Surprise Question Associated with Risk of Hospitalization in Patients with CKD

San Diego—Patients with chronic kidney disease (CKD) are at high risk for hospitalization. The surprise question (SQ), “Would you be surprised if this patient died within the next 12 months?,” is a provider-based subjective health measure. There have been studies examining the association between the SQ and the risk of mortality in patients with advanced CKD; however there are few data on the relationship between the SQ and the risk of hospitalization.

Nicolas Baddour and colleagues at the Vanderbilt University School of Medicine, Nashville, Tennessee, conducted  a study to examine this association. Results were reported during a poster session in a poster titled Surprise Question and Time-to-Hospitalization in Advanced CKD.

The study included 488 ambulatory patients with stage 4 or 5 CKD. Inclusion criteria were age ≥60 years and treatment in a nephrology clinic. Following each office visit, the providers were asked the SQ using a binary and 5-point Likert scale. Linkage to a statewide health database provided data on hospitalizations; linkage with national or local public health records provided data on deaths. Chart review provided data on other covariates using structured forms. Cox proportional hazards regression was used to determine hazard ratios for time to first hospitalization.

In addition to the SQ Likert scale, the multivariable model included age, sex, race, marital status, insurance status, diabetes, hypertension, cardiovascular disease, heart failure, malignancy, liver disease, chronic lung disease, number of hospitalizations in the prior year, number of medications, body mass index, estimated glomerular filtration rate, serum albumin, and hemoglobin.

During the study period, 60% of the total cohort (n=293) were hospitalized. In patients with SQ binary “surprised” answer, the median time to first hospitalization was 591 days compared with 277 days in the “not surprised” counterparts. Three point three percent of patients died before a hospitalization.

There was a significant association (P=.01) between the Likert SQ response and the time to a hospitalization. Hazard ratios (95% confidence interval) were 1.67 (0.88-3.17), 2.25 (1.46-3.5), 1.60 (1.08-2.37), and 1.51 (1.01-2.26) for 1 (not at all surprised), 2, 3, and 4, respectively (versus reference: 5, very surprised).

In summary, the researchers said, “Provider response to the SQ is associated with time-to-hospitalization in older adults with stage 4 and 5 CKD. Future studies should examine whether the SQ is useful in identifying patients with high healthcare utilization.”

Source: Baddour N, Robinson-Cohen C, Lipworth L, Bian A, Stewart TG, Abdel-Kader K. Surprise question and time-to-hospitalization in advanced CKD. Abstract of a poster (TH-PO1144) presented at the American Society of Nephrology Kidney Week 2018, October 25, 2018, San Diego, California.