An infographic is available for this article at : https://links.lww.com/MPG/C426
What Is Known/What Is New What Is Known
Reliable and valid assessment of stool consistency is important for evaluating defecation patterns in toddlers and diagnosing gastrointestinal disorders.
Reliability and validity has not been evaluated in current stool scales on defecation patterns in toddlers that may include stools from a diaper as well as from a toilet/potty.
What Is New
The modified Bristol Stool Form Scale was reliable, valid, and user-friendly to use to evaluate stool consistency in both toilet- and nontoilet-trained toddlers in the Netherlands.
To the best of our knowledge, this is the first stool scale that has been validated to score stool consistency in both toilet/potty as well as in diapers. This will be of value for monitoring bowel habits in young children in clinical research for food ingredients, medicines, or lifestyle changes that can impact this parameter at a critical age of toilet training.
The reliability and validity of the modified Bristol Stool Form Scale has been assessed in caregivers.
Alterations in stool frequency and stool consistency are associated with numerous organic and functional gastrointestinal (GI) disorders in children, such as irritable bowel syndrome (IBS), inflammatory bowel disease (IBD), and functional constipation (FC). These diseases may be so serious that they affect quality of life of the children and their parents (1–3) constipation (4,5) Bristol Stool Form Scale (BSFS) and the modified-BSFS (m-BSFS) but also scales that allow for scaling stools in diapers, such as the Amsterdam infant stool scale (AISS) and Brussels infant and toddler stool scale (BITSS) (6–9) https://links.lww.com/MPG/C373
MATERIALS AND METHODS
Translation and Cultural Adaptation of the Dutch Modified Bristol Stool Form Scale
Published guidelines were followed to translate and to culturally adapt the English version of the m-BSFS to Dutch (10–13)
Participants
The Medical Ethical Reviewing Committee of Wageningen University (METC-WU) reviewed the research file and concluded that this research does not fall within the remit of the Dutch ‘Medical Research Involving Human Subjects Act’. Following International Conference on Harmonization—Good Clinical Practice (ICH-GCP) guidelines, however, informed consent was obtained. Participants were eligible if they were a parent of at least 1 child of 1 to 3 years old (12–36 months), day childcare employees working with toddlers of 1 to 3 years old, or grandparents of at least 1 grandchild of 1 to 3 years old. We sought to include approximately n = 100 participants, based on published recommendations, for the 32 general stool pictures in the ratio 3 : 1 : 1 for parents, grandparents, and day childcare employees, respectively (11)
Inter- and Intra-rater Reliability
The same 32 color pictures of stools as were used to initially evaluate and re-evaluate the m-BSFS by pediatric gastroenterologists and children of 3 to 18 years of age by Chumpitazi et al (9) (14) (8)
For the validation, interrater reliability was used as a measure for agreement between raters and intra-rater reliability as a measure for agreement within 1 person between the first and second time of scoring the pictures. To assess inter- and intra-rater reliability of the general stool pictures, participants were asked to complete a questionnaire that was built in the online platform Castor EDC (15)
For the stools in diaper pictures, inter- and intra-rater reliability was assessed in participants for 7 focused, close-up color images of bowels in diapers. Participants were asked to fill out the questionnaire twice with at least 1 week between the first and second scoring and again the order of the pictures was different for both questionnaires to avoid bias.
Paper Versus Online Use and User-friendliness in a Real-life Situation
For the scoring of the stools in diaper pictures, the participants were randomly assigned to scoring the stools in diapers either in a paper version or in an online version to investigate whether this would impact the ICCconsistency . Participants were also asked to fill out a 1-week diary, in which they scored all bowel movements of the child. Thus the parent, grandparent, or day childcare employee scored the bowel movements of the child(ren) for which they were present. Moreover, 3 statements were added to assess the user-friendliness, clarity of the instructions, and feasibility to use the m-BSFS on fresh stool samples in real-life situations. Each category for user-friendliness was scored on a 5-point Likert scale ranging from strongly agree to strongly disagree (16)
Statistical Analysis
Statistical analyses were performed with R software, version 3.6.1 using the “irr” package version 0.84.1 and the “ICC.Sample.Size” package version 1.0 (17–19) consistency was used for inter-rater reliability with the “icc’ function, whereas Cohen κ (function “kappa2” was used for intra-rater reliability (11,20) (19) (21)
RESULTS
Translation and Cultural Adaptation of the Dutch Modified Bristol Stool Form Scale
The original English m-BSFS and the final translated Dutch version of the m-BSFS as used in the validation study are presented in Figure 1 .
