fa of inappropiate mealtime behaviors

7/28/2019 FA of Inappropiate Mealtime Behaviors

1/18

187

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2003, 36, 187204 NUMBER 2 (SUMMER 2003)

FUNCTIONAL ANALYSIS OF INAPPROPRIATEMEALTIME BEHAVIORS

CATHLEEN C. PIAZZA AND WAYNE W. FISHERMARCUS AND KENNEDY KRIEGER INSTITUTES AND

JOHNS HOPKINS UNIVERSITY SCHOOL OF MEDICINE

KIMBERLY A. BROWN AND BRIDGET A. SHOREKENNEDY KRIEGER INSTITUTE AND


MEETA R. PATELMARCUS INSTITUTE AND

EMORY UNIVERSITY SCHOOL OF MEDICINE

RICHARD M. KATZ, BART M. SEVIN, AND CHARLES S. GULOTTAKENNEDY KRIEGER INSTITUTE AND


AND

AUDREY BLAKELY-SMITHKENNEDY KRIEGER INSTITUTE

The purpose of the current investigation was to apply the functional analysis describedby Iwata, Dorsey, Slifer, Bauman, and Richman (1982/1994) to the inappropriate meal-time behaviors of 15 children who had been referred to an intensive program for theassessment and treatment of severe feeding disorders. During Study 1, we conducteddescriptive assessments of children and parents during meals. The results of Study 1showed that parents used the following consequences for inappropriate mealtime behav-iors: coaxing and reprimanding, allowing the child to periodically take a break from oravoid eating, and giving the child preferred food or toys following inappropriate behavior.The effects of these consequences were tested systematically in Study 2 when we con-ducted analogue functional analyses with the children. During alternating meals, one ofthe consequences typically used by parents consistently followed inappropriate child be-havior. Results indicated that these consequences actually worsened behavior for 10 ofthe 15 children (67%). These results suggested that the analogue functional analysisdescribed by Iwata et al. may be useful in identifying the environmental events that playa role in feeding disorders.

DESCRIPTORS: descriptive assessment, feeding disorders, functional analysis, neg-ative reinforcement

A feeding disorder is identified when achild is unable or refuses to eat or drink suf-

This investigation was supported in part by Grant1 K24 HD01380-01 from the Department of Healthand Human Services, the National Institute of ChildHealth and Human Development.

Requests for reprints should be addressed to Cath-leen C. Piazza, Marcus Institute, 1920 Briarcliff Road,

Atlanta, Georgia 30329.

ficient quantities to maintain nutritional sta-tus, regardless of etiology (e.g., Babbitt,Hoch, & Coe, 1994; Budd et al., 1992).

The complications from feeding problemsrange from mild (e.g., missed meals) to se-vere (e.g., malnourishment, lack of growth,or failure to thrive; Polan et al., 1991). De-spite the potential seriousness of the prob-lem, little is known about the etiology of


2/18

188 CATHLEEN C. PIAZZA et al.

feeding problems or factors that maintainfeeding problems. Rather, the vast majorityof research on feeding problems has focusedon treatments to increase acceptance and de-

crease problematic mealtime behavior (e.g.,Ahearn, Kerwin, Eicher, Shantz, & Swear-ingin, 1996; Cooper et al., 1999; Hoch,Babbitt, Coe, Krell, & Hackbert, 1994; Kel-ley, Piazza, Fisher, & Oberdorff, 2003; Patel,Piazza, Kelly, Ochsner, & Santana, 2001;Patel, Piazza, Martinez, Volkert, & Santana,2002; Patel, Piazza, Santana, & Volkert,2002; Piazza et al., in press; Riordan, Iwata,Finney, Wohl, & Stanley, 1984; Riordan,Iwata, Wohl, & Finney, 1980). These studieshave shown that treatments based on oper-

ant consequences (e.g., escape extinction) areeffective for increasing consumption in chil-dren with feeding problems (Ahearn et al.;Cooper et al.; Hoch et al.; Patel et al., inpress).

Based on the success of operant conse-quences in the treatment of feeding prob-lems, a number of investigators (Ahearn etal., 1996; Cooper et al., 1999; Hoch et al.,1994) have hypothesized that feeding prob-lems are, at least in part, learned behaviorsthat develop as a result of a childs interac-

tions with the environment (e.g., throughnegative reinforcement, such as escape fromeating, or through positive reinforcement,such as attention or access to tangible items).Nevertheless, relatively little research on therole of environmental events on the main-tenance of feeding problems has been re-ported. Understanding the functional char-acteristics (i.e., reinforcing effects of environ-mental events) of feeding problems may behelpful in developing treatments that moreprecisely match the sources of reinforcement

that maintain the problem.Experimental functional analyses have

been used to quantify precisely the reinforc-ing functions of destructive behavior such asself-injury. For example, Iwata, Dorsey, Sli-fer, Bauman, and Richman (1982/1994) ex-

posed individuals to a series of conditions inwhich antecedent events and reinforcingconsequences for self-injurious behavior(SIB) were manipulated. The results sug-

gested that experimental functional analysiswas useful in identifying the environmentalcontexts in which SIB was likely and un-likely to occur. Functional analysis was orig-inally developed to assess environmental var-iables that influence SIB; however, thismethod has been used subsequently with a

wide array of behavior problems, includingaggression (Thompson, Fisher, Piazza, &Kuhn, 1998), elopement (Piazza et al.,1997), pica (Piazza, Hanley, & Fisher,1996), psychotic speech (Fisher, Piazza, &

Page, 1989; Mace & Lalli, 1991), and tan-trums (Vollmer, Northup, Ringdahl, Le-Blanc, & Chauvin, 1996).

