protective equipment: continuous and contingent application ...

JOURNAL OF APPLIED BEHAVIOR ANALYSIS

PROTECTIVE EQUIPMENT: CONTINUOUS AND CONTINGENTAPPLICATION IN THE TREATMENT OF SELF-INJURIOUS

BEHAVIOR

MICHAEL F. DORSEY, BRIAN A. IWATA,DENNIS H. REID, AND PATRICIA A. DAVIS

TEMPLE UNIVERSITY-WOODHAVEN PROGRAM, THE JOHN F. KENNEDYINSTITUTE AND THE JOHNS HOPKINS UNIVERSITY SCHOOL OF MEDICINE,

AND NORTHERN INDIANA STATE HOSPITAL AND DEVELOPMENTALDISABILITIES CENTER

This study evaluated the use of protective equipment in treating self-injurious behavior(SIB) exhibited by three retarded persons. In Experiment 1, the equipment was firstapplied continuously during 20-min sessions in individual multiple baseline designsacross settings. Results showed substantial reductions in head hitting, eye gouging, andhand biting. Brief periods of time-out with the protective equipment were later madecontingent on SIB and combined with a differential reinforcement procedure. Reducedlevels of SIB was maintained with all subjects. Additionally, the amount of time duringwhich the equipment was applied decreased as the SIB diminished. Experiment 2evaluated the use of contingent protective equipment (the final condition in Experi-ment 1) when applied directly in the subjects' living uints during the day. DuringExperiment 2, SIB remained at or below the levels found at the termination of Experi-ment 1. Finally, in an effort to assess the long-term effectiveness of the procedure,responsibility for implementation was given to the staff who were typically assigned toprovide therapy to the subjects. Follow-up probe observations conducted up to 104days after termination of the final experimental condition showed continued low levelsof both SIB and equipment usage. Results of these experiments suggest that contingentprotective equipment and differential reinforcement may be effective in reducing chronicself-injury.DESCRIPTORS: retardation, self-injurious behavior, protective equipment, restraint,

time-out

The reduction and maintenance of self-injuri-ous behavior present major problems, both in

This study was approved by both university andinstitutional human subjects committees, as well asthe administrative staff of the institution and localschool district. We thank Robert Crow for his con-tinued administrative support, Marilyn Bezeredy, SteveColeman, and Thomas Pritzel for their assistance inconducting the study, and Arthur Snapper, PaulWeiner, and Wayne Fuqua for their assistance in pre-paring an earlier dissertation version of the manu-script. Reprints may be obtained from either MichaelF. Dorsey, Temple University, Woodhaven Program,2900 Southampton Road, Philadelphia, Pennsylvania19154; or Brian A. Iwata, Division of BehavioralPsychology, The John F. Kennedy Institute, 707North Broadway, Baltimore, Maryland 21205.

terms of habilitative programming as well asprotecting individuals from self-inflicted harm.The use of mechanical restraint devices has be-come the most prevalent method of control andprotection of self-injurious persons clients(Schroeder, Note 1). Unfortunately, many formsof restraint have been noted to cause long-termnegative side effects such as demineralization ofbones, shortening of tendons, and arrested motordevelopment, secondary to disuse of the re-strained individual's limbs (Lovaas & Simmons,1969). The restraint of physical movement mayalso interfere with other client training goals(Rojahn, Schroeder, & Mulick, 1980). Problemssuch as these have become the basis for a num-

217

NUMBER 2 (SUMMER 1982)1982, 15. 217-230

MICHAEL F. DORSEY et al.

ber of legal and regulatory controls limiting theuse of restraint (Joint Commission on Accredi-tation of Hospitals, 1975).

Despite the attention paid to restraint devicesin applied settings, research regarding their usein programs designed to eliminate self-injury hasbeen relatively limited. A review of this litera-ture suggests a variety of therapeutic uses forrestraint, including applications as consequence(Favell, McGimsey, & Jones, 1978; Favell, Mc-Gimsey, Jones, & Cannon, 1981; Hamilton,Stephens, & Allen, 1967; Rapoff, Altman, &Christophersen, 1980) as well as antecedentevents (Parrish, Aguerrevere, Dorsey, & Iwata,1980; Rojahn, Mulick, McCoy, & Schroeder,1978), and both as punishing (Hamilton et al.,1967; Rapoff et al., 1980) as well as reinforcingstimuli (Favell et al., 1978, 1981). Hamiltonet al. (1967) reported the use of a time-out pro-cedure involving a padded restraint chair forperiods ranging from 30 min to 2 hrs, with fiveindividuals who engaged in self-injurious andaggressive behaviors. Rapoff et al. (1980) pre-sented a case study of a 7-yr-old blind child whoengaged in self hitting. A 30-sec period of physi-cal restraint was used successfully in the elimina-tion of her self-injury. Favell et al. (1978), onthe other hand, described the successful use ofrestraint, contingent upon the absence of self-injury, for three profoundly retarded persons.Because their results suggested that restraintfunctioned as a reinforcer, Favell et al. (1981)conducted a subsequent experiment in whichcontingent restraint was found to increase ap-propriate toy play.Two studies have examined the antecedent

effects of restraint upon self-injury. Rojahn et al.(1978) reported the successful use of adaptivejackets with large pockets, designed to allowappropriate "self-restraint," in the reduction ofself-injury. Parrish et al. (1980) investigated theeffects of a padded football helmet on rates ofhead hitting in a profoundly retarded client.During a series of brief reversal periods, an aver-age of 60 responses per min occurred during

non-helmet periods as compared with 5 re-sponses per min during periods with the helmeton.

