Objective. To determine if children with stool toileting refusal have more behavior problems than matched children who are toilet trained.
Design. Case-control study.
Setting. Suburban private pediatric practice.
Participants. Children 30 to 48 months old who had achieved bladder control but refused to defecate on the toilet were identified as cases. Controls were sex- and age-matched children who were fully toilet trained.
Measures. Total behavior problems were assessed using a semi-structured behavior screening interview with the child's parents. The parents also completed the Child Behavior Checklist for ages 2 to 4 and either the Toddler Temperament Scale (30 to 36 months old) or the Behavioral Style Questionnaire (36 to 48 months old). Child compliance with adult instructions was measured during a room clean-up task.
Results. Children with stool toileting refusal were not found to have more behavior problems than the matched children who were toilet trained. There were no differences between the two groups in compliance during the room clean-up task. There was a trend toward children with stool toileting refusal having a more difficult temperament, and these children were reported to have more problems with constipation and painful bowel movements than the controls.
Conclusions. Children with stool toileting refusal do not have more behavior problems than controls who are toilet trained. Parents do report higher rates of constipation and painful defecation, but it is not clear whether this is a cause or effect of stool toileting refusal.
Problems with toilet training are among the most common behavioral concerns for parents of young children.1-3 Although most children are trained by 30 months of age, studies generally find that 2% to over 10% of children are not trained by age 4.4 During the toilet training process most children achieve bowel control before or simultaneously with bladder control.5,6 The group of children who achieve daytime bladder control before bowel control varies from 1.7% to over 20% of the population.5-7 This group of children have been referred to as having “toileting refusal”8 or the “battle of the bowels”9 and have been of interest to clinicians because of the distress they cause parents and the possibility that stool toileting refusal may be related to the later development of encopresis.8,10 It has been suggested that a variety of different factors may be related to the development of stool toileting refusal including early toilet training, excessive parent-child conflict, irrational fears or anxieties around toileting, a difficult temperament, and hard or painful stools as a result of chronic constipation or an anal fissure.9,11,12 Despite these hypothesized causes of stool toileting refusal, there have been no systematic studies of the etiology of this behavior.
This article reports the results of a case-control study investigating the hypothesis that behavior problems in general, and oppositional or noncompliant behaviors in particular occur more frequently in children with stool toileting refusal than in children who are successfully toilet trained. We also investigated the hypothesis that children with stool toileting refusal would have a more difficult temperament.
Twenty-seven children with stool toileting refusal and 27 matched controls were recruited from the private pediatric practice of one of the authors (B.T.). The practice is in a suburb of a major metropolitan area and serves predominately middle and upper-middle class families. The mean Hollingshead occupational score13for the fathers was 7.0 (range, 2 = unskilled workers to 9 = executives and professionals). Twenty-eight of the 54 mothers were employed full-time and five were employed part-time. The mean Hollingshead occupational score13 for the mothers was 5.8 (range, 3 = semiskilled workers to 9).
Children in this practice were being followed prospectively as part of a study of toileting training experiences.7 Children were identified as having stool toileting refusal if they met the following criteria: age between 30 and 48 months, daytime bladder control with 2 or fewer accidents in the previous week, and refusal to defecate on the toilet for more than 1 month after achieving daytime bladder control. Parents of 35 children with stool-toileting refusal were asked to participate in this study. Twenty-seven (77%) completed the study, one refused to participate and seven agreed to participate, but dropped out before completing the study. Each case was matched to the next child in the study of toilet training experiences who met the following criteria: same sex as the case, age within 1 month of the case, no history of stool toileting refusal, and successful completion of toilet training for both urine and stool. If the parents of that child did not agree to participate then the parents of the next child who met the criteria were asked. Parents of 29 children were asked to participate to obtain the 27 controls.
A sample size of 27 per group was determined based on the number of subjects required to detect a large effect size with power = .90 and P < .05 (two-tailed).
The study was approved by the Institutional Review Board of the Children's Hospital of Philadelphia.
Procedures and Dependent Measures
After written informed consent was obtained by one of the authors (B.T.) from the child's parents, this author asked the parents information about when they began toilet training their child and about any history of painful bowel movements or hard stools. All other members of the study team were blind to the child's toilet training history. The parents were then given the rating scales discussed below to complete and scheduled to return for the behavioral interview and observation. The interview and behavioral observation took place at their pediatrician's office during a time when the office was closed for business.
