Reflections and Critical Directions for Toilet Training in Applied Behavior Analysis

Abstract

Achieving toileting independence is a critical skill that yields several benefits of pressing social, developmental, and health-related importance. The seminal behavioral approach to toilet training established the conceptualization of continence and framework for toileting research thereafter. Contemporary researchers continue to evaluate toilet training procedures that produce efficacious outcomes for young children that closely align with current applied behavior analysis (ABA) ethics and standards of practice. Despite the overall success of behavior-analytic toileting approaches, there are critical directions still worthy of consideration and investigation. In this paper, we acknowledge the seminal roots and many of the contemporary contributions. We also critically reflect on current practices while proposing necessary areas to advance behavior-analytic toilet training research.

Achieving toileting independence is a critical developmental milestone (Center for Disease Control and Prevention [CDC], 2021a). Promoting continence in young children increases general hygiene practices, which may prevent (or minimize) infection, discomfort, or risk for abuse (Kroeger & Sorensen-Burnworth, 2009; Sapi et al., 2009). Enuresis or encopresis (i.e., lack of bladder or bowel movement control) can produce high levels of personal and parental distress (Macias et al., 2006). Additionally, many families simply cannot afford to keep their children in diapers (National Diaper Bank Network, 2022) or possibly maintain costs associated with incontinence (e.g., elevated water or electricity bills due to laundry). Lacking toileting independence (or diaper dependency) is not only cost prohibitive but also prevents children from accessing school placements (Cicero & Pfadt, 2002). Therefore, identifying effective toilet training strategies is of pressing social importance.

Five decades of toilet training research should generate reflections on past and present methodology to move toward essential next steps. The current discussion aims to propose critical areas to advance toilet training research that involves children with a lack of toileting independence, most of whom have neurodevelopmental diagnoses. Some recent reviews discuss distinctions between toilet training approaches (a child-oriented versus structured behavioral approach; de Carvalho Mrad et al., 2021) and the overall effectiveness of procedures evaluated with children with developmental disabilities or autism spectrum disorder (Saral & Ulke-Kurkcuoglu, 2020; Simon et al., 2022). We will generally cover seminal and contemporary work that targeted in-toilet eliminations and toileting accidents (i.e., eliminations that occur off the toilet) to provide a foundation for our proposed directions. We mostly cite empirical studies published within the last decade, and tie in earlier work that complements crucial mechanisms and procedures to advance behavior-analytic toilet training research. Therefore, the primary focus of the current discussion is to propose a wave of research moving forward, rather than exhaustively reviewing the toilet training literature.

Seminal Roots to Present Day

An early contribution of applied behavior-analytic research involved targeting incontinence with institutionalized adults (Azrin & Foxx, 1971). Toileting had long been regarded as purely reflexive behavior produced by private stimuli; however, Azrin and Foxx demonstrated that toileting involves a complex chain of responses that are sensitive to operant contingencies (in accordance with the conceptualization of Ellis, 1963). The influence of their rapid toilet training method cannot be overstated. Before its publication, no empirically validated procedure established and maintained independent toileting in institutionalized adults with intellectual disabilities. The efficient acquisition of comprehensive toileting behavior occurred using reliable measurement systems (e.g., apparatus) and adoptable procedures. In all, their seminal work advanced behavioral conceptualization of continence and established a framework for toilet training research thereafter.

Some of the components in the seminal rapid toilet training method described by Azrin and Foxx (1971) remain active in contemporary research practices, whereas others have not endured the test of time. For example, changing undergarments, scheduling toilet trips, offering fluids, and providing differential consequences for in-toilet eliminations versus toileting accidents are present in recent research practices (e.g., Greer et al., 2016; Lomas Mevers et al., 2018; Perez et al., 2020). However, other components such as the use of tepid showers, delaying meals, removing opportunities to use chairs, bodily shakes, requiring clients to clean soiled clothing, floors, or chairs, and sitting clients close together to allow observation of other clients using the restroom do not conform to current ethical standards of behavior-analytic practices or research (Behavior Analyst Certification Board [BACB], 2020). Therefore, the rapid toilet training method in its entirety would not be recommended or used by contemporary, ethical behavior-analytic toilet training researchers.