The results of the Parrish et al. (1980) studylend some support to the position that manyforms of behavior, which appear to provide noexternal consequence, may be maintained bytheir sensory-stimulating properties (Carr,1977). Research by Rincover, Cook, Peoples,and Packard (1979) and Rincover (1978) withchildren exhibiting self-stimulatory behaviorsuggests that an apparatus designed to attenuateor mask sensory stimulation derived from a spe-cific response may create an extinction effect,causing a reduction in retarded children's self-stimulatory behavior. This analysis seeminglymay be applied to self-injury. Clearly, manyforms of self-injury are topographically similarto self-stimulatory behavior, and may occur inthe absence of identifiable reinforcement contin-gencies. For example, in assessing the differentialeffects of environment on the self-injurious be-havior of nine subjects, Iwata, Dorsey, Slifer,Bauman, and Richman (1982) found that fourexhibited higher levels of self-injury in a situ-ation where opportunities for social stimulationand reinforcement were absent, and externalsources of physical stimulation were minimized(i.e., toys and other manipulable items were un-available). The major difference between self-injury and self-stimulatory behavior is that theimmediate or long-term effect of self-injury issome form of physical trauma. Thus, a procedurethat either attenuates or masks the self-stimula-tory components of self-injury and, at the sametime, protects the client from the deleterious re-sults of the behavior, might prove to be bothclinically effective and medically safe.

The initial focus of the present research wasto extend previous work by Dorsey, Iwata, Ong,and McSween (1980), in assessing the effects of apunishing stimulus-a water mist applied to theface-across time and settings. This procedurewas found ineffective in reducing self-injury dur-ing the initial stages of treatment with one sub-

218

PROTECTIVE EQUIPMENT FOR SELF-INJURY

ject. Rather than using more intrusive measures

(e.g., electric shock), the purpose of the researchwas altered to evaluate the effects of equipmentdesigned to protect the individual from self-in-flicted harm, and potentially attenuate the sen-

sory stimulation which occurred as a result of theself-injurious behaviors. Additionally, the studyinvestigated the maintenance of reduced levelsof self-injury via the use of response-contingentapplication of the apparatus combined with a

differential reinforcement procedure.

EXPERIMENT 1

METHOD

Participans

Three retarded clients of a state residentialfacility participated in this study. Selection was

based upon a display of a high rate of behaviorsconsidered to be self-injurious. Each resident wasconsidered to be a chronic self-abuser whose be-havior resulted in some form of tissue damage.Previous unsuccessful attempts at eliminatingthese behaviors included differential reinforce-ment, overcorrection, restraint, time-out, andvarious drugs.Ron was a 16-yr-old male, institutionalized

since the age of 2. His primary diagnosis was

profound mental retardation due to encephalo-pathy, secondary to a prenatal injury that causedanoxia. He had impaired hearing and vision andwas nonambulatory. His medical records indi-cated a history and variety of SIBs, with headhitting and hand biting being predominant.Physical damage, consisting of scalp nodules (i.e.,subdural hematomas) and abrasions of the skinresulted from his SIB.

Margie was a 16-yr-old female, institutional-ized since the age of 6. Her primary diagnosiswas severe mental retardation of undeterminedcauses, combined with congenital glaucoma. Shewas ambulatory and seemed to have normalhearing. Medical records indicated a long historyof self-injurious behaviors, including insertingher fingers into her eye sockets, hand biting, and

head hitting. Superficial cuts, callouses, and scartissue resulted from the hand biting behaviorwhile reddened areas and scar tissue around herface and eyes occurred as a result of the otherbehaviors.

James was a 14-yr-old male, institutionalizedsince the age of 4. His primary diagnosis wassevere mental retardation secondary to rubelladuring pregnancy, combined with a severe lossof vision. He was ambulatory and had a hearingimpairment. His target behavior was eye goug-ing-inserting his index finger into the eyesocket between the eye ball and eye lid to ap-proximately the second knuckle. This resultedin swelling to both the eyeball and the tissuewithin the eye socket. Medical records indicatedthat James had fractured a cataract as a resultof this behavior.

Setting

Two daily sessions were conducted individu-ally with each client. Afternoon sessions wereheld in the day area of the institution's livingunit, while morning sessions were conducted inthe day area for Ron and James, and in a specialeducation classroom at a local public school forMargie. The day areas measured approximately5.8 X 5.8 m, and were used as activity areas inwhich residents spent time while not engagedin other structured activities. Sessions were con-ducted by both institution psychology staff andparaprofessionals hired through the Comprehen-sive Employment and Training Act (CETA)program. Immediate supervision was providedby the first author.

ObservationResponse definitions were as follows:

1. Head hitting. Striking the head with anyportion of the hand, or (during treatment ses-sions) contact of a glove with the helmet (Ronand Margie).

2. Hand Biting. Insertion of the hand into the

219


mouth, or (during treatment sessions) contact ofa glove with the face mask (Ron and Margie).

3. Eye Gouging. Contact of the fingers withthe skin within the orbit of the eye, or insertionof the fingers into the eye socket; or (duringtreatment sessions) contact of a glove with theface mask (Margie and James).

