Visualizing Neighborhood COVID-19 Levels, Trends, and Inequities in Wastewater: An Equity-Centered Approach and Comparison to CDC Methods

Abstract

Context:

Monitoring neighborhood-level SARS-CoV-2 wastewater concentrations can help guide public health interventions and provide early warning ahead of lagging COVID-19 clinical indicators. To date, however, U.S. Centers for Disease Control and Prevention’s (CDC) National Wastewater Surveillance System (NWSS) has provided methodology solely for communicating national and state-level “wastewater viral activity levels.”

Program:

In October 2022, the Boston Public Health Commission (BPHC) began routinely sampling wastewater at 11 neighborhood sites to better understand COVID-19 epidemiology and inequities across neighborhoods, which vary widely in sociodemographic and socioeconomic characteristics. We developed equity-centered methods to routinely report interpretable and actionable descriptions of COVID-19 wastewater levels, trends, and neighborhood-level inequities.

Approach and Implementation:

To produce these data visualizations, spanning October 2022 to December 2023, we followed four general steps: (1) smoothing raw values; (2) classifying current COVID-19 wastewater levels; (3) classifying current trends; and (4) reporting and visualizing results.

Evaluation:

COVID-19 wastewater levels corresponded well with lagged COVID-19 hospitalizations and deaths over time, with “Very High” wastewater levels coinciding with winter surges. When citywide COVID-19 levels were at the highest and lowest points, levels and trends tended to be consistent across sites. In contrast, when citywide levels were moderate, neighborhood levels and trends were more variable, revealing inequities across neighborhoods, emphasizing the importance of neighborhood-level results. Applying CDC/NWSS state-level methodology to neighborhood sites resulted in vastly different neighborhood-specific wastewater cut points for “High” or “Low,” obscured inequities between neighborhoods, and systematically underestimated COVID-19 levels during surge periods in neighborhoods with the highest COVID-19 morbidity and mortality.

Discussion:

Our methods offer an approach that other local jurisdictions can use for routinely monitoring, comparing, and communicating neighborhood-level wastewater levels, trends, and inequities. Applying CDC/NWSS methodology at the neighborhood-level can obscure and perpetuate COVID-19 inequities. We recommend jurisdictions adopt equity-focused approaches in neighborhood-level wastewater surveillance for valid community comparisons.

Introduction

Context

Monitoring levels, trends, and inequities in SARS-CoV-2 wastewater concentrations can provide an accurate proxy of population-level COVID-19 burden independent of testing practices or clinical severity, provide early warning ahead of lagging COVID-19 clinical indicators, and help guide equitable public health interventions.1-6 Viral concentrations in wastewater, measured in units of RNA copies/mL, represent pooled results from individuals contributing to the wastewater sampling area. Currently, there are no standard methods to translate wastewater concentration values into readily interpretable metrics, such as the number of infected cases or expected hospitalizations. This stems in part from the complex and variable relationship between viral wastewater concentrations and COVID-19 clinical indicators (ie, number of cases, hospitalizations, deaths), which is expected to vary over time and across settings based on numerous factors (eg, sewerage infrastructure, precipitation, population immunity and sociodemographics, and testing practices).4,7-9 In this way, wastewater viral concentration values have limited interpretability for public health without necessary details to contextualize and communicate these results, such as quantifying current levels relative to prior time points, and trends over time.

To date, existing standard protocols have focused on reporting results at broad geographies. For example, the U.S. Centers for Disease Control and Prevention’s (CDC) National Wastewater Epidemiology Surveillance System (NWSS), proposed in December 2023, offers technical documentation communicating national and state-level “wastewater viral activity levels.” CDC/NWSS methodology uses past-year concentration values at each NWSS sampling site to calculate site-specific thresholds to classify “wastewater viral activity levels” (Very High, High, Moderate, Low, and Minimal).10,11 This approach allows for consistent measurement and reporting of trends within NWSS sites, which vary widely by the laboratory and sampling processes employed. However, this approach does not allow for comparison of viral burden between locations, including those with the same laboratory and processing methods, since a viral activity level classified as “High” may be classified as “Moderate” elsewhere due to differing thresholds between sites. In the absence of alternative approaches, local jurisdictions may look to adapt methods used in other settings, such as those implemented by CDC/NWSS. Whether the CDC/NWSS methodology can be used to monitor and detect inequities in wastewater viral activity at the neighborhood-level within cities has not been evaluated. Further work is needed to develop methods that allow for meaningful for monitoring of levels, trends, and inequities between sites in neighborhood-level wastewater data.