Even though functional analysis has beenapplied to a wide array of behavior prob-lems, it has been used infrequently in theassessment of feeding problems. The ratio-nale for applying functional analysis to theassessment of feeding disorders is that re-gardless of the original etiology of a feedingdisorder, parents use a variety of consequenc-es to motivate their children to eat. For ex-

ample, children may display food refusal be-haviors such as crying, head turning, andbatting at the spoon. Faced with this exhi-bition, parents may terminate the meal and

wait for the child to calm down beforecontinuing. Parents also may provide in-creased attention following problematicmealtime behavior. For example, a parentmay coax the child to eat (e.g., you likethis, its good for you) or alter the quantityor quality of attention (e.g., play games withthe child such as here comes the airplane)

following food refusal. Some parents mayprovide the child with a more preferred food(e.g., peanut butter and jelly sandwich)

when the child refuses a less preferred food(e.g., brussel sprouts) so that the child eatssomething rather than nothing. Parents also


3/18

189INAPPROPRIATE MEALTIME BEHAVIORS

may provide toys during the meal to calmor distract the child. From a parental per-spective, consequences such as terminatingthe meal, coaxing, or distracting are logical

because they may produce the immediate ef-fect of temporarily stopping the undesirablebehavior (i.e., the parents behavior may bemaintained by negative reinforcement).However, from a functional perspective,such consequences can worsen mealtimeproblems over the long term if they functionas reinforcement.

The purpose of the current study was touse functional analysis to assess the problembehavior of children with pediatric feedingdisorders. The first goal of the study was to

conduct naturalistic assessments to develophypotheses about behavioral influences onmealtime behavior. The second goal was touse experimental functional analysis to eval-uate the effects of each of the consequencesused by the parents to determine if inappro-priate behavior improved or worsened dur-ing meals.

METHOD

Participants

Six children (Allison, Craig, Todd, Rob-ert, Tom, and Paul) and their parents par-ticipated in Study 1. Nine additional chil-dren participated in Study 2 (Maya, Nora,Peter, Sheila, Sally, Matt, Colin, Adam, andKyle), for a total of 15 children in Study 2.The participants were patients in a pediatricfeeding disorders program, and 1 child(Matt) was a patient on a neurobehavioralunit. The participants exhibited feeding

problems that resulted in failure to thrive,inadequate nutritional status, or severe be-havior problems at mealtimes that signifi-cantly interfered with food intake (partici-pants ages, sex, diagnoses, and specific prob-lem behaviors are listed in Table 1).

Evaluation of Medical Status

The participants underwent a thoroughinterdisciplinary evaluation to assess under-lying physical causes for the feeding prob-

lem. The medical team reviewed the partic-ipants history and conducted indicated di-agnostic studies. The team also assessed thesafety of oral feeding by ensuring that theparticipant had the appropriate physiologicalstatus and skills to swallow and prevent as-piration of ingested foods. In addition, par-ticipants nutritional and metabolic needs

were evaluated to maintain or promote ap-propriate weight gain for age and size. Whenthe interdisciplinary evaluation was com-pleted, the team made recommendations re-

garding all aspects of the childs feedingproblems that required treatment (e.g., phar-macological or surgical intervention, therapyto address oral motor issues, and nutritionalsupport).

Setting and Materials

All analyses were conducted in rooms (3m by 3 m) that were adjacent to roomsequipped with one-way observation andsound. Each child used age-appropriate seat-ing arrangements (e.g., high chair, booster

seat) and eating utensils. Children wereserved foods from the standard hospital traysfor children who ate foods at a pureed tex-ture or higher or a variety of jarred babyfoods for children whose diet was limited tobaby food. Food items were selected arbi-trarily (i.e., we rotated through a variety offruits, starches, vegetables, and meats) with-out consideration of the childs food prefer-ences.

Response Measurement and Data Collection

Parent measures. During Study 1, datawere collected on both parent and child be-haviors. Parent behaviors included the deliv-ery of escape, attention, and tangible items

within 10 s of an inappropriate behavior. Es-cape was scored when the parent removed


4/18


Table 1

Participant Descriptions

Participant Gender

Age(years

months) Diagnoses Inappropriate mealtime behaviors

Maya F 19 Failure to thrive (FTT), food allergies Batting, head turning, negative vocali-zations

Nora F 12 Gastroesophageal reflux (GER) Batting, negative vocalizationsPeter M 211 Cerebral palsy (CP), developmental

delay (DD), FTT, bilateral hearingloss

Batting, head turning, negative vocali-zations

Sheila F 18 DD, FTT, GER Batting, negative vocalizationsAllison F 26 Delayed gastric emptying, GER, FTT,

oral motor dysfunctionBatting, head turning, aggression,

throwing food, covering face, nega-tive vocalizations

Craig M 6 DD, seizure disorder, GER, micro-cephaly, reactive airway disease, hy-potonia, dysmorphism

Batting, head turning, negative vocali-zations, aggression, self-injury, handmouthing

Sally F 16 CP, DD, GER Batting, head turning, negative vocali-

zationsMatt M 7 Autism, attention deficit hyperactivity disorder, severe mental retardation

Batting, aggression, throwing food

Todd M 3 Spastic colon, chronic constipation,history of oral motor dysfunction

Batting, gagging, negative vocalizations

Colin M 23 FTT, GER Batting, head turning, negative vocali-zations

Robert M 4 CP, poor vision, autism Batting, head turning, negative vocali-zations

