Dogs vs. Seal Robots vs. Plush Toy Cats: Do Any of These Improve Sleep and Long-Term Mental State in Nursing Home Elderly?

For researchers at Aarhus University in Denmark, the answer to this question was a resounding no. They enrolled 124 elderly participants living in nursing homes to participate in a study to test the effectiveness of twice weekly visits by dogs on depression, cognition, and sleep.  However, results indicate that there were no statistically significant decreases in depression scores or weight/BMI, while there were increases in cognitive decline and disability over the period of the study. The only positive change participants experienced were longer sleeping periods when a dog visited compared to the robot seal or toy cat. (You can access this article here:

While I applaud the authors for publishing negative results, I feel that this article raised several questions that impede my ability to accurately assess the results. While I have no doubt that the authors are correct in their analysis of the data, this article is a prime example of how difficult it is to evaluate research in this field and make conclusions about the potential benefits or harms of animal assisted therapies (AAT). Here are some of the issues that need to be more fully flushed out in this article in order for us to better understand the methodology and results of the study:

  1.  The authors note this study is “part of a larger study of the acute and long-term affects of biweekly dog visits”. In this study – dogs visited nursing home residents twice weekly for six weeks (12 visits total). This seems like a very short time period for an intervention like this, especially if one is trying to determine if this intervention can decrease depression and improve sleep. Why did they choose this six week period? Is the larger study using a longer period? What in the literature led them to use this short time-frame, or were they impeded by other constraints such as funding or manpower to do the study?
  2. Duration of the visits with the dog, toy cat or robot seal were limited to ten minutes. The researchers did this noting the individuals with severe cognitive impairment are often unable to focus for much longer, but they also note in the limitations section of the paper that this may have been too short a duration to have had any meaningful effect.
  3. Results do not indicate whether the participants were depressed at the beginning of this study. If they were, and their scores improved (as is indicated in the data although these improvements were not statistically significant), this could be a potentially important finding
  4. The researchers used a randomized block design in which participants were randomly assigned to receive visits from either a large breed dog and his/her handler, a robot seal (you can see videos of these seals here: or here:, or a plush toy cat. The question I am left with, is why the researchers omitted a true control arm for this study, i.e. a group who received no visits at all? This could have been very useful in determining the effects of these visits.
  5. The researchers used actigraphy to assess sleep time, sleep efficiency, and restlessness. I have been involved in studies using this technology in studies related to depression and they are very effective tools at assessing how interventions are affecting sleep, a key factor that impacts depression. However, in this study, participants wore the watches only four nights during the six week study period, once at basleine, twice on the evenings after a visit, and once following the intervention period (i.e. in week seven). I’m not sure why researchers chose only to look at one night, given that sleep patterns are best aassessed over a longer period of time, and so I’m not even sure what their data tell us about the sleep improvements in this population. However, given the expense of this technology and the N in this study, my inclination is to think that the researchers simply were constrained by not having enough devices to give to participants for longer periods than one night. The researchers do note that they are the first to use this technology in looking at an AAT and that cautious interpretation of their results is warranted. Further, they suggest that more research is needed to determine if this is an appropriate way to measure the effectiveness of AAT.

The researchers do note other limitations of their study, namely that having dogs visit nursing homes with their owners could have made them more confident which could have impacted study results; only large dogs were used in the study; and low intensity/low level of interaction with the dogs during the visits. These are of course, issues that most researchers in AAT deal with in some way, i.e. how do you standardize animals as interventions in research studies? And again, I would suggest that we need to be creative and think outside of the box, incorporating the expertise and techniques from multiple disciplines in order to overcome these issues.

Nevertheless, although this was an interesting article about a unique study to assess the impact of dogs on nursing home residents, many questions remain. Although critics have pointed to this study as an example of how animals are not beneficial in ways we would like to think that they are (, examining the details of the research suggest we need to be just as critical of the negative results in AAT studies as we are of the positive ones.