Benefits and qualities of groups in human services

RESEARCH ARTICLE

Utilizing Group-Based Contingencies to Increase Hand Washing in a Large Human Service Setting

Lynn G. Bowman1,2 & Samantha L. Hardesty1,3 & Sigurdur O. Sigurdsson3 & Melissa McIvor1 & Phillip M. Orchowitz1 & Leaora L. Wagner1 & Louis P. Hagopian1,2

Published online: 23 January 2019 # Association for Behavior Analysis International 2019

Abstract Hand washing is the most important preventative measure for the reduction of contagious disease. Although hand washing is easy to perform, non-adherence is a ubiquitous problem. Several studies have demonstrated the effectiveness of multi-component intervention packages to improve hand washing among employees; however, interventions are limited to acute settings, are often implemented for a short period of time, and rarely, if ever, include information on long-term effectiveness. The purpose of the current study was to utilize a behavior analytic approach to determine the stimulus conditions under which hand washing should occur, and to assess and then implement a long-term monitoring system among direct care workers in a large, non-acute inpatient unit. A single-case repeated measures reversal design was used to evaluate the effectiveness of two interventions aimed at improving hand washing adherence. A lottery was found to be effective in increasing hand hygiene for 2-years with 170 staff.

Keywords Hand washing . OBM . Lottery . Standard precautions . Stimulus control

In human service and health care settings, the importance of patient safety has received considerable attention, and one of the highest priority goals set forth by the World Health Organization (WHO) was to reduce hospital-acquired infections (Gould, Drey, Moralejo, Grimshaw, & Chudleigh, 2008). Hand washing is a relatively simple behavior to perform and is the single most important preventative measure for the reduction of contagious disease (Centers for Disease Control and Prevention [CDC], 2016). A meta-analysis of the effects of hand hygiene on infectious disease risk showed that improved hand-hygiene resulted in decreased rates of gastrointestinal and respiratory illnesses (Aiello, Coulbourn, Perez, & Larson, 2008). Despite its importance, health care workers often fail to practice good hand hygiene practices and adherence to best practice guidelines remain poor (e.g., Universal Precautions [UP] from

Occupational Safety and Health Administration [OSHA], 1999; Centers for Disease Control and Prevention, 2016; The Joint Commission [TJC], 2009). A review of 96 empirical arti- cles found the median adherence rate of healthcare workers to be 40%, with lower adherence rates associated with high activity and those observations with which a physician was involved (Erasmus et al., 2010). Given the importance of good hand hygiene practices, and the consensus across studies that adher- ence is poor, it is not surprising that hand hygiene continues to be a focus of extensive research.

Antecedents for Hand Washing

Several studies have examined specific situations, sometimes called indicators or critical antecedents that are likely to lead to better hand hygiene practices in the absence of intervention components. Across healthcare settings, antecedents identi- fied by the CDC and WHO include: 1) before patient contact, 2) before starting an invasive procedure, 3) after contact with blood, body fluids or excretions, mucous membranes, non- intact skin, and wound dressings, 4) after removing gloves, 5) when moving from a contaminated patient body site to a clean site during care, 6) after contact with inanimate objects

* Lynn G. Bowman [email protected]

1 Neurobehavioral Unit, Kennedy Krieger Institute, 707 N. Broadway, Baltimore, MD 21205, USA

2 Johns Hopkins University School of Medicine, Baltimore, MD, USA 3 University of Maryland, Baltimore County, Baltimore, MD, USA

Behavior Analysis in Practice (2019) 12:600–611 https://doi.org/10.1007/s40617-018-00328-z

or medical equipment close to the patient, and 7) after patient contact (TJC, 2009).

It is important to note that antecedents for hand hygiene are likely to differ from setting to setting and that staff adherence may differ across antecedents. This variation in setting and across antecedents suggests the need for an individualized approach to the measurement of hand washing. For example, Raboud et al. (2004) observed hand washing in a small group of nurses employed on a hospital unit and found that nurses were more likely to wash their hands following high risk situations (e.g., contact with bodily fluid or patient skin) than for other reasons (e.g., contact with equipment). Creedon (2005) found that hand washing practices were better prior to the initiation of an invasive procedure, and at the onset of patient care, while Mayer, Dubbert, Miller, Burkett, and Chapman (1986) found adherence following patient contact to be the highest. The identification of organizational-specific antecedents can lead to a more targeted intervention, which is ultimately more successful, efficient, and cost effective.

Although hand washing is an observable behavior amena- ble to the principles of applied behavior analysis, most re- search on this topic has been published in journals dedicated to medical and infection control practices (e.g., American Journal of Infection Control, Infection Control and Hospital Epidemiology, American Journal of Public Health, Epidemiology). Notable exceptions include a few small N studies published in behavioral science journals (e.g., Journal of Organizational Behavior Management, Journal of Applied Behavioral Analysis). There is some agreement with respect to the effectiveness of certain intervention com- ponents. Specifically, educational interventions were found to have a very short-term influence on hand hygiene behavior (Dubbert, Dolce, Richter, Miller, & Chapman, 1990), and the use of strategically placed reminders, or prompts from patients and staff, have had only a modest effect on hand hygiene adherence (Khatib, Ghassan, Abdallah, & Ibrahim, 1999). Arranging the environment to make hand washing easier (e.g., automated sinks, moisturized soaps) led to minimal im- provements (Larson et al., 1991). Feedback on performance was found to be successful at increasing hand washing, but if delivered intermittently the effect was not maintained over- time (Conly, Hill, Ross, Lertzman, & Louie, 1989; Mayer et al., 1986). Multi-component approaches which combined education with written materials, reminders, and continued feedback seemed to have the most impact on good hand hy- giene practices (Naikoba & Hayward, 2001).