4. Toy Play. Physical contact with availabletoys in which a toy was elevated manually fromthe floor or table by the participant a minimumof 1 in (Ron).

Occurrences of these behaviors and the use ofprotective equipment were recorded during non-continuous intervals in which the observer re-corded the behavior for five consecutive 10-secintervals and rested during the sixth, using a par-tial interval observation procedure (Powell,Martindale, & Kulp, 1975). A cassette tape con-taining prerecorded prompts was used to indicatethe beginning of each interval. The percentageof intervals during which the target responsesand/or use of the equipment occurred was ob-tained by dividing the positively scored intervalsby the total number of intervals and multiplyingby 100. All sessions were of a constant 20-minlength for each participant.

Observations were conducted by the trainerassigned to the session. Observers were trainedthrough instructions and modeling, and eachachieved a minimum criterion of 90% reliabilitywith the first author prior to formal data col-lection.

ReliabilityInterobserver agreement on the target behav-

iors and protective equipment usage was assessedduring 23% of the total sessions, distributedacross all participants and conditions. Duringsessions in which reliability was assessed, datawere collected by both the trainer assigned tothat session and an independent observer. Agree-ment percentages were calculated on an intervalby interval basis by dividing the number ofagreements on the occurrence or nonoccurrenceof the behavior and/or equipment usage by the

number of agreements plus disagreements andmultiplying by 100. Scores ranged from 87%to 100% (mean = 90%) for agreements of oc-currence and 80% to 100% (mean = 95%)for agreements of nonoccurrence for Margie;68% to 100% (mean = 91%) for agreementsof occurrence and 33% to 100% (mean =97%) for agreements of nonoccurrence for Ron;and 88% to 100% (mean = 99%) for agree-ments of occurrence and 66% to 100% (mean= 99%) for agreements of nonoccurrence forJames. Lower occurrence agreement scores wereobtained during sessions in which relatively fewoccurrences of the target behaviors were re-corded (e.g., 68% occurrence reliability for Ronin session 61 [unit] with SIB = 28%).

Procedure

Baseline. Target behaviors were observed andrecorded during individual sessions for each par-ticipant. No contingencies were applied to thetarget responses. Clients were not involved ineducational or recreational activities at this time,with the exception of Margie during school ses-sions, and the observers did not interact with theclients. Margie's teacher was instructed to treather as she had in the past relative to the targetbehaviors, as well as all academic tasks.

Reinforced toy play plus verbal reprimand.Four to five toys, selected from those availableon the clients' living units, were placed withinreach of each participant during this condition.Social praise and edibles (e.g., cookies, M&Ms)were provided on a 30-sec schedule contingentupon toy play, and the absence of SIB. Each oc-currence of self-injurious behavior was followedby a verbal "no" from the trainer in a forcefulbut normal speaking voice (American Sign Lan-guage was used with James).

Reinforced toy play plus verbal reprimandplus mist. Social and edible reinforcement wereprovided contingent on contact with toys in amanner identical to that described previously.In addition, the verbal reprimand provided inthe previous condition was paired with a fine

220

PROTECTWVE EQUIPMENT FOR SELF-INJURY

mist of water directed toward the client's facecontingent upon the occurrence of a target SIB.The mist was delivered from a standard plantsprayer and the method of application was iden-tical to that described by Dorsey et al. (1980).

Continuous protective equipment. At the be-ginning of each session, an apparatus was placedon the client designed to prevent injury and pos-sibly attenuate the tactile stimulation received asa result of the SIB. Due to the topographical sim-ilarities of the behaviors, a combination of foam-padded gloves and a football helmet lined withadditional foam padding was used for all partici-pants. The equipment did not fully prevent par-ticipants from engaging in the target responses,but did prevent injury from resulting. For exam-ple, Ron could continue to engage in a head hit-ting and/or hand biting response with the ap-paratus on by striking the football helmet withthe foam-padded gloves, or biting the gloves.The apparatus remained in place throughouteach session during this condition.

Two-minute protective time-out plus sensorystimulating toy play. Contingent upon the occur-rence of a target SIB, the protective equipmentwas applied for a period of 2 min, and the toysprovided were removed. The contingent use ofthe protective equipment was based on resultsobtained during the continuous protective equip-ment condition and the hypothesis that if antece-dent application suppressed SIB, consequent ap-plication might also be effective. The equipmentwas left in place until 30 sec had elapsed withno SIB. Due to the fact that the clients werephysically restrained from responding during theapplication and removal of the apparatus, datawere not collected during these periods and thesession length was increased to compensate forthis lost time. In the absence of SIB, continuousaccess was provided to toys designed to providesensory stimulation within the same sensory mo-dality as their SIB. Again, due to the topographi-cal simularities of Ron and Margie's SIBs, thesame types of toys were provided to both par-ticipants (e.g., a hand puppet with a battery op-erated vibrator enclosed, Busy Box), while James

was provided with visually stimulating toys (e.g.,a flashlight, mirrors).