Tom M 13 CP, DD, GER, oral motor dysfunction Negative vocalizationsPaul M 3 DD, right hemiplegia, cortical blind-

ness, hydrocephalus, microcephalyBatting, head turning, negative vocali-

zationsKyle M 22 FTT, brain cyst, cortical blindness, hy-

potonia, DDHead turning, negative vocalizations

Adam M 110 GER, FTT, reactive airway disease,asthma

Batting, head turning, negative vocali-zations

the bite presentation or terminated the meal.Attention was scored when the parent deliv-ered attention in the form of reprimands(dont do that), verbal or physical redirec-tion (look at daddy, hes eating his peas),or coaxing (you like green beans). Deliveryof a tangible item was scored when the par-ent gave the child a preferred food or a toy.Observers indicated the occurrence or non-

occurrence of each type of parent responsefollowing the occurrence of each inappro-priate child behavior. Data for parent re-sponses were calculated by dividing thenumber of consequences following inappro-priate behaviors by the number of inappro-

priate behaviors multiplied by 100%. Forexample, if the child engaged in 10 inappro-priate behaviors during the meal and theparent responded by providing attention fol-lowing five of those behaviors, then the levelof attention delivery was 50%. The numberof meals observed varied for each participant(Allison, 3; Craig, 2; Todd, 2; Robert, 4;Tom, 1; Paul, 4).

Child measures. Data were collected onboth appropriate and inappropriate meal-time behaviors for all children during Stud-ies 1 and 2. The appropriate behavior wasacceptance of food (the child accepting thebite or drink into his or her mouth within


5/18


5 s of the bite or drink presentation). Dataalso were collected for expelling food, gag-ging, and vomiting. Inappropriate mealtimebehaviorsdiffered for each child and are list-

ed in Table 1; however, each child displayedat least one of the following inappropriatemealtime behaviors: head turning, batting atthe spoon, throwing food or utensils, out ofseat, negative vocalizations, self-injury, or ag-gression.

Expulsions were defined as emitting foodor liquid larger than the size of a pea pastthe plane of the lips. Gaggingwas defined asretching via movement of stomach, chestand mouth, with or without audible noises.Vomitingwas defined as emitting contents of

the esophagus or stomach, consisting of pre-viously digested food or liquid, past theplane of the lips. Head turning was definedas moving the head away from midline afterspoon presentation. Battingat the spoon wasdefined as the child pushing the spoon orthe therapists hand away from his or hermouth. Throwing food or utensilswas definedas the child throwing food or utensils acrossthe room, at people, or onto the floor. Outof seat was defined as the childs bottom nottouching the base of the chair for 3 s or

more. Negative vocalizations were defined asscreaming, crying, or making negative state-ments (e.g., I dont want it, no) abovenormal conversational levels that lasted for 3s or more. Self-injurywas defined as behaviordirected towards the self that caused or hadthe potential to cause tissue damage (e.g.,head banging, hand biting). Aggression wasdefined as behavior directed towards othersthat caused or had the potential to cause tis-sue damage (e.g., hitting, kicking, biting,scratching).

Data on child target responses were col-lected via event or frequency recording usinga paper-and-pencil data-collection method.

Acceptances, expulsions, gagging, vomiting,and inappropriate behaviors for Maya andTom were collected via event recording (i.e.,

either the behavior occurred or did not oc-cur during each bite presentation) and wererecorded once for each bite presentation.Data for event-recording measures were con-

verted to percentage of trials by dividing thenumber of occurrences of the behavior bythe number of bite presentations multipliedby 100%. Inappropriate behaviors for allother participants were recorded as the fre-quency of occurrence. Data for inappropri-ate behaviors then were converted to a rateby dividing the total number of responses bythe total number of minutes per session.

Interrater Agreement

Interrater agreement was obtained by hav-

ing two observers record data simultaneouslybut independently on a mean of 47% and32% of the sessions during Studies 1 and 2,respectively. Interrater agreement for appro-priate and inappropriate behaviors recordedas event measures was calculated by dividingthe number of agreements plus disagree-ments and multiplying by 100%. An agree-ment was defined as two observers agreeingthat the behavior either occurred or did notoccur, and a disagreement was defined as

one observer scoring the occurrence of thebehavior and one observer not scoring theoccurrence of the behavior. For inappropri-ate behaviors recorded as frequency mea-sures, the total number of inappropriate be-haviors per bite presentation was comparedby dividing the smaller frequency by thelarger frequency and multiplying by 100%.The mean agreement coefficient for parentbehaviors for all participants in Study 1 was97% (range, 92% to 100%). The meanagreement coefficient for inappropriate child

behaviors for all participants was 95%(range, 87% to 100%) during Study 1 and92% (range, 76% to 100%) during Study 2.The mean agreement coefficient for accep-tances was 98% (range, 91% to 100%) dur-ing Study 2.


6/18


Figure 1. Mean percentage of consequence delivery for Allison, Craig, Todd, Robert, Tom, and Paul.

STUDY 1: OBSERVATIONS OFMEALS FED BY PARENTS

Procedure

Parents were observed feeding their chil-

dren in rooms at the program site. Theywere asked to feed the child as they normallywould at home and to use similar materials(e.g., foods, utensils, toys). They also wereasked to respond to the childs behaviors asthey typically did at home. Prior to theseobservations, we asked parents to describethe procedures they used at home duringmeals. Specific tangible items were availablein the room if the parents reported provid-ing such items to the child during meals.Meals varied in length and ended when theparent terminated the meal (i.e., as they

would do at home).

