2012 Symposium on Human Factors and Ergonomics in Health Care  

A Multidisciplinary Human-Centered Approach to the Design of Incubation Systems

Thomas Ferris and Mardelle Shepley Center for Health Systems & Design

Texas A&M University College Station, TX

Graduate students in a Cognitive Systems Engineering course collaborated with those in Architectural Design on a project that followed a human-centered design approach to the incubation system for a Neonatal Intensive Care Unit (NICU). This approach emphasized the needs of three user groups: infant patients, family members, and care givers. Cognitive task analyses with NICU nurses at a local hospital and heuristic evaluations of commercial equipment helped highlight problem areas which informed the design of prototype incubators and supporting equipment. This paper introduces the methods used and lessons learned by following this multidisciplinary approach, and highlights some of the analysis findings and proposed redesign features.

INTRODUCTION

The “incubation system” that is central to the care of ill

and/or pre-term infants in Neonatal Intensive Care Units (NICUs) includes neonatal incubators, interfacing peripheral equipment such as physiological monitors, mechanical ventilators, and fluid/nourishment/medication pumps. This system must meet the infant patient’s basic care needs by controlling the temperature and moisture content of the air; protecting them from drafts, auditory noise, infection, airborne irritants, and excessive handling; and supporting provision of oxygen, nourishment, fluids, and medication. It also must support the activities of family members and care givers by providing direct means of observing the patient, supporting parental bonding via manual interaction and nursing the child, and allowing ease of transportation and a range of access to the infant in case an emergency intervention is necessary. While most NICU incubation systems adequately support most of these basic functions, a number of design issues remain.

A human-centered design approach means carefully analyzing the needs of the system users and designing the technology to fit those needs (e.g., Wickens, Lee, Liu, & Gordon-Becker, 2004). In the case of the incubation system, the needs of three types of human users – infant patients, family members, and care givers – must be considered. It may be the case that some design features which are beneficial to some user types are detrimental to others, therefore the needs of all three groups must be considered together. For example, it is important that the system limits the infant’s exposure to bright light, and controls the timing of light/dark cycles which stimulate the production of hormones that are critical to growth and development. However, adequate lighting is critical for care givers to be able to observe the patient, and observation may be required often enough that it disrupts light/dark cycles. Other equipment that aids care givers in their responsibilities may inadvertently harm the patient, for example, by increasing the ambient sound level or exposing the patient to additional electromagnetic fields, both of which can adversely affect the patient’s development.

Pre-term and/or ill infants may have a limited ability to thermo-regulate and retain fluids, a compromised immune system, hypersensitivity to environmental conditions, and be

in a state of general discomfort. The incubator protects this vulnerable patient from inappropriate temperatures, drafts, noise, infection, air born irritants, and excessive handling. It may also include sensor and display functionality, either natively or via peripheral add-on technologies, to monitor physiological health parameters of the infant (such as blood pressure, cardiac activity, and respiratory activity) and to control the delivery of oxygen, nourishment, fluids, and medication. While most basic needs of the infant are met by the incubation system, there are remaining issues in controlling the infant’s exposure to ambient light, sound, and electromagnetic fields that should be improved in future incubation systems (e.g., Antonucci, Porcella, & Fanos, 2009; Cermakova, 2003).

Design of incubation systems has largely neglected consideration of the families of the patient. Parents of neonates can be under great amounts of physical and emotional stress, having recently endured hours of painful labor in delivery and possibly surgical intervention, and they may be severely sleep-deprived. Designs could do better to mitigate errors that people under these conditions tend to make, such as forgetting to wash or sanitize their hands before interacting with the child. Parents are also undoubtedly suffering from anxiety for the health of the infant. While the numerous sensors, wires, and tubes attached to and inserted in the infant help meet the infant’s basic needs and help care givers maintain awareness of their physiology, the image that this creates can have a de-humanizing effect that does little to ease the family members’ anxiety.

