Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jan 1.
Published in final edited form as: Int J Drug Policy. 2020 Nov 1;87:103018. doi: 10.1016/j.drugpo.2020.103018

Spatiotemporal Trends in Discarded Needle Reports in San Francisco Over a 10-year Period, 2010–2019

Laila Fozouni a, Saira Khan b, Benjamin Bearnot c,d
PMCID: PMC7940553  NIHMSID: NIHMS1642817  PMID: 33142160

Abstract

Background:

To describe the geographic and spatiotemporal distribution of needle reports in San Francisco, and examine spatial relationships between needle reports and needle disposal boxes, needle disposal kiosks, and homeless shelters.

Methods:

We conducted multiple geospatial analyses of a crowdsourced database of non-emergency service requests. We describe changes in discarded needle and bulky item reports across San Francisco between 2010 and 2019, and compared changes in the reporting patterns of these items, while 200-meter Euclidean buffers captured needle reports in close proximity to needle deposit boxes, needle deposit kiosks, and homeless shelters in 2019.

Results:

34,912 needle reports were included. Yearly needle reports increased by 3827.1%, with a markedly different geospatial distribution from bulky item reports. 45.6% of needle reports originated in the five downtown neighborhoods with the highest needle report density, and 33.8% were identified within 200 meters of boxes, kiosks, or homeless shelters.

Conclusions:

Reports of discarded needles in San Francisco increased dramatically over the last decade, and more than one third of 2019 reports were adjacent to harm reduction and homeless shelter locations. Needle reports provide an opportunity to understand changes in public injection drug use and target harm reduction services.

Keywords: Public injection drug use, overdose, opioids, improperly discarded needles, homelessness

Introduction

Drug overdoses are the leading cause of accidental death in the United States for individuals under 50 years old, and have contributed to decreasing life expectancy among Americans.1 Overdoses from opioids and stimulants, including methamphetamine and cocaine, have increased in part due to the contamination of the drug supply with potent synthetic opioids, such as illicitly manufactured fentanyl and its analogs, which are often short-acting and require frequent use to avoid withdrawal states.2 Despite the fact that some people who inject drugs (PWID) choose to inject in public to increase the chances of an overdose reversal, public injection is associated with behaviors that increase morbidity and mortality including rushed drug preparation, syringe sharing, and higher risk of overdose.3–6 A leading risk factor for public injection is homelessness, as individuals experiencing housing instability frequently lack private spaces to use more safely.3, 4 Studies suggest that the location of public drug use is highly correlated with both the location of drug acquisition as well as the location of discarded syringes.7

Crowdsourced and publicly available databases of municipal response services, such as 311, present an important opportunity to understand the geospatial distribution of reported discarded needles and syringes. Over 200 cities and municipalities in the United States have programs like 311 for reporting non-emergency concerns.8 In 2007, a 311 reporting service was established in the city of San Francisco (SF) for telephone reporting, and by 2010 reports could be made by web, social media, and a mobile application (app). Prior analysis of 311 needle reports in Boston identified two neighborhood hot spots, and an association between markers of poverty, including homeless shelters, and increased numbers of needle reports.9 To our knowledge, no such analysis has been conducted with 311 needle reports in SF.

Compared with the rest of the United States, SF has experienced relatively fewer overdose deaths from prescription opioids and heroin for a variety of reasons, including widely available harm reduction services and local variation in opioid source forms.10, 11 SF has worked to reduce the burden of publicly discarded needles and syringes, in part by installing small 24 hour needle disposal boxes and larger 24 hour needle disposal kiosks, comparable in size to postal boxes. The city has also worked to shelter its growing population of individuals with unstable housing, who are at higher risk of injecting publicly.12 To achieve this, the city has opened new shelters and navigation centers, which aim to transition individuals into permanent housing.13 However, the city has still faced increasing numbers of overdoses, which rose notably in 2018. The primary objective of this study was to describe the spatiotemporal distribution of needle reports in SF over a 10-year period, between January 2010 and December 2019. This analysis also aimed to understand the contribution of increased usage of 311 services to the change in number of needle reports by comparing the temporal change in number of needle reports to the temporal change in number of bulky item reports, which were the most common report category in the 311 database. Finally, this study aimed to understand the relationship between needle reports and the proximity to needle boxes, needle kiosks, and homeless shelters.

