Skip to main content
CLP 2024 Annual Meeting Main banner
Final Program


CLP 2024 Program
Back

Brief Oral Papers

Neuropsychiatry

Brief Oral Papers Session 03 - Delirium

Dlin/Fischer Clinical Research Award: Assessing the Diagnostic Accuracy of the Stanford Proxy Test for Delirium (S-PTD) in a Critically Ill Population

Friday, November 8, 2024
11:05 AM – 12:20 PM EST
Location: Jasmine
Abstract:

Background: Delirium is the most common psychiatric disorder in general settings, with an incidence of up to 87% in intensive care units. While many delirium screening tools report high sensitivities in controlled research environments, these seem to decrease under real-life conditions.

Methods: A triple-blinded study assessed the efficacy of the two delirium screening tools utilized at our institution, CAM-ICU and SPTD, against a neuropsychiatric assessment based on DSM-5 criteria. All patients had been admitted to the adult cardiothoracic-ICU at Stanford Medicine. Patient data was pooled from an ongoing clinical trial intended to determine the accuracy of an EEG-based diagnosis for delirium. The CAM-ICU and SPTD were compared against a DSM-5 based neuropsychiatric exam, considered the gold standard for delirium diagnosis in our study and performed by a critical care psychiatrist.

Results: As of this submission, we have successfully enrolled 142 (age range 29 – 92 y/o, 66.2% male; 90.7% surgical cases) patients over 32 weeks. Our objective is to enroll a total of 200 patients over the course of the study. By the time of the conference in November 2024, we anticipate having recruited the entire study population and have completed data analysis. Partial analysis of the initial 32 weeks of the study is as follows: 429 clinical assessments utilizing the various delirium screening tools were conducted. Of these 115 (26.8%) yielded a positive diagnosis of delirium as per the DSM-5-based assessment. Forty-five CAM-ICU exams were unable to be assessed due to patient uncooperativeness; all of which were clearly identified by the blinded psychiatric assessment. There were no instances when the SPTD exam was not able to be performed; The CAM-ICU displayed a sensitivity of 22.8% (failing to detect 78.6%, 69.4%, and 100.0% of hypoactive, mixed, and hyperactive delirium cases, respectively), specificity of 99.7%, PPV of 96.3%, NPV of 88.1%, and accuracy of 79.4%. The SPTD demonstrated a sensitivity of 94.2% (failing to detect 9.2%, 0.0%, and 0.0% of hypoactive, mixed, and hyperactive delirium cases, respectively), specificity of 82.1%, PPV of 64.1%, NPV of 97.9%, and an accuracy of 85.2%.

Discussion: In our sample, that breakdown is as follows: hypoactive (57.7%), mixed (29.3%), and hyperactive (6.5%), which is consistent with previously published data. In addition to classic delirium, the DSM-5 based assessment identified 19 cases of subsyndromal delirium; of those 15 (78.9%) were appropriately identified by the SPTD, all of which were missed by the CAM-ICU. Although both tools demonstrate similar accuracy, the clinical relevance of a screening tool relies on its sensitivity, clearly superior for the S-PTD.



Conclusion: Among the most common delirium screening tools at our institution, the SPTD seems more accurate than traditional tools. It is the only validated tool against the DSM-5 criteria, demonstrating a higher sensitivity and overall accuracy in our sample, along with accommodating patients unable to participate in traditional clinical assessments.

Sponsored by the Critical Care Psychiatry SIG.