New associations of the Multidimensional Prognostic Index

Z Gerontol Geriatr. 2019 Aug;52(5):460-467. doi: 10.1007/s00391-018-01471-6. Epub 2018 Nov 7.


Background: The multidimensional prognostic index (MPI) is a validated, sensitive, and specific prognosis estimation tool based on a comprehensive geriatric assessment (CGA). The MPI accurately predicts mortality after 1 month and 1 year in older, multimorbid patients with acute disease or relapse of chronic conditions.

Objective: To evaluate whether the MPI predicts indicators of healthcare resources, i.e. grade of care (GC), length of hospital stay (LHS) and destination after hospital discharge in older patients in an acute medical setting.

Material and methods: In this study 135 hospitalized patients aged 70 years and older underwent a CGA evaluation to calculate the MPI on admission and discharge. Accordingly, patients were subdivided in low (MPI‑1, score 0-0.33), moderate (MPI-2, score 0.34-0.66) and high (MPI-3, score 0.67-1) risk of mortality. The GC, LHS and the discharge allocation were also recorded.

Results: The MPI score was significantly related to LHS (p = 0.011) and to GC (p < 0.001). In addition, MPI-3 patients were significantly more often transferred from other hospital settings (p = 0.007) as well as significantly less likely to be discharged home (p = 0.04) than other groups.

Conclusion: The CGA-based MPI values are significantly associated with use of indicators of healthcare resources, including GC, LHS and discharge allocation. These findings suggest that the MPI may be useful for resource planning in the care of older multimorbid patients admitted to hospital.

Keywords: Aging medicine; Clinical decision making; Comprehensive Geriatric Assessment; Grade of care; Prognosis.

MeSH terms

  • Activities of Daily Living*
  • Aged
  • Female
  • Geriatric Assessment / methods*
  • Hospitalization
  • Humans
  • Male
  • Patient Admission
  • Patient Discharge
  • Predictive Value of Tests
  • Prognosis
  • Risk Assessment / methods*
  • Risk Factors