New Book! Coming out late 2014 / early 2015

Implementing Medical Device Connectivity with Electronic Health Records and Data Warehouses: A Practicum for Hospital Enterprise Clinical and Information Technology Implementation Teams. ... Read More

View Next Project View Complete Portfolio

Dictionary of Computer Science, Engineering, and Technology

Zaleski, JR, (contributing Author), Dictionary of Computer Science, Engineering, and Technology, (CRC Press, Phil Laplante, Editor-in-Chief). ... Read More

View Next Project View Complete Portfolio

Integrating Device Data into the Electronic Medical Record

The seminal book on the interoperability of medical device data and health information technology systems. John Zaleski, Ph.D. A book on empirical practice of medical device interoperability, based on years of experience in the field. A Developer’s Guide to Des... Read More

View Next Project View Complete Portfolio

Medical Device Data and Modeling for Clinical Decision Making

This work combines much of the experience learned in medical device interoperability and clinical informatics I have gained over the course of the past 20+ years. I have leveraged work from my  PhD and experience in product management of critical care. The devic... Read More

View Next Project View Complete Portfolio

Weaning from Postoperative Mechanical Ventilation

Modeling Post-Operative Respiratory State in Coronary Artery Bypass Graft Patients: A Method for Weaning Patients from Mechanical Ventilation. This PhD research developed a model for real-time assessment of patient postoperative recovery and viability for weaning f... Read More

View Next Project View Complete Portfolio

Latest happenings from the blog

View All Posts In My Blog »

Tracking Blood Glucose, Part 1



Diabetes diagnosis is typically associated with the following [1, 2]:

  • Two or more consecutive fasting glucose levels ≥ 126 mg/dL;
  • Any single ad-hoc measurement of blood glucose > 200 mg/dL;
  • A1c test result ≥ 6.5%; or,
  • A 2-hour oral glucose measurement value > 200 mg/dL.

Normal blood glucose levels during fasting (e.g.: usually morning levels, for which fasting has been undertaken for 8 hours nominally) should range between 70 and 99 mg/dL. Normal post-meal blood glucose (2 hours or so after eating) < 140 mg/dL. An important part of diabetes management is measurement: the old adage, you cannot control what you cannot measure, applies. Hence, in the process of managing one’s diabetes (through diet, exercise, insulin and oral medications), measurement of the blood glucose is key to determining, empirically, whether behaviors and diet are working (i.e., “good” behaviors) or a not working (i.e., “bad” behaviors). When it comes to diabetes management, the patient is the only one who can control it. Physicians, dieticians, and other members of the care team can facilitate, prescribe, assist and guide. But, ultimately, it is only the patient—the individual diagnosed with the ailment—who has absolute control over the effects of diabetes [3]. As a diabetic myself, I know that I am in control of my diet; I am in control of my weight & exercise regimen; and, I am in control of my medications—adhering to the regimens that work is completely in my control, and is my responsibility. The linked white paper provides some empirical assessment of glucose measurement in a single patient over a year, to illustrate the types of trending and potential benefits of following glucose closely for the diabetic patient.

[1] “What are normal blood glucose levels?”

[2] “How the Blood Sugar of Diabetes Affects the Body.”

[3] R. K. Bernstein, MD. “Dr. Bernstein’s Diabetes Solution: A Complete Guide to Achieving Normal Blood Sugars.” Little, Brown & Co. 1997. Page 3.


Mathematical Techniques for Mitigating Alarm Fatigue



Alarm Fatigue

“Hospital staff are exposed to an average of 350 alarms per bed per day, based on a sample from an intensive care unit at the Johns Hopkins Hospital in Baltimore.”[1]

From the same survey, almost 9 in 10 hospitals indicated they would increase their use of patient monitoring, particularly of Capnography and pulse oximetry, if false alarms could be reduced.[2]

“Of those hospitals surveyed that monitor some or all patients with pulse oximetry or Capnography, more than 65 percent have experienced positive results in terms of either a reduction in overall adverse events or in reduction of costs.”[3]

The problem with attenuating alarm data is achieving the balance between communicating the essential, patient-safety specific information that will provide proper notification to clinical staff while minimizing the excess, spurious and non-emergent events that are not indicative of a threat to patient safety. In the absence of contextual information, the option is usually to err on the side of excess because the risk of missing an emergent alarm or notification carries with it the potential for high cost (e.g.: patient harm or death).