Experimental DesignsMultiple baseline designs (Baer, Wolf, & Ris-

ley, 1968) across settings were used for eachclient. Following initial baselines in both morn-ing and afternoon sessions, the reinforced toyplay and verbal reprimand condition was intro-duced simultaneously in both sessions for Ron.Next, the verbal reprimand was combined withthe water mist during afternoon sessions whilea return to baseline was introduced in the morn-ing sessions. These conditions were implementedfor two reasons: (a) to establish the verbal repri-mand as a conditioned punisher (first setting)and (b) to allow for a functional relationship be-tween the treatment and behavior to be demon-strated later. Next, the continuous protectiveequipment condition was implemented, first inthe afternoon and later in the morning sessions.This was done to evaluate the effects of the pro-tective equipment prior to beginning the sensorystimulating toy-play condition. Finally, the twominute protective time-out plus sensory stimulat-ing toy play condition was introduced in a multi-ple baseline fashion across sessions.Two deviations from this basic design were

made for Margie. The reinforced toy play plusverbal reprimand plus mist condition was notconducted. Additionally the continuous protec-tive equipment condition was implemented onlyin the first setting. These changes were made fortwo reasons. First, the water mist procedure wasdropped because Margie had a history of aggres-sive reactions to water. Second, the continuousprotective equipment condition was withheldfrom the second setting to control for possiblesequence effects between this and the final con-dition.

Finally, one additional deviation from theoriginal design was made for James. Prior to theuse of the protective equipment, access to sensorystimulating toys was made available to Jamesthroughout the session as long as no SIB oc-curred. The toys were removed contingent upon

221

222 ~~~~~MICHAEL F. DORSEY et al.

the occurrence of SIB, plus an additional 30 sec

change over delay. This change was imple-mented to control for the possibility that the sen-

sory stimulating toys alone would act to controlSIB, without the need for protective equipment.

RESULTS AND DiscuSSIONData for Ron, Margie and James, expressed as

percent intervals of total SIB, are presented inFigures 1, 2, and 3, respectively. Participants typ-ically exhibited high rates of SIB during base-line. Ron's SIB was highly variable throughoutbaseline, and no particular events could be iden-tified to account for either high or low levels ofresponding. Inconsistent changes in respondingwere noted during the initial treatment condi-tions. The mean changes in behavior were asfollows: Ron's SIB increased from 33.8% and34.4% in the afternoon and morning baselinesessions to 8 1.8% and 60% upon implementa-tion of the verbal reprimand and reinforced toyplay condition. Margie's SIB, however, decreasedfrom 87.7%7 to 79.1% with the introduction ofthe same procedure. Use of the water mist pro-cedure reduced Ron's SIB from 8 1.8% to34.6%o. Finally, the use of contingent access to

sensory stimulating toys reduced James' SIBfrom 92% to 73%6. Similar changes were noted

in other conditions, in that none of the initialattempts at eliminating the clients' SIB resultedin clinically significant reductions.Upon implementation of the continuous pro-

tective equipment conditions, SIB decreased no-

ticeably for all three participants. For example,Ron's SIB was reduced from 34.6%6 to 8.3%6in the afternoon sessions and from 90%9 to 3.6%9in the morning sessions. Margie's SIB decreasedfrom a mean of 79.1%9 to 18%9, and James'from 73%6 to 4%6. Across all participants, SIBwas reduced from the mean of the prior condi-tions of 69.4%6 to a treatment mean of 8.5%6.During the last five days of treatment, SIB acrossall participants averaged 1.25%9 with a range of0.0%9 to 6%9.

The final change in treatment to the two min-ute protective time-out plus sensory stimulatingtoy play condition was effective in maintaininglow levels of SIB, and reduced the percentage oftime which clients were exposed to the protec-tive equipment. SIB for this condition averaged11.7%6 in the afternoon sessions and 18.6%6 inthe morning sessions for Ron, 14.7%6 on theunit and 12% in school for Margie, and 1.2%6in the afternoon and 0.7%6 in the morning forJames. The last five days of treatment for Ronshowed SIB averaging 2.6%6 in the afternoon

100-

Bo-

SO-

40-

20-

0-

REINFORCED TOYPLAY + VERBAL REINFORCED TOY PLAY.* CONTINUOUS EUPETTMOJ

PROTECTIVEEQIMNTMEUBASELINE REPRIIMAND VERBAL REPRIMAND, MIS1T EQUIPMENT 4 SENS STIM TOY PLAY

Fig. 1. Percentage of intervals of SIB (closed circles) and protective equipment usage (open circles) across

experimental conditions for Ron (Experiment 1).

5 lD 152025 303o5 40 4Sri550 56 657075' 6065 iO 95 160010511013S120O125 1301355140 1456NSESSIONS

9z

It t :: It.i 6 #";6a

222

jevt;5

W)-ja

mg

LeJui

I9I--zI'lu

X!Mui da.

t;z


VERBAL REPRIMAND +

BASELINE REINFORCED TOY PLAY

100-,080 .

