Preventing Hospital Acquired Infections through a Workﬂow-based

Cyber-physical System

Maria Iuliana Bocicor

1,2

, Arthur-Jozsef Molnar

1,2

and Cristian Taslitchi

SC Info World SRL, Bucharest, Romania

Faculty of Mathematics and Computer Science, Babes-Bolyai University, Cluj-Napoca, Romania

Faculty of Automatic Control and Computers, University Politehnica of Bucharest, Bucharest, Romania

Keywords:

Hospital Acquired Infection, Nosocomial Infection, Outbreak, Clinical Workﬂow Monitoring, Cyber-physical

System.

Abstract:

Hospital acquired infections (HAI) are infections acquired within the hospital from healthcare workers, pa-

tients or from the environment, but which have no connection to the initial reason for the patient’s hospital

admission. HAI are a serious world-wide problem, leading to an increase in mortality rates, duration of hos-

pitalisation as well as signiﬁcant economic burden on hospitals. Although clear preventive guidelines exist,

studies show that compliance to them is frequently poor. This paper details the software perspective for an

innovative, business process software based cyber-physical system that will be implemented as part of a Eu-

ropean Union-funded research project. The system is composed of a network of sensors mounted in different

sites around the hospital, a series of wearables used by the healthcare workers and a server side workﬂow en-

gine. For better understanding, we describe the system through the lens of a single, simple clinical workﬂow

that is responsible for a signiﬁcant portion of all hospital infections. The goal is that when completed, the

system will be conﬁgurable in the sense of facilitating the creation and automated monitoring of those clinical

workﬂows that when combined, account for over 90% of hospital infections.

1 INTRODUCTION

Hospital acquired infections (HAI) or nosocomial in-

fections are deﬁned as infections ”acquired in hospi-

tal by a patient who was admitted for a reason other

than that infection. An infection occurring in a pa-

tient in a hospital or other healthcare facility in whom

the infection was not present or incubating at the time

of admission. This includes infections acquired in the

hospital but appearing after discharge, and also occu-

pational infections among staff of the facility” (World

Health Organization, 2002). Thus, in addition to de-

creasing the quality of life and increasing mortality

rates, the duration of hospitalisation as well as the

costs of medical visits for patients, HAI represent a

direct occupational risk for healthcare workers, which

signiﬁcantly affects costs and has the potential of cre-

ating personnel deﬁcits in case of an outbreak.

Existing research shows that HAI are prevalent

across the globe, regardless of geographical, politi-

cal, social or economic factors (World Health Orga-

nization, 2002), (World Health Organization, 2010).

Even more compelling is the fact that while the so-

phistication of medical care is constantly increas-

ing, reported HAI rates have not seen meaningful

decrease (Tikhomirov, 1987), (Coello et al., 1993),

(World Health Organization, 2010), (European Cen-

tre for Disease Prevention and Control, 2015).

Studies undertaken between 1995 and 2008 in

several developed countries have revealed infection

rates between 5.1% to 11.6% (World Health Organi-

zation, 2010); data from lesser developed countries

is in many cases limited and deemed of low quality.

According to the European Centre for Disease Pre-

vention and Control, approximately 4.2 million HAI

occurred in 2013 alone in European long-term care fa-

cilities, the crude prevalence of residents with at least

one HAI being 3.4%; this translates to more than 100

thousand patients on any given day (European Cen-

tre for Disease Prevention and Control, 2015). The

total cases of HAI in Europe amount to 16 million

extra hospitalisation days, 37.000 attributable deaths

and an economic burden of e7 billion in direct costs

(World Health Organization, 2010). A study con-

ducted in the United Kingdom (Plowman, 2000) con-

cluded that patients who developed HAI stayed in

Bocicor, M., Molnar, A-J. and Taslitchi, C.

Preventing Hospital Acquired Infections through a Workﬂow-based Cyber-physical System.

