Strategies used for childhood chronic functional constipation: the SUCCESS evidence synthesis

Planned stakeholder events and changes due to COVID-19

The original plans had been to hold three face-to-face meetings, each occurring over 1 or 2 whole days, in Glasgow. Each of these events would have incorporated a series of meetings, discussions, presentations and practical exercises interspersed with breaks and opportunities for informal social interaction. The first of these planned events was held (January 2020), but due to COVID-19 restrictions and concerns all remaining meetings were conducted online via Zoom. The move to online meetings led to changes to the planned meeting format, with more frequent, shorter, meetings.

Stakeholder meetings

Appendix 1 provides an overview of the SG meetings held throughout the SUCCESS project.

Meeting participants

The majority of meetings were open to all SG and research team members; however, one meeting was held (13 May 2020) at which only the public members and one researcher attended. This was held soon after the move to online meetings (due to COVID-19 lockdown restrictions) and one of the reasons for this meeting was to explore whether separate meetings for public members would ensure that the public members had sufficient opportunity to contribute. The decision was made that it was preferable to have all SG members invited to all meetings, but it was highlighted that careful chairing was required to ensure that everyone had an equal opportunity to contribute at meetings. One further meeting was held with the public members only (12 April 2021); this was held specifically to explore an idea that public members had about leading a publication relating to PPI in this project.

Meeting conduct

Meetings adhered to key principles of research co-production,29 creating an environment that recognised everyone’s contributions and in which people worked together to achieve a shared understanding. Agendas and meeting material were circulated prior to the meetings. Meetings were chaired by one of the research team. Agendas were planned to ensure substantial time for discussion, and during Zoom meetings the ‘hands-up’ function was used to ensure everyone got the opportunity to speak. The ‘chat’ function was also used to support ‘side’ conversations, and to gain additional comments and questions, during the meetings. Meetings were recorded and a member of the research team listened back to the discussion when writing meeting notes, capturing the outcomes of activities and/or writing sections of the final report.

Electronic searches

We developed a comprehensive search strategy based on previous published searches. 19 The search strategy was peer-reviewed by an information specialist (JC) in accordance with peer review of electronic search strategies (PRESS) guidelines44 and the search string for medical literature analysis and retrieval system online (MEDLINE) is shown in Appendix 3. Other searches are available in Report Supplementary Material 2.

Searches were adapted for each of the following major electronic databases:

• MEDLINE and PREMEDLINE (Medical Literature Analysis and Retrieval System Online);

• Cochrane library databases including CENTRAL (Cochrane Central Register of Controlled Trials), CDSR (Cochrane Database of Systematic Reviews);

• EMBASE (Excerpta Medica database);

• CINAHL (Cumulative Index to Nursing and Allied Health Literature);

• PsycINFO;

• AMED (Allied and Complementary Medicine Database);

• Archived databases: DARE (Database of Abstracts of Reviews of Effects), HTA (Health Technology Assessment) database, NHS EED (NHS Economic Evaluation Database) will be accessed at www.crd.york.ac.uk/CRDWeb/.

Clinical trial registries:

• World Health Organisation (WHO) International Clinical Trials Registry Platform (ICTRP) (www.who.int/clinical-trials-registry-platform);

• OpenTrials (https://opentrials.net);

• NIH US National Library of Medicine ClinicalTrials.gov (https://clinicaltrials.gov).

Electronic bibliographic databases and clinical trial registers were searched from January 2011 (unless otherwise indicated) to 10 March 2020. No language restrictions were applied. The electronic database search date was set from January 2011 following explicit comments from funders during the development of the research proposal. In order to avoid substantive duplication of searching previously conducted for NICE (2012) updated guidelines,19 which had a search date of 3 February 2012, the search date of January 2011 was agreed on.

Searching other resources

We conducted several supplementary searches in order to identify other potentially relevant studies. Dates of when each of the searches were conducted and any date limitations applied to this grey literature is documented in Report Supplementary Material 2. This included searching the following grey literature:

• Google Scholar (https://scholar.google.com/);

• Grey Matters (www.cadth.ca/grey-matters-practical-tool-searching-health-related-grey-literature);

• Open Grey Repository (www.opengrey.eu/);

• Physiotherapy Evidence Database (PEDro) (https://pedro.org.au/);

• OTseeker (www.otseeker.com/);

• PROSPERO International Prospective Register of Systematic Reviews (www.crd.york.ac.uk/prospero/);

• Online social media platforms identified by members of our SG including Netmums, Scope, Challenging Behaviour Foundation forum.

We also searched relevant journals, conferences, guidelines and websites. This included:

• International Continence Society (www.ics.org);

• Digestive Disease Week (https://ddw.org);

• United European Gastroenterology Week and the European Society for Paediatric Gastroenterology (https://ueg.eu);

• European Society for Paediatric Gastroenterology, Hepatology and Nutrition (www.espghan.org);

• North American Society of Pediatric Gastroenterology and Nutrition archives (https://naspghan.org);

• British Journal of School Nursing (www.magonlinelibrary.com/loi/bjsn);

• ProQuest Dissertations and Theses.

We also contacted experts, authors of eligible papers, manufacturers, national and international professional organisations and bodies (e.g. British Dietetic Association, School and Public Health Nurses Association) who are involved in the management of CFC to identify other relevant published and unpublished studies. Finally, we looked for studies included within published guidelines, including NICE guidelines1,2,19,20 and European guidelines,45 relevant SRs and reference lists.

Excluded systematic reviews

Any SRs which did not meet the inclusion criteria were excluded and listed in a table of excluded studies. Papers which had been tagged as SRs during the scoping review, but on appraisal, were found not to have SR methods (i.e. a comprehensive search, systematic selection, data extraction and ROB assessment) were automatically considered to be high ROB (or unclear ROB), and were classed as ‘narrative reviews’ and excluded.

Excluded randomised controlled trials

Randomised controlled trials judged found not to meet the inclusion criteria were excluded, with details listed in a table of excluded studies.

Ongoing randomised controlled trials

Where the identified paper was a protocol or study registration, a search was done for any reports of the completed study. If no reports were found, the study was excluded from the review, and details listed in a table of ongoing studies. The search for reports of completed studies was repeated (between 1 May 2022 and 1 June 2022) during the evidence synthesis write-up stage to ensure completed study data were not missing from the final synthesis.

Excluded primary research studies

Where a non-randomised primary research study was judged to address the same, or a similar, question as an included SR or as a number of RCTs found, that study was excluded, and not included in the narrative synthesis.

What is the effect of nurse-led models of care as compared to alternative models of care?

Four studies58,59,60,61 addressed this question (see Table 2). Burnett (2004)58 randomised 102 children for follow-up (after an initial paediatric gastroenterology clinic appointment) by either a nurse-led clinic or a physician-led clinic. Results did not reach statistical significance; however, 65.4% were ‘cured’ in the nurse-led clinic, compared to 50.0% in the physician-led clinic, with a median time to cure of 18.0 months (95% CI 8.5 to 27.5) in the nurse-led clinic and 23.2 months (95% CI 17.3 to 29.2) in the physician-led clinic. Faramarzian (2018)59 allocated 120 children to either additional nurse-led input or usual care. There were no statistically significant differences between groups at 1, 2 or 3 months after the start of the intervention. Ismail (2011)60 explored the impact of a nurse-led clinic for children who had not made satisfactory progress while attending a general paediatric clinic. Of the first 50 children attending the nurse-led clinic, significant self/parent reported improvement was made after 3–4 months. Tappin (2013)61 found no significant differences between 75 children who attended either nurse-led or physician-led clinics for any of the reported outcomes. In summary, there is some very-low-quality evidence that nurse-led clinics are feasible and could result in equivalent (or possibly better) outcomes than traditional physician-led clinics.

What is the effect of a constipation care pathway/algorithm used in primary care/community settings?

Three studies62–65 (see Table 2) explored the effect of a constipation care pathway/algorithm used in primary care/community settings. Mallon (2015)63 found that, compared to before the circulation of a management algorithm, after implementation there were significantly fewer children with faecal impaction and a smaller proportion of children referred on to secondary care (gastrointestinal specialists). Bellesheim (2018)62 explored the effect of a practice pathway, comprising a framework of flowcharts covering assessments, interventions, medication, referrals and follow-up, for children with constipation and autistic spectrum disorder (ASD). The refined pathway led to 85% of 82 children with ASD and constipation achieving at least one personal ‘goal’. Norbedo (2017)64 explored the effects of a constipation care pathway/algorithm used for children presenting in emergency departments (EDs), aiming to determine incidence of functional constipation in children presenting at the ED. Information on treatment and treatment outcomes are limited, although 734/1020 were treated with an enema. In summary, there is very limited evidence that an algorithm, or care pathway, used in primary care settings to guide the management and referral of children with constipation (including children with ASD) may be beneficial. There is insufficient evidence on which to reach conclusions relating to care pathways within ED.

What are the effects of specialist (Level 2) services and models of service delivery?

Six included studies65–70 addressed this question (see Table 2). A RCT69 investigated the effect of a model of care which involved 86 children with, and families, receiving personalised dietary advice from a registered dietitian. Effect on constipation-related outcomes is not reported. Gonring (2019)68 described the delivery and outcomes of an interdisciplinary, ‘carer-assisted medical-behavioural, group-based intervention’. The results demonstrate improvements in stooling patterns (frequency and incontinence) and suggest this programme may be beneficial to children who have failed ‘traditional medical management’. Gabr (2020)67 evaluated the effect of a Bowel Management Program comprising laxatives, diet modification and toilet training delivered as part of a comprehensive management pathway. A statistically significant beneficial change in a QoL score and faecal incontinence score was found after the programme. Costigan (2019)66 described the profile of children attending a bowel management clinic, with a focus on the use of transanal irrigation (TAI) devices. Only 15 of the 192 included children had CFC. However, 10 of the children with CFC who attended the service were able to discontinue washout use and achieve faecal continence. Athanasakos (2020)65 explored the effect of having input from a specialist physiology service, providing scientific investigations, to inform decisions of the multidisciplinary team. The authors reported a significant improvement in subjective condition severity (physical and emotional) scores at 3 months post intervention. Poenaru (1997)70 explored the effect of a multidisciplinary team running a paediatric bowel management clinic. Defaecation frequency, stool consistency and abdominal pain significantly improved but there was no effect on faecal incontinence.

In summary, consistent findings from studies with some limitations provide very-low-quality evidence that specialist services may have a beneficial impact on outcomes of children with chronic constipation, but further research is required.

What is the effect of different follow-up regimes following appointments with specialists?

We identified one RCT71 which randomised 235 children to receive either a phone follow-up, access to a website or standard care. Results at 12 months showed there were 68.4% successfully treated children in the phone follow-up group, 78.5% in the website access group and 72.7% in the control group. The percentage with faecal incontinence in the three groups were 21.5%, 15.6% and 13.0%, respectively. The authors conclude that provision of access to web-based information may be more beneficial for the recovery from constipation.

In summary, low-quality evidence from one RCT suggests that access to web-based information, following an appointment with a specialist, may be more beneficial that a follow-up phone call or no follow-up.