Methods

Understanding the spatiotemporal characteristics of 311 needle reports

Needle reports from January 2010 to December 2019 were obtained from the 311 database available via the SF Open Data website.14 The dataset was cleaned to exclude incorrect, incomplete, or duplicate entries. Needle reports were excluded from the study if they did not include x,y coordinates or fell outside of San Francisco city boundaries. The remaining entries were geocoded using ArcMap 10.6.1 (ESRI, Redlands, CA) to examine the spatiotemporal distribution of needle reports. The dataset contains the date and time of the initial report, the agency responsible for cleanup, and the street address and neighborhood where the disposed needle(s) was reported. In some instances, the neighborhood field had no value, in which case the x and y coordinates enabled the geolocation of points to correctly label the neighborhood in which the needle was reported.

We identified the yearly geographic center of needle reports by calculating the average center point of all data entries using their x and y coordinates.15 Needle reports were summed by the neighborhood in which they were reported and the reporting mechanism. Additional analyses were conducted by spatially joining needle reports to a 100 meter (m) by 100m (100m2) grid overlying a map of the city. Needle reports were analyzed both over the entire study period as well as on an annual basis.

Comparing needles and bulky item reports

Needle reports were compared to public reports of bulky items in order to understand reporting biases and changes in reporting trends between 2010 and 2019. Bulky items, which describe solid waste that does not fit within a container, including household appliances, furniture, carpets, mattresses, and oversized yard waste, were chosen for comparison because they were the most frequently reported category in the SF 311 database.16 Just as with needle reports, bulky item reports were cleaned to exclude incomplete data entries and joined to a 100m2 grid overlying the city to calculate reporting density.

Spatial characterization of needle reports, kiosks, boxes, and homeless shelters

In 2019 SF had 22 needle disposal sites in the city to supplement ongoing needle collection efforts. These sites include 12 small disposal boxes as well as 10 larger kiosks. Data regarding locations of needle disposal boxes and kiosks were obtained from the SF Department of Public Health.

We included homeless shelters in our analysis given the strong association between homelessness and public injection drug use. Furthermore, many shelters do not allow for either new or used needles to be brought into the shelter space, potentially leading to an increase in improperly discarded needles found immediately outside of shelters.

To examine the spatial relationship between discarded needle reports and needle boxes, kiosks, and homeless shelters in SF, the latter three variables were geocoded using Google Earth Pro (Google, Mountain View, CA) and imported into ArcMap 10.6.1. To capture spatial variability, the minimum, maximum, and average distance between the needle boxes, kiosks and, homeless shelters was calculated for each dataset using the Point Distance tool.17

200m Euclidian buffers were generated around each point of interest to capture the number of reported needle points within these distances. 200m was chosen as the smallest possible distance to account for the large size of some homeless shelters, as the buffer is generated around the centroid of the shelter, but still provides useful information about the distribution of needle reports in close proximity to points of interest.

Results

Needle Reports

Between January 1, 2010 and December 31, 2019, a total of 35,935 needle reports were made in SF via the 311 service, and 34,912 (97.2% of total needle reports) were included in this analysis. The majority of needle reports were made through the 311 mobile app (60.2%), while 34.0% of needle reports were made over the phone, 4.6% were made over the Web, and 1.3% of reports were made from other sources, including Twitter.

The number of yearly needle reports increased dramatically over the study period, see Figure 1. In 2010, there were a total of 221 needle reports, which rose to 9,373 reports in 2018. There was a small decrease to 8,679 needle reports in 2019. In the broader context of San Francisco’s total population, in 2010 there were 0.27 needle reports per thousand SF residents, compared to 2019 when there were 9.8 needle reports per thousand residents. The number of needle reports increased by 3827.2% over the 10-year study period. When considering the change in density of needle reports over time, there was a maximum of 10 needle reports in a single 100m2 grid in 2010, a maximum of 110 needle reports in a 100m2 grid in 2018, and a maximum of 65 needle reports in a single 100m2 grid in 2019.

Figure 1.

Figure 1.

Open in a new tab

Annual 311 reports of bulky items and discarded needles in San Francisco between 2010 and 2019.

Bulky Item Reports

Between January 2010 and December 2019, a total of 505,028 bulky item reports were made, and 501,034 (99.2%) were included in this analysis. The plurality of bulky item reports were made over the phone (47.7%); 33.5% of needle reports were made using the 311 mobile app, 10.3% were made through an integrated agency, 8.0% were made over the web, and 0.6% of reports were made from other sources, including Twitter.

The number of bulky item reports also increased over the 10 year study period, see Figure 1. In 2010, there were 16,301 bulky item reports; by 2019 there were 100,870 bulky item reports. The number of bulky item reports increased by 518.8% over the 10 year study period. The number of 311 needle reports increased out of proportion to the number of bulky item reports: in 2010, there were 0.01 needle reports per bulky item report, while in 2019 there were 0.09 needle reports per bulky item report. When considering the change in density of bulky item reports over time, there was a maximum of 53 bulky item reports in a single 100m2 grid in 2010 and a maximum of 345 bulky item reports in 2019.

