Inclusion criteria followed the diagnostic criteria for childhood functional abdominal pain in Rome III [2]: (1) episodic or continuous abdominal pain at least once per week for at least 2 months before diagnosis; (2) insufficient criteria for other FGIDs; and (3) no evidence of an inflammatory, anatomic, metabolic, or neoplastic process that explains the patient’s symptoms.

The following criteria were tested to exclude other illnesses: (1) physical examination including normal growth pattern and (2) laboratory screening, including serological screening with immunoglobulin A antibodies against type 2 (tissue) transglutaminase (TGA-IgA), complete blood cell count, erythrocyte sedimentation rate or C-reactive protein, and urine analysis. A fecal calprotectin test was also included if the girl reported symptoms such as diarrhea.

Diagnostic criteria for IBS in Rome III [2] were used. First, at least once per week for at least 2 months before diagnosis, the patient must have had abdominal discomfort or pain associated with 2 or more of the following at least 25% of the time: (1) improved with defecation, (2) onset associated with a change in the frequency of stool, or (3) onset associated with a change in the form (appearance) of stool. Second, there must be no evidence of an inflammatory, anatomic, metabolic, or neoplastic process that explains the patient’s symptoms.

The following criteria were tested to exclude other illnesses: (1) physical examination including normal growth pattern and (2) laboratory screening including serological screening with TGA-IgA, complete blood cell count, erythrocyte sedimentation rate or C-reactive protein, and urine analysis. A fecal calprotectin test was also included if the girl reported symptoms such as diarrhea.

Persistent pain after examination at the pediatric center was measured with a pain diary at baseline. Girls who reported one or more episodes of pain with a pain score of 4 or higher on the faces pain scale-revised (FPS-r) [76,77] during a full week were eligible for the study.

The exclusion criteria for this trial were (1) contemporaneous celiac or inflammatory bowel disease; (2) difficulty following oral instructions, such as hearing impairment, mental retardation, or language difficulties; (3) simultaneous treatment with CBT; and (4) severe psychological symptoms for which other treatment is needed.

Participants were recruited from outpatient clinics, primary health care services, and the public.

We recruited participants from the outpatient clinics of the pediatric departments at the university hospitals in Örebro, Karlskoga, and Lindesberg, Sweden. In the second recruitment year, participants were also recruited from the pediatric outpatient clinic in Västerås, Sweden. All girls aged 9 to 13 years (9 to 12 years during the first year of the study) who had visited one of the outpatient clinics during the previous 2 years because of abdominal pain and had received a diagnosis of IBS, functional abdominal pain, constipation, or abdominal pain received an information letter asking whether they wanted to join the study. All participants were found via local diagnosis registries at the included hospitals. After written consent was received from the legal guardians, the diagnosis according to the inclusion criteria was verified in the medical records or by examination by a pediatrician.

In addition, we recruited girls from primary health care services in the region of Örebro and Västerås. Information letters were delivered to potential participants and their legal guardians in primary care, the counselling unit, and school health services. Interested girls and their legal guardians contacted the research team, who distributed an information letter and the consent form. After written consent was obtained, the research team booked an appointment for an examination at the pediatric clinic (Örebro) or research clinic (Västerås).

Information about the study project was provided by several information channels, such as social media, ordinary media, and websites. Interested girls and their legal guardians contacted the research team, who distributed an information letter and the consent form. After written consent was obtained, the research team booked an appointment for an examination at the pediatric clinic (Örebro) or research clinic (Västerås).

Girls and their legal guardians who consented to the study and were eligible for the study according to the inclusion criteria completed the baseline measurement.

When the baseline measurement was completed and the eligibility check was accomplished, the sample was randomized into an intervention group and a control group. Randomization was performed by an external statistician using minimization based on pain intensity and age at baseline [78].

The total number of required study participants was calculated for the primary outcome (decreased maximal pain after 8 months). The calculation was based on several assumptions. First, we considered the expected proportion of participants who decreased their maximal pain after 8 months in the intervention group. We estimated that 30% of the intervention group would decrease their maximal pain after 8 months based on previous research with CBT, in which approximately 50% of participants were pain free immediately after the intervention [79]. Second, we calculated the expected proportion of participants whose maximal pain would decrease after 8 months in the control group. In previous CBT studies, approximately 25% of control participants reported being pain free when the intervention ended [79]. Since only girls with a long duration of abdominal pain were included in the present study, we estimated that the placebo effect would be smaller and that 10% of the girls in the control group would report decreased maximal pain at the 8-month follow-up. Third, dropout is a common problem that can bias outcomes. A 20% to 50% dropout rate is commonly reported in studies of participants with long-lasting pain [80]. In the present study, a 20% dropout rate was estimated.