Results

The results of Study 1 are depicted in Fig-ure 1. Each parent used a variety of conse-

quences during meals. All parents providedattention in the form of reprimands (e.g.,dont throw food), soothing comments(e.g., its okay, dont cry), or coaxing (e.g.,

you like this) when their children dis-played inappropriate behavior. All parentsremoved bites of food (allowed the child totake a break or escape eating) following in-appropriate behavior. Three of the 6 parentsgave their child a tangible item (i.e., a pre-ferred food, drink, toys) when the child en-gaged in inappropriate behavior.

STUDY 2: FUNCTIONAL ANALYSIS

The results of Study 1 suggested that par-ents used a variety of potentially functional

consequences when their child engaged ininappropriate behavior during mealtime. Be-cause the consequences differed from parentto parent and because all parents used morethan one consequence when their child en-gaged in inappropriate behavior, it was not


7/18


possible to determine the effects of any oneconsequence on inappropriate behavior inStudy 1. Therefore, the individual effects ofeach consequence (escape, tangible, atten-

tion) on inappropriate behavior were evalu-ated systematically in Study 2.

General Procedure

Each session was 10 min long, and bitesof food were presented once every 30 s. Dur-ing each presentation, a spoon was held 2.5cm from the childs lips and the child wasinstructed to take a bite. Trained therapistsfed all meals. One food item from each offour food groups (protein, starch, fruit, veg-etable) was offered during each session, and

bite presentations rotated randomly amongthe four food groups. Selection of food tex-ture and bite size was based on the childsskills, previous experiences with eating, andthe recommendations of the occupationaltherapist. Three to four session blocks wereconducted each day, with two to three 10-min functional analysis sessions (20 to 30min of total eating time) per session block(for a total of 6 to 12 sessions per day). Briefbreaks (5 to 10 min) were provided betweeneach session, during which the child was al-

lowed to leave the therapy room and play.Session blocks were spaced approximately 2to 3 hr apart (e.g., 8:30 a.m., 11:30 a.m., 2:00 p.m., 4:00 p.m.). Acceptance of bites offood resulted in brief praise (e.g., good eat-ing) across all experimental and baselineconditions and presentation of another biteof food at the next 30-s interval. Inappro-priate behaviors resulted in one of threetypes of consequences (negative reinforce-ment, positive reinforcement in the form ofattention, or positive reinforcement in the

form of tangible items), depending on thecondition in effect.

Specific Procedure

Baseline (play). Toys were available on thetray of the high chair or the table, and the

therapist interacted (e.g., engaged in conver-sation) with the child throughout the ses-sion. If the child engaged in inappropriatebehavior, the therapist did not provide a dif-

ferential consequence (i.e., the therapist con-tinued interacting with the child as if thebehavior did not occur). The spoon re-mained 2.5 cm from the childs lips follow-ing inappropriate behavior and for the du-ration of the 30-s interval. At the end of the30-s interval, the therapist removed thespoon and presented a new bite of food tothe child. The purpose of this condition wasto observe the frequency of inappropriatebehavior when the child had free access toattention and preferred items. Low levels of

inappropriate behavior in this conditionwould indicate that the child was not mo-tivated to engage in inappropriate behavior

when preferred interaction and items wereavailable continuously.

Escape. If the child engaged in an inap-propriate behavior, the therapist removedthe spoon of food for the remainder of the30-s interval. Another bite of food was pre-sented at the next 30-s interval. The thera-pist did not provide any other differentialconsequence following the childs inappro-

priate behavior (e.g., the therapist did notreprimand the child). Toys were not avail-able during these sessions. The purpose ofthis condition was to simulate a situation in

which a parent removed the bite (i.e., al-lowed escape) when the child engaged in in-appropriate behavior. High levels of inappro-priate behavior in this condition would sug-gest that the childs behavior was sensitive toescape as reinforcement.

Attention. Inappropriate behavior resultedin brief (5 to 10 s) attention, in the form of

coaxing (e.g., you like this) and statementsof concern (e.g., dont cry, youll be okay).The spoon remained 2.5 cm from the childslips following inappropriate behavior and forthe duration of the 30-s interval. At the endof the 30-s interval, the therapist removed


8/18


the spoon and presented a new bite of foodto the child. Toys were not available duringthese sessions. The purpose of this condition

was to simulate a situation in which the par-

ent provided attention or coaxing to moti-vate the child to eat. High levels of inappro-priate behavior in this condition would sug-gest that the childs behavior was sensitive topositive reinforcement in the form of atten-tion.

Tangible. Inappropriate behavior resultedin the presentation of a tangible item (pre-ferred toys, foods, or drinks) for the remain-der of the 30-s interval. The therapist placedpreferred toys or edible items on the tray andallowed the child to interact with the toys

or consume the edible items. However, thespoon with the original bite of food re-mained 2.5 cm from the childs lips follow-ing inappropriate behavior and for the du-ration of the 30-s interval. The therapist didnot provide any other differential conse-quence following inappropriate behavior anddid not interact with the child if the childhad a toy. At the end of the 30-s interval,the therapist removed the spoon and thetangible item and presented a new bite offood. The tangible items were selected based

on direct observations and parent report ofthe items they presented to their child whenthe child did not eat. Tangible items for in-dividual children included chocolate icecream, bacon, toys, milk, and television. Thepurpose of this condition was to simulate asituation in which parents attempted to mo-tivate their child to eat by providing them

with preferred items, activities, or food.High levels of inappropriate behavior in thiscondition would suggest that the childs be-havior was sensitive to positive reinforce-

ment in the form of the tangible item.

