Applying the Science of Habit Formation to Evidence-Based Psychological Treatments for Mental Illness

Abstract

Habits impact nearly every domain of our physical and mental health. While the science of habit formation has long been of interest to psychological scientists across disciplines, we propose that applications to clinical psychological science have been insufficiently explored. More specifically, EBPTs are interventions targeting psychological processes that cause and/or maintain mental illness and that have been developed and evaluated scientifically. An implicit goal of EBPTs is to disrupt unwanted habits and develop desired habits. Yet, there has been insufficient attention given to habit formation principles, theory and measures in the development and delivery of EBTPs. Herein we consider if outcomes following the receipt of an EBPT would greatly improve if the basic science of habit formation were more fully leveraged. We distill six ingredients that are central to habit formation and demonstrate how these are relevant to EBPTs. We highlight practice points and an agenda for future research. We propose that there is an urgent need for research to guide the application of the science of habit formation and disruption to the complex “real-life” habits that are the essence of EBPTs.

A habit is defined as a learned action that is performed with minimal cognitive effort (Lally & Gardner, 2013; Wood & Neal, 2007). The importance of habits for our everyday lives cannot be overstated. Habits impact every domain of our physical and mental health including our exercise frequency and intensity, when and how we sleep, what we eat, the content and patterning of our thinking, how our attention is allocated and captured, whether we tend toward approach or avoidance and so much more. Therefore, our habits hold a key to the major health challenges we all face as we attempt to fashion our lives to incorporate healthy habits and reduce or—even better—eliminate unhealthy habits (Gardner, Rebar, & Lally, 2019; Verplanken, 2018).

The science of habits has long been of interest to psychological scientists. In 1890, William James wrote this statement about the critical importance of habits: “The more of the details of our daily life we can hand over to the effortless custody of automatism, the more our higher powers of mind will be set free for their own proper work” (James, 1983; p. 34). The behaviorists demonstrated that habits develop and are strengthened through repetitions of associative learning and reinforced responses (Hull, 1943; Skinner, 1938; Thorndike, 1998). Cognitive scientists showed the importance of automaticity in the formation of habits (Schneider & Schiffrin, 1977; Wason & Evans, 1974). Neuroscience, computational approaches and social and health psychology have also played critical roles in advancing the field. While the study of habits clearly crosses interdisciplinary boundaries (Verplanken & Orbell, 2003; Wood, 2017; Wood & Rünger, 2016), the central concern raised herein is that applications to clinical psychological science have been insufficiently explored.

The past several decades have been marked by exciting advances in evidence-based psychological treatments (EBPTs) for mental illness and a broad range of psychological health problems. EBPTs are interventions targeting psychological processes that cause and/or maintain mental illness and that have been developed and evaluated scientifically (Barlow, Bullis, Comer, & Ametaj, 2013; Chambless & Hollon, 1998). EBPTs are clearly effective and are considered to be front-line treatments for many mental illnesses (Division 12 of the American Psychological Association, 2016; Layard & Clark, 2014; National Institute for Health and Care Excellence, 2020). However, there is room for improvement (Kazdin, 2018). In this paper, we propose that applying the science of habit formation to EBPTs may greatly improve outcome. It is clear that an implicit goal of many EBPTs is to disrupt unwanted habits and to develop new desired habits. However, (a) the extent to which existing EBPTs can successfully disrupt and build habits is unknown and (b) there has been insufficient attention to habit formation principles, theory and measures in the development and application of EBPTs. The goal of this paper is to distill ingredients of habit formation that could be relevant to EBPTs in a way that is transdiagnostic (i.e., relevant across disorders) and pantreatment (i.e., relevant across EBPTs).

Our motivation for considering a deeper application of the science of habit formation within clinical psychological science is that many researchers and clinicians consider the process of habit formation to be a “passive phenomenon,” or “a ‘natural’ outcome of the behavior change process” rather than a process that can be specifically planned for and guided (Stokes & Baer, 1977; p. 349). In contrast, as will become evident, there are clear principles and strategies we can draw from, adapt and infuse into existing EBPTs to more intentionally incorporate the science of habit formation. Therefore, we propose the study of habits is an important, fertile, creative and fascinating domain for future research.