Program

Since March 2020, the Massachusetts Water Resources Authority (MWRA) has monitored wastewater for SARS-CoV-2 in the greater-Boston metro area.12 MWRA data represent composite results from the 43 cities and towns covered by MWRA’s Deer Island Wastewater Treatment Plant (WWTP) and does not provide data on specific municipalities in the greater-Boston area or specific neighborhoods within cities, and hence cannot be used to monitor city-level and neighborhood-level inequities.

To supplement MWRA regional data, in October 2022, the Boston Public Health Commission (BPHC) began testing wastewater for COVID-19 at 11 neighborhood sampling sites to better understand COVID-19 epidemiology and inequities across Boston neighborhoods.13 In total, the 11 sampling sites covered approximately 18% of Boston’s population (119 178 residents), with individual sites ranging in estimated population from 1869 to 34 412 people.14 Samples were collected twice weekly and reported as effective concentrations normalized by the human fecal biomarker, pepper mild mottle virus (PMMoV), to account for differences in population and environmental factors (eg, precipitation) between sites and over time.15,16 Results from the first year of this program demonstrated large inequities in SARS-CoV-2 concentrations across neighborhoods, with patterns consistent with longstanding health and social inequities.14 Building on these results, BPHC, in collaboration with academic partners, aimed to develop methods for routinely reporting wastewater concentration levels, trends, and inequities that are both interpretable and actionable at the neighborhood-level.

In this public health practice report, we present our equity-oriented methodology developed for neighborhood-level wastewater data analysis and reporting in the city of Boston. To validate our approach, we compared the estimated COVID-19 wastewater levels and trends to neighborhood COVID-19 clinical indicators. Finally, we compared our methodology to the CDC/NWSS framework introduced in December 2023 for reporting on national- and state-level wastewater viral trends and levels.

Approach and Implementation

Neighborhoods play an important role in influencing health, and granular, neighborhood-level data are crucial for guiding local public health efforts to advance health equity.17-20 Boston remains economically and racially segregated, and social and environmental factors that structure risk for SARS-CoV-2 infection and severe COVID-19 vary widely across Boston neighborhoods (Supplementary Digital Content, Table S1, available at: http://links.lww.com/JPHMP/B425).14 A focus on neighborhood-level health equity was thus central to our approach—both in data visualizations to communicate levels, trends, and inequities, and linking these results to neighborhood-specific recommendations and resources. Therefore, we aimed to balance methodological rigor with simplicity and interpretability. These methods were developed collaboratively and iteratively via input from community organizations, clinical partners, internal BPHC working groups, and academic public health researchers.

Data reporting, visualization, and communication

Figure 1 displays our approach to report current citywide wastewater COVID-19 levels and 2-week trends (ie, the population-weighted average across sites) alongside neighborhood levels and trends. In this visualization (Figure 1A), we emphasized the qualitative description of the level and direction of trend for quick interpretation (eg, Level: “Very High”; 2-week trend: Increase), but also included quantitative results for each (eg, Level: 1864 copies/mL; 2-week trend: + 351 copies/mL [+23%]). The accompanying levels and trends map (Figure 1B) depicts each neighborhood site as a circle placed at the centroid of each catchment area, with the most recent COVID-19 wastewater level depicted by the color of the circle and the 2-week trend shown as an arrow icon within the circle. An additional table (Figure 1C) lists each neighborhood ordered from highest to lowest SARS-CoV-2 concentration and shows the corresponding level and trend in that neighborhood consistent with the map in Figure 1B. Inequities across neighborhoods are readily identifiable.

FIGURE 1.

Example of First Iteration of Updated SARS-CoV-2 Wastewater Levels and Trends Report Including (A) Boston Citywide COVID-19 Level and 2-week Trend* across All 11 Neighborhood Sites; (B) Map; and (C) Table Ranking and Providing Details About Current SARS-CoV-2 Levels and Trends for Each Neighborhood Sampling Site. This figure is available in color online (www.JPHMP.com).

*2-week trend categories: large increase: ≥+500 copies/mL; increase: +150 to +500 copies/mL; stable: −150 to +150 copies/mL; decrease: −500 to −150 copies/mL; large decrease: < −500 copies/mL.