In Proceedings of the 11th International Conference on Evaluation of Novel Software Approaches to Software Engineering (ENASE 2016), pages 63-68

ISBN: 978-989-758-189-2

hospital 2.5 times longer and the hospital costs (for

nursing care, hospital overheads, capital charges, and

management) tripled. In the USA, in 2002 alone, 1.7

million patients were affected by HAI and the annual

economic impact was approximately US$6.5 billion

in 2004. Besides all these resource expenses, what

is even worse is that HAI are responsible for an in-

crease in mortality rates, as these infections lead to

death in 2.7% of cases (World Health Organization,

2009). A more recent study (Magill and et al., 2014)

conducted in 183 US hospitals revealed that 4% of the

patients had one or more HAI, which lead to an esti-

mated number of 648.000 patients with HAI in acute

care hospitals in the US, in 2011. According to a re-

cent study (M.L. et al., 2015), the prevalence of HAI

in Southeast Asia between 2000 and 2012 was 9%,

the excess length of stay in hospitals of infected pa-

tients varied between 5 and 21 days and the attributed

mortality was estimated between 7% and 46%. In

Canada, more than 220.000 HAI result in 8000 deaths

a year, making infections the fourth leading cause of

death in the country, with $129 million in extra costs

incurred in 2010 (Government of Newfoundland and

Labrador. Department of Health and Community Ser-

vices, 2013), (Canadian Union of Public Employees,

2014). In Australia, there is an estimated number of

200.000 cases of HAI per year, resulting in 2 mil-

lion hospitalization days (Australian Commission on

Safety and Quality in Health Care, 2008).

The most common target sites of HAI are the uri-

nary and respiratory tracts and areas involved in in-

vasive procedures or catheter insertion areas. While

the methodology for prevention exists, it is often ig-

nored due to lack of time, unavailability of appropri-

ate equipment or because of inadequate staff training.

Research shows that the most important transmission

route for HAI are members of staff coming into con-

tact with patients or contaminated equipment with-

out following proper hygiene procedures (Hammer,

2013).

During the twentieth century, several speciﬁc

measures were taken to prevent the occurrence and

spread of infections, which have been translated into

a series of instructions for controlling the vectors that

propagate infection as well as to properly manage out-

breaks and epidemics. Originally, these instructions

were established as guidelines for healthcare workers,

later they were also transformed into prevention rules

included in the documentation concerning workplace

safety, and in recent years they were incorporated into

various software or cyber-physical solutions to moni-

tor and ensure compliance. An unquestionable beneﬁt

and substantial improvement in this respect has been

brought by the Internet of Things (IoT) technologies,

which are currently employed in healthcare and many

other areas of life. Since the use of various IoT moni-

toring systems, signiﬁcant improvement has been ob-

served regarding compliance to hygiene regulations

and prevention standards, as well as decreases in the

rates of infections.

In this paper we present an IoT-based cyber-

physical system that targets HAI prevention, on which

development has started under funding from the Eu-

ropean Union. Integrating a network of sensors

that monitor clinical workﬂows and ambient condi-

tions with monitoring software, the system will pro-

vide real-time information and alerts. The system

will monitor general processes known to affect HAI

spread such as cleaning and equipment maintenance,

together with clinical processes at risk for HAI such

as catheter insertion, postoperative care or mechanical

ventilation. In this paper, we will illustrate the sys-

tem using a clinical workﬂow that is often involved in

HAI transmission, together with a motivating exam-

ple that details a software perspective regarding how

the system ensures compliance to established preven-

tive guidelines.

2 RELATED WORK

Several automated solutions have been implemented

to avert and reduce HAI and their impact. The un-

derlying idea used by several of these systems is con-

tinuous monitoring of healthcare workers’ hand hy-

giene and real-time alert generation in case of non-

compliance with established guidelines. The ﬁnal

purpose is to modify human behaviour towards bet-

ter hand hygiene compliance. This is usually accom-

plished using wearable devices worn by healthcare

workers, which interact with sensors placed in key

hospital locations in order to record personnel activi-

ties and hand hygiene events. When a hand hygiene

event is omitted or performed, the device provides

visual, auditory or haptic notiﬁcation. The Intelli-

gentM (Ryan, 2013) and Hyginex (Hyginex, 2015)

systems use bracelet-like devices that are equipped

with motion sensors to ensure that hand sanitation is

correctly performed. Biovigil technology (BIOVIGIL

Healthcare Systems, Inc., 2015) and MedSense (Gen-

eral Sensing, 2014) are designed for the same pur-

pose, only in these cases bracelets are replaced with

badges worn by the healthcare workers. The Biovigil

device uses chemical sensors to detect whether hand

hygiene is in accordance with established standards.