Foxx and Azrin (1973) extended a variant of the rapid toilet training method to young children. They omitted many of the previously mentioned procedures and specified 16 components, a portion of which are still used (e.g., offering fluids, checking pants, positive reinforcement for in-toilet eliminations). However, the procedures still involved minimizing or removing access to sources of positive reinforcement (e.g., “distraction-free environment,” time-out from social interaction and preferred items contingent on toileting accidents), positive practice (repeatedly practicing going to and from the bathroom), reprimands, and remaining in soiled clothing. A subsequently well-cited and more socially acceptable toilet training model for children with developmental disabilities surfaced approximately three decades later (LeBlanc et al., 2005). This method included several components: (a) a progressive sit schedule, (b) differential consequences for successful urinations and self-initiations (i.e., requests for the toilet), (c) programmatic fluid intake, (d) prompting communication (for the toilet), (e) a urine alarm, (f) socially mediated punishment (reprimands and positive practice) for urinary accidents, and (g) underwear (Hanney et al., 2012; LeBlanc et al., 2005). Therefore, the procedure still involved the use of potentially aversive contingencies (e.g., urine alarm, reprimands, and positive practice).

In the past decade, some empirical extensions have either added socially mediated punishment components only after observing minimal progress (e.g., Lomas Mevers et al., 2018) or did not use these components at all (e.g., Greer et al., 2016; Perez et al., 2020), and still produced successful toileting outcomes with young children. In some cases, simplified procedures involving isolated (e.g., wearing underwear; Greer et al., 2016) or few (underwear, differential reinforcement, and a dense sit schedule; Greer et al., 2016; Perez et al., 2020) components have resulted in increased in-toilet urinations. Therefore, current research efforts seemingly further align with ethical considerations by prioritizing reinforcement-based and feasible procedures (code 2.15, BACB, 2020).

Contemporary toilet training components (some of which are described below) abide by ethical guidelines and are likely successful due to their conceptual basis. For example, taking a child to the bathroom on a sit schedule increases the probability of an elimination occurring in the toilet. Over time, the toilet should begin to exert control over eliminations given the correlation between sitting and the removal of a full (or partially full) bladder. In this case, a full bladder is conceptualized as an aversive stimulus and removal (or relief) of a full bladder negatively reinforces eliminating in the toilet. Other events such as the termination of the toilet sit might also fit with a negative reinforcement account (e.g., Luiselli, 2007). Additionally, differential reinforcement aims to increase in-toilet eliminations and decrease toileting accidents. Often, in-toilet eliminations are followed by access to positive reinforcers (e.g., preferred toys, fluids, or snacks and praise; Greer et al., 2016; LeBlanc et al., 2005; Lomas Mevers et al., 2018; Perez et al., 2020) and toileting accidents are followed by changing of soiled undergarments (e.g., Greer et al., 2016; Perez et al., 2020) or a neutral statement (e.g., Cagliani et al., 2021; Cocchiola et al., 2012). Adding positive practice (a positive punishment procedure) might decrease urinary accidents when enuresis persists after attempting reinforcement-based procedures alone (e.g., Lomas Mevers et al., 2018), but adding this component is apparently only necessary in some cases. Moreover, placing a child in underwear might serve a discriminative function, or establish toileting accidents as an aversive event (Greer et al., 2016; Simon & Thompson, 2006; Tarbox et al., 2004). For example, soiled clothing might function as positive punishment and losing access to preferred activities or underwear (when changing to clean clothes) might function as negative punishment for urinary accidents (as discussed by Greer et al., 2016). Another common component involves offering fluids on a time-based schedule, which typically corresponds with an increase in urinations and presumably establishes discriminative internal stimuli to signal an elimination is forthcoming. An increase in urinations should increase the probability of contacting the differential consequences for in-toilet urinations versus urinary accidents. Additionally, establishing an event (a full bladder via increased fluids) in combination with communication training, might result in self-initiating (asking or labeling) necessity for the bathroom. Finally, communication training involves prompting an individual to ask to use the bathroom to establish a relation between a communicative response and entry to the bathroom. Ideally, a portion of these prompted communicative responses correspond with in-toilet eliminations to facilitate transferring control from instructional prompts to internal stimuli (e.g., full bladder). Establishing communicative responses under the relevant stimulus conditions facilitates independent toileting when the sit schedule is thinned or discontinued (e.g., Perez et al., 2020, 2021). The aforementioned components are not intended to be exhaustive but were selected for discussion given they are present even in component-lean procedures (e.g., Greer et al., 2016; Lomas Mevers et al., 2018; Perez et al., 2020). Collectively, procedural shifts in toilet training research practices are rooted in present-day ethical standards and conceptual systems.