What are the effects of highly specialist (Level 3) services and models of care?

We found one study72 which explored the effect of a recovery protocol for children who underwent colorectal surgery. The majority of children had inflammatory bowel disease, with only seven with a diagnosis of constipation. Compared to pre protocol, the length of stay significantly decreased, with no significant differences for complication and readmission rates. In summary, there is very-low-quality evidence that a recovery protocol may benefit outcomes following colorectal surgery. This evidence does not relate specifically to the population of children with CFC.

What is the effectiveness of probiotics?

Harris (2019)74 pooled evidence from 14 RCTs80,106–118 which compared probiotics with any control (of which 9 had an ‘active’ and 5 an ‘inactive’ control intervention); we added 2 RCTs to the ‘active’ control comparison75,76 (see Table 3), so there were data from a total of 16 RCTs (1021 participants). Chao (2017)77 did not provide any relevant data. For the outcome of defaecation frequency (in bowel movements/week), the addition of the new trials did not change the direction or significantly change the size of the effect: the weighted mean difference (WMD) between probiotics and active control was 0.14 (−0.14, 0.43) (see Figure 12). Results for the meta-analysis for a measure of ‘treatment success’ remained the same, as neither of the new trials included data relating to this outcome [relative risk (RR) between probiotics and active control: 1.16 (0.93, 1.44)]. The conclusion of Harris (2019)74 that the adverse events observed during treatment with probiotics were few and were balanced between groups, suggesting safety, is not affected by the new trials. In summary, evidence suggests that probiotics may not be more beneficial than control at improving outcomes in children with constipation, but there is no suggestion that probiotics are not safe. We have moderate certainty in this finding.

FIGURE 12.

Forest plot: probiotics vs. control. Defaecation frequency.

What is the effectiveness of additional dietary fibre?

Piccoli de Mello (2018)78 pooled evidence from 9 RCTs119–127 (680 participants) which compared dietary fibre with either a placebo (4 studies) or laxatives (5 studies); we added one pilot placebo-controlled RCT to this comparison79 (see Table 3), so there were data from a total of 10 RCTs (702 participants). The WMDs between fibre and control were 0.18 (−0.10, 0.46) for defaecation frequency (bowel movements per week) (see Figure 13) and 0.04 (−0.33, 0.41) for stool consistency (see Figure 14). The addition of the new trial did not change the direction or significantly change the size of the effect for either outcome. Four further new trials80–83 were relevant to this review but none reported sufficient data for inclusion in the updated meta-analysis. Thus, updating the review of fibre with new evidence from one new RCT did not change conclusions which can be drawn about fibre. In summary, there is very-low-quality evidence that additional dietary fibre is not more beneficial than control or laxatives in improving clinical outcomes.

FIGURE 13.

Forest plot: dietary fibre vs. control. Defaecation frequency.

FIGURE 14.

Forest plot: dietary fibre vs. control. Stool consistency.

What are the effects of different milk formula in infants?

Six studies84–89 address this question (see Table 3). Two small RCTs, both with methodological concerns, each compared two different formula. 84,85 Four non-randomised studies also explored different milk formulas. 86–89 The heterogeneity between these studies – particularly in terms of the type of milk formula given, but also in terms of whether or not infants had diagnosed constipation, or the length of the reported symptoms – and methodological concerns mean that it is not possible to reach generalised conclusions. In summary, there is insufficient evidence to support any generalised conclusions about the relative effect of different milk formula.

What is the effect of a cow’s milk-free diet?

Four studies90–93 addressed this question (see Table 3). Three were RCTs, all of which recruited children who were referred to gastrointestinal clinics with chronic constipation, having had no benefits from laxatives (total n = 275). In two of these, a cow’s milk free-diet was compared with a diet including cow’s milk, with children in both groups receiving the same dose of PEG. 90,93 The other RCT compared children taking cow’s milk and soy milk, with neither groups receiving laxatives. 91 We also identified one small cohort study. 92 In all studies, there were improvements in symptoms of constipation in the children receiving a cow’s milk-free diet. Two crossover trials (reported in one paper),128 including 52 children, were excluded from our synthesis because the majority of included children had cow’s milk allergy. In summary, there is low certainty that cow’s milk-free diet may be beneficial to outcomes, in children for whom laxatives have been unsuccessful. It is important to note that evidence from studies which included participants who had a diagnosis of cow’s milk allergy were excluded from our review. It will be important to consider this wider body of evidence relating to cow’s milk allergy to inform decisions relating to exclusion of cow’s milk from diet.

What is the effectiveness of sugars (brown sugar, figs syrup, black sugar molasses)?

Three studies94–96 addressed this question (see Table 3). Tajik (2018) compared a daily dose of brown sugar with a daily dose of Fijian figs, in children aged 2–10 years with CFC. 96 Both interventions had a beneficial effect on defaecation frequency and pain, faecal incontinence and abdominal pain. Brown sugar had a greater improvement than Fijian figs on faecal incontinence and abdominal pain. Dehghani (2019) compared black sugar molasses with PEG syrup. 95 Similar outcomes were observed for both groups. Beleli (2015)94 reported that the prebiotic 4ʹ-galactooligosaccharide might be more beneficial than placebo. The heterogeneity between these studies – particularly in terms of the type of sugars and the adjunct treatments given – and methodological concerns mean that it was not possible to reach generalised conclusions. In summary, there is insufficient evidence to support any generalised conclusions about the effectiveness of sugars (including brown sugar, Fijian figs and black sugar molasses).

What are the effects of selenium supplements?

One RCT97 randomised children (from a boarding school) who had constipation to receive a daily selenium supplement or a placebo for 2 weeks. Twenty-three per cent of 530 school students were found to have constipation and were recruited to the study. There were statistically significant differences between groups for frequency of defaecation at days 14 and 21 (but not day 7), and for stool consistency and abdominal pain severity at days 7, 14 and 21. The proportion of children from the boarding school with constipation varies from reported prevalence in other populations, limiting certainty in the findings. In summary, there is very low certainty that selenium supplements may improve outcomes of defaecation frequency, stool consistency and abdominal pain.

What is the effectiveness of other/alternative dietary intake?

Three studies98–100 addressed this question (see Table 3). Hassanein (2021)98 randomised 100 children with cerebral palsy to receive either oral magnesium sulphate or a placebo for 1 month. The oral magnesium sulphate improved outcomes of defaecation frequency and stool consistency compared to placebo. Stepurina (2018) randomised 95 children to receive a ‘magnesium-containing water’ in addition to standard care, or standard care alone, over an 18-day treatment period. 100 Modaresi Saryazdi (2013) randomised 58 children to receive paraffin oil in addition to a combined treatment. 99 Limited information prevents generalised conclusions being drawn from either Stepurina (2018)100 or Modaresi Saryazdi (2013),102 but authors of both studies conclude that the interventions may be safe and beneficial. In summary, there is insufficient evidence relating to support generalised conclusions relating to other/alternative dietary intake.

What is the effect of fluid intake on constipation?

One RCT101 explored the effect of increasing water intake and found no effect. The authors conclude that ‘Advising parents of constipated children to increase liquid intake is not helpful and should not be recommended unless history suggests that the child’s liquid intake is inadequate for a normal child of that age and activity level’. In summary, there is insufficient evidence to support any routine change in fluid intake for children with constipation, unless ‘history suggests that the child’s liquid intake is inadequate for a normal child of that age and activity level’.

What are the effects of educational interventions (delivered in addition to routine care)?

We identified three studies102–104 (see Table 3). Two RCTs102,103 investigated the effect on an internet-based intervention (www.ucanpooptoo.com/), which gave parents and children access to web-based educational resources and tools designed to support ‘enhanced toilet training’ in children with encopresis. One RCT104 investigated the effect of an information leaflet for parents of children with CFC. These three RCTs concluded that there was evidence of a beneficial effect on the number of bowel movements.

In summary, there is some limited evidence that educational interventions – particularly web-based interventions – used by parents/carers in addition to standard care may have a beneficial effect on clinical outcomes. Due to the small evidence base, limitations in the reporting of these studies and some methodological limitations, we have very low certainty in this finding.

What is the effect of combined dietary and behavioural interventions?

One study105 addressed this question (see Table 3), providing 25 children with ‘dietary and behavioural rules’ and 27 children with PEG. 105 Parents were involved in deciding which treatment children should receive, and in the intervention delivery and outcome assessment. In summary, there is insufficient evidence to support any generalised conclusions relating to combined dietary and behavioural interventions.

What are the effects of laxatives?

Gordon (2016)133 pooled evidence from 25 RCTs121,124–126,148–168; we identified 10 more recent relevant RCTs110,129–132,135–137,140,139 (see Table 4). Four135,136,139,140 investigated comparisons which were explored by Gordon (2016):133

• PEG versus placebo. Gordon (2016)133 included two studies (101 children). We added one more RCT,136 adding 102 participants. For defaecation frequency, Gordon (2016)133 reported a WMD of 2.61 (CI 1.15 to 4.08) (low certainty). Adding data from Modin (2018)136 changed this to 1.94 (CI 0.44 to 3.44) (see Figure 16). This addition did not change the direction or significantly change the size of the effect.

• PEG versus lactulose. Gordon (2016)133 included six studies (465 children). We identified two more RCTs,135,139 with a further 148 participants, and were able to add data from 1135 (83 children) to the meta-analysis for defaecation frequency, adding 83 participants. Gordon (2016)133 reported a WMD of 0.7 (CI 0.1 to 1.31) (low certainty). Adding data from Jarzebicka (2019)135 changed this to 1.08 (CI 0.12 to 2.05) (see Figure 17). This addition did not change the direction or significantly change the size of the effect.

• PEG versus liquid paraffin. Gordon (2016)133 included three studies (299 studies), and pooled data from two, finding no difference in defaecation frequency between groups (low certainty). We identified one further RCT140 but there were no suitable data to combine.

FIGURE 16.

Forest plot: PEG vs. control. Defaecation frequency.

FIGURE 17.

Forest plot. PEG vs. lactulose. Defaecation frequency.

Six of the 10 newly identified studies110,129–132,137 investigated comparisons which were not explored within the trials included in the SR. These were:

• Polyethylene glycol 3350 (PEG3350) plus electrolytes versus polyethylene glycol 4000 (PEG4000). Bekkali (2018)129 randomised 97 children to receive PEG3350 plus electrolytes or PEG4000 and assessed outcomes after a year. The results demonstrated no difference between PEG3350 plus electrolytes and PEG4000 for a parent/child reported symptom score or adverse events.

• Lactulose versus placebo. Cao (2018)131 randomised 100 children to receive lactulose or placebo for 6 weeks. Lactulose was more beneficial than placebo for defaecation frequency and stool consistency, and there were no differences in adverse events.

• Oral bisacodyl versus rectal bisacodyl. Pranoto (2016)137 randomised 99 children to receive an oral or rectal stimulant laxative (5 mg bisacodyl). No differences were found between groups.

