JOURNALOFAPPLIEDBEHAVIORANALYSIS 2009, 42, 381–385 NUMBER2 (SUMMER2009) CHEWING GUM AS A TREATMENT FOR RUMINATION IN A CHILD WITH AUTISM DENISE RHINE AND JONATHAN TARBOX CENTERFORAUTISMANDRELATEDDISORDERS Rumination involves regurgitation of previously ingested food, rechewing the food, and reswallowingit.Inthecurrentstudy,achildwithautismdisplayedchronicrumination,resulting inthe decay andsubsequent removal ofseveral teeth. Afterseveral treatments failed, including thickenedliquidsandstarchsatiation,theparticipantwastaughttochewgum.Hisrumination decreasedsignificantlywhengumwasmadeavailable.Resultssuggestthataccesstochewinggum maybean effectivetreatment for rumination insome individuals. DESCRIPTORS: autism,chewing gum, rumination _______________________________________________________________________________ Rumination involves regurgitation of previ- METHOD ously ingested food, rechewing the food, and Client and Setting reswallowing it. The potential negative effects Reggie was a 6-year-old boy with a diagnosis on the quality of life of individuals who display of autism who attended a home-based behav- rumination include weight loss, malnutrition, dental decay, halitosis, and electrolyte abnor- ioral intervention program for children with malities (Chial, Camilleri, Williams, Litzinger, autism during afterschool hours. Reggie was & Perrault, 2003). Early research on the able to communicate his basic needs via three- reduction of rumination in individuals with to four-word mands; other than that, he developmental disabilities often involved aver- engaged in very little unprompted speech. sive procedures, such as contingent delivery of According to a recent language assessment aversivetastes(Sajwaj,Libet,&Agras,1974).A (Preschool Language Scales–4), Reggie scored small number of less aversive procedures (e.g., in the 1st percentile on expressive and receptive starchsatiation)havealsobeendemonstratedto language, indicating a very significant language be effective. Starch satiation involves giving an delay. According to parental report, Reggie had individual unlimited access to starchy foods, been ruminating regularly throughout each day typically immediately after meals, and has been for approximately 1 year, and there were no demonstrated to be an effective treatment in specific activities in which rumination did not several studies (e.g., Rast, Johnston, Drum, & occur. No formal data were collected to assess Conrin, 1981), in some cases producing long- which contexts rumination occurred in most lasting effects (Dunn, Lockwood, Williams, & frequently, but his caregivers reported that it Peacock, 1997). In the current study, we occurred in the morning, evening, and after- evaluated the effects of a novel treatment noon; both in school and at home; and during procedure, access to chewing gum, on the both educational activities and play time. rumination of a young boy with autism for Reggie’s dentist asserted that the decay caused whom starch satiation was ineffective. by his rumination had necessitated the removal of five of his teeth. Prior treatments for We thank Jennifer Purcell for her assistance with data rumination included thickening of liquids and collection. AddresscorrespondencetoJonathanTarbox,Centerfor starch satiation, neither of which produced a Autism and Related Disorders, 19019 Ventura Blvd, 3rd reduction in rumination. In the current study, Floor, Tarzana, California 91356 (e-mail: j.tarbox@ all intervention procedures were implemented centerforautism.com). doi:10.1901/jaba.2009.42-381 and all data were collected as a part of his 381 382 DENISE RHINE and JONATHAN TARBOX regularly scheduled behavioral intervention 93%, (range, 0% to 100%; only one session sessions, which occurred in his home. The produced an agreement of 0%, which consisted primary therapeutic activities during his regular of theprimaryobservertallyingonerumination behavioral intervention sessions consisted of in a 30-min session and the secondary observer naturalistic and discrete-trial approaches to tallying none). teaching a variety of adaptive skills, including language, socialization, preacademic skills, and Experimental Design and Sequence activities of daily living. These regularly sched- A multielement design was used to compare uled therapeutic activities continued for the the effects of continuous noncontingent access duration of the study, and all procedures were to chewing gum to the absence of gum. Blocks implemented by his regular therapy team. of sessions were scheduled several times per week,basedontherapistandparentavailability. Response Measurement and For each block of sessions, a coin toss was used Interobserver Agreement to determine the sequence of conditions within Data were collected on the frequency of that block, with the rule that no condition rumination throughout the study and were could be conducted more than three consecu- summarized as an hourly rate. Rumination was tivetimes.Thesequenceof sessionswithineach defined as the presence of food material in block of sessions was determined with a coin Reggie’s mouth at a time other than when he toss so that inadvertent scheduling of particular was eating. He was not prompted to open his conditions before or after meals could not bias mouth in order to facilitate data collection; one condition or the other toward higher or however, his mouth