Home

Prediction of Mortality in Community-living Frail Elderly People with Long-Term Care Needs

Jump to>> Article Citation Bottom Lines Methods Setting Conclusion Comments
Printer-friendly versionPrinter-friendly version    
Article Citation
Carey, EC et al “Prediction of Mortality in Community-living Frail Elderly People with Long-Term Care needs” JAG 2008 56:68-75
PubMed ID: 
http://www.ncbi.nlm.nih.gov/pubmed/18031487
   
Clinical Bottom Lines

The 7 identified risk factors independently predict death in frail elderly population.

  • Age >75 [ 75-79/80-84/>85]
  • Male
  • Partially or fully dependent in toileting or dressing
  • Malignant neoplasm
  • CHF
  • COPD
  • Renal Failure/insufficiency

The more risk factors that the patients had, the higher their risk of mortality became.

   
Methods
Type of Study: 
Clinical Prediction Rule
Study Design: 
Cohort study
Follow-up Period: 
Patients were followed for a mean of 2.5 +/- 1.6 years. All data was collected into a centralized and standardized DataPACE program to ensure reliability.
Setting: 
Patients enrolled in total of 11 sites of the Program of All Inclusive Care for the Elderly (PACE) from 5/1988 – 12/1996.
Patient Population: 

Mean age for both groups 79, with 94-95% eligible for Medicaid,  49% diagnosed with dementia, 50-52% had been partially or fully dependent in >3 ADL’s.

Each group was examined by an interdisciplinary healthcare team. Participants were followed throughout the study period.

The evaluation included a complete history and physical, review of records and available labs.

They were evaluated for the following predictors:

  • Demographics: age, sex, marital status
  • Diseases and conditions: active medical problems, problems controlled with medications, if they contributed to participant’s disability
  • Cognitive impairment : evaluated using Short Portable Mental Status Questionnaire (SPMSQ)
  • Hospitalization hx within the last 6 months.
  • Functional Status: evaluated dependence for ADL’s and IADL’s
  • Caregiver status: if formal or informal
  • Sensory deficit: hearing, vision, communication

Development group: participants in Northern California, Colorado, Oregon, Texas

Pts were evaluated for the above medical problems. The risk factors that showed an independent risk for mortality, were assigned a point value. Risk score was then calculated for each of the patients. Pts were then divided into three risk stratified groups (0-3/4-5/>5) to approximate tertiles of risk.

Validation Group: participants form Massachusetts, South Carolina, Wisconsin, New York and Michigan

Patients were evaluated for the above predictors. Each patient was assigned a risk score based on the scores developed in the development group. Mortality rates were compared to the development group at year 1 and 3.

Significant Exclusions: 

Patients who “were missing major predictor variables”

Intervention/Exposure: 

Cox regression was performed to determine the relationship between the risk factors and mortality.  Multivariate analysis of the development group results used stepwise elimination of the risk factors to identify the predictors for the prognostic index.

Outcome Measures: 

Time to death from time of enrollment

Participant Follow-up: 
Unknown
   
Conclusion
Results: 

Variables that had independent association with Mortality:

  • Age >75 [ 75-79/80-84/>85]
  • Male
  • Partially or fully dependent in toileting or dressing
  • Comorbidities: Malignant neoplasm, CHF, COPD, Renal Failure/insufficiency
  1 year (%) 1 year (%) 2 year (%) 2 year (%) 3year (%) 3year (%)
Risk group
 D V
 D V D V
0-3 6.4 6.8 13.8 14.5 20.9 18.1
4-5 12.1 10.8 24.3 24.8 36.2 35.7
>5 20.6 22.2 39.7
 40.5 54.1 55.1

D= development group
V= validation group

Concerns Regarding Methodology, Applicability to Older Adults, etc.: 
  • The baseline demographics are different in the 2 groups as expected due to geography, but the tool was still validated.
  • May have wanted to include in the baseline demographics ie BMI or fall history to give us a better sense of the frailty of the patients as well as include other risk factors for mortality.
  • This may be helpful to give peace of mind to families, but mortality at 3 years isn’t really too helpful, as you cannot refuse the resources. 1 year mortality may be more helpful.
  • The outcomes were only mortality and so it does not give you a better handle of the concern of families as to what morbidity to be expected especially in terms of the caregiver burdens.
  • Important to note the population that these are patients who already are eligible for LTC and these are different than many of our ambulatory patients. 1/3 of the patients enrolled in PACE died within 3 years which is higher than the US life tables for 79 year olds, so this begs the question if the study population from the outset was too specific to be generalized.
  • The validation group was again the “same” population – a PACE program, just in a different geographical location. It would have been better for validity to have a different population used to validate the rule.
Funding Source and Role: 
None
Created By: 
Shira Goldberg, MD - Geriatric Fellow Brookdale Department of Geriatrics, Mount Sinai School of Medicine