The Bureau of Labor Statistics reported a seasonally adjusted quits rate of 1.7% for health care and social assistance in May 2026. In the same table, the sector had 412,000 quits. BLS defines the quits rate as quits during the month divided by employment.
What the JOLTS finding measures
| Finding | Period and population | Limit |
|---|---|---|
| 1.7% quits rate | Health care and social assistance, May 2026, seasonally adjusted | Industry measure, not home-care occupation data |
| 412,000 quits | Same industry and month | Estimate is not an agency count |
| 2.0% quits rate | Same industry, May 2025 comparison in the table | Historical comparison can be revised |
Health care and social assistance includes more than home care. The JOLTS result therefore provides a labor-market context signal, not a measured caregiver turnover rate for a particular agency.
Why denominator and scope matter
An agency retention study should state the worker population, starting headcount, observation period, voluntary versus involuntary exits, and whether transfers are included. A monthly sectoral quits rate cannot answer those questions. It also cannot identify why a worker quit or whether client outcomes changed.
Source register and limits
JOLTS is a monthly establishment survey with industry-level estimates and published definitions. The data can be revised, and seasonal adjustment changes interpretation. Use the original table and release date when comparing periods.
FAQ
Is 1.7% the caregiver turnover rate?
No. It is the seasonally adjusted quits rate for the broader health care and social assistance industry.
Can a lower quits rate prove better retention?
No. It measures quits only. Other exits, staffing mix, hours, and workforce composition still matter.
What should an agency measure locally?
Track a defined worker population, hires, quits, other separations, tenure, hours, and the observation period with a stated denominator.
Related research
See the direct-care workforce openings and home care caregiver retention process benchmark.