Because BPHC’s wastewater program emphasizes neighborhood-specific results, we additionally presented results specific to each neighborhood with accompanying text to interpret levels and trends (Figure 2A). For each neighborhood, the graph shows the smoothed time series overlayed over the raw sample values for that neighborhood as well as the citywide average trend over time (Figure 2A). The text provides data on the COVID-19 level; its smoothed concentration value; whether the neighborhood level is higher, similar, or lower to the citywide average; and the direction of the current 2-week trend and both the percentage and absolute change of the current value compared to 2 weeks prior (Figure 2A).

FIGURE 2.

(A) Example of Neighborhood-Specific Results and Interpretation Included in the First Iteration of Updated COVID-19 Wastewater Levels and Trends Report for the Dorchester Neighborhood Sampling Site; (B) Example of Recommendations and Resources Based on COVID-19 Level Included in the First Iteration of Updated COVID-19 Wastewater Levels and Trends Report; Results Shown are for the “Very High” Level.This figure is available in color online (www.JPHMP.com).

To guide neighborhood-specific action, the visualization also provided a link to recommendations and resources based on the current level in that neighborhood (Figure 2B). Recommended actions for any given level—including those specific to each COVID-19 wastewater level and general guidance recommended across all levels (eg, staying up-to-date on vaccinations)—were adapted from CDC, Massachusetts Department of Public Health (MA DPH) and internal BPHC guidance; the links to resources include neighborhood-specific vaccination locations, testing sites and treatment information, as well as links to more general educational materials about COVID-19 (Figure 2B).

BPHC’s website also provided detailed maps and tables containing further quantitative results for levels and trends (Supplementary Digital Content, Figure S2, available at: http://links.lww.com/JPHMP/B425). Key differences between detailed results and simplified results in Figure 1 are: (1) the map uses a continuous color scale to depict the current concentration at each neighborhood site rather than a discrete color scale representing the five risk levels; (2) the accompanying table includes corresponding quantitative values for both current COVID-19 level and 2-week trend; and (3) the colored circles are sized proportionately to the size of population covered.

Since December 2023, BPHC has posted these general and detailed visualizations and data on neighborhood-level COVID-19 wastewater levels and trends in twice-weekly reports and shared these reports with BPHC’s partners (eg, community organizations, Boston hospitals, Community Health Centers, and Boston Public Schools).

Data processing and analysis steps

Underlying our equity-centered methodology are four data processing and analytic steps, involving: (1) smoothing raw values; (2) classifying current COVID-19 wastewater levels; (3) classifying current trends; and (4) reporting and visualizing results. These methods were designed to be easily implemented by local health practitioners using standard software; additional methodological details can be found in BPHC’s accompanying technical documentation.21

Step 1: smoothing

Wastewater data are inherently variable due to factors such as daily fluctuations in flow rates and differences in population size between sites.22 Thus, wastewater values are often smoothed (eg, using moving averages) to describe and visualize levels and temporal trends in wastewater concentrations to address sources of variability.23

To smooth the data, we applied smoothing splines to SARS-CoV-2 effective concentrations, with parameters chosen to approximate three-sample (8-day) trailing average values.21,24,25 This procedure produced visually smoothed estimates that were recent enough to reflect current values, and allowed for comparison between greater-Boston metro area wastewater data and other COVID-19 clinical indicators, reported as 7-day trailing averages (Supplementary Digital Content, Figure S3, available at: http://links.lww.com/JPHMP/B425).21

Step 2: COVID-19 wastewater-level classification

After smoothing, wastewater concentration values were categorized into qualitative levels—Very High, High, Moderate, Low, and Very Low—to help with interpretation. Because there was no established set of values/cut points to define these levels across jurisdictions, we established cut points using the distribution of values observed at the city-level over the first year of the program (ie, October 2022 to September 2023) to allow BPHC to prepare for the upcoming respiratory season. Unlike in prior years, COVID-19 rates over the first year of the wastewater program varied widely, with periods of very low transmission (eg, summer) and periods of high levels of SARS-CoV-2 transmission (in winter). In addition, BPHC’s citywide wastewater concentrations over the first year closely tracked with trends observed in regional wastewater data for the greater-Boston area (MWRA) and were strongly correlated with subsequent trends in clinical indicators with varying lead-times over the first year of the program (Supplementary Digital Content, Table S4, available at: http://links.lww.com/JPHMP/B425).14

As starting values for cut points between levels, we used quintiles observed over the first year citywide. We then adjusted these quintile cut points to ensure an equal range of values within each level to preserve interpretation/context of levels relative to each other.21 Differences between the initial and adjusted categories were slight, with the latter improving interpretability. In general, these adjusted levels reflect time periods of known low and high SARS-CoV-2 concentration, with the highest levels observed during winter months and lowest observed during May/June (Supplementary Digital Content, Figure S5, available at: http://links.lww.com/JPHMP/B425). In addition, wastewater-level categories showed good concordance with lagged clinical indicators (see Supplementary Digital Content for additional details, http://links.lww.com/JPHMP/B425). “Very High” wastewater levels generally corresponded with peak levels of COVID-19 hospital admissions, inpatient hospitalizations, and deaths observed during winter surges with wastewater levels leading corresponding clinical indicators (Supplementary Digital Content, Figure S6A-C, available at: http://links.lww.com/JPHMP/B425).