FIGURE 1: Left: Dutch version of the modified
Bristol Stool Form Scale , right: the original English m-BSFS. Data from
(13) . m-BSFS = modified
Bristol Stool Form Scale .
Participants
In total, 93 participants completed the questionnaire on the general stool pictures, of whom 69% were parents, 20% grandparents, and 11% day childcare employees (Fig. 2 ). A total of 16 participants, constituting 69% parents, 19% grandparents, and 12% day childcare employees, completed the stools in diaper pictures and the user-friendliness questions in full. Three participants did not complete the questionnaire without giving reasons and 1 participant indicated that viewing and rating the pictures caused nausea. Participants reported to look after 1 to 2-year-old (52%) and 2 to 3-year-old toddlers (48%), 55% girls, and 45% boys, of whom 15% were completely toilet-trained, 7% only during the day, 1% only for urine, and 77% were nontoilet-trained.
FIGURE 2: Flow chart of participants in the validation study.
Inter- and Intra-rater Reliability
Out of a total of 4505 ratings for both general stool pictures and stools in diaper pictures, 3505 (77.8%) were in agreement with the most commonly chosen rating, and 4272 (94.8%) were within 1 form type of the most commonly chosen rating for each stool picture. More specifically, for the general stool pictures, 3349 out of 4288 ratings (78.1%) were in agreement with the most commonly chosen rating, while this was the case for 156 out of 217 ratings (71.8%) for the stools in diaper pictures. For the general stool pictures, 4055 out of 4288 (94.6%) were within 1 rating from the most commonly chosen rating, compared with 217 out of 217 (100.0%) for the stools in diaper pictures.
The proportions of exact agreement of each individual picture are presented in Figure 3 A and B for the general stool and stools in diaper pictures, respectively. Of the 32 general stool pictures, 3 pictures were most commonly scored as type 1, 8 as type 2, 8 as type 3, 8 as type 4, and 5 as type 5 (Fig. 3 ). For the 7 stools in diaper pictures, 1 was most commonly scored as type 1, 2 as type 2, 1 as type 3, 1 as type 4, and 2 as type 5 (Fig. 3 B). Concerning the percentage of ratings that were in concordance with each other, the 3 weakest performing general stool pictures corresponded to m-BSFS type 4 (43%), type 5 (46%), and type 3 (47%). For the stools in diaper pictures, the 3 weakest performing pictures were m-BSFS type 4 (48%), type 3 (52%), and type 1 (52%).
FIGURE 3: Pictures are ordered to agreement on type. Participants received the pictures in random order. Colors of this figure are a gradient ranging from dark blue as highest values, light blue as lowest values. (A) Proportions of exact agreement for each of the 32 general stool pictures. (B) Proportions of exact agreement for each of the 7 stools in diaper pictures. m-BSFS = modified Bristol Stool Form Scale .
The a posteriori power calculation shows that, for a power of 0.8 and an α of 0.05, sample sizes for both general stool pictures and the stools in a diaper pictures numbers were above the sample size as used in this study (Table 1 ). To further support this, the subject to item ratio was calculated. For the general stool pictures, this subject to item ratio was 2.8 for the first time and 2.4 for the second time and for the diaper-specific pictures this was 2.3 for the first time and 2.1 for the second time.
