Deschasse et al reported scores for predicting mortality of a geriatric patient following discharge from an acute geriatric unit in the hospital. These can help to identify a patient who should have advanced care planning. The authors are from multiple institutions in France.
Patient selection: geriatric patient being discharged from an acute geriatric unit, age >= 75 years
Outcome: 3-month (90-day) mortality
Parameters:
(1) Katz ADL score at home
(2) age in years
(3) serum albumin in g/L
(4) Charlson comorbidity index (CCI)
(5) time interval between hospital admission and admission to acute geriatric unit (AGU) in days
(6) length of delay in days (presumably delay in discharge)
(7) appropriate length of stay in days
(8) hospitalized in previous 6 months
(9) living at home
(10) male sex
(11) body mass index (BMI) in kg per square meter
(12) community-acquired infection
(13) number of drugs taken at home
(14) number of psychotropic medications
|
Parameter
|
Finding
|
Points
|
|
Katz ADL
|
|
-0.191 * (Katz score)
|
|
age in years
|
|
0.033 * (years)
|
|
serum albumin
|
|
-0.066 * (albumin)
|
|
Charlson comorbidity
|
|
0.144 * (CCI)
|
|
time interval
|
< 2 days
|
0
|
|
|
>= 2 days
|
0.366
|
|
length of delay
|
|
0.042 * (delay)
|
|
appropriate length of stay
|
yes
|
0.059 * (days)
|
|
hospitalized in past 6 months
|
no
|
0
|
|
|
yes
|
0.344
|
|
living at home
|
no
|
0
|
|
|
yes
|
-0.525
|
|
sex
|
female
|
0
|
|
|
male
|
0.279
|
|
body mass index
|
|
-0.051 * (BMI)
|
|
community acquired infection
|
no
|
0
|
|
|
yes
|
-0.415
|
|
number of drugs at home
|
|
0.044 * (number)
|
|
number of psychotropic drugs
|
|
-0.359 * (number)
|
value of X =
= [SUM(points for all of the parameters) - 1.637] * 0.864
probability of 3 month mortality =
= 1 / (1 + EXP((-1) * X))
where:
• 0.864 is the shrinkage factor estimated using bootstrap validation.
Performance:
• The c-statistic was 0.73.
