Scoring heart rate increases as a surrogate arousal marker on portable monitor studies for obstructive sleep apnea: Impact on diagnostic accuracy and clinical decision-making.

Cortical arousal-related hypopneas are not scored on type 3 home devices, which therefore limits their diagnostic accuracy for obstructive sleep apnea. The objective of this study was to evaluate whether scoring heart rate accelerations as surrogate markers of arousal improves type 3 portable monitor diagnostic agreement compared with polysomnography and improves therapeutic decision-making. We prospectively recruited patients evaluated for obstructive sleep apnea to undergo in-laboratory simultaneous full polysomnography + type 3 portable monitoring. Hypopnea events were scored on portable monitor studies with and without autonomic scoring, which was defined as an associated increase in pulse oximetry-derived heart rate ≥6 beats per min (autonomic hypopnea). Portable monitor diagnostic agreement compared with polysomnography with and without autonomic hypopnea scoring was assessed. We also evaluated whether reporting autonomic hypopnea scoring improves portable monitor clinical treatment decision agreement after four physicians reviewed clinical data and sleep study results (polysomnography, portable monitor with autonomic hypopnea, portable monitor without autonomic hypopnea). Eighty-two participants completed simultaneous polysomnography and in-laboratory portable monitor studies. Scoring autonomic hypopnea resulted in a decreased mean difference between in-laboratory portable monitor respiratory event index and polysomnography apnea-hypopnea index in Bland-Altman analysis (mean difference 14.6 per hr without versus 6.1 per hr with autonomic hypopnea scoring [p ˂ 0.01]), and increased intraclass correlation from 0.769 to 0.844. Inclusion of autonomic hypopnea scoring resulted in better accuracy between portable monitor and polysomnography expert's treatment decisions, and ultimately resulted in 24% fewer additional polysomnographies requested. The addition of pulse oximetry heart rate increases for autonomic hypopnea scoring during portable monitor resulted in better diagnostic agreement, improved clinical decision-making and reduced additional polysomnography testing.

A pharmacist-physician intervention model using a computerized alert system to reduce high-risk medication use in primary care.

PURPOSE: Potentially inappropriate medications (PIMs) have been associated with a greater risk of adverse drug events and hospitalizations. To reduce PIMs use, a family health team (FHT) implemented a knowledge translation (KT) strategy that included a pharmacist-physician intervention model based on alerts from a computerized alert system (CAS). METHODS: Our pragmatic, single-site, pilot study was conducted in an FHT clinic in Quebec, Canada. We included community-dwelling older adults (≥ 65 years), with at least 1 alert for selected PIMs and a medical appointment during the study period. PIMs were selected from the Beers and STOPP criteria. The primary outcome was PIMs cessation, decreased dose, or replacement. The secondary outcome was the clinical relevance of the alerts as assessed by the pharmacists. RESULTS: During the 134 days of the study, the CAS screened 369 individuals leading to the identification of 65 (18%) patients with at least 1 new alert. For those 65 patients, the mean age was 77 years, men accounted for 29% of the group and 55% were prescribed 10 or more drugs. One or more clinically relevant alerts were generated for 27 of 65 included patients for an overall clinical relevance of the alerts of 42%. Of the 27 patients with at least 1 relevant alert, 17 (63%) had at least 1 medication change as suggested by the pharmacist. CONCLUSION: An interdisciplinary pharmacist-physician intervention model, based on alerts generated by a CAS, reduced the use of PIMs in community-dwelling older adults followed by an FHT.

Expected clinical utility of automatable prediction models for improving palliative and end-of-life care outcomes: Toward routine decision analysis before implementation.

OBJECTIVE: The study sought to evaluate the expected clinical utility of automatable prediction models for increasing goals-of-care discussions (GOCDs) among hospitalized patients at the end of life (EOL). MATERIALS AND METHODS: We built a decision model from the perspective of clinicians who aim to increase GOCDs at the EOL using an automated alert system. The alternative strategies were 4 prediction models-3 random forest models and the Modified Hospital One-year Mortality Risk model-to generate alerts for patients at a high risk of 1-year mortality. They were trained on admissions from 2011 to 2016 (70 788 patients) and tested with admissions from 2017-2018 (16 490 patients). GOCDs occurring in usual care were measured with code status orders. We calculated the expected risk difference (beneficial outcomes with alerts minus beneficial outcomes without alerts among those at the EOL), the number needed to benefit (number of alerts needed to increase benefit over usual care by 1 outcome), and the net benefit (benefit minus cost) of each strategy. RESULTS: Models had a C-statistic between 0.79 and 0.86. A code status order occurred during 2599 of 3773 (69%) hospitalizations at the EOL. At a risk threshold corresponding to an alert prevalence of 10%, the expected risk difference ranged from 5.4% to 10.7% and the number needed to benefit ranged from 5.4 to 10.9 alerts. Using revealed preferences, only 2 models improved net benefit over usual care. A random forest model with diagnostic predictors had the highest expected value, including in sensitivity analyses. DISCUSSION: Prediction models with acceptable predictive validity differed meaningfully in their ability to improve over usual decision making. CONCLUSIONS: An evaluation of clinical utility, such as by using decision curve analysis, is recommended after validating a prediction model because metrics of model predictiveness, such as the C-statistic, are not informative of clinical value.