Geriatric Events Among Older Adults Undergoing Nonelective Surgery Are Associated with Poor Outcomes

Abstract

Older adults undergoing nonelective surgery are at risk for geriatric events (GEs: delirium, dehydration, falls/fractures, failure to thrive, and pressure ulcers), but the impact of GEs on postoperative outcomes is unclear. Using the 2013 to 2014 National Inpatient Sample, we analyzed nonelective hospital admissions for five common operations (laparoscopic cholecystectomy, colectomy, soft tissue debridement, small bowel resection, and laparoscopic appendectomy) in older adults (aged ≥65 years) and a younger referent group (aged 55-64 years). Nationally weighted descriptive statistics were generated for GEs. Logistic regression controlling for patient, procedure, and hospital characteristics estimated the association of 1) age with GEs and 2) GEs with outcomes. Of 471,325 overall admissions, 64.7 per cent were aged ≥65 years. The rate of any GE in older adults was 26.9 per cent; GEs varied by age and procedure (P < 0.001). After adjustment, the probability of any GE increased with age category (P < 0.001); having any GE was associated with higher probability of all outcomes (P < 0.001): mortality (4.5% vs 0.8%), postoperative complications (61.7% vs 24.9%), prolonged length of stay (24.3% vs 7.9%), and skilled nursing facility discharge (46.6% vs 10.3%). In addition, there was a dose-response relationship between GEs and negative outcomes. GEs are prevalent in the nonelective surgery setting and associated with worse clinical outcomes. Quality improvement efforts should focus on addressing GEs.

Figure 1:. Adjusted Probability of Outcomes by Number of Geriatric Events

Probabilities are derived using multivariable logistic regression models using each outcome (eg, inpatient mortality) as the dependent variable, number of geriatric events (0, 1, 2, 3+) as the predictor, adjusted for patient (age, gender, race, comorbidities), procedure, and hospital characteristics (size, location, teaching status, region, ownership). A single category is depicted for 3 or more geriatric events given the small sample of observations with 4 (n = 265) and 5 geriatric events (n = 5). Prolonged length of stay is defined as the procedure-specific top decile of length of stay.