Geospatial distribution of needle reports and bulky item reports over 10 years

The geospatial distribution of needle and bulky item reports differed over the study period. A total of 3,814 grid cells (29.8%) contained one or more needle reports between 2010 and 2019. The highest density of needle reports was in downtown SF, primarily in the Tenderloin, Civic Center, Hayes Valley, South of Market (SoMa), and Castro neighborhoods, see Figure 2. These five neighborhoods contained 15,922 (45.6%) of total needle reports. Over the 10 year period, there was a maximum of 352 needle reports in a single 100m2 grid, which was found in SoMa. Despite rising needle reports between 2010 and 2019, the yearly geographic center of reports were within 294.4m of each other in the Southwestern portion of the SoMa neighborhood.

Figure 2.

Figure 2.

Open in a new tab

Distribution of all discarded needles 311 reports per 100m2 in San Francisco between 2010 and 2019, based on 311 reports.

In contrast, bulky item reports were distributed more broadly across SF, with 8,889 100m2 grids (69.5%) containing at least one bulky item report over the 10 year period. When considering the 100m2 grids with the highest number of bulky items reports, the reports were more spatially dispersed across the city than needle reports. The highest density of bulky item reports was in the Nob Hill, Tenderloin, Cole Valley, Lower Nob Hill, and Chinatown neighborhoods, see Figure 3. These five neighborhoods contained 58,219 (11.6%) of total bulky item reports. There was a maximum of 1,210 bulky item reports in a single 100m2 grid over the 10-year study period, which was found in the Tenderloin neighborhood. The geographic center of all bulky items reports was in Mission Dolores, approximately 1.6 kilometers (km) southwest from the geographic center of the needle reports over the same period. Over 10 years, this center remained fairly stable, though more variable than the geographic center for the needles, with a maximum distance of 473.6m between any two annual mean centers.

Figure 3.

Figure 3.

Open in a new tab

Distribution of all bulky items reports per 100m2 in San Francisco between 2010 and 2019, based on 311 reports.

Needle reports in relation to needle boxes, needle kiosks, and homeless shelters

The geocoded dataset included 12 needle boxes, 10 needle kiosks, and 47 homeless shelters that existed in 2019, see Figure 4. The mapped needle boxes were located in Tenderloin, Civic Center, SoMa, Castro, and Mission neighborhoods. Needle kiosks were also found in these five neighborhoods, in addition to Central Waterfront and Produce Market. The distance between any 2 individual needle boxes ranged from 0.6m to 2.8km, with an average distance of 850m. The distance between individual needle kiosks ranged from 68.8m to 4.6km, with an average distance of 1.7km. Unlike the safe needle drop off services, homeless shelters were more widely dispersed across the city. The distance between homeless shelters ranged from 0m, indicating that there was more than one shelter located at the same street address, to 10.5km, with an average distance of 3.0km between shelters. Both needle boxes and kiosks were clustered in close proximity to each other, and in areas with some of the highest density of needle reports. While there were also clusters of homeless shelters near high density needle report areas, homeless shelters had a wider geospatial distribution around the city and appear less densely concentrated.

Figure 4.

Figure 4.

Open in a new tab

Distribution of all discarded needles 311 reports per 100m2 in San Francisco in 2019 based on 311 reports, overlaid with known locations of homeless shelters, needle boxes and kiosks.

When looking at 200m Euclidian buffers, 265 (9.7%) of needle reports were found within 200m of 24 hour needle disposal boxes, which comprised 0.7% of SF’s total area; 369 (12.0%) of needles were reported within 200m of 24 hour needle disposal kiosks, which comprised 0.9% of SF’s total area. Approximately 724 (26.1%) of needle reports originated within 200m of homeless shelters, which comprised 3.6% of SF’s total area. Collectively, 33.8% of needle reports were identified within 200 meters of needle disposal boxes, kiosks, or homeless shelters, which comprises 4.3% of SF’s total area.

Discussion

Crowdsourced, publicly available 311 databases provide a novel lens through which to examine the evolving addiction and drug overdose crises in the US. In this descriptive study, we found that needle reports in SF dramatically increased over time, and were concentrated in the Tenderloin, Civic Center, Hayes Valley, SoMa, and Castro neighborhoods. This was in contrast to the distribution of bulky items reports, which were more evenly distributed throughout the city. The rising proportion of needle reports to bulky item reports over time suggests that the rise in needle reports is not wholly explained by increased knowledge or usage of the 311 app alone, and instead reflects a true increase in number of publicly discarded needles related to injection use of opioids or stimulants. Future studies should aim to investigate potential causes of increasing 311 needle reports, such as a rise in homelessness or evolving local drug markets.

