Factors Associated With Difficult Toilet Training
* Children’s Hospital, Division of General Pediatrics, and Harvard Medical School, Boston, Massachusetts
Brazelton Touchpoints Center, Children’s Hospital, Boston, Massachusetts
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Objective. To identify temperament and behavioral patterns in children with difficult toilet training and to compare those children with same-aged toilet-trained children.
Methods. We compared 46 referred clinic patients who were difficult toilet trainers (DTT) with 62 comparison children (CC) using the Carey-McDevitt Behavioral Style Questionnaire, the Parenting Scale, and a questionnaire of toilet-training history.
Results. CC were more likely to have easy temperaments (odds ratio [OR]: 33.51). DTT were more likely to be less adaptable (OR: 3.12), more negative in mood (OR: 2.79), less persistent (OR: 2.97), and lower in approach (OR: 1.85). DTT were more likely than CC to be constipated (OR: 3.52), although 55% of CC were constipated. DTT were likely to hide to stool (74%) and to ask for pull-ups in which to leave stool (37%). Parenting styles did not differ between the groups.
Conclusions. Although the referral population may be inherently biased, these data suggest that difficult toilet training is associated with difficult temperamental traits and constipation in affected children.
Key Words: toilet training • temperament • constipation
Abbreviations: DTT, difficult toilet trainers • CC, comparison children • BSQ, Behavioral Style Questionnaire • PS, Parenting Scale • OR, odds ratio
Children typically toilet train by 3 years of age in the United States.1 However, many children have difficult toilet training, a concern often brought to the attention of their pediatrician.2 Pediatricians have little evidence-based information regarding these struggles, and it is unclear what distinguishes this group from children who toilet as expected.3
At our Developmental Medicine Center within a large, tertiary-care hospital, 750 outpatient visits per year take place for toileting difficulties, including constipation, encopresis, enuresis, and failure to toilet train. We designed this study to determine differences between children with difficult toilet training and children who toilet as expected.
We hypothesized that children who struggle with toilet training would be more temperamentally difficult, likely to be constipated, and likely to have had an anxiety-provoking event associated with training. We hypothesized that their parents would have more dysfunctional parenting styles when compared with children who train as expected.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Subjects/Setting
Children with difficult toilet training were identified from consecutive patients who were referred to the Pains and Incontinence Program. The Pains and Incontinence Program is a specialty clinic at Boston’s Children’s Hospital, a university-based regional center with patients from Boston as well as a wide surrounding area. Children with enuresis, encopresis, and failure to toilet train are evaluated and treated. Of 182 consecutive 3- to 7-year-olds, full sets of temperament, parenting, and toileting questionnaires were completed for 139 children, yielding a 76% response rate. Of these 139, 46 were identified as currently having difficulty toilet training, thus called "difficult toilet trainers" (DTT), using the following criteria: 1) 3.5 to 4 years old and not trained after 6 months of trying or 2) older than 4 and younger than 7 years and refuses to use the toilet or has not completed daytime urine or stool training. Children with reported developmental delays or underlying medical causes to their constipation were excluded.
Comparison children (CC) were recruited from 5 local preschools. Of 222 preschoolers asked, 81 (36%) participated. From the 81 preschool participants, 62 CC were identified, after eliminating children who were not yet toilet trained or had developmental delays.
Design
We performed a cross-sectional, descriptive study comparing convenience samples of clinic patients and community preschoolers. Parents of patients and preschoolers completed a Carey-McDevitt Behavioral Style Questionnaire (BSQ), the Parenting Scale (PS), and a toileting history questionnaire. A letter of the voluntary nature of this study was included and approved by the Committee for Clinical Investigation as implied consent.
Measures
Parents completed full sets of temperament, parenting, and modified toileting questionnaires. The BSQ is a 100-item, 6-point Likert-type, 25-minute tool for 3.0- to 7.0-year-olds that is widely used in pediatric and psychology literature.4 Nine temperament categories are identified (activity, adaptability, approach, distractibility, intensity, mood, persistence, rhythmicity, and threshold [Table 1]). Categories cluster into 4 temperaments: easy, slow to warm up, difficult, and intermediate. Higher scores in 5 categories (rhythmicity, approach/withdrawal, adaptability, intensity, and mood) reflect more difficult styles. "Difficult temperament" criteria are met when a child is scored higher than the mean in 4 of these 5 categories, 1 of which must be the intensity category, and with 2 of these traits scoring >1 standard deviation (SD) above the mean. "Easy temperament" criteria are met in a child who is not rated above the mean in >2 of these 5 categories, and neither of these 2 can be >1 SD above the mean (Table 2). Total score test-retest reliability has been shown to be 0.89. Internal consistency for the instrument was 0.84. The major temperament clusters of the BSQ are similar to those found in the New York Longitudinal Study.5
|
|
The PS is a 5- to 10-minute, 30-item, 7-point Likert scale that measures 3 dysfunctional discipline styles in parents of young children: laxness, overreactivity, and verbosity.6 The laxness factor relates to permissive discipline, such as a tendency for parents to allow rules to go unenforced. Overreactivity reflects parental tendency toward displays of anger, meanness, and irritability. Verbosity entails lengthy verbal responses and a reliance on talking even when talking is ineffective. Previous testing has shown test-retest reliability of this measure to be 0.84 for the total score and internal consistency to be 0.84 for the total score. Initial development of the PS found scores to correlate significantly with observational measures of dysfunctional discipline and child misbehavior.