The systems can be conﬁgured to remind clinicians

to disinfect their hands before entering patient rooms,

or before procedures such as intravenous infusions or

ENASE 2016 - 11th International Conference on Evaluation of Novel Software Approaches to Software Engineering

catheter insertion. Furthermore, these systems record

hygiene events, centralise them and enable analysis,

visualisation and report generation. Unlike the solu-

tions presented so far, the SwipeSense (Swipe Sense,

2015) system employs small devices that are alcohol-

based and easy to use gel dispensers which can be

worn by medical personnel. This exempts healthcare

workers from interrupting their activities in order to

walk to a sink or a disinfectant dispenser (Simonette,

2013).

A different type of system that is of assistance in

the ﬁght against infection is Protocol Watch (Philips,

2015), a decision support system used to improve

compliance with the ”Surviving Sepsis Campaign”

international guidelines (Surviving Sepsis Campaign,

2012) for the prevention and management of sepsis.

Protocol Watch regularly checks certain medical pa-

rameters of the patients (Philips, 2015), its main goal

being to reduce the time period between the debut of

sepsis (the moment when it is ﬁrst detected) and the

beginning of treatment. If the system detects that cer-

tain conditions that may cause sepsis are met, it alerts

the medical staff and indicates which tests, observa-

tions and interventions must be performed, according

to prevention and treatment protocols.

A different approach for the prevention of infec-

tion is taken by Xenex (Xenex, 2015): the Xenex

”Germ-Zapping Robot” can disinfect a room by us-

ing pulses of high-intensity, high-energy ultraviolet

light. The robot must be taken inside the room to

be disinfected and in most cases, the deactivation of

pathogens takes place in ﬁve minutes. After disinfec-

tion, the room will remain at a low microbial load, un-

til it is recontaminated by a patient, healthcare worker

or through the ventilation system.

Another relevant issue regards the problem of

identifying control policies and optimal treatment in

infection outbreaks, as introduced in (Curtis et al.,

2013). The authors propose a comprehensive ap-

proach using electronic health records to build health-

care worker contact networks with the objective of

putting into place efﬁcient vaccination policies in case

of outbreaks. Relevant software systems developed to

also improve treatment policies in case of outbreaks

and epidemics are RL6: Infection (RL Solutions,

2015) - a software solution developed to assist hos-

pitals in the processes of controlling and monitoring

infections and outbreaks, and Accreditrack (Excelion

Technology Inc., 2013) - a software system designed

to ensure compliance with hand hygiene guidelines

and to verify nosocomial infection management pro-

cesses.

3 MOTIVATING EXAMPLE -

HAND HYGIENE

Research shows that most cases of hospital infection

are tightly connected to certain clinical workﬂows

(World Health Organization, 2002). Thus, we pro-

pose a cyber-physical system which monitors work-

ﬂows considered relevant in HAI propagation. Al-

though there are other ICT automated solutions that

target the prevention of HAI, there are no other sys-

tems that use the clinical workﬂow monitoring based

approach, to our knowledge. In order to successfully

prevent infection and outbreaks, the system must fo-

cus on the typical sites where these infections usu-

ally occur (urinary and respiratory tracts, sites of

surgery or invasive procedures (World Health Orga-

nization, 2002)), as well as track workﬂows that are

not infection-site speciﬁc, but which have a signiﬁ-

cant contribution to HAI rates, such as hand hygiene

and transmission from the environment.

A hardware-centric motivating example of the

proposed system was presented in (Shhedi et al.,

2015). The authors of (Shhedi et al., 2015) focus

on the hardware components of the system, namely

types and location of sensors used to monitor clini-

cal activities and the wearable technology employed

by the medical staff for identiﬁcation, monitoring and

Figure 1: A typical hospital room. Figure adapted from

(Shhedi et al., 2015).

Preventing Hospital Acquired Infections through a Workﬂow-based Cyber-physical System

alerting. The purpose of our paper is to illustrate the

software-side of the proposed cyber-physical system

using a motivating example based on one of the most

relevant, and often occurring clinical workﬂows.

