This study was conducted at the May Center for Child Development, Randolph, Massachusetts. We appreciate the skilled participation of Adela Scancarova.
FUNCTIONAL ANALYSIS AND INTERVENTION FOR CHRONIC RUMINATION
Article first published online: 1 APR 2013
© Society for the Experimental Analysis of Behavior
Journal of Applied Behavior Analysis
Special Issue: Special Issue on Functional Analysis: Commemorating Thirty Years of Research and Practice
Volume 46, Issue 1, pages 328–332, Spring 2013
How to Cite
Woods, K. E., Luiselli, J. K. and Tomassone, S. (2013), FUNCTIONAL ANALYSIS AND INTERVENTION FOR CHRONIC RUMINATION. Jnl of Applied Behav Analysis, 46: 328–332. doi: 10.1002/jaba.24
- Issue published online: 1 APR 2013
- Article first published online: 1 APR 2013
- Manuscript Accepted: 17 DEC 2012
- Manuscript Received: 16 MAY 2012
- functional analysis;
- automatic reinforcement;
- noncontingent reinforcement
We conducted a functional analysis and treatment evaluation of chronic rumination in a 19-year-old man with intellectual disabilities. Outcomes of the functional analysis suggested that rumination was maintained by automatic reinforcement. Results of the intervention evaluation suggested that (a) noncontingent access to food after meals reduced rumination more effectively than did noncontingent access to inedible stimuli, (b) a particular type of food was associated with lower levels of rumination than other types of food, and (c) both presession and continuous access to food reduced levels of rumination more effectively than did fixed-time access to food.
Rumination is defined by repeated regurgitating, chewing, and reswallowing previously ingested food. Serious medical problems can result from chronic rumination, such as electrolyte imbalance, malnutrition, weight loss, esophagitis, respiratory complications, and decreased resistance to disease (Boris & Dalton, 2007). Although the earliest behavioral intervention research on rumination featured punishment procedures (Sajwaj, Libet, & Agras, 1974), later studies evaluated the effects of less restrictive intervention by manipulating food quantity and consumption (Rast, Johnston, Drum, & Conrin, 1981), supplemental feedings (Thibadeau, Blew, Reedy, & Luiselli, 1999), and noncontingent access to food or liquids (Kenzer & Wallace, 2007; Kliebert & Tiger, 2011). Despite the importance of identifying maintaining reinforcers to inform intervention, relatively few published studies on rumination included functional analyses (Kenzer & Wallace, 2007; Lyons, Rue, Luiselli, & DiGennaro, 2007; Wilder, Register, Register, Bajagic, & Neidert, 2009). Accordingly, the present study featured (a) a functional analysis of rumination that suggested automatic reinforcement as the source of control and (b) a subsequent evaluation of stimuli that were intended to compete with rumination.
Participant and Setting
Aaron was a 19-year-old man who had been diagnosed with congenital encephalopathy, global developmental delay, gastroesophageal reflux, and intellectual disability. He did not speak but communicated through gestures, body language, vocalizations, and a picture-point communication book. Aaron had displayed rumination for many years. Although he was previously a participant in earlier research (Lyons et al., 2007), he required further intervention because the problem had reemerged. He weighed 54 kg (3rd percentile) and was 175 cm tall.
A classroom teacher with several years of experience conducted functional analysis and intervention sessions with Aaron each weekday. Only the teacher and observers were present during sessions, which were conducted in a partitioned area of Aaron's classroom at a private residential school. Before the start of the study, the first author trained the teacher to implement functional analysis and intervention procedures using instructions, written guidelines, demonstration, rehearsal, and performance feedback (Moore et al., 2002).
Response Measurement and Interobserver Agreement
Rumination was defined as any instance of gurgling (upward movement of throat), swishing (side-to-side mouth movement with lips closed), or gagging (retching sound) with or without puffed cheeks or drooling. An observer collected partial-interval rumination data during 5-min functional analyses (three conducted per day) and treatment sessions (one conducted per day) that were divided into consecutive 15-s intervals. Aaron typically ruminated immediately after meals, so we began sessions approximately 2 min after he consumed his lunch in the school's cafeteria. Interobserver agreement was assessed during 60% of functional analysis sessions and 27% of intervention evaluation sessions by having a second person record rumination simultaneously with the primary observer. Interobserver agreement (interval agreements divided by interval agreements plus disagreements and converted to a percentage) averaged 98% (range, 80% to 100%) for the functional analysis and 96% (range, 88% to 100%) for the intervention evaluation. Procedural integrity was assessed during 38% of functional analysis sessions and 34% of intervention evaluation sessions by having a second person record the teacher implementing the written procedural guidelines. Procedural integrity (procedures implemented correctly divided by procedures implemented correctly and incorrectly and converted to a percentage) was assessed for each session and averaged 98% and 97%, respectively (range, 80% to 100%).