Design and Independent Variables

We used a reversal design in which eachtest condition was compared to the baseline(play) condition. All analyses began with the

baseline condition. The order of test condi-tions was selected randomly. The test con-ditions with the highest rates of inappropri-ate mealtime behaviors were repeated to es-

tablish experimental control.

RESULTS

The means and standard deviations ofmeasures of inappropriate behaviors for thefour conditions across all participants wereattention (M 5.04, SD 4.7), escape (M 3.14, SD 1.47), tangible (M 5.42,SD 4.7), and play (M 0.7, SD 0.52).The results of the functional analyses appearin Figure 2 for Maya and Nora; in Figure 3

for Peter, Sheila, and Allison; in Figure 4 forCraig, Sally, and Matt; in Figure 5 for Todd,Colin, and Robert; and in Figure 6 for Tom.Three participants (data not shown) dis-played low to zero rates of inappropriate be-haviors across all conditions.

Ten (Maya, Nora, Peter, Sheila, Allison,Craig, Sally, Matt, Todd, and Tom) of the 15participants (67%) displayed high levels ofinappropriate behavior during one or moreof the test conditions relative to the baselinecondition, suggesting that environmental var-

iables played a role in the childs feedingproblem. Inappropriate behavior was high inthe escape condition for 9 (60%) of the par-ticipants (Maya, Nora, Peter, Sheila, Allison,Craig, Sally, Todd, and Tom). Inappropriatebehavior was high in the attention conditionfor 8 (53%) of the participants (Maya, Peter,Sheila, Allison, Craig, Sally, Matt, and Tom)and high in the tangible condition for 2 par-ticipants (Todd and Tom).

Tom (Figure 6) displayed two types of be-haviors, inappropriate mealtime behavior

(top panel) and expulsion (bottom panel).Expulsions occurred across all conditions,

whereas inappropriate mealtime behavior oc-curred at higher levels during the test con-ditions (demand, attention, and tangible).These results suggested that Toms inappro-


9/18


Figure 2. Percentage of intervals of inappropriate mealtime behaviors for Maya and inappropriate mealtimebehaviors per minute for Nora.

priate mealtime behaviors were influencedby environmental factors (i.e., attention, es-cape, and access to tangible items), whereashis expulsion of food was not.

Table 2 depicts the results of the natural-istic observations (the consequences that theparents used following inappropriate meal-time behavior), the analogue functionalanalysis (the reinforcers that maintained in-appropriate mealtime behavior), and thetreatment outcomes based on percentages ofacceptance in baseline and the treatment in

use at discharge. The results of the natural-istic observations and the analogue function-al analyses yielded identical results for 3 ofthe 6 participants (Allison, Craig, and Tom).Todds parents provided escape, attention,and tangible items. However, the results of

his analogue functional analysis suggestedthat his inappropriate behavior was sensitiveto negative reinforcement (escape) and pos-itive reinforcement (tangible items), but notattention. Roberts parents provided escapeand attention following inappropriate meal-time behaviors; however, Roberts analoguefunctional analysis was undifferentiated (i.e.,his inappropriate behavior was not differ-entially sensitive to escape or attention as re-inforcement). Pauls parents provided escape,attention, and a tangible item following in-

appropriate mealtime behaviors; however,Paul (data not shown) did not exhibit in-appropriate behavior in the analogue func-tional analysis (i.e., his inappropriate behav-ior was not differentially sensitive to escape,attention, or tangible items as reinforce-


10/18


Figure 3. Inappropriate mealtime behaviors per minute for Peter, Sheila, and Allison.

ment). Treatment resulted in increases in ac-ceptance above 80% for all participants.

We also evaluated the extent to which therecommended treatments matched the re-sults of the functional analysis. Positive re-inforcement in the form of attention and ex-

tinction for attention (ignoring) were usedin treatment for 88% and 100%, respective-ly, of participants who demonstrated sensi-tivity to attention as reinforcement. Negativereinforcement in the form of escape and es-cape extinction (either nonremoval of the


11/18


Figure 4. Inappropriate mealtime behaviors per minute for Craig, Sally, and Matt.

spoon or physical guidance) was used for

11% and 67%, respectively, of participantswho demonstrated sensitivity to escape as re-inforcement. Positive reinforcement in theform of tangible items and extinction for be-haviors maintained by tangible items wereused in treatment for 100% of participants

who demonstrated sensitivity to tangible

items as reinforcement.

DISCUSSION

The results of the current investigationsuggested that environmental variables play


12/18


Figure 5. Inappropriate mealtime behaviors per minute for Todd, Colin, and Robert.

a role in the occurrence of feeding disorders.The results of Study 1 showed that parents

provided a variety of consequences whentheir children engaged in inappropriate be-havior that may be functionally related tothat behavior. In the natural environment, it

would be very difficult (if not impossible) toevaluate which combination of consequences

(if any) affect behavior (Lerman & Iwata,1993). Therefore, the functional analysis

conducted in Study 2 tested the extent towhich each of the consequences functionedas reinforcement for inappropriate behaviorfor individual children. These data suggestedthat functional analysis could be used toidentify the reinforcers for feeding problems.


13/18


Figure 6. Percentage of intervals of negative vocalizations (top panel) and expulsions (bottom panel) forTom.