With a power of 80% and a significance level of .05, we calculated the sample size to be N=150 (75 intervention + 75 control participants), including a 20% dropout. To decrease bias, a minimum of 50 participants in each group is recommended in trials studying intervention effects on pain, according to previous literature [80].

The intervention is called Just in TIME and addresses the importance of early intervention for this target group. The word “TIME” stands for “Try, Identify, Move, and Enjoy,” which also characterizes the key aspects of the intervention, which aims to decrease FAPDs among 9- to 13-year-old girls through dance and yoga. Essential elements of the intervention included a focus on enjoyment, socialization, and playful creativity in an undemanding environment. We chose a combination of dance and yoga because of experiences from an earlier study [38] that targeted adolescent girls with internalizing problems and evaluated a dance intervention. The study participants appreciated the intervention and rated it to be a positive experience [38], but at the end of the intervention period, they started to request yoga as an add-on at the end of the sessions. They believed that it would be a valuable closure to the session to help them tune in to the relaxation. These requests were a form of evaluation that the research team took into account when designing the current study.

Informational material was distributed during the training of the intervention providers. It consisted of written course materials and visually recorded practical choreography and sequences, all provided by the intervention coordinator.

The intervention was performed as a group activity twice a week after school hours during an 8-month period. The duration of the class was 60 minutes, comprising 30 minutes of dance practice, 25 minutes of yoga including relaxation, and 5 minutes of short reflection (Table 1). Throughout the intervention period, the participants were encouraged to practice their favorite dances, yoga poses, or relaxation techniques at home if they wanted to. Both centers followed the same routine.

Each group included 7 to 14 girls and was under the guidance of 2 instructors, one at a time. The instructors had a profession in either health care or pedagogy, had experience working with young people, and were educated in teaching children or adolescents in dance (instructor training) and yoga (ie, registered yoga teacher 200-hour training). Prior to the start of the intervention, all instructors attended a 2-day course administered by the research team. This course consisted of practical instructions from the intervention coordinator about the dance and yoga session according to the standardized program design, including dance choreographies and yoga sequences adjusted to the target group as well as teaching style (nonjudgmental and supporting approach). Lectures about theories and underlying mechanisms and about guidelines and ethical aspects regarding children with FAPDs were also given.

Various styles of dance with a focus on enjoyment were performed during the intervention period. The yoga practice consisted of both playful movements (such as creative yoga storytelling) and calm physical postures with a focus on breathing and attention, which were performed individually, in pairs, and as a group. The intervention (further described in Table 1) was developed by the intervention coordinator (author AD). To ensure standardization in program delivery across both centers and between instructors over time, booster sessions were given 3 times during the intervention period (in addition to the initial 2-day course). The participants in the intervention were advised to wear practical soft clothes suitable for the activity, and no special shoes were needed.

The intervention took place in studios located in the center of Örebro and Västerås, Sweden, which were easily accessible by bus or walking. The studios had yoga mats and music equipment.

The 8-month period was chosen because it corresponded to 2 school semesters. No classes were held during holidays; thus, 50 dance classes were held over 25 weeks. The practice can be classified as moderate-to-vigorous physical activity.

To address modifications, both dance and yoga movements were introduced in steps to include all participants regardless of previous experience. Alternatives were given when needed. No changes were made in the intervention.

Adherence to the dance and yoga intervention was noted by the instructors. A strategy to maintain and improve fidelity was to consciously focus on relatedness and provide a feeling of social inclusion, supportiveness, and the chance to meet new friends. To keep the dance and yoga sessions interesting over the time period, a variety of styles and themes were presented, all distributed by the intervention coordinator. The intervention was delivered as planned throughout the 3 intervention years.

