Kidney Week

Abstract: PO2066

Customizing PROMIS-Depression Computer Adaptive Testing Stopping Rules for Patients with Kidney Failure

Session Information

• Transplantation: Clinical - Allocation, Evaluation, Prognosis, and Viral Onslaughts
  November 04, 2021 | Location: On-Demand, Virtual Only
  Abstract Time: 10:00 AM - 12:00 PM

Category: Transplantation

• 1902 Transplantation: Clinical

Authors

• Ahmed, Tibyan, University Health Network, Toronto, Ontario, Canada

Tibyan Ahmed

• Edwards, Nathaniel, University Health Network, Toronto, Ontario, Canada

Nathaniel Edwards,

• Aghamohammadi, Setareh, University Health Network, Toronto, Ontario, Canada

Setareh Aghamohammadi,

• Liu, Kaiyi, University Health Network, Toronto, Ontario, Canada

Kaiyi Liu,

• Yanga, Nawang, University Health Network, Toronto, Ontario, Canada

Nawang Yanga,

• Mucsi, Istvan, University Health Network, Toronto, Ontario, Canada

Istvan Mucsi,

Group or Team Name

• Kidney Health Education and Research Group

Background

The Patient Reported Outcomes Measurement Information System Depression (PROMIS-D) computer adaptive testing (CAT) allows precise and tailored assessment of depressive symptoms. Due to the default stopping rules, many respondents without depression may need to answer 10-12 items. The maximum number of items required by the stopping rule can be reduced which could improve efficiency when the tool is used for screening. We assess the screening performance of customized CAT stopping rules in patients with kidney failure.

Methods

A cross-sectional convenience sample of adults with kidney failure treated with dialysis or kidney transplant completed PROMIS-D CAT as well as the Patient-Health Questionnaire-9 (PHQ-9). Moderate/severe depressive symptoms were defined as a PHQ-9 cut-off score ≥10. Sociodemographic and clinical characteristics were obtained from self-report and medical records. All patients completed CAT using the original stopping rule (CAT_o) that requires a reliability of >90% or maximum 12 items. We compare this to three simulated CAT customizations with maximum 8, 6 and 4 items (CAT₈, CAT₆ and CAT₄) respectively. Reliable T score range (reliability is >90%), sensitivity and specificity of each version were assessed.

Results

Of the 336 patients, the mean SD age was 55(16), 63% were male, 49% were Caucasian and 32% were on dialysis. Based on PHQ-9, 16% reported moderate/severe depressive symptoms. Using a PHQ-9≥10 as a reference for moderate/severe depressive symptoms, sensitivity and specificity of a T score of 55 with CAT_o was 79% and 81% respectively.
CAT₈ presented no change in the reliable range (T-score 41 to 84), while CAT₆ and CAT₄ presented a small reduction in the reliable range (41-76 and 45-73 respectively) compared to CAT_o. Sensitivity and specificity of the modified CAT versions remained essentially the same.

Conclusion

Customizing PROMIS-D CAT stopping rules have the potential to improve efficiency of screening for moderate/severe depressive symptoms. This reduces question burden without change in the discrimination of the T score. A PROMIS-D CAT with modified stopping rule (maximum 6 or even 4 items) could be used for screening for depressive symptoms among patients with kidney failure.