Feedback is an effective and frequently used intervention in the field of Organizational Behavior Management (OBM) (VanStelle et al., 2012). With respect to increasing hand hy- giene behaviors, various types of feedback were found to be effective. DeVries, Burnette, and Redmon (1991) improved glove usage among nurses using bi-weekly feedback sessions consisting of goal setting and graphic feedback. Babcock,

Sulzer-Azaroff, and Sanderson (1992) improved the feedback of hand hygiene practices by supervisory nurses with training, weekly feedback meetings with goal setting, and letters of recognition. Stephens and Ludwig (2005) found training, set- ting group goals, and posting individualized-graphic feedback effective at improving nurses’ adherence to UP. Finally, Luke and Alavosius (2011) demonstrated the effectiveness of per- sonalized performance feedback consisting of a combination of verbal and written feedback across 3 health-care workers and improvement maintained for several months following the intervention.

Although performance feedback has been demonstrated to be effective, there are limitations to the long-term implemen- tation of such components, particularly for a behavior that needs to occur with such frequency as hand washing. For example, the training, time, and resources necessary to sustain individualized-level performance feedback systems are likely to be too cumbersome for most facilities to implement. For example, Luke and Alavosius (2011) estimated that it took approximately 250-man hours over 6 months to create mate- rials, complete observer training, conduct observations, and provide feedback for only 3 participants. Implementing a sim- ilar program across more individuals, or within in a large facility, may not be feasible.

Despite the effective strategies identified within these multi-component hand hygiene programs, several gaps con- tinue to exist within both behavioral and non-behavioral liter- ature. First, little is known regarding the maintenance or long- term effectiveness of multi-component, hospital-wide pro- grams as follow-up data are rarely, if ever published. Secondly, although frequent feedback (delivered either verbal- ly, graphically, or a combination of the two) has been found to be effective, little to no discussion has been made regarding the resources (time, personnel, or associated costs) needed to implement such a program, especially in large human service settings. Interventions that are equally or more effective as frequent feedback, as well as those that require fewer re- sources and maintain over time, are needed.

A lottery system can provide an organization with the abil- ity to provide reinforcers (monetary or non-monetary) on a lean schedule of reinforcement without deleterious effects on the intervention. Mayer et al. (2011) evaluated the use of a multi-component intervention that included 1) in-service pre- sentations, 2) one-on-one discussions with staff, 3) convenient positioning of soap dispensers and sanitizers, 4) the formation of a hand hygiene committee that monitored progress, and who generated catchy ideas for posters, jingles, and motiva- tors, and 5) delivery of motivators identified by a committee paired with feedback on hand hygiene practices. Specifically, feedback and reinforcement components were embedded within a series of programs in which employees, trained as hand hygiene monitors, “caught (hand washers) in the act” pg. 61 and immediately distributed incentives (i.e., chocolate bars,

Behav Analysis Practice (2019) 12:600–611 601

pizza party, and entry into monthly drawings) to those staff observed washing their hands following identified anteced- ents. The authors noted long-term and sustained improve- ments in hand washing practices following this program. However, a component analysis was not conducted to deter- mine which intervention(s) were responsible for hand washing improvement, and there was a lack of sufficient detail on how incentives were identified, how frequently incentives needed to be delivered to maintain effects, and whether or not hand washing occurred in the absence of monitoring. Although Mayer et al. (2011) did not describe the incentive component in these terms, entry into a drawing is analogous to a lottery. Although incentives and lottery systems are not frequently used in organizational settings (notable exceptions include Iwata, Bailey, Brown, Foshee, & Alpern, 1976; Luiselli et al., 2009 and Miller, Carlson, & Sigurdsson, 2014), they have been shown to be a low-cost option for motivating be- havior of individuals in large participant pools.

The purpose of the current study was to extend the research on hand hygiene practices by implementing a long-term mon- itoring system on hand washing adherence among direct care workers in a large, non-acute human service setting. Hand washing adherence was measured following antecedents outlined in hospital policies, and based on those results, a lottery system was implemented to increase and sustain hand washing for almost two years. In addition, the effects of ob- server presence and observer absence on hand washing adher- ence was measured.

Method

Participants and Setting

Participants included approximately 170 direct care staff (full- time and part-time) employed on a 16-bed inpatient unit with- in a rehabilitation hospital from September 2009 through August 2011. A minimum requirement for this position was a high-school diploma or equivalent (i.e., GED); though most staff had some college (some college or bachelor’s degree, 85%; working on graduate degree, 6%; unknown or no col- lege, 9%). The inpatient unit provided services to children and young adults diagnosed with an intellectual and developmen- tal disability (IDD) who exhibited severe behavioral prob- lems. Patients resided on the unit for approximately 4-5 months, and the unit functioned more like a school/home en- vironment than a typical hospital environment as the patients were not physically ill, but rather were admitted due to a severe behavior disorder. Patient behavioral intensity necessi- tated a minimum of one-to-one direct care staffing during all waking hours and staff worked day shifts (7:00 a.m. to 3:30 p.m.), evening shifts (3:00 p.m. to 11:30 p.m.), and night shifts (11:00 p.m. to 9:00 a.m.). Many other employees entered the

unit during the day as well (e.g., nursing staff, supervisors, behavioral treatment teams, faculty, physicians, etc.), and it was not uncommon for approximately 120-150 staff to visit the unit on any given day. However, data were only collected on hand washing adherence of the direct care staff.

The inpatient unit was approximately 6,000 square feet (557.42 m2). It consisted of one large open room and two smaller, classroom-sized rooms. These common areas were connected by a long hallway that was lined with 4 bedrooms and 4 bathrooms. Unit supervisors had an office at the main entrance of the unit. Direct care staff and patient dyads were not permitted in bedrooms outside patient scheduled sleep times, unless otherwise indicated in the behavioral plan (i.e., behavioral safety or medical need for isolation). In the large common area there were enough couches, tables, and chairs to accommodate most staff and patients. To facilitate hand wash- ing, 10 sanitizer dispensers and 9 sinks with soap dispensers were on the unit and were predominately located in the hall- way and the common areas.

