Skip to content
FourthSpace

What connection is worth: loneliness, health and Sacramento

How strongly is social connection tied to health and survival? How common is loneliness in the U.S. and in Sacramento County, and how many Sacramento adults does that represent?

Independent analysis, not commissioned

Written as a policy brief for readers at Sacramento County Public Health.

Analysis by Andrew Ly (opens in a new tab) · Updated

Key finding

32.1%

Adults who feel lonely always usually or sometimes

Weak connection goes with higher mortality in every meta-analysis.

Mortality ratio for the less connected group, log scale. Pooled effect sizes with 95% confidence intervals.

  • Loneliness
  • Social isolation
  • Living alone
  • Weak social ties (broad measures)

forest_plot_data.csv

Findings

  1. Weak connection goes with higher mortality in every meta-analysis.

    Holt-Lunstad and colleagues (2010, 148 studies, 308,849 people) found people with stronger social relationships had 50% higher odds of survival (OR 1.50, 95% CI 1.42 to 1.59). Wang and colleagues (2023, 90 studies, 2.2 million people) found social isolation tied to a 32% higher risk of death (HR 1.32) and loneliness to 14% (HR 1.14). These are observational pooled estimates.

  2. One in three U.S. adults feels lonely at least sometimes.

    32.1% in 2023 and 31.7% in 2024 across states that asked. 7.4% feel lonely always or usually. 23.1% lack the support they need at least sometimes (2023). California's 2023 rate was 33.5%.

  3. Loneliness tracks closely with mental distress.

    54% of adults who are always lonely report 14 or more mentally unhealthy days a month, against 6% of those who are never lonely. After controls, lonely adults have a 21.2-point higher rate of frequent mental distress (95% CI 20.5 to 21.9) and a 22.1-point higher rate of diagnosed depression.

  4. Sacramento County is lonelier than most of the country.

    PLACES puts its age-adjusted loneliness rate at 39.5% (95% CI 35.3 to 44.2). That is higher than 95.5% of the 2,300 U.S. counties with an estimate and second highest of California's 58 counties. Lack of support (28.9%) is higher than 89.7% of counties.

  5. That is about 477,000 Sacramento adults (estimate).

    38.9% crude prevalence times 1,226,535 adults, with a range of 426,000 to 532,000.

  6. Loneliness is spread across the county.

    Tract estimates range from 30.5% to 49.6%, with a median of 39.2%. The loneliest fifth of tracts holds 22.2% of the estimated lonely adults, so a tract-targeted program reaches only a slice. Project 07 maps these tracts.

  7. Global context.

    The WHO Commission on Social Connection (June 30, 2025) estimates loneliness is linked to about 871,000 deaths a year and affects 1 in 6 people worldwide.

Options

  1. Option A. Measure it

    Add loneliness and lack of support to the next Community Health Needs Assessment and track the PLACES measures each release.

  2. Option B. Social prescribing pilot

    Let primary care clinics refer lonely patients to community programs, with a navigator to follow up. The UK runs this at scale. U.S. evidence is still thin.

  3. Option C. Extend third place hours where loneliness is highest

    My pick

    Fund evening and weekend hours at libraries, community centers and park facilities in the loneliest tracts from Project 07, then track visit counts.

  4. Option D. Public awareness campaign

    Normalize talking about loneliness and point people to local options. Cheap to run, hard to measure.

  5. Option E. Partner with connection platforms

    Publish open data on free public spaces and events so apps (FourthSpace included, which is a conflict to disclose) can surface them.

Default pick: C. It uses buildings the county and city already run. Visit counts give a cheap way to measure results, and the tract data in this repo shows where to start. Evidence that any intervention reduces loneliness is weaker than the evidence that loneliness harms health, so the pilot should be small, measured, and paired with A.