When considering the needs of the care staff, the functions of the incubation system can be analyzed from the perspective of cognitive systems engineering, which concerns the individual and interacting roles of human and machine cognitions for tasks conducted in a work domain (e.g., Woods & Roth, 1988). This perspective involves describing characteristics of the users, the design environment, and the required work tasks in ways that allow identification of design interventions so that tasks can be conducted more safely, efficiently, and/or with greater satisfaction.

Following this perspective, system users (caregivers) vary in their areas of expertise and experience levels. They suffer from various amounts of job-related stress and fatigue issues that can and do affect performance (e.g., Carayon & Gurses,

Cop

yrig

ht 2

012

Hum

an F

acto

rs a

nd E

rgon

omic

s S

ocie

ty. A

ll rig

hts

rese

rved

. 10.

1518

/HC

S-2

012.

9452

8940

1.03

2

187

2012 Symposium on Human Factors and Ergonomics in Health Care  2008; DeLucia, Ott, & Palmieri, 2009). In addition to accommodating these differences, the incubation system must support a range of cognitive states, which can include challenging states induced by time pressure, high workload, and the divided attention requirements of being responsible for multiple patients.

A key environmental issue that will impact the design of an incubation system is whether it will be located in a private room or shared room setting (e.g., Shepley, Harris, & White, 2008). These two settings will have different effects on ambient sound, temperature, interaction, family and caregiver stress, infection control, and access to auxiliary equipment.

Physiological monitoring is a key task conducted by caregivers within the incubation system, and can be characterized by a high demand for visual and attentional resources. In practice, alarms designed to alleviate these demands have a well-documented set of issues (e.g., Woods, 1995). The vast majority (up to 90%) of these alarms do not require a clinical response (Imhoff & Kuhls, 2006), and the resultant “cry wolf” effect limits their effectiveness. In the NICU, there are fewer clues for quickly verifying the validity of an alarm because an infant patient cannot verbally describe how they feel and visible signs of discomfort may be less observable in pre-term patients. The NICU also limits the attention-directing qualities of alarms because of the sensitivity of patients to loud noises and bright lights. Additionally, the floor layout of a NICU – in particular, whether the patients are in private rooms or a shared room setting – can significantly influence which strategies personnel use to maintain awareness of a patient’s physiology.

This report documents a multidisciplinary user-centered design approach, combining the fields of Cognitive Systems Engineering and Architectural Design to take a wider perspective of design issues for the incubation system, including psychosocial, ergonomic, usability, safety, and sustainability issues.

METHOD

During the Fall of 2011, 27 graduate students at Texas

A&M University, from two graduate classes – Cognitive Systems Engineering (Instructor: Thomas Ferris) and Architectural Design (Mardelle Shepley) formed multidisciplinary teams for this semester-long project. The teams focused on identifying problems encountered by, and designing to support the needs of, the three primary user groups: infants, family members, and care personnel.

The general approach taken by each group in this project was as follows: 1) initial data collection and identification of group design focus areas; 2) cognitive task analysis with

NICU caregivers; 3) usability analysis and evaluation of existing systems according to design principles; and 4) propose redesign solutions.

The initial data collection consisted of an extensive literature review to understand the history of incubation systems and some of the historical and current challenges to the care and safety of the infant. Additionally, the groups conducted interviews via Skype with domain experts (a neonatologist, a NICU architectural designer, and family members of infants who spent significant time in the NICU). Following the completion of this initial stage of data collection, groups made decisions regarding 1) the type of NICU model (e.g., open-bay/pod, or single-family room) that will be considered the context for their analysis and design; and 2) the subset of provider tasks (e.g., feeding, programming infusion pumps, administering various types of therapy, physiological monitoring) that they focused on to narrow the scope for the remainder of the project.

The second phase of the project involved using Cognitive Task Analysis methods to come up with a description for each groups’ task subset that could be used to identify problems and provide a roadmap for addressing them. The analysis was conducted by the student groups during visits to a local NICU at St. Joseph’s Regional Health Center in Bryan, TX. Each student group meeting was conducted with different NICU staff members and during different shifts/times of day. Task analysis methods used during these meetings included observation with questioning, structured interviews, and the critical incident technique (e.g., Kirwan & Ainsworth, 1992). Following this data collection, each group created a Hierarchical Task Analysis (HTA) tree that could be used to answer questions about how and why certain activities of interest were performed (see Figure 1).