Procedure

A single-case repeated measures reversal design (ABCAC) was used to evaluate the effectiveness of two interventions aimed at improving hand washing adherence. Prior to baseline and throughout the study, a traditional hospital-wide hand hygiene campaign was in place. This program consisted of annual competencies related to good hand hygiene practices and posters reminding staff to “wash your hands” strategically placed in the hospital (e.g., in bathrooms, on elevators). During staff orientation on the inpatient unit, staff received additional training on the importance of hand hygiene, ante- cedents when hand washing was necessary, and staff respon- sible for training new employees modeled appropriate hand hygiene behavior.

Dependent Variable and Measurement

Data were collected by trained observers on staff adherence with hand hygiene. Hand hygiene adherence was defined as: 1) using sanitizer (wall-mounted or portable), or 2) using soap and water to wash hands. Five antecedents of hand washing were identified based on hospital infection control policy prior to the start of data collection, and hand washing adherence was only monitored and documented following one of the identified antecedents. Critical antecedents included: 1) enter- ing the unit, 2) exiting the unit, 3) patient-to-patient contact (before making contact with a new patient if contact was pre- viously made with another patient), 4) after taking a patient to the bathroom, and 5) before providing a patient with his or her meal. During some antecedents (e.g., taking a patient to the bathroom) staff was required to wear gloves; however, they

602 Behav Analysis Practice (2019) 12:600–611

were still required to wash their hands following glove removal.

Codes were created on the data sheet to indicate the type of hand washing observed (sanitizer, soap and water, portable sanitizer, or none), and which of the 5 critical antecedents was observed. If hand washing occurred within 3-5 s of one of the antecedents, or before another antecedent occurred (whichever came first), the staff person was scored as adhering to hygiene policy (and specific type recorded). If hand wash- ing did not occur, “none” was recorded. Given the large num- ber of direct care staff that could potentially be observed dur- ing an observation period, data were aggregated across obser- vations of employees, and the percentage of adherence was calculated by dividing the number of observations where ad- herence was observed by the total number of observations during each observation period. In addition, covert data col- lectors recorded whether a supervisor associated with the lot- tery program was within eyesight of an employee who en- gaged in hand washing (supervisor present) or not within eye- sight (supervisor absent).

Data were collected during planned observation periods conducted at variable times and locations on the unit between the hours of 7:30 a.m. and 10:30 p.m. Observation periods ranged from approximately 15 min in duration to a maximum of one hour. Data were only included if the data collector(s) observed five or more critical antecedents within an observa- tion period. On average, 10.22 critical antecedents were scored (range 5–37 critical antecedents) per observation period.

Observation Procedure Throughout the study, overt and co- vert data monitoring was conducted on hand washing adher- ence. During overt observation periods, data were collected by one (or two when reliability data were collected) of three trained observers, all of whom were supervisors of the direct care staff. During these observations, the supervisors walked around the hospital unit holding clipboards while collecting data on direct care staff’s hand hygiene. During covert obser- vation periods, data were collected by one (or two when reli- ability data were collected) of four trained observers who were not supervisory staff. Covert data collectors discretely collect- ed data while they were seemingly performing an assigned task on the unit (e.g., interacting with a patient, replacing unit materials, reviewing protocol changes, cleaning, and helping with patients). Probably due to the large number of employees on the unit at any given time, there was no indication that direct care staff were aware that their hand hygiene behavior was being observed covertly. During covert monitoring obser- vations, supervisors who conducted overt data collection were purposefully not present on the unit.

Because of the physical layout of the unit (a straight, wide hallway lined with sinks and sanitizers connecting three large common areas), all data collectors (overt and covert) were

easily able to view each other as well as observe the hand washing behavior of multiple staff, often simultaneously.

Reliability Observations For both overt and covert observa- tions, two observers made independent and simultaneous re- cordings of hand washing across critical antecedent condi- tions. To ensure data were collected on the same employee, observers discretely signaled to each other (e.g., made eye contact, nodded, or gestured) to indicate the onset of an ob- servation. These observations were compared to establish ob- servation reliability. Agreement was scored if both observers denoted the: 1) same type of adherence or “none”, 2) the same critical antecedent observed, 3) for the same individual and the same observation time, 4) and whether the supervisor was present or absent. A percentage was then calculated by divid- ing the total number of observations in which agreement was achieved by the total number of observations. Due to sched- uling conflicts among the overt data collectors and many ob- servations resulting in no data recorded (given five or more critical antecedents were not observed), total reliability was assessed for only 19% of overt observations and 26% of co- vert observations. Agreement averaged 93% for hand hygiene adherence (across all types of antecedents) during overt ob- servations and 99% during covert observations.

Critical Antecedent Analysis It was hypothesized by the au- thors that hand washing was likely to occur more often fol- lowing some critical antecedents (e.g., after assisting a patient in the bathroom) than following others (e.g., entering the unit). Therefore, an analysis was conducted to measure adherence following each critical antecedent. In total, 280 observations were conducted across all antecedent conditions. All critical antecedent analysis observations were conducted by the overt data collectors. The purpose of this analysis was to inform intervention planning. Standard hospital and unit practices were in place during this phase as well as all subsequent phases. In addition, prior to collecting data on hand washing, a memo was read at all shift changes to direct care staff reminding them of the importance of hand washing as well as describing when staff should wash their hands.

Baseline Data obtained during the critical antecedent analysis served as the initial baseline observations. During a return to baseline phase, all direct care employees were informed by memo that the lottery was no longer in effect. Staff was con- gratulated for their improved hand hygiene adherence follow- ing the critical antecedents. In addition, staff was encouraged to keep up the good work and reminded of the critical ante- cedents in which hand washing should occur.