On the software side, the system will model and

encode clinical workﬂows using the Business Process

Modelling and Notation (BPMN) (Object Manage-

ment Group, 2015) standard and it will be compati-

ble with leading medical informatics standards such

as HL7 V3 (Health Level Seven International, 2016),

thus allowing seamless interconnection to hospital in-

frastructure via HL7-compliant Hospital Information

Systems (HIS). In addition, it will integrate a network

of hardware sensors, able to identify ongoing clini-

cal processes, provide location information and track

the use of hospital equipment and materials. Health-

care workers will use wearable devices that continu-

ously monitor their location and activity. The central

hub of the system will be a server side engine able to

load and execute BPMN-based workﬂows. Integra-

tion with the HIS enables the retrieval of key informa-

tion regarding patients, such as admissions, transfers

and discharges as well as records of past of planned

invasive interventions, which can be used to deter-

mine the level of risk. Hence, infections and outbreak

management could be improved so that in case of sus-

picion, the locations of previous admissions, as well

patients of members of staff considered at risk can be

contacted.

The modelled clinical workﬂows will be executed

using a BPMN engine. When the workﬂow leads to

a state that is a risk of HAI, the engine will gener-

ate alerts that are received directly by the involved

healthcare workers. In this paper we illustrate the in-

terplay between the hardware and software compo-

nents of the proposed system using a motivating ex-

ample based on hand hygiene, which remains one of

the most common pathways of HAI transmission. We

must note that the provided example is only used to

portray the workﬂow-based system, and that the ﬁnal

implementation will allow the creation and monitor-

ing of a variety of clinical workﬂows, that will be ex-

ecuted by the BPMN engine whenever necessary.

We illustrate our motivating example using a typ-

ical scenario: a healthcare worker enters a room with

two beds, and interacts with both inpatients before ex-

iting. Both common as well as highly relevant to hand

hygiene, our example is illustrated in Figure 2, using

simpliﬁed BPMN-like notation to reduce the number

of decision points and emphasize readability.

According to established hand hygiene guidelines

(World Health Organization, 2016), upon entering a

patient room, workers must perform hand disinfec-

tion. If this procedure is skipped or performed in-

adequately (e.g. without disinfectant, shorter wash-

ing time than recommended), the system will gener-

ate an alert to warn the clinician about the detected

non-compliance. After each patient contact, and be-

fore leaving the room, hand disinfection should again

be correctly performed. All these events and alert re-

ports will be persisted to enable later analyses, such as

identiﬁcation of an outbreak’s patient zero and route

of transmission.

Figure 1, which was adapted from (Shhedi et al.,

2015) illustrates a typical patient room, with two

beds, a sink and a bathroom. As soon as the healthcare

worker enters it, the Radio-Frequency Identiﬁcation

(RFID) tag and motion sensor combination detect this

and identify them. The system records and interprets

the data received from the sensors and the BPMN en-

gine starts a new instance of the relevant workﬂows,

including the one for hand-hygiene. According to the

hand hygiene workﬂow, the clinician should perform

the hand disinfection procedure before going near a

patient and before leaving their surrounding. In our

example, this is achieved using the sink or the bath-

room sink and the disinfectant dispenser. All of them

have inexpensive sensors together with RFID tags and

Bluetooth Low Energy transceivers; their role is to

provide input to the workﬂow engine. The workﬂow

engine records received data into the persistent repos-

itory. Furthermore, by running the workﬂow, the soft-

ware engine ascertains that hand hygiene guidelines

were observed. In our example, the healthcare worker

performs hand disinfection before contact with the

ﬁrst patient, which is recorded by the sensors inte-

grated with the sink and the disinfectant dispenser.

However, the worker can move to the second patient

directly, as shown in Figure 2. In this case, the system

records their proximity to the second bed via RFID;

if a disinfection event that is compliant with guide-

lines was not recorded before the contact, the system

interprets this as non-compliance, and emits an alert

that is recorded and received by the healthcare worker

through their wearable device. Once they become

compliant by undergoing hand disinfection, they can

resume contact with the patient. Information that is

persisted is planned to be reused at later dates and in

the context of more than one workﬂow, including at

least all the workﬂows active during that time. As an

example, the room entry and exit events from Figure

2 can be used for ﬁnding the source of an outbreak, or

tracking its propagation. In our example, the current

instance of the hand hygiene workﬂow ends once the

healthcare worker exists the patient room.