The toileting history questionnaire for DTT contained 115 items, whereas that for CC had 64 items, requiring 25 minutes and 15 minutes to complete, respectively. This data-gathering tool has not been validated. It is based on the previsit questionnaire used in our clinic to preview patient information before the first visit for clinical purposes. Questions that are pertinent only to clinic visits were omitted from the toileting questionnaires of CC to increase likelihood of response. Many questions asked of the DTT were not appropriate for CC, such as, "What are the issues you hope will be addressed by our program?" The 64 remaining items had the same wording and sequence as found in the clinic questionnaire. Questions related to timing and methods of toilet training, as well as medical and developmental history. For example, parents were asked with checklists, yes/no, and multiple-choice questions to rate the difficulty experienced when toilet training their child, whether the child was rewarded or praised for accomplishing toileting goals, and how often the child stools into the toilet or potty. Parents of DTT were asked whether the child hides when stooling, asks for pull-ups or a diaper and then stools into them, hides dirty underwear, or has been teased because of accidents. Constipation was defined as the endorsement of hard stool, painful stooling, or stools so large that they clog the toilet.
Data Analysis
Data were dually entered into Excel worksheets and compared, and all discrepancies were reconciled by checking the original data source. The cleaned data were then imported into SPSS (SPSS Inc, Chicago, IL) for analyses. Frequencies, cross-tabulations, 
2 tests, and t tests were performed when appropriate. Differences in categorical outcomes were assessed using multivariable binomial logistic regression, controlling for gender and age. For all statistical tests, a 2-tailed P < .05 was considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Demographics
The mean age of DTT (n = 46) was 5.1 years and of CC (n = 62) was 4.5 years (P = .001). DTT were 39% female and 61% male; CC were 52% female and 48% male (P = .54). All regression analyses controlled for age and gender. Additional demographic information was not collected.
Temperament Traits and Clusters
CC were significantly more likely than DTT to have easy temperament clusters: 42% of CC were "easy" compared with only 2% of DTT (odds ratio [OR]: 33.55; P = .001). Four temperament traits differed significantly between the 2 groups of children. DTT were less adaptable (OR: 3.12; P = .001), had more negative mood (OR: 2.79; P = .01), were less persistent (OR: 2.97; P = .01), and were lower in approach (OR: 1.85; P = .03; Table 3). DTT did not have significantly more "difficult" or "slow to warm up" temperament clusters (Fig 1). Although the DTT did not meet the strict definition of "difficult temperament" as defined by the BSQ, 3 of the significantly different traits (adaptability, mood, and approach) contribute to more difficult temperamental natures in the DTT.
|
|
Parenting Style
Parenting style did not differ significantly in laxness, verbosity, or overreactivity between the 2 groups (Table 4). Although in the DTT group 2 of the 3 domains trended toward more dysfunctional parenting styles, they did not reach statistical significance.
|
Constipation
Seventy-eight percent of DTT were constipated, compared with 55% of CC (OR: 3.54; P = .01).
Associated Factors
There were no significant differences in frequency of anxiety-provoking events associated with toileting, such as a public accident, parental separation or divorce, new sibling’s birth, or interruption of toilet training. However, 53% of DTT were reported to have had fears about toilet training, compared with 26% of CC (OR: 0.41; P = .04).