Reflections and Critical Directions

Toilet training has come a long way since the inception of the rapid toilet training method (Azrin & Foxx, 1971; Foxx & Azrin, 1973). However, the possibilities for extensions are vast given the gaps in the current behavior-analytic toilet training literature. As a result, we have reflected on current and prior research to identify critical areas to address moving forward. We consolidated these critical areas into four categories: (a) the identification and measurement of relevant variables, (b) consideration for medical conditions and collaboration, (c) attending to possible conditioning history with toileting stimuli, and (d) the development of cost-effective (or cost-efficient) models. Importantly, we propose these critical areas to expand the lens through which behavior-analytic researchers conceptualize and design toilet training evaluations. Most of the examples we provide rely on the breadth of research targeting urinary continence; however, we also incorporate and suggest a portion of these areas may be applicable to bowel movement continence.

Identification and Measurement of Relevant Variables

Many variables influence toilet training outcomes, including participant characteristics or repertoires, dependent measures selected for analyses, and secondary or related behavior. Commonly reported participant characteristics include demographics (e.g., age, neurodevelopmental diagnosis, communication modality), previously attempted but ineffective toileting strategies, and therapeutic or educational enrollment status (e.g., LeBlanc et al., 2005; Greer et al., 2016; Perez et al., 2020). The inclusion of comprehensive demographic information (e.g., race, sex) is particularly important for behavior-analytic research (review Jones et al., 2020). On some occasions, researchers have reported diagnostic and verbal skills assessment scores (e.g., LeBlanc et al., 2005). Reporting the presence of relevant diagnoses or medical conditions (e.g., avoidant restrictive food intake disorder, delayed gastric emptying) would highlight for whom we have identified effective toilet training procedures, and when to integrate other disciplines for collaboration. Additionally, researchers have considered the presence of “readiness” skills (Schum et al., 2002) or scores on the Profile of Toileting Issues (Matson et al., 2011). Specific verbal or “readiness” skills might predict a need for individualized treatment or differential sensitivity (or responsiveness) to the programmed contingencies. These same measures can permit identification of prerequisite skills, which researchers can use to develop inclusion criteria for study enrollment. Further, disclosing the heterogeneity of participant characteristics can inform the generality of findings and promote adoptability in practice.

Coarse dependent measures such as frequency of accidents (i.e., eliminations that occur outside of the toilet), frequency of successes (i.e., eliminations that occur in the toilet), percentage of successes, percentage of self-initiated bathroom trips, frequency or rate of self-initiations, and percentage of self-initiation correspondence (i.e., requests for bathroom followed by successes) provide evidence of toileting acquisition (e.g., Cagliani et al., 2021; Call et al., 2017; Greer et al., 2016; Lomas Mevers et al., 2018, 2019; Perez et al., 2020). Although a graphical display of in-toilet eliminations and toileting accidents is common (e.g., Call et al., 2017; Greer et al., 2016; Lomas Mevers et al., 2019), calculating the difference between total eliminations between baseline and treatment might indicate sudden (and possibly unfavorable) changes in elimination patterns. For example, if a child has a daily average of five urinations during baseline and zero urinations during the first few days of treatment, then this may indicate the child is withholding urination and would more likely benefit from an individualized treatment package (e.g., pull-up fading; Lomas Mevers et al., 2018; Luiselli, 1996a, b).

Additionally, when targeting toileting continence in outpatient settings, researchers should consider analyzing fine-grained measures such as nutritional intake (e.g., fluids, food), volume of urination, inter-urination-intervals (IUIs) or time between urinations, days between and texture of bowel movements, latency to the first onsite elimination, latency to elimination during a sit (Doan & Toussaint, 2016), percentage of sits followed by successes (i.e., sit-to-success correspondence), and status of undergarments upon arrival (i.e., type of undergarments, dry or wet). These additional measures can indicate: (a) individualized elimination schedules (e.g., the number of fluid ounces can inform when to schedule a toilet sit; Kroeger & Sorensen-Burnworth, 2009), (b) the presence of illness (e.g., a high frequency of urinations or short IUIs can indicate a urinary tract infection; CDC, 2021b), (c) evidence of stimulus control (e.g., high sit-to-success correspondence, short latency to elimination after sit), (d) individualized modifications to procedural components (e.g., latency to first on-site urination or average IUI can inform sit schedule, latency to urination during a sit can inform the sit duration), and (e) recommendations to caregivers or changes in caregiver behavior (e.g., if children consistently arrive to the clinic in dry diapers, caregivers can start keeping their children in underwear during car rides).