TABLE 1 -
Intraclass Correlation Coefficient
consistency and Cohen κ for inter- and intra-rater reliabilities of the modified
Bristol Stool Form Scale
ICC/kappa
95% CI
Sample size boundary for power = 0.8
General stool pictures
First time, n = 89
0.716
0.617 to 0.818, P < 0.001
57 to 64
Second time, n = 77
0.793
0.710 to 0.872, P < 0.001
20 to 22
Kappa, n = 77
0.706
P < 0.001
n.a.
Stools in diaper pictures
First time, n = 16
0.925
0.828 to 0.984, P < 0.001
6
Second time, n = 15
0.934
0.847 to 0.986, P < 0.001
6
Kappa, n = 15
0.769
P < 0.001
n.a.
CI = confidence interval; ICC = intraclass correlation coefficient.
In conclusion, all inter- and intra-rater reliability scores were above the most commonly used thresholds of >0.61 or >0.7 (Table 1 ). Inter- and intra-rater reliabilities were found to be higher for the stools in diaper pictures, compared with the general stool pictures.
Paper Versus Online Use and User-friendliness in Real-life Situations
A comparison of the paper (n = 10) versus the online (n = 6) ICCconsistency for the stools in diaper pictures revealed strong ICCconsistency of 0.94 for both the online version (0.94) and paper version (0.93). To assess significance of the difference between these 2 ICCs, a 2-sided Fischer r-to-z transformation was used, resulting in a P value of 0.90, indicating no statistical significance between the ICC of the paper and the online version.
In terms of user-friendliness by means of the 5-point Likert scale, all 16 participants, experienced the m-BSFS as user friendly based on the responses for clarity of instructions, which were “neutral” for 10%, whereas 60% answered “agree” and 30% “strongly agree.” In response to the following translated statements, “I think it is demanding to use the m-BSFS. For example, it takes effort to remember to use the m-BSFS or it was demanding to compare fresh stools to the m-BSFS,” 20% of the participants answered with “strongly disagree,” 20% with “disagree” and 60% with “neutral.” In conclusion, both the paper and online version of the m-BSFS were largely considered user-friendly by the study participants.
DISCUSSION
The objective of this study was to assess whether the m-BSFS is reliable, valid, and user-friendly to use by parents, grandparents, and day childcare employees to evaluate stool consistency in both toilet- and nontoilet-trained toddlers in the Netherlands. Overall, the m-BSFS was successfully translated and culturally adapted to Dutch, and showed to have a high degree of inter-rater reliability, intra-rater reliability and user-friendliness, regardless of whether this was on paper or online. This scale is the first that can be used for rating stools both from a diaper as well as from a potty or toilet.
We showed that both inter- and intra-rater reliability were above thresholds of >0.61 or >0.7, as recommended by published guidelines (11,22–24) (1,14) (9) (14)
To the best of our knowledge, this is the first stool scale that has been validated for scoring stool consistency in stools in a potty/toilet and also for stools in a diaper. The added value of this validation is that data of potty-trained children can be directly compared with nonpotty-trained children and can also be used in a period when children are being potty-trained. Evaluation of the stools in diaper pictures resulted in a high ICC. Related to these findings concerning our 7 stools in diaper pictures and the findings by Lane et al (14) (25) (20,25) et al to our data but also for comparing the ICCs of the general stool pictures to the stools in diaper pictures. As the ICCs for our stools in diaper pictures are, however, very high, ranging from 0.925 to 0.934, we are confident that, even with the possible bias described above, our questionnaire is valid and well above the thresholds of 0.61 or 0.7 as the computations did not show a bigger difference than 0.06 (11,22–24)
Moreover, it has been suggested that stool form, as measured, for example, by the m-BSFS, is a proxy for colonic transit rate (6,26)
In addition, most stool scales have previously been completed on paper while currently there is a shift towards more online-based methods (27) (27) consistency . In short, when the m-BSFS is used in practice, caregivers can use the paper or online version, or even mix and match during research or management of GI diseases or functional GI disorders to the method that suits best at that moment. Furthermore, in general, participants indicated that the m-BSFS is user friendly, with clear user instructions and not very demanding to use.
