited to urban centers. Most individuals lived further than a four-hour round-trip drive to the nearest methadone treatment program.

CONCLUSIONS

Identifying Wyoming counties with high opioid overdose vulnerabilities and limited access to overdose treatment can inform public health and harm reduction responses.

Food Choice With Economic Scarcity and Time Abundance: A Qualitative Study

Consumers with low income in the United States have higher vulnerability to unhealthy diets compared with the general population. Although some literature speculates that scarcity is an explanation for this disparity, empirical evidence is lacking. We conducted a qualitative study of food choice to explore whether scarcity-related phenomena, such as tunneling and bandwidth tax, may contribute to unhealthy dietary choices. We used participant-driven photo elicitation (n = 18) to investigate the food choice behaviors of individuals living in the greater Boston area who met the federal guidelines for poverty. Participants took photos at the point of food acquisition for 1 month, after which we interviewed them using a semistructured interview guide with the photos as prompts. Thematic coding was used for analysis. Respondents had relative time abundance. Two major themes emerged: participants used a set of strategies to stretch their budgets, and they highly prioritized cost and preference when making food choices. The extreme focus on obtaining food at low cost, which required time and effort, was suggestive of tunneling. We found no evidence of the bandwidth tax. Our findings raise the hypothesis of scarcity as a continuum: when individuals experience multiple resource constraints, they experience scarcity; whereas people with very limited finances and relative time abundance may instead be in a prescarcity condition, with a hyperfocus on a scarce resource that could lead to tunneling as constraints increase. Additional studies are needed to understand whether and how tunneling and bandwidth tax emerge, independently or together, as people face different levels and types of scarcity.

The Health and Housing Study: Nutrition-Related Outcomes Among Low- to Moderate-Income Individuals Living in or Seeking to Live in New Subsidized Chinatown-Based Housing

Objectives

Housing stability, food security and healthy eating are interrelated. In 2015, a housing lottery for newly developed subsidized housing in Chinatown, Boston, MA was held for low- and moderate- income families who were randomly selected from a waitlist. Four years later (19/20), our study explores the associations between housing status (living in the new subsidized building (NSB) or being on the lottery waitlist) and Food Insecurity (FI); use of food safety net programs (SN), defined as use of SNAP, WIC and/or food pantries; weekly fruit and vegetable consumption (FV); weekly soda consumption; and monthly fast food consumption (FF).

Methods

Questionnaire links were sent to NSB (n = 95) and waitlist (n = 2498) households. Logistic regression was used to explore the association between housing status and (1) FI (2-item screener) and (2) use of SN (self-reported use in the past year). Linear regression with robust standard error estimation was used to explore the association between housing status and weekly consumption of (1) FV, (2) soda, and (3) monthly consumption of FF, (assessed through health behavior questions from the California Health Interview Survey). Models were adjusted for age, sex, ethnicity, household size, education, income, employment status and distance to grocery store.

Results

138 respondents completed the survey; NSB = 36, waitlist = 102. The two housing status groups were demographically similar. Respondents in the NSB showed non-statistically significant lower odds of experiencing FI (OR 0.65, 95% CI [0.3, 1.6]) and of using SN (OR 0.45, 95% CI [0.1, 1.6]). Respondents in the NSB showed non-statistically significant higher weekly consumption of FV (β 2.03, 95% CI [−1.5, 5.6]); lower weekly consumption of soda (β −0.39, 95% CI [−1.9, 1.1]) and lower monthly consumption of FF (β −1.19, 95% CI [−3.9, 1.5]), when compared to respondents on the waitlist.

Conclusions

In this small sample, main outcomes were not significantly different. Dietary intakes among NSB residents were consistently positive compared to waitlist residents. Future studies should continue to explore mechanisms through which living in a NSB may impact nutrition and health outcomes.

Funding Sources

Tufts Collaborates Seed Grants. Conclusions do not represent any U.S. Department of Agriculture office or Government determination. This research was not part of MI's work at USDA.