Antecedent-Based Intervention The antecedent-based inter- vention involved the provision of portable hand sanitizer bot- tles to direct care staff. Bottles were distributed to staff during

Behav Analysis Practice (2019) 12:600–611 603

randomly assigned shifts. The purpose of this arrangement was to compare hand washing adherence across shifts with and without hand sanitizer bottles. This intervention was cho- sen based on discussions with supervisory staff and anecdotal observations made during the critical antecedent analysis. It was hypothesized that staff was engaged in several competing responsibilities (e.g., maintaining the safety of a difficult pa- tient, holding data sheets and patient materials, etc.) that might make walking to a sink or wall sanitizer cumbersome. Portable bottles were selected in an effort to: 1) increase ac- cessibility of a cleaning agent, 2) reduce response effort asso- ciated with hand washing, and 3) help prompt staff to wash their hands. The sanitizer bottles were 2 oz. (59.18 ml) and clipped onto break-away lanyards that staff already wore. Covert data collection began during this phase.

Group-Based Lottery Due to the large size of the participant group, it was not feasible or economical to implement an individualized reinforcement schedule for each staff member. As an alternative, a lottery system was devised. To determine lottery prizes, a survey was administered to all direct care staff. On the survey, three monetary prizes were listed including $25 gift cards to local restaurants and stores (e.g., coffee, depart- ment stores), as well as three non-monetary prizes (e.g., selecting a patient to work with for multiple shifts, choosing break times, and being drawn to go home). Staff then ranked prize preference for all six items. Results indicated the most preferred item was a $25 gift card to popular department store©. Two days prior to starting the lottery, a memo was distributed to all staff describing the details of the lottery, a reminder of the critical antecedents, what constituted adher- ence with hand hygiene (i.e., using soap and water, or sanitizers), and the rules associated with administration of the lottery prize. Specifically, staff members were told 1) they would be entered into the lottery drawing each time they were observed washing their hands following a critical antecedent, 2) they could be entered into the lottery multiple times during an observation, and 3) lottery drawings were to be conducted on pre-determined, randomly selected shifts. Participants were only entered into the lottery during overt observations.

During the initial lottery phase, drawings were randomly scheduled to occur across the day shift, evening shift, and night shift approximately every two monitoring periods (i.e., approximately once per week). Numbers were generated for all lottery participants eligible for the prize, and one number was randomly selected by a supervisor otherwise not involved in the study. To ensure hand washing continued for the re- mainder of the shift, drawings were conducted no more than 1 h prior to the end of the shift. One of the three overt data collectors was responsible for announcing the winner (in the presence of other co-workers), providing verbal praise to the winner, and delivering the gift card. Throughout the lottery phase, the memo detailing the procedures was read before

each shift. Current winners (those that won within the last two weeks) were announced and congratulated during shift changes, and names of recent winners were also posted on the main hospital unit.

Following a return to baseline and while a steady trend of hand washing was observed, the number of times the lottery distribution schedule was reduced to approximately every third monitoring period (i.e., approximately once every 2 weeks) to minimize costs.. All other procedures remained the same. Overt and covert hand hygiene observations contin- ued for 10 months. Overt observation data suggested that ad- herence improved and that improvements were observed fol- lowing each of the five critical antecedents. However, covert observation data suggested that adherence had not improved beyond baseline levels.

Generalization of Lottery To increase hand hygiene adherence when overt data collectors were not present, 13-unit supervi- sors who were routinely present on the unit across all 3 shifts were recruited to hand out prizes for hand washing. During this phase, data continued to be collected by the overt and covert data collectors (to ensure integrity and to minimize added duties); however, each of the 13 supervisors was sys- tematically selected to participate in the drawing and trained by one of the overt data collectors to provide verbal praise in front of co-workers, and deliver gift cards to winners in the same way the overt data collectors delivered gift cards.

Social Validity and Infection Control Data

A short treatment acceptability questionnaire was adminis- tered to direct care staff following the antecedent-based inter- vention assessing their acceptance of the portable hand sanitizers. Within this survey, staff’s knowledge of critical antecedents was also examined, and suggestions for additional intervention strategies were requested (see Appendix).

Concurrent to this study, data were collected quarterly across the entire hospital (including the inpatient unit) on staff hand hygiene behavior by “secret shoppers” who served on the hospital's Infection Control Advisory Board. It is important to note that these personnel had no knowledge of the current study, and the authors had no knowledge of the identity of the secret shoppers. Data were summarized each quarter for the respective units and were distributed via e-mail to administrators. When indicated, administrators were urged to address adherence issues with their staff; however, no specific guidance was provided on how to improve hand washing behavior.

Results

Figure 1 shows the percentage of opportunities with hand hygiene adherence during observation periods across baseline,

604 Behav Analysis Practice (2019) 12:600–611

the antecedent-based intervention, and lottery phases. During baseline, hand washing was low across all critical antecedents (see Fig. 2), making it necessary to target all 5 antecedents. Initiation of the antecedent-based intervention (portable hand sanitizers) did not improve adherence. Hand washing in- creased and remained high with the initiation of the group- based lottery (consequent intervention) and remained high during schedule thinning and generalization of lottery phases.

Baseline During the initial baseline phase, hand washing ad- herence averaged 11% (range 0% to 25%). When the lottery was removed during the return to baseline phase, levels of hand washing adherence immediately returned to low levels (M =15%, range 0% to 29%).

Antecedent-Based Intervention Hand washing adherence remained low during shifts in which portable hand sanitizers were distributed to staff (M= 17%, range 0% to 29%).

Group-Based Lottery Following the initiation of the lottery (consequent intervention), hand washing adherence imme- diately increased across all situations and remained high (M = 63%, range 50% to 80%) throughout the initial phase in which gift cards were distributed following ap- proximately every 2 observations. Following the return to baseline phase, the lottery was reinstated and hand wash- ing adherence again increased (M =76%, range 50% to 100%). Hand washing adherence remained high, even

when the density of reinforcement was thinned from fol- lowing approximately every 2 observations to following approximately every 3 observations (M =72%, range 38% to 100%).

