van Tilburg, Miranda A.L.*; Squires, Megan*; Blois-Martin, Nanette*; Williams, Cathy†; Benninga, Marc A.‡; Peeters, Babette‡; Ulshen, Martin†
In the Unites States, almost 1 in 4 children seen in primary clinics is reported to have constipation (1). The majority of constipated patients (79%) experience reduced bowel movements in combination with fecal incontinence (FI) (2). FI is usually caused by overflow around a fecal impaction due to stool holding. Although FI can occur in the absence of fecal retention, we will focus on FI with constipation and stool withholding. It has been suggested that parents play an important role in the development and persistence of FI (3).
For parents, FI may be one of the most burdensome consequences of constipation in their children. FI creates stress because of the burden of cleaning clothes, constant vigilance to bowel functions, and the child's perceived laziness and unwillingness to cooperate (4,5). In our own clinical experience, a focus on children's recalcitrance as the cause for FI delays treatment seeking in many families. Families may feel FI is the child's fault and, therefore, medical intervention is not necessary. Postponing treatment not only increases the discomfort for the child but also complicates future treatment success. FI is created by a complex vicious cycle, both psychological and physiological, which becomes increasingly difficult to break as time passes (6). In a study among small children, duration of constipation of <3 months before referral, compared with long-lasting symptoms, correlated with better outcomes (7). Early intervention is therefore of utmost importance.
The aim of the present study was to assess parental knowledge and misconceptions about FI in their children and how this relates to the care and treatment of FI. We determined parental knowledge of FI and how this knowledge is related to treatment choices. We also assessed whether parental knowledge of FI changes after physician treatment.
Participants consisted of parents of children ages 4 to 18 years with constipation and FI. FI was defined as passage of feces in underwear on average 1 time per week or more. To capture retentive FI in association with constipation, at least 2 of the following had to be present in addition to FI: ≤2 bowel movements in the toilet per week in the last 2 months; history of painful or hard bowel movements; a history of large-diameter stools, which may obstruct the toilet. Children with a physician-diagnosed neurological, mental, or developmental disorder that can cause soiling were excluded.
Parents were recruited through the University of North Carolina pediatric gastroenterology and general pediatric clinics, the Duke University pediatric gastroenterology clinics, and various Web sites (including the UNC Center for Functional GI and Motility Disorders, International Foundation for Functional Gastrointestinal Disorders, and various other gastrointestinal disorders support groups, as well as parent forum groups), to ensure a mix of patients who had not seen a doctor, consulted a primary care physician, or consulted a pediatric gastroenterologist.
To generate items, in-depth semistructured interviews were conducted with 15 parents of children who consulted for FI at our pediatric gastroenterology office. Parents were asked about beliefs, worries, causes, and treatment preferences. From these interviews, an initial pool of items was generated. Items were evaluated by a group of clinicians (10 pediatric gastroenterologist and 6 pediatric nurse practitioners) and parents of patients (N = 5) to determine content validity, that is, evaluate any omissions and item content and understandability of the items. Changes were made according to feedback. A total of 69 items were generated, which were answered on a 5-point scale (never/rarely, sometimes, often, most of the time, always).
To determine subscales and initial reliability and validity, a new group of parents (N = 251) of children with FI were invited to complete the developed questionnaire twice. All of the questionnaires were completed online.
Last, to examine whether the newly developed scale is responsive to change in parental knowledge of FI, a small pilot study was conducted among 30 parents of new patients at the UNC and Duke clinics (with 5 practicing pediatric gastroenterologists each). These families had not seen a pediatric gastroenterologist for this condition previously and were asked to complete the questionnaire before (T1) and after meeting with the physician (T2) and 2 months later (T3).
The following analyses were performed with SPSS 17.0 (SPSS Inc, Chicago, IL):
1. Violation of normal distribution was determined by assessing skewness, kurtosis, and range. Items that are in clear violation were dropped from the questionnaire.
2. To determine subscales and shorten the questionnaire, principal component analysis was run (8). Principal component analysis is useful to examine whether some variables are related to each other and, as such, may measure the same construct. At the same time, it will identify redundant variables and can be used to shorten the questionnaire. First, a Scree test was performed to determine the number of constructs also called factors. Once the number of factors was determined, the principal component analysis was rerun with a varimax rotation to maximize high correlation and minimize low correlations between factors and variables.
3. Reliability was assessed by Cronbach α (internal reliability) and Pearson correlations between the first and second administration of the online questionnaire (test-retest reliability).
4. To establish initial validity of the questionnaire's scales, Mann-Whitney U tests were performed to compare sum scores of the subscales by treatment groups (no treatment, primary care, tertiary care) and treatments (use of over the counter and prescription medication).