The location of publicly discarded needles may indicate locations of public injection drug use, as previous studies have identified a strong relationship between locations of public injection and discarded needles and syringes.7, 18, 19 The geographic distribution of needle reports aligns closely with the geospatial distribution of overdose deaths in SF in 2016 and 2017 which were concentrated in the Tenderloin, SoMa, and Mission neighborhoods.20 These findings further support an emerging link between the location of publicly discarded needles and the risks associated with public injection drug use. Furthermore, there was a strong relationship between needle reports with the location of homeless shelters, supporting the well-established link between public injection drug use and homelessness. However, an alternative reason for the increased needle reports in close proximity to homeless shelters could be due to shelter policy barring new or used needles from being brought inside, potentially leading to an increase in improperly discarded needles. These data suggest the need for additional harm reduction services tailored to unhoused and housing unstable populations. They also highlight homeless shelters as important and underutilized spaces for harm reduction interventions, including needle collection, syringe exchange, and low-threshold opportunities to enter addiction treatment. Future studies should aim to understand how such interventions, including changes to shelter policy around needles, could promote harm reduction and increase rates of proper needle disposal. The 311 database could serve as a powerful tool in understanding the relationship between discarded needles and areas of frequent public injection and associated overdose.

We also found that needle boxes and kiosks were generally located nearby other boxes and kiosks, respectively. These needle disposal sites tended to be in areas with a high density of needle reports, and 200m buffers around boxes and kiosks contained a disproportionately high percentage of needle reports in a relatively small area. Future studies should aim to identify the impact of installation of needle boxes and kiosks on the number of needle reports. Furthermore, there were several identified areas of high needle report density that did not have a needle box or kiosk. Utilizing 311 service data could inform future targeted sites for harm reduction.

An important limitation in using 311 service data is the risk of reporting bias. These potential biases include temporal and geographic changes in the knowledge about and use of the 311 service. In order to directly assess the risk for these potential biases, we compared the geospatial and temporal distribution of needle reports to bulky item reports. Our findings indicate that, while both needle reports and bulky item reports generally rose over time, there were important temporal and geographic differences in needle reports as compared to those of bulky items. These differences indicate that there may be meaningful temporal and geospatial information regarding public injection drug use embedded in 311 reports that should be further investigated in future studies. Another limitation of this study was that, due to the fact that some needle reports were marked as being within the shelter, we generated buffers around the centroid of shelters. Due to the variation in shelter size, this will result in underreporting the number of needles that would fall 200 meters from the larger shelters, since a larger proportion of the buffer would be occupied by the shelter itself.

Notably, San Francisco has implemented additional methods for collecting improperly disposed needles and syringes beyond the needle boxes and kiosks included in this study. In July 2018, the city signed a contract with the SF AIDS Foundation to create a syringe pick-up team.21 Needles collected through this program are reported through a separate portal on the SF AIDS Foundation Website, and may not be reflected in the 311 database. This may, in part, account for the decrease in 311 reports of discarded needles in 2019.

Our study is the first to describe public reporting of discarded needle data in San Francisco, using crowdsourced and publicly available 311 data. We believe these data could be used to understand the evolving trends in public injection drug use and more effectively target harm reduction services for opioid and stimulant use. Our study provides assessment of the temporal changes in the number of needle reports as a proxy for understanding changing trends in injection drug use, and uses reports of bulky items as a proxy to account for increased knowledge and usage of the 311 service. Furthermore, identifying areas with high density of needle reports affords a unique understanding of areas with high interaction between individuals reporting needles and those discarding them, areas which have the potential to be more fraught with police enforcement and potential increased public health risks as a consequence. Finally, our study has important implications regarding how publicly available and crowdsourced data can be used to inform public health interventions globally, and in particular how such databases can be used to bolster public health responses in low and middle income countries, where centralized data collection and data cleaning processes may not be available.

Crowdsourced and publicly available, 311 data shows promise as a tool to understand the overdose epidemic. Our study found that there was meaningful geospatial information regarding the overdose epidemic in 311 needle reports in SF, with needle reports clustered near areas of high social stress and coinciding with areas of high rates of overdose. While this study focused on SF, 311 services exist throughout the U.S. and similar studies should be conducted in other municipalities in order to guide the future implementation of geographic and temporal targeting of needed harm reduction services.

Highlights.

  • Discarded needle reports increased dramatically in San Francisco over 10 years

  • Downtown San Francisco had highest density of needle reports

  • Reports were clustered near needle disposal sites and homeless shelters

  • Needle reports may serve as a proxy for locations of public injection drug use

Acknowledgements:

Dr. Bearnot’s work is supported by NIH and NIDA under award number K12DA043490.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Works Cited

RESOURCES