One important contribution of functionalanalysis is that it has provided behavior an-alysts with a tool with which to improve ourunderstanding of the influences of operantmechanisms on behavior disorders (Fisher,DeLeon, & Kuhn, 2000). For example,Carrs (1977) hypotheses that self-injuriousbehavior could be reinforced by positive re-inforcement (attention), negative reinforce-ment (escape from tasks), or by the sensory

consequences automatically produced by theresponse largely have been validated throughepidemiological investigations using func-tional analysis (Derby et al., 1992; Iwata,Pace, Dorsey, et al., 1994). Similarly, thecurrent results provide partial support for

the hypothesis proposed by a number of in-vestigators that feeding problems are main-tained by negative reinforcement (Ahearn etal., 1996; Cooper et al., 1999; Hoch et al.,1994; Patel et al., 2002). Specifically, of the10 children who demonstrated sensitivity toone or more of the reinforcers tested duringthe functional analysis, 90% demonstratedsensitivity to escape as reinforcement, sug-gesting that negative reinforcement contrib-

utes to the maintenance of feeding problemsfor many children.

Although the current results documentthe role of negative reinforcement in themaintenance of feeding problems, perhapsthe more surprising and interesting finding


14/18


Table 2

Results from Naturalistic Observation, Analogue Functional Analysis, and Treatment

Participant Naturalistic observations Functional analysis

% Acceptance

Baseline Treatment

MayaNoraPeterSheila

AllisonCraig

Escape and attentionEscape and attention

Escape and attentionEscapeEscape and attentionEscape and attentionEscape and attentionEscape and attention

180

60006

1009992

1009995

SallyMattToddColinRobert

Escape, attention, and tangible

Escape and attention

Escape and attentionAttentionEscape and tangibleNo differentiationNo differentiation

040104938

968797a

9996

Tom Escape, attention, and tangible Escape, attention, and tangible 81 99Paul Escape, attention, and tangible No behaviors 81 b

a Data for Todd reflect the percentage of bites he consumed (number of bites swallowed divided by number of bitespresented).

b No programmed treatment was implemented for Paul because he consumed food in baseline.

was that positive reinforcement contributedto the maintenance of inappropriate meal-time behavior in over half the cases. In ad-dition, tangible items functioned as rein-forcement for 13% of the children. The re-sults of the naturalistic observations suggest-ed that parents provide attention (in theform of coaxing) or tangible items (in theform of toys or preferred foods) following

inappropriate mealtime behavior. The resultsof the experimental functional analysis sug-gested that for some children, these conse-quences function as reinforcement for inap-propriate behavior.

A second important contribution of func-tional analysis has been to provide assess-ment data that categorize aberrant behavioraccording to its functional rather than its to-pographical properties. Previous studies haveshown that topographically dissimilar re-sponses can be members of a common func-

tional response class (e.g., Lalli, Mace,Wohn, & Livezey, 1995) and that the sametopographical response (e.g., aggression) canbelong to multiple functional response clas-ses (e.g., Thompson et al., 1998). The cur-rent results suggest that inappropriate meal-

time behavior may similarly be multiplycontrolled. In addition, functional analysisresults may help to distinguish between be-haviors associated with a lack of motivationto eat versus other factors (e.g., skill deficits).For example, Tom displayed two types ofproblems during meals, expulsions and in-appropriate behavior. Expulsions occurredacross all conditions, whereas inappropriate

behavior occurred at higher levels whenthese behaviors produced attention, toys, orescape from the bite. Thus, the results fromToms analysis suggested that his inappropri-ate behavior was influenced by the tested en-vironmental factors (i.e., attention, escape,and access to tangible items), whereas his ex-pulsion of food was not.

A third important contribution of func-tional analysis has been the development ofhighly specific and effective interventionsthat directly address the function or func-

tions that maintain the problem behavior(e.g., Iwata, Pace, Cowdery, & Miltenberger,1994). For example, escape extinction wouldbe an indicated treatment and time-out

would be contraindicated for inappropriatemealtime behavior maintained by negative


15/18


reinforcement. By contrast, time-out couldbe an appropriate treatment for children

who engage in inappropriate behavior thatis functionally related to attention. Treat-

ments also may be developed that targetboth motivational and skill deficits. For ex-ample, Toms treatment consisted of a func-tion-based component (noncontingent rein-forcement with attention and toys) and acomponent to address his oral motor deficits(recommended by the occupational thera-pist) to facilitate swallowing.

Overall, a function-based treatment wasdeveloped for 8 of the 10 participants whosefunctional analysis results were differentiat-ed. At least one of the functional reinforcers

(escape, attention, tangible items) was usedin a reinforcement-based treatment (i.e., ei-ther differential positive, differential nega-tive, or noncontingent reinforcement) for100% of the participants. An escape extinc-tion procedure was used with 6 of the 9(67%) participants for whom escape wasidentified as a reinforcer for inappropriatemealtime behavior. Inappropriate behavior

was ignored for 100% of participants whodemonstrated sensitivity to attention as re-inforcement. All of the participants who re-

ceived function-based treatments met theirtreatment goals for oral intake. However, theunique contribution of the functional anal-ysis results could not be isolated in most cas-es because a multicomponent treatment ap-proach was implemented (e.g., differentialreinforcement plus escape extinction).

Although the results of the current inves-tigation suggested that functional analysishas a number of advantages for understand-ing, assessing, and treating feeding problems,the study also has a number of limitations,

which may have implications for functionalanalysis of feeding problems. One limitation

was that the functional analyses conductedin the current investigation were somewhatartificial because (a) the events that occurredduring the meal sessions were controlled

carefully, (b) the meals were conducted bytrained therapists rather than parents, and(c) the meals were conducted in a therapyroom rather than at home or in other nat-

ural settings. In addition, the analyses usedhere were carried out with children whowere referred to an intensive, hospital-basedprogram for the assessment and treatment offeeding disorders. Thus, our results may notbe representative of a broader sample of chil-dren with feeding disorders. However, theresults suggest that under controlled condi-tions, the effects of environmental events onfeeding problems can be assessed using func-tional analysis.