• PEG versus cassia fistula. Esmaelidooki (2016)132 randomised 109 children to receive either PEG4000 or cassia fistula. This trial is included within the synthesis reported in Section Evidence of effectiveness: complementary/alternative medicines.

• PEG versus probiotics. Hashemi (2015)110 randomised 120 children to receive PEG, probiotics or PEG plus probiotics. No data are reported but the authors conclude that PEG plus probiotics is superior to probiotics alone.

• Lubiprostone versus placebo, and comparison of doses. Benninga (2022)132 conducted two studies, involving the same children. One study randomised 606 children to lubiprostone or placebo. No differences were found between groups for defaecation frequency (p = 0.22). The second study compared a 36-week treatment of two different doses of lubiprostone.

Gordon (2016)133 identified a number of comparisons which we were not found in other studies. This includes a comparison of PEG with milk of magnesia, including four studies (261 children) and concluding that PEG was superior to milk of magnesia for defaecation frequency (low certainty). Gordon (2016)133 also identified single trials which reported no difference on key outcomes between PEG and enema, PEG and flixweed, PEG and dietary fibre mix, lactulose and lactitol, lactulose and partially hydrolysed guar gum, dietary fibre mix and lactulose, and senna and lactulose; two studies which found liquid paraffin to be more effective than lactulose, and one which found a difference between high- and low-dose PEG. Quality of this evidence was considered low to very low.

We identified a SR138 focused specifically on children aged less than 24 months. Data from two RCTs and three retrospective studies were included,150,129,169–171 and the authors concluded that there is insufficient evidence to determine the optimal dosage of PEG.

In summary, new trials do not change key conclusions from previous reviews, and quality of evidence remains low.

Note: Studies primarily include children without additional needs. There may be high risks of aspiration when using PEG or liquid paraffin in children with additional needs. Guidelines for children with neurological impairment specifically state that laxative prescription for this population should be ‘as in typically developing children, unless there is a risk of aspiration of polyethylene glycol or liquid paraffin’. 172 A review of side effects in children using senna concluded that senna was safe, although a rare side effect of skin blistering may occur where there is skin exposure. 173

Head-to-head comparisons of laxatives

Table 5 illustrates where there are trials with head-to-head comparisons of different laxatives and provides an indication of the relative effectiveness (across multiple outcomes). There have been direct comparisons of PEG with lactulose, milk of magnesia, liquid paraffin and placebo; lactulose with milk of magnesia, senna, liquid paraffin and placebo; and lubiprostone with placebo. PEG appears superior to lactulose and milk of magnesia, milk of magnesia superior to lactulose, and lactulose equivalent to senna. Quality of evidence is low to very low for all findings. Empty cells indicate that no RCT evidence has been identified.

What are the effects of laxatives plus domperidone?

We found one RCT which investigated the effect of domperidone in 52 children with cerebral palsy134 (see Table 4). After 2 weeks, slightly more number of children receiving both PEG and domperidone had a satisfactory response than children receiving PEG only. Importantly, in 2014 the Medicines and Healthcare Products Regulatory Agency (MHRA) in the UK advised that, due to a small increased risk of serious side effects, domperidone should only be used for specific conditions, and should not be used in children under 16 years of age.

What is the effect of physical exercise (focused on pelvic floor muscles)?

A cohort study investigated the effect of an exercise (walking in squatting) designed to exercise the pelvic muscles, twice a day for 8 weeks in 44 children with CFC, for whom other treatments – including laxatives and toilet training – had failed141 (see Table 4). Statistically significant changes were found for measures of overall improvement in symptoms, defaecation frequency and stool consistency. In summary, there is low certainty that physical exercise (focused on pelvic floor muscles) may improve outcomes.

What is the effect of a combined pharmacological, diet and behavioural programme?

Four studies explored combined treatment approaches, including education – including dietary and behavioural – and pharmacological interventions, including children with and without additional needs144,146,147 (see Table 4). These studies provide some very limited data which suggest that a combined pharmacological, dietary and behavioural programme may have some benefits for children with CFC. One small RCT145 randomised 20 children with ASD and encopresis to 10 appointments at a nurse-led behavioural programme compared to a waiting-list control. The study reports a large increase in the number of faecally continent children after treatment for children who received the intervention (treatment group 6/10 continent post-intervention and 5/10 at follow-up; control group 0/10 and 1/10 respectively). Findings from Axelrod (2016)142 are in agreement with this RCT. In summary, there are some very limited data which suggest that a combined pharmacological, dietary and behavioural programme may have some benefits. We have very low certainty in this finding due to the quantity and quality of available studies.

What is the effect of rectal enemas in children with severe constipation?

Bongers (2009)191 randomised 102 children to receive an addition of 3 rectal enemas a week for 3 months or PEG only. Defaecation frequency improved in both groups at 26- and 52-week follow-up; however, the enema group had significantly higher frequency of defaecation. Abdominal pain, painful defaecation and faecal incontinence improved in both groups, but there were no significant differences between groups. The majority (76%) of the children in the enema group reported to never/seldom feeling worse after the application of the enema, while 11% reported sometimes and 13% reported always. A rectal enema was perceived as very to extremely terrible in 15% of children. Bekkali (2009)148 randomised 90 children to 6 consecutive days of rectal enemas or oral PEG3350 with electrolytes. There were no statistically significant differences in outcomes at follow-up for measures of successful disimpaction, defaecation frequency, stool consistency, faecal incontinence and abdominal pain. Successful disimpaction was achieved in 37 (80%) children receiving enemas and 30 (68%) receiving oral laxatives (p = 0.28). Garcia (2016)204 randomised 58 hospitalised children, to a combined oral and rectal enema 1-day regimen (which included Fleet enema, bisacodyl suppository, and castor oil) or a 3-day Fleet enema regimen. Disimpaction was successful in both groups, but the combined group reported significantly fewer complaints of abdominal pain after treatment compared to the Fleet only group (p = 0.019). Yoo (2017)176 evaluated the efficacy and safety of combined oral (PEG3350 electrolyte solution) and enema solution in children with CFC who were hospitalised after failed attempts at disimpaction as an outpatient (n = 28). Defaecation frequency was found to significantly increase after disimpaction. Abdominal pain was also reduced from 15 children before therapy to 4 children post therapy.

In summary, the addition of regular rectal enemas may increase defaecation frequency in children with severe constipation, but not have any effect on overall treatment success or other outcomes, and may cause discomfort or distress to some. Enemas and high-dose laxatives may be equally effective at reducing rectal faecal impaction. Combined oral and enema therapy may be beneficial for hospitalised children with faecal impaction. We have very low certainty in these findings. There is insufficient evidence to reach any conclusions about the relative effect of different types of enemas or the effectiveness of specific regimens.

Note: A narrative review205 reports that data provided by the Medicines Control Agency (MCA) showed 24 potential serious reactions to phosphate enemas. Further, in 2014 the US Food and Drug Administration (FDA)206 issued a warning that sodium phosphate, delivered orally or rectally in an enema, may cause rare complications and even death. Given that there is currently insufficient research evidence about the relative effect of different types of enemas, these reports support the avoidance of phosphate enemas.

What is the difference in effectiveness of microenemas and oral laxatives for functional constipation in infants?

Strisciuglio (2021)193 randomised 158 infants, aged 6–48 months, with functional constipation, to receive 2 weeks of oral PEG4000 or of Promelaxin (a microenema containing honeys and polysaccharides). For stool frequency, the group receiving Promelaxin had a better outcome [16.5% (CI 1.55% to 31.49%)]. There were no statistically significant differences between groups for other outcomes and no adverse events identified in either group. In summary, there is evidence that Promelaxin microenemas and oral laxatives are equally effective in the treatment of functional constipation in infants (6–48 months). Due to the quantity and quality of the evidence, we have low certainty in this finding.

What is the difference in effectiveness of an enema and a soft suppository for disimpaction?

Ormarsson (2016)192 randomised 54 children receive either low-dose LP101 free fatty acids suppository (n = 23) or Klyx enema (n = 33). In a second phase, a further 21 children received a higher-dose suppository (and were compared to the earlier randomised groups). The higher-dose suppository was equally effective as the Klyx enema for bowel emptying and symptom relief. In summary, there is evidence that enemas and high-dose suppositories may be equally effective at promoting bowel emptying. Due to quantity and quality of evidence, we have low certainty in this finding.

What is the effect of transcutaneous electrical stimulation?

A Cochrane review (Ng, 2016)174 evaluated the effectiveness and safety of transcutaneous electric stimulation. This review included one study (Chase, 2015)207 which randomised 46 children to receive either TES or sham TES. One more recent relevant RCT175 randomised 90 children to receive either interferential electrical stimulation and pelvic floor muscle exercises or sham stimulation and pelvic floor muscle exercises. The number with diagnosed CFC reduced more in TES group (38/45) than the sham group (19/44) after treatment (p < 0.0003). We updated analyses for QoL and improvement in faecal soiling (see Report Supplementary Material 7 for exploration of outcomes available for pooling). For QoL, using the optimal method for estimating SDs from medians,208 the pooled effect size (SMD) is 0.49 (95% CI 0.13 to 0.85) (see Figure 19), demonstrating a statistically significant result in favour of TES. This changes the results presented by Ng (2016) which found no significant difference between TES and sham. For improvement in faecal soiling, the risk ratio is 2.78 (95% CI 1.63 to 4.74) (see Figure 20); this is statistically significant and increases the effect size reported by Ng (2016). In summary, there is very low certainty that TES may reduce the number of soiling episodes and improve self-reported QoL, as compared to sham TES. [Note: evidence relating to SNS is included within Level 3 synthesis – see Evidence of effectiveness: interventions delivered by consultant-led teams (Level 3/highly specialist tertiary care services)].

FIGURE 19.

Forest plot. Transcutaneous electric stimulation vs. placebo. QoL.

FIGURE 20.

Forest plot. Transcutaneous electric stimulation vs. placebo. Faecal incontinence.

What is the effect of transanal irrigation?

Five studies,177–181 investigating the effect of TAI in a small number of children with CFC (see Table 6), provide very limited evidence about the effectiveness of TAI. These studies provide some very-low-certainty evidence that TAI may be safe, feasible and effective for children with intractable symptoms which have not resolved with long-term conventional laxatives and management.

What is the effect of biofeedback?

Seven studies182,183,194,196–198 addressed this question (see Table 6).

What is the effect of physiotherapy, in combination with conventional treatment?

Three RCTs explore the effectiveness of different physiotherapy regimes199–201 (see Table 6). Silva (2013)199 reported a significant difference in favour of the physiotherapy group, for defaecation frequency at 6-week follow-up (p = 0.01). There was no significant difference between groups for any other outcomes measured. Van Engelenburg (2017)200 concluded that outcomes improved more for the physiotherapy group, for numbers with CFC, use of laxatives, measures of treatment success and QoL. Van Summeren (2020)201 found no statistically significant differences in success rates or QoL between treatment groups, but a difference in favour of physiotherapy for global perceived treatment effect (adjusted RR 1.40; 95% CI 1.00 to 1.73). A subgroup analysis comparing outcomes of children with and without chronic laxative use indicated that children with chronic laxative use may have had some benefit from physiotherapy. The authors compare the results of their study with the findings of Silva (2013) and Van Engelenburg (2017)200 and conclude that there is no evidence of benefit of providing physiotherapy to the ‘whole group of children with CFC consulting in primary care’, but that it was possible that physiotherapy may be beneficial in a subgroup of children with longer duration symptoms. In summary, evidence relating to the effectiveness of physiotherapy is inconsistent. Routine referral to physiotherapy for all children with constipation seen within primary care is not supported. There is some limited evidence that physiotherapy may be beneficial for a subgroup of children, but further research is required to confirm (or refute) this. We have very low certainty in this finding, due to the quantity and quality of studies.