opened regularly through- lowerratesofrumination.Priortotheinitiation out the day, typically to vocalize in some of the gum-chewing treatment, Reggie had manner, so therapists were able to look into it reportedly never chewed gum and apparently and observe regurgitated food frequently. did not possess the skill to do it (e.g., when Reggie also frequently chewed and manipulated handed a piece of gum, he would look at it and regurgitated food with his tongue, and his donothing).Therefore,beforethetherapyteam mouth regularly opened during the course of initiated the gum-chewing treatment, the such movements, thereby giving his therapists a speech and language pathologist in Reggie’s natural opportunity to observe the presence or special education classroom designed and absence of food. Data were collected on the implemented a skill-acquisition program to presence of regurgitated food in the mouth, as teach him to chew gum. Instruction consisted opposed to the act of regurgitating, because it of shaping the behavior by reinforcing compli- was nearly impossible to detect regurgitation. ance with gradually increasing requirements, Virtually no neck or head movement (i.e., starting with placing a piece of gum in the retching) occurred when Reggie regurgitated. mouth, then biting it, then chewing it twice, Twenty-two percent of sessions were video- andsoon,untilReggiewasabletochewapiece taped, and a second observer viewed the of gum for several minutes. videotapes at a later time and collected data for the purposes of assessing interobserver Procedure agreement. Interobserver agreement was calcu- During Sessions 1 through 17, in both gum- lated using the total method, in which the chewing and no-gum conditions, the therapist smaller frequency recorded across a whole gave Reggie a maintenance task, such as a book session was divided by the larger frequency to look at or a worksheet to complete. The recorded, and the resulting fraction was con- duration of these initial sessions depended on verted to a percentage. Mean agreement was how long it took Reggie to complete the task RUMINATION 383 Figure1. Hourlyratesofruminationingum-chewing(filledcircles)andno-gum(opensquares)conditions.Session durationwasgraduallyincreasedto120minacrossthecourseoftreatment.Thefinalthreedatapointsrepresentfollow- updataat 1,2, and3months posttreatment. and ranged from approximately 3 min to noncompliant with the initial instruction to approximately 10 min. Starting with Session chew the gum and rarely expelled the gum. 18,thedurationofsessionswasfixedat15 min The size of the piece of gum was gradually and was then gradually increased to 120 min increased as session duration was increased, over the course of the study. Sessions were movinguptoafullpieceofgumonSession27. conducted 3 to 5 days per week, two to five Sessionsoftheno-gumconditionwereidentical times per day. During gum-chewing sessions, in every way, except that Reggie was not given Reggie’s therapist presented him with one gum to chew. Rumination produced no quarter of a piece of chewing gum and told reaction from the therapists during any phase him to chew it. If he did not accept the gum, of the study. the therapist held it within 0.3 m of Reggie’s mouth. The therapist again told him to chew RESULTS AND DISCUSSION the gum and repeated this prompt approxi- matelyevery3 suntilReggieacceptedthegum. Figure 1 depicts rates of rumination in both If he expelled the gum at any point during a the gum-chewing and no-gum conditions. The session,thetherapistre-presentedthegum(orif gum-chewing condition produced consistently the gum fell on the floor, the therapist re- lower rates of rumination than the no-gum presented a fresh piece) and asked him to chew condition. Furthermore, rates of rumination it. Systematic data were not collected on remainedlowwhenthedurationofsessionswas compliance with the directive to chew gum, gradually increased to 120 min. Reggie contin- but therapists reported that Reggie rarely was ued to demonstrate low rates of rumination 384 DENISE RHINE and JONATHAN TARBOX duringfollow-upsessions at1,2,and3months therefore did not have the capacity to record posttreatment.Hevisitedhisdentist1yearafter data across multiple intervals of short duration this evaluation was completed, and his dentist (e.g., 10 s), as would be required for more reported that Reggie’s tooth decay had ceased. stringent analyses. Due to the consistent and This study provides initial evidence that often large differences in the rate of rumination providing access to chewing gum may be an across conditions, we do not feel that any effective treatment for reducing rumination in breachesinmeasurementagreementaffectedthe some individuals with autism or other develop- validity of the relation demonstrated. Never- mental disorders. Providing access to chewing theless, future research on treating rumination gum is a relatively inexpensive, low-effort, and via chewing gum should attempt to employ nonintrusive treatment. In addition, it may agreement procedures of greater rigor. entail the ingestion of fewer calories than a Third, no systematic data were collected starch satiation procedure in cases when while Reggie was initially taught how to chew excessive weight gain is a concern. gum, and a precise technological description of Several limitations of the current study the teaching procedure