In summary, the literature on functional

analysis unambiguously demonstrates that itis a powerful technology that can identifyenvironmental influences on a broad rangeof problem behaviors. In the current study,the range of functional analysis was extendedto potentially life-threatening problems suchas food refusal and failure to thrive. Despitethe limitations of our study, the extent andseriousness of these feeding problems, thelimitations of organismic accounts (e.g.,medical, skill deficits), the clear results weobtained provide a strong rationale for con-

tinuing investigations of this sort.The results raise a number of important

questions about the use of functional anal-ysis in the assessment and treatment of pe-diatric feeding problems that should be thefocus of future investigations. First, the re-sults of the current study suggest that par-ents use a variety of consequences when theirchildren exhibit feeding problems. Future re-search should focus on observing children atrisk for the development of feeding prob-lems to understand how parental responses

to child behavior at meals affect the devel-opment of feeding problems. Second, the re-sults of the current investigation suggest thatnegative reinforcement plays a significantrole in the maintenance of feeding problems.Therefore, future studies should focus on


16/18


understanding the role of escape extinctionin the treatment of feeding problems. Forexample, under what conditions is escape ex-tinction effective in treatment, and do dif-

ferential or noncontingent reinforcementprocedures contribute to treatment effects?Several procedures have been conceptualizedas escape extinction (Ahearn et al., 1996;Coe et al., 1997; Cooper et al., 1999), butit is not clear which procedures constituteescape extinction for individual children.

Also, as indicated by the literature onfeeding problems taken as a whole, there isa continuum of seriousness. The problemsaddressed here were on the high end, as in-dicated by the intensiveness of the treatment

sought by parents (e.g., hospital admission).Whether functional analysis would be asuseful with problems on the lower end ofthe continuum (e.g., picky eating) is un-known, yet the literature on functional anal-ysis suggests that it might be. At first, func-tional analysis was used with life-threateningbehavior problems (e.g., SIB; Iwata et al.,1982/1994). After successes with these prob-lems were reported, functional analysis tech-nology was applied with less intense prob-lems (e.g., noncompliance; Northup et al.,

1994) and similar levels of success were re-ported. Thus it seems plausible that use offunctional analysis with a broader range offeeding problems, beginning with seriousproblems as in this study and migrating toless serious problems in studies to come,

would follow a similar trajectory of success.In conclusion, our results underscore andadd to the importance of the role functionalanalysis can play in the assessment of prob-lematic behavior, and they supply supportfor expanding investigations into a new

realm, feeding problems.

REFERENCES

Ahearn, W. H., Kerwin, M. E., Eicher, P. S., Shantz,J., & Swearingin, W. (1996). An alternating

treatments comparison of two intensive interven-tions for food refusal. Journal of Applied Behavior

Analysis, 29, 321332.Babbitt, R. L., Hoch, T. A., & Coe, D. A. (1994).

Behavioral feeding disorders. In D. Tuchman &R. Walters (Eds.), Pediatric feeding and swallowing

disorders: Pathology, diagnosis, and treatment. SanDiego, CA: Singular Publishing Group.

Budd, K. S., McGraw, T. E., Farbisz, R., Murphy, T.B., Hawkins, D., Heilman, N., et al. (1992). Psy-chosocial concomitants of childrens feeding dis-orders. Journal of Pediatric Psychology, 17, 8194.

Carr, T. E. (1977). The motivation of self-injuriousbehavior: A review of some hypotheses. Psycholog-ical Bulletin, 84, 800816.

Coe, D. A., Babbitt, R. L., Williams, K. E., Hajmi-halis, C., Snyder, A. M., Ballard, C., et al. (1997).Use of extinction and reinforcement to increasefood consumption and reduce expulsion. Journalof Applied Behavior Analysis, 30, 581583.

Cooper, L. J., Wacker, D. P., Brown, K., McComas,

J. J., Peck, S. M., Drew, J., et al. (1999). Use ofa concurrent operants paradigm to evaluate posi-tive reinforcers during treatment of food refusal.Behavior Modification, 23, 340.

Derby, K. M., Wacker, D. P., Sasso, G., Steege, M.,Northup, J., Cigrand, K., et al. (1992). Brieffunctional assessment techniques to evaluate ab-errant behavior in an outpatient setting: A sum-mary of 79 cases. Journal of Applied Behavior Anal-

ysis, 25, 713721.Fisher, W. W., DeLeon, I. G., & Kuhn, D. E. (2000).

Functional analysis of precurrent contingenciesbetween mands and destructive behavior. In J. C.Leslie & D. Blackman (Eds.), Issues in the exper-imental and applied analyses of human behavior

(pp. 212234). Reno, NV: Context Press.Fisher, W. W., Piazza, C. C., & Page, T. (1989). As-sessing independent and interactional effects ofbehavioral and pharmacologic interventions for aclient with dual diagnoses. Journal of BehaviorTherapy and Experimental Psychiatry, 20, 241250.

Hoch, T. A., Babbitt, R. L., Coe, D. A., Krell, D. M.,& Hackbert, L. (1994). Contingency contacting:Combining positive reinforcement and escape ex-tinction procedures to treat persistent food refusal.Behavior Modification, 18, 106128.

Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K.E., & Richman, G. S. (1994). Toward a func-tional analysis of self-injury. Journal of Applied Be-

havior Analysis, 27, 197209. (Reprinted fromAnalysis and Intervention in Developmental Dis-abilities, 2, 320, 1982)

Iwata, B. A., Pace, G. M., Cowdery, G. E., & Mil-tenberger, R. G. (1994). What makes extinction

work: An analysis of procedural form and func-tion. Journal of Applied Behavior Analysis, 27,131144.