What is the effectiveness of physical therapy for children with cerebral palsy?

Two studies investigate the effect of physical therapy on constipation in children with cerebral palsy184,202 (see Table 6). Awan (2016)184 reported that defaecation frequency and severity of constipation improved significantly after the programme (p < 0.05). Awan (2021)202 reported differences at baseline between groups for defaecation frequency and constipation severity score, and no significant differences between groups after the first intervention period, concluding that there was a significant improvement over the whole intervention period (p < 0.001). Eisenberg (2009)185 compared the use at home of a standing frame and a walker in children with severe cerebral palsy. There was a reduction in that proportion of children with constipation among the walker group, but not in the standing group (p = 0.02). In summary, there is some very-low-quality evidence to suggest that constipation in children with cerebral palsy may be improved with physical therapy. However, evidence is insufficient to support generalised conclusions.

What is the effect of dietary exclusion of fructose and lactose?

One study,188 explored the effect of expert health professional advice and a diet which excluded an identified positive sugar, for 6–12 months (see Table 6). After exclusion of sugars, consistency of stools and severity of constipation were significantly improved. Seventy-two per cent of parents reported difficulties implementing the diet. In summary, there is insufficient evidence to support any conclusions. One small study provides very-low-certainty evidence that exclusion of fructose and lactose could reduce severity of constipation. However, implementation of this diet was challenging and further research is required to support any conclusions.

What is effect of a combined therapeutic programme?

Four studies187–189,203 explored a range of combined programmes (see Table 6). Borowitz (2002)203 found no significant differences between groups, but children receiving enhanced toilet training achieved these outcomes whilst using significantly less laxatives than the group receiving intensive medical therapy only. Loening-Baucke (1989)187 reports that 43% of children had recovered at 12 months, and 57% not to have recovered, based on clinical symptoms. Data analysis compares those who did and did not recover. Loening-Baucke (1993)188 interviewed parents of children who had not recovered but does not provide data supporting conclusions about effectiveness of treatment. Modin (2016)189 provides data which demonstrated the impact of constipation on behavioural difficulties, but the study design is not appropriate to support conclusions relating to treatment effectiveness. In summary, there is insufficient evidence to support specific conclusions relating to the effect of a combined treatment programme, but some very-low-certainty evidence that combined programmes may be beneficial for some children.

What is the effect of botulinum toxin?

Five studies209–213 (see Table 7) provide very-low-quality evidence that botulinum toxin injection for chronic constipation may be an effective method in managing children with CFC. Due to methodological limitations and limitations in the reporting of these studies, we have very low certainty in this finding.

What is the effect of sacral neuromodulation?

Nine studies214–221 provide some very-low-quality evidence that sacral neuromodulation may be effective in treating the symptoms of CFC. Aspects commonly reported across studies include less abdominal pain, improvement in symptoms, complications.

What is the effect of anorectal myectomy?

Four studies222–225 provide some very-low-quality evidence that suggests anorectal myectomy may be effective at treating CFC, in children who have not responded to medical treatment. However, due to methodological limitations and low number of studies, the evidence is insufficient to support generalisable conclusions. It should be noted that these studies were conducted prior to the availability of botulinum toxin, which may provide a non-surgical alternative.

What is the effect of antegrade continence enema/Malone antegrade continence enema?

Seventeen studies226–242 provide very low certainty that the use of ACE/MACE may be effective for CFC. Although studies were consistent in their findings that ACE/MACE is effective, many of the outcomes of interest were not addressed by the studies. Complications (e.g. granulation, leakage, additional surgery required) arising from ACE/MACE use were common.

What is the effect of MACE compared to caecostomy button?

One study243 provides some very-low-quality evidence that caecostomy button may have less complications than MACE, but this is insufficient to support generalisable conclusions.

What is the effect of ACE compared to sacral nerve stimulation?

Two studies244,245 provide some very limited evidence that SNS may be superior to ACE for faecal incontinence, ACE superior to SNS at improving defaecation frequency in those with reduced bowel movements, and that SNS may have less complications than ACE. This evidence is insufficient to support generalisable conclusions.

What is the effect of colonic resection?

Two studies246,247 provide some very-low-quality evidence which suggests outcomes were mixed for the use of colonic resection to treat the symptoms of CFC. There is insufficient evidence to suggest the use of colonic resection is safe and effective for the treatment of CFC. The population of children to whom this evidence relates is limited to those with CFC for whom other treatments, including ACE, have failed.

What is the effect of colonic resection combined with Malone appendicostomy?

One study248 provides some very-low-quality evidence which suggests colonic resection combined with Malone appendicostomy may be safe and effective to treat the symptoms of CFC. There is only one small study, providing insufficient evidence to support generalisable conclusions.

What is the effect of surgical intervention [ileostomy, colostomy or (sub)total colectomy]?

One study249 assessed outcomes of surgical intervention [ileostomy, colostomy or (sub)total colectomy] in children with CFC. There provides insufficient evidence to support generalisable conclusions about these surgical interventions.

What is the effect of abdominal massage (in children with or without disabilities)?

We found two small studies of abdominal massage. 261,263 Bromley (2014)261 report on 25 children with an identified disability provided with daily abdominal massage by parents, following training. Constipation was reported to improve by 87.5% of parents, diet by 41%, sleep by 37% and school attendance by 19%. Mostamand (2018)263 reported the results of manometric tracings in children who had received 5 minutes of abdominal massage, and report improvements. In summary, abdominal massage may be beneficial for children with CFC, but there is insufficient evidence to support generalised conclusions.

What is the effect of connective tissue manipulation and kinesio taping in children with cerebral palsy?

One study255 randomised 45 children with cerebral palsy and CFC to receive either connective tissue manipulation (CTM) to the sacral and cervical regions, kinesio taping (KT) to the lumbar sacral area, or control. Outcomes improved in the two treatment groups compared to control. In summary, there is low certainty from one RCT that physiotherapy techniques of CTM and KT may be beneficial components of a programme for children with cerebral palsy who have constipation and are receiving physiotherapy.

What is the effect of chiropractic or osteopathic manipulation?

Chase (2011)251 included three case studies (five children) in a SR of chiropractic manipulation. One of the studies combined spinal manipulation with changes to diet, and another with abdominal massage. In summary, there is insufficient evidence to support conclusions relating to the effectiveness of manipulation interventions.

What is the effect of dry cupping therapy compared to laxatives?

Shahamat (2016)256 compared dry cupping therapy of the abdominal wall with laxative therapy in 120 children with CFC. Findings suggest that cupping therapy was as effective as laxatives. In summary, there is very low certainty that dry cupping therapy of the abdominal wall may be as effective as laxatives.

What is the effect of reflexology?

Chase (2011)251 included one study which demonstrated some benefits of reflexology in 48 children with CFC. We identified three later studies. 253,254,262 Canbulat Sahiner (2017)253 randomised 37 children with CFC to receive reflexology or control group. There were no significant differences between groups. Duymaz (2020)254 and Elbasan (2018)262 explored the effect of the addition of reflexology to neurodevelopmental therapy session in children with cerebral palsy. There were statistically significant differences between groups, in favour of reflexology, for constipation-related outcomes. In summary, studies provide mixed results. There is insufficient evidence to support generalised conclusions relating to the effect of reflexology.

What is the effect of acupuncture?

Chase (2011)251 included one study (n = 10) which investigated the effect acupuncture to points in children with CFC and demonstrated some benefits. This is insufficient evidence to support conclusions about the effect of acupuncture.

What is the effectiveness of cassia fistula?

Aslam (2021)252 combined results of two studies (n = 132) of cassia fistula. Meta-analyses demonstrated a beneficial effect on outcomes of defaecation frequency (MD 4.22, 95% CIs, CI 2.78 to 5.66),¹ severity of pain during defaecation (MD −4.84, 95% CI −8.28 to −1.41) and stool consistency (MD −9.51, 95% CI −13.52 to −5.51), but not for outcomes of faecal incontinence and retentive posturing. In summary, there is some very limited evidence that suggests cassia fistula may have some beneficial effects, but this is insufficient to support any generalised conclusions. We have very low certainty in this finding.

What is the effect of herbal and/or traditional medicines?

Six studies investigated the role of other herbal/traditional medicines. 257–260,264,265 Cai (2018)257 explored the effect and safety of Xiao’er Biantong (XEBT) granules in 487 children with CFC. The study found that XEBT can increase frequency of defaecation and shorten defaecation interval time as well as positive impact on a range of other factors. Qiao (2021)260 randomised 100 children with CFC to receive a Chinese herbal medicine or placebo. There were significant differences between treatment groups for all outcomes, in favour of the herbal medicine, but a small number of side effects in the herbal medicine group. Tavassoli (2021)259 randomised 140 children to receive the medical herb viola flower or PEG4000. After 4 weeks of intervention, there were no differences between the groups for clinical outcomes including of stool consistency, defaecation frequency, faecal retention or soiling. There was a significant difference between group in the number of minor side effects (p = 0.03). Nimrouzi (2015)258 compared the medical herb Flixweed (Descurainia sophia L.) with PEG4000 in 120 children. This RCT is also included in the SR of laxatives which is included in the Level 1 synthesis [see What is the effectiveness of assessment and intervention by primary care services (Wider Children’s workforce/Level 1 interventions)?]. No significant differences were found between the groups after 8 weeks of treatment. Zadpe (2020)265 studied the effect of Shunthyadi syrup in 30 children with CFC, concluding that Shunthyadi was effective at improving outcomes. Babaei (2018)264 studied the effect of behaviour modification and herbal drugs (based on traditional Persian medicine) on the symptoms of constipation in six children.

In summary, collectively, there is very low quality of evidence that herbal/traditional medicine for children with CFC may result in equivalent or improved outcomes. However, studies investigated different interventions, making it difficult to support clinical decisions. Studies provides evidence that XEBT granules and XiaojiDaozhi Decoction may be promising interventions, and that Flixweed may be a suitable alternative to PEG, but these studies had limitations, limiting conclusions.

What is the effect of behavioural therapy techniques delivered by specialist practitioners?