40-ll

20- 1l

CONTINUOUSPROTECTIVEEQUIPMENT

EQUIPMENT/TIMEOUT4 SENS STIM TOY PLAY

55 60 65 70 75 80 85 90 95 100 105 110 115SESSIONS

Fig. 2. Percentage of intervals of SIB (closed circles) and protective equipment usage (open circles) acrossexperimental conditions for Margie (Experiment 1).

sessions and 9.4% in the morning sessions, whileMargie had a mean of 11.2% SIB for the lastfive days of treatment on the unit and 7.8% inschool, while James remained at 0.0% duringboth a.m. and p.m. sessions over the last five daysof treatment. Similar results can be noted for thepercentage of time the participants were exposedto the protective equipment. The mean percent-

age of time that clients spent wearing the equip-ment during this condition was: 29.9% in theafternoon sessions and 39.7% in the morningsessions for Ron, 50.9% on the unit and 35.4%at school for Margie, and 9.7% in the afternoonand 6.0% in the morning for James. As with theSIB data, these means do not accurately reflectdescending trends noted for all three partici-pants. For example, the means for the last fivedays of treatment for Ron were 9.4% in the af-ternoon sessions and 13.6% in the morning ses-

sions.One possible explanation for the procedure's

suppressive effect on behavior is that the weightof the gloves may have increased the "responseeffort" of SIB. Alternatively, the wearing and/orcorresponded to reductions in SIB. Ron was instriking of the helmet alone may have had

punishing properties. In order to evaluate theseconfounding variables, several pre- and posttreat-

ment probes were conducted with Ron, on ses-

(z

LiJ

z

Li

IU

0.

100

0so60

II"

20~

00

z

CONTINUOUSSENS STIM PROTECTIVE EQUIPMENT/TIMEOUT

BASELINE TOY PLAY EQUIPMENT + SENS STIM TOY PLAY

1'III 1rte EUblt'...1 In ~I

L -1

SESSIONS

Fig. 3. Percentage of intervals of SIB (closed cir-cles) and protective equipment usage (open circles)across experimental conditions for James (Experi-ment 1).

223

z

U,

Li

:wz

uj

I-

z


sion days 113, 114, and 115, in which only thegloves or only the helmet was applied. In allcases, Ron engaged in his particular SIB at rates

comparable to those observed during the initialbaseline, averaging 73% SIB with the helmetalone and 92% with the gloves alone across thethree sessions.

Within the present experiment little data can

be provided to demonstrate the establishment ofbehavioral alternatives to SIB (i.e., toy play).Although this was an initial goal of the present

research, final use of the data collected towardthis end was not possible due to a particular flawin the observation procedure. Specifically, "toyplay" was initially defined as "physical contact

with available toys involving their manual ele-vation from the floor or wheelchair table a min-imum of 1 in." The purpose of this particulardefinition was to exclude instances in which par-

ticipants merely rested an arm or hand upon thetoy. However, such a definition precluded therecording of many appropriate "toy play" re-

sponses in subsequent sensory stimulating toy

play conditions. That is, when presented with a

toy that provided the client with the same stimu-lation regardless of the topography of his or herinteractions with it, the highest probability re-

sponse would be the one with the least effort(i.e., resting an arm or hand on the toy). In orderto prevent confounding the data by changingdefinitions across conditions, this particular re-

sponse was eliminated from the observationalsystem at the conclusion of treatment for Ron(both Margie and James were involved at some

earlier point in the study at the time the defini-tional problem was discovered, eliminating thepossibility of correcting the problem for eitherof them). Data using this definition of "top play"for Ron are presented on Table 1, expressed as

the mean percentage of intervals across condi-tions. As the table shows, Ron engaged in thehighest levels of toy play during the final pro-tective time-out plus sensory stimulating toy playcondition during both morning and afternoonsessions. Although problems do exist with thesedata, it is interesting to note that SIB and toy

play appeared to be inversely related. Finally,subjective observations of all three participantssuggested that interactions with the various toys

did increase as a function of the introduction ofsensory stimulating toys in the final condition.

Results of Experiment 1 indicate that protec-

tive equipment may be a useful treatment forSIB. However, the results are limited becauseof the short duration of experimental sessions,as well as the specialized training of the staffwho implemented the procedure. Experiment 2was designed to assess the effects of the treatment

implemented in more naturalistic settings across

a longer period of the day by staff normally em-

ployed on institutional wards.

EXPERIMENT 2

METHOD

Participants and Settings

Ron, Margie, and James also participated inthis study. Daily sessions were conducted indi-

Table 1

Mean percentage of toy play and self-injurious be-havior by condition for Ron.

Afternoon SessionsCondition Toy Play SIB

Baseline No Toys AvailableReinforced Toy Play +

Verbal Reprimand 2 81.8Reinforced Toy Play +

Verbal Reprimand+ Mist 14 34.6

Continuous ProtectiveEquipment No Toys Available

Two Minute ProtectiveTime-out + SensoryStimulating Toy Play 19 11.7

Morning SessionsCondition Toy Play SIB

Baseline No Toys AvailableReinforced Toy Play +

Verbal Reprimand 3 91Baseline No Toys AvailableContinuous Protective

Equipment No Toys AvailableTwo Minute Protective

Time-out + SensoryStimulating Toy Play 12 18.6

224


vidually within their day area, bedroom, and din-ing room, beginning approximately 1 mo afterthe conclusion of Experiment 1. Sessions were

run Monday through Friday, from the time eachclient returned to the center from school (3:00p.m.) until bedtime (8:00 p.m.). No attempt

was made to isolate the participants from envi-ronmental distractions or to restrict their normaldaily schedule. Therapists were three CETA/grant employees, one assigned to each partici-pant. During follow-up observations, procedureswere conducted by direct care staff of the insti-tution.

Observation

Response definitions were identical to thoseused in Experiment 1. The occurrence of thesebehaviors and the application of the protectiveequipment were recorded using a partial intervalobservation procedure (Powell et al., 1975).Observations were conducted daily, with behav-ior sampled for six continuous 10-sec intervalsevery 15 min.

