Abstract Hand hygiene compliance is the most significant, modifiable cause of hospital-acquired infections, yet national averages for compliance rates remain unsatisfactory

SYSTEMS-LEVEL QUALITY IMPROVEMENT

A Hand Hygiene Compliance Check System: Brief Communication on a System to Improve Hand Hygiene Compliance in Hospitals and Reduce Infection

Tracey S. Hong1,2 & Emily C. Bush1,3 & Morgan F. Hauenstein1 & Alec Lafontant1 & Chen Li1 & Jonathan P. Wanderer5,4 & Jesse M. Ehrenfeld4,5,6,7

Received: 13 March 2015 /Accepted: 30 April 2015 /Published online: 12 May 2015 # Springer Science+Business Media New York 2015

Abstract Hand hygiene compliance is the most significant, modifiable cause of hospital-acquired infections, yet national averages for compliance rates remain unsatisfactory. Noncompliance can contribute to patient mortality, extended hospital stays, higher re-admission rates, and lower reimburse- ment for hospitals under the Patient Protection and Affordable Care Act. Although several hand sanitizing tracking systems currently exist, they pose problems of personal tracking, workflow interference, system maintenance concerns, among others. Considering these barriers, we created a prototype sys- tem that includes compliance rate tracking, real-time sanitiza- tion reminders, and a data archive for future studies.

Keywords Hand hygiene compliance . Hand sanitation .

Handwashing . Nosocomial infections . Reinforcement system . Compliance rate tracking . Rreadmittance rates

Introduction

Over 98,000 deaths per year in the United States can be attributed to hospital-acquired infections – many of which can be attributed to methicillin-resistant Staphylococcus aureus (MRSA) [1]. The frequency of MRSA infections is inversely proportional to hand hygiene compliance, which is the most significant, modifiable cause of noso- comial infections in hospitals [2]. Hand hygiene compli- ance is defined as properly washing one’s hands with soap and water or an antiseptic agent before and after all pa- tient or patient environment contact [3]. Vanderbilt University Medical Center (VUMC) uses manual observa- tion, education, and incentives to address this problem, but these are time and resource-intensive methods of en- forcement. Other hospitals have employed methods of compliance rate enforcement and tracking but they have proved to be unsatisfactory, as they are either subject to human error, are extremely costly, or infringe on employee privacy.

Non-compliance with hand hygiene is a critical prob- lem because it impacts our ability to provide optimal care, leads to increased transmission of nosocomial diseases, and can lead to significant additional costs incurred by patients and hospitals. These risks could be mitigated through the implementation of sensing technology that is designed around the shortcomings of existing methods. In the 2013 Centers for Disease Control and Prevention (CDC) Antibiotic Resistance Threats report, it was esti- mated that the direct healthcare cost of antibiotic

This article is part of the Topical Collection on Systems-Level Quality Improvement

Tracey S. Hong and Emily C. Bush contributed equally to this work.

* Tracey S. Hong tracey.s.hong@Vanderbilt.Edu

1 Department of Biomedical Engineering, Vanderbilt University, Nashville, TN 37232, USA

2 School of Medicine, Vanderbilt University, 201 Light Hall, Nashville, TN 37232, USA

3 Institute of Imaging Science, Vanderbilt University, Nashville, TN 37232, USA

4 Department of Anesthesiology, Vanderbilt University, Nashville, TN 37232, USA

5 Department of Biomedical Informatics, Vanderbilt University, Nashville, TN 37232, USA

6 Department of Surgery, Vanderbilt University, Nashville, TN 37232, USA

7 Department of Health Policy, Vanderbilt University, Nashville, TN 37232, USA

J Med Syst (2015) 39: 69 DOI 10.1007/s10916-015-0253-z

resistance was near $20 billion annually. This figure does not include the additional estimated $35 billion cost to society attributed to lost productivity [4]. Additionally, the largest cost for patients is often due to extended hos- pital stays. Research shows that increasing sanitization rates can reduce mortality rates and length-of-stay. The North Carolina Children’s Hospital Pediatric Intensive Care Unit implemented sanitization strategies that cut death rates by 2.3 %, decreased hospitals stays by 2.3 days and cut costs by over $12,000 per case [5].

A sustainable accountability and compliance checking sys- tem is needed to address the aforementioned problems. We therefore designed and implemented a prototype system that includes the ability to provide compliance rate tracking, real- time behavior modification, and a mechanism to archive in- formation to be used for future studies comparing infection and compliance rates.