Intervention Results Figure 3 depicts the averaged percentage of opportunities with hand hygiene adherence across baseline, the antecedent-based intervention, the overt observations from the group-based lottery, and the covert observations (supervi- sor absent and supervisor present). A total of 3097 observa- tions of hand washing were observed throughout the study (2332 overt and 765 covert observations). Data were aggre- gated within each phase.

Figure 4 isolates the covert data obtained throughout all phases of the study. Although data obtained from overt obser- vations suggested that hand washing improved, data obtained from covert observations did not show improvement, suggest- ing that participants’ hand washing was impacted by their awareness of being observed by individuals associated with the lottery (supervisor-absent observations, M =16%, range 0% to 60%). When the intervention was generalized to trans- fer stimulus control of hand washing compliance to the larger group of supervisors, and they became responsible for the distribution of gift cards, data obtained from covert observa- tions showed improvement over time but only when the su- pervisor(s) were visible to staff (supervisor-present observa- tions, M =48%, range 0% to 100%). The total cost in gift cards over the approximately 2-year study was 1,000.00 dollars.

0

10

20

30

40

50

60

70

80

90

100

Observation Periods

Antecdent Manipulation

Memo

Lottery Baseline

Sanitizer

bottles No bottles

Baseline Lottery Lottery: Schedule Thinning

P e r c e n

ta g e o

f O

p p

o r tu

n it

ie s

W it

h H

a n

d H

y g ie

n e A

d h

e r e n

c e

Fig. 1 Percentage of opportunities with hand hygiene adherence during baseline phases, the antecedent intervention, and lottery phases

Behav Analysis Practice (2019) 12:600–611 605

Intervention Acceptability Questionnaire (for Portable Hand Sanitizers) Although the return rate on the treatment accept- ability questionnaire was low (n= 30), staff who responded overwhelming did not like the portable sanitizers, indicating

the bottles were difficult to carry due to size, and described the hand sanitizers as “heavy”, “bulky”, and “a hassle” (Appendix). In general, staff was able to list the critical ante- cedents, particularly those that included higher risk situations

0

10

20

30

40

50

60

70

80

90

100

Entering Unit Exiting Unit After Toileting

Patient

Before Feeding

Patient

Following Patient

Contact

Critical Antecedent

A v e r a g e P

e r c e n

ta g e H

a n

d H

y g ie

n e A

d h

e r e n

c e

Fig. 2 Averaged percentage adherence with hand hygiene during critical antecedents

0

10

20

30

40

50

60

70

80

90

100

Baseline Antecedent-Based Intervention Group-Based Lottery Covert (Supervisor Absent) Covert (Supervisor Present)

ecnerehd A

eneigy H

dna H

hti Wseitinutropp

Ofo egatnecreP

egarev A

Phases of Study

Fig. 3 Averaged percent of observations with hand hygiene adherence during baseline phases, the antecedent-based intervention, the overt group- based lottery observations, and the covert observations (supervisor absent and supervisor present)

606 Behav Analysis Practice (2019) 12:600–611

(i.e., after toileting a patient, 100% of staff who responded; prior to feeding a patient, 86% of staff who responded). Antecedents that one might consider less risky were listed less frequently (i.e., walking on and off the unit, 13% of staff who responded; following patient contact, 50% of staff who responded). Only a few suggestions were made to influence staff hand washing such as offering hand moisturizer and re- quiring that staff collect data on their own hand washing behavior.

Statistical Analysis

A chi-square goodness of fit indicated that there was no dif- ference between hand washing adherence in baseline phases 1 and 2; χ2 (N=525, 2) = 1.405, p= 0.495. There was also no difference between baseline observations and observations made during the antecedent manipulation χ2 (N=496, 2) = 2.627, p= 0.269. A significant difference was observed in hand washing adherence between baseline observations and the lottery intervention χ2 (N=2169, 1) = 562.368, p<0.001.With respect to covert and overt data, a difference was also observed during the lottery intervention χ2 (N=1569, 1) = 339.099, p<0.001 suggesting that the intervention was only effective when individuals associated with data collec- tion and the lottery were present to observe hand hygiene. When the intervention components were generalized to all supervisory staff, a difference was observed in hand washing

adherence when comparing conditions when supervisors were present to conditions when they were absent χ2 (N=519, 1) = 69.376, p<0.001.

Infection Control Data

Prior to the lottery intervention, hand washing adherence was historically lower on the inpatient unit than the overall hospital average (Fig. 5). Data obtained quarterly by “secret shoppers” suggested that hand washing improved during the lottery in- tervention on the unit. As part of standard practices, the hos- pital set a goal of 100% adherence in hand washing, with a minimum acceptable goal of 90%. The inpatient unit met these standards during the lottery phase and failed to do so when the lottery was not in place.

Discussion

This study demonstrates long-term maintenance of hand hy- giene behavior among direct care staff in a non-acute hospital setting. A bi-weekly lottery was effective at immediately in- creasing the hand hygiene behavior of approximately 170 di- rect care staff, and that increase maintained for approximately 2 years. A preference survey was conducted prior to the onset of the lottery and a $25 gift card from popular department store © combined with praise was delivered to one employee during each drawing. The effectiveness of the lottery was

A n te

c e d e n t

M a n ip

u la

ti o n P

h a s e

L o

tt e

r y P

h a

s e

B a

s e

li n

e P

h a

s e

L o

tt e

r y P

h a

s e

0

10

20

30

40

50

60

70

80

90

100

1 1 1

2 1

3 1

4 1

5 1

6 1

7 1

8 1

9 1

1 01

1 11

1 21

1 31

1 41

1 51

1 61

1 71

1 81

1 91

2 01

2 11

2 21

2 31

2 41

2 51

2 61

2 71

2 81

2 91

3 01

3 11

3 21

3 31

3 41

3 51

3 61

3 71

3 81

3 91

4 01

4 11

Observation Periods

Generalization of Lottery

C ov

er t

(S up

er vi

so r

A b

se n

t)