Step 3: trends classification

After smoothing and classifying COVID-19 wastewater concentrations into levels, our final step was to quantify the current magnitude and direction of change in neighborhood wastewater concentrations over the past 14 days (this was not designed to predict future levels). We calculated the difference between the current smoothed estimate and the smoothed estimate from the prior 14 days (described in Step 1), as this difference can be easily understood and interpreted as either increasing, decreasing, or stable over that time period.21

We categorized the differences in smoothed estimates into five categories:

• Large Increase: ≥+ 500 copies/mL,

• Increase: +150 to +500 copies/mL,

• Stable: −150 to +150 copies/mL,

• Decrease: −500 to −150 copies/mL,

• Large Decrease: <−500 copies/mL.

These categories allowed us to distinguish between large increases observed at the start of respiratory disease season and smaller, steadier increases observed during other times (eg, late summer/back-to-school season) (Supplementary Digital Content, Figure S7, available at: http://links.lww.com/JPHMP/B425).21

Evaluation

We conducted several descriptive analyses to: (1) assess whether our approach accurately reflected COVID-19 epidemiology and neighborhood-level inequities in COVID-19 disease burden, and (2) compared our methodology to the COVID-19 wastewater level monitoring methodology introduced by CDC/NWSS in December 2023.

Levels and trends over time

Generally, neighborhood levels and trends varied both across neighborhoods and over time. When citywide and regional wastewater SARS-CoV-2 concentrations were at their highest (December/January) and lowest (May/June), levels and trends observed across sites were relatively consistent (Figure 3). For example, on December 21, 2022, COVID-19 wastewater levels were “Very High” or “High” across all 11 sites, and most sites had increasing trends (Figure 3A). Similarly, on May 7, 2023, COVID-19 levels were “Very Low” with “stable” trends across all 11 sites (Figure 3B). In contrast, during periods where citywide and regional levels were not at their highest and lowest, there was more variation in COVID-19 levels and trends observed across neighborhoods sites (Figure 3C).

FIGURE 3.

COVID-19 Wastewater Levels and Trends Across Sites Over Time at Three Selected Timepoints. (A) December 21, 2022 (High COVID-19 Levels Citywide),* (B) May 7, 2023 (Low COVID-19 Levels Citywide), and (C) September 17, 2023 (Moderate COVID-19 Levels Citywide)

*In Panel A, on December 21, 2022, Roslindale/West Roxbury (RS/WR) and Mattapan (MT) neighborhood sampling sites had just begun sample collection and therefore did not have 2-week trend information available to display; 2-week trend categories: large increase: ≥+500 copies/mL; increase: +150 to +500 copies/mL; stable: −150 to +150 copies/mL; decrease: −500 to −150 copies/mL; large decrease: <−500 copies/mL; neighborhood abbreviations can be found in table in Figure 1C.This figure is available in color online (www.JPHMP.com).

Comparison to CDC/NWSS methodology

Finally, to assess whether CDC/NWSS methodology can be applied to neighborhood-level wastewater data, we compared the BPHC methodology described above to the CDC/NWSS methodology. Both BPHC and NWSS methodology classified “wastewater viral activity levels” into five categories (very high, high, moderate, low, and minimal). According to NWSS, “…the value associated with the Wastewater Viral Activity Level is the number of standard deviations above the baseline, transformed to the linear scale,” then categorized into five levels (Minimal: <1.5, Low: [1.5, 3.0], Moderate: [3.0, 4.5], High: [4.5, 8.0], very high: ≥8.0).11 However, both baseline values and standard deviation used are calculated separately for each NWSS sampling site using data over the past year (and re-calculated every 6 months in July and January), resulting in a different set of cut points for viral activity levels for each NWSS site (Supplementary Digital Content, Table S8, available at: http://links.lww.com/JPHMP/B425). By design, catchment areas with higher levels of viral activity in the previous year (and higher COVID-19 morbidity and mortality) had higher cut points for the five levels.