5. Furthermore, Mann-Whitney U tests, χ2 tests, and Pearson correlations were used to assess how the subscales were related to child age, child sex, parent sex, and duration of disease.
6. Mann Whitney U tests were run to compare subscale scores before visiting a doctor to after visiting a doctor (T1 compared to T2) and after 2 months of medical treatment (T1 compared to T3). Because these last analyses are run on a small sample size, no statistical significant differences are expected to be found. Rather, the results were reviewed for trends.
A total of 251 parents completed the online questionnaire. From this group, we excluded parents who reported that their child had a medical condition that can explain the FI such as Hirschsprung disease or developmental delays (N = 7); and did not soil in the last week (N = 6); and age <4 years (N = 6). This left us with a total of 232 parents of which 196 (84.5%) completed the 2-week follow-up. Demographics are given in Table 1.
Determining Subscales and Reliability
First, it was determined whether scores on items followed a normal symmetrical, bell-shaped distribution. Those who do not were eliminated from future analyses. All of the items had a minimum of 1 and a maximum of 5. Skewness (a measure of the degree of asymmetry of the normal distribution, with a value of 0 denoting symmetry) was between −2 and 2 for all of the items. Kurtosis (a measure of whether the data are peaked or flat relative to a normal distribution) was ≥2 for 4 items and these were deleted. Normal distribution was found for the remaining items.
Next principal component analysis was run to reveal whether items group together in similar constructs or components and to reduce the number of items. Principal component analysis revealed the existence of 2 factors in the data (Table 2). Factor 1 deals with the parental belief that FI occurs to get attention from others or in reaction to stimuli such as stress or relaxation. Parents punish their child when they have accidents. We label this factor “Blame and Punish” (Cronbach α 0.91; 2-week test-retest reliability 0.68). Factor 2 is related to the belief that FI has long-term negative consequences and the child may be forgetful about or not aware of having a bowel movement. Parents help their child by encouraging appropriate toilet use. We label this factor “Worry and Help” (Cronbach α 0.88; 2-week test-retest reliability 0.69). To make both subscales equal in length, 1 item with the lowest factor loading was removed from Blame and Punish, which did not change the internal consistency. Now both scales contain 15 items.
Establish Validity: Does Parental Knowledge Vary With Physician Consultation?
We expected the 2 subscales to vary with treatment choices. Consulting a physician was associated with increased scores on Worry and Help in comparison to no consultation (Fig. 1). In addition, scores on Worry and Help increased from primary to tertiary care. No significant differences were found on the Blame and Punish subscale.
In addition, the 2 subscales varied with medication use. Worry and Help was increased in those who report using over the counter laxatives (mean 50.2 vs 44.1; P < 0.001), fiber supplements (mean 49.2 vs 44.4; P < 0.01), enemas (mean 50.1 vs 45.7; P < 0.01), and prescription medication (mean 51.4 vs 43.1; P < 0.001), whereas Blame and Punish was decreased in those using fiber supplements (mean 26.1 vs 30.1; P < 0.05) or enemas (mean 24.3 vs 29.3; P < 0.05). Because most of these medications are used for the treatment of constipation, we assessed whether parents thought their child was constipated. As expected, parents who acknowledged their child was constipated (N = 138) scored higher on Worry and Help (mean 48.5) than those who did not believe that their child was constipated (N = 94; mean 41.0; P < 0.0001). No differences between these 2 groups were found for Blame and Punish.
The 2 subscales were not associated with most demographics such as child's age, sex, and duration of FI, which indicates that scores can be compared across all of the children. Compared to fathers, mothers scored higher on Worry and Help (mean 46.6 vs 41.6, P < 0.0001) and lower on Blame and Punish (mean 26.1 vs 32.5, P < 0.05); thus, scores in mothers may not be comparable to scores in fathers.
Validity: Does the Scale Capture Change in Parental Knowledge?
A new sample of 30 parents (93% mothers) was recruited for this part of the study. One parent was excluded because of failure to complete the baseline questionnaire at T1. A total of 29 patients completed T1 and T2 questionnaires and 24 completed all 3 questionnaires (82.8%). Child age ranged from 4 to 16 years (mean 7.8; SD 3.2). Children were predominantly boys (82.8%) and white (72.4% white, 24.1% African American). Mean FI duration was 4.7 years and mean number of FI was 6.6/week. Before treatment, 62.1% used fiber supplements, 82.1% laxatives, and 75.9% of parents believed their child was constipated.
Because the study 2 questionnaire was based on an interim analysis in study 1, not all of the items of the final 2 subscales were included. The Blame and Punish subscale consisted of 8 items and the Worry and Help of 11 items. In addition, 3 items in the Blame and Punish questionnaire showed low variability. More than 90% of parents endorsed “never” on 3 items, meaning there was no possibility to find reduction in scores after treatment. Therefore, we deleted these 3 items. Due to the difference in length between the questionnaires, mean scores were calculated rather than total sum scores.