ENASE 2016 - 11th International Conference on Evaluation of Novel Software Approaches to Software Engineering

Figure 2: Example of hand-hygiene relevant clinical workﬂow (simpliﬁed for readability).

4 CONCLUSIONS

This paper is centred on technology-driven preventive

measures for a serious public health issue, namely

hospital acquired infections. While most transmis-

sion routes are well understood and standard guide-

lines exist to curb them, available research shows that

in practice, they are either not applied, or not ap-

plied thoroughly. As such, we propose an innova-

tive technological solution for preventing HAI and

outbreaks. Our proposed approach will employ a

standards-compliant, workﬂow based, cyber-physical

system able to monitor and enforce compliance with

clinical workﬂows that are associated with over 90%

of HAI instances, in accordance with current guide-

lines and best practices.

We have illustrated an overall picture of the sys-

tem using a software workﬂow perspective through

a motivating example based on hand-hygiene guide-

lines. However, our aim is to enable the creation,

conﬁguration and execution of location-speciﬁc clin-

ical workﬂows using BPMN-like notation. As ex-

isting research shows that infections within the uri-

nary tract, respiratory tract, surgical sites, or skin and

soft tissue (World Health Organization, 2002) account

for over 75% of the total number, we aim to also

evaluate the system using workﬂows related to those

sites. Another category is represented by workﬂows

that are not infection-site speciﬁc, but which repre-

sent transmission vectors for numerous types of infec-

tions: hand hygiene (such as presented in this paper)

and transmission from the environment.

Another future functionality, that currently sits be-

yond the scope of our research is a graphically inter-

active, conﬁgurable component that allows the man-

agement of the monitored workﬂows, by taking into

account the speciﬁc infrastructure of the deployed

sensors and of the unit of care. Furthermore, addi-

tional components that can be added to the system in-

clude advanced analysis tools for the gathered data

that allow building risk maps and contact networks

enabling new ways of pinpointing elusive reasons of

hospital infection that occur even when best practices

are adhered to.

In addition to the software-based functionalities,

another issue that must be addressed is related to the

legal aspects regarding the deployment of such a sys-

tem, more precisely those regarding data protection.

Considering that the system involves tracking of pa-

tients and healthcare workers alike, the software side

of such systems must remain compliant with regula-

tions regarding the protection of highly-sensitive per-

sonal data.

ACKNOWLEDGEMENT

This work was supported by a grant of the Romanian

National Authority for Scientiﬁc Research and Inno-

vation, CCCDI UEFISCDI, project number 9831

https://www.eurostars-eureka.eu/project/id/9831

Preventing Hospital Acquired Infections through a Workﬂow-based Cyber-physical System

REFERENCES

Australian Commission on Safety and Quality in Health

Care (2008). Reducing harm to patients from health

care associated infection: the role of surveillance.

BIOVIGIL Healthcare Systems, Inc. (2015). Biovigil and

our team. http://www.biovigilsystems.com/about/.

Accessed: 2015-12-02.

Canadian Union of Public Employees (2014). Health care

associated infections: backgrounder and fact sheet.

http://cupe.ca/health-care-associated-infections-

backgrounder-and-fact-sheet. Accessed: 2016-01-07.

Coello, R., Glenister, H., Fereres, J., Bartlett, C., Leigh, D.,

Sedgwick, J., and Cooke, E. (1993). The cost of infec-

tion in surgical patients: a case-control study. Journal

of Hospital Infections, 25:239–250.

Curtis, D., Hlady, C., Kanade, G., Pemmaraju, S., Pol-

green, P., and Segre, A. (2013). Healthcare worker

contact networks and the prevention of hospital-

acquired infections. Plos One. DOI: 10.1371/jour-

nal.pone.0079906.

European Centre for Disease Prevention and Control

(2015). Annual epidemiological report. antimicrobial

resistance and healthcare-associated infections. 2014.

http://ecdc.europa.eu/en/publications/Publications/

antimicrobial-resistance-annual-epidemiological-

report.pdf. Accessed: 2016-01-06.

Excelion Technology Inc. (2013). Accreditrack. http://

www.exceliontech.com/accreditrack.html. Accessed:

2015-12-04.

