Study summary · 2024 · Observational
CCM in diabetes: lower hospital use with substantial confounding risk
CCM receipt was associated with fewer admissions and deaths; patient selection and missing clinical data prevent a causal interpretation.
Summary by Vironix Health · Source review: November 11, 2026
How we select and interpret evidence
Original publication
Dongzhe Hong; Charles Stoecker; Yixue Shao; Elizabeth Nauman; Vivian Fonseca; Gang Hu; Alessandra N. Bazzano; Edmond K. Kabagambe; Lizheng Shi
Tulane University and collaborating research institutions
Journal of General Internal Medicine · Online: · Issue: August 2024
Journal version of record. · Full text consulted.
What was studied
- Design
- Observational
- Population
- Medicare fee-for-service beneficiaries with type 2 diabetes and another chronic condition
- Sample
- 118,643 eligible; 8,415 received CCM. Main outcome models: 114,973, including 8,341 CCM recipients.
- Intervention
- At least one billed non-face-to-face CCM service
- Comparator
- No billed CCM
- Follow-up
- Mean 1.46 years (SD 1.01)
- Study/data period
- 2014–2018
- Setting
- Louisiana, United States
Methods
Retrospective claims analysis using propensity-score weighting and prior-utilization trajectories.
Findings
Adjusted admissions differed by −0.012 per patient-month (95% CI −0.014 to −0.011). Outpatient visits excluding CCM were unchanged (−0.001; −0.025 to 0.023). The mortality coefficient was −0.078 (−0.081 to −0.074), not a treated-group death rate. Stroke showed no clear difference.
Limitations
Unmeasured disease severity, engagement, hospice status and provider factors may bias results. Claims lack biomarkers. Louisiana data and years overlap the related Shao study; independence of patients is unestablished.
What this means for care delivery
The mortality association is not proof that CCM prevents deaths. Added outpatient encounters largely represent the CCM services themselves.