Child Behavior Checklist (CBCL)
The 2- to 4-year-old version of the CBCL is a 99-item questionnaire completed by parents about their child's behavior during the previous 6 months. It provides measures of behavioral difficulties along the dimensions of externalizing behaviors and internalizing behaviors. The externalizing dimension assesses defiant, overactive, aggressive, and destructive behaviors, while the internalizing dimension assesses withdrawn and depressed behaviors. It has test-retest reliability of r = .88 and 1-year stability of r ≥ .70 on both dimensions.14 Children referred for evaluation of behavior problems score higher than nonreferred children on all subscales.14
Due to the ages of the children in this study two temperament questionnaires were used. The Behavioral Style Questionnaire15, developed for children 3 to 7 years old, was used for all children over 36 months old and the Toddler Temperament Scale16, developed for children 1 to 3 years old, was used for all children between 30 and 36 months old. Both questionnaires have nine subscales corresponding to the nine temperament characteristics identified by Thomas and colleagues17 in the New York Longitudinal Study and the two questionnaires have many of the same items. The Behavioral Style Questionnaire has test-retest reliability of r = .89 and the Toddler Temperament Scale has test-retest reliability ofr = .81. The child's overall temperament can be classified on both scales as easy, intermediate-low, intermediate-high, or difficult.
A semi-structured behavioral screening interview for preschool children18 was used to assess for problem behaviors in the areas of feeding, sleeping, activity level, attention span, peer and sibling relationships, dependence on parents, difficulty with discipline, temper tantrums, moods, worries or anxiety, and fears. Questions about difficulty toilet training were eliminated from the interview so that the interviewer would remain blind to the child's toilet training history. Each area assessed is scored by the interviewer on a scale from 0 to 2, with higher scores representing increasing behavior problems. The total behavior problem score is the sum of the scores for each area and can range from 0 to 24. In this study the maximum score was 22 because the assessment of toilet training was eliminated from the interview. The interview has been used to screen populations of three-year old children for problem behaviors.1,2
All interviews were videotaped and 16 (30%) were selected at random and scored by an independent observer. Interrater reliability was calculated as the correlation between the scores of two raters (r = .82; P < .001).
Observation of the parent-child interaction occurred during a clean-up task. One parent (usually the mother) and the child were in a 3.4 × 2.3 meter waiting room with only two chairs and two tables. Ten books and 10 toys were placed on the floor throughout the room. The child's parent was instructed to play for 5 minutes with the child leading the play. At the end of the play time the parent was instructed to have child clean-up the room by putting all the books on one table and all the toys on the other table. The parents were told that they were not allowed to pick up the toys or books themselves and that the clean-up task would end when either the child had picked up all the toys or books or after 5 minutes if the child had not completed the task.
All clean-up sessions were videotaped and the videotapes were scored for the presence of α-commands, β-commands, compliance, and parental reaction to compliance. The coding system was a modification of the system described by Forehand and McMahon19 and used fewer codes than their system. Briefly, α-commands are commands that require a specific motor or verbal response, whereas β-commands are vague commands or commands that are followed within 5 seconds by other statements or commands so the child is not given an opportunity to comply. Child compliance is scored only in response to α-commands and is coded if the child initiates compliance with an α-command within 5 seconds of the command. If the child complies with a parental request the parental response is coded as either attending to the child if they praise or provide attention to the child for complying or ignoring if they do nothing to acknowledge the child's compliance or just issue another command. The total number of commands issued (sum of α- and β-commands) and the time to complete the task (maximum = 300 seconds) were also recorded.
Seventeen (31%) of the videotapes were scored by two observers. Interobserver agreement was calculated for each subject as the number of agreements divided by the number of agreements plus disagreements multiplied by 100%. Interobserver agreements on α-commands was 92.3% (58.9% to 100%); β-commands was 82.0% (33.3% to 100%); child compliance was 95.1% (66.7% to 100%); and parent attending was 90.2% (66.7% to 100%).
The toilet training characteristics of the sample are shown in Table 1. There was a tendency for children with stool toileting refusal to wet their pants more during the day and a higher percentage wet their bed at night (at least once a week) although neither of these reached statistical significance. There was a trend toward parents of children with toileting refusal reporting that they began toilet training at a later age, but there was no difference in the age at which the two groups became dry during the day. On the CBCL parents are asked to rate the frequency of constipation, painful bowel movements, and resists toilet training on a scale from 0 (not true) to 2 (very or often true). Constipation was more frequent among children with stool toileting refusal (54% vs 7%; P < .001) as were painful bowel movements (60% vs 15%; P < .01). Parents reported that 7% of the controls resisted toilet training (no score higher than 1) as compared to 96% of the children with stool toileting refusal (P < .001).
We could find no evidence that children with stool toileting refusal had more behavior problems than our control group (Table2). Parent reports of both externalizing and internalizing behavior problems on the CBCL, and total behavior problems during the interview were not different between groups. A slightly higher percentage of children in the toileting refusal group completed the 5-minute clean-up task (not statistically significant) and their was no difference in the percent compliance to parental directions between the groups. Parents of the children with stool toileting refusal were more likely to attend to their children in response to child compliance than were the parents of controls. There was a trend toward parents of children with stool toileting refusal rating their children's overall temperament as more difficult than that of the control group (Table 3), but there were no statistically significant differences on any of the nine temperament subscales (data not shown).