ReliabilityReliability on the occurrence of the target

behaviors and the use of the protective equip-ment was assessed during all conditions by as-

signing two observers to record independent ob-

servations a minimum of four days per week,overlapping 2-3 observation periods per partici-pant. Results were calculated as described previ-ously. Scores ranged from 91% to 97% (mean= 94%) for occurrence and 80% to 100%(mean = 97%) for nonoccurrence.

Procedure

Baseline. Target behaviors were recorded dailyfor each participant during regularly scheduledactivities. Examples included games, crafts, out-

side play, and involvement with nursing or re-

habilitation staff in various therapies. No experi-menter-controlled contingencies were in effectfor the target responses during this condition.Staff were informed that data were being col-lected and were instructed to interact with the

clients as they had previously. Examples of con-tingencies currently in effect were: nonsystematicdifferential reinforcement procedures, responseinterruption, verbal reprimands, and physicalprompting of other behaviors.Two minute protection time-out plus contin-

gent sensory stimulating toy play. Contingentupon the occurrence of SIB, the apparatuses usedin Experiment 1 were placed on the clients fora period of 2 min, as described previously. Con-tingent upon the absence of a target SIB, sensory-stimulating toys were made available in the samemanner as described previously. Implementationof these procedures was accomplished by theCETA staff member assigned to each participant.

Follow-up. During follow-up sessions, contin-gencies remained essentially the same as in theprevious condition. However, implementation ofthe treatment program was turned over exclu-sively to the direct care staff assigned to the cli-ents' living unit. Feedback was given to the staffregarding the application of the procedure viathe unit program supervisor (fourth author) ina manner similar to that concerning other pro-grams conducted on the living unit.

Experimental Design

Following the collection of baseline data forall three participants, the two minute protectivetime-out plus sensory stimulating toy play con-dition was implemented in a multiple baselinedesign across Ron and Margie, and in an A-Bfashion for James, The follow-up condition wasimplemented for Margie and James to providean assessment of the procedure when run com-pletely by direct care staff on a long-term basis(i.e., over a period of 104 and 100 calendar days,respectively). Follow-up was not possible forRon, due to his transfer to a nursing home dur-ing the final experimental condition.

RESULTS AND DISCUSSION

Data for Ron and Margie are presented inFigure 4, and data for James are presented inFigure 5. All three clients typically exhibitedrates of SIB during baseline equivalent to those

225


100

80

z0

CO-jcc

Lii

z

60

40

20

0 -

BASELINEEQUIPMENT/TIMEOUT+ SENS STIM TOY PLAY

III

LI_ _1

II f'

I EQUIPMENT

&I 0

zosc,# , 's p , . S

I : ' b01 1 I0 b

IFOLLOW -UP

II k

,1XII..

5 o 15 20 25 30 35 4 450 85 120

SESSIONSFig. 4. Percentage of intervals of SIB (closed circles) and protective equipment (open circles) across experi-

mental conditions for Ron and Margie (Experiment 2).

seen at the onset of Experiment 1. Upon imple-mentation of treatment, rapid reductions were

noted for all three participants. Mean changesfrom baseline to treatment were as follows: Ron,48% to 7.2%; Margie, 47% to 6.4%; James,63.2% to 1.9%. Additionally, the percentage

of time each client wore the equipment also de-creased throughout the treatment condition andcorresponded to reductions in SIB. Ron was inprotective equipment an average of 13.1% of hisdays, Margie 33.6%, and James 4.2%. Means

for the final five days of treatment showed fur-ther reductions in equipment usage. During thisperiod, Ron was in the protective equipment6%, Margie 13.8%, and James 0.0%. Thus,both the reductions in SIB as well as the percent-

age of time in protective equipment were very

close to the results found in Experiment 1.A major goal of Experiment 2 was to evaluate

the procedure when implemented by nonprofes-sional staff. This was accomplished during thefollow-up phase of treatment with Margie and

J-LiiCO

100

80

rln-

0

I-4

zL)JL-)

LUJ0- or

142 40-

20-

0' 0 10 a 0 v R a E. E

226

I


EQUIPMENT/TIMEOUT+ SENS STIM

BASELINE TOY PLAY FOLLOW-UPIIIII

.<v~~IIA I ~~~~~~~~~~~~I

SIB III II I

IEQUIPMENT A

I

5 10 15 20 25 45 50 85 90

SESSIONSFig. 5. Percentage of intervals of SIB (closed circles) and protective equipment usage (open circles) across

experimental conditions for James (Experiment 1).

James. Although Ron was discharged to a nurs- these use of the equipment observed during onlying home, his discharge was contingent upon the one of the follow-up probes over a period of 100development of a procedure that would success- calendar days following the termination of ex-

fully control his SIB. Several visits to the nursing perimenter-implemented treatment.

home conducted by the institution's social servicestaff over a period of 6 mo suggested that the GENERAL DISCUSSION

procedure was being carried out and that Ron'slevel of SIB was being maintained at acceptably Results suggest that the use of protectivelow levels. equipment may affect both rapid and substantial

Follow-up data for Margie and James indi- decreases in SIB. In addition, the data fromcated that the direct care staff assigned to work Experiment 2 suggest that the combination ofwith them were able to implement the procedure contingent protective equipment and access to

successfully. Results of the first probe conducted sensory-stimulating toys may support the main-with Margie 48 calendar days after the termina- tenance of treatment gains.tion of the treatment condition showed an in- The use of the protective equipment pro-

crease in SIB, while data for the second and third cedure in the reduction of SIB seems justifiedprobes conducted 55 and 104 calendar days post- relative to many aversive procedures currentlytreatment were comparable to those found at the available. The development of a hierarchy oftermination of experimenter-implemented treat- techniques, based upon the model of "less restric-ment. Similar results were noted for James, with tive" procedures (May, Risley, Twardosz, Fried-