C ov

er t

(S u

p er

vi so

r P

re se

nt )

Covert Observations During All Phases of the Study

P er

ce n

ta ge

o f

O p

p or

tu n

it ie

s W

it h

H an

d H

yg ie

n e

A d

h er

en ce

Fig. 4 Covert data collected on the percentage of opportunities with hand hygiene adherence during baseline, antecedent intervention, and lottery interventions while a supervisor was absent or was present

Behav Analysis Practice (2019) 12:600–611 607

robust with little associated costs (1,000.00 dollars spent dur- ing lottery implementation for incentives). The cost of just one in-service for staff training would have been more than the total cost of the lottery. Additionally, data collected from “se- cret shoppers” further validate the findings of the study.

Adherence to good hand hygiene practices is of utmost importance and this study offers a thorough examination of hand hygiene in a human service setting not typically studied in the literature. Even though hand washing is an easy behav- ior to perform, non-adherence with hand hygiene is a ubiqui- tous problem (~40% rates) even in intensive care environ- ments with nurses and physicians (Erasmus et al., 2010). Numerous studies have been published in medical and infec- tion control journals, and some results of the current study match results obtained from prior research. For example, stra- tegically placed reminders (signs to wash hands) and educa- tional interventions (staff in-services and trainings memos) were present prior to and during the study and appeared to have no impact on staff hand hygiene. Review articles sum- marizing the effectiveness of various hospital-wide programs aimed at improving hand hygiene adherence (Naikoba & Hayward, 2001; Gould et al., 2008; Erasmus et al., 2010) agree that in general, studies lack sufficient detail regarding intervention content (Gould et al., 2008) and the interventions did not result in robust or long-term effects (Naikoba & Hayward, 2001). Hand washing adherence was high and sta- ble when individuals associated with the lottery were present but was low when covert data collection occurred during times when staff associated with the lottery were not on the unit. Given that participants needed to be seen washing their hands and have their name entered into the lottery to win, it is likely that the overt data collectors in this study served as discrimi- native stimuli for reinforcement, signaling the opportunity to be entered in, and potentially win, the lottery. In the absence of

the overt data collectors or the supervisors associated with the lottery, staff was less inclined to wash their hands.

While direct observation is considered best practice by TJC and the CDC, clearly other strategies that do not require the presence of a known observer are needed. In our study, asso- ciating the supervisors with the lottery system improved hand washing even during covert data collection. Other, less staff- intensive, data collection methods might include the use of high-tech hand hygiene monitoring systems that notify and detect hand washing via sensors in staff badges. However, it is unclear if these expensive systems are feasible in residential, school, clinics, or treatment units such as ours where anteced- ents for hand hygiene may not be as salient as those in acute care settings (i.e., entering an intensive care patient’s bed- room). More frequent use of covert or “secret data” collectors who collect data either live or via video monitoring coupled with the delivery of consequences (e.g., feedback, incentives, or both) may be necessary. Alternatively, permanent products of hand washing (i.e., measurement of sanitizer or soap use, use of black light on staff hands) may be another viable option in which to deliver consequences and are worthy of further examination.

This study adds to the current literature by providing a unique analysis of hand washing behavior across important antecedents of appropriate hand hygiene prior to intervening. Utilization of performance diagnostic tools such as identifying organizational/individualized critical antecedents is important because it allows for the identification of the most appropriate intervention(s) which will hopefully lead to more effective and efficient recommendations. The critical antecedent analy- sis in this setting found that hand hygiene adherence was poor across all antecedents. Based on the results of the pre- intervention assessment, the first intervention included the use of an antecedent modification (i.e., increasing accessibility

0

10

20

30

40

50

60

70

80

90

100

1 2 3 4 5 6 7 8 9 10 ecnerehd

A eneigy

H dna

H hti

WffatSfo egatnecreP

egarev A

Quarterly Infection Control Observations

Baseline Lottery Baseline Lottery Maintenance Fig. 5 Averaged percent of staff

with hand hygiene adherence for the inpatient unit obtained by “secret shoppers” from the hospital’s Infection Control Advisory Board

608 Behav Analysis Practice (2019) 12:600–611

to hand washing by providing portable sanitizers to staff). However, this intervention was found to be ineffective and staff did not prefer it. Therefore, the antecedent intervention was replaced with a consequence-based intervention which included a group-based lottery.

Although this study demonstrated that an innovative group-based reinforcement program was effective at improv- ing and sustaining hand washing adherence across a large number of direct care staff, there are some other limitations worthy of discussion. First, the lottery phase included two components: task clarification (i.e., memo reminding staff and outlining antecedents) and the consequence for adherence with hand washing (i.e., lottery). Since a component analysis was not conducted, it is impossible to say which aspect of that intervention led to the change in hand washing behavior. However, a similar memo reminding staff to practice good hand hygiene and the critical antecedents to handwashing was read to staff prior to the baseline phase, and hand washing was low. An additional limitation of the study is that treatment integrity data was not obtained on the delivery of the lottery when it was generalized to supervisors. While the authors of this study were responsible for identifying the winners, writ- ing and publishing the memo, the supervisors were responsi- ble for delivering the gift cards. Winners signed a form and indicated the gift card number and the date they received the gift; card however, reliability data denoting that cards were delivered to winners in the same manner as the covert data collectors would have added strength to the study.

