Original Contributions Applications of Ecological ...€¦ · Original Contributions 74 Journal of Healthcare Information Management — Vol.18, No.4 Introduction Since the publication

Original Contributions

74 Journal of Healthcare Information Management — Vol. 18, No. 4

IntroductionSince the publication of the Institute of Medicine report

“To Err is Human: Building a Safer Health System”1 and thesubsequent report, “Crossing the Quality Chasm,”2 there hasbeen a growing understanding of the role that humanfactors can play in preventing errors and improvinghealthcare delivery. The latest Institute of Medicine report,“Keeping Patients Safe: Transforming the Work Environmentof Nurses,”3 recommends an increase in the developmentand use of computer-supported clinical decision-supportsystems as a way to improve patient safety.

A recent large-scale project by O’Neill et al4 called “The Nurse Computer Decision Support Project” (N-CODES)describes a prototype for a prospective decision-supportsystem that integrates the best available evidence into thedecision-making process. The project staff consists of foursenior researchers, two computer engineers, two nurses,and seven research assistants (three engineering graduate

students and four nursing graduate students). The projecthas not yet produced a working prototype, although thegroundwork has been completed and presumably the task of integrating and presenting the information now will be addressed.

One approach that can be used to design interfaces forcomplex systems is the ecological interface design5,6

approach. This paper will describe ecological interfacedesign (EID); the underlying cognitive work analysis (CWA)7

that is done before the design phase; an overview of workthat has been done in this area in the last two decades,with an emphasis on work in the healthcare environment;and a discussion of four possible applications of EID to thenursing process: patient monitoring, database design,decision support tools, and the measurement and analysisof nurse-sensitive outcomes. The challenges facing theapplication of EID to the nursing process also are reviewed.

Applications of EcologicalInterface Design in Supporting

the Nursing ProcessKathryn Momtahan, RN, PhD; and Catherine Burns

A B S T R A C T

Today’s nursing environment is complex, with many sources of data that are often poorly displayed.

Ecological interface design (EID) is a systematic approach to designing interfaces to complex systems.

EID has been used to design interfaces for aviation displays, power plant monitoring and control,

human hemodynamic monitoring, anesthesia monitoring, and neonatal intensive care monitoring

and diagnosis. EID makes critical relationships easily visible, eliminating the mental workload of

integrating, calculating, or remembering multiple values.This paper reviews past experimental

studies of EID in healthcare applications and discusses the application of EID to a decision-support

tool for diabetic patients using personal digital assistants.The authors also discuss other

contributions that EID could make to the nursing process in the areas of physiological monitoring,

decision support, database design, and the measurement and analysis of nurse-sensitive outcomes,

including patient safety outcomes.

K E Y W O R D S

■ Ecological interface design ■ Cognitive work analysis

■ Nursing process ■ Personal digital assistants


Journal of Healthcare Information Management — Vol. 18, No. 4 75

Overview of Ecological Interface DesignEcological interface design is a systematic approach to

interface design for complex systems. EID uses a methodcalled work domain analysis to determine the informationrequirements for ecological displays. Work domain analysisis an important component of cognitive work analysis.

Work domain analysis results in a five-level hierarchicalmodel of the system of interest or work domain. This five-level hierarchy is called an abstraction hierarchy, originallyconceptualized by Rasmussen.8,9 This model ranges from theconcrete, with a description of equipment and its capabil-ities, to the abstract, with a description of purposes to beachieved with that equipment. In this way, work domainanalysis specifies information requirements in a way that isrelevant to goals.

Figure 1 outlines the levels of the abstraction hierarchy.The functional purpose layer describes the purpose forwhich the system is to be built. In computer-based systemsfor which this theoretical framework has been developed,this layer most commonly will map to the application layerof the system. The abstract function level maps out thecausal structure of the system. The generalized functionlevel provides a view of the processes by which theabstract functions are carried out. The physical functionlayer describes the components and physical implementa-tions of the processes in the generalized function layer. Inthe physical form layer, the components are described bytheir appearance, location, and physical form.