The children in this study were followed until toilet trained as part of the longitudinal study of toilet training mentioned above.7 Fourteen children in this study had stool toileting refusal for at least 6 months and 13 of the 14 were not toilet trained for stool by 42 months of age. Thirteen trained for stool less than 6 months after becoming dry during the day. Posthoc analysis comparing those with stool toileting refusal for less than 6 months with those who had it more than 6 months revealed no differences on the internalizing or externalizing dimension of the CBCL or the total behavior problems score on the interview. There was a tendency for parents to issue more α-commands to those children with stool toileting refusal for less than 6 months (14.4 ± 8.7 vs 8.1 ± 5.5; P < .03), but when we corrected for the fact that those who had stool toileting refusal for less than 6 months were slightly younger at enrollment in the study (36 ± 2.6 month vs 39.5 ± 4.5 months, P = .02) the difference was no longer statistically significant.
Problems with toilet training are very common and there have been many hypothesized causes. In the 1950s and 1960s it was believed that coercive toilet training practices were a common cause of toilet training problems and the efficacy of a child-oriented approach was demonstrated.5 Although this child-oriented approach to toilet training has become widely accepted,20 problems with toilet training remain common. Current hypotheses about causes of one type of toilet training problem, stool toileting refusal, have focused on the behavioral characteristics of the child, the parent-child interaction, or the child's experiences when defecating (ie, hard or painful bowel movements).
In this study we could not find any evidence that children with stool toileting refusal had more behavior problems than a group of age- and sex-matched children who were successfully toilet trained. Specifically, parents did not report more externalizing or internalizing behavior problems on the CBCL or more overall behavior problems during a semistructured parent interview. During a 5-minute clean-up task there was no difference in the number of directions parents issued or the child's compliance with these directions. There was a trend toward children with stool toileting refusal being rated as having a more difficult temperament, but no specific temperament characteristic distinguished the two groups. The parents of the children with stool toileting refusal were more likely to praise their children when they complied, but the significance of this finding is not clear.
The question of why clinicians perceive children with stool toileting refusal to be more difficult and have more behavior problems than other children remains open. It is possible that the parents who are most likely to request help from professionals are parents whose children have toilet training difficulties and other behavior problems. In a study conducted by Richman et al,1 30% of the children with toilet training difficulties had total behavior problem scores in the clinically significant range. Alternatively, because toileting refusal resolves in many cases,7 it is possible that the parents who seek professional help are the ones whose children have stool toileting refusal of a longer duration. However, posthoc analysis of the group in this study did not suggest that the duration of toileting refusal was related to the severity of other behavior problems. The sample size used in this study gave us a very high likelihood of detecting a large difference between the two groups. We cannot rule out that the clinician's impressions are formed by detecting small or moderate size differences between the two groups that may be detected with a larger sample. Finally, the data do suggest a trend toward children with stool toileting refusal having a more difficult temperament than the controls. It is possible that an interaction between a difficult temperament and family or other factors not measured in this study cause stool toileting refusal in some cases.
We did find that parents of children with toileting refusal reported that their children had constipation and/or painful defecation more frequently than did parents of the controls. However, our study was not specifically designed to investigate this possible cause of toileting refusal. Thus, although this finding is consistent with previous studies,21 it is limited, as were the previous studies, in that the questions about constipation and painful defecation were asked only after the toileting refusal had begun. Thus, one cannot determine if constipation or painful defecation are a cause or an effect of stool toileting refusal. In either case, constipation or painful defecation may be important in maintaining the behavior. Given the high frequency of constipation and painful defecation in children with stool toileting refusal, it is recommended that dietary changes or medications to soften the stool be used as part of treatment for children with stool toileting refusal.12
We thank Paul R. Gallagher, MA, for his assistance with the statistical analysis and William B. Carey, MD, for his comments and helpful suggestions on an earlier version of this manuscript.
- Received December 11, 1995.
- Accepted March 11, 1996.
Reprint requests to (N.J.B.) Children's Seashore House, 3405 Civic Center Blvd, Philadelphia, PA 19104.
- CBCL =
- Child Behavior Checklist
- Mesibov GB,
- Schroeder CS,
- Wesson L
- Berk LB,
- Friman PC
- Brazelton TB
- ↵Taubman B. Toilet training and toilet training refusal for stool only: A prospective study. Pediatrics. 1997;99;54–58
- Schmitt BD
- Levine MD
- ↵Hollingshead AB. Four Factor Index of Social Status. New Haven, CT: Yale University; 1975
- Fullard W,
- McDevitt SC,
- Carey WB
- ↵Thomas A, Chess S, Birch HG, Hertzig ME, Korn S. Behavioral Individuality in Early Childhood. New York, NY: New York University Press; 1963
- ↵Forehand RL, McMahon RJ. Helping the Noncompliant Child. A Clinician's Guide to Parent Training. New York, NY: The Guilford Press; 1981
- Partin JC,
- Hamill SK,
- Fischel JE,
- Partin JS
- Copyright © 1997 American Academy of Pediatrics