The CDC and WHO (CDC, 2016; WHO, 2009) outline specific components of hand washing that should be complet- ed to best prevent the spread of bacteria and viruses (e.g., using soap, vigorously scrubbing hands for a minimum of 20 s, etc.). Due to the large number of staff observed and the way in which data collection occurred, it was also not feasible to parse out these components or to measure the duration of hand washing. In this study, adherence was scored when staff were observed to place soap and water or sanitizer on their hands following a critical antecedent; we did not measure duration or other characteristics of appropriate hand washing. Another limitation to this study is that other performance di- agnostic tools more recently used in OBM (e.g., Performance Diagnostic Checklist-Human Service; Carr, Wilder, Majdalany, Mathisen, & Strain, 2013) were not utilized prior to implementing the interventions.

Another limitation to this study was that, the personnel and resources necessary to collect the data were immense. Not only were data collectors required to spend time directly ob- serving the hand washing behavior of employees, but to en- sure the correct names were entered into the lottery, data col- lectors also needed to be knowledgeable of all staff names. Since data collectors were also employees of the inpatient unit, in almost all cases, the names of staff were known. However, in some cases, data collectors needed to ask for

names of hand washers (overt only) or confirm names through the electronic patient assignment document which outlined patient-staff assignments for that shift (covert only). Although the open layout of this inpatient unit and our setting allowed for direct observation of hand washing among all staff, such an intensive data collection and observation system may not be plausible in every human service setting (i.e., those with more private/divided living areas). Third, the current study addressed only one antecedent intervention (the portable hand sanitizers). Perhaps other antecedent interventions re- ported in the literature may have resulted in similar results as the lottery system.

While hand hygiene is an important behavior for all human service settings, it is rarely studied with vigor in settings more analogous to ours (i.e., school-based, residential programs, or child-care facilities) as it is studied with acute inpatient hospi- tal units. On our unit, patients were admitted for a longer duration of time than in most hospital units, on average 4-5 months, and to ensure comfort during that time every effort was made to emulate a more home and school-like environ- ment. As such, critical antecedents to hand washing may not have been as obvious to staff (i.e., staff were not treating wounds or entering isolation areas, and the patients were not sick). Unlike intensive care units or standard hospital units, patients (and staff) on this inpatient unit spent the clear major- ity of their time in shared/common areas and frequently interacted with one another. That is, although staff was assigned to one patient per shift, staff was encouraged to in- termingle and engage with other patients and staff throughout the shift, particularly during individual and group-led activi- ties (i.e., circle time, dance parties, academics). Furthermore, during every shift, several staff (typically 4) was assigned as rotational staff. These individuals were responsible for cover- ing breaks and assisting other patient/staff dyads when indi- cated, which required him/her to switch frequently from one patient to another.

Competing job duties of the participants in the current study are also likely to be more intense than in other hospital related positions. Competing duties that might interfere with hand washing within this unit included anything from staff needing to protect themselves and the patient from severe problem behavior, to intensive data collection requiring staff to document frequency counts of 6-8 different topographies of problem behavior every half hour, to needing to transport numerous materials (e.g., treatment items, data sheets, patient information binder, protective equipment) from one location to another while simultaneously caring for an ambulatory pa- tient. Given that the consequences associated with failure to complete many of these competing activities are likely to be more immediate and severe (i.e., patient or staff gets hurt, punitive feedback for incomplete data) than those conse- quences loosely associated with hand washing, it is not sur- prising that staff would allocate responding to those duties over

Behav Analysis Practice (2019) 12:600–611 609

adherence to good hand washing practices. Future research may include evaluating different antecedent manipulations (e.g., hand sanitizer on key chains rather than on a lanyard may be more acceptable to staff) and replicating these findings in set- tings that are more representative of a typical hospital given the importance of hand hygiene in acute health care settings.

Acknowledgments Special thanks to the many data collectors over the years:

Alex Arevalo, Anna Armstrong, Thomas Banz, Mariana Castillo, Monica Lugo, Shari Pincus, Rebecca Stern, Monica Urich

Compliance with Ethical Standards

Conflict of Interest All authors declare that he/she has no conflict of interest.

Ethical Approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institu- tional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed Consent Informed consent was not necessary as the project served as a quality improvement project for staff employed in the hospital.

1. Do you like the portable bottles of sanitizer?

YES NO N/A (have not had)

2. Please indicate why you DO or DO NOT like the sanitizers.

________________________________________________

________________________________________________

________________________________________________

3. Would you like to have a portable sanitizer with you during shifts?

YES NO N/A (have not had)

4. Did the portable sanitizers help remind you to wash your hands?

YES NO N/A (have not had)

5. Please list situations in which you think you should wash your hands.

_______________ _____________ ___________

_______________ _____________ ____________

_______________ _____________ ____________

_______________ _____________ ____________

6. How often (# times) per shift would you say that you wash your hands? _______________________________________________________

7. Do you have any ideas of things we could do or materials we could have that

would help you to wash your hands? ______________

______________________________________________________

______________________________________________________

______________________________________________________

Appendix

Staff survey about hand washing practices and social va- lidity of portable sanitizers

610 Behav Analysis Practice (2019) 12:600–611

Publisher’s Note Springer Nature remains neutral with regard to jurisdic- tional claims in published maps and institutional affiliations.

References

Aiello, A., Coulbourn, R., Perez, V., & Larson, E. (2008). Effect of hand hygiene on infectious disease risk in the community setting: A meta- analysis. American Journal of Public Health, 98, 1372–1381.

Babcock, R. A., Sulzer-Azaroff, B., & Sanderson, M. (1992). Increasing nurses’ use of feedback to promote infection-control practices in a head-injury treatment center. Journal of Applied Behavior Analysis, 25, 621–627. https://doi.org/10.1901/jaba.1992.25-621.

Carr, J., Wilder, D., Majdalany, L., Mathisen, D., & Strain, L. (2013). An assessment-based solution to a human-service employee perfor- mance problem: An initial evaluation of the Performance Diagnostic Checklist-Human Services. Behavior Analysis in Practice, 6, 16–32. https://doi.org/10.1007/BF03391789.