The primary outcome is the proportion of girls in each group who have reduced maximal pain, measured immediately after the intervention (8-month follow-up) with the FPS-r [76,77] in a pain diary in which the girls register their abdominal pain 3 times a day for 1 week. The psychometric features for the FPS-r are reported to be good [77].

To assess covariates, we asked about (1) demographics; (2) parental abdominal and gastrointestinal morbidity; (3) the girl’s background, including mode of delivery (caesarean or vaginal), antibiotic consumption early in life, dance and yoga experiences, menarche, etc; and (4) safety at home, in school, during leisure time activities, on social media, and at other places were the girls spend their time.

The Just in TIME study follows the intention-to-treat paradigm. Complementary analyses will be performed per protocol and according to the number of sessions the participant attends.

For baseline statistics between the groups, descriptions will be constructed using frequencies and proportions for categorical data and means and standard deviations for continuous variables. Baseline characteristics will be compared using the Fisher exact test for dichotomous outcomes, the Mantel-Haenszel chi-square test for ordered categorical outcomes, and the Mann-Whitney U test or unpaired t test for continuous outcomes. A P value of <.05 for the 2-tailed test will be considered statistically significant for all outcomes. When deemed necessary, correction for multiple significance will be performed.

The primary outcome analysis will initially be performed using a Fisher exact test to evaluate differences between groups. To analyze the change in scores from baseline between groups for approximately normally distributed variables, a repeated-measures covariance pattern mixed model will be used, adjusting for significant differences at baseline. To identify predictive factors at baseline that might be associated with the primary outcome, a univariable logistic regression analysis followed by a stepwise multiple logistic regression analysis will be performed.

The distribution of continuous variables will be described using the mean and standard deviation or median and interquartile range. Categorical variables will be described with numbers and percentages.

Study participants who choose to leave the intervention before the study is finished are encouraged to complete all assessments. Missing data will be handled with multiple imputations or within the mixed model.

To study the difference between morning and evening values, the evening-morning cortisol quotients will be calculated by dividing each girl’s evening cortisol value by the morning cortisol value. To study potential changes in evening-morning cortisol quotients between baseline and 4 and 8 months, the median change score will be calculated by subtracting baseline cortisol morning samples from 4-month morning cortisol samples and 8-month morning cortisol samples, and the same procedure will be performed for the evening samples.

Participants received a kit containing everything needed to perform the fecal test at home. The test was stored in the freezer at home for approximately 1 week and was subsequently delivered to the research team in a small freezer bag. The samples are stored at –80 ⁰C at the university hospital in Örebro, and analyses of the intestinal microbiota and short-chain fatty acids will be performed.

The frequency of the accelerometer collecting data will be set to 30 Hz. Data will be processed in ActiLife (version 6.13.4; ActiGraph). Raw data from the accelerometers will be converted to 10-second epochs. Valid wear time will be considered to be at least 480 minutes per day, and minimum valid days will be set to 4 days, including 2 weekdays and 1 weekend day. Total counts per day, sedentary time, and activity of moderate or higher intensity will be measured.

The interviews will be transcribed verbatim and analyzed with inductive content analysis, as described by Elo and Kyngäs [109]. The analysis process will initially include getting to know the material, generating codes, and identifying categories. Data will be analyzed using the NVivo software program (QSR International) for qualitative data analysis [110]. All of the authors will discuss the coding process until it is agreed upon and the codes will be grouped into subcategories that reflect the core message of the interviews. Thereafter, the subcategories will be abstracted into generic and main categories [109].

The health economic evaluation will be performed as a cost-utility analysis using individual data [111]. Societal costs (including health care, informal care, and school health care) will be considered. Gained quality-adjusted life-years (QALYs) will be used to measure the effects. QOL will be measured with Kidscreen-10 [104], and an algorithm for mapping the Kidscreen-10 index onto the CHU9D utility scores will be used [105]. Questions about resource use and QOL are asked to both girls and their legal guardians at all follow-ups. Cost-effectiveness ratios will be based on the changes in QALY and net costs for the intervention group compared with the control group. The results will be presented as an incremental cost-effectiveness ratio (ICER), which is expressed as ICER = (Ca – Cb)/(Ea – Eb), where “Ca” is the cost of the intervention, Cb is the cost of the comparator, Ea is the effectiveness of the intervention, and Eb is the effectiveness of the comparator.