Methods

Our study protocol was approved by the Vanderbilt Institutional ReviewBoard. During the design phase, we iden- tified a series of goals, constraints, and assumptions, which are as follows:

Goals

1. Integrate foam dispenser sensors and a door sensor into a dual-sensing system with a warning alarm to create a sus- tainable device that encourages active avoidance learning.

2. Accurately track hand hygiene compliance rates with the dual-sensing system at VUMC.

3. Store hand hygiene compliance rates in a centralized database.

Constraints

1. Our system has no personal tracking. This is an important feature because of personal privacy concerns which might otherwise impede end-user adoption. [6, 7]

2. The systemmust have easily serviceable parts. A hospital- wide system could include thousands of units, therefore it is critical that those units would not require constant main- tenance. It is also paramount that the system does not interrupt workflow.

3. Finally, to assist with compliance rates moving forward, this system should have the ability to archive data for compliance studies and trend determination. [8, 9]

Assumptions

1. Only one person enters at a time and all entries are human. This assumption is made to exclude the entry of carts, gurneys, and other medical equipment.

2. Everyone must sanitize prior to entry. There are a few instances in which a person is not required to sanitize (if hands are still wet from a prior sanitization), but we are excluding those events.

3. The patient is already in the room. 4. The hand sanitizing canisters are full and users know how

to correctly use them. 5. Sanitization occurs in a timely manner. After observing

patient rooms, we have set this time to be 2 s. 6. The system resets after 2 s. Specifically, this means that

the next event cannot take place until 2 s following the first.

It should be noted that we focused primarily on the sim- plest environment: a single patient room. This made testing the sensing system for accuracy and reliability straightfor- ward because there is a single entryway, fixed sanitizing stations, and one patient per room. Additionally, it helped to solidify our assumptions and constraints for the prototype.

System Design

Our system prototype is comprised of a doorway sensor, two sanitizer dispenser sensors, a microprocessor and an alarm. The doorway sensor, used to track room entry, is an ultrasonic sensor (Maxbotix LV-EZ1 Ultrasonic Sensor, Max Range 6.45 m, Digital Output) attached to the outside of the doorframe. It works by creating continuous distance readings, which are constant in an empty doorway, by outputting correlated pulse widths (Fig. 1c). When the beam is broken (Fig. 1d), the change in distance is detected, which signals the activa- tion of doorway entry. The hospital’s existing hand san- itizing units were equipped with a BFlatback U-Ring^ (Fig. 1a), a device we created to house small IR beam detectors (SHARP Infrared Sensor (GP2D120XJ00F), Range: 4-30 cm, Analog Output). These are used to determine if the hand sanitizer has been used due to a hand breaking the beam under the nozzle (Fig. 1b). The sensors are wired to an Arduino microprocessor (Arduino Uno Micro-processing Board), housed on the wall, where the sensing information is processed. Using built-in Wi-Fi capabilities, the Arduino can send the data to the hospital’s wireless network, allowing the data to be captured in one centralized location for anal- ysis. The Arduino sends counts of both compliant and noncompliant events, and these numbers can be

69 Page 2 of 4 J Med Syst (2015) 39: 69

recorded to create compliance rates by department. In addition to sending data, when a non-compliant event occurs, the Arduino uses a built-in speaker to sound a short, audible alarm, reminding the person entering the room to sanitize.

The scenarios that our system can handle are shown in Fig. 2.

1. Scenario 1: The user would sanitize, activating the infrared sensor on the dispenser, and walk through the door within two seconds, activating the ultrason- ic sensor on the door. This would register as a com- pliant event.

2. Scenario 2: The user walks through the door first; activat- ing the ultrasonic beam and then uses the hand sanitizer inside the patient’s room within two seconds. This also registers as a compliant event. If the hand sanitizer inside the room is bypassed or activated after two seconds post room entry, an alarm will sound, signaling a non- compliant event.

3. Scenario 3: If a hand sanitizer on either side of the door is activated and there is no entry/exit within two seconds, no event is recorded.

This method accounts for most probable scenarios, even when multiple users are involved. Additionally, the user never interacts with the system because it is automated. The only interaction necessary would be for maintenance purposes (i.e. battery replacement or system repair).