Recent correlational, difference, and risk-ratio analyses demonstrated improvements in secondary toileting behavior with the emergence of urinary continence (Perez et al., 2021), suggesting that comprehensive toileting interventions may be sufficient to observe changes in toileting behavior for many individuals. However, targeted interventions may be required for some individuals when improvements do not occur. Even if urinary continence is the target, it seems necessary to monitor secondary toileting behavior (such as bowel movements) to determine if any potentially unfavorable collateral changes are occurring (e.g., stool withholding). Researchers have recommended monitoring problem behavior (e.g., aggression, self-injurious behavior, disruptive behavior, negative vocalizations) as it may yield differential acceptability and feasibility of the procedures as well as delay or inhibit continence (Greer et al., 2016; Perez et al., 2020, 2021). Future researchers should report how often problem behavior prevents or results in discontinuation of study enrollment. These data would mark the prevalence of problem behavior prohibiting continence and identify when to intervene on problem behavior if it does not decrease with toileting acquisition (as observed in Perez et al., 2021). Relatedly, a functional analysis of toileting behavior may be necessary as it is possible an individual engages in problem behavior because of certain antecedent events (e.g., being wet or being taken to the restroom when they do not need to void), to escape or avoid the bathroom (i.e., negative reinforcement), to access a positive reinforcer (e.g., maintaining contact with attention or tangible items during play), or some combination. The assessment of problem behavior associated with toileting could therefore inform different function-based treatments. Along these lines, it will be crucial to evaluate a transition from the use of programmed positive reinforcers to naturally occurring negative reinforcement associated with toileting. In fact, it is presently not known whether the use of positive reinforcers is critical for the acquisition of continence, even at the outset for some individuals.

Consideration for Medical Conditions and Collaboration

Consideration of biomedical variables is crucial in the toileting process (Christophersen & Friman, 2004; Friman et al., 2006) and can facilitate when to collaborate with medical professionals or modify toilet training procedures to account for common illnesses (e.g., diaper dermatitis, urinary tract infections). The BACB ethics code suggests assessing and addressing medical needs under reasonable circumstances depending on the referred behavior (code 2.12, BACB, 2020). Specific to toileting, biomedical events may be acute or chronic, which may indicate the necessity for initial, periodic, or routine medical clearance or consultation. Medical clearance and collaboration (or a multidisciplinary approach) may be warranted when targeting encopresis (Call et al., 2017; Friman et al., 2006; Lomas Mevers et al., 2019). Chronic constipation can impact urinary continence, which supports awareness of gastrointestinal and urinary tract symptoms (e.g., urgency to void) and integration of medical specialists when focusing on toileting continence (Deshpande et al., 2012; Halachmi & Farhat, 2008).

Undoubtedly, researchers encounter common acute illnesses such as urinary tract infections, diarrhea, and diaper dermatitis (among others) while arranging contingencies to promote continence (CDC, 2021b; Mayo Clinic, n.d.; Klunk et al., 2014). It seems that researchers should familiarize themselves with signs or symptoms of possible acute illnesses directly related to toileting behavior. Behavior-analytic researchers are also well suited to arrange measurement systems to detect sudden changes in urinary or bowel movement patterns that may indicate illness. However, diagnosing and treating these illnesses is beyond behavior analysts’ scope of competence (code 1.05, BACB, 2020). Researchers should plan for modified toileting procedures when acute illness is suspected or when working with children who exhibit chronic illness (Frauman & Brandon, 1996). Additionally, researchers should discuss conditions during which coordinating multidisciplinary interventions is practical or likely to favor successful outcomes (e.g., Call et al., 2017; Lomas Mevers et al., 2019). Future researchers should consider assembling medically informed resources on preparing for and responding to illness during toilet training and disseminating these approaches (similar to Friman et al., 2006). They should also consider reporting the prevalence of these illnesses during research participation and the extent to which such illness may have impacted acquisition.