Behavior Patterns
Several behavior patterns were identified from the DTT (questions regarding these patterns were not asked of the comparison preschool group): 74% of DTT hide to stool occasionally or often, 37% ask for pull-ups in which to stool, 24% of DTT hide dirty underwear, and 22% have been teased because of stooling accidents.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
DTT have been the subject of very few studies. Our findings fill a gap in understanding who these children are and temperamental and medical tendencies that should be considered when addressing their struggle with toileting. For instance, Blum et al7 found that 27 children with stool toileting refusal had no more behavior problems than matched, toilet-trained control subjects. Refusers had higher rates of constipation and painful defecation, although whether this was cause or effect was unclear. A trend toward more difficult temperament in the refusers was found. More recently, Blum et al8 found that neither constipation nor stool toileting refusal is associated with earlier toilet training. Kuhn et al9 published treatment guidelines for treating stool toileting refusal, emphasizing the need to address disruptive behavior problems that may obstruct successful toileting. Additional reports examine treatment strategies for children with stool toileting refusal.10–12 Thus, previous data indicate that toilet-training struggles are not correlated with age of initiation or behavior problems and begin to consider how to address difficult toileting, but our study is the first to document the significant positive correlation among difficult toilet training, constipation, and temperament.
We found that DTT were significantly more constipated than CC. However, many of the CC were constipated as well, confirming that this is a common finding in this age group. Although constipation is 1 part of the constellation of findings in children with difficult training, it is unclear whether this is cause or effect; the child may be constipated, feel pain with defecation, and begin to withhold and refuse. Alternatively, the child may withhold and then become constipated. Regardless, constipation must be considered and treated when present in children with toilet refusal.
A descriptive cross-sectional study of toilet training found that older age, female gender, nonwhite race, and single parenthood were more strongly associated with toileting success than cognitive development or temperament.13 Included were 496 children aged 15 to 42 months; thus, older children with ongoing toileting refusal were not studied. The need for additional analysis comparing temperament traits to age of toilet training was specified. Our study begins to meet this need.
Our initial hypotheses did not include the role of race in difficult toilet training, as we focused on temperament. As the role of nonwhite race was implicated in the age of toilet training after our data collection was complete,13 that demographic information was not gathered in our study.
Previous data does not confirm varying temperament distribution across races.14 In addition, we prefer not to ask for demographic information before the first clinic visit.
Temperament profiles of DTT in our study differ from those of CC. DTT were found to be less adaptable, more negative in mood, less persistent, and lower in approach than comparison children. Thus, DTT are a more difficult group temperamentally, even without meeting full "difficult temperament" criteria.
The BSQ has stringent criteria for "difficult temperament," requiring intensity to be 1 SD above the mean. In our study, DTT tended not to score above the mean for intensity, and there was no significant difference in intensity between the 2 groups. (see Table 3)
It is interesting that the CC were notably more likely to have "easy temperaments" than the DTT. DTT had a high proportion of difficult traits, allowing only 2% to be clustered as "easy"; this may be attributable to the skewed population that seeks guidance at our referral center. However, the proportion of children with easy temperaments in the CC group (42%) is similar to that in the group that normed the BSQ (33%) as well as the original New York Longitudinal Study (40%). The response rate in the CC group was only 36%, but these comparable rates of easy temperament suggest that the proportion of temperaments among them is typically distributed.
DTT demonstrated common behavior patterns that provide insight into the nature of their difficulties. The frequency of their "hiding to stool" and "asking for pull-ups" indicates that a large number of DTT have full bowel control and control where they will or will not defecate. They often have the developmental skills necessary for toilet training, a fact that can be useful when developing a plan with the family. That these children hide to stool may reflect their understanding of the private nature of toileting, suggesting appropriate social awareness.
The frequency of hiding dirty underwear and being teased for stooling accidents highlights the child’s simultaneous difficulty in consolidating this skill, despite having control of stooling. The child may feel overwhelmed and ashamed, unable to meet the full demands of using the toilet. Parental expectations and reactions combined with peer humiliation may cause the child to feel even more overwhelmed in the face of this challenge.
Parents of DTT did not seem to differ in parenting style as measured by the PS, despite our expectations. Although ineffective interactions and overreactive or lax parenting styles may be observed around issues of toileting, this may be specific to toileting behavior and not to the parent as a whole. Parents may be relieved that their own parenting styles do not seem to correlate with toilet-training difficulties, alleviating their guilt and enhancing their ability to work constructively on the toileting issues. Alternatively, appreciating the child’s temperamental contribution to toileting issues may improve the "fit" between parent and child, allowing parents to develop better strategies for the conflicts that have developed.15
Fears were more likely to be reported for DTT than CC, but this may represent a recall bias. Additional prospective study is needed to clarify this factor.
