SONNET Newsletter: December 2021

SONNET is a national network of applied researchers who help design, evaluate, and implement effective social health interventions to improve member health across Kaiser Permanente. SONNET's work is made possible with funding from KP's National Community Health Program.

December 14, 2021

Understanding the impacts of loneliness on health  


KP researchers conduct the largest known study on loneliness – assessing data from 24,000 KP members – to better understand the impact of loneliness on health, factors that contribute to loneliness, and what can help stop loneliness.

An interview with John F. Steiner, MD, MPH, senior clinician investigator at the Kaiser Permanente Colorado Institute for Health Research (IHR) and former SONNET director; Tammy S. Kelly, MA, senior innovations and life care planning manager at Kaiser Permanente Colorado; and Alena Wall, MA, CHES, public affairs director for the Marin-Sonoma Kaiser Permanente, and previously community health manager at Kaiser Permanente in Northern California.  

What is this study about? 

Steiner: This study is using data from Kaiser Permanente’s annual Medicare Total Health Assessment (MTHA), which captures self-reported health information and can assess changes in behavioral health over time. In the Colorado region, Kaiser Permanente has offered the MTHA survey to members 65 and older each year since 2012. This allows us to look at changes in loneliness in relation to changes in other health outcomes, like depression, anxiety, self-perception of general health.  

Kelly: From prior research, we already know that loneliness is associated with poorer health outcomes. But by using annual MTHA data, this study gives us the unique opportunity to look at self-reported changes in loneliness at different time points. 

Wall: Kaiser Permanente’s National Social Health Practice has also looked at social drivers of health outcomes enterprise wide. In a representative sample of KP members, we found that loneliness was the second most prevalent social health need, so that really deepened our interest in going further in this study. 

Steiner: Our overall goal was to examine the statistical relationships that might contribute to loneliness. But throughout the study, Tammy and Alena have really pushed us to look deeper than the numbers and to bring this down to the individual patient level. How can we look beyond what the electronic health record can provide and better understand what causes loneliness? We wanted to understand what that experience feels like from the patient’s perspective, including the consequences for their health and the implications for health care decision-makers.  

To do that, we identified 4 distinct groups within a cohort of 24,000 KP Colorado members who completed the MTHA twice over a time period of 6 to 18 months.  

  • Group 1 experienced loneliness at both time points (about a year apart on average).
  • Group 2 experienced loneliness at the first time point. 
  • Group 3 experienced loneliness at the second time point, but not the first. 
  • Group 4 did not experience loneliness at either time point.  

What has the study found out so far? 

Steiner: More than 70% of members in the study reported they were not experiencing loneliness at either time point. About 6% of members in the study experienced increasing loneliness, and another 6% experienced decreasing loneliness over time. 

We found that individuals who were lonely were most likely to experience depression, anxiety, or adverse general health, while individuals who were not lonely had the lowest rate of depression, anxiety, and adverse general health.  

Kelly: We also found that individuals who saw an improvement in loneliness had an associated reduction in depression and anxiety. This is extremely hopeful. Understanding factors that contribute to the onset of loneliness, as well as those that help resolve loneliness, is important. If we understand what some of those factors are, we might be able to provide additional resources or interventions and have a better understanding of how we can help people experiencing similar situations in the future. 

What makes this study different from previous research on loneliness and health? 

Steiner: With longitudinal data from 24,000 people, this is the largest patient sample we know of in which loneliness has been assessed and changes have been measured over time.  

Kelly: In addition to having an unusually large cohort, this study is based in a clinical setting, which means we can connect MTHA data to health outcomes and get more insight into what’s going on at the individual level. 

For example, I noticed the accumulation of different sorts of negative events that can happen in a person's life. Maybe a patient is struggling with a longtime behavioral health condition, then also dealing with complicated family interactions and declining health. It’s clear that their ability to compensate is limited as additional challenges come up for them. This is the population that I really want to help – people who are persistently lonely and likely experiencing other issues in their life. This population requires a big picture understanding of what's contributed to this throughout their life, which should inform how we develop interventions or programs to help them.  

I recall a volunteer who worked in the senior programs department who would call members who self-reported as lonely. The volunteer would do individual outreach and chat with them on a regular basis. After a while some people would say, “I’m actually feeling like I’m okay now.” This is the type of intervention I would like to see that could be fairly simple and potentially have a big impact.   

What’s next? 

Kelly: Future studies should take a deeper dive to examine factors that contribute to loneliness from a qualitative perspective. We need to learn more from first-hand observation, as well as doing more interviews, questionnaires, or focus groups with patients. Additionally, there are current programs in place, like the Program of All-Inclusive Care for the Elderly (PACE), which may have an impact on the relationship between self-reported loneliness and health outcomes such as depression and anxiety. PACE is an example of a clinical program that provides all-inclusive care to older adults, including medical, social, and cognitive components.  

Wall: One of the things we’re trying to understand is next steps for a pilot for member-facing interventions. Using information collected through this project, we are interested in looking more deeply at people who reported newly experienced loneliness as we consider what kind of interventions would be most effective in turning those feelings of loneliness around.  

We’re finishing up a set of analyses that look at the same four groups in relation to changes in health care utilization, ambulatory care, primary and specialty care, emergency department hospital care, hospitalization rates. We’re also looking at the duration of hospitalizations under the rationale that patients who have more loneliness and social isolation might have more trouble with being discharged from the hospital after they're admitted, resulting in longer hospital stays. These results, which are expected to be completed by early 2022, have important organizational implications.  

Steiner: There are many interesting questions left to explore, such as the relationship between loneliness and changes in cognition in older adults. Do individuals who report more loneliness experience greater cognitive decline?  

Medicare health assessments are important ways to gather information that isn’t typically available in the electronic health record. The MTHA is an absolute treasure trove of information for clinical practices and research. It has enabled us to look at the relationship between food insecurity and health outcomes, along with a whole host of other questions that we can address to improve care. 

This study reflects our aspirations to continue to grow as a learning health system. Certainly, the scientific value of this research is significant, but I think the hallmark of a learning health system is that scientific knowledge influences decisions. With the partnerships we've had at Kaiser Permanente at every level, we have a lot more confidence that the knowledge from this study will really help achieve that goal of informing decisions to improve care.

SONNET-loneliness-project-team.pngClockwise: Tammy Kelly, Andrea Paolino, John Steiner, and Alena Wall

More news from this issue

Larnette Slade

Award-winning innovation that comes from the heart

Community resources specialists (CRSs) are members of the primary care team who help connect patients to community-based resources for social needs such as housing, transportation, caregiving support, and healthy food access. In this blog, CRS Larnette Slade writes about what her job means to her and how it feels to be part of an award-winning team at KP Washington. READ MORE.

Contact SONNET

Email us at to learn more about our work or request a consultation.