Next, the groups interacted with incubation equipment that was loaned by GE Healthcare for the purpose of this project (the GE Giraffe OmniBed: http://www3.gehealthcare.com/en/Products/Categories/Maternal-Infant_Care/Giraffe_Omni-Bed_Incubator_and_Warmer). Following the subset of tasks described in the HTA, the students performed a usability heuristic analysis using Nielsen’s method and list of heuristics (http://www.useit.com/papers/heuristic/heuristic_evaluation.html) and/or a set of heuristics developed specifically for medical devices (e.g., Zhang, Johnson, Patel, Paige, & Kubose, 2003). Individual group members interacted with the system to identify their own set of design issues, which they then compiled with their group members and rated in terms of severity (see Figure 2). The most severe problems were highlighted to be addressed in redesign efforts.

188

 2012 Symposiu

Figure 2: E

um on Human F

Figure 1: Exam

Example excer

Factors and Er

mple excerpt fr

rpt from usabil

rgonomics in H

from HTA tree

lity heuristic an

Health Care

(credit: Wil-Jo

nalysis (credit:

ohneen Ardoin

Ivan Alexand

n, Li Cao, Kara

er, Nannan Ch

an Shukla)

hen, No Young You)

189

 

S

showdesigthe de

FigurSinghShukl

FigurShriv

2012 Symposiu

Schematics for n below, in Fign are given in esign highlight

re 3: POD Inch, Kara Wetzla)

re 4: ibaby Invastava, Kevin

um on Human F

RESUL

the proposed rgures 3 - 6. DeCHSD, 2012. Tts, often expres

cubator (creditel, Wil-Johne

ncubator (crediWalsh, Yunya

Factors and Er

LTS

redesigns for etailed descriptiTable 1 describssed in multiple

: Cameron Chen Ardoin, L

it: Ninad Athaan Li, Shiyan Y

rgonomics in H

ach group are ions for each bes some of e designs.

hristian, TanyaLi Cao, Karan

avale, SuruchiYang)

Health Care

a n

i

FigurSongNo Y

FigurKubi

re 5: Eggshellg, Fang Xu, GuYoung You)

re 6: Stork Iiak, Lauren Sob

l Incubator (cruang Yang, Iv

Incubator (crebecki, Kristen

redit: Katie Bvan Alexander,

dit: Cameron Miller, Jonpau

Bodolus, Yuting, Nannan Chen

Burke, Lancul Zelik)

g n,

e

190

2012 Symposium on Human Factors and Ergonomics in Health Care  Table 1: Examples of design highlights and design rationale Design highlights Rationale

Pivoting, height-adjustable bed

Support infant access and interaction for parents during various stages of recovery (e.g., in bed, chair, standing).

Personalizing through message screens and “crib” decorations

Support family members’ bonding with the infant. Reduce stress on family members by making incubator look more “crib-like”.

Color-coded visual sound meter

Supports awareness of ambient sound level, and captures attention when noise must be reduced.

Video display of infant

Supports observation of patient in low-light conditions, avoid exposing infant to bright lights or disrupting sleep cycle.

Motion-sensitive controls, such as alarm dismissal

Reduce manual interaction with the system, which reduces risk of infection.

Cable and tube management techniques (e.g., tensioners, wireless sensors)

Addresses confusion among tubes/wiring which can be simply frustrating but also can lead to serious injury in the case of tube confusion. Also reduces likelihood of tangling that occurs when maneuvering the child; can pull off sensors and lead to alerts.

Advanced, integrated physiological displays

Allows smarter algorithms to more accurately detect when critical intervention is needed, filters out artifact and reduces nuisance alarms.

Analog alarms

Supports workers in calibrating trust in the alarm system. Urgency of required intervention can be mapped to signal and easily interpreted by caregivers, aiding in decisions about when to take action.