Strengths of this study includes that parents, grandparents, and day childcare employees were divided as proposed on forehand (3 : 1 : 1, respectively). Moreover, division of the toddlers in terms of gender and age ranges was close to 50 : 50, indicating a good representation of the population in which this stool scale might be used. Moreover, we did an a posteriori sample size calculation to check whether the number of participants was sufficient. By using our own data, that is, the actual ICCs found in this study, the power calculation is even more reliable as it is not an estimate based on comparable studies or study populations. We used the function ‘calculateIccPower’ from the ICC.Sample.Size package in R, which calculates a post hoc power from an ICC. This function demonstrates the additional power gained by increasing the number of subjects or the number of subjects needed to increase power. In addition, it determines the number of participants needed for a specific power. Consequently, we can conclude, based on the results from the “calculateIccPower” function, that the sample size boundary for a power of 0.8 would be 6 participants, as shown in Table 1 . Our conclusion is, therefore, valid based on the a posteriori power calculation as well as confidence intervals of which even the lowest boundary is well above the most commonly used thresholds of 0.61 or 0.7. In addition, the subject to item ratios were between 2.8 and 2.1 in this study, which is according to Anthoine et al (21)
Considering limitations of this study, the 32 pictures generously provided by Chumpitazi et al are at least 10 years old and a few of which are relatively low in resolution and quality. The extra stools in diaper photographs helped to address this. This potentially leads to a lower ICCconsistency or Cohen κ, not because of a true difficulty to discriminate the type of stool but because of the resolution and quality of the images. Possibly in line with this, we observed a higher ICCconsistency and Cohen κ for the 7 stools in diaper samples. The latter pictures were more comparable to each other in terms of size, lighting and photographic composition but were also of higher resolution and quality, which might be a factor influencing ICCconsistency and Cohen κ. Demographic data and socioeconomic status of the participants in our study are lacking, and it is, therefore, unknown if this sample is representative of the general population. We, however, do not expect that differences in demographics will significantly change the results of this study.
In conclusion, the modified m-BSFS, as paper or online version, is reliable, valid and user-friendly to use for Dutch-speaking parents, grandparents, and day childcare employees to evaluate defecation parameters in toddlers whether in diapers or toilet-trained. This validated m-BSFS is likely to prove useful in both clinical and research settings as a validated measure to record stool form from both diapers and toilet or potty. The m-BSFS can be used in clinical practice and clinical trials as tool for diagnosis, management, and evaluation of bowel patterns in healthy toddlers as well as in several disorders, such as functional constipation (FC), functional diarrhea (FD), or irritable bowel syndrome (IBS).
Acknowledgments
The authors would like to thank all participants, Bruno P. Chumpitazi for generously providing us with the 32 general stool pictures, Maria Reumerman and Floriane Brunel for their assistance during this study and the translators Esther van Raamsdonk, Peter van Elswijk, James C. Dykstra, and Michael Rose.
REFERENCES
1. Hyams JS, Di Lorenzo C, Saps M, et al. Childhood
functional gastrointestinal disorders : child/adolescent.
Gastroenterology 2016; 150:1456–1468. e2.
2. Herzog D, Fournier N, Buehr P, et al. Swiss IBD Cohort Study Group. Age at disease onset of inflammatory bowel disease is associated with later extraintestinal manifestations and complications.
Euro J Gastroenterol Hepatol 2018; 30:598–607.
3. Drossman DA.
Functional gastrointestinal disorders : history, pathophysiology, clinical features, and Rome IV.
Gastroenterology 2016; 150:1262–1279. e2.
4. Wegh CA, Schoterman MH, Vaughan EE, et al. The effect of fiber and prebiotics on children's gastrointestinal disorders and microbiome.
Expert Rev Gastroenterol Hepatol 2017; 11:1031–1045.
5. Firmansyah A, Chongviriyaphan N, Dillon DH, et al. Fructans in the first 1000 days of life and beyond, and for pregnancy.
