Edited by: Federica Pirrone, University of Milan, Italy
Reviewed by: Margaret Ruth Slater, Texas A&M University, United States; Louise Anne Buckley, University of Edinburgh, United Kingdom; Linda Jacobson, Toronto Humane Society, Canada
This article was submitted to Veterinary Humanities and Social Sciences, a section of the journal Frontiers in Veterinary Science
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
In this retrospective, exploratory study, intake and outcome data were compiled from 1,373 U.S. animal shelters for which such data were reported consistently across a five-year study period (2016–2020). Linear regression analysis was used to examine the five-year trends and the impacts of the first year of the COVID-19 pandemic (2020) on the overall trends in intake and outcomes in U.S. animal shelters. The results of the analysis reveal that total intake and euthanasia for both dogs and cats significantly decreased over the study period. The adoption, return-to-owner, return-to-field, and transfer (for cats) categories as a percentage of intake all showed significant increases. Live release rates as a function of total intakes and total outcomes for both dogs and cats showed significant increases over the study period. The findings from this study address a critical gap in the field by summarizing emerging trends at the national level in how cats and dogs are being served in U.S. animal shelters.
The efforts to standardize data collection on sources of intake and outcomes for animals in the care of U.S. animal shelter and rescue organizations have been underway since the 1980s. However, relatively little research to date has assessed the overall trends in intake and outcomes nationally. This gap in data collection and comprehensive program evaluation is particularly notable given the growing recognition of the importance of pet keeping on individual and community health and the increasing number of essential pet support services these organizations offer to their communities.
When animal sheltering began in the United States, as early as 1866 with the founding of the American Society for the Prevention of Cruelty to Animals in New York City, animal control efforts were concentrated on removing dogs and cats from city streets to reduce the threat of rabies (
Recognizing that such uncertainty made it difficult, if not impossible, to measure improvements in a particular program's effectiveness—and in the animal sheltering system's capacity to support the community's animals, more generally—researchers called for a more careful accounting of shelter data. In a 1992 editorial, Rowan (
Although many of the larger animal shelters were beginning to keep “comprehensive statistics on the number of animals handled and their disposition (euthanasia, adoption, and redemption),” the practice was not universal (
As recently as 2008, Scarlett (
“progress has been made toward eliminating Rowan's ‘statistical black hole'… basic data still elude us, including: the actual number of animal shelters in the country, national shelter estimates of impoundments and dispositions (euthanasia, adoption, redemption), and the effectiveness of programs (e.g., spay/neuter, adoption counseling) in reducing euthanasia.”
An important step in satisfying the need for “basic data” was addressed in 2004 with the adoption of the Asilomar Accords by industry leaders who agreed to a series of definitions that would “provide a standard way to categorize the dogs and cats who comprise the shelter population of the various organizations each year” (
To build on the efforts of SAC, Best Friends Animal Society (BFAS) began compiling shelter data in 2016, an effort that first necessitated the identification of thousands of organizations across the country considered to be animal shelters (see definition below). Within the BFAS dataset, shelter metrics from SAC were combined with those shared directly with BFAS, as well as those obtained from other sources (e.g., public records) (
Intake and outcome data were obtained from SAC and BFAS to generate a nationally representative sample of animal sheltering organizations that reported intake and outcome data consistently over the study period of 2016–2020. Since all data were publicly available or obtained from the organizations with permission to use for research and evaluation purposes, no human subject protection oversight or other forms of ethical approvals were required. For the purposes of this study, a
The number of dogs and cats taken into the shelters and the outcomes for those animals were collected according to the industry standards established through the Shelter Animals Count Basic Animal Data Matrix (
The categories for outcomes included euthanasia: animals that were euthanized by the facility other than those categorized as owner-intended euthanasia or other non-live outcomes; adoption: animals that were adopted, having permanently left the agency's possession, including barn cat programs resulting in adoption (this does not include animals in foster care or “trial” stays); returned to owner (RTO): stray or owner relinquished animals who are returned to their legal owner; transferred out: animals that were transferred to another facility, either locally or in a different state or territory; returned to field (RTF): animals included in intake, already altered, or altered after intake, and returned to stray capture location to be released (often referred to as shelter-neuter-return or SNR); and other non-live outcome: animals that died in care, were lost in care, or were euthanized as a result of an owner-intended euthanasia request; and other live outcomes: live outcomes not captured in the above (not captured in the outcome subtypes, an example would be the barn cat programs in some shelters).
