The “Per Thousand People” Metrics
The most common measure of shelter performance is the live release rate, which measures the live outcomes of an individual shelter against its euthanasias. There is another measure of shelter performance in common use, and that is Pets Per Thousand People (PPTP). Although it says “pets,” what it really measures is shelter intake per thousand people. Additional “per thousand people” measures include euthanasia (EPTP) and adoption (APTP). Proponents of the “PTP” metric (which was invented by Merritt Clifton) argue that we must standardize shelter statistics against human population to be able to compare communities to each other in a meaningful way.
I think that the PPTP and its associated measurements suffer from a serious flaw in that they assume that human population is a limiting factor on a shelter’s ability to save lives. In fact, there seems to be little correlation between PPTP and shelter success. In my research on shelter stats, I have seen many successful No Kill shelters with higher than average PPTP, and some less-than-successful shelters with low PPTP. In spite of this, I think that two of the PTP measures – PPTP and APTP – can provide very useful information in some circumstances. Shelter directors may want to be familiar enough with PPTP and APTP to derive those numbers for their own shelters and to know what the numbers mean for their shelter’s performance. Explaining why is going to take a few paragraphs.
Let’s start by setting out the limitations of the PTP numbers. Standardizing numbers is a good thing to do if you have a relevant variable that is not accounted for by the numbers coming out of the shelter itself. For example, if community A and community B both have 5,000 homeless animals each year, but community A has 900,000 people while community B has 100,000 people, then the task of finding homes for all the animals seems as though it would be considerably easier for community A. Thus it seems unfair to judge community A’s shelter performance as “better” than community B’s performance if the shelter in community A rehomes 4800 of its 5,000 homeless animals and the shelter in community B rehomes 4200.
In the real world, though, it doesn’t seem to be true that having higher intake per person is a handicap. I have a spreadsheet where I’ve entered data on the 100 shelters that were listed on this blog at the end of 2013 as having live release rates of 90% or more (serving over 200 communities). The PPTP for the communities served by 82 of those shelters for which I had 2012 intake data ranged from 3 to 139. I don’t think there is anything significantly wrong with the data going in, because the average intake for the shelters I list was 33 per thousand people, and that is near the estimated average for the United States of 15 to 30 pets per thousand people. The communities I list on my blog are a homogeneous group in that they are all very good at successfully placing the great majority of the animals their shelters take in. I verified this not by the live release rate alone, but by my research on each community. If PPTP is a relevant variable for a shelter’s performance, then one would think that these shelters, which are all performing at a similar high level, would all have a similar PPTP, but instead the variation in PPTP is huge.
Why is human population not a fully relevant variable for shelter performance? One reason is that the relationship of total human population to potential pet-owning population varies a lot from one community to another. Let’s look at an example. The community in the United States that has the lowest or near-lowest PPTP number is New York City. New York City is far from the best in terms of the live release rate, though, since the live release rate was at only 77% in 2013 and 81% in 2014. So which is it? Is the New York City community the best in the nation at saving homeless animals, or is it only somewhat above average? The problem with trying to standardize shelter killing against population for New York City is that, as a big city, it has lots of small apartments and a very high cost of living. In other words, New York City probably has a higher percentage of people who are not pet owners and are not prospective pet owners than any other city in the United States. These people are not going to affect shelter performance. They are not going to affect the intake rate because they are not going to give up a pet, fail to have a pet sterilized, or allow a pet to stray. The low PPTP is simply an artifact of the type of housing in New York. So using PPTP to compare New York City – or any other large, highly urbanized, expensive city – to a surburban or rural community is not very meaningful for shelter performance.
Another reason that human population in a community is not a limiting factor on shelter performance is that today we are living in the age of transports. Transports, by taking animals from an area of low demand for shelter pets to an area of high demand, can level the playing field between communities. Going back to our example of Community A and Community B, if community B transports 4000 of its 5000 animals to Community A each year, then each shelter will have to place 10 animals per thousand people. Yet Community B will still have a much higher PPTP than Community A, because PPTP is calculated on intake without subtracting out transfers. In theory, the PPTP could be modified to account for transports, but I have not seen this suggested by any of its proponents.