Freeman (2014)267 included 10 studies in a SR of behavioural interventions for children with faecal incontinence with constipation (see Report Supplementary Material 11 for a summary of the interventions included). Data from four RCTs102,103,203,271 were pooled within meta-analysis, and suggest behavioural interventions are more beneficial than control interventions at achieving ‘author-defined success’. We identified one further RCT266 relevant to this review and were able to combine data on treatment success with the data pooled within Freeman (2014)267 review for this outcome. This did not change the direction of evidence but changed the risk ratio from 1.78 (95% CI 1.25 to 2.55) to 1.68 (95% CI 1.22 to 2.32) (see Figure 24). There were no data suitable for inclusion in other meta-analyses. The pooled data in Freeman (2014)267 indicate that behavioural interventions are more beneficial than control interventions at reducing soiling frequency (MD −2.81, 95% CIs −5.04 to −0.58; low certainty) and defaecation frequency (MD −0.57, 95% CI −2.90 to 1.75; low certainty). In summary, behavioural therapy techniques delivered by specialist practitioners may be beneficial. Our certainty in this finding is very low due to limitations in the reporting of completed studies.

FIGURE 24.

Forest plot. Behavioural interventions vs. control. Treatment success.

What is the effect of externalising treatment, compared to other behavioural interventions?

Silver (1998)269 conducted a retrospective audit of 108 children treated for soiling problems, including constipation, using externalising treatment or other more traditional, behavioural approaches. Some benefits of externalising therapy were found (42/47 stopped soiling in externalising group compared to 30/43 other; 24/29 found externalising therapy helpful compared to 10/30 other). There is insufficient evidence to support conclusions about the benefits of externalising treatment, compared to other behavioural interventions.

What is the effect of psychotherapy, given in addition to other behavioural therapy?

Taitz (1986)270 compared children with CFC who did and did not receive psychotherapy and found no difference in results between the two groups, but factors such as psychosocial background had an influence on success. There is insufficient evidence to support conclusions about the benefits of providing psychotherapy in addition to other behavioural therapy.

Cost-of-illness studies

The majority of included studies were categorised as cost-of-illness studies which estimate resource use and calculate costs from a variety of sources (n = 20)261,273–277,282–291 or SRs. 17,270 Most of these studies reported costs from a health sector perspective; 4/20 considered both health and societal perspectives274,275,277,286; however, data about the indirect impact of childhood constipation (e.g. missed days at school, QoL, academic achievement) were sparse (Report Supplementary Material 13).

Reporting of cost/resource descriptions also varied widely across these studies; from those reporting limited information about cost savings or burden of illness in children with constipation (e.g. see references177,282,285,295,296) to those that provided more detailed descriptions of costs associated with treating constipation in children. 276,280,281,290,292,293 Of those studies that provided a richer description of costs, most relied on data sources that are now 5–15 years out of date.

Economic evaluations

Of the remaining eight studies, seven were economic evaluations using decision models1,224,278,279 and one was an economic evaluation conducted alongside a RCT. 294 The reporting of these studies was assessed using the CHEC list. The RQ, economic importance and type of economic evaluation were clearly stated in all of the studies. However, reporting outcomes for the data collection were less reliable, and analysis and interpretation of results for the CHEC criteria were only reported in sufficient detail in one study224 (see Report Supplementary Material 13).

No economic evaluations were identified for interventions delivered by parents/carers; interventions delivered by the wider children workforce; models of delivery or psychosocial/complementary interventions (see Figure 26).

FIGURE 26.

Studies reporting economic evidence mapped across levels of care provision.

Four studies were categorised as partial economic evaluations (i.e. cost minimisation analysis). 1,278,279 These studies all assessed the clinical and economic impact of treating faecal impaction using laxatives or combinations of laxatives. The studies drew on a variety of sources to develop the models including retrospective case notes, interviews with clinicians, clinical expert opinion and a hypothetical case study. These studies all assumed equal effectiveness between the comparators, essentially evaluating the cost differences between the treatment options. All of the studies concluded that PEG plus electrolytes (PEG3350) provided a cost-effective option.

Four studies conducted full economic evaluations; two studies reported in the NICE 2010 guidelines considered the cost effectiveness of disimpaction by dose of PEG3350, and during the maintenance phase following disimpaction1; one evaluated the cost effectiveness of sacral neuromodulation224 and one considered the cost effectiveness of physiotherapy plus conventional treatment. 294 The decision models used in the NICE guidelines reported two cost-effectiveness/utility analysis evaluations,1 which were based on data from a RCT297 and clinical expert opinion. In the disimpaction model, it was concluded that PEG3350 provided cost-savings, but in the maintenance model, the cost of drugs in the treatment alternatives had a greater impact on the total of care than hospitalisation (see Report Supplementary Material 13).

The cost-effectiveness evaluation of sacral neuromodulation employed a Markov model based on a cohort of 30 children with refractory constipation. The authors concluded that the intervention was both effective and cost-effective (see Report Supplementary Material 13). The cost-effectiveness study of physiotherapy plus standard care (education, dietary advice, toilet training and laxative) was based on a concurrent RCT study. 207 Cost-effectiveness data for this study have not been fully reported, but preliminary findings published in an abstract concluded that physiotherapy as first-line treatment for all children with constipation in primary care was not cost-effective compared to standard care.

Intervention characteristics

The evidence base (or lack of) behind ‘successfulness’ of the intervention was an important factor. Sixty-five studies described evidence uncertainties and a limited evidence base as a significant barrier to implementation. 63,64,69,74,84,92,102,103,119,132,133,136,143,145,149,153,154,161,174,178,185,203,211,214,232,243,255,258,270,282,302,303,306–337 Evidence was reported as hampered by variation in diagnostic assessment and inconsistent application of CFC definitions, different outcomes, poor adherence, small sample sizes and general lack of high-quality studies.

Conversely, implementation was described as successful in 41 studies61,65,84,91,102,133,143,145,148–151,153,154,178,185,232,243,246,255,270,282,298,303,305,314–317,319,322,330,333,337–343 when the evidence underpinning an intervention was perceived as ‘credible’ or built on a strong evidence base resulting in a confident clinical message being conveyed. 4–6,8,11,12,15,18–21,23,24 For example, Hankinson (2018) states that

In addition to providing efficacious care, the multidisciplinary model may also help correct the presumption that childhood constipation, and incontinence should be viewed as primarily a medical or behavioral problem. This message may be given to parents of children with constipation and incontinence issues when there are actually many medical issues that may be contributing … the myth that this is solely a physical health or behavioral condition is implicitly debunked. Both of these treatment modalities have empirical support based on previous literature addressing constipation and faecal incontinence (p39).

Intervention adaptability and flexibility was described critical to successful implementation in 38 studies. 61,65,84,92,102,132,143,145,148,149,151,154,178,203,212,215,233,270,282,303,306,314,317,322,324,325,328,331,339,344–350 Nineteen studies24,130,133,149,179,186,246,268,299,303,308,309,312,319,323,326,332,336,342 described interventions that could not be adapted, for example medications with a specific dosage or delivery route, and/or had significant side effects.

Interventions or services that could be tailored to meet the needs of the child and/or family circumstances were more likely to be implemented. Tailoring the intervention to the child and/or family included:

• multidisciplinary input to provide ‘bespoke’ interventions;65,282,303

• mode of delivery, for example home-based versus more traditional outpatient clinic based on children and family’s needs;65

• modifying or planning alternative approaches. 61,65,84,92,102,132,143,145,148,149,151,154,178,203,212,215,233,270,282,303,306,314,315,317,322,325,328,331,339,344–350

Intervention adaptability was often linked to the complexity of the treatments. Interventions which were perceived as simple (i.e. did not involve a lot of steps), not too time-consuming or disruptive to daily life were seen as a facilitator. 63,103,143,149,178,243,294,300,306,314,323,331,337,339,351 These interventions were commonly available, for example foods that could be bought at a supermarket to improve fibre intake in a child’s diet or being able to access the intervention over the counter or at a health food store. Attending appointments was another ‘simple’ intervention as described by Hankinson (2018) ‘the combination of medical and behavioural treatments in a single visit has the potential to address multiple detrimental symptoms associated with constipation in a concise and accessible manner’.

Stakeholders’ perception of what improved as a result of using the intervention compared with an alternative solution was identified as both a barrier and facilitator. This ‘relative advantage’ was reported as a barrier in 33 studies24,74,84,132,154,162,178,247,301,302,306,308,312,314,315,319,321,322,328,330,331,337,343,348–350,352–356 and a facilitator in 47 studies. 24,61,65,102,103,133,148,149,153,154,161,167,145,132,136,143,178,232,243,247,255,294,306–310,313,315,319,322,323,327,328,330,331,334,337,341,348–350,353–355,357,358 Studies where the intervention was perceived to have more severe side effects or adverse events as a consequence of the intervention were more frequently reported as having less of a relative advantage. Studies also weighed the relative advantage based on the target population. For example, additional caution was advised when a child who had additional needs was using the intervention, as described in Romano (2018) who states that ‘reports of severe pneumonia due to aspiration of laxatives and therefore use of macrogol or liquid paraffin must be provided with caution in children with NI [Neurological Impairment] with high risk of aspiration’ (p255). Studies describing fewer side effects or adverse events were frequently reported as having an advantage and were selected for implementation (see Appendix 6, Table 20).

Interventions that were perceived to have good quality, design and packaging of an intervention that fitted the needs of children and their families were reported as a facilitator in 13 studies. 61,65,103,153,162,302,305,313,322,333,341,347 Studies reported better engagement with the intervention when the children and families were able to ‘enjoy themselves’, or ‘go at their own pace’ or used innovative approaches (e.g. video games or internet resources) to receive information. Studies also reported the importance of supporting children and families emotionally, reducing the isolation that many families felt when dealing with constipation. For example, Sullivan (2006) reported that a nurse-led clinic was successful because of the degree of empathy shown by the nurse prescriber and the quality of the information provided.

Outer setting

The needs of children and young people living with CFC were often unmet or overlooked. This presented a major obstacle in 27 studies. 65,69,135,143,161,246,298,300,302,303,307,309,314,315,319,320,328–333,338,342,344,347,351,359 Studies reported the failure of clinicians to recognise or listen to what families were saying as negatively impacting on implementation. Other studies describe a misplaced emphasis on dealing with the physical symptoms over other relevant factors as observed by Athanasakos (2020) who states that the ‘perceived cause and effect relationship between physical and psychological pathologies was deemed to be less important than correctly diagnosing both and effectively treating both in tandems’.

Other studies highlighted the twin impact of stigma and embarrassment on seeking help for constipation, with parents often confused as to who they should talk to and when. Gaps in knowledge and practices of healthcare professionals and knowing who was responsible for management meant that constipation was frequently undetected or identified too late, compounding the initial problem. As Trajanovska (2020) noted that ‘delays in management may result in patients suffering from psychological and social complications arising from their condition. For families experiencing disadvantage, the need for timely and low-cost care in the public setting is important to ensure that all children receive quality care and that socio-economic disparities do not impact child outcomes’. A lack of specific integrated care pathways for children and young people with disabilities, particularly during the transition from paediatric to adult care, was also highlighted.

Multiple studies described the importance of working in partnership with families, children and building close collaboration with other involved agencies (e.g. school/educators) to overcome many of these barriers. Several examples included encouragement of parents, building their confidence in delivering interventions or providing opportunities for children and young people to improve their own self-management (where possible). Many of the examples cited in these studies are closely tied to the constructs of self-efficacy and adaptability discussed earlier. As stated in O’Connor (2012) ‘You have to get the family on board and ask them what is going to work for their child and come up with such a plan’ (p1).

