SUPPORT Datasets | Datasets Archived at ICPSR

The support dataset is a random sample of 1000 patients from Phases I & II of SUPPORT (Study to Understand Prognoses Preferences Outcomes and Risks of Treatment). This dataset is very good for learning how to fit highly nonlinear predictor effects, imputing missing data, and for modeling 4 types of response variables:
  • Binary (hospital death)
  • Ordinal (severe functional disability - see below)
  • Continuous (total hospital costs or possibly study length of stay)
  • Time until event (death) with waning effects of baseline physiologic variables over time. Patients are followed up to 5.56 years.

See the following reference for more information about the SUPPORT study (funded by the Robert Wood Johnson Foundation): Knaus WA, Harrell FE, Lynn J et al. (1995): The SUPPORT prognostic model: Objective estimates of survival for seriously ill hospitalized adults. Annals of Internal Medicine 122:191-203.

You may want to use the following normal values that have been found to be satisfactory in imputing missing baseline physiologic data:

Baseline Variable Normal Fill-in Value
Serum albumin 3.5
PaO2/FiO2 ratio (pafi) 333.3
Bilirubin 1.01
Creatinine 1.01
BUN 6.51
White blood count 9 (thousands)
Urine output 2502

The ordinal functional disability variable sfdm2 has the following levels (in order):

Level Meaning
1 Patient lived 2 months to be able to get 2 month interview, and from this interview there were no signs of moderate to severe functional disability
2 Patient was unable to do 4 or more activities of daily living at month 2 after study entry. If the patient was not interviewed but the patient's surrogate was, the cutoff for disability was ADL ³ 5
3 Sickness Impact Profile total score at 2 months ³ 30
4 Patient intubated or in coma at 2 months
5 Patient died before 2 months after study entry

There are 159 patients surviving 2 months for whom there were no patient or surrogate interviews. These patients have missing sfdm2.

support2 Dataset

Data for all 9105 SUPPORT patients are available. These files contains all variables in the above sdd/support plus the following (see variable labels for more documentation). As before, most of the patient assessments are made on the patient's third study day.

VariableSorted ascending Meaning
aps APACHE III day 3 physiology score
ca Patient has cancer
dementia Dementia as a comorbidity
diabetes Diabetes as a comorbidity
dnr Patient had DNR order
dnrday Day of DNR order (<0 if before study)
hday Day in hospital at which patient entered study
prg2m Physician's 2-month survival estimate for pt.
prg6m Physician's 6-month survival estimate for pt.
sps SUPPORT day 3 physiology score
surv2m SUPPORT model 2-month survival estimate
surv6m SUPPORT model 6-month survival estimate

To develop models without using findings from previous models, be sure not to use aps, sps, surv2m, surv6m as predictors. You also will probably not want to use prg2m, prg6m, dnr, dnrday.
Topic attachments
I Attachment Action Size Date Who Comment
SUPPORTmodels.pdfpdf SUPPORTmodels.pdf manage 72.0 K 19 May 2010 - 14:53 FrankHarrell Prognostic models from SUPPORT (without typos in published models)
Topic revision: r2 - 19 May 2010, FrankHarrell
 

This site is powered by FoswikiCopyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback