Sie sind auf Seite 1von 7

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/282517661

Usability of Incident Reporting Systems: Preliminary Results of a Case Study

Article · June 2015


DOI: 10.1177/2327857915041031

CITATIONS READS
2 222

2 authors, including:

Esa Rantanen
Rochester Institute of Technology
110 PUBLICATIONS   664 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Course on I/O Psychology View project

All content following this page was uploaded by Esa Rantanen on 07 March 2016.

The user has requested enhancement of the downloaded file.


2015 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 168

USABILITY OF INCIDENT REPORTING SYSTEMS:


PRELIMINARY RESULTS OF A CASE STUDY
Paul Wiele and Esa Rantanen
Department of Psychology, Rochester Institute of Technology, Rochester, NY

Although low participation in incident-reporting systems contributes to the problem of poor error preven-
tion, little research has focused on improving participation. The present study is a case study of a hospital’s
incident reporting practices, seeking to increase both the number and quality of their incident reports by ex-
amining the usability of the reporting system itself as well as organizational factors impacting attitudes to-
wards reporting. A snowball sample of hospital staff was recruited to identify organizational and software
usability factors that both impede and promote frequent and useful incident reporting. Preliminary results
suggest that time pressure is a key impediment to reporting, and that hospital staff in certain units are al-
ready attempting innovative shortcuts to cut time demands without reducing their reporting rates.

INTRODUCTION Is Incident Reporting Working?

Thanks in large part to the “Quality Chasm Series” of re- It is very difficult to judge the extent to which an incident
ports (Institute of Medicine, 1999, 2001, 2004), recent efforts reporting system is working because the true number of inci-
to improve patient safety in the American healthcare system dents is always unknown (Bleich, 2005). Reporting rates vary
have focused on adverse events, or cases of preventable acci- in nonrandom ways that suggest that they are not representa-
dental harm to patients. Adverse events, along with near miss- tive of the true incident rates. Even in facilities with proce-
es, which could plausibly have caused harm but did not, to- dures in place for making reports, they likely capture only a
gether make up the general category of “incidents”. Figure 1 small fraction of the incidents. For example, Jha et al. (1998)
illustrates this taxonomy and reporting protocols in New York assessed patient charts and incident reports for evidence of ad-
State. verse drug events (ADEs) and found that only 9 out of 166 in-
Near misses generally stem from the same underlying cidents classified as “preventable ADEs” had been voluntarily
problems and processes as actual incidents (Institute of Medi- reported.
cine, 2004). They are also much more numerous than incidents Murff, Patel, Hripcsak, and Bates (2003) built on this and
Copyright 2015 Human Factors and Ergonomics Society. All rights reserved. DOI 10.1177/2327857915041031

that result in harm, allowing for more reliable detection of pat- other studies to compare voluntary incident reporting with
terns and trends through statistical means. Hence, in addition other methods for detecting incidents such as chart review and
to learning from adverse events, a system designed to mini- usability testing. They concluded that all methods have limita-
mize future harm to patients would also do well to learn from tions and no method alone can detect all incidents. Other
near misses. sources of variation in incident reporting include geographical
differences, including within states that require some form of
incident reporting (State of New York Department of Health,
2007), and differences by medical specialty (Busse & Wright,
2000; Johnson, 2002). A possible explanation for this is that
policy and safety culture differences on the level of individual
facilities or units contribute to the willingness of healthcare
workers to report incidents.
There has been some recent progress from the Partnership
for Patients initiative, which has been credited with accelerat-
ing the recent decline in hospital-acquired conditions (Agency
for Healthcare Research and Quality, 2013). Nevertheless, the
healthcare system in general still lacks formal ways of sharing
safety information that would allow safety improvements to
flow freely through the system.
The highly successful Aviation Safety Reporting System
(ASRS) offers an example of how such a system could be im-
plemented, and a starting point for the criteria that make a re-
porting system effective. Most importantly, the reporting sys-
tem should allow confidentiality to protect the reporter from
professional or legal risks as well as feedback that encourages
Figure 1. A taxonomy of incidents in health care and reporting future reporting (Billings, 1998). ASRS also offers us the in-
protocols in New York State. triguing and unusual insight that minimizing harm may not re-
quire a complete or even representative record of incidents,

Downloaded from hcs.sagepub.com at HFES-Human Factors and Ergonomics Society on January 6, 2016
2015 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 169

but instead the ability for “sharp end” workers to identify and of the analysis results. In this paper we report the results from
report those issues they reasonably believe to be the most im- the first stage of the study.
portant (Beaubien & Baker, 2002).
METHODS
Approaches to Safety
Participants
To accomplish the goal of identifying the most important
types of incidents, our research focused specifically on the ap- Participants were recruited via “snowball” sampling
proach of increasing participation in the incident reporting among staff in any department or role at the hospital. Initial e-
process. This requires an understanding of those factors that mails were sent by the hospital’s Patient Safety department to
inhibit reporting and those that actively encourage it—in other likely participants. Those who volunteered were then asked to
words, as Hollnagel, Wears, and Braithwaite (2015) would put recruit coworkers who they thought would be interested in
it, a “Safety-II” attitude of understanding why a system suc- participating.
ceeds in addition to why it fails. These inhibitions and encour-
agements can come from the reporting system itself or from Materials
the practices of the organization that implements it. The re-
search reported here examines both in the case study of Roch- Participating in the first stage of the study entailed a usa-
ester General Hospital (RGH), which uses a web-based appli- bility test and a structured interview. So far two usability test
cation to report incidents. and interview sessions have been conducted, with a goal of a
Employees’ willingness to report incidents within any or- total of 10. Additional preliminary usability data were gath-
ganization is partially a product of that organization’s safety ered by the authors in a pilot test.
culture, that is, its attitudes and practices regarding improving In the usability testing phase, participants were recorded
safety. An organization with a weak safety culture will blame entering an incident report about a hypothetical scenario based
individual employees for incidents, and consider reporting on a published case report (Gupta & Cook, 2013) using the
those incidents “out of line”, while one with a strong safety Quantros training module. This module duplicates the form
culture will credit employees for providing useful information used to report incidents, without actually submitting the report
that can be learned from (Taylor, 2012). Safety culture indica- when complete. The recordings are reviewed so that any prob-
tors show that better teamwork (Erler, et al., 2013) and strong lems the participants encounter may be noted and classified
safety culture on the level of the whole organization (Kagan & according to usability guidelines for web forms (Bargas-Avila
Barnow, 2013) predict a higher incident reporting rate and a et al., 2010) as well as the more common and general- purpose
lower actual rate of incidents occurring. heuristic analysis (Nielsen, 1994).

Current Practices Interviews

Rochester General Hospital (RGH) uses multiple special- The interview phase began with general questions as-
ized methods to collect information about incidents. Billing sessing the participant’s views on the organizational culture.
data may be used to measure factors like time of stay, which The questions are based on those used in existing safety cul-
can be influenced by the quality of care provided. The nursing ture surveys (Commission on Accreditation of Medical
department submits data to the National Database of Nursing Transport Systems, 2012; Agency for Healthcare Research
Quality Indicators. Finally, hospital staff in general report in- and Quality, 2014b, 2014a), one study regarding the imple-
cidents using an incident reporting web application produced mentation of a health information exchange system (Feldman,
by Quantros, Inc. (this app is commonly referred to by users as Schooley, & Bhavsar, 2014), and one study on eliciting the
“Quantros” for short). Because Quantros alone among these user requirements of a proposed software project (Browne &
methods identifies both adverse events and near misses, is Rogich, 2001). The remainder of the interview phase consists
used by all departments, and contains details about not just of two “walkthroughs” using the critical incident technique
what kinds of incidents occur but also how they occur, this re- (Crandall, Klein, & Hoffman, 2006). In the first critical inci-
search will focus on the usability and organizational practices dent walkthrough, the participant is asked to guide the inter-
surrounding Quantros. viewer through an example of a time they decided not to make
an incident report, and in the second a time they decided to
Purpose of the Study make one, in both cases describing the circumstances and
thought processes that led to those outcomes.
The purpose of this research is to examine incident report-
ing in a large general hospital in a comprehensive and system- PRELIMINARY RESULTS
atic way. This project has three distinct stages: (1) Examina-
tion of the user experiences with the incident reporting inter- Pilot testing of the application’s usability using the au-
face through a focus group study, (2) investigation of the gen- thors as naive users detected potential usability problems in a
eral safety culture in the hospital through in-depth interviews variety of different categories. Categories are described in Ta-
and surveying of hospital staff members, and (3) examination ble 1 for the web form guidelines from Bargas-Avila et al.
of the analyses of the reported incident data and dissemination (2010) and in Table 2 for the general usability heuristics from

Downloaded from hcs.sagepub.com at HFES-Human Factors and Ergonomics Society on January 6, 2016
2015 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 170

Nielsen (1994). However, tests on experienced users have on- additional fields to appear contingent on those answers, but
ly corroborated the total length of the form and the large num- sometimes produces pop-ups with their own "save" and "can-
ber of options on menus as being actual inhibitors to use. cel" buttons (Table 1, "form content"; Table 2, "consistency
Form length is in the “form content” category under the web and standards"). Some menus lack a way to indicate "not ap-
form-specific guidelines and the “visibility of system status” plicable" when such a response would be expected, such as
category under the general heuristics. The large number of op- when answering the question "was another area involved?"
tions is in the “input types” category under the web form when describing where the incident took place (Table 1, "input
guidelines but is not accounted for by the general heuristics. types"). The software provides no feedback to the user to con-
Further tests will provide further evidence for or against the firm when a report in progress has been saved (Table 2, "visi-
importance of other categories of usability issues. bility of system status"). Perhaps most alarmingly, other than a
Some problems encountered in pilot usability testing but tutorial video, there is no help documentation, so troubleshoot-
not corroborated so far in tests on experienced users remain ing requires contacting a support center in ways that would
causes for concern. Responding to questions has inconsistent not guarantee anonymity (Table 2, "help and documentation").
results “downstream” in the form. The system often causes

Table 1
Reporting Application’s Compliance with Bargas-Avila and Colleagues’ (2010) Guidelines for Web Forms

Category Description Violations: Pilot Main


Form content Forms should be short and intuitively-ordered, allow content flexibility in answers, and clearly Yes Yes
distinguish between required and optional questions.
Form layout Fields for answers should be labeled above, listed one layout per row, and sized appropriately No No
for the length of the answer.
Input types Menus, checkboxes, buttons, or open response fields are appropriate for different numbers and Yes Yes
types of answers. Options should be in an intuitive order and limited in number if possible.
Error handling Answers should have their expected formats clearly indicated and should not be cleared by er- No No
rors. Error messages should be polite, informative, embedded in the form, and easily noticed.
Form submission Submission buttons should only be usable once and - not be confusable with “reset” buttons. Yes No
Submissions and how they will be used should be confirmed.

Table 2:
Reporting Application’s Compliance with Nielsen’s (1994) General Usability Heuristics.

Heuristic Description Violations: Pilot Main


Visibility of system status The system should provide appropriate and timely feedback Yes Yes
Match between system and the The system should present things in familiar terms similar to those encoun- Yes No
real world tered in other contexts
User control and freedom The system should allow users to undo and redo actions Yes No
Consistency and standards Terms should used consistent and conventional definitions Yes No
Error prevention The system should be designed to avoid errors rather than simply address Yes No
them
Recognition rather than recall The user should have easy access to references and not need to store and Possibly No
recall information from one part to another
Flexibility and efficiency of use There should be customization and shortcut tools available for experienced Possibly No
users to adapt the system to their needs
Aesthetic and minimalist design Dialogues should avoid displaying unnecessary information No No
Help users recognize, diagnose, Error messages should be presented in terms the user can understand and No No
and recover from errors suggest ways to address the error
Help and documentation Help tools should be readily available, searchable, and explain things step by Yes No
step

The structured interviews revealed generally positive ity of feedback provided to staff varies between units. Some
opinions of the hospital’s safety culture, although with the ca- units were seen as much more punitive than others, although
veat that participants believe other, unspecified units to be less there is an official effort in the form of an in-hospital newslet-
safety-conscious and less dedicated to learning from incidents ter to remind staff of the usefulness of incident reporting in
than their own. Participants reported that the amount and qual- general and near miss reporting in particular. Coordination be-

Downloaded from hcs.sagepub.com at HFES-Human Factors and Ergonomics Society on January 6, 2016
2015 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 171

tween units when major policy changes need to be made is dif- written narratives accompanied by stickers containing patient
ficult, and there does not seem to be a procedure in place to information fills the same niche. In both cases, the narratives
collect and compare data from different units to look for simi- are submitted to designated staff members, who transcribe and
lar incidents or solutions to them elsewhere in the hospital. classify them in the reporting application. Participants esti-
Both the emergency department and a specialized cardi- mated that this cut the time for the typical staff member to
ology unit reported suffering serious time pressure problems, submit a report from 15–20 to 1–2 minutes, while also pre-
with staff feeling that they did not have sufficient time to serving enough information for unit managers and other ana-
make incident reports without being interrupted. As a result, lysts to make minor local policy changes. Information is not
both units developed shortcuts to try to make incident report- available at this time to judge the effectiveness of these local
ing faster without sacrificing quality. In the ED, with the hos- changes. Figure 2 describes the information flow within the
pital’s official support, a telephone hotline has been intro- incident reporting system in place at RGH, based on our re-
duced, while in the cardiology unit, a system involving short search on the issues.

Figure 2. The flow of information within the incident reporting system. Dashed lines indicate problematic areas.

DISCUSSION ways by not facilitating sharing of data between units. It is al-


so worth noting the other information flow represented by a
The above descriptions, in addition to some preliminary dotted line, from the incident reporting newsletter. The news-
information provided by the Patient Safety department before letter provides a highly condensed form of feedback, and we
data collection began, allow us to describe the methods in have not yet been able to evaluate whether this level and for-
place for reporting incidents (Fig. 2). An incident may be re- mat of feedback is encouraging.
ported directly into the web app or via one of the unit shortcut Usability heuristic assessment and comparison to web
methods. Designated quality coordinators receive reports and form guidelines point to several likely factors discouraging
ensure that the unit in which the incident took place is notified system use: time constraints due to the amount of information
to follow up on it. Unit managers then decide how to respond required by Quantros, lack of customizability and shortcuts for
to the incidents in their department, and request additional as- experienced users, and inconvenient error handling/technical
sistance from upper management when needed. Although par- support. Comparison of known organizationally-influenced
ticipants reported that this is sometimes extremely helpful—it features of how Quantros is used to those of a known success-
was through this process that the emergency hotline shortcut ful reporting system (ASRS) suggests additional likely dis-
became an official hospital policy—it also falls short in other couraging factors: lack of evidence that incident reports will

Downloaded from hcs.sagepub.com at HFES-Human Factors and Ergonomics Society on January 6, 2016
2015 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 172

be used to improve patient safety, fear of punishment, and lack ble web form design. In R. Mátrai (Ed.), User interfaces
of sufficient organizational encouragement to continue partic- (pp. 1–10). Rijeka, Croatia: InTech.
ipating. Additionally, common work experience suggests that Beaubien, J. M., & Baker, D. P. (2002). A review of selected
the issue of lack of sufficient time to provide a report could al- aviation human factors taxonomies, accident/incident re-
so be due to time not being available in the potential user’s porting systems and data collection tools. International
work schedule. Journal of Applied Aviation Studies, 2(2), 11–36.
Billings, C. (1998). Incident reporting systems in medicine
Further Research and experience with the aviation safety reporting system.
In R. Cook, D. Woods, & C. Miller (Eds.), A tale of two
At this point, we have not gathered sufficient data to de- stories: contrasting views of patient safety. Report from a
termine trends that contribute to more or less, or better or workshop on assembling the scientific basis for progress
worse, reporting. Additional interviews and usability tests are on patient safety. McLean, VA: National Patient Safety
needed, and the participant pool must include both unit man- Foundation at the AMA.
agers, to determine whether they are lacking important details Bleich, S. (2005). Medical errors: five years after the IOM re-
needed to understand the causes of incidents, and ordinary port. Issue Brief (Commonwealth Fund), 830, 1–15.
staff members, to determine whether the software and the or- Browne, G., & Rogich, M. (2001). An empirical investigation
ganization are truly providing the support they need to report of user requirements elicitation: Comparing the effective-
everything that is worth reporting. These will provide the ness of prompting techniques. Journal of Management In-
“depth” of the study, identifying more notably bad and notably formation Systems, 17, 223–249.
good factors affecting incident report frequency and useful- Busse, D., & Wright, D. (2000). Classification and analysis of
ness. The “breadth” will come in a final stage of data collec- incidents in complex medical environments. Topics in
tion, a hospital-wide survey gauging the staff’s experience health information management, 20(4), 1–11.
with these different factors. Commission on Accreditation of Medical Transport Systems.
The ultimate goal of this study is to promote the use of in- (2012). Safety culture survey [survey instrument]. Re-
cident reporting in such a way that “sharp end” staff feel not trieved from http://www.camts.org/Safety-Culture-
only able but actively encouraged to identify and fix the most Survey.html
important safety problems they encounter. To this end, we will Crandall, B., Klein, G. A., & Hoffman, R. R. (2006). Working
use the survey results to categorize issues by their common- minds: A practitioner’s guide to cognitive task analysis.
ness and severity, which we can then use to recommend spe- MIT Press.
cific improvements based on a review of the literature to iden- Erler, C., Edwards, N. E., Ritchey, S., Pesut, D. J., Sands, L.,
tify where others have already succeeded with similar prob- & Wu, J. (2013). Perceived patient safety culture in a crit-
lems. It is our hope and goal that this will encourage im- ical care transport program. Air Medical Journal, 32(4),
provement both locally and at other facilities in the “Safety-I” 208–215.
sense of learning from past incidents and in the “Safety-II” Feldman, S., Schooley, B., & Bhavsar, G. (2014). Health in-
sense of noting where and why successes occur. formation exchange implementation: Lessons learned and
critical success factors from a case study. Journal of Med-
REFERENCES ical Internet Research Medical Informatics, 2(2), e19.
Gupta, K., & Cook, T. (2013). Accidental hypoglycaemia
Agency for Healthcare Research and Quality. (2013). Interim caused by an arterial drug flush error: A case report and
update on 2013 annual hospital-acquired condition rate contributory causes analysis. Anaesthesia, 68, 1179–1187.
and estimates of cost savings and deaths averted from Hollnagel, E., Wears, R., & Braithwaite, J. (2015). From Safe-
2010 to 2013 (Report No. 15-0011-EF). Retrieved from ty-I to Safety-II: A white paper. Authors.
http://www.ahrq.gov/professionals/quality-patient-safety/ Institute of Medicine. (1999). To Err is Human: Building a
pfp/interimhacrate2013.pdf safer health system. Committee on Health Care in Ameri-
Agency for Healthcare Research and Quality. (2014a). Hospi- ca. Institute of Medicine. Washington, DC: National
tal survey on patient safety culture: Items and dimensions. Academy Press.
[Supplemental materials for data collection instrument]. Institute of Medicine. (2001). Crossing the quality chasm: A
Rockville, MD. Retrieved from http://www.ahrq.gov/ pro- new health system for the 21st century. Washington, DC:
fessionals/quality-patient-safety/patientsafetyculture/ hos- National Academies Press.
pital/index.html Institute of Medicine. (2004). Patient safety. Washington, DC:
Agency for Healthcare Research and Quality. (2014b). Hospi- National Academies Press.
tal survey on patient safety. [Data collection instrument]. Jha, A., Kuperman, G., Teich, J., Leape, L., Shea, B., Ritten-
Rockville, MD. Retrieved from http://www.ahrq.gov/ pro- berg, E., & Bates, D. W. (1998). Identifying adverse drug
fessionals/quality-patient-safety/patientsafetyculture/ in- events: Development of a computer-based monitor and
dex.html comparison with chart review and stimulated voluntary
Bargas-Avila, J., Brenzikofer, O., Roth, S., Tuch, A., Orsini, report. Journal of the American Medical Informatics As-
S., & Opwis, K. (2010). Simple but crucial user interfaces sociation, 5(3), 305–314.
in the world wide web: Introducing 20 guidelines for usa- Johnson, C. (2002). Reasons for the failure of incident report-
ing in the healthcare and rail industries. In Components of

Downloaded from hcs.sagepub.com at HFES-Human Factors and Ergonomics Society on January 6, 2016
2015 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 173

system safety: Proceedings of the 10th safety-critical sys- Nielsen, J. (1994). Heuristic evaluation. In J. Nielsen & R.
tems symposium (pp. 31–57). Springer. Mack (Eds.), Usability Inspection Methods (pp. 25–62).
Kagan, I., & Barnoy, S. (2013). Organizational safety culture New York, NY: John Wiley & Sons.
and medical error reporting by Israeli nurses. Journal of State of New York Department of Health. (2007). New York
Nursing Scholarship, 45(3), 273–280. Patient Occurrence Reporting and Tracking System—
Murff, H., Patel, V., Hripcsak, G., & Bates, D. (2003). Detect- NYPORTS—2005, 2006, 2007 Report (Tech. Rep.). Alba-
ing adverse events for patient safety research: A review of ny, NY: State of New York Department of Health.
current methodologies. Journal of Biomedical Informat- Taylor, J. B. (2012). Safety culture: Assessing and changing
ics, 36, 131–143. the behaviour of organisations. Gower Publishing, Ltd.

Downloaded from hcs.sagepub.com at HFES-Human Factors and Ergonomics Society on January 6, 2016
View publication stats

Das könnte Ihnen auch gefallen