When applied to BPHC’s neighborhood sites, CDC/NWSS cut points using data through July 2023 varied widely across neighborhood sampling sites (Supplementary Digital Content, Figure S9, available at: http://links.lww.com/JPHMP/B425). For example, the CDC/NWSS threshold for “Very High” COVID-19 levels was 3.5 times higher at the Allston-Brighton neighborhood site compared to the Charlestown site (respectively, ≥2918 copies/mL and ≥822 copies/mL). Consequently, a concentration of 850 copies/mL would be considered “Very High,” in Charlestown, but only “moderate” at five other neighborhood sites that experienced higher COVID-19 burdens over the past year. Neighborhoods that had higher thresholds for “Very High” COVID-19 levels also tended to be neighborhoods with higher values on the CDC/ASTDR Social Vulnerability Index (SVI) (Supplementary Digital Content, Table S10, available at: http://links.lww.com/JPHMP/B425). Conversely, the CDC/NWSS cut points signaled higher levels of concern at lower viral concentrations than BPHC cut points in neighborhoods with lower SVI values.

CDC/NWSS wastewater levels were consistent with BPHC levels for 41% of samples (490/1189) across all sites; however, the correspondence ranged across sites (Figure 4; Supplementary Digital Content, Table S10, available at: http://links.lww.com/JPHMP/B425). In 38% (454/1189) and 11% (134/1189) of samples, the CDC/NWSS level was one and two levels higher than the BPHC level, respectively; this was most common in summer months during periods of lower SARS-CoV-2 transmission (Figure 4; Supplementary Digital Content, Table S10, available at: http://links.lww.com/JPHMP/B425). In contrast, BPHC levels were higher than CDC/NWSS levels for 9% of samples (111/1189), largely occurring during winter surge periods and only observed in 6 of 11 neighborhood sampling sites. In five neighborhood sampling sites—Allston-Brighton, Roxbury, Hyde Park, Mattapan, and Dorchester—BPHC levels were higher than CDC/NWSS levels for more than 10% of all samples. For these neighborhoods, CDC/NWSS methods lagged behind BPHC levels in indicating “very high” levels at the start of the winter 2022-2023 surge (Supplementary Digital Content, Table S11, available at: http://links.lww.com/JPHMP/B425). For example, for two neighborhood sampling sites—Roxbury and Dorchester—“Very High” levels using CDC/NWSS methodology were not observed until 28 and 31 days after “very high” BPHC levels, respectively.

FIGURE 4.

Comparison of BPHC’s Methodology for Calculating Wastewater Level to Methods Used by CDC/NWSS. (A) Levels Calculated Using CDC/NWSS Methodology Compared to Levels Calculated Using BPHC Methodology Across 11 Neighborhood Sampling Sites and (B) Difference between CDC/NWSS Levels and BPHC Levels. This figure is available in color online (www.JPHMP.com).

In panel (B), orange colors indicate sampling dates when the CDC/NWSS level was lower than the corresponding BPHC level, and purple colors indicate sampling dates when CDC/NWSS levels were higher than the corresponding BPHC level. Neighborhood sampling sites are ordered by their threshold for “very high” activity level from highest (Allston-Brighton [A/B]) to lowest (Charlestown [CH]).* In panel (B), colors using “>>” and “<<” denote two levels difference, and columns using “>” and “<” indicate one level difference. Summary statistics describing agreement between BPHC and CDC/NWSS methodology can be found in Supplementary Digital Content, Table S10, available at: http://links.lww.com/JPHMP/B425.*Neighborhood abbreviations can be found in table in Figure 1C.

Importantly, these neighborhoods, where CDC/NWSS methodology consistently underestimated COVID-19 levels and signaled a later start to the winter surge, experienced some of the highest COVID-19 morbidity and mortality. These neighborhoods also exhibit higher levels of social conditions that increase the risk for SARS-CoV-2 exposure and severe COVID-19 outcomes (Table 1). Neighborhoods where CDC/NWSS methods consistently underestimated COVID-19 levels were in the 86th percentile of SVI, compared to the 66th percentile in other Boston neighborhoods, with higher values also observed across three of four SVI subdomains (socioeconomic status, household characteristics, and racial and ethnic minority status) (Table 1). Compared to other Boston neighborhoods, neighborhoods where CDC/NWSS levels systematically underestimated COVID-19 burden had higher percentage of Black, Indigenous and other people of color, lower levels of educational attainment, higher household poverty and lower median income, more crowded and rent-burdened households, more residents without health insurance or insured through Medicaid, and a higher percentage of residents working in-person and relying on public transit to get to work (Table 1).