This work domain analysis is used to guide the computerinterface design. Originally designed to support the design ofprocess control displays,10 it more recently has been used to develop aviation displays,11 displays for a document-retrieval system for public libraries,12,13 network management,14

hemodynamic monitoring,15 neonatal intensive caremonitoring,16 and anesthesia monitoring.17,18,19 Most of theanalyses in the medical domain have modified the hierarchysomewhat, dispensing with generic-level labeling, such asabstract function and generalized function, in favor of levelsmore tightly tailored to medical applications. Table 1 showsthese modifications by comparing the layers of theabstraction hierarchy developed for designing displays forpower plant monitoring, aviation, and anesthesia.

Until now, the investigation of the application of EID to healthcare has been limited to several studies in anesthesiology monitoring, hemodynamic monitoring, andneonatal monitoring. These environments relate well to thepower plant and aviation domains because they requireextensive monitoring and highly trained personnel. Table 2summarizes these studies and their findings.



Patient MonitoringWhile some researchers using the EID framework have

focused on hemodynamic monitoring of critically ill patientsthat may be performed by nurses or physicians,15,16 othershave focused on the role of the anesthetist and theanesthesia workstation.17,18,19,20

Although these patient monitoring activities seem fairlysimilar, in fact, there are several differences. In theoperating room, the anesthetist is the sole monitor of onepatient. In the intensive care setting, although the nurse isthe main monitor, others may be involved in the care of apatient, such as the intensivist. The nurse also may bemonitoring more than one patient; for example, in arecovery room setting, a nurse may take care of severalpatients at once. In addition, the anesthetist is activelyinducing a state of disequilibrium (the anesthesia) so theoperation can proceed. In an intensive care situation, theintent is for the patient to achieve and maintain a steadystate. The EID displays in the studies presented in Table 2were compared against existing displays in use in thoseareas at the time of the studies.

Diabetic MonitoringEID was used to design a tool for diabetic monitoring for

patients using personal digital assistants (PDAs). While notyet tested for its performance, it provides a good demon-stration that EID concepts may be useful at the patient leveland applicable to small screen displays.

Novel visualizations for diabetic health monitoring5 weredeveloped, concentrating on visually portraying the keyrelationships for overall health by showing energy balancesbetween exercise, basal metabolic rate and carbohydrateintake. The monitor shows insulin action curves and therelationship between insulin injections and blood insulinlevels. Both of these relationships are difficult for patients to understand for various reasons; for example, the energybalance requires multiple calculations, and the blood insulin level involves the prediction of a time lag. Currentpatient tools, such as log books or digital applications, do not provide this type of calculated and predictive information even though both of these relationships arealso important if the patient is to maintain tight control oftheir metabolic system.

Study Tested with Improved performance with EID display?

Experts or novices performed better?

Accuracy Speed Hemodynamic Monitoring Effkin et al. (1997) Experiment 1

19 undergraduate psychology students Between subject design

Yes No Not tested

Effkin et al. (1997) Experiment 2

13 undergraduate psychology students Within subject design

Yes Yes Not tested

Effkin et al. (1997) Experiment 3

6 experienced critical care nurses 6 nursing students

Yes Yes

Yes Yes

Some evidence novices better supported

Sharp and Helmicki (1998) 16 anesthetists (attending fellows and residents)

Yes Not measured Least experienced physicians performed better with the EID display

Anesthesia Monitoring Jungk (2000) 20 anesthetists Yes No Not tested Blike (1999) 3 anesthesia residents,

8 anesthetists Yes Yes No difference

Blike (2001) 5 anesthesia residents, 2 anesthetists

Mixed results Mixed results Not tested

Jungk (2001) Experiment 1

16 anesthetists No Mixed results Not tested

Jungk (2001) Experiment 2

8 anesthetists Yes Yes Not tested

Zhang (2002) Study 1

12 anesthetists Mixed results Mixed results Not tested

Zhang (2002) Study 2

12 undergraduate bioengineering students

Yes No Not tested

Table 2. Summary of the results of healthcare-related studies where the EID display was compared with conventional displays.



These variables appear when a work domain analysis ofblood glucose metabolism is conducted, and, notably, theyappear at the higher levels where processes, balances, andpurposes are described. Table 3 shows some of thevariables that are identified. Many of these variables—while they can be calculated, estimated, or predicted with a decision aid—are never available at the patient level (see Table 4). Figure 2 shows the interface developed for this system.

The menu includes entries for energy balance and bodymass index, which are unique to the EID approach. Overallblood glucose state is quickly shown through too high(left), normal (middle), and too low (right) with color andshape coding.

Applications of EID to NursingThe nursing process involves the evaluation of a

patient’s condition, the identification of problems, and the

implementation of interventions by nurses. Problems areidentified based on the patient interview and physicalassessment, prior knowledge, the patient’s medical information (such as the admitting diagnosis), and tests and procedures performed to determine a course of action.This process is recursive.

The frequency with which re-assessments are performedand interventions evaluated and modified will depend on apatient’s acuity level. In an intensive care unit, the processmay occur every 15 minutes; in a stable patient ready to gohome from the hospital, the process may occur perhapsevery 12 hours and with less intensity.

The nursing process now also incorporates best evidenceand practice. The nurse must collect and analyze data aboutthe individual patient and then cognitively integrate thatinformation with what they understand to be the bestevidence to determine an intervention.

As with all healthcare-related processes, the nursing

Table 3. Some variables identified from the work domain model.



process has become increasingly complex, with a proliferation of sophisticated equipment that either canoperate on a stand-alone basis or integrate with clinicalinformation systems.

As Burns and Hajdukiewicz5 point out, there are severalchallenges to applying EID to a healthcare domain. Thesechallenges include:

• Choosing the system boundary.

• Deciding when to stop decomposing of the part-whole analysis.

• Integrating medical information.

• Dealing with issues associated with sensor limitations and availability.

• Determining information requirements for different roles in the medical environment.

• Extending EID to other sensory modes (such as auditory).

There are several opportunities toapply ecological interface design tothe nursing process. Four possibleapplications will be discussed—physiological monitoring, decisionsupport, database design, andmeasuring and analyzing nursing-sensitive outcomes.

To focus on a design, however, acognitive work analysis first needs tobe done in the area of interest. The

focus here will mainly be on hospital-based nursing, butthese applications of EID to physiological monitoring,decision support, and the measuring and analyzing ofnursing-sensitive outcomes could all be used in primarycare as well, particularly if the applications were to bedeveloped for use with PDAs, which are popularly used asportable computerized devices.

Analysis of the Work that Nurses DoUntil recently, the major focus of the efforts to measure

the work that nurses do has focused on nursing workloadmeasurement, so nursing productivity levels—sometimesreferred to as utilization levels—could be calculated basedon time-motion studies to facilitate the adjustment ofstaffing levels. For a comparison of four workload measurement systems developed for this purpose,

Table 4. Variables identified from the work domain model, with availability.



Malloch21,22 and Conovaloff23 have reviewed the evolution of these systems.

It is now the norm for hospitals to have patient classification systems and nursing workload measures.Originally designed as an economic tool to optimize staff-mix ratios, there has been a shift toward optimizingstaff-mix and nurse-patient ratios to improve patientoutcomes,24,25,26 including patient safety outcomes.27

Nursing work also has been measured in the past todetermine the percentage of time that nurses spend doingpaperwork versus providing patient care. For example, suchmeasurements are typical before and after implementationof a clinical information system so healthcare executivescan measure the potential decrease in the amount of timenurses do paperwork after a CIS is implemented. Typically,either work-sampling methods28 or a combination of time-motion and work-sampling methods29 are used to measurethe impact on nursing work before, after, and during theimplementation of a CIS.

More recently, there has been some work done toaccount for other factors involved in the work that nursesdo. Potter et al30 combined a time-motion analysis with adescriptive analysis of nursing work. By combining this datawith a link analysis, the authors developed what they call a“cognitive pathway.” Other researchers have incorporated ameasure of environmental complexity into their nursingwork measurement and analysis.31,32 On a more global scale,there are recent efforts to identify nurse-sensitiveoutcomes,33 integrate reference terminology in nursing,34,and move this forward by mapping the nursing process tofacilitate nursing information capture and utilization withinHL7 standards.35

Physiological MonitoringThe most obvious application of EID to physiological

monitoring based on cognitive work analysis is hemody-namic monitoring of critically ill patients in operating roomsand intensive care units, which has already been reviewed.The developing application of EID to diabetic monitoringalso was discussed earlier. However, although this has beenintended for use by diabetics themselves, this informationcould be shared with the diabetic’s physician or theadvanced practice nurse in the diabetic clinic.

Diabetics are not the only group of patients whosecondition requires continuous monitoring and adjustment ofactions and treatment. Heart failure patients are anotherlarge group of patients whose physiological state needs tobe closely watched. The monitoring of this informationalone is important, but being able to share the informationwith physicians and heart failure clinic nurses would bevery valuable.

In a six-month prospective interventional crossover studyof 19 patients, Tsang et al36 were able to demonstrate asignificant improvement in glycemic control when patients

used a monitoring system with a handheld, touchscreenelectronic diary (with a mean HbA1C reduction of 0.825percent). Electronic diaries, which are becoming popular forthe management of diabetes mellitus,37 cannot bythemselves be considered a computerized decision supporttool if they are used solely for logging patient information.

Decision SupportAlthough far from common, computerized decision

support for clinical practice has been used long enough sothat systematic reviews of their implementation and use are available.

In their systematic review of 63 controlled trials ofcomputer-based clinical decision support systems on physician performance and patient outcomes, Hunt et al38

concluded that “published studies of CDSSs are increasingrapidly and their quality is improving. The CDSSs canenhance clinical performance for drug dosing, preventivecare, and other aspects of medical care, but not convinc-ingly for diagnosis. The effects of CDSSs on patientoutcomes have been insufficiently studied.”

In a more recent assessment of the situation followingtheir review of the literature, Trowbridge and Weingarten39

concluded that, “The preponderance of evidence suggeststhat clinical decision support systems are at least somewhateffective. Their highest utility has been demonstrated in theprevention of medical errors, especially when coupled witha computerized medical record and directly intercalated intothe care process.”

Computerized decision support for clinical nursingpractice is less common,40 although there are somepromising results related to the implementation and use ofdecision support in nursing. Mueller et al41 reported thepositive impact of a decision support tool—a diuretictreatment algorithm, designed and implemented formanaging heart failure patients—used by advanced practicenurses. Researchers were able to demonstrate a significantreduction in the 30-day readmission rate and a 50 percentreduction in heart failure-related hospitalizations in 271patients during a four-year period.

There is a growing interest in the integration of information from diverse sources to develop decision supporttools that support the nursing process4 Brekke et al42 describe an interesting approach to integratingpopulation-based statistics to estimate the impact of coronaryheart disease treatments on disease outcomes for popula-tions. They created event-decision-intervention-outcome flowpathways to simulate risk of a clinical event and expectedintervention outcome for heart failure, tachyarrhythmia, stableangina pectoris, acute coronary syndrome, bradycardia, post-myocardial infarction, and post-coronary artery bypassgrafting. This type of modeling using population-basedstatistics for nursing interventions could enhance decisionsupport tools for nurses and policy makers.



Database DesignIn a recent paper, Burns and Zhou43 reported on a

project using work domain analysis as a foundation for thedesign of a health information database for blood glucosemanagement. They compared the content of four existingcommercial databases for blood glucose management withthe database they had designed. The study’s benchmarkwas the sum of the data elements generated from the fivedatabases, for a total of 300 possible unique data elements.

The database generated through work domain analysiscovered 81.7 percent of the elements. The next closestdatabase contained 43 percent of the elements. Importantly,some of the elements that the work domain analysis-generated database did not cover were related to eventssuch as fasting, because the work domain analysis wouldrecord energy balances and calculate whether or not thepatient was losing mass.

Event-based knowledge will need to be accounted for infuture designs of health-related databases. Similarly, thecontributions of other sources of data must be factored in.The tracking and treating of patients usually includes inputfrom the patients themselves but also from a variety ofhealthcare professionals. In the case of hospital databases,the main source of information about the patient will bereceived from the healthcare professionals.

The ubiquity of data sources brings up another importantpoint in using cognitive work analysis and the results of thisanalysis to support the design of interfaces to integrate anddisplay information: how to account for interdisciplinarywork. A good example of shared documentation in ahospital is the patient problem list; patients’ problems canbe added or removed by any healthcare professionaltreating the patient.

Another example is clinical pathways. Dykes et al44

recently studied the adequacy of national standardizedterminologies in the United States for interdisciplinarycoded concepts in an automated clinical pathway for heartfailure. They found that of the 260 unique pathwayconcepts identified, 239 of them—or 91.9 percent—were represented by one or more of the standardizedcoding systems.

Nurse-Sensitive OutcomesThere is a great deal of interest in the measurement and

analysis of nursing-sensitive outcomes.33 Although this is an

important activity for researchers who want to link the workthat nurses do with patient outcomes, it is essential to findways to not only capture and categorize this informationbut to integrate and present the information back to nursesto enhance their practice. The standardization of termi-nology34,35 and the capturing of information are necessaryparts of this process.

The current interest in improving patient safety whileunder the care of medical professionals has resulted in adiscussion of nurse-sensitive patient safety indicators, asubset of nurse-sensitive patient outcomes. In their reviewof the literature, White and McGillis Hall45 found medicationerrors, nosocomial infections, patient falls, pressure ulcers,and mortality to be the most common safety outcomeslinked to nursing practice.

ConclusionOther than the cognitive work analysis that has been

done related to hemodynamic monitoring,15 there has yet tobe a cognitive work analysis performed to model the workthat nurses do to design computerized systems that wouldfacilitate the nursing process.

The task analysis and redesign of an intravenous infusionpump interface performed by Lin et al46 is a good exampleof how an interface design can affect cognitive workloadand patient safety, even though it is a simple example andis not based on a cognitive work analysis and EID becauseit deals with a single piece of equipment and not acomplex system. Lin et al were able to demonstrate anincrease in the speed and accuracy of the programming ofan intravenous infusion pump by experienced nurses whenthe interface was redesigned.

Before the introduction and universal use of intravenousinfusion pumps, nurses counted the number of drops perminute when they set up and monitored intravenousinfusions for patients. Now, not only are almost all intra-venous fluids infused through programmed pumps, buthospitals typically have a variety of pumps that are in usefrom several different manufacturers. It can be assumed thatthe human factors design of at least some of these pumps isnot optimal. What is needed to design better computerizedsystems to support nurses is cognitive work analysis andthe design of interfaces that integrate all the informationthat nurses use and need in a way that decreases theircognitive workload. The expectation is that the speed andaccuracy with which nurses perform their work would beenhanced and that this, in turn, would lead to a saferenvironment and improved outcomes for patients.

There are many challenges involved in using ecologicalinterface design to support the nursing process. In terms ofdesign, there has been a sparse amount of work domainanalyses, except for research that has been done for physio-logical monitoring of patients. Most of the work to date hasbeen tested on few subjects, and there is at least some

“Ecological interface design is a systematic

approach to interface design for

complex systems.”



evidence that individual differences exist in the way peopleprocess information from displays.47,48 Elements of thedesign that affect the performance of novices versus expertusers is still largely unknown and may differ depending onthe domain and the user group. The role of the visualdesigner has not previously been taken into account.Because display design is partly an analytic and partly anartistic endeavor, it is conceivable the same type of displaydesigned by two different visual designers might yielddifferent results.

On the nursing side, the emerging classification of dataelements means that a system designed now may notconform to the eventual form of the final standard. Also,because visual display interpretation has not been as exten-

sively tested with nurses, little is known about the best wayto present information to nurses.

About the AuthorsKathryn Momtahan, RN, PhD, is the research scientist for

nursing at the University of Ottawa (Ontario, Canada) HeartInstitute. A cardiovascular nurse, she has a PhD in humanfactors psychology and worked for several years as ahuman factors specialist in industry. She can be reached [email protected].

Catherine Burns directs the Advanced Interface DesignLab at the University of Waterloo in Ontario, Canada. Herresearch looks at human performance support through EIDand advanced visualization.

References1. Kohn LT, Corrigan, J, Donaldson MS (Eds). To Err is Human:

Building a Safer Health System. Washington: National AcademyPress, 2000.

2. Institute of Medicine. Crossing the Quality Chasm: A New HealthSystem for the 21st Century. Washington: National Academy Press, 2001

3. Institute of Medicine. Keeping Patients Safe: Transforming the WorkEnvironment of Nurses Prepublication Copy. Washington: NationalAcademy Press, 2004.

4. O’Neill ES, Dluhy NM, Fortier PJ, Michel HE. 2004. Knowledge acquisition, synthesis, and validation: a model for decision supportsystems. J Adv Nurs, 47(2), 134-142.

5. Burns C, Hajdukiewicz JR. Ecological Interface Design. Boca Raton,FL: CRC Press, 2004.

6. Vicente, KJ. 2002. Ecological interface design: Progress and challenges. ]Human Factors, 44(1), 62-78.

7. Vicente KJ. Work domain analysis: Toward safe, productive, and healthy computer-based work. Mahwah, NJ: LawrenceErlBaum, 1999.

8. Rasmussen J. 1985. The role of hierarchical knowledge representationin decision making and system management. IEEE Transactions onSystems, Man and Cybernetics, 15, 234-243.

9. Rasmussen J. Information Processing and Human-MachineInteraction: An Approach to Cognitive Engineering. Series Volume12. New York, NY: North-Holland, 1986.

10. Vicente KJ. 1992. Memory recall in a process control system: A measure of expertise and display effectiveness. Memory &Cognition, 20, 356-373.

11. Dinadis N, Vicente KJ. 1999. Designing function visualizations foraircraft systems status displays. Int J Aviation Psychology, 9(3), 241-269.

12. Pejtersen AM. 1992. The BookHouse: An icon based database system for information retrieval from libraries. In B. Cronin (Ed),The Marketing of Library and Information Services 2, pp. 472-591.London: ASLIB.

13. Pejtersen AM, Goodstein LP. Beyond the desktop metaphor: information retrieval with an icon based interface. In MJ Tauber &P Gorny (Eds), Proceedings of the 7th Interdisciplinary Workshopon Informatics and Psychology on Visualization in HumanComputer Interaction. Berlin: Springer-Verlag, 1990.

14. Burns CM, Kuo J, Ng S. Ecological interface design: a new approachfor visualizing network management. Computer Networks: theInternational Journal of Distributed Informatique, 43(3), 369-388.

15. Effken JA, Kim N-G, Shaw RE. 1997. Making the constraints visible:testing the ecological approach to interface design. Ergonomics,40(1), 1-27.

16. Sharp TD, Helmicki. 1998. The application of the ecological inter-face design approach to neonatal intensive care medicine.Proceedings of the Human Factors and Ergonomics Society 42ndAnnual Meeting, 350-354.

17. Blike GT, Surgenor SD, Whalen K. 1999. A graphical object displayimproves anesthesiologists’ performance on a simulated diagnostictask. J of Clinical Monitoring and Computing, 15(1), 37-44.

18. Blike GT, Surgenor SD, Whalen K, Jensen J. 2000. Specific elementsof a new hemodynamics display improves the performance ofanesthesiologists. J of Clinical Monitoring and Computing, 16(7),485-491.

19. Jungk A, Thull B, Hoeft A, Rau G. 2000. Evaluation of two new eco-logical interface approaches for the anesthesia workplace. J ofClinical Monitoring and Computing. 2000;16(4), 243:258.

20. Zhang Y, Drews FA, Westenskow DR, Foresti S, Agutter J, BermudezJC, Blike G, Loeb R. 2002. Effects of integrated graphical displayson situation awareness in anaesthesiology. Cognition, Technology,& Work. 4, 82-90.

21. Cockerill R, Pallas L, Bolley H, Pink G. 1993. Measuring nursingworkload for case costing. Nurs Econ, 11(6), 342-9.

22. O’Brien-Pallas L, Cockerill R, Leatt P. 1992. Different systems, different costs? An examination of the comparability of workloadmeasurement systems. J Nurs Adm, 22(12), 17-22.

23. Malloch K, Conovaloff A. 1999. Patient Classification Systems, Part I:The Third Generation. JONA, 29(7/8), 49-56.

24. Needleman J, Buerhaus P, Mattke S, Stewart M, Zelevinsky K. 2002.Nurse-staffing levels and the quality of care in hospitals. N Engl JMed, 346(22), 1715-1722.

25. McGillis Hall L, Doran D. 2004. Nursing staffing, care deliverymodel, and patient care quality. J Nurs Care Qual, 19(1), 27-33.

26. Sochalski J. Is more better? The relationship between nurse staffingand the quality of nursing care in hospitals. Medical Care, 42(2 suppl), II-67-II-73.



27. McGillis Hall L, Doran D, Pink GH. Nurse staffing models, nursinghours, and patient safety outcomes. JONA, 34(1), 41-45.

28. Pabst MK, Scherubel J, Minnick A. 1996. The Impact ofComputerized Documentation on Nurses’ Use of Time. Computers in Nursing. 14(1), 25-30.

29. Bosman RJ, Rood E, Oudemans-van Straaten HM, Van der Spoel JI,Wester JP, Zandstra DF. 2003. Intensive care information systemreduces documentation time of the nurses after cardiothoracic surgery. Intensive Care Med, 29(1):83-90.