General Sensing (2014). Medsense clear. hand

hygiene compliance monitoring. http://

www.generalsensing.com/medsenseclear/. Accessed:

2015-12-11.

Government of Newfoundland and Labrador. Department

of Health and Community Services (2013). Hai re-

port 2009-2012. http://www.health.gov.nl.ca/health/

publichealth/cdc/hai/hai 2012.pdf. Accessed: 2016-

01-07.

Hammer, S. (2013). Hand washing: Reducing nosoco-

mial infections. http://iwsp.human.cornell.edu/ﬁles/

2013/09/Hand-Washing-Reducing-Nosocomial-

Infections-2j1mlfb.pdf. Accessed: 2016-01-06.

Health Level Seven International (2016). Hl7 version

3 product suite. https://www.hl7.org/implement/

standards/product brief.cfm?product id=186. Ac-

cessed: 2016-01-07.

Hyginex (2015). Introducing hyginex generation 3. http://

www.hyginex.com/. Accessed: 2015-12-04.

Magill, S. S. and et al., J. R. E. (2014). Multistate point-

prevalence survey of health careassociated infections.

The New England Journal of Medicine, 370:1198–

1208. DOI: 10.1056/NEJMoa1306801.

M.L., L., A., A., and G., M. (2015). The burden

of healthcare-associated infections in southeast asia:

A systematic literature review and meta-analysis.

Clinical Infectious Diseases, 60:1690–9. DOI:

10.1093/cid/civ095.

Object Management Group (2015). Business process model

and notation. http://www.bpmn.org/. Accessed: 2016-

01-07.

Philips (2015). Protocolwatch - ssc sepsis. http://

www.healthcare.philips.com/main/products/patient

monitoring/products/protocol watch/. Accessed:

2015-11-25.

Plowman, R. (2000). The socioeconomic burden of hospital

acquired infection. Eurosurveillance, 5.

RL Solutions (2015). The rl6 suite / infection surveillance.

http://www.rlsolutions.com/rl-products/infection-

surveillance. Accessed: 2015-12-02.

Ryan, J. (2013). Medtech proﬁles: Intelligentm a

simple yet powerful app to dramatically reduce

hospital-acquired infections. https://medtechboston.

medstro.com/proﬁles-intelligentm/. Accessed: 2015-

12-02.

Shhedi, Z. A., Moldoveanu, A., Moldoveanu, F., and

Taslitchi, C. (2015). Real-time hand hygiene monitor-

ing system for hai prevention. In The 5th IEEE Inter-

national Conference on E-Health and Bioengineering

- EHB 2015, Grigore T. Popa University of Medicine

and Pharmacy, Iai, Romania.

Simonette, M. (2013). Tech solutions to hospital acquired

infections. http://www.healthbizdecoded.com/2013/

06/tech-solutions-to-hospital-acquired-infections/.

Accessed: 2015-12-03.

Surviving Sepsis Campaign (2012). International guide-

lines for management of severe sepsis and sep-

tic shock: 2012. http://www.sccm.org/Documents/

SSC-Guidelines.pdf. Accessed: 2015-11-25.

Swipe Sense (2015). Hand hygiene. redeﬁned. https://

www.swipesense.com/. Accessed: 2015-12-04.

Tikhomirov, E. (1987). Who programme for the control of

hospital infections. Chemioterapia, 6:148–151.

World Health Organization (2002). Prevention of

hospital-acquired infections - a practical guide.

http://www.who.int/csr/resources/publications/

whocdscsreph200212.pdf. Accessed: 2015-11-25.

World Health Organization (2009). WHO guidelines on

hand hygiene in health care. http://apps.who.int/iris/

bitstream/10665/44102/1/9789241597906 eng.pdf.

Accessed: 2016-01-06.

World Health Organization (2010). The burden of

health care-associated infection worldwide.

http://www.who.int/gpsc/country work/summary

20100430 en.pdf. Accessed: 2016-01-07.

World Health Organization (2016). Clean care is safer care

- ﬁve moments for hand hygiene. http://www.who.int/

gpsc/tools/Five moments/en/. Accessed: 2016-01-07.

Xenex (2015). Xenex germ-zapping robots. http://

www.xenex.com/. Accessed: 2015-12-11.

ENASE 2016 - 11th International Conference on Evaluation of Novel Software Approaches to Software Engineering