Artificial Intelligence (AI) has diffused into many areas of our private and professional life. In this research note, we describe exemplary risks of black-box AI, the consequent need for explainability, and previous research on Explainable AI (XAI) in information systems research. Moreover, we discuss the origin of the term XAI, generalized XAI objectives, and stakeholder groups, as well as quality criteria of personalized explanations. We conclude with an outlook to future research on XAI.

Background There is no consensus on the instruments for diagnosis of post-intensive care syndrome (PICS). We present a proposal for a set of outcome measurement instruments of PICS in outpatient care. Methods We conducted a three-round, semi-structured consensus-seeking process with medical experts, followed each by exploratory feasibility investigations with intensive care unit survivors (n1 = 5; n2 = 5; n3 = 7). Fourteen participants from nine stakeholder groups participated in the first and second consensus meeting. In the third consensus meeting, a core group of six clinical researchers refined the final outcome measurement instrument set proposal. Results We suggest an outcome measurement instrument set used in a two-step process. First step: Screening with brief tests covering PICS domains of (1) mental health (Patient Health Questionnaire-4 (PHQ-4)), (2) cognition (MiniCog, Animal Naming), (3) physical function (Timed Up-and-Go (TUG), handgrip strength), and (4) health-related quality of life (HRQoL) (EQ-5D-5L). Single items measure subjective health before and after the intensive care unit stay. If patients report new or worsened health problems after intensive care unit discharge and show relevant impairment in at least one of the screening tests, a second extended assessment follows: (1) Mental health (Patient Health Questionnaire-8 (PHQ-8), Generalized Anxiety Disorder Scale-7 (GAD-7), Impact of Event Scale – revised (IES-R)); (2) cognition (Repeatable Battery for the Assessment of Neuropsychological Status (RBANS), Trail Making Test (TMT) A and B); (3) physical function (2-Minute Walk Test (2-MWT), handgrip strength, Short Physical Performance Battery (SPPB)); and (4) HRQoL (EQ-5D-5L, 12-Item WHO Disability Assessment Schedule (WHODAS 2.0)). Conclusions We propose an outcome measurement instrument set used in a two-step measurement of PICS, combining performance-based and patient-reported outcome measures. First-step screening is brief, free-of-charge, and easily applicable by health care professionals across different sectors. If indicated, specialized healthcare providers can perform the extended, second-step assessment. Usage of the first-step screening of our suggested outcome measurement instrument set in outpatient clinics with subsequent transfer to specialists is recommended for all intensive care unit survivors. This may increase awareness and reduce the burden of PICS. Trial registration This study was registered at ClinicalTrials.gov (Identifier: NCT04175236; first posted 22 November 2019).

No abstract