Limitations
This cross-sectional study has several limitations. Parents who bring their children to the specialty clinic may be more likely to describe their children as "difficult," resulting in recall bias. Parent descriptions of their child’s toileting difficulties or their own parenting styles may also be skewed after the difficult experience of toilet refusal. In addition, we cannot confirm that the 2 groups are comparable. Only 36% of the surveyed comparison families responded, and we cannot determine factors that influenced them to respond or what distinguishes responders from those who did not respond. For example, families with easier toileting experiences may have been more likely to complete the forms. CC were identified from communities within our referral base, including both urban and suburban populations, but no socioeconomic data were available for comparison. Finally, our DTT group is a referral population, so their traits may not generalize to the greater population of toilet trainers.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Our hypotheses were mostly supported. Children in our referral population with difficult toilet training were more likely to have difficult temperamental traits and were more likely to be constipated. Their parents, however, did not differ in style from parents of those who toilet train as expected. To confirm these findings, a prospective study would be helpful. Primary care providers can use these findings to anticipate children who are at risk for difficult toilet training and to work with parents collaboratively during the toilet-training process.
|
ACKNOWLEDGMENTS |
|---|
We thank Nathan Blum, MD, for review of the manuscript and Leonard Rappaport, MD, for assistance and support during this study.
|
FOOTNOTES |
|---|
Received for publication Aug 11, 2003; Accepted Jan 26, 2004.
Reprint requests to (A.S.) Children’s Hospital, Fegan 10, 300 Longwood Ave, Boston, MA 02115. E-mail: alison.schonwald{at}childrens.harvard.edu
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
- Brazelton TB, Christopherson ER, Frauman AC, et al. Instruction, timeliness, and medical influences affecting toilet training.
Pediatrics.1999; 103
:1353
–1358
[Free Full Text] - Christopherson ER. Toileting problems in children. Pediatr Ann.1991; 20 :240 –244[Web of Science][Medline]
- Brooks RC, Copen RM, Cox DJ, et al. Review of the treatment literature for encopresis, functional constipation, and stool-toileting refusal. Ann Behav Med.2000; 22 :260 –267[Web of Science][Medline]
- McDevitt SC, Carey WB. The measurement of temperament in 3–7 year old children. J Child Psychol Psychiatry.1978; 19 :245 –253[Web of Science][Medline]
- Chess S. Developmental theory revisited: findings of longitudinal study. Can J Psychiatry.1979; 24 :101 –112[Web of Science][Medline]
- Arnold D, O’Leary S, Wolff L, Acker M. The Parenting Scale: a measure of dysfunctional parenting in discipline situations. Psychol Assess.1993; 5 :137 –144[CrossRef]
- Blum NJ, Taubman B, Osborne ML. Behavioral characteristics of children with stool toileting refusal.
Pediatrics.1997; 99
:50
–53
[Abstract/Free Full Text] - Blum NJ, Taubman B, Nemeth N. Relationship between age at initiation of toilet training and duration of training: a prospective study.
Pediatrics.2003; 111
:810
–814
[Abstract/Free Full Text] - Kuhn BR, Marcus BA, Pitner SL. Treatment guidelines for primary nonretentive encopresis and stool toileting refusal. Am Fam Physician.1999; 59 :2171 –2178, 2184–2186[Web of Science][Medline]
- Taubman B, Buzby M. Overflow encopresis and stool toileting refusal during toilet training: a prospective study on the effect of therapeutic efficacy. J Pediatr.1997; 131 :768 –771[CrossRef][Web of Science][Medline]
- Luxem MC, Christopherson ER, Purvis PC, et al. Behavioral-medical treatment of pediatric toileting refusal. J Dev Behav Pediatr.1997; 18 :34 –41[CrossRef][Web of Science][Medline]
- Scmitt BD. Toilet training refusal: Avoid the battle and win the war. Contemp Pediatr.1987; 4 :32 –50
- Schum TR, McAuliffe TL, Simms MD, et al. Factors associated with toilet training in the 1990s. Ambul Pediatr.2001; 1 :79 –86[CrossRef][Web of Science][Medline]
- Carey WB, MCDevitt SC. Coping With Children’s Temperament: A Guide for Professionals. New York, NY: BasicBooks; 1995:21
- Zuckerman B. Healthy parenting choices: children’s temperament styles and parent’s expectations. Pediatr Basics.1995; 74 :2 –7
PEDIATRICS (ISSN 1098-4275). ©2004 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  C. Joinson, J. Heron, A. von Gontard, U. Butler, J. Golding, and A. Emond Early Childhood Risk Factors Associated with Daytime Wetting and Soiling in School-age Children J. Pediatr. Psychol., August 1, 2008; 33(7): 739 - 750. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. R. Fleisher Understanding Toilet Training Difficulties Pediatrics, June 1, 2004; 113(6): 1809 - 1810. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