17/18


Iwata, B. A., Pace, G. M., Dorsey, M. F., Zarcone, J.R., Vollmer, T. R., Smith, R. G., et al. (1994).The functions of self-injurious behavior: An ex-perimental-epidemiological analysis. Journal of Ap-

plied Behavior Analysis, 27, 215240.Kelley, M. E., Piazza, C. C., Fisher, W. W., & Ober-

dorff, A. J. (2003). Acquisition of cup drinkingusing previously refused foods as positive and neg-ative reinforcement. Journal of Applied Behavior

Analysis, 36, 8993.Lalli, J. S., Mace, F. C., Wohn, T., & Livezey, K.

(1995). Identification and modification of a re-sponse-class hierarchy. Journal of Applied Behavior

Analysis, 28, 551559.Lerman, D. C., & Iwata, B. A. (1993). Descriptive

and experimental analyses of variables maintainingself-injurious behavior. Journal of Applied Behavior

Analysis, 26, 293319.Mace, F. C., & Lalli, J. S. (1991). Linking descriptive

and experimental analyses in the treatment of bi-zarre speech. Journal of Applied Behavior Analysis,

24, 553562.Northup, J., Wacker, D. P., Berg, W. K., Kelly, L.,Sasso, G., & DeRaad, A. (1994). The treatmentof severe behavior problems in school settings us-ing a technical assistance model. Journal of AppliedBehavior Analysis, 27, 3347.

Patel, M. R., Piazza, C. C., Kelly, M. L., Ochsner, C.A., & Santana, C. M. (2001). Using a fadingprocedure to increase fluid consumption in a child

with feeding problems. Journal of Applied BehaviorAnalysis, 34, 357360.

Patel, M. R., Piazza, C. C., Martinez, C. J., Volkert,V. M., & Santana, C. M. (2002). An evaluationof two differential reinforcement procedures withescape extinction to treat food refusal. Journal of

Applied Behavior Analysis, 35, 183186.Patel, M. R., Piazza, C. C., Santana, C. M., & Volk-

ert, V. M. (2002). An evaluation of food typeand texture in the treatment of a feeding problem.

Journal of Applied Behavior Analysis, 35, 363374.Piazza, C. C., Hanley, G. P., Bowman, L. B., Ruyter,

J. M., Lindauer, S. E., & Saiontz, D. M. (1997).Functional analysis and treatment of elopement.

Journal of Applied Behavior Analysis, 30, 653672.Piazza, C. C., Hanley, G. P., & Fisher, W. W. (1996).

Functional analysis and treatment of cigarettepica. Journal of Applied Behavior Analysis, 29,

437450.Piazza, C. C., Patel, M. R., Santana, C. M., Goh, H.,

Delia, M. D., & Lancaster, B. M. (in press). Anevaluation of simultaneous and sequential presen-tation of preferred and nonpreferred food to treatfood selectivity. Journal of Applied Behavior Anal-

ysis.Polan, H. J., Kaplan, M. D., Kessler, D. B., Shindle-

decker, R., Newmark, M., Stern, D. N., et al.(1991). Psychopathology in mothers of children

with failure to thrive. Infant Mental Health Jour-nal, 12, 5564.

Riordan, M. M., Iwata, B. A., Finney, J. W., Wohl,M. K., & Stanley, A. E. (1984). Behavioral as-sessment and treatment of chronic food refusal in

handicapped children. Journal of Applied BehaviorAnalysis, 17, 327341.Riordan, M. M., Iwata, B. A., Wohl, M. K., & Finney,

J. W. (1980). Behavioral treatment of food refusaland selectivity in developmentally disabled chil-dren. Applied Research in Mental Retardation, 1,95112.

Thompson, R. H., Fisher, W. W., Piazza, C. C., &Kuhn, D. E. (1998). The evaluation of aggres-sion maintained by attention and automatic re-inforcement. Journal of Applied Behavior Analysis,31, 103116.

Vollmer, T. R., Northup, J., Ringdahl, J. E., LeBlanc,L. A., & Chauvin, T. M. (1996). Functionalanalysis of severe tantrums displayed by children

with language delays: An outclinic assessment. Be-havior Modification, 20, 97115.

Received November 13, 2001Final acceptance February 18, 2003

Action Editor, Patrick Friman

STUDY QUESTIONS

1. Why might it be helpful to conduct a functional analysis of childrens inappropriate mealtimebehavior?

2. Briefly describe the two categories of dependent variables measured in Study 1.

3. What establishing operations and consequences were manipulated in each of the functionalanalysis conditions?


18/18


4. In Study 2, what might have accounted for the gradual decreases in inappropriate behaviorsobserved during baseline for 8 of the 12 participants?

5. Briefly summarize the results of Study 2. Provide an alternative explanation for the high

rates of inappropriate behavior observed during both the attention and tangible conditions.

6. Results of the naturalistic observations (Study 1) revealed that parents provided a variety ofconsequences for problem behavior observed during feeding situations. Why was it importantto conduct subsequent functional analyses?

7. How might one determine whether inappropriate behaviors that were high in both the escapeand attention conditions during Study 2 were actually multiply maintained?

8. Describe several ways in which the results of this study contribute to the literature on theassessment and treatment of feeding problems.

Questions prepared by David Wilson and Carrie Dempsey, The University of Florida

fa of inappropiate mealtime behaviors

Documents