Centers for Disease Control and Prevention (2016). Hand hygiene in healthcare settings.

Conly, J., Hill, S., Ross, J., Lertzman, J., & Louie, T. (1989). Handwashing practices in an intensive care unit: The effects of an educational program and its relationship to infection rates. American Journal of Infection Control, 17, 330–339. https://doi.org/10.1016/ 0196-6553(89)90002-3.

Creedon, S. (2005). Healthcare workers’ hand decontamination practices: Compliance with recommended guidelines. Issues and Innovations in Nursing Practice, 51, 208–216. https://doi.org/10.1111/j.1365- 2648.2005.03490.x.

DeVries, J., Burnette, M., & Redmon, W. (1991). AIDS prevention: Improving nurses’ compliance with glove wearing through perfor- mance feedback. Journal of Applied Behavior Analysis, 24, 705– 711. https://doi.org/10.1901/jaba.1991.24-705.

Dubbert, P., Dolce, J., Richter, W., Miller, M., & Chapman, S. (1990). Increasing ICU staff hand washing: Effects of education and group feedback. Infection Control and Hospital Epidemiology, 11, 191– 193. https://doi.org/10.1177/1054773805282445.

Erasmus, V., Daha, T., Brug, H., Richardus, J., Behrendt, M., Vos, M., & Beeck, E. (2010). Systematic review of studies on compliance with hand hygiene guidelines in hospital care. Infection Control and Hospital Epidemiology, 31, 283–294. https://doi.org/10.1086/ 650451.

Gould, D., Drey, N., Moralejo, D., Grimshaw, J., & Chudleigh, J. (2008). Interventions to improve hand hygiene compliance in patient care. Journal of Hospital Infection, 68, 193–202. https://doi.org/10.1016/ j.jhin.2007.11.013.

Iwata, B. A., Bailey, J. S., Brown, K. M., Foshee, T. J., & Alpern, M. (1976). A performance-based lottery to improve residential care and training by institutional staff. Journal of Applied Behavior Analysis, 9, 417–431. https://doi.org/10.1901/jaba.1976.9-417.

Khatib, M., Ghassan, J., Abdallah, A., & Ibrahim, Y. (1999). Hand wash- ing and use of gloves while managing patients receiving mechanical ventilation in the ICU. Chest, 116, 172–175. https://doi.org/10.1378/ chest.116.1.172.

Larson, E., McGreer, A., Quraishi, Z., Krenzischek, D., Parsons, B., Holdford, J., & Hierholzer, W. (1991). Effect of an automated sink on hand washing practices and attitudes in high risk units. Infection Control and Hospital Epidemiology, 12, 422–428. https://doi.org/ 10.2307/30148304

Luiselli, J., Reed, F., Christian, W., Markowski, W., Rue, H., St. Amand, C., & Ryan, C. (2009). Effects of an informational brochure, lottery- based financial incentive, and public posting on absenteeism of direct-care human service employees. Behavior Modification, 33, 175–181. https://doi.org/10.1177/0145445508320624.

Luke, M., & Alvosius, M. (2011). Adherence with universal precautions after immediate personalized performance feedback. Journal of Applied Behavior Analysis, 44, 967–971.

Mayer, J. A., Dubbert, P. M., Miller, M., Burkett, P. A., & Chapman, S. W. (1986). Increasing handwashing in an intensive care unit. Infection Control, 7, 259–262. https://doi.org/10.1017/ S0195941700064171.

Mayer, J. A., Mooney, B., Gundlapalli, A., Harbarth, S., Stoddard, G., Rubin, M., Eutropius, L., Brinton, B., & Samore, M. (2011). Dissemination and sustainability of a hospital-wide hand hygiene program emphasizing positive reinforcement. Infection Control and Hospital Epidemiology, 32, 59–66. https://doi.org/10.1086/ 657666.

Miller, M., Carlson, J., & Sigurdsson, S. (2014). Improving treatment integrity in a human service setting using lottery-based incentives. Journal of Organizational Behavior Management, 34, 29–38.

Naikoba, S., & Hayward, A. (2001). The effectiveness of interventions aimed at increasing hand washing in healthcare workers: A system- atic review. Journal of Hospital Infection, 47, 173–180. https://doi. org/10.1056/jhin.2000.0882.

Occupational Safety and Health Administration (1999). Occupational exposure to bloodborne pathogens. (Publication No. 29, Code of Federal Regulations Part 1910). Washington: Office of the Federal Register, National Archives and Records Administration.

Raboud, J., Saskin, R., Wong, K., Moore, C., Parucha, G., Bennet, J., Green, K., Low, D., Loeb, M., Simor, A., & McGeer, A. (2004). Patterns of handwashing behavior and visits to patients on a general medical ward of healthcare workers. Infection Control and Hospital Epidemiology, 25, 198–202. https://doi.org/10.1086/502377.

Stephens, S., & Ludwig, T. (2005). Improving anesthesia nurse compli- ance with universal precautions using group goals and public feed- back. Journal of Organizational Behavior Management, 25, 37–71. https://doi.org/10.1300/J075v2502-02.

The Joint Commission (TJC, 2009). Measuring hand hygiene adherence: Overcoming the challenges. Monograph produced as part of con- sensus Management in Hand Hygiene (CMHH) project. Printed in USA.

VanStelle, S., Vicars, S., Harr, V., Miguel, C., Koerber, J., Kazbour, R., & Austin, J. (2012). The publication history of. Journal of Organizational Behavior Management, 32, 93–123. https://doi. org/10.1080/01608061.2012.675864.

World Health Organization. (2009). WHO guidelines on hand hygiene in health care: First global patient safety challenge: Clean care is safer care. Geneva: World Health Organization, Patient Safety.

Behav Analysis Practice (2019) 12:600–611 611