Procedure and Design
We exposed Aaron to five experimental conditions in a multielement design (Iwata, Dorsey, Slifer, Bauman, & Richman, 1994). Three conditions were presented each day based on a random sequence. In the attention condition, Aaron's teacher gave him a preferred leisure activity (access to a computer) for the duration of the session, did not interact with him, and delivered a disapproving comment when he ruminated. In the ignore condition, the teacher sat out of Aaron's view, did not interact with him, allowed him free access to the classroom, and did not react when rumination occurred. During the escape condition, the teacher presented Aaron with three learning tasks (one task per session) from his individualized educational program. When Aaron ruminated, the teacher removed the task and stopped interacting with him for 15 s. Finally, two toy play conditions assessed whether different leisure items that Aaron frequently chose during free time would compete with rumination. During the Play 1 condition, he had continuous access to his most preferred leisure activity (computer) to watch videos. The same computer that was used during the attention condition was used in this phase. During the Play 2 condition, he had continuous access to puzzles, board games, and letter and number books. In both play conditions, the teacher spoke to Aaron every 30 s (e.g., “You're watching a favorite video.” “You have a blue letter.”)
Three staff members who were familiar with Aaron were asked independently to identify foods and liquids that he liked to consume. During 10-min preference assessment sessions (Pace, Ivancic, Edwards, Iwata, & Page, 1985), Aaron's teacher presented him with a single food or liquid for 60 s. An observer recorded the percentage of food that he consumed in quantities of 10 (e.g., 10 potato chips, a granola bar broken into 10 pieces). Foods and liquids were randomized and presented once over a 2-day period. Liquid and pureed food amounts were recorded in ounces, and one ounce was presented to him at a time. The following foods and liquids that had 100% consumption were subsequently included in the intervention evaluation: soft granola bar, cheese crackers, applesauce, and graham crackers.
Results of the functional analysis suggested that rumination was maintained by automatic reinforcement. Accordingly, the purpose of the intervention evaluation was to examine the effects of noncontingent access to preferred stimuli (Phases 1 and 2) and different stimulus-presentation formats (Phase 3) on rumination. As noted, we compared different intervention conditions in an alternating treatments design by scheduling one 5-min session each day. Similar to the functional analysis, intervention evaluation sessions started approximately 2 min after Aaron consumed his lunch. In baseline, the procedures were identical to the ignore condition of the functional analysis. The purpose of Phase 1 was to compare the effects of auditory, visual, and food stimuli on rumination. During the auditory condition, Aaron had continuous access to music played on a CD player. In the visual condition, he used a DVD player to watch brief clips from his favorite movies and television shows. He could switch between songs and video clips without restriction in both conditions. In the food condition, Aaron had continuous access to small pieces (5 g per piece) of one of his highly preferred foods, which was a soft granola bar.
Based on the results of the first intervention phase, the purpose of Phase 2 was to evaluate foods that differed in texture and taste. In separate sessions, the teacher presented Aaron with goldfish crackers, applesauce, granola bars, and graham crackers. We selected these foods for their nutritional content (i.e., lower calorie foods preferred over high-calorie foods). As in all conditions in Phase 1, the different foods were available continuously throughout the session.
The purpose of Phase 3 was to compare different formats for presenting the most effective food identified in Phase 2 (soft granola bar). In the presession presentation condition, which occurred 2 min after Aaron consumed his lunch, he had continuous access to food for 15 min, followed immediately by a 5-min intervention evaluation session without food. Under fixed-time presentation, the teacher gave Aaron a piece of the food every 60 s throughout the session. The third condition, continuous presentation, was identical to the continuous access to food condition in Phase 1. The same brand-name granola bar was used in Phases 2 and 3.
RESULTS AND DISCUSSION
In the functional analysis (Figure 1, top), rumination was relatively stable across attention (M = 43%), ignore (M = 50%), escape (M = 57%), Play 1 (M = 31%), and Play 2 (M = 35%) conditions. For the intervention evaluation (Figure 1, bottom), rumination averaged 53% in baseline. In Phase 1, rumination decreased during the auditory stimulation condition (M = 40%) and the visual stimulation condition (M = 25.5%), with the largest reduction during the food condition (M = 4%). In Phase 2, rumination was lowest for the soft granola bar (M = 6%), followed by applesauce (M = 10%), goldfish crackers (M = 14%), and graham crackers (M = 20%). When the soft granola bar was presented exclusively in Phase 3, rumination occurred at similar low percentages during the presession continuous access condition (M = 2%) and the in-session continuous access condition (M = 3%) with an accelerating trend in the fixed-time condition (M = 33%). Aaron consumed an average of 13 entire bars per session (range, 3 to 17) during this phase.