Seventeen studies described a lack of external policy and/or incentives as a barrier to implementation. 24,63,65,148,149,185,314,316,319,330,331,333,337,350,356,359 Studies commented on the lack of guidance or consensus about what constitutes a ‘normal’ bowel motion particularly in younger children (Eicher, 2006). Other studies observed that the complexity of constipation is not fully appreciated as Mallon (2015) states that ‘constipation involves more complex physical and behavioural components and lacks a straightforward objective measure of severity’ (e1304). Studies highlighted the real-world challenges of using the ROME criteria for the diagnosis of CFC in clinical practice. Yang (2013) argues that ‘given the lack of familiarity with NASPGHAN guidelines, finding ways to increase awareness of these guidelines among pediatricians would be beneficial and likely increase consistency in management’.

Inner setting

Culture was identified as a barrier to successful implementation in nine studies,24,69,148,330,347,360–362 and is closely linked to an earlier construct – child’s needs and resources. The taboo nature of constipation and the reluctance of children, parents, healthcare professionals and wider society to openly engage in discussion about the subject negatively impacts on the implementation of interventions. Studies cited examples of families delaying help because of the stigma – hoping that it will resolve spontaneously. The culture of ‘not talking’ about constipation was identified across all levels of healthcare provision; from individual healthcare professional across the wider organisation; and was mirrored by the limited attention that constipation receives at the level of public health. A recent publication from a multidisciplinary group highlighted the costs of constipation, arguing that awareness of constipation urgently needs to be escalated to ‘encourage increased understanding of the condition and start implementing solutions that can begin to alleviate the problem’ (Cost of Constipation, p3). 277

Supporting clinicians caring for children and families who are living with CFC by providing a suitable climate to deliver a seamless ‘integrated’ service was identified as a facilitator in eight studies. 65,69,149,246,282,294,347,362 This frequently involved the vertical and horizontal integration of services involving the collaboration of professionals across disciplines. This was supported by good communication and strong networks described in two studies. 294,362

Another important consideration in the success of an intervention was how compatible the intervention was with the needs of the child. Poor intervention compatibility, as a result of poor tolerance of the intervention was reported in five studies as a barrier. 103,151,243,246,305 Conversely, interventions that were perceived as compatible were described as successful in 11 studies and could result in an improved QoL. 65,102,149,143,243,302,303,325,328,332,359 This frequently involved making sure that the right professional was available at the right time as highlighted by Hankinson (2018) ‘Having a nurse practitioner- and psychologist-led clinic allowed for patients to be seen more frequently (as frequent as monthly) to thoroughly evaluate medical symptoms and adherence to medical and behavioral treatments’ (p10). 143

Understanding the tension for change (i.e. why clinicians and families felt that the changes were needed now) was identified as a facilitator in 19 studies. 61,65,103,148–151,153,143,178,294,303,306,331,348,359,363,364 In most of these studies, it was the failure of treatments to improve CFC which acted as a trigger for clinicians and/or families to consider other treatment options as noted by Eicher (2007) ‘We realized we needed to take drastic measures to get him help’ (p2). 359 Sometimes the driver for change was having the expertise on board to push forward change as O’Connor reported that ‘When United Kingdom specialist paediatric HV and an expert in bowel management in children, Jackie Wade, came to work on the ward, she provided Smith with the impetus and the necessary practice guidelines to make a case for a specialist nursing service’ (p1). 347

The readiness for implementation was described as a barrier and facilitator. Several studies74,305,307,319,330,333,356 highlighted the lack of access to knowledge and information within and across multiple organisations as a barrier to readiness for implementation. Torres (2015) pointed out that ‘knowledge gaps regarding pediatric constipation may be a worldwide problem that should help in educational planning for pediatric residency programs and the pediatrics curricula of medical schools’ (p78). 333 The family role was also highlighted here as ‘the desire for information about their child’s care is usually very high. In some cases, families do not believe that their doctors are providing them with all the information they want or need’ (p646). 305 Improving access to education and knowledge about ‘multimodal’ CFC treatments and raising the awareness of the complex nature of CFC across physical, psychological and sociological factors as well as including families in decision-making was reported as facilitator in six studies. 63,305,328,330,333,342 Yang (2014) argues that we need to ‘… encourage pediatrics and GI societies to establish simple guidelines to be implemented during the course of medical school education’ (p78). 354

Limited available resources including expertise (behavioural and medical), capacity and workload were all cited as impeding implementation in five studies. 145,305,307,333,347 Trajanovska (2020) stated that ‘… most significant obstacle in the provision of timely management for constipation, a relatively low morbidity chronic condition, is wait time. Several factors may contribute to increased wait times, such as high volumes of incoming referrals, referral management and acceptance parameters, triage and scheduling procedures and prioritisation based on severity’ (p301). 307

Seven studies highlighted the importance of adequate resourcing for implementation success. 56,61,145,307,344,346,347 This included ensuring equipment was provided and schools having access to private bathrooms for children to use uninterrupted. Other studies commented on the benefits of redesigning care pathways to maximise available resources which often included streamlining referral and triage processes.

Characteristics of individuals

The ability to engage in the intervention and ensure successful implementation was cited as a barrier in 14 studies. 149,151,154,185,270,300,305,322,323,328,331,361,365,366 Multiple reasons for a lack of compliance or adherence with the intervention were cited across studies and included:

• poor communication with the child and/or parent;185,305

• motivation of the child (and family);185,305

• lack of observed efficacy of the treatment;154,305

• child’s refusal to take the medication305,322 because of taste167,361 or adverse events;302

• lack of time to deliver the intervention305,323 or failure to prioritise treatment ‘I forgot’;302,305

• financial difficulties;302

• Family concern about safety and side effects of drugs. 365

Self-efficacy was reported as a key component to the success of implementation in 33 studies,61,69,103,136,143,151,153,161,167,178,185,203,215,244,255,282,394,302,305,306,309,318,319,322,328,333,341,345–347,349,350,360 with several studies reporting innovative strategies to overcome barriers to compliance and improving treatment adherence. These included:

• tangible incentives such as star charts;143,203,350

• provision of individualised easy-to-follow instructions detailing treatment protocols to families;69,136,178,203,255,294,347

• involving family in the decision-making process;350

• sending e-mail reminders;305

• increasing knowledge and improving understanding61,102,333,350,365 [e.g. providing tailored education demonstration of the intervention to the parent and child using of models, visuals (e.g. Bristol stool charts)];136,203

• normalising the use of any devices needed to improve CFC by encouraging the use in everyday circumstances;185

• use of bowel diaries;69,153,143,313,318,319,322,328,345,346,360

• disguising the taste of medications to make them more palatable;341,349

• supporting the use of more complex interventions178 by means of facilitating home management179,292 and/or free samples to improve compliance;294

• involving other key stakeholders (e.g. school). 333

Eight studies identified knowledge and beliefs about the intervention as a barrier. 136,307,320,342,343,360,365,367 Studies described a lack of knowledge about current constipation guidelines and what recommendations are for optimal treatment and care plans; a lack of agreed terminology and definitions were also mentioned as a barrier, with one study reporting that how constipation is defined was often a limiting factor in it being operationalised in a clinical setting. As a result, studies reported that care providers relied on other diagnostic measures rather than adopting the clinical features described in the ROME criteria.

A lack of knowledge by some providers also provided a very real barrier in accessing timely care or referral which could exacerbate the condition as Modin (2018)136 stated ‘varying therapeutic approaches, with some children receiving inappropriately short or long treatments, which may cause maladaptation to toilet routines and create unnecessarily complex and costly situation’.

Several studies argued that raising awareness of guideline recommendations and developing clearer treatment protocols would reduce the variability in practice and improve the management of CFC. Studies also called for a clearer and simpler message about treatment. Other studies took a more pragmatic approach, calling for healthcare professionals to start a conversation and actively look for opportunities to engage with, and listen to families, so that the problems with constipation were not dismissed or overlooked. 347

‘As healthcare professionals, I think we have a responsibility to have this conversation with people as often as possible, as a poor functioning gut can have such a negative impact on wellbeing. Everyone should pay attention to their bowel habits to know what their ‘normal’ is and when something is not quite right – and, yes, that means turning around to see what it looks like!’ (p32). 296

Personal (individual) characteristics of healthcare professionals were described as barriers to implementation in seven studies. 167,300,302,320,334,350,365 Some studies argued that healthcare professionals were reluctant to change their habits, particularly their prescribing, or lacked motivation to deal with CFC. Several studies also described the unique challenges of timely diagnosis and clear communication from professionals to families about the importance of regular medication adherence in children. These communication challenges left particularly vulnerable children at a greater risk of having issues with CFC. 350

Process

The involvement of informal champions, such as a parent or carer, or a more formal champion, for example, a nurse or another clinician, was frequently reported as facilitating implementation. 61,63,92,102,136,143,145,148,150,151,154,161,185,204,232,282,294,305,311,317,333,334,346,347,350,359,363,365 Family involvement was reported in these studies as critical for the successful implementation. Parental involvement in their role as a ‘champion’ was extensive and included completing questionnaires, providing detailed histories, attending appointments (remote and face to face), documenting symptoms and recording information about their child’s bowel motions, delivering interventions and increasing input into decision-making about their child’s treatment. Formal ‘champions’ had roles in supporting children and young people, by building their confidence and self-esteem and did this in a variety of ways. For example, the establishment of a code word or signal was reported in one study, so that a child could indicate that they needed to leave a classroom.

We noted that there were little data reported about the process of planning for implementation, or the opportunity for reflecting and evaluating within studies. Some studies touched on the value of forward management planning, but only one described this in detail, highlighting the importance of consultation with all of the involved professionals as successful treatment for constipation is often protracted. 346

Interventions which may be beneficial

Our evidence synthesis highlights some interventions which may be beneficial. These include:

• Nurse-led clinics are feasible and could result in equivalent (or possibly better) outcomes than traditional physician-led clinics. We have very low certainty in this finding. Current guidelines do not make any recommendations relating to nurse-led models of care (but recommend further research to investigate effectiveness).

• An algorithm, or care pathway, used in primary care settings to guide the management and referral of children with constipation (including children with ASDs) may be beneficial. We have very low certainty in this finding and further research is required. The European Guidelines18 contains algorithms for the evaluation and treatment of children with CFC.

• Specialist (Level 2) services may have a beneficial effect on outcomes, but we have very low certainty in these findings and further research is required. The specialist services investigated included specialist dietitians, psychologists, physiologists and teams providing specific, generally multidisciplinary, bowel management programmes, including elements such as laxatives, diet and toilet training. NICE guidelines1,2,19,20 refer to combined treatment approaches. There is disagreement with European guidelines18 which do not recommend multidisciplinary treatment.

• Web-based information, following an appointment with a specialist, may be more beneficial than a phone call or no follow-up. We have low certainty in this finding, which is based on a single RCT. Current guidelines do not specifically recommend use of web-based information.

