Pediatric tube weaning: A meta-analysis of single-subject research studies

Background: Receiving adequate nutrition is essential for child growth and development (Sharp, Jaquess, Morton, & Herzinger, 2010). However, as many as 45% of typically developing children and 80% of children with medical or developmental differences experience feeding challenges (Chatoor, 2002; Levy, et al., 2009; Manikam & Perman, 2000). While some difficulties resolve on their own, between 3-10% require medical treatment, including gastrostomy tube placement (Manikam & Perman, 2000). Some patients, when medically appropriate, may transition from tube to oral feeding; however, interventions widely vary and often utilize a multidisciplinary approach, including behavioral, nutritional, and medical components (Sharp, Volkert, Scahill, McCracken, & McElhanon, 2017). Further, the providers involved, setting (e.g., inpatient, outpatient), and dosage of interventions is highly mixed, making it difficult to compare interventions. Given that lengthy interventions can increase cost and burden to families, it is crucial to understand which intervention components lead to successful tube weaning. Purpose: The current meta-analysis sought to examine intervention components that significantly contribute to treatment success (i.e., number of children weaned from tubes; percent kilocalories consumed orally) within single-subject research design studies published from 1987-2018. While treatment protocols vary, many include similar components, including multidisciplinary teams and the use of hunger provocation (Gardiner, Vuillermin, & Fuller 2017) Design: We conducted a systematic review and meta-analysis of single-subject research designs (N=24) to examine whether treatment components (i.e., setting, treatment modality, multidisciplinary teams) resulted in (1) a significantly higher percentage of patients weaned from a tube and (2) a higher percentage of kilocalories consumed orally. Results: Overall, treatments were successful in weaning 55.5% of children from feeding tubes. When evaluating factors contributing to patients weaned, analyses found that children treated in an inpatient or day treatment setting were 3.6 times more likely to wean from feeding tubes than those in outpatient settings (χ2 = 3.98, p = .046). Hunger provocation did not significantly predict tube weaning success (χ2 = .75, p = .39), nor did having a multidisciplinary approach (χ2 = 1.62, p = .20). This result should be interpreted with caution given that some studies utilized individuals from different disciplines for multiple components (e.g., speech-language pathologists using both oral motor and behavioral techniques). The second set of analyses focused on factors impacting proportion of kilocalories consumed orally (as opposed to total proportion of the sample fully weaned). Results indicate that there was not a significant effect of treatment setting on the percent of calories consumed orally following treatment, controlling for baseline oral consumption [ F(1,25) = .14, p =.71, partial eta squared =.01]. Hunger provocation was not examined statistically as very few studies with hunger provocation reported on kilocalories consumed (n=2). Use of multidisciplinary teams was not statistically significant in predicting kilocalories consumed, [F(1,14) = 3.53, p =.085], but the effect size was large, partial eta squared = .23. Although these findings differ slightly from the results focused on the proportion of the sample fully weaned, it is important to note that kilocalories consumed orally is not consistently reported in tube weaning literature. Conclusion: Weaning procedures used in single-subject research design studies are largely effective, possibly more effective in inpatient and day treatment settings. It is important to consider differences between the inpatient protocols and outpatient protocols that led to greater tube weaning. Although some factors examined did not yield statistically significant findings, it may be that components were not clearly defined or described within the literature. Future research should clearly define treatment components to assist with clinical applications. In addition, treatment components should be evaluated in randomized clinical trials with larger samples.