Routine pre‐operative modified telephone interview for cognitive status screening: a reply

Van Zuylen poses important questions about how and in whom peri-operative cognitive screening should be implemented [1]. We reported that the modified telephone interview for cognitive status (TICS-M) was a feasible, remotely administered screening tool and that baseline scores were associated with cognitive classification by an expert panel at 12- and 24-month follow-up. We recommended its use for older adults aged ≥65 years in the peri-operative setting as a risk stratification tool [2]. This suggestion was intended to support best practice guidelines published by several independent bodies, including the Perioperative Neurotoxicity Working Group and the American Geriatrics Society Expert Panel [3], that recommend pre-operative cognitive screening as a tool to identify risk of postoperative delirium. Although the TICS-M showed utility in our study population, we agree that there are many tools that may be appropriate and effective. We agree that the question of how and in whom peri-operative cognitive screening should be implemented is complex and that feasibility alone is not sufficient justification for broad adoption. Our study included patients with subjective cognitive complaints or previously identified mild cognitive impairment, reflecting a higher risk group than may be routinely encountered in a peri-operative setting. We acknowledge, both here and in our study, that this limits generalisability to older adults without cognitive concerns. We also agree that the TICS-M is a screening rather than diagnostic instrument and should never be used in isolation to ‘diagnose’ peri-operative neurocognitive disorders. Instead, neuropsychological tests interpreted in the context of patient and informant reports and relevant medical history by trained clinicians should serve as the gold standard. Nonetheless, these assessments are likely beyond the capacity and scope of most peri-operative settings. In the absence of a gold standard assessment, van Zuylen highlights appropriately the importance of ensuring that screening tools, such as the TICS-M, measure constructs that are clinically relevant. Our study assists in answering this question by showing an association between TICS-M scores and real-world cognitive classifications according to an expert panel of neurologists, geriatricians and neuropsychologists at multiple time points. In our study, the consensus panel made use of a battery of paper and pencil neuropsychological tests measuring episodic and semantic memory, attention, language and executive functions. These were complemented by functional interviews and questions about participants' activities of daily living and are perhaps as close an approximation to a gold standard assessment as feasible in a peri-operative research setting. Only weak to moderate correlations between the TICS-M and neuropsychological tests across a series of cognitive domains were found by van Zuylen et al. [4]. Such a finding is common in the field of psychology when measuring latent constructs such as cognition, in which even tests measuring the same construct (i.e. executive function) reveal only moderate associations [5]. Weak associations such as those reported by van Zuylen et al. do not invalidate the utility of a test. An important consideration, which we welcome the opportunity to discuss, is that regardless of the test, the optimal implementation strategy for peri-operative cognitive screening will be influenced by the priorities of the individual service. Universal screening of all older patients having surgery is likely to identify more instances of cognitive impairment but will inevitably increase false positives. Conversely, selective or targeted screening based on subjective complaint or clinician concern will reduce false positives but fail to detect cases. Both approaches have implications for patients, clinicians and health systems and warrant explicit evaluation of the trade-off between costs and benefits. Our findings suggest that the TICS-M offers a feasible, low-burden option that could be integrated into peri-operative pathways to support risk stratification and clinical decision-making, particularly in settings where in-person cognitive assessment is impractical. It should not replace comprehensive assessment for diagnostic purposes, and it may not be suitable for all services. Nonetheless, brief assessments for older adult patients are recommended in best practice clinical guidelines and our study may assist clinicians and health systems to decide whether the TICS-M is suitable for their unique setting.