TABLE 1.

Sociodemographic Characteristics for Neighborhoods Where BPHC Wastewater Levels Exceed CDC/NWSS Levels for More than 10% of Observations Compared to All Other Boston Neighborhoods

	Boston (Citywide)	Neighborhoods with Levels BPHC > CDCa	Other Boston Neighborhoods
2020 Census Redistricting Data
Total population	675 647	313 968	361 679
Race/Ethnicity (U.S. Census categories)
American Indian and Alaska Native	989 (0.1%)	600 (0.2%)	389 (0.1%)
Asian	75 588 (11.2%)	33 401 (10.6%)	42 187 (11.7%)
Black or African American	129 264 (19.1%)	102 168 (32.5%)	27 096 (7.5%)
Hispanic or Latino	126 113 (18.7%)	64 574 (20.6%)	61 539 (17.0%)
Native Hawaiian and Other Pacific Islander	251 (0.0%)	110 (0.0%)	141 (0.0%)
Some Other Race	9257 (1.4%)	6334 (2.0%)	2923 (0.8%)
Population of two or more races	32 721 (4.8%)	19 271 (6.1%)	13 450 (3.7%)
White	301 464 (44.6%)	87 510 (27.9%)	213 954 (59.2%)
Group Quarters (GQ)
Total GQ population	46 574 (6.9%)	7498 (2.4%)	39 076 (10.8%)
Institutionalized GQ population	5196 (0.8%)	1450 (0.5%)	3746 (1.0%)
Adult correctional facilities	915 (0.1%)	0 (0.0%)	915 (0.3%)
Juvenile facilities	705 (0.1%)	573 (0.2%)	132 (0.0%)
Nursing facilities/Skilled-nursing facilities	3257 (0.5%)	821 (0.3%)	2436 (0.7%)
Other institutional facilities	319 (0.0%)	56 (0.0%)	263 (0.1%)
Non-institutionalized GQ population	41 378 (6.1%)	6048 (1.9%)	35 330 (9.8%)
College/University student housing	35 750 (5.3%)	4005 (1.3%)	31 745 (8.8%)
Military quarters	146 (0.0%)	0 (0.0%)	146 (0.0%)
Other non-institutional facilities	5482 (0.8%)	2043 (0.7%)	3439 (1.0%)
2021 American Community Surveyb
Total population	672 814	317 402	355 412
Educational Attainmentc
Total aged ≥25 years	466 959	216 535	250 424
Less than high school diploma	55 289 (11.8%)	30 019 (13.9%)	25 270 (10.1%)
High school diploma or Associate’s degree	168 221 (36.0%)	100 110 (46.2%)	68 111 (27.2%)
Bachelor’s degree or higher	243 449 (52.1%)	86 407 (39.9%)	157 042 (62.7%)
Household Poverty & Incomed
Population in households	630 121	310 389	319 732
Income below the poverty level	110 995 (17.6%)	63 108 (20.3%)	47 887 (15.0%)
Median household income	$85 451	$62 847	$104 539
Household Crowdinge
Occupied housing units	271 950	119 844	152 106
>1 occupant per room	9514 (3.5%)	4430 (3.7%)	5084 (3.3%)
1.01 to 2.0 occupants per room	629 (0.2%)	293 (0.2%)	336 (0.2%)
2.01 or more occupants per room	8885 (3.3%)	4136 (3.5%)	4749 (3.1%)
≤1 occupant per room	262 436 (96.5%)	115 414 (96.3%)	147 022 (96.7%)
Household Rent Burdenf
Renter-occupied housing units	177 371	80 856	96 515
Household rent costs ≥30% income	82 807 (46.7%)	41 106 (50.8%)	41 701 (43.2%)
Household rent costs 30-50% income	39 735 (22.4%)	18 475 (22.8%)	21 260 (22.0%)
Household rent costs ≥50% income	43 72 (24.3%)	22 631 (28.0%)	20 441 (21.2%)
Household rent costs <30% income	85 567 (48.2%)	35 696 (44.1%)	49 871 (51.7%)
Household rent costs not reported	8997 (5.1%)	4054 (5.0%)	4943 (5.1%)
Health Insurance Coverageg
Total non-institutionalized population	667 612	315 987	351 625
No health insurance coverage	22 847 (3.4%)	13 267 (4.2%)	9580 (2.7%)
Employer-based health insurance	338 489 (50.7%)	137 322 (43.5%)	201 167 (57.2%)
Direct-purchase health insurance	48 112 (7.2%)	18 666 (5.9%)	29 446 (8.4%)
Medicare	23 335 (3.5%)	12 458 (3.9%)	10 877 (3.1%)
Medicaid/means-tested public insurance	135 860 (20.4%)	81 625 (25.8%)	54 235 (15.4%)
VA, TRICARE, or military health insurance	1954 (0.3%)	1214 (0.4%)	740 (0.2%)
Two more types of health insurance	97 015 (14.5%)	51 435 (16.3%)	45 580 (13.0%)
Transportation to Workh
Workers aged 16 years and over	365 746	163 383	202 363
Public transportation (excluding taxicab)	102 376 (28.0%)	50 132 (30.7%)	52 244 (25.8%)
Worked from home	44 981 (12.3%)	14 650 (9.0%)	30 331 (15.0%)
Other transportation to work	218 389 (59.7%)	98 601 (60.3%)	119 788 (59.2%)
2020 CDC/ATSDR Social Vulnerability Indexi
Overall Social Vulnerability Index (SVI)	0.757	0.865	0.663
SVI Theme/Domainj
Socioeconomic Status	0.635	0.789	0.501
Household Characteristics	0.496	0.681	0.336
Racial and Ethnic Minority Status	0.829	0.904	0.764
Housing Type and Transportation	0.891	0.884	0.896

