15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

potential confounding impact of the human handler on study outcomes is considered. (Friedmann et al. (2019) in their research address this methodological weakness by including an equally matched attention control group.) Gee and Mueller (2019) present a systematic review of the evidence in HAI and aging with a focus on the different outcomes based on pet ownership or pet interaction. In other words, does a person need to own a pet, or is interaction with a companion animal sufficient to reap any potential benefits? Overall this review concludes that the strongest evidence for the benefits of pets to people comes from animal-assisted intervention (AAI) research, and indicates that the research methodologies that compare older adults interacting with an animal with older adults interacting with a human (or some other control condition) provide the strongest and clearest evidence of the benefits of HAI. The pet ownership results are fewer in number, include more mixed or null results, and generally provide weaker evidence. Even so, most people still want to know if owning an animal is a healthy choice. Research on the topic of pet ownership has several inherent challenges. Arguably the biggest challenge is selection bias, in that pet owners self-select for pet ownership status; people want to decide for themselves whether to own a pet and if so, which species they prefer. They generally are not in favor of researchers randomly assigning them to a particular pet ownership status. Thus, it is challenging to separate out whether pet ownership makes people healthier or whether healthier people choose to own pets. Some studies do include random assignment of participants to conditions; this automatically provides a higher level of potential evidence by moving it from a quasi-experimental design to a true experimental design. However, in the general population it is likely that a greater number of people own pets than participate in animal-assisted interventions, so pet ownership studies are Gee and Galik 285 Anthrozoös likely to be more broadly applicable to real world settings, something researchers will want to consider when taking an epidemiological approach. Thorpe and Kelly-Moore (2019) provide a critical review of epidemiological research involving companion animals and older adults, including a systematic review of pet ownership information that is publicly-available in existing population studies and could be accessed for research. The authors provide a table that interested researchers are likely to find helpful in identifying useful data sources, and they make clear recommendations regarding future research (which we highlight in our conclusions). In a recent methodological review paper, Friedmann and Gee (2018) evaluated and discussed a variety of other HAI research approaches investigating pet ownership, brief interactions with animals, and animal-assisted activities/interventions. In their candid discussion, several limitations involved in this research are reported, some of which echo the concerns raised by others in this thematic issue (e.g., Gee & Mueller, 2019; Thorpe & Kelly-Moore, 2019). For ease of consideration and accessibility we present a consolidated list of recommendations for future research on HAI and human aging, based on the limitations and suggestions noted by the authors in this thematic issue: Longitudinal Research Designs. Longitudinal research is needed because a cross-sectional approach does not permit the assessment of temporality, sequencing or ordering of events; this is important because life events may trigger pet ownership or the need to rehome a pet. Thorpe and Kelley-Moore (2019) make the important point that pets may not play the same role in a person’s life over time; a longitudinal approach allows one to better understand the types of roles pets play at various life stages (trajectory of change) as well as their potential influence on the development and progression of chronic disease. Including HAI Questions in Existing Large-scale Panel Studies. As Thorpe and Kelly-Moore (2019) report, this has been done in studies like the Health and Retirement Study, but they suggest that these sorts of HAI modules should be included more widely in other studies such as the Baltimore Longitudinal Study of Aging (BLSA); we can now report that is in progress. A new study led by Erika Friedmann and funded by the WALTHAM Centre for Pet Nutrition has just placed an HAI module in the BLSA. At the time of this writing, the NICHD/Mars-WALTHAM Public Private Partnership is also in the process of adding an HAI module into the General Social Survey. We encourage researchers to include HAI questions in other large-scale panel studies, based on the clear advantage to adding such questionnaires to existing large (and expensive) populationbased studies. Given the ubiquitous nature of pet ownership in our society, it is heartening that these studies will now include at least some pet ownership-related questions. Deeper Exploration of Pet Attachment. Because the level or type of human attachment to, and involvement with, pets may be critical in determining if, or how, pets impact human health, it is important to include a reliable and valid measure of pet attachment and involvement in future research. Greater Numbers of Participants. Sufficiently large numbers should be