C))-J

z

LLJ0

w0Hw0nw

100

80

60

40-

20

0

0

o Q0 a

227

I

I

MICHAEL F. DORSEY et 4l.

man, Bijou, Wexler et al., 1975), would seem to

include contingent protective equipment as a

more desirable procedure than even "mild" aver-

sive stimulation. The techniques does not require

that the client be exposed to stimuli which, sub-jectively, cause pain to the individual. In addi-tion, the procedure affords the client some degreeof protection from self-inflicted injury duringtreatment, but does not include many of delete-rious components of common forms of restraint(e.g., arm splints, camisoles). However, two is-sues should be considered prior to its general use.

First, precautions should be taken to ensure

the safety of the client, as well as others withinhis or her environment when the procedure isused. Care should be taken to protect the clientfrom extreme bursts of responding which may

occur initially as a function of the procedure.That is, one should go beyond the intensity ofthe response as it exists in baseline when con-

sidering the type of apparatus to be used. Al-though this aspect may be critical for a smallsegment of the total treatment duration, the finaleffectiveness of the procedure may rest upon thisspecific issue. It is possible that the use of an

apparatus that does not attenuate severe levelsof responding may serve only to intensify theinitial level of responding. Additionally, physicalfeatures of the apparatus must be considered inrelation to its potential use as an instrument ofaggression. Devices such as helmets or face-guards may be potentially used by the client to

increase the effectiveness of aggressive behavior.Whenever possible, the apparatus selectedshould be designed so as to take these potentialproblems into consideration.

Second, learning principles from which thistreatment is derived may be important whenconsidering its use. The procedure was originallydesigned, based upon the work of Rincover(1978) and Rincover et al. (1979) with self-stimulatory behaviors. Rincover suggested thatcertain classes of behavior directly produce rein-forcement of a sensory nature, and that if one

could mask or attenuate this source of stimula-tion, the response would decrease as a function

of extinction. Although the present study did notprovide data to support such a position, clearly,the behaviors had many similar characteristics tothose behaviors studied by Rincover, in that theywere highly repetitious and seemed to occur inthe absence of any external consequences. Be-cause of the risk of physical injury these behav-iors posed to the participants, however, manipu-lations such as those conducted by Rincover werenot attempted. In the absence of such data, onecan only speculate as to the actual conceptualbasis of the procedure. However, one means ofempirical support can be seen through a within-session analysis of responding during treatment.Skinner (1938) described a typical extinctioncurve as having "a high rate of elicitation main-tained for a short time" subsequent to the termi-nation of the delivery of reinforcement contin-gent upon responding, a fluctuating response ratelater within the session, and a smooth flattenedcurve finally developing. Continued reexposureto such a situation, as was created in this experi-ment, should have caused progressively flattercurves each time the client came in contact withthe protective equipment. That is, a discrimina-tion would be established between the within-session equipment condition and the between-session reinforcement condition. The recordspresented in Figure 6 show cumulative, within-session responding for Ron during sessions 91,98, and 107 within the continuous protectiveequipment condition of Experiment 1. The dataappear to match the typical pattern of respond-ing described by Skinner and add some supportto the position that the phenomenon observed inthis experiment was an extinction effect. If, infact, such a conceptual basis is a correct assump-tion, the procedure would provide an alternativewhose focus is directly in line with those vari-ables responsible for the maintenance of someforms of SIB. Such an approach in the selectionof a treatment technique would seem to have ahigher probability of success than an attempt toreduce SIB through the manipulation of contin-gencies irrespective of the variables responsiblefor the behavior.

228


CO-Jcc:

LU 100

~ 75W so"o"0,0X .. 0.""_O0O00

25oodls&.:pitOiooooiooooooooooooo0ooooooo0ooooooooo0ooo0oooooo oooooooo@LU_ 25 000O0 go

10 20 30 40 50 60 70 80 90 100U INTERVALS

Fig. 6. Cumulative number of intervals in which Ron engaged in SIB during sessions 91, 98, and 107 (Ex-periment 1).

An alternative explanation for the procedure'seffectiveness is that the placement of equipmenton the client and removal of access to toys com-bined with systematic ignoring is simply a time-out procedure in which the client is protectedfrom physical injury. The reduction in SIBwould then be attributed to the contingent re-moval of social attention and/or preferred ac-tivities. A third position that must be consideredis that the application of the apparatus acts as anaversive stimulus, causing a decrease in respond-ing simply as a function of punishment. The de-termination of which, if any, of these three po-tential variables are responsible for the pro-cedure's effectiveness seems to be an importantissue and one that should be resolved throughfuture research. The~tesults of such an analysisshould not detract from the procedure's technicalor ethical significance. In fact, if those resultstended to support either the time-out or punish-ment paradigms, as opposed to the sensory ex-tinction model, the technique's generalized ap-plicability would seem enhanced.