^aNeighborhoods where BPHC wastewater levels exceeded CDC’s wastewater levels for more than 10% of observations; Allston/Brighton, Hyde Park, Roxbury, Mattapan, and Dorchester.

^bMost recent available data: American Community Survey 5-year Estimates, 2017-2021.

^cAmong population aged ≥25 years.

^dAmong populations in households, over the last 12 months.

^eAmong occupied housing units.

^fAmong renter-occupied housing units.

^gAmong non-institutionalized population.

^hAmong workers aged 16 years and older.

ⁱThe U.S. Centers for Disease Control and Prevention (CDC) and the Agency for Toxic Substances and Disease Registry (CDC/ATSDR) Social Vulnerability Index (SVI) is a place-based measure developed for use in health and disaster preparedness35,36; values are percentiles based on distribution of U.S. Zip Code Tabulation Areas (ZCTAs) and range from 0 to 1, with higher values indicating higher levels of vulnerability, both overall and for subdomains.

^jSubdomain categories and names listed as reported by CDC/ATSDR. See CDC/ASTDR documentation36 for variables included in each subdomain.

Discussion

This public health practice report, using wastewater data from across 11 neighborhoods in Boston, Massachusetts, describes a new method to categorize and communicate neighborhood-level SARS-CoV-2 levels and trends. This method produces interpretable and actionable results, including monitoring of neighborhood inequities in SARS-CoV-2 burden. Our approach includes steps to smooth, classify, visualize, and communicate interpretable COVID-19 levels and 2-week trends linked to neighborhood-specific public health actions and resources (eg, vaccinations, testing, and treatment). This equity-centered approach may be especially applicable to other jurisdictions implementing wastewater SARS-CoV-2 monitoring at the neighborhood level. In contrast to our approach, applying CDC/NWSS methodology to neighborhood sites resulted in vastly different thresholds for wastewater levels between neighborhoods, which obscured inequities between neighborhoods, and systematically underestimated COVID-19 levels during surge periods in neighborhoods with the highest COVID-19 morbidity and mortality.

In descriptive analyses, we found that the COVID-19 wastewater levels corresponded well to lagged clinical indicators, including COVID-19 hospital admissions and COVID-19 deaths, with “Very High” COVID-19 wastewater levels generally corresponding to winter surges in severe COVID-19 outcomes. Our results, alongside prior work, suggest while the relationship between wastewater clinical indicators may change over time, wastewater levels may still be useful in anticipating and preventing increases in severe COVID-19 outcomes.16,26-28

During periods of high and low transmission citywide, there were relatively consistent levels and trends across neighborhood sites. However, when citywide COVID-19 levels were not at their highest and lowest points, we observed substantial variation across sites. This variation would otherwise be obscured in WWTP-level data, emphasizing the importance of neighborhood-level sampling in understanding COVID-19 epidemiology and neighborhood-level inequities.

