# Blog Viewer

### Evolving Role of Analytics and Ethics in Bedside Medicine if ($('.thread-icon-follow').length !== 0) {$('.thread-icon-follow').on('click', function () { var $this =$(this); hl_common_ui_blockUI(); HigherLogic.Util.post( '/higherlogic/ui/mvc/Directory/Following/ToggleFollower', JSON.stringify({ itemKey: $this.data('item-key') }), 'json' ).done(function() { var icon =$this.find('i'); icon.toggleClass('fal').toggleClass('fas'); icon.prop('title', icon.hasClass('fal') ? 'Subscribe' : 'Unsubscribe'); \$this.attr('aria-label', icon.hasClass('fal') ? 'Subscribe' : 'Unsubscribe'); hl_common_ui_unBlockUI(); }); }); }

#### By John Zaleski posted 02-06-2018 03:56

The Certified Analytics Professional Code of Ethics (See CAP Ethics) is very important to me personally and professionally in my role as a Chief Analytics Officer and in my conduct with clients in the healthcare domain. I work closely with medical professionals in defining and implementing methods for guiding intervention at the point-of-care per the practice of medicine. It occurred to me that while there are many analytics professionals who ply their practice to healthcare, it might be useful to share a particular application relative to the real-time intervention and guidance of interventions relative to patient care.

An area of increasing importance and focus over the past 8-10 years has been in the area of clinician alarm fatigue and reducing false, non-actionable alarm signals at the bedside. Clinical alarm signals are issued by real-time monitoring and therapeutic equipment (e.g.: physiologic monitors, mechanical ventilators, anesthesia machines, pulse oximeters, etc.) Analytics professionals are most probably aware and well-versed in the concepts of Type-I and Type-II error, and the identification and quantification of the null hypothesis. The null hypothesis concept is very important and applicable in medicine relative to the identification of patients who are experiencing real adverse events. Informing clinicians when a patient is deteriorating or is otherwise in crisis is extremely important as early intervention can and does make the difference between life and death. Unfortunately, it is estimated from studies that upwards of 99% of monitor-issued electrocardiograph (ECG) alarm signals are non-actionable (See Patient Safety Institute; AAMI Clinical Alarm Compendium). That is, most alarm signals have no clinical relevance or important in identifying true crises and this fact can result in "the boy who cried wolf" syndrome in which clinicians become "snow blind" to true events, oftentimes ignoring true crises as they become insensitive to them or becoming exhausted responding to non-actionable alarm signals issued by medical devices at the bedside.

Actionable ECG alarm signals from real-time physiologic monitors fall into many categories in terms of true crises and warnings, and the issuing of these alarm signals is dependent upon many clinical factors, some of which are controllable by the bedside clinicians. Examples of these factors include identifying limit thresholds on true tachycardia, bradycardia and atrial fibrillation, and identification of true asystole, ventricular fibrillation and ventricular tachycardia. Hospitals establish and convene clinical alarm committees focused on quantifying the numbers or counts of alarm signals that are issued in each department or unit, particularly those units in which high acuity monitoring is conducted (e.g.: intensive care, telemetry, emergency, post-anesthesia care). As part of this assessment, clinicians rely upon data analyzed and presented to provide a holistic view of the department, patient and alarm signal situation at hand so that adjustments to bedside medical devices can be conducted safely to reduce the quantities of issued alarm signals while maintaining the safety of the patient. To this end, the analytics professional (i.e., data scientist or analyst) will have access to raw, real-time data and has a very important role: understanding, cleaning, and assessing data accurately; presenting truthful and un-biased results; and protecting the identities of the patients involved.

The impact of the assessment made by the analytics professional will have direct impact on the clinical decision-making process. Ergo, adherence to high standards involving the data and respect and treatment of the patient stands out as major tenets of analytical practice in the healthcare environment.

The monitoring, therapy, and management of patients will increasingly rely on the use of both real-time and historical data. Consequently, clinicians and health systems are relying more and more on those who can analyze the data and present information in ways that are clear, actionable, accurate, and un-biased. Hence, the growing role of the analytics professional in healthcare will necessitate an increasing focus on high ethical standards.

John R. Zaleski, Ph.D., CAP, CPHIMS
Chief Analytics Officer
Bernoulli Health
http://bernoullihealth.com
jzaleski@bernoullihealth.com
+1 203-343-9225 (O)
+1 484-319-7345 (M)