Method

  1. 01_forest_plot.py draws the published all-cause mortality estimates from three meta-analyses on one log scale.
  2. 02_brfss_extract.py pulls 21 variables from the 2023 and 2024 BRFSS files (890,993 records).
  3. 03_brfss_analysis.py keeps the states that asked the loneliness module and uses CDC's definitions. Lonely means feeling lonely always, usually or sometimes. Lacking support means getting needed social and emotional support sometimes, rarely or never. It estimates prevalence, the outcome rate at each answer level, and adjusted gaps controlling for age, sex, race, education, income, marital status, urban or rural residence and year.
  4. 04_places_sacramento.py places Sacramento County against all counties with an estimate and computes the count of Sacramento adults affected. That count is an estimate: PLACES crude prevalence times the PLACES adult population, with the PLACES 95% interval carried through.
  5. 05_charts.py draws the charts.
  6. Validation. Our BRFSS 2023 loneliness rate for module states (32.1%) sits beside CDC's published 2022 figure for a closely related item (31.9% social isolation or loneliness, MMWR 2024).

Tools

Python 3.13, pandas, NumPy, pyreadstat (BRFSS SAS transport files), pyarrow, matplotlib, Plotly, and the shared design-based variance module scripts/survey.py (Taylor linearization with BRFSS strata and PSUs).

Caveats

  1. Every BRFSS and meta-analysis estimate is an association. Lonely people may be lonely because they are sick, and illness and loneliness share causes such as poverty and disability.
  2. PLACES values are model-based small-area estimates. They borrow strength from demographics and state-level survey data, so a county's estimate partly reflects its population makeup. Sacramento's direct survey sample is too small for a county estimate on its own.
  3. The BRFSS loneliness question is in an optional module. Only 33 to 35 states asked it, so "U.S." figures here mean those states.
  4. The "never gets support" group has lower distress than the "rarely" group. Some people in it may need little support, which makes that category hard to read.
  5. PLACES (38.9% crude) and our direct BRFSS California figure (33.5%) differ because one is a county model estimate and the other is a statewide survey estimate.
  6. Mortality ratios from meta-analyses carry heterogeneity across studies and countries. They describe risk. They do not measure what an intervention would change.

Downloads

These files come from US federal agencies and are in the public domain. Every other source has its own terms, so it is linked at the publisher under Sources.

Sources

  1. Published meta-analyses of social connection and mortality, hand-entered with page-level checks (data/raw/meta_analysis_effect_sizes.csv)

    Year
    2010-2023
    Geography
    International cohorts

    Holt-Lunstad et al. 2010 https://doi.org/10.1371/journal.pmed.1000316 (opens in a new tab) · Holt-Lunstad et al. 2015 https://pubmed.ncbi.nlm.nih.gov/25910392/ (opens in a new tab) · Wang et al. 2023 https://www.nature.com/articles/s41562-023-01617-6 (opens in a new tab)

    Publisher's terms

  2. CDC Behavioral Risk Factor Surveillance System, Social Determinants and Health Equity module

    Year
    2023, 2024
    Geography
    Adults in 35 states (2023) and 33 states (2024) incl. California

    https://www.cdc.gov/brfss/annual_data/annual_data.htm (opens in a new tab)

    Public domain

  3. CDC PLACES 2025 release, county and census tract estimates (based on BRFSS 2023)

    Year
    2023 data
    Geography
    U.S. counties; Sacramento County tracts

    https://www.cdc.gov/places (opens in a new tab)

    Public domain

  4. WHO Commission on Social Connection report (cited)

    Year
    2025
    Geography
    Global

    https://www.who.int/news/item/30-06-2025-social-connection-linked-to-improved-heath-and-reduced-risk-of-early-death (opens in a new tab)

    Publisher's terms

Interactive charts

  • Every major meta-analysis finds higher mortality among less connected adults

  • Half of adults who are always lonely report frequent mental distress

  • Sacramento County is lonelier than 96% of U.S. counties with an estimate

Keep reading

Third place supply: what we have, what we lost, where we spend our time

Has the U.S. lost its third places, the spots outside home and work where people gather? Which kinds of places grew and which shrank since 1998, nationally and in Sacramento County? And do Americans still use the ones that remain?

Next study
Coffee shops per person nearly doubled since 1998, and civic clubs fell 40%

Be first through the door.

We're opening in Sacramento first. Join the list and we'll email you when we open.

Fields marked * are required.

Places you'd go (optional)
How we use your details

Study03/07Join the waitlist