Our novel comparison of our equity-centered methods to those employed by the CDC/NWSS to report national and state COVID-19 levels revealed important discrepancies when applied at the neighborhood-level. Specifically, for nearly half of BPHC’s neighborhood sites, applying the CDC/NWSS methodology systematically underestimated neighborhoods’ COVID-19 burden and lagged behind BPHC methodology in signaling “very high” levels during the 2022-2023 winter surge. Moreover, this underestimation largely occurred in neighborhoods that have jointly (1) borne the greatest burden of COVID-19 morbidity and mortality, and (2) had higher prevalence of crowded housing conditions, precarious working conditions, and higher reliance on public transportation. These inequities in infectious disease outcomes are maintained in part by structural racism in the form of racialized socioeconomic residential segregation, which concentrates conditions that increase risk of both infectious disease exposure and severe COVID-19 outcomes.29-31

Methodological approaches for wastewater data that allow for meaningful comparisons between locations with different levels of SARS-CoV-2 are, as of now, lacking and are thus urgently needed to understand and mitigate inequities. Yet, as our comparisons of methods reveals, the CDC/NWSS approach using thresholds specific to each sampling site should be avoided at the neighborhood-level, as it impedes the assessment of neighborhood-level inequities and effectively sets different standards for concern across neighborhoods, which undermines equitable policymaking and resource prioritization. However, creating an interpretation of risk conditional on broad levels of geography (eg, “Very High” levels “for a given site”) risks perpetuating structural injustices via statistical methodology.32 The impact of this methodological choice is similar to prior CDC/NCHS weighting methods that consistently underestimated racialized inequities COVID-19 mortality, by factoring out key neighborhood-level influences that drive health inequities, including racialized economic residential segregation.33 The net implication is that applying CDC/NWSS methodology, at not only the neighborhood-level but at any sub-national level (eg, state, county, or city), has the potential to exacerbate inequities via crucial delays in recommendations for increased COVID-19 precautions and/or mismatched resource allocation. These potential applications underscore the importance of continuously involving impacted communities in decision-making—including during planning, implementation, data analysis, and communication and actioning of results—to ensure that data are used ethically, aligned with community priorities, and to mitigate misuse including stigmatization and punishment.8,34

A limitation of our methodology is that it assumes different areas have the same proportional relationship between SARS-CoV-2 wastewater levels and COVID-19 burden; we believe this is a reasonable assumption within the neighborhoods of a single city but might not be reasonable for comparing larger areas, such as states. In addition, our approach also requires that laboratory processing methods and factors influencing SARS-CoV-2 wastewater concentrations are uniform across sites. While likely partly addressed through normalization with PMMoV, differences in SARS-CoV-2 concentrations observed might also be influenced by infrastructural, environmental, and sociodemographic/epidemiologic factors (eg, features of the sewer network, rainfall, and differences in viral shedding). Looking ahead, BPHC plans to routinely revisit this methodology, given potential for new variants and changes in population immunity (due to both vaccination and infection) and growing abandonment of non-pharmacological interventions.

Conclusions

The methods we describe to report neighborhood-level wastewater levels, trends, and inequities in Boston enabled a valid and meaningful equity-centered comparison across neighborhood sites, which may be applicable to other jurisdictions implementing similar programs. As we demonstrated, applying the CDC/NWSS methodology at the neighborhood-level (and potentially also other sub-national geographic levels) risks masking COVID-19 inequities and therefore perpetuation of inequities in these neighborhoods. We recommend that jurisdictions implementing neighborhood-level wastewater sampling use equity-centered methods that allow for valid and readily interpretable comparison of levels and trends across sites, in ways that communicate results clearly alongside actionable steps to reduce risk.

Implications for Policy & Practice

• This public practice report presents novel, equity-centered methods to smooth, classify, visualize, and communicate readily interpretable COVID-19 wastewater levels and 14-day trends.

• Our neighborhood-level focused methods surfaced inequities across sites and show how localized data can be used to guide neighborhood-specific public health actions and resource allocation (eg, vaccinations, testing, and treatment).

• Our study additionally shows how adapting the standardized CDC/NWSS methodology to the neighborhood level tended to underestimate and obscure local inequities in COVID-19 disease burden, particularly in areas with higher social vulnerability, and therefore should be avoided.

• Our study therefore underscores the importance of neighborhood-level wastewater-based epidemiology and calls for the establishment of standardized methods at the neighborhood scale to guide localized public health intervention efforts.

• Further research is needed to validate our neighborhood-level approach in other communities and contexts.