Results

During a pilot testing period, 40 compliant and 40 noncom- pliant events were simulated in a single patient pre-operative holding room. With a successful detection of 97.5 % of the compliant events and 100 % of the non-compliant events, our system delivered an overall accuracy of 98.75%. The Arduino device exported the data to a local spreadsheet every 5 s. Our system also has the capability of wirelessly transmitting data to a centralized hospital database via broadcasted Wi-Fi net- works. The data spreadsheet has the ability to automatically calculate compliance rates for shifts, days, months, etc. based on specific areas of the hospital. The archived compliance data can be compared to the existing gold standard of measuring compliance rates (e.g. secret shoppers), to see whether the alarm system has a positive impact on hand hygiene compliance.

Future Directions and Conclusions

Future applications would include more complex situa- tions such as trauma bays, preoperative rooms and re- covery rooms, which have large quantities of beds sep- arated by curtains and mobile sanitizing stations. With the time and materials at hand, our proof of concept prototype achieved our goals and exhibited the potential for future development. The integration of high perfor- mance wireless sensors with an alternate micro- processing platform could allow for a noninvasive and sustainable hospital-wide sensor system. We have suc- cessfully demonstrated the temporal integration of IR and ultrasonic sensors to detect hand sanitizer use and doorway entry, as well as developed an alarm mecha- nism to increase compliance rates in hospitals. Our sys- tem has created a platform for exporting and archiving data to be saved and compared with existing gold stan- dards, such as secret shoppers. With the push for an increase in hand hygiene compliance rates, and decrease in hospital acquired infections and re-admittance rates, our device demonstrates the practicality and need for a system that can solve these problems.

Fig. 1 Hand sanitizing unit equipped with BFlatback U-Ring^ (a); Hand breaking IR beam on the unit during hand sanitization (b); Door with US beam mounted to frame (c); US beam on door being interrupted by human entry to room (d)

Fig. 2 Decision tree outlining the different operational scenarios which could occur

J Med Syst (2015) 39: 69 Page 3 of 4 69

References

1. Cummings, K. L., Anderson, D. J., and Kaye, K. S., Hand hygiene noncompliance and the cost of hospital-acquired methicillin-resistant Staphylococcus aureus infection. Infec Control Hospital Epidemiol Off J Soc Hospital Epidemiol Am 31:357–364, 2010. doi:10.1086/ 651096.

2. Lederer, J. W., Best, D., and Hendrix, V., A comprehensive hand hygiene approach to reducing MRSA health care-associated infec- tions. Joint Comm J Quality Patient Safety Joint Com Res 35:180– 185, 2009.

3. Scheithauer, S., Rosarius, A., Rex, S., Post, P., Heisel, H., Krizanovic, V., et al., Improving hand hygiene compliance in the anesthesia work- ing room work area: more than just more hand rubs. Am. J. Infect. Control 41:1001–6, 2013. doi:10.1016/j.ajic.2013.02.004.

4. Centers for Disease Control and Prevention (U.S.); National Center for Emerging Zoonotic and Infectious Diseases (U.S.); National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention (U.S.); National Center for Immunization and Respiratory Diseases (U.S.) (2013). Antibiotic Resistance Threats in the United States,

2013. Retrieved from http://www.cdc.gov/drugresistance/threat-re- port-2013/pdf/ar-threats-2013-508.pdf

5. Harris, B. D., Hanson, C., Christy, C., Adams, T., Banks, A., Willis, T. S., and Maciejewski, M. L., Strict Hand Hygiene And Other Practices Shortened Stays And Cut Costs And Mortality In A Pediatric Intensive Care Unit. Health Aff. 30(9):1751–1761, 2011. doi:10.1377/hlthaff.2010.1282.

6. Rosenbaum, B. P., Radio frequency identification (RFID) in health care: privacy and security concerns limiting adoption. J. Med. Syst. 38(3):19, 2014. doi:10.1007/s10916-014-0019-z.

7. Martínez-Pérez, B., de la Torre-Díez, I., and López-Coronado, M., Privacy and security in mobile health apps: a review and recommen- dations. J. Med. Syst. 39(1):181, 2015. doi:10.1007/s10916-014- 0181-3.

8. Yang, C. T., Liao, C. J., Liu, J. C., Den, W., Chou, Y. C., and Tsai, J. J., Construction and application of an intelligent air quality monitor- ing system for healthcare environment. J. Med. Syst. 38(2):15, 2014. doi:10.1007/s10916-014-0015-3.

9. Yüregir, O. H., Oral, M., and Kalan, O., A decision support system for preventing Legionella disease. J. Med. Syst. 34(5):875–81, 2010. doi:10.1007/s10916-009-9302-9.

69 Page 4 of 4 J Med Syst (2015) 39: 69

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.