Do Key BH Metrics Differ More by Facility or Patient?

Key behavioral health performance metrics such as patient satisfaction are influenced by many factors. Some operate at the patient level (such as symptom severity or engagement), while others operate at the facility level (such as staffing models, care processes, and culture). These influences also differ depending on the metric being examined.

For example:

• • Patient satisfaction is closely linked to whether patients feel safe, respected, and supported during care¹.
  • Clinical improvement is often related to how severe symptoms are at admission, whether evidence‑based treatments are used, and how well care is coordinated^{2, 3}.
  • Safety outcomes, such as seclusion and restraint, may reflect both patient acuity and how facilities manage escalating behaviors.

When metrics are compared across facilities, an important descriptive question arises:

Do metrics differ more between facilities, or more between patients treated within the same facility?

This is not a question about cause. Instead, it helps us understand where data variation is concentrated, which informs how we interpret comparisons and where improvement efforts may be most useful.

EXAMINING VARIATION

Statistical methods allow us to determine what percentage of a metric’s data variation is due to:

1. Between‑facility variation: that is how different overall scores are facility to facility. High percentages indicate facility scores are very different from facility to facility, suggesting facility related factors could be key to improvement.
2. Variation between a facility’s patients: that is how different patient scores are within individual facilities. High percentages indicate patient scores are very different within individual facilities, suggesting patient related factors could be key to improvement.

These percentages add up to 100. A metric with low between-facility variation would also have high variation between a facility’s patients, and vice versa.

THE METRICS

Using MHO’s data on inpatient discharges from 2025, we examined common BH metrics:

• • Patient Satisfaction
  • Likelihood to Recommend
  • Clinical Outcomes
    • BASIS‑32™ change
    • PHQ‑9 change
    • CABA‑Y change
  • Seclusion and Restraint Rates

PATIENT SATISFACTION AND CLINICAL OUTCOMES

For satisfaction, likelihood to recommend, and all three clinical outcomes change scores, the percentage of variation due to differences between a facility’s patients were high (generally above 80%) while the percentage due to facility differences were low.

What this means in practical terms:

• • Most variation occurs between individual patients, not between facilities
  • Two patients treated in the same facility may have very different experiences or levels of improvement
  • Facility to facility overall scores are broadly similar and their comparison alone may oversimplify comparative performance

Why this matters clinically and operationally:

• • Facility‑level averages are not the full picture; patient-level data are helpful
  • A facility who scores particularly strong or weak against other facilities may have hidden important differences among patients
  • Improvement efforts may benefit from:
    • • Identifying which patients are doing better or worse
      • Targeting interventions to specific patient groups
      • Understanding which clinical or engagement factors matter most

SECLUSION AND RESTRAINT RATES

Seclusion and restraint rates showed a very different pattern, with over 70% of variation due to differences in rates between facilities.

What this means in practical terms:

• • Facilities’ rates differ substantially from one another
  • Patients do not differ substantially within facilities, as most patients do not experience seclusion or restraint

It’s important to note this does not necessarily mean differences are driven by facility practices alone. Some variation may reflect differences in patient populations, such as higher rates of aggression, greater acuity, or more involuntary admissions.

Why this matters clinically and operationally:

• • Comparing facilities is more informative for these metrics than for others
  • Improvement efforts may benefit from reviewing facility:
    • Policies and procedures
    • Staffing models
    • De‑escalation approaches

KEY TAKEAWAYS

Patient satisfaction and clinical outcomes:
There is substantial variation in patient scores while facility averages are largely similar. Exploring patient-level data and focusing on who needs improving is more likely to yield actionable changes precisely where needed.

Restraint and seclusion rates:
Facility averages are substantially different while individual patients are largely similar. Comparing your facility’s overall performance to a benchmark is an informative indicator of performance, and investigating and improving on internal factors such as policy, procedure, or culture may be more likely to improve performance.

References

1. Lu L, Hu Q, Walter Z, Huo D, Zhang H. Assessing patient satisfaction in healthcare: Integrating ratings of service attributes and BERT-based analysis of comments. International Journal of Engineering Business Management. 2025;17. doi:1177/18479790251384339
2. Setkowski K, Boogert K, Hoogendoorn AW, Gilissen R, vanBalkom AJLM. Guidelines improve patient outcomes in specialised mental health care: A systematic review and meta-analysis. Acta Psychiatr Scand. 2021;144:246–258. https://doi.org/10.1111/acps.13332
3. Lim, C. T., Caan, M. P., Kim, C. H., Chow, C. M., Leff, H. S., & Tepper, M. C. (2022). Care management for serious mental illness: A systematic review and meta-analysis. Psychiatric Services,73(2), 180–187. https://doi.org/10.1176/appi.ps.202000473

MHO Health Outcomes

April 21, 2026

Categories:Special Analysis

4 Likes