The best estimates identify 4,400 animal shelters across the U.S.; however, data were available from SAC and BFAS for only 3,330 of these organizations (76%) during 2020 and datasets for previous years included fewer shelters (
Trends in the aggregated total, intake subtype, and outcome subtype data (e.g., stray, owner relinquished, adoption, and euthanasia) for the described metrics across the five-year study period were identified by linear regression analysis. Prior to conducting analyses, the assumptions of linear regression were tested for our count and percentage/ratio data by examining normal Q–Q plots, scale location plots, and residual leverage plots. It was determined that the data met the assumptions of linear regression. This exploratory analysis assessed simple increases or decreases over time with the assumption that systemic heteroscedasticity was not an issue and that any changes over the interval were primarily monotonic.
To illustrate the linear nature of the intake and outcome data,
To acknowledge the possibility that the COVID-19 pandemic, which began in 2020, may have skewed shelter operations, beginning in 2020, outliers in the dataset were determined using a two-sided Dixon outlier test. An additional linear regression analysis was then completed to explore the trends in the data with any data categories containing significant outliers in 2020 removed for analysis.
No adjustments for the multiplicity of testing were incorporated owing to the exploratory nature of the analyses; thus, the overall type I error could have been greater than the α value (i.e., 0.05) for individual tests. The total change, final value (predicted year 2020 value), and percentage change for the trend line over the study period were calculated for all trends that had slopes significantly different from 0. Total change was calculated as the slope multiplied by the number of years of change that were analyzed (5 years) in the regression analysis. The final value of the trend line was calculated by adding the total change to the y-intercept, and percentage change was calculated as 100 times the total change divided by the y-intercept. For trend lines with slopes not significantly different from 0, the final value was assumed to be the same as the y-intercept.
Data were reported primarily as predicted values from linear regression analyses (not as actual/observed values). This was done because our interest was in examining trends over the entire five-year study period rather than year-to-year changes; reporting only observed values can be misleading since doing so ignores year-to-year variation and may not account for baseline data. Therefore, values for the first year in each analysis (2016) were reported as the predicted y-intercept for the linear regression, and values for subsequent years were calculated from the y-intercept and slope.
Trends in intake were assessed for all animals, dogs, and cats, based on total intake and intake subtype categories (
Results of linear regression analysis of total intake and outcome data reported by animal shelter organizations in the United States (2016–2020).
Dogs | −117,771 | 0.07 | 1,704,366 | NC | 1,233,284 | NC |
Cats | −55,833 | 0.23 | 1,483,317 | NC | 1,259,985 | NC |
−226,154 | 0.07 | 3,473,481 | NC | 2,568,864 | NC | |
Dogs | −110,818 | 0.06 | 1,507,634 | NC | 1,064,361 | NC |
Cats | −64,251 | 0.15 | 1,362,842 | NC | 1,105,838 | NC |
−51,342 | 0.35 | 1,489,711 | NC | 1,284,342 | NC | |
Dogs | −37,682 | 0.13 | 658,880 | NC | 508,151 | NC |
Cats | −15,381 | 0.47 | 635,547 | NC | 574,022 | NC |
−21,860 | 0.31 | 681,063 | NC | 593,622 | NC | |
Dogs | −16,015 | 0.15 | 308,526 | NC | 244,468 | NC |
Cats | −7,865 | 0.25 | 295,354 | NC | 263,895 | NC |
10,091 | 0.20 | 164,953 | 40,364 | 205,316 | NC | |
Dogs | −629 | 0.83 | 86,731 | −2,515 | 84,216 | NC |
Cats | 4,792 | 0.32 | 58,838 | 19,167 | 78,005 | NC |
1,007 | 0.93 | 348,655 | NC | 352,682 | NC | |
Dogs | −6,952 | 0.36 | 196,733 | NC | 168,923 | NC |
Cats | 8,418 | 0.11 | 120,476 | NC | 154,147 | NC |
−37,782 | 0.46 | 1,638,195 | −151,128 | 1,487,067 | NC | |
Dogs | −39,154 | 0.15 | 748,958 | −156,615 | 592,343 | NC |
Cats | 6,529 | 0.71 | 690,344 | 26,114 | 716,458 | NC |
−14,240 | 0.32 | 418,055 | −56,959 | 361,096 | NC | |
Dogs | −15,027 | 0.16 | 319,418 | −60,107 | 259,311 | NC |
Cats | −361 | 0.78 | 36,172 | −1,442 | 34,730 | NC |
6,617 | 0.19 | 69,967 | 26,469 | 96,437 | NC | |
Dogs | 622 | 0.09 | 852 | 2,488 | 3,340 | NC |
Cats | 5,732 | 0.19 | 61,441 | 22,931 | 84,372 | NC |
−5,395 | 0.66 | 514,061 | −21,580 | 492,481 | NC | |
Dogs | −11,292 | 0.10 | 261,637 | −45,169 | 216,469 | NC |
Cats | 2,327 | 0.62 | 168,230 | 9,308 | 177,539 | NC |
Dogs | −6,216 | 0.11 | 67,770 | −24,864 | 42,906 | NC |
Cats | −7,080 | 0.16 | 77,172 | −28,319 | 48,853 | NC |
−6123 | 0.74 | 117,106 | −24,492 | 92,614 | NC | |
Dogs | −3,668 | 0.70 | 47,896 | −14,674 | 33,223 | NC |
Cats | −1,230 | 0.83 | 38,512 | −4,920 | 33,592 | NC |
Trends in outcomes were assessed for all animals, dogs, and cats, based on total outcomes and outcome subtype categories (
Based on our sample of 1,373 animal shelters, the total number of animals euthanized as a percentage of intake in shelters across the country decreased by 44% from 2016 to 2020 (
Results of linear regression analysis of outcomes as a percentage of intake reported by animal shelter organizations in the United States (2016–2020).