The purpose of this study (downloadable here) is to look at the mathematics and some of the techniques and options available for evaluating real-time data. The objective is to suggest a dialog for further research and investigation into the use of such techniques as appropriate. Clearly, patient safety, regulatory, staff fatigue and other factors must be taken into account in terms of aligning on a best approach or practice (if one can even be identified). These aspects of alarm fatigue are intentionally omitted from the discussion at this point (to be taken up at another time) so that a pure study of the physics of the parameter data and techniques for analyzing can be explored.


[1] Ilene MacDonald, “Hospitals rank alarm fatigue as top patient safety concern”, Fierce Healthcare. January 22, 2014.

[2] Wong, Michael; Mabuyi, Anuj; Gonzalez, Beverly; “First National Survey of Patient-Controlled Analgesia Practices.” March-April 2013, A Promise to Amanda Foundation and the Physician-Patient Alliance for Health & Safety.

[3] Ibid.


Early identification of impending events through physiologic surveillance



In the October 2006 issue of Resuscitation, Smith etal. [1] published an article on the benefits of early warning associated with the monitoring and physiologic surveillance of patients in the hospital (ICU, principally). From the abstract of that publication:

“Hospitalised patients, who suffer cardiac arrest and require unanticipated intensive care unit (ICU) admission or die, often exhibit premonitory abnormalities in vital signs…”

“…It is possible for raw physiology data, early warning scores (EWS), vital signs charts and oxygen therapy records to be made instantaneously available to any member of the hospital healthcare team via the W-LAN or hospital intranet…”

From another source [2]:

“Physiologic monitoring systems measure [pulse], blood pressure, …other vitals…Data about adverse events in hospitalized patients indicate…a majority of physiologic abnormalities are not detected early enough to prevent the event, even when some…abnormalities are present for hours before…[occurrence].”

Early warning and physiologic surveillance are not new concepts, whether in the ICU or elsewhere. What, perhaps, has evolved over the past 10 years or so since the formal introduction of the electronic medical record (EMR) is that the automated and complete collection of data normally charted within the EMR is necessary to support such early warning protocols (particularly outside of the ICU environment) so long as the data available are part of an integrated delivery system [3]. That is, complete and contextual data are necessary to promote accurate early warning notifications that can be developed from multiple sources of data, inclusive of physiologic, laboratory, and demographic.

Early warning score predictors can include:

a. vital signs

b. laboratory test results

c. severity of illness scores

d. longitudinal chronic illness burden scores

e. transpired length of hospital stay, and

f. care directives

Escobar [3] reported that of 4,036 events from a cohort of 102,422 patients, modified early warning score c-statistics of 0.709 at 95% confidence. In comparison to EMR-based models, which had a c-statistic of 0.845. Best early warning performance was detected amongst those patients with gastrointestinal (GI) diagnoses (0,841) and worst amongst those with congestive heart failure (CHF) (0.683).

While performance was less robust among the modified early warning models compared with EMR-based models, the performance correlation between the two is encouraging. Perhaps a place exists for the use of early warning protocols and methods which can be based less on the availability of sophisticated information and more on the availability of data readily collected from the bedside.

[1] Smith GB, etal., “Hospital-wide physiological surveillance–a new approach to the early identification and management of the sick patient.” Resuscitation. 2006 Oct;71(1):19-28. Epub 2006 Aug 30.

[2] Yoder-Wise, Patricia S., Leading and Managing in Nursing: fifth Edition. Elsevier-Mosby. 2014. ISBN: 978-0-323-24183-0. Page 201

[3] Escobar etal., Early Detection of Impending Physiologic Deterioration Among Patients Who are Not in Intensive Care: Development of Predictive Models Using Data From an Automated Electronic Medical Record.” Journal of Hospital Medicine. 2012 Society of Hospital Medicine. DOI 10.1002/jhm.1929.



Wordpress SEO Plugin by SEOPressor

Switch to our mobile site