Larry Fennigkoh

Data Mining CMMSs: How to Convert Data into Knowledge

Although the healthcare technology management (HTM) community has decades of accumulated medical device-related maintenance data, little knowledge has been gleaned from these data. Finding and extracting such knowledge requires the use of the well-established, but admittedly somewhat foreign to HTM, application of inferential statistics. This article sought to provide a basic background on inferential statistics and describe a case study of their application, limitations, and proper interpretation. The research question associated with this case study involved examining the effects of ventilator preventive maintenance (PM) labor hours, age, and manufacturer on needed unscheduled corrective maintenance (CM) labor hours. The study sample included more than 21,000 combined PM inspections and CM work orders on 2,045 ventilators from 26 manufacturers during a five-year period (2012-16). A multiple regression analysis revealed that device age, manufacturer, and accumulated PM inspection labor hours all influenced the amount of CM labor significantly (P < 0.001). In essence, CM labor hours increased with increasing PM labor. However, and despite the statistical significance of these predictors, the regression analysis also indicated that ventilator age, manufacturer, and PM labor hours only explained approximately 16% of all variability in CM labor, with the remainder (84%) caused by other factors that were not included in the study. As such, the regression model obtained here is not suitable for predicting ventilator CM labor hours.

Fabrication of a nursing manikin overlay for simulation of chest drainage management

Medical manikins are a type of patient simulator used to train medical staff. The benefits of using a medical manikin simulator include the suspension of disbelief, simultaneous team and individual learning, allowable failure, personalized scenarios, frequent repetition and a focus on the needs of the learner rather than the patient. Unfortunately, one of the limitations of medical manikins is that they cannot simulate every problem that medical staff comes across. The SimMan (Laerdal) manikins do not have the ability to simulate when excess air or fluid is in the pleural space of the lung or thorax, the drainage of which must be managed by nurses. These three symptoms are called a pneumothorax, pleural effusion and excess thoracic blood, respectively. The purpose of this research was to create an overlay for the SimMan manikin to simulate a pneumothorax, pleural effusion or excess of thoracic blood for nurses to learn and practice responsibilities during chest drainage. The process began by obtaining a scan of a manikin from the Milwaukee School of Engineering nursing department. Simultaneously developed, a pneumothorax, pleural effusion and excess thoracic blood simulator became the basis for the layout of the overlay. Beginning with the computer scan, a mold of the overlay was shaped using computer automated design programs. This mold was customized to accommodate all necessary components and ports of the simulator. The mold was printed using a stereolithography (SLA) machine and created by layering silicone to make the housing for the simulator. The overlay resulting from this research is currently being used to train nursing students at the Milwaukee School of Engineering.

The complexities of the human-medical device interface.

Human factors, as a highly interdisciplinary and systems science, has already established the tremendously important role associated with the human-machine interface in ensuring safe, usable, and human-friendly products. For users of medical devices, however, this interface is significantly more complex because the patient is often, quite literally, an extension of the device. Through their direct electrical (e.g., electrocardiogram), fluidic (e.g., arterial blood pressure line), and/or pneumatic (e.g., ventilator) connections to various medical devices, patients become one with the technology. Not only are the machines capable of altering patient physiology, but the patient is also capable of altering the machine, e.g., through the triggering of alarms. Such an alliance between patient and machine is not only dynamic, it may also be highly fragile and unstable depending on the changing pathology of the patient. This complexity places a significantly increased cognitive burden on the clinical users of medical devices. Correspondingly, designers of medical devices need to be acutely aware of this burden when designing and testing the effectiveness and quality of their product interface. Hospital-based biomedical equipment technicians (BMETs) and clinical engineers will also do well to stay mindful of these complexities during device troubleshooting and, especially, when introducing new technology into the hospital environment. The focus of this paper is on the subtle, almost subliminal nature associated with many aspects of the human-medical device interface, some of which include the simple use or misuse of color, obscure panel labeling, or switch behavior. Collectively, however, they all may affect how the machine and clinician user will eventually “get along” with each other—and how effectively and safely the device will be used. The profoundly-instrumented, acutely-ill patient represented in Figure 1 continues to be more the norm rather than the exception in many hospital critical care units. Such patients may have multiple indwelling catheters, be connected to multiple infusion pumps, have direct electrical connections to physiological monitors and/or electrosurgical ground electrodes, and be intubated and connected to a ventilator as well. Obtaining or delivering clinically useful, real-time information or therapeutics is, in fact, precisely why these patients are instrumented in the first place. These obvious physical connections to and from the patient, however, do not represent the only interface. As further illustrated in Figure 2, there are additional “connections” between the clinician and the device and between the clinician and patient. These create, in effect, a closed-loop system. The overall safety and effectiveness of this patient-device-clinician system may also be profoundly affected by the environment in which they all reside. It is in this total systems context that the full impact, importance, and complexities of these multiple interfaces become apparent. Most of the medical errors related to such systems ocLarry Fennigkoh, PhD, PE, CCE is a professor within the biomedical engineering program at the Milwaukee School of Engineering. Email: fennigko@msoe.edu Figure 1. Through their direct connection to many medical devices, critical care patients become extensions of these devices.