Cats | 0.001 | 0.14 | 0.024 | 0.003 | NC | NC |
Dogs | 0.006 | 0.09 | 0.151 | 0.025 | NC | NC |
−0.0002 | 0.97 | 0.031 | −0.001 | NC | NC | |
Dogs | −0.001 | 0.87 | 0.029 | −0.004 | NC | NC |
Cats | 0.00003 | 0.99 | 0.026 | 0.00001 | NC | NC |
Trend lines indicated that the live release rate (LRR) for dogs as a function of intakes increased by 15% from 2016 to 2020 (
Results of linear regression analysis of dog, cat, and total live release rates as a function of total annual intakes and as a function of total outcomes by animal shelter organizations in the United States from 2016 to 2020.
The outlier test indicated that there were several categories of shelter operations in 2020 that differed significantly from previous years. The year of 2020 was an outlier in the dataset for the following categories: total intake, total community-based intake, total adoptions, total return to owner, and total transfers out (
Results of linear regression analysis of intake and outcome subtypes (2016–2019), with intake and outcome subtypes in which 2020 was an outlier.
−72,503 | 0.09 | 3,669,558 | NC | 3,379,546 | NC | |
Cats | 8,166 | 0.41 | 1,419,318 | NC | 1,451,982 | NC |
−89,928 | 0.10 | 3,337,321 | NC | 2,977,609 | NC | |
Cats | −6,434 | 0.53 | 1,305,025 | NC | 1,279,289 | NC |
Dogs | −4,746 | 0.55 | 714,550 | NC | 695,566 | NC |
6,439 | 0.08 | 397,376 | NC | 423,132 | NC | |
Dogs | −1,357 | 0.66 | 319,418 | NC | 313,990 | NC |
Dogs | −3,475 | 0.31 | 253,821 | NC | 239,921 | NC |
This study included a sample of 1,373 animal shelter organizations across the U.S. While this dataset represents the most representative and accurate estimate of shelter intake and outcome on the national basis to date, there is still a need for increased participation by shelter organizations in reporting data to these national repositories. For example, only 2,386 of the 4,400 animal shelters known to exist in the U.S. (44%) self-reported a full year of data to SAC in 2020 and there was a lack of participation from the Midwestern and Southern regions (
The overall trends observed in this study indicate that total intake and euthanasia are decreasing for both dogs and cats. Understanding trends in intake across the U.S. is important for assessing the overall capacity and resources of the sheltering system that could be made available to address community-specific needs. Previous studies have utilized geographic information system (GIS) mapping of intake sources for specific communities to inform program development and resource allocation, particularly for communities with high rates of intake (
Many animal shelter practitioners are eager to examine how the COVID-19 pandemic has impacted the animal shelter system within the U.S. While there was an overall decrease in the trend in total intake from 2016 to 2020, it is notable that an estimated 2,622,682 million dogs and cats entered shelters in 2020, which represents a decrease of 25% from the 3,489,598 million total intakes reported in 2019. This stands in contrast to the modest change documented between 2018 and 2019, when admissions increased by 0.07%. The restricted services imposed by many shelters during the COVID-19 pandemic may have been a key factor in the decreased admissions recorded in this dataset during 2020. For example, in a statement outlining recommendations for animal control operations during the COVID-19 pandemic, the National Animal Care and Control Association emphasized the importance of reducing shelter admissions:
“Animal control agencies should take active measures to reduce non-essential shelter intake. Measures taken should include returning pets in the field instead of impounding them, suspending non-emergency owner surrender intake, and encouraging owners who are ill to keep their pets at home whenever possible” (
Many organizations also embraced a “community-supported sheltering” model during the COVID-19 pandemic and created new programs or increased the availability of existing programs that proactively address the most common reasons for shelter intake (e.g., housing insecurity, access to veterinary care, and access to pet food and supplies). Examples of these programs include pet food and supply banks (
This study explored whether the first year of the pandemic (2020) functioned as an outlier in the overall trends in animal shelter intake and outcomes over the last 6 years. The results indicate that 2020 was, in fact, an outlier for the following categories of animal shelter intake and outcomes: total intake, community-based intake, adoptions, return to owner, and transfers out. Any differences in trends that have been observed with or without the 2020 data may be initial indicators of the impacts of the COVID-era programs that have been implemented by shelters in recent years. However, it is important to note that one would not typically remove any single year of data from a trends analysis solely based on the results of any one statistical test. Further investigation is recommended once the additional years of data following the pandemic (2021, 2022) are available, to assess the extent to which 2020 has been or will continue to be an outlier in the dataset. Further, future research should aim to measure the long-term sustainability and effectiveness of these programs while also identifying any remaining gaps in pet support services, particularly in historically marginalized communities.
