Use of Intensive Care Units for Patients with Spinal Disorders

This study describes the characteristics of patients with spinal disorders admitted to our ICUs and also assesses the value of two Illness Severity Scoring (ISS) systems and one therapeutic intervention scoring system (TISS) in identifying patients with low severity of illness who may not require ICU care.

Among all ICU admissions (1996–98) with a primary diagnosis of a spinal disorder (N=266), patients less than 17 years of age and those with trauma involving a second organ system other than spine were excluded. Data extracted included demographics, clinical details, vitals and laboratory values at admission and most abnormal value during the first 24 hours of ICU admission for 17 specific physiological variables.

The amount and type of monitoring and treatment received by each patient was recorded using the TISS. Among the 76 TISS tasks, 20 interventions were identified as ICU specific (16 treatment, 4 monitoring). ISSing was based on Acute Physiology and Chronic Health Evaluation (APACHE) 11 and the Mortality Prediction Model (MPMO) 11 methods.

Patients with predicted mortality risk of less than 5% were classified as having low severity of illness. Differences between these and other patients with regard to age, sex, admission type, diagnosis, LOS, total TISS score and use of ICU specific intervention were examined for statistical significance.

Among the 184 patients (114 males, 70 females, mean (±SD) age 48 (± 17) yrs) meeting our inclusion criteria, 54% had significant neurological deficits, 40% had comorbidities, and 61% were elective postoperative admissions. The median length of hospital and ICU stay (LOHS, LOICUS) for the group was 9 and 3 days respectively. The actual mortality for the whole group was 9.2%, which was not statistically different from that, predicted by the APACHE 11 and MPM 11 methods. Both ISS systems had good discriminating ability (area under receiver operating characteristic curve of 0.85 and 0.84) and excellent goodness of fit (Hosmer Lemeshow C statistic, p 0.5).

Patients with low severity of illness accounted for 64% of all admissions including 84% of elective postoperative and 5% of non–operative admissions. These patients were significantly younger, admitted electively, were postoperative, had lower LOHS, LOICUS & TISS scores and did not receive ICU specific intervention. A large proportion of patients with spinal disorders admitted to ICU has a low probability of death and does not receive ICU specific interventions. Routine use of ISSing systems can identify patients with low severity of illness who could be cared for out with the ICU.