Background

Machine learning (ML) has captured the attention of many clinicians who may not have formal training in this area but are otherwise increasingly exposed to ML literature that may be relevant to their clinical specialties. ML papers that follow an outcomes-based research format can be assessed using clinical research appraisal frameworks such as PICO (Population, Intervention, Comparison, Outcome). However, the PICO frameworks strain when applied to ML papers that create new ML models, which are akin to diagnostic tests. There is a need for a new framework to help assess such papers.

Objective

We propose a new framework to help clinicians systematically read and evaluate medical ML papers whose aim is to create a new ML model: ML-PICO (Machine Learning, Population, Identification, Crosscheck, Outcomes). We describe how the ML-PICO framework can be applied toward appraising literature describing ML models for health care.

Conclusion

The relevance of ML to practitioners of clinical medicine is steadily increasing with a growing body of literature. Therefore, it is increasingly important for clinicians to be familiar with how to assess and best utilize these tools. In this paper we have described a practical framework on how to read ML papers that create a new ML model (or diagnostic test): ML-PICO. We hope that this can be used by clinicians to better evaluate the quality and utility of ML papers.

Subtle alterations in how cortical network dynamics are modulated by different behavioral states could disrupt normal brain function and underlie symptoms of neuropsychiatric disorders, including Fragile X syndrome (FXS). Using two-photon calcium imaging and electrophysiology, we recorded spontaneous neuronal ensemble activity in mouse somatosensory cortex. Unanesthetized Fmr1(-/-) mice exhibited abnormally high synchrony of neocortical network activity, especially during the first two postnatal weeks. Neuronal firing rates were threefold higher in Fmr1(-/-) mice than in wild-type mice during whole-cell recordings manifesting Up/Down states (slow-wave sleep, quiet wakefulness), probably as a result of a higher firing probability during Up states. Combined electroencephalography and calcium imaging experiments confirmed that neurons in mutant mice had abnormally high firing and synchrony during sleep. We conclude that cortical networks in FXS are hyperexcitable in a brain state-dependent manner during a critical period for experience-dependent plasticity. These state-dependent network defects could explain the intellectual, sleep and sensory integration dysfunctions associated with FXS.

As the Coronavirus disease 2019 (COVID-19) pandemic spread to the US, so too did descriptions of an associated coagulopathy and thrombotic complications. Hospitals created institutional protocols for inpatient management of COVID-19 coagulopathy and thrombosis in response to this developing data. We collected and analyzed protocols from 21 US academic medical centers developed between January and May 2020. We found greatest consensus on recommendations for heparin-based pharmacologic venous thromboembolism (VTE) prophylaxis in COVID-19 patients without contraindications. Protocols differed regarding incorporation of D-dimer tests, dosing of VTE prophylaxis, indications for post-discharge pharmacologic VTE prophylaxis, how to evaluate for VTE, and the use of empiric therapeutic anticoagulation. These findings support ongoing efforts to establish international, evidence-based guidelines.