Co-located alarm dismissal functions and (when available and appropriate) functionality to give solution procedure instructions

Addresses difficulty in immediately identifying the problem when an alert sounds, and having trouble locating the alarm “silence” button, thus exposing the infant to heightened noise levels and stressful stimuli longer than necessary. Also, giving fast feedback on recommended steps to resolve the problem when they are requested.

Advanced displays and alerts, such as remote monitors and tactile alerts

To address the need to keep lighting and sound levels low in the NICU, designs proposed ways to display information remotely so the stimuli would not disturb the infant, yet still could be reliably linked to the location of the infant. Tactile alerts, such as a vibrating belt or wristband, can capture and direct attention without adding significant auditory stimuli to the environment.

Mobile/web access to realtime infant video and data

A secure connection can allow family members to see their baby even when away from the hospital. Caregivers can use mobile or web applications to get information and alerts about the infant when away from the NICU.

CONCLUSION

This project offered an exciting topic for the students to learn and practice several important human factors methods. The pairing of expertise between Cognitive Systems Engineering and Architectural Design students allowed the human-centered design approach to consider how the technological and architectural aspects of the incubation system contribute to addressing the various needs of the three types of users. The approach outlined in this paper can be repeated by students and/or professionals to inform design of incubation systems and NICU environments that consider a wide perspective of issues, including psychosocial, ergonomic, usability, safety, and sustainability.

ACKNOWLEDGMENT

The authors would like to thank the students of the two

classes for their outstanding efforts in this project. Also thanks to: Dr. Robert White and Dr. Debra Harris for hosting group interviews and reviewing designs, Tamara Congdon and the NICU nursing staff at St. Joseph’s Regional Health Center for facilitating task analyses, and Wendy Slatery and GE Health Systems for providing the Giraffe technology. Finally, the authors would like to thank the Industrial & Systems Engineering and Architecture departments for in-kind support.

REFERENCES Antonucci, R., Porcella, A. & Fanos, V. (2009). The infant incubator in the

neonatal intensive care unit: unresolved issues and future developments. J Perinat Med, 37(6), 587-598.

Carayon, P. & Gurses, A.P. (2008). Nursing Workload and Patient Safety – A Human Factors Engineering Perspective. In: Hughes, R. (Ed.), Safety and Quality: An Evidence-Based Handbook for Nurses. Agency for Healthcare Research and Quality. Rockville, MD.

Center for Health Systems & Design (CHSD) (2012). Incubator Design. Technical report, CHSD, Texas A&M University. January.

Cermakova, E. (2003). Study of extremely low frequency electromagnetic fields in infant incubators. International Journal of Occupational Medicine & Environmental Health, 16(6), 215-220.

DeLucia, P.R., Ott, T.E., & Palmieri, P.A. (2009). Performance in Nursing. Reviews of Human Factors and Ergonomics, 5(1), 1-40.

Imhoff, M. & Kuhls, S. (2006). Alarm algorithms in critical care monitoring. Anesthesia & Analgesia, 102, 1525-1537.

Kirwan, B. & Ainsworth, L.K. (1992). A guide to task analysis. CRC Press. Nielsen, J. Usability heuristics and evaluation.

http://www.useit.com/papers/heuristic/heuristic_evaluation.html Shepley, M.M., Harris, D.D., & White, R. (2008). Open-bay and single-family

room Neonatal Intensive Care Units: Caregiver satisfaction and stress. Environment and Behavior, 40(2), 249-268.

Wickens, C.D., Lee, J.D., Liu, Y., & Gordon-Becker, S.E. (2004). An Introduction to Human Factors Engineering (2nd Ed). Prentice Hall.

Woods, D.D. & Roth, E.M. (1988). Cognitive systems engineering. In Helander, M. (ed.) Handbook of Human-Computer Interaction. North-Holland: Elsevier. 3-43.

Woods, D.D. (1995). The alarm problem and directed attention in dynamic fault management. Ergonomics. 38(11), 2371-2393.

Zhang, J., Johnson, T.R., Patel, V.L., Paige, D.L., Kubose, T. (2003). Using usability heuristics to evaluate patient safety of medical devices. J Biomedical Informatics, 36, 23-30.

191