Finally, the results of the present study shouldbe considered in relation to other research find-

ings on the use of restraint procedures. As notedpreviously, restraint may have reinforcing as wellas punishing properties. Although the presentstudy adds yet another example of the punishingeffects, it should be noted that the devices usedwere different from those described by Favell etal. (1978, 1981), in that they did not restrictthe subjects' movement during treatment.Rather, the served only to protect the individu-al's from self-inflicted harm. Although this dif-ference may have contributed to a divergence inthe results of the two studies, a number of othervariables may function to establish restraint aseither a reinforcing or punishing event, includ-ing increased adult attention and/or increasedphysical comfort during periods of restraint(Favell et al., 1978), the opportunity to escapefrom a more aversive environment (Carr, 1977),or the discriminative properties of restraint as a"safety signal" from response requirements oraversive events (Jones, Simmons, & Frankel,1974). Nevertheless, it is apparent that restraintcan have similar effects on the occurrence of acommon target behavior when used in differentways. This observation underscores the impor-

229

230 MICHAEL F. DORSEY et al.

tance of improving our understanding of themotivational variables responsible for a client'sbehavior prior to attempting to treat the client(e.g., Iwata et al., 1982), while taking advan-tage of the idiosyncratic nature of reinforcers andpunishers.

REFERENCE NOTE

1. Schroeder, S. The analysis of self injurious be-havior: Pathogenesis and treatment (DDSA H.I.P.Grant No. 5 1-P-205 1). Chapel Hill: University ofNorth Carolina, 1977.

REFERENCES

Baer, D. M., Wolf, M. M., & Risley, T. R. Somecurrent dimensions of applied behavior analysis.Journal of Applied Behavior Analysis, 1968, 1,91-97.

Carr, E. G. The motivation of self-injurious be-havior: A review of some hypotheses. Psycho-logical Bulletin, 1977, 84, 800-816.

Dorsey, M. F., Iwata, B. A., Ong, P., & McSween,T. E. Treatment of self-injurious behavior usinga water mist: Initial response suppression andgeneralization. Journal of Applied Behavior Anal-ysis, 1980, 13, 343-354.

Favell, J. E., McGimsey, J. F., & Jones, M. L. The useof physical restraint in the treatment of self-injuryand as positive reinforcement. Journal of AppliedBehavior Analysis, 1978, 11, 225-242.

Favell, J. E., McGimsey, J. F., Jones, M. L., & Cannon,P. R. Physical restraint as positive reinforce-ment. American Journal of Mental Deficiency,1981, 85, 425-432.

Hamilton, J., Stephens, L., & Allen, P. Controllingaggressive and destructive behavior in severelyretarded institutionalized residents. AmericanJournal of Mental Deficiency, 1967, 71, 852-856.

Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K.E., & Richman, G. S. Toward a functional anal-ysis of self-injury. Analysis and Intervention inDevelopmental Disabilities, 1982, 2, 3-20.

Joint Commission on Accreditation of Hospitals.Standards for residential facilities for the mentally

retarded. Chicago: Accreditation Council for Fa-cilities for the Mentally Retarded, 1975.

Jones, F. H., Simmons, J. Q., & Frankel, F. An ex-tinction procedure for eliminating self-destructivebehavior in a nine-year-old autistic girl. Journalof Autism and Childhood Schizophrenia, 1974,4, 241-250.

Lovaas, I. O., & Simmons, J. Manipulation of self-destruction in three retarded children. Journal ofApplied Behavior Analysis, 1969, 2, 143-157.

May, J. G., Risley, T. R., Twardosz, S., Friedman, P.,Bijou, S. W., Wexler, D. et al. Guidelines forthe use of behavioral procedures in state programsfor retarded persons. MR Research, 1975, 1.

Parrish, J. M., Aguerrevere, L., Dorsey, M. F., &Iwata, B. A. The effects of protective equipmenton self injurious behavior. The Behavior Ther-apist, 1980, 3, 28-29.

Powell, J., Martindale, A., & Kulp, S. An evaluationof time-sampling measures of behavior. Journalof Applied Behavior Analysis, 1975, 8, 463-469.

Rapoff, M. A., Altman, K., & Christophersen, E. R.Elimination of a blind child's self-hitting by re-sponse-contingent brief restraint. Education andTreatment of Children, 1980, 3, 231-236.

Rincover, A. Sensory extinction: A procedure foreliminating self-stimulatory behavior in develop-mentally disabled children. Journal of AbnormalChild Psychology, 1978, 6, 299-310.

Rincover, A., Cook, R., Peoples, A., & Packard, D.Sensory extinction and sensory reinforcementprinciples for programming multiple adaptivebehavior change. Journal of Applied BehaviorAnalysis, 1979, 12, 221-234.

Rojahn, J., Mulick, J. A., McCoy, D., & Schroeder,S. R. Setting effects, adaptive clothing, and themodification of head banging and self-restraint intwo, profoundly retarded adults. Behavioral Anal-ysis and Modification, 1978, 2, 185-196.

Rojahn, J., Schroeder, S. R., & Mulick, J. A. Eco-logical assessment of self-protective devices inthree profoundly retarded adults. Journal ofAutism and Developmental Disorders, 1980, 10,59-66.

Skinner, B. F. The behavior of organisms. Engle-wood Cliffs, N.J.: Prentice-Hall, 1938.

Received April 22,1 981Final acceptance November 24, 1981

protective equipment: continuous and contingent application ...

Documents

Transcript of protective equipment: continuous and contingent application ...