Asia Pac J Clin Nutr 2016; 25:652.
6. Lewis SJ, Heaton KW. Stool Form Scale as a Useful Guide to Intestinal Transit Time.
Scand J Gastroenterol 1997; 32:920–924.
7. Ghanma A, Puttemans K, Deneyer M, et al. Amsterdam Infant Stool Scale is more useful for assessing children who have not been toilet trained than Bristol Stool Scale.
Acta Paediatr 2014; 103:e91–e92.
8. Huysentruyt K, Koppen I, Benninga M, et al. BITSS working group. The Brussels Infant and Toddler Stool Scale: a study on interobserver reliability.
J Pediatr Gastroenterol Hepatol Nutr 2019; 68:207–213.
9. Chumpitazi BP, Lane MM, Czyzewski DI, et al. Creation and initial evaluation of a Stool Form Scale for children.
J Pediatr 2010; 157:594–597.
10. Lohr KN. Assessing health status and quality-of-life instruments: attributes and review criteria.
Qual Life Res 2002; 11:193–205.
11. Terwee CB, Bot SDM, de Boer MR, et al. Quality criteria were proposed for measurement properties of health status questionnaires.
J Clin Epidemiol 2007; 60:34–42.
12. Tsang S, Royse CF, Terkawi AS. Guidelines for developing, translating, and validating a questionnaire in perioperative and pain medicine.
Saudi J Anaesth 2017; 11: (Suppl 1): S80–S89.
13. Sousa VD, Rojjanasrirat W. Translation, adaptation and validation of instruments or scales for use in cross-cultural health care research: a clear and user-friendly guideline.
J Eval Clin Pract 2011; 17:268–274.
14. Lane MM, Czyzewski DI, Chumpitazi BP, et al. Reliability and validity of a
modified Bristol Stool Form Scale for children .
J Pediatr 2011; 159:437.e1–441.e1.
15. Castor Electronic Data Capture, Ciwit BV, Amsterdam, The Netherlands. 2019.
16. Likert R. A technique for the measurement of attitudes.
Arch Psychol 1932.
17. R Core Team. (2019) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
https://www.R-project.org/ . [Accessed 25 September 2019]
18. Matthias Gamer, Jim Lemon, Ian Fellows and Puspendra Singh. irr: Various Coefficients of Interrater Reliability and Agreement. R package version 0.84.1.
https://CRAN.R-project.org/package=irr . 2019. [Accessed 25 September 2019]
19. Alasdair Rathbone SS, Kumbhare D. ICC.Sample.Size: Calculation of Sample Size and Power for ICC. R package version 1.0.
https://CRAN.R-project.org/package=ICC.Sample.Size . 2015. [Accessed 25 September 2019]
20. Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research.
J Chiropr Med 2016; 15:155–163.
21. Anthoine E, Moret L, Regnault A, et al. Sample size used to validate a scale: a review of publications on newly-developed patient reported outcomes measures.
Health Qual Life Outcomes 2014; 12:176.
22. Landis JR, Koch GG. The measurement of observer agreement for categorical data.
Biometrics 1977; 33:159–174.
23. Cicchetti DV, Sparrow SA. Developing criteria for establishing interrater reliability of specific items: applications to assessment of adaptive behavior.
Am J Ment Defic 1981; 86:127–137.
24. Fleiss JL, Levin B, Paik MC. Statistical Methods for Rates and Proportions. New York: John Wiley & Sons; 2013.
25. Bartko JJ. The intraclass correlation coefficient as a measure of reliability.
Psychol Rep 1966; 19:3–11.
26. Degen L, Phillips S. How well does stool form reflect colonic transit?
Gut 1996; 39:109–113.
27. Dewaele JM. Online Questionnaires. In: Phakiti A., De Costa P., Plonsky L., Starfield S. (eds) The Palgrave Handbook of Applied Linguistics Research Methodology. Palgrave Macmillan, London. 2018.
https://doi.org/10.1057/978-1-137-59900-1_13 .