As noted above, this study is one of only a few that conducted a functional analysis of rumination. In addition, a caregiver implemented the functional analysis in a real-world (nonsimulated) setting, with good fidelity and high acceptability. One potential limitation of our analysis is that we concluded that rumination was maintained by automatic reinforcement even though responding was undifferentiated across conditions (i.e., rumination was not highest in the ignore condition, occurred in a condition of relatively high stimulation [escape], and did not appear to be stable in some of the test conditions; Hagopian et al., 1997). Nonetheless, the data suggested that rumination was not maintained by a social variable.
Our intervention evaluation demonstrated a systematic approach for developing an effective treatment via noncontingent access to preferred stimuli. We compared the effects of continuous access to different types of sensory stimulation, different types of one sensory stimulus (food), and different food presentation formats. The principal finding was that the participant ruminated less frequently when he had access to a particular food (soft granola bars) that was available either during or immediately before sessions. The presession condition in Phase 3 also suggested that continuous access to food may have a lingering reductive effect on rumination notwithstanding the possible confounding effect of passage of time. However, we did not isolate the mechanisms responsible for the treatment outcome. For example, continuous access to food may have reduced rumination through satiation or oral and pharyngeal stimulation.
Results should be interpreted cautiously for several reasons. Specifically, we included only a single participant, the amount of food he ate at lunch was not documented, rumination was not recorded during the presession condition, and the intervention evaluation was confined to a brief postlunch period rather than other times of the day when rumination occurred. Continuous access to food also could produce weight gain, which, as with Aaron, may not be an early-stage concern in cases of chronic rumination but a possible treatment contraindication long term. Despite these concerns, our findings demonstrate the advantages of conducting a functional analysis of rumination to select an effective intervention that can be applied successfully by caregivers outside a formal research setting.
- 2007). Vegetative disorders. In R. M. Kliegman, R. E. Behrman, H. B. Jenson, & B. F. Stanton (Eds.), Nelson textbook of pediatrics (18th ed., pp. 113–115). Philadelphia, PA: Elsevier. , & (
- 1997). Toward the development of structured criteria for interpretation of functional analysis data. Journal of Applied Behavior Analysis, 30, 313–326. doi: 10.1901/jaba.1997.30-313 , , , , , & (
- 1994). Toward a functional analysis of self-injury. Journal of Applied Behavior Analysis, 27, 197–209. doi: 10.1901/jaba.1994.27-197 (Reprinted from Analysis and Intervention in Developmental Disabilities, 2, 3–20, 1982) , , , , & (
- 2007). Treatment of rumination maintained by automatic reinforcement: A comparison of extra portions during a meal and supplemental postmeal feedings. Behavioral Interventions, 22, 297–304. doi: 10.1002/bin.249 , & (
- 2011). Direct and distal effects of noncontingent juice on the rumination exhibited by a child with autism. Journal of Applied Behavior Analysis, 44, 955–959. doi: 10.1901/jaba.2011.44-955 , & (
- 2007). Brief functional analysis and supplemental feeding for postmeal rumination in children with developmental disabilities. Journal of Applied Behavior Analysis, 40, 743–747. doi: 10.1901/jaba.2007.40-743 , , , & (
- 2002). Teacher acquisition of functional analysis methodology. Journal of Applied Behavior Analysis, 35, 73–77. doi: 10.1901/jaba.2002.35-73 , , , , , & (
- 1985). Assessment of stimulus preferences and reinforcer value with profoundly retarded individuals. Journal of Applied Behavior Analysis, 18, 249–255. doi: 10.1901/jaba.1985.18-249 , , , , & (
- 1981). The relation of food quantity to rumination behavior. Journal of Applied Behavior Analysis, 14, 121–130. doi: 10.1901/jaba.1981.14-121 , , , & (
- 1974). Lemon-juice therapy: The control of life-threatening rumination in a six-month-old infant. Journal of Applied Behavior Analysis, 7, 557–563. doi: 10.1901/jaba.1974.7-557 , , & (
- 1999). Access to white bread as an intervention for chronic ruminative vomiting. Journal of Behavior Therapy & Experimental Psychiatry, 30, 137–144. doi: 10.1016/S0005-7916(99)00012-9 , , , & (
- 2009). Functional analysis and treatment of rumination using fixed-time delivery of a flavor spray. Journal of Applied Behavior Analysis, 42, 977–882. doi: 10.1901/jaba.2009.42-877 , , , , & (
Action Editor, Michael Kelley