Attending to Possible Conditioning History with Toileting (or Related) Stimuli

A well-established conceptualization of achieving continence involves both respondent and operant conditioning (Azrin & Foxx, 1971; Ellis, 1963). Accordingly, an individual’s history with toileting (or related) stimuli can, therefore, contribute to differential acquisition of continence. For example, a history of urinary tract infections (painful stimulation) might establish the toilet as a conditioned aversive stimulus, which might evoke problem behavior in the context of toileting (even in the current absence of a urinary tract infection). Relatedly, the increased prevalence of hyperacusis (hypersensitivity to sounds others deem tolerable) in the autism community (e.g., Williams et al., 2021) may account, in part, for typical bathroom acoustics (e.g., hand dryers, toilet flush) delaying approaches to the bathroom. Additionally, the prevalence of abuse experienced by incontinent children merits consideration for trauma informed care (Rajaraman et al., 2022; Sapi et al., 2009). In some very unfortunate cases, toilet trainers (adults) or the stimulus conditions surrounding toileting might be conditioned aversive stimuli, which necessitates specialized care and collaboration with other professionals. Therefore, attending to the individual’s history and considering procedures to attenuate the aversive properties of toileting stimuli seems warranted.

Relatedly, when children begin toileting procedures, they often have a history with different undergarments than those prescribed for treatment (e.g., changing from pull-ups to underwear; Perez et al., 2020). Therefore, the history of solely eliminating in a diaper or pull-up might require individualized interventions to transfer stimulus control from the undergarments to the toilet (e.g., pull-up fading; Lomas Mevers et al., 2018; Luiselli, 1996a, b). In sum, it is not known how often and to what extent these variables prevent or delay the onset of toilet training. Gaining more information on the influence of conditioning histories may influence treatment selection. For example, if certain components of the bathroom are conditioned aversive stimuli, pairing those stimuli with preferred events and items may be necessary for counterconditioning prior to the onset of more standardized toileting procedures.

Developing Cost-Effective and Efficient Models

Effective toilet training procedures are essential; however, their viability may, in part, depend on their cost-effectiveness and efficiency in producing meaningful outcomes.

Both the cost of resources necessary to produce the outcome and the clinical relevance of the outcomes inform a cost-effectiveness analysis (Dams, 1997). One factor to consider is the cost of undergarments. For example, toilet training procedures commonly involve switching from diapers (or pull-ups) to underwear (e.g., Cagliani et al., 2021; Cicero & Pfadt, 2002; Greer et al., 2016; LeBlanc et al., 2005; Perez et al., 2020). Diapers are particularly costly, and some families depend on federal assistance programs to supply their children with them (e.g., Temporary Assistance for Needy Families). Researchers should consider completing analyses that project cost savings when a child no longer uses diapers or pull-ups due to acquiring continence in underwear. Indeed, such an evaluation would be consistent with ethical standards (code 2.17; 2.18, BACB, 2020). Additionally, researchers might consider how long they keep children in diapers or pull-ups before implementing treatment. The multiple baseline across participants design is often used to demonstrate experimental control in toilet training studies (e.g., LeBlanc et al., 2005; Perez et al., 2020). Therefore, some participants remain in diapers or pull-ups for extended periods, which results in an increased cost for their participation. It would be advantageous to identify the minimum number of sessions or exposure to contingencies (e.g., sit trials, differential consequences for types of eliminations) necessary to inform when to implement treatment. Researchers may consider using consecutive controlled case series or randomized clinical trials to support the generality of findings in toilet training evaluations (Hagopian, 2020; Lomas Mevers et al., 2018, 2019). Moving forward with efficient models may inherently reduce the cost of toileting procedures and, therefore, increase acceptability and adoptability in practice.

Concluding Remarks

Toilet training is a pressing area with social relevance in the behavior-analytic literature and practice. We presented an account of seminal, contemporary, and crucial future research in relation to toilet training empirical evaluations. Our reflections on prior research and proposals for future research are in no way intended to be exhaustive. The specified critical directions attempted to go beyond well-documented prospective areas of toilet training research (e.g., generating conceptually sound technology for other relevant behavior such as self-initiations and bowel movements; Perez et al., 2021; identifying effective components; Greer et al., 2016). Instead, we hope the current discussion emphasizes an acknowledgment of behavior analysis’ five-decade history and its current research practices, as well as generates critical reflections of and advancements for toilet training technology. The next decades to come of toilet training research are critical and rich with possibilities and importance.

Data availability

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

Declarations

Conflicts of interest

We have no conflicts of interest to disclose.