is not available. Thus, warrant discussion. First, functional assessment replication of that procedure is likely to be of challenging behaviors prior to implementa- difficult for future investigators. In future tion of behavioral interventions is widely studies, researchers should document and considered best practice (Hanley, Iwata, & provide a more technological description of McCord, 2003), and we did not conduct a procedures used to teach gum chewing to functional analysis of Reggie’s rumination prior facilitate replication. to the intervention. The rumination displayed Fourth, the procedures used in this study do by some individuals may be maintained by not allow any definitive conclusions regarding automatic reinforcement; until recently, rumi- the behavioral processes that were responsible nation research generally did not include for the reduction in rumination observed in the functionalanalyses(foranexception,seeLyons, gum-chewing condition. It seems most likely Rue, Luiselli, & DiGennaro, 2007), but the that chewing the gum produced an alternative lack of a functional assessment in the current formoforalstimulationthatcompetedwiththe study makes any assumption regarding main- oral stimulation produced by ruminating. This taining reinforcers tenuous. However, the fact possible interpretation is supported by previous that rumination did not decrease over the research on other automatically reinforced course of 62 no-gum sessions, in which there behaviors that has shown that providing access were never any socially mediated consequences to an alternative behavior that results in oral for rumination, provides evidence that rumina- stimulation similar to that produced by the tion was at least partially maintained by target problem behavior reduces the target automatic consequences. behavior (Piazza et al., 1998). Second, the method of interobserver agree- Finally, we did not extend the duration of ment calculation employed in this study, the sessions beyond 120 min to an all-day analysis total method, is generally considered to be of the effects of gum chewing on rumination an insensitive measure of agreement (Repp, (see Lindberg, Iwata, Roscoe, Worsdell, & Deitz,Boles,Deitz,&Repp,1976).Weelected Hanley, 2003, for an example of this sort of tousetotalagreementbecausethetherapistwho temporally extended evaluation). Nevertheless, collected data was also responsible for the the use of 120-min sessions may be viewed as a implementation of teaching trials and data strength of the current study, considering the collection for a variety of other behaviors, and common practice of 10- or 15-min session RUMINATION 385 durations common in applied behavioral re- Dunn,J.,Lockwood,K.,Williams,D. E.,&Peacock,S. (1997).Aseven-yearfollowupoftreatingrumination search. Chewing gum also may not be a viable with dietary satiation. Behavioral Interventions, 12, interventionduringmanyeducationalactivities, 163–172. inwhichthepresenceofguminthemouthmay Hanley, G. P., Iwata, B. A., & McCord, B. (2003). interfere with oral motor activities (i.e., those Functional analysis of problem behavior: A review. Journal ofAppliedBehavior Analysis, 36,147–186. that require vocal behavior). Although no Lindberg,J.S.,Iwata,B.A.,Roscoe,E.M.,Worsdell,A. formal data were collected to evaluate this S., & Hanley, G. P. (2003). Treatment efficacy of possible relation, a large proportion of Reggie’s noncontingent reinforcement during brief and ex- tended application. Journal of Applied Behavior educational activities that occurred during this Analysis,36,1–19. study required vocal responses on his part, and Lyons,E.A.,Rue,H.C.,Luiselli,J.K.,&DiGennaro,F. his therapists did not report any increased D.(2007).Brieffunctionalanalysisandsupplemental difficultieswiththesetasksduringgum-chewing feeding for postmeal rumination in children with developmentaldisabilities.JournalofAppliedBehavior sessions. Analysis,40,743–747. In conclusion, the primary contribution of Piazza, C.C., Fisher, W.W.,Hanley, G.P.,LeBlanc,L. this study is that a treatment for intractable A., Wordsell, A. S., Lindauer, S. E., et al. (1998). Treatment of pica through multiple analyses of its rumination that was novel, nonintrusive, and reinforcing functions. Journal of Applied Behavior highly effective for our participant was identi- Analysis,14,121–130. fied. Future researchers should continue to Rast,J.,Johnston,J.M.,Drum,C.,&Conrin,J.(1981). investigate treatments for rumination that are Therelationoffoodquantitytoruminationbehavior. Journal ofAppliedBehavior Analysis, 14, 121–130. practicalinreal-lifesettingsandshouldbeginto Repp,A.C.,Deitz,D.E.D.,Boles,S.M.,Deitz,S.M., identify the behavioral processes that occur & Repp, C. F. (1976). Differences among common when these interventions effect appreciable methods for calculating interobserver agreement. Journal ofAppliedBehavior Analysis, 9,109–113. changes in rumination. Sajwaj, T., Libet, J., & Agras, S. (1974). Lemon-juice therapy:Thecontroloflife-threateningruminationin a six-month-old infant. Journal of Applied Behavior REFERENCES Analysis,7,557–563. Chial,H.J.,Camilleri,M.,Williams,D.E.,Litzinger,K., & Perrault, J. (2003). Rumination syndrome in Received November 19,2007 children and adolescents: Diagnosis, treatment, and Final acceptanceMarch 6,2008 prognosis. Pediatrics,111,158–162. Action Editor,Gregory Hanley