Last week, I wrote a blog post entitled "Why We Love Our Pets: They're Just Like Us!" summarizing some of the most popular and headline-grabbing scientific findings about pet ownership. I was particularly intrigued by a couple of well-designed long-term studies showing benefits of pet ownership on “hard” medical outcomes, including cardiovascular disease and stroke. A fellow Psychology Today blogger and pet psychology expert, Hal Herzog, Ph.D. wrote a response to my blog post. He provided links to scientific studies, which, he claimed, did not support a link between pet ownership and improved health. I’m always up for an evidence-based challenge, so here is my response to Dr H.
In his blog post entitled Puppies, Kittens, and Human Health: Science Versus Wishful Thinking (“What do scientists really know about "the pet effect?"), Dr Herzog said the following:
“The file cabinet in my office contains a two-inch thick folder of studies Greenberg would like. These studies found beneficial effects of living with pets. The bad news is that I have another two-inch stack of studies, which found that pets had no effect on human health and happiness or that pet owners were actually worse off than people without pets.”
Having thus thrown down the gauntlet, he provided links to many studies purporting to demonstrate these non-effects. However, after a careful methodological review, I still stand by my original conclusion that best evidence supports a beneficial effect of pet ownership on health (meaning physical health, as distinct from psychological adjustment). I found issues with the methodology of most of Dr H's cited studies that limit conclusions or ability to properly address the question of health effects for pets.But, before I get to critiquing the research he preented, I'd like to review the evidence supporting a health benefit of owning pets.
Two studies that looked at survival (vs death) in the year following admission to a hospital for a cardiac problem as the outcome, found clinically meaningful and statistically significant differences in the odds of survival between pet owners and non-owners.
One study reported these massive effects:
“Dog owners (n = 87, 1 died) are significantly less likely to die within 1 year than those who did not own dogs (n = 282, 19 died; p < 0.05)” (Friedmann & Thomas, 1995).
In a 2011 study of community-living heart attack survivors, after controlling for heart disease severity, depression, and other factors,
"not owning a pet was the only significant independent predictor of mortality." (Friedmann et al., 2011).
In a 2008 study, Researchers at the University of Minnesota's Stroke Research Center looked at 4,435 people, aged 30 to 75 years, who were participating in ongoing national government health research from the National Health and Nutrition Examination Study. They found that over a 20-year period, those who had never owned a cat had a 40 percent greater risk of death due to heart attack and a 30 percent higher risk of death due to any cardiovascular disease than previous or current cat owners.
In several studies, over time frames from 1 to 20 years, pet ownership was associated with reduced risk of dying from a cardiovascular disease or event. These findings have great clinical and social importance.
The benefits of pet ownership have been demonstrated on many different types of “hard” health outcomes, including
These are hard scientific outcomes, reliably measured. Most of these effects had real-life importance. For example, a recent literature eview concluded that:
“dog owners were 50% more likely to achieve the nationally recommended 3 hours per week of exercise.”
In addition, many experimental studies have found that presence of pets reduce blood pressure and heart rate when experimental participants are exposed to stressors such as mental arithmetic, putting one’s hand in ice water, or making a speech.
In contrast to the strong methodology of studies supporting a health benefit for pets, studies not finding a benefit pf pets were mostly just comparisons of the two groups at one time point, and had questionnaire or other 'soft" outcomes. When hard scientific outcomes were used, they tended to be assessed in less reliable and clear ways.
Most of the studies Dr Herzog cited compared the responses of pet owners and non-owners on self-report questionnaires measuring states like loneliness, perceived health, or depression. These studies found pet owners were more lonely or perceived their health as poorer than non=owners. There are a number of problems with this strategy. How you perceive your health is strongly affected by your mood. Measures of perceived health don’t capture physical health as well as objective medical outcomes, such as your lab results or whether or not you have heart disease, or even whether or not you die. Loneliness and depression are psychological, rather than physical health states. You can be sick and not lonely (e.g., a cancer patient with a large family living nearby) or lonely but not sick (e.g., a homesick college student). Therefore, studies with these types of outcomes aren’t asking the right question with respect to the effect of pets on actual health status.