No significant changes were observed between pre- and postvisit (T1–T2 and T1–T3) for the 2 subscales, but changes are in the expected direction. As seen in Figure 2, Worry and Help increased and Blame and Punish decreased after consulting a pediatric gastroenterologist. Given that our sample was small, we were underpowered to find significant differences; however, a trend was found for the Blame and Punish mean scores to reduce after treatment (P = 0.08). Clearly, data collection in a larger sample is needed before any definitive conclusions can be drawn.
The aim of the present study was to assess parental knowledge about FI and to determine whether parental knowledge is associated with treatment choices. We developed a 30-item questionnaire that consists of 2 subscales: Blame and Punish and Worry and Help. These subscales have good initial reliability, content validity, and seem to be amenable to change.
Blame and Punish is related to the belief that children respond to cues in their environment with soiling. Parents believe that children may soil to upset others, get attention, or because they are stressed, have emotional problems, ate the wrong thing, are relaxed, and so on. Children are seen as stubborn and uncooperative. Thus, parents believe children share some of the blame for their FI episodes and respond with punishment to try to change their child's behavior. Blame and Punish was not different between parents who did and did not seek care for FI. This may suggest that shying away from seeking medical help does not seem to be caused by a belief that the child is to blame for FI. Alternatively, the data may indicate that once parents do come in for treatment, physicians do not actively counsel parents to reduce punishment and blaming of the child. Unfortunately, the cross-sectional data do not allow us to determine which one of the above is more likely to happen.
Worry and Help captures parental anxiety about the effects of FI on their child's emotional and physical health as well as the fear that FI will continue for a long time. This was coupled to a belief that the child often ignores the urge to defecate and is not aware of soiling. These last two items may seem to show conceptual overlap with Blame and Punish: the parent is acknowledging that the child's behavior causes FI. However, in this case, parents do not punish the child but rather react with concern. Parents report to try to help the child by reminding them about using the toilet and rewarding toilet sitting or being clean. This finding, in addition to the association between Worry and Help and acknowledging the role of constipation in FI, suggests that parents understand the child may be in pain or feels anxious about having a bowel movement and therefore reacts with concern and help, rather than blaming the child for the FI.
As expected, Worry and Help was higher in families who consulted a physician for their child's FI. Although we have no knowledge about the information parents received from physicians, we can expect most physicians to educate parents on the role of constipation, the child's fearful reactions to bowel movements, and to encourage appropriate parental coping such as the use of consistent toilet sitting and positive reward. These components of physician education were reflected in Worry and Help. In addition, physician education may increase parental worry. Physicians often warn parents that treatment of FI will take months to years with many relapses and that stool impaction can lead to changes in the rectum, worries that were reflected in the Worry and Help subscale; however, it could also be argued that those parents who worry about long-term consequences and chronicity of symptoms are more likely to visit a physician. Clearly, worry about FI, a tendency to help the child, and physician consultation are linked, but cause and effect cannot be determined from the data.
The present study raises interesting associations between parental knowledge of FI and treatment choices as well as outcomes; however, several limitations restrict the interpretation of the study's findings. For the online study, the recruitment method may have skewed the sample toward parents who are interested in finding out about their child's FI. Indeed, we only had a limited number of parents who did not consult a doctor for their child's FI, indicating we may have missed a large group of parents who are less likely to consult any source of information in FI. Second, FI was determined in this online group by parental self-report; no physician diagnosis was available. Last, the online study had a large sample, but the retrospective nature of the study limits any conclusions about cause and effect.
In a small pilot study in a tertiary care clinic, we showed that our questionnaire is able to capture change in parental knowledge with physician education. Although changes seemed to occur mainly in Blame and Punish, several limitations of the present study preclude any definitive conclusion about the change in beliefs of parents. The present study is limited by its small sample size and short duration of follow-up. In addition, the education was not standardized. Pediatric gastroenterologists were instructed to provide their usual care. The content of the education may have varied widely across providers. Also, the providers were aware of the study, and it is possible that they may have provided more education and/or time with the patient than usual. Last, parents in tertiary care already have received education about FI from their primary physician. Therefore, replication is needed before any definitive conclusions can be drawn. For now, all of the data suggest that our questionnaire is able to measure changes in parental knowledge over time.
In sum, parental knowledge of FI centers on the belief either that children can be blamed for FI and should be punished for accidents or that FI is caused by ignoring the urge to defecate, and parental reminders as well as rewards can prevent long-term negative consequences. It is likely that most parents oscillate between these somewhat opposite views of FI. The present studies suggest that with physician consultation, understanding and helping the child is increased. These findings can help in developing educational materials for parents to encourage early diagnosis and treatment and prevent chronic problems that are difficult to manage.