Evidence gaps

We found studies which investigated the effect of models of primary care services and the use of a constipation care pathway/algorithm within EDs, but limitations in the evidence meant that we were unable to reach generalisable conclusions.

We found no evidence which addressed questions relating to the key components and features of an effective team involved in care of children with CFC, or how a child’s ‘journey’ through the different levels of care could be made as seamless and direct as possible.

Interventions which may be beneficial

Our evidence synthesis highlights some interventions which may be beneficial:

• A cow’s milk-free diet may be beneficial (but our certainty in these findings is low). This is in partial agreement with NICE (2010) (CG99, 1.5.5)1 which recommends that ‘in children with idiopathic constipation, start a cow’s milk exclusion diet only on the advice of the relevant specialist services’. Further research into which children should be considered for a trial of a cow’s milk-free diet, the best way to implement this, the advice and information required by parents to do this safely, and the role of specialist services (and how parents can access this) is required.

• Education interventions for parents may be beneficial, particularly web-based interventions focused on education around toilet training (but our certainty in these findings is very low). This is in partial agreement with NICE (2010) (CG99, 1.5.4 and 1.8)1 and NICE (2019)22 which recommends information and advice for parents and children, and in agreement with European Guidelines (2014)18 which recommends (based on expert opinion) guidance on toilet training. Our evidence synthesis highlights that there is a gap in knowledge relating to the optimal content and mode of delivery of information to parents.

• Selenium supplements may be beneficial (very low certainty). The current guidelines do not make any recommendations relating to use of selenium. Further research is warranted.

Interventions which may not have an effect

Our evidence synthesis identified that the following interventions may not be beneficial:

• Probiotics (moderate certainty of evidence). These findings are in agreement with current guidelines18 which do not recommend routine use of probiotics.

• Additional dietary fibre (low certainty of evidence). There is some dissonance with the NICE (2017)20 guidance which states that ‘limited evidence for dietary interventions suggests that fibre can improve constipation’, but agreement with other guidance1,18–20,22 which recommends ‘adequate’ fibre, but not fibre supplements or unprocessed bran. Recommendations for daily fibre intake in children are available369,370; however, research to clearly define ‘adequate’ fibre for children, and to provide parents with this knowledge in a way which is accessible and understandable, is required. Future studies should consider the severity of constipation, and whether additional dietary fibre provided within home settings (i.e. Level 0) may benefit children with mild symptoms.

• Increased fluid intake. This finding is in agreement with current guidance, which recommends ‘adequate’, but not extra, fluids. 1,18–20,22

Evidence gaps

Our synthesis demonstrated that, although there have been some studies, there remains insufficient evidence to support conclusions relating to different milk formula, sugars, other/alternative dietary intake (such as cassia fistula emulsion, paraffin oil and magnesium-containing mineral water) and combined dietary and behavioural interventions as an alternative to laxatives. No recommendations relating to these interventions are made within current guidelines.

There were no studies identified, and therefore an absence of evidence, relating to a number of topics. These included gaps relating to diet (e.g. breast milk, gluten-free diet, alternatives to cow’s milk), physical activity and exercise, optimal content and format of information and support for children and families, recognition of constipation by children and families, and continence-related equipment and products.

Interventions which may be beneficial

Our evidence synthesis highlights some interventions which may be beneficial:

• Laxatives may be beneficial and are recommended by clinical guidelines. Our update of evidence relating to laxatives does not change the conclusions that can be drawn about laxatives and is in agreement with current guidelines. Our updated analysis re-states the benefit of PEG compared with placebo and with lactulose. However, it should be noted that this finding remains limited by the lack of new head-to-head evidence across multiple treatments, lack of evidence relating to effectiveness and safety of some laxatives, and lack of evidence relating to effectiveness of combinations of laxatives.

• Physical exercise focused on pelvic floor muscles may improve overall symptoms, defaecation frequency and stool consistency (low certainty of evidence). This is in partial agreement with current guidelines which recommend ‘daily physical activity that is tailored to the child ….’ (NICE, 2010, CG99 1.5.6)1 and normal physical activity levels. 18 Our synthesis identified limited evidence which supports physical exercise, focused on pelvic floor muscles, in additional to normal physical activity. Further research is required to investigate this.

• A combined pharmacological, diet and behaviour programme may be beneficial for children with and without typical development, including children with ASDs. Current guidelines do not explicitly address combined programmes, or the specific population of children with ASDs. Pragmatic studies to investigate the effect of combined programmes delivered by primary care services are required.

Interventions which may be harmful

Our evidence synthesis identified one trial which suggested that the combination of PEG plus domperidone may be more beneficial than PEG only in children with cerebral palsy. This study does not present data on side effects. There is a mismatch between the aim of the identified study and guidance relating to safety; MHRA advise against use of domperidone in children under 12 years, due to serious side effects. 371

Evidence gaps

There were no studies identified, and therefore an absence of evidence, relating to a number of topics including early identification and diagnosis, behavioural interventions combined with laxatives and information provision.

Interventions which may be beneficial

Our evidence synthesis highlights some interventions which may be beneficial:

• Combined oral and enema therapy in hospitalised children may benefit faecal disimpaction, but there is insufficient evidence to reach conclusions about the relative effectiveness of specific regimes. This is in partial agreement with current guidelines which list enemas as an option for disimpaction. 20 Further research is required to explore the optimal regime.

• Transanal irrigation may be safe, feasible and effective in children who have symptoms which have failed to resolve with conventional treatments. This finding is in partial agreement with the NICE (2018) Exceptional Surveillance21 which concludes that there are ‘promising results’, but further evidence required before the guideline on this topic is updated. Our finding is in agreement with a NICE Medial Technology Guidance372 report which concluded that, while data for children were limited, there was some evidence to suggest TAI may be beneficial for some children. Although we conclude that there is some very low certainty evidence that TAI may be effective, the quality of evidence remains poor and further research required.

• Biofeedback may be beneficial for children with abnormal defaecation dynamics (very low-certainty evidence). There is dissonance between these finding, and current clinical guidelines which conclude that evidence does not support the use of biofeedback. 1,18 However, how to effectively examine for abnormal defaecation dynamics within clinical practice remains an important uncertainty.

• Combined treatment programmes, incorporating behavioural therapy with pharmacological, and possibly dietary, interventions may be beneficial (very low certainty), but the evidence is insufficient to support specific conclusions relating to the effective components of a combined programme. Further research is required; however, in exploring combined programmes it is important to consider child-centred care. Further understanding of the pathophysiology of CFC is required in order to develop targeted treatment programmes.

Interventions which may not have an effect

Our evidence synthesis also identified that the following interventions may not be beneficial:

• Biofeedback for children with normal defaecation dynamics. This is in agreement with current clinical guidance, which does not recommend the use of biofeedback. 1,18 As stated above, how to effectively examine for abnormal defaecation dynamics within clinical practice remains an important uncertainty.

Interventions which may be equally effective

Our evidence synthesis provides some low certainty evidence that promelaxin microenemas may be equally effective to oral laxatives in infants (6–24 months), rectal enemas (dioctyl sulfosuccinate sodium) may be equally effective to oral laxatives, and that a soft suppository (LP101 free fatty acids suppository) may be equally effective to an enema. These findings have some agreement with current clinical guidelines which conclude that oral osmotic laxatives may have a similar effect to rectal enemas but are different from the NICE (2010) guidance1 which states that rectal medications should only be used for disimpaction if all oral medications have failed. Given the barriers to implementation of oral laxatives, further research to establish suitable alternatives to oral laxatives merit further investigation. This could include investigation of efficacy of combined therapies.

Interventions for which evidence is mixed

Our evidence synthesis identified some low-certainty evidence that the addition of regular rectal enemas may increase defaecation frequency in children with severe constipation, but not have any effect on overall treatment success or other outcomes and may cause some discomfort or distress. We found insufficient evidence to reach any conclusions about the relative effect of different types of enemas delivered within an ED. These findings have partial agreement with guidelines which concludes that enemas are an option for disimpaction if all oral medications have failed. 20 The retrospective studies included in our evidence syntheses report use of sodium phosphate enemas; it is important to note that potential serious reactions, including death, have been linked to phosphate enemas. 206 Given that there is currently insufficient research evidence about the relative effect of different types of enemas, these reports support the avoidance of phosphate enemas.

We found mixed findings relating to physiotherapy for children with CFC; current evidence does not support the routine referral to physiotherapy for all children with CFC. Guidelines do not currently make recommendation relating to physiotherapy, and the European guidelines18 state that routine multidisciplinary treatment is not recommended.

Evidence gaps

We found some very-low-quality evidence relating to the effect of TES, the use of physical therapy for children with severe cerebral palsy and dietary exclusion of fructose and lactose, but this was insufficient to support generalisable conclusions. Further research is required to determine whether these interventions have any benefits.

Key evidence gaps arising from where we found mixed, or limited, evidence relate to the relative effect of different interventions. This includes questions such as whether TAI may be an alternative to rectal interventions, and the relative effectiveness of different rectal interventions. There were no studies identified, and therefore an absence of evidence, relating to physiotherapy interventions and the use of tibial nerve stimulation.

Interventions which may be beneficial

Our evidence synthesis highlights some interventions which may be beneficial:

• Botulinum toxin may be effective in the management of CFC (very low certainty). Botulinum toxin is not mentioned in current clinical guidelines,1,18 but the NICE (2018) Surveillance21 does identify the need for further evidence relating to this intervention.

• ACE/MACE may be safe and effective, but there were high rates of (minor) complications, and our certainty in these findings are very low. This finding is in partial agreement with current guidelines which recommend referring children with unresolved symptoms to specialist centres for assessment of suitability for ACE/MACE.

• Sacral neuromodulation may be safe and effective, but we have very low certainty in these findings. Current guidelines do not address this intervention.

Further high-quality studies are required to investigate all these interventions.

Evidence gaps

We found some studies which addressed a number of other interventions, but for which the evidence was insufficient to support any generalisable conclusions. These included ACE compared to caecostomy button, ACE compared to SNS, and a range of surgical interventions (including colonic resection, colonic resection combined with appendicostomy, anorectal myectomy, ileostomy, colostomy, colectomy). Current clinical guidelines did not make any recommendations relating to these interventions. There was an absence of evidence relating to long-term implications of surgical procedures, effect of ACE/MACE in children with special needs and manual evacuation.

Interventions which may be beneficial

Our evidence synthesis highlights some interventions which may be beneficial:

• Connective tissue manipulation as a component of a programme for children with cerebral palsy and are receiving physiotherapy (low certainty).

• Herbal/traditional medicines may improve outcomes in children with CFC, but studies investigate a range of interventions and there is insufficient evidence to reach conclusions about the effect of specific interventions. There is a need for high-quality research studies; these should investigate safety and adverse events, as well as effectiveness.

Interventions which may be equally effective

Evidence from one small RCT suggests that dry cupping therapy of the abdominal wall could be equally effective as laxatives. Further research is required to investigate this finding.