There are a number of emerging best practices in the animal sheltering field that likely inform the observed trends in intake and outcomes over the study period. The significant decreases in the trends of euthanasia for cats and dogs likely represent a focused effort of both local and national organizations to implement best practices both within the shelter organizations and in the surrounding community. Across the 1,373 animal shelters examined in this study, there were an estimated 265,578 dogs and cats euthanized in 2020, which represents a decrease of 44% from the 475,489 euthanasia outcomes reported in 2019. This stands in contrast to the modest change documented between 2018 and 2019, when euthanasia decreased by 9%. The literature indicates that these best practices for reducing shelter euthanasia include providing specialized medical and behavior care for animals in shelter care, reducing barriers to adoption (e.g., eliminating adoption fees), engaging in triage and appointment-based admissions, and increasing stakeholder engagement in shelter services (e.g., foster caregiving, partnerships with local private practice veterinarians) (
This study also observed statistically significant decreases in intake by owner-intended euthanasia and for the “no reason given” category. There has been limited examination of the issue of owner-intended euthanasia, so ongoing assessment of this trend should be studied in future research (
Although the decreases observed in RTO rates for both dogs and cats were not statistically significant, the considerable differences in RTO as a percentage of intake by species are worth highlighting. While RTO as a percentage of intake increased over the study period by 13% for both all animals and dogs, there was no significant increase, there was no significant increase in RTO as a percentage of intake for cats. By contrast, RTF as a percentage of intake significantly increased over the study period by 65% for cats and 696% for dogs. It is worth nothing that the RTF data for dogs are likely erroneous, almost certainly representing pet dogs returned directly to their owners “in the field” by enforcement staff without bringing the dogs to the shelter. The observed increases in canine RTO and feline RTF rates, each as a percentage of intake, may be a reflection of more organizations returning lost animals to the community where they were found, rather than keeping them in shelter care to be reclaimed through the traditional RTO process. This innovation in lost/found programs for both cats and dogs was implemented within the shift toward community-support sheltering models and effort to reduce shelter intake during the COVID-19 pandemic and was further justified by previous studies that documented RTO rates of 7% or less for cats (
While not statistically significant, the decreasing trends in community-based intake should be monitored on a national basis as a promising indicator of how a collaborative animal welfare system, growing emphasis on surrender prevention, and increasing access to pet support services might be making a positive impact on animal welfare outcomes. Transfers are another important strategy for optimizing the shelter system's capacity on the local, regional, and national level that should be monitored in future research. Over the study period, transfers as a percentage of intake increased by 27% for all animals and by 26% for cats, with no change for dogs. Transfers can help facilitate live outcomes for animals, particularly when the organization receiving the transfer has a higher degree of specialization in addressing the medical or behavioral challenge of the animal or has access to a larger population of potential adopters. Future research should monitor efforts to standardize health and safety protocols for transfer partnerships, including best practices such as mandatory quarantine or medical treatment prior to or post-transport. Further, transfers across state lines should be studied at the state or regional level to understand the extent to which this source of intake may impact the community's risk of disease (
A number of studies have documented trends in animal shelter intake and outcomes on the individual organization or state level (
The findings of this study have several limitations. The methods used in recording, compiling, and analyzing data from a national sample of sheltering organizations are not without their shortcomings. One potential limitation is that all shelter data included in the study are self-reported by each organization and assumed to be as accurate as can be reasonably expected. The best estimates identify 4,400 animal shelters across the United States (
The datasets presented in this article are not readily available. Requests to access the datasets should be directed to:
PW, JR, SMH, and KM conceived the study design. JD and SH compiled and analyzed the original dataset. JR performed the statistical analysis. JR, JD, SH, PW, SMH, and KM contributed to the drafting of the manuscript. All authors contributed to the article and approved the submitted version.
JR's position is supported by a grant from an anonymous donor to the University of Denver Graduate School of Social Work, and SMH and KM's positions are partially funded by the latter's American Humane Endowed Chair research fund.
JD, SH, and PW were employed by Best Friends Animal Society, Kanab, UT, United States. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
The authors would like to thank the staff and volunteers of Best Friends Animal Society, who gathered and organized the data used in this study, as well as the staff and volunteers of the many shelters that collected and shared their data.