Proponents of the PTP metrics argue that the live release rate is unfair to the shelters that are doing the best job, because the best shelters are those that are diverting healthy and treatable animals and only taking in vicious dogs and hopelessly ill animals. This idea makes some sense in theory, but in all my researches I have yet to see an example of a public shelter that has a low live release rate because it is taking in only tough cases and diverting the easy cases. Instead, it appears that a central shelter that impounds animals before diverting them to rescues, treats most of the treatable ones in-house, and centralizes return-to-owner and lost-and-found, is the more efficient arrangement. And for that type of shelter the live release rate is an accurate measure. Even if shelters do begin to divert large numbers of animals, such as cats who go to a return-to-field program, it would be an easy matter to calculate a live release rate that reflects the diversions.
Having said all that, I nevertheless think that two of the three PTP measures, PPTP and APTP, are useful in limited circumstances. The average Pets Per Thousand People (remember this is really “shelter intake” not “pets”) in the United States is estimated at 15 to 30. There are many examples of successful No Kill shelters with PPTP higher than 30. As the PPTP number gets higher though, more and more of those shelters are achieving their high live release rates by transports. Although human population is not a limiting factor for No Kill success, at some point it is a limiting factor for being able to place animals within the community. Exactly where that limiting factor is no doubt depends on the individual community, but I think a shelter with PPTP of over 40 that is struggling to maintain a high live release rate may need to think about transports. (They also need to figure out why intake is so high and see if there is anything that can be done to address it, but that is likely to be a longer-term solution.)
Of the three PTP metrics, APTP is the one that has the best argument for statistical validity, because it makes intuitive sense that the number of adoptions will depend on the number of people in the community. Yet this is not entirely true, because the real limiting factor on adoptions is market share, not the number of people in a community. In all but a handful of communities in the United States, the total market for pets far exceeds the total number of pets needing homes. There are currently some 180 million pets in the United States, and only about 5 million healthy or treatable shelter animals needing homes each year. With sufficient market share, replacement alone would provide more than enough homes for shelter pets. Furthermore, about half of the animals needing homes are cats, and many of them can be returned to field, further reducing the gap between the market for pets and the supply of shelter animals. Because market share, not human population, is the real limiting factor for adoptions, shelters can control their own destiny in terms of adoptions.
Some people argue that there is a ceiling to APTP and that adoption rates cannot exceed 4 to 10 APTP on a sustained basis. In fact, statistics show that No Kill shelters regularly exceed 10 APTP, which demonstrates that 10 APTP is not a ceiling but a floor. The reporting shelters for the state of Colorado, for example, averaged 17 APTP in 2014. The Nevada Humane Society has historically averaged over 20 APTP. The value of the APTP is that if a shelter is under 10, you know it has a lot of room for improvement in adoption marketing. In fact, if a shelter has an APTP below the mid-teens, it very probably can do better and it should be able to raise its live release rate without resorting to transports.
I have never been able to discover a use for the EPTP, which relates the number of animals killed in a community to the number of human residents. The idea behind EPTP is that we need it to accurately evaluate public shelters that take in only vicious and terminally ill animals because they are diverting all the healthy and treatable animals. This is ill-considered since, as mentioned above, today we do not have any such public shelters. In addition, EPTP is flawed because it fails to take into account all the ways animals can be live releases without going into a home within the community. Those ways include transports, TNR, and return to field.
The way I think of PPTP and APTP is not as measures of shelter performance, but as information for shelters that are struggling. Shelters that have PPTP of 30 or less and APTP of over 10 probably do not need to be concerned with the PTP metrics. But for those that do not meet those standards and are having trouble achieving a high live release rate, PPTP and APTP may provide some useful information.
So how do you calculate the PPTP and APTP? The easiest way is simply to divide the intake number or adoption number by the human population of the jurisdiction served by the shelter minus three zeros. For example, if you have intake of 15,000 animals per year in a city of 1 million people, you divide 15,000 by 1,000 for a result of 15 PPTP. The APTP for a city of 40,000 people where the shelter adopted out 400 animals in a year would be 400 divided by 40, for a result of 10.