Evidence gaps

We found some studies which addressed a number of other interventions, but for which the evidence was insufficient to support any generalisable conclusions. These included abdominal massage, chiropractic or osteopathic manipulation, reflexology and acupuncture. We found no studies relating to the effect of baby massage on constipation in infants.

Interventions which may be beneficial

Our evidence synthesis highlights that behavioural therapy techniques delivered by specialist practitioners may be beneficial (very low certainty). This is in partial agreement with the recommendations within current clinical guidelines; however, these are based on expert opinion and the lack of high-quality evidence is acknowledged. The European Guidelines18 recommend behavioural toilet training, but not the routine use of an intensive programme. The NICE (2010)1 guidance recommends referral to specialist psychological support where there is psychological distress.

Evidence gaps

We found studies which compared externalising treatment to other behavioural interventions, and which investigated the addition of psychotherapy to other behavioural therapies, but for both of these the evidence was insufficient to support any generalisable conclusions.

We found no studies which addressed a range of topics, including the best ways to support children and families to deal with the psychological effects of constipation.

Stakeholder involvement

Our SG comprised a small number of people, meaning that input could have been biased and thus not reflective or representative of a national or international viewpoint. We did not take any specific steps to ensure diversity within our SG or to involve people from under-served groups. The limitations relating to our stakeholder involvement have been discussed elsewhere (see Chapter 3).

Focus on treatments for chronic functional constipation

As defined by the NIHR call, this project focused on the effectiveness and implementation of interventions, and combinations and sequences of interventions, for childhood CFC (see Research questions). Consequently, we did not search for or synthesise evidence specifically focused on the recognition and diagnosis of CFC. Recognition and diagnosis of CFC are clearly as essential first step to successful management, and our conclusions are limited as we have not reviewed evidence relating to this key step. Our stakeholders raised concerns about this omission but accepted that this was beyond the scope of the SUCCESS project. Stakeholders highlighted that a child’s parents, carers or family are likely to be the first people who attempt to recognise the cause of a child’s discomfort, altered bowel habit, soiling, straining or changed behaviour, and that the lack of awareness of the problem of constipation, and societal taboos around engaging in open discussion about this, are barriers to identification and early treatment.

Stakeholders highlighted that there was an evidence gap relating to how parents, carers and children can recognise constipation, the relationship between early recognition and outcomes, how health professionals could facilitate the early identification and diagnosis of constipation, and education of professionals across the workforce to support this. Our review did include some limited evidence that suggested that educational interventions, particularly web-based interventions may be beneficial. Further research to identify the optimal content and format of providing information to children and families, including information around what is ‘normal’ and when to seek professional input is needed. This is supported by a recent SR, which concludes that lack of knowledge and understanding is a key issue faced by families/carers of children with CFC, and research to identify effective educational/informational strategies a priority. 392

Our focus was on management of CFC in children, and – while we considered psychosocial and behavioural interventions – we did not consider the psychosocial and behavioural factors which may contribute to poor outcomes. For example, there is evidence that some children may actively avoid using school toilets395 and concern has been raised about the impact of this on children with bowel problems. 396

Methodological approach and rigour

We adopted a step-wise approach to study selection and inclusion. A key limitation of this approach was that it is novel and there is a lack of best practice guidance and few examples to guide implementation. Our protocol lacked sufficient clarity relating to inclusion of different study designs, and as a result we included all relevant RCTs and primary studies for the majority of our evidence syntheses. This was inclusive and resulted in a comprehensive synthesis of evidence, but it took considerable time, and resulted in inclusion of many very-low-quality studies, which have limited generalisability. While our methods benefited from the use of independent reviewers to check study inclusion and data extraction, we failed to collect data which enabled us to calculate and report inter-rater reliability.

Following our step-wise approach, and specifically design to avoid duplication and research waste, where we identified a SR which we judged to be low ROB we included this within our evidence synthesis, rather than the original primary studies. There are a number of established debates relating to the methodological approaches adopted by SRs and meta-analyses. This includes debates relating to the merits of ‘lumping’ versus ‘splitting’ of interventions, comparators and outcomes. 397,398 We did not make any judgement relating to the methodological approach of a review, assuming that we assessed the review as low ROB. This meant that our synthesis could potentially include SRs and meta-analyses which had taken different approaches, and this may have impacted on the specificity of conclusions that could be drawn. For example, we included a review of probiotics [Harris, 2019,74 see Section Evidence of effectiveness: lifestyle interventions delivered by family/carers (Level 0)] which combined (‘lumped’) evidence from trials with a range of active and inactive comparators but, following completion of our review, a Cochrane review of probiotics397 was published which ‘split’ the same body of evidence. Whilst there remains to be consensus within this debate, our ‘neutral’ stance of considering the most up-to-date review, regardless of approach, seems justifiable, but may have impacted on the conclusions that could be drawn from the evidence.

• Our methodological approach involved the use of one comprehensive search, implemented for our scoping review, with studies then ‘tagged’ for inclusion within for focused reviews of effectiveness and factors relating implementation. This was designed as an efficient approach, avoiding duplication of searching across different evidence synthesis. To avoid duplication of searches from previous evidence syntheses, we included studies identified from the searches for the NICE guidelines and limited our searches to after this time period. There are arguably risks associated with this approach; we assumed that searches for previous projects were suitably comprehensive to address our broad RQ, and we utilised a broad search (focused on terms for constipation and children) which may have potentially lacked the sensitivity to find studies of some select interventions. This approach to searching meant that our electronic searches were from January 2011 onwards, but that we attempted to integrate findings from research published before 2011 by including evidence used to develop national and international guidelines. However, these guidelines often limited their inclusion criteria to SRs or RCTs only. For example, NICE guidelines1,19 only included other study designs if no SRs, meta-analyses or RCTs were identified, and did not search grey literature. Further, systematically identifying and integrating grey literature into our search results was challenging as there is little guidance in this area. Studies published before 2011 represent 15% of the total number of studies included within our review.

We followed a predefined protocol for decisions around statistical analyses and conducted narrative syntheses where results could not be pooled statistically. We extracted and reported details relating to the children included in studies, but we did not systematically group these according to the details of the populations studied. This limits the generalisability of our narrative syntheses. Due to the quality and heterogeneity of studies we were limited in the number of analyses that we could conduct. We did not conduct any subgroup analysis.

Quality of evidence

Findings from this project are limited by the quality of research evidence available. This includes a lack of high-quality studies and poor reporting. Our review could be criticised for synthesising studies which have used a wide variety of study designs. This means we have included all studies that may report some, albeit limited, evidence related to CFC. However, while many of these studies report piecemeal evidence, they do present a realistic picture of the evidence that is currently available in this field. They also highlight the gaps in evidence and underscore the need for enhanced methodological quality.

There are limitations in our approach to judging certainty of evidence. We grouped included studies according to the question that they addressed and judged our certainty in the evidence relating to this question. We used an approach based on the GRADE56,57 approach, in which we applied downgrades where we had concerns relating to our confidence in the findings of the studies addressing each question. The broad nature of our evidence syntheses, including a wide range of different study designs and including studies which addressed heterogeneous outcomes, meant that we were unable to reach judgements using the established GRADE approach, following specific guidance for individual domains of study limitations, inconsistency, indirectness and imprecision. As a consequence, our judgement of certainty in the evidence is arguably a crude categorisation, which falls short of the rigour which would be obtained by using GRADE approach.

Prioritisation of interventions to be addressed by review

In our project proposal we stated that we would work in partnership with stakeholders to identify the top priority interventions for which we should conduct focused evidence syntheses, anticipating prioritisation of a list of interventions, and/or intervention combinations. However, the stakeholders chose to focus on broad RQs rather than specific interventions/intervention combinations.

The stakeholders ranked the broad questions into high, medium and low priority for completion of evidence syntheses. However, the research team found it difficult to conduct ‘light touch’ syntheses for the lower priority questions, as they did not wish to compromise methodological rigour. Consequently, the completed evidence syntheses have been inclusive and comprehensive, providing a broad overview of all research in this field. The limitation of this is that a large volume of evidence has been synthesised, potentially reducing the accessibility of information and impacting on the level of detail that could be provided for individual studies.

Prioritisation of outcomes

A strength of our project was that our stakeholders reached consensus on the outcomes which were of greatest importance to include within our SR of effectiveness. However, our protocol did not clearly specify how we would use these prioritised outcomes to inform selection of studies for inclusion or synthesis of evidence within our review. Further, our synthesis of evidence relating to these outcomes was limited by the fact that studies often presented clinical outcomes as combined scores, for example reporting ‘constipation scores’ or ‘improvement scores’. Consequently, our narrative syntheses report findings relating to a wide range of different outcomes, as reported by the included studies. Our evidence synthesis may have been more concise and informative had we strictly adhered to reporting the prioritised outcomes.

Timelines and search dates

Conducted between January 2020 and May 2022, the project was subject to the working practice impacts of the COVID-19 pandemic and associated restrictions. During 2020–1 we experienced delays to the literature identification and retrieval of full texts due to the temporary closure of the British Library. Clinical trial registers (e.g. WHO ICTRP) and SR registers (e.g. PROSPERO) were frequently unstable as they struggled to cope with the volume of COVID-19-related studies. There were some changes to research staff, and a number of periods of staff absence over the course of the project, creating challenges to continuity. As a result, the project end date was extended. However, in order to facilitate completion within the available resources, it was agreed with our funders (e-mail correspondence 10 May 2021) that an update to the searches (conducted in March 2020) was not essential, meaning searches are more than 12 months out of date. However, our comprehensive search of trials registers meant that we had knowledge of relevant ongoing RCTs. In May 2022, we searched for completed RCTs and updated our syntheses accordingly. This means that we are confident that we are not missing any key evidence which is likely to impact on our conclusions. We may be missing some recent non-randomised studies, but the results of these are unlikely to impact on our conclusions or certainty in the findings. A small number of studies remain as ‘awaiting assessment’ meaning some data may be missing from the synthesis.

Assigning studies to levels based on our Pyramid

Our evidence syntheses were structured around the Pyramid developed by the SG. This introduced a number of challenges, particularly in relation to assigning individual studies to a level on the Pyramid. Studies of some interventions, which could feasibly be delivered within Level 0 or Level 1, were conducted by highly specialist (Level 2 and 3) teams. It was often difficult to judge whether the level at which the intervention was delivered within the study was a factor related to study conduct (e.g. aiding study recruitment by identifying participants through a consultant-led clinic) or was a key factor of the intervention. However, there was strength in the involvement of the wide group of stakeholders to inform consensus decisions.

Recommendations for future research

To address RQ4 (see Research questions) we identified and produced a list of evidence gaps (see Evidence gaps). We explored with our SG whether it was possible to generate and/or prioritise specific recommendations for future research from this collated list of evidence gaps. However, as reported in section Evidence gaps, stakeholders informed our decision not to generate more specific, focused recommendations for future research. The limitations in our SG members (see above) supported this decision. Consequently, our key recommendation for future research is that a ‘robust research prioritisation exercise involving a wide range of stakeholders’ is conducted (see Implications for research), and this project has not generated more specific recommendation for future research.