30. Potter R, Boxerman S, Wolf L, Marshall J. 2004. Mapping the nursingprocess: A new approach for understanding the work of nursing.JONA, 34(2), 101-109.

31. O’Brien-Pallas L, Thomson D, McGillis Hall, L, Pink G, Kerr, M,Wang S, Li X, Meyer R (2003). Evidence Based Standards forMeasuring Nurse Staffing and Performance. Report for theCanadian Health Services Research Foundation, December, 2003.Project #RCI-721-06.

32. Hedberg B, Satterlund Larsson U. 2004. Environmental elementsaffecting the decision-making process in nursing practice. J ClinNurs, 13, 316-324.

33. Doran D (Ed). Nursing-Sensitive Outcomes: State of the Science.Sudbury, MA: Jones & Bartlett, 2003.

34. International Standards Organization (ISO). 2003. Health informatics—integration of a reference terminology model for nursing. ISO Standard 18104. International Standards Organization. Geneva, Switzerland.

35. Goossen WT, Ozbolt JG, Coenen A, Park HA, Mead C, Ehnfors M,Marin HF. 2004. Development of a provisional domain model forthe nursing process for use within the Health Level 7 referenceinformation model. J Am Med Inform Assoc, 11(3), 186-194.

36. Tsang MW, Mok M, Kam G, Jung M, Tang A, Chan U, Chu CM, Li I,Chan J. 2001. Improvement in diabetes control with a monitoringsystem based on a hand-held, touch-screen electronic diary. J Telemed Telecare, 7(1), 47-50.

37. Kerkenbush JL, Lasome CE. 2003. The emerging role of electronicdiaries in the management of diabetes mellitus. AACN Clin Issues,14(3), 371-378.

38. Hunt, DL, Haynes RB, Hanna SE, Smith K. 1998. Effects of computer-based clinical decision support systems on physicianperformance and patient outcomes: a systematic review, JAMA,280(15), 1339-1346.

39. Trowbridge R, Weingarten S. 2001. Clinical decision support systems.In Making health care safer: A critical analysis of patient safety practices (AHRQ Publication 01-E058). Rockville, MD:AHRQ. 2001, 589-594.(http://www.ahrq.gov/clinic/ptsafety/pdf/ptsafety.pdf, accessedApril 25, 2004).

40. Tooey MJ, Mayo A. 2003. Handheld technologies in a clinical setting:State of the technology and resources. AACN Clinical Issues, 14(3),342-349.

41. Mueller TM, Vuckovic KM, Knox DA, Williams RE. 2002.Telemanagement of heart failure: A diuretic treatment algorithmfor advanced practice nurses. Heart & Lung, September/October,340-347.

42. Brekke MJ, Kottke TE, Brekke LN, Wu LA. 2003. Design and devel-opment of a coronary heart disease decision support tool. Intl J Technology Assessment in Healthcare, 19(3), 555-558.

43. Burns CM, Zhou Y. Work domain analysis as a step in a systematicdatabase design methodology.

44. Dykes PC, Currie LM, Cimino JJ. 2003. Adequacy of evolving nation-al standardized terminologies for interdisciplinary coded conceptsin an automated clinical pathway. J Biomed Inf, 36, 313-325.

45. White P, McGillis Hall L. 2003. Patient safety outcomes. In Doran D(Ed), Nursing-Sensitive Outcomes: State of the Science. Sudbury, MA: Jones & Bartlett.

46. Lin I, Vicente KJ, Doyle DJ. 2001. Patient safety, potential adversedrug events, and medical device design: A human factors engineering approach. J Biomed Inform, 34, 274-284.

47. Torenvliet GL, Jamieson A, Vicente KJ. 2000. Making the most ofecological interface design: the role of individual differences.Applied Ergonomics, 31, 395-408.

48. Freer Y, Ferguson L, Ewing G, Hunter J, Logie R, Rudkin S,McIntosh N. 2002. Mismatched concepts in a neonatal intensivecare unit (NICU): further issues for computer decision support? J Clin Monit Comput, 17(7-8), 441-447.

Documents

Original Contributions Applications of Ecological ...€¦ · Original Contributions 74 Journal of Healthcare Information Management — Vol.18, No.4 Introduction Since the publication