Some of these studies measured pet ownership and health outcomes, such as blood pressure or obesity, at the same time. This raises the “chicken and egg” issue. We don’t know what causes what. If pet owners report more loneliness, it may be that pets make people lonelier (Why would this be????) or that lonelier people are more likely to buy pets (reverse causality) or that another factor, such as having no family nearby, both makes people lonely and makes them more likely to buy a pet for companionship (third variable confound). We simply can’t tell why the variables are related, so this type of design doesn’t properly address the question of whether pets make us healthier.
Dr. Herzog cited one study that compared blood pressure between pet owners and non-owners and concluded that
“after adjustment for age and other confounders, pet ownership was not associated with systolic or diastolic blood pressure, pulse pressure, mean arterial pressure or risk of hypertension.”
Sounds convincing, right? A closer look at the methodology shows, however, that, although they measured lots of people, they only assessed blood pressure at a single clinic visit. It turns out that one measure, especially one taken at a clinic, doesn’t mean much. Why? Blood pressure readings are all over the place—they are influenced by situational factors; such as the room you are in, who is doing the test, whether you exercised before, what you ate, how stressed you are, etc. People generally have higher blood pressure in the clinic than at home, which is known as “white coat hypertension.’ In other words, visiting doctors is stressful or makes you worry about your health, which sends your blood pressure soaring above normal levels. Therefore, national research guidelines suggest taking 2 or 3 blood pressure readings at different time points or having subjects wear an ambulatory blood pressure cuff that stays on all day and takes continuous measures. One clinic measure simply doesn't cut it.
The first study cited by Dr. Herzog was described by its authors as follows:
This was a prospective study of 424 patients admitted to a cardiac unit with an acute coronary event. Participants completed questions regarding pet ownership in hospital. Rates of cardiac death and readmission were assessed at 1 year following hospitalization. Results showed that pet owners were more likely to experience a death or readmission to the hospital, after controlling factors such as diabetes and depression occurring post-admission. However, when dog and cat owners were considered separately, only cat owners had increased risk of death or readmission.
This study measured people at two different time points following pet ownership and hospital admission using hard medical outcome measures (death or hospital readmission), and results appear inconsistent with a pet health benefit. A closer reading reveals, however, that the researchers in this study combined two dissimilar outcomes—dying and being readmitted to the hospital. The researchers did not show any evidence that these outcomes were related, so as to justify combining them. Hospital readmission could have delayed death, or made it more likely—we don’t know. It’s like combining apples and oranges—we don’t know what the combined measure really means in terms of health.
Their Limitations section stated that:
The study used a combined endpoint of cardiac death and readmission and is therefore not a direct replication of the two previous studies that have examined cardiac deaths only.”
In fact, in this study, since less than 4% of participants (about 16 of 412) actually died, there weren’t enough dead people to validly answer the main research question of whether pet ownership predicted subsequent death. With so few people, there's not enough statistical power to detect group differences. In addition, the authors noted
“the study endpoint did not include death and readmissions that occurred in the month following the admission for acute coronary syndrome.”
This gap in data collection also make it more difficult to interpret the results.
Thus, at close inspection, the evidence for no health benefits or negative effect of pet ownership on health is not at all strong. As you will see below the evidence in favor of a pet health benefit is a lot more compelling.
I thank Dr Herzog for the opportunity to debate this fascinating issue. He is a well-respected expert in the area of pet effects. I, on the other hand, don't know all that much about pets, but do know a lot about assessing health outcomes. We look at the same data with different eyes, and there may be more studies that Dr H. has up his sleeve that add to this disciussion. If this is the case, I look forward to hearing about them.
Cheers for now! Gotta go walk my dog....
About the Author
Melanie Greenberg, Ph.D. is a Clinical Psychologist, and expert on Mindfulness and Positive Psychology. Dr Greenberg provides workshops and speaking engagements for organizations, life, weight loss, or career coaching, and psychotherapy for individuals and couples.
Visit my website:
Read my Psychology Today blog & personal blog