The great potential for leveraging existing knowledge on habit formation has already been recognized in subfields. For example, habit is included within the formulation of specific problems, such as depressive rumination (Watkins, 2018; Watkins & Nolen-Hoeksema, 2014), obsessive compulsive disorder (Gillan et al., 2014) and hair pulling, tics, nail biting and skin picking (Azrin & Nunn, 1973). In addition, addictions have been conceptualized as habits. However, more recently the conceptualization of addition has expanded to recognize a broader range of issues (Everitt & Robbins, 2016). In this paper, we consider if the outcomes from a broader spectrum of EBPTs might improve if we were to more fully embrace the science of habit formation and train providers in these skills.

It will become evident that there is a great need for naturalistic studies to delineate the contributors to knowledge about the multiple complex “real-life” habits that are tackled in EBPTs. In the absence of these, we can only conjecture about applications to EBPTs. Nonetheless, possible implications for EBPTs will be summarized in Table 1. Although points will be illustrated with examples from various EBPTs, most often the examples will be drawn from cognitive behavior therapy for insomnia (CBT-I) (Morin & Espie, 2003; Perlis, Jungquist, Smith, & Posner, 2005) and other related sleep interventions (Harvey & Buysse, 2017). This focus was selected for several reasons: sleep problems are highly prevalent, sleep health behaviors are ideal targets for habit formation and CBT-I is the frontline treatment (Qaseem, Kansagara, Forciea, Cooke, & Denberg, 2016; Riemann et al., 2017; Trauer, Qian, Doyle, Rajaratnam, & Cunnington, 2015). Moreover, the existing literature on sleep interventions seek to disrupt several unhelpful habits. For example, sleep problems are often maintained by habits such as irregular bed and wake times, excessive and poorly timed caffeine consumption or technology use or insufficient or poorly timed light exposure during the day.

Table 1.

Summary of Habit-Promoting Tips for Providers of EBPTs

1. Habits are independent of goals

Goal setting, motivational interviewing and psychoeducation are initial steps

Within psychoeducation, include education on the building blocks of habit formation

2. Habits are cued by specific contexts

Set up cues that are salient, accessible and perceptible

Facilitate the repetition of the desired new habit in a stable context

Set-up explicit environmental cues for the habits that patients wish to develop and/or explicitly identify and remove cues for habits that patients wish to disrupt

Frame the functional analysis process as a method to uncover the cues to undesirable habits and to develop new desired habits

3. Habits are learned via repetition

Measure habit formation as part of progress monitoring

Ensure patients have sufficient practice/repetition to build new habits

Provide patients with the skills and knowledge to modify the habits that are formed when needed so that repetition and habit formation can continue

Use implementation intentions to promote repetition/practice

Provide reminders to engage in repetition/practice

4. Habits are automatic

Functional analysis and self-monitoring can be used to “unpack” the automatic processes of an unwanted habit

Consider using behavioral experiments

Consider substituting a new more helpful habit in response to the same cues as the unwanted habit. Include a theoretical rationale for the substitution strategy and for selecting the specific substitute behaviors

Guide patients to conceptualize and select substitute behaviors that are enjoyable and consistent with their values

Implementation intentions can be useful to develop automaticity to the substitute

Measure the development of automaticity as part of progress monitoring

5. Reinforcers promote habits

For disrupting unwanted habits, use functional analysis to discover reinforcers

Reinforcers may be helpful at the beginning of the habit formation process

Consider how to optimize the schedule of reinforcement for the habit formation process

Consider how reinforcers can be delivered following the completion of a course of treatment, such as patient self-delivered reinforcers

Examine the impact of reinforcers a regular basis and modify according to the individual’s subjective evaluation of the reinforcing properties

Build positive associations with the new desired habits

6. Habits Take Time to Develop

Consider delivering booster sessions to promote habit formation

Consider developing “bundles” of desired behaviors that can be chunked together as the unit for habit formation

Theory

Although the understanding of habits varies across scholars and disciplines—and there are ongoing debates (Gardner, Abraham, Lally, & de Bruijn, 2012; Trafimow, 2018; Wood & Neal, 2007)—the theoretical grounding for this paper is drawn from health psychology, due to its conceptual proximity to clinical psychological science. In the health psychology literature, habit formation is understood to be a learned process whereby a behavior (the desired habit) becomes paired with a stable context cue and, via repetition, come to trigger an automatic impulse to engage in the habit (Gardner, 2015). Repetition reinforces the behavior-context association. Reinforcement motivates and strengthens repetition. With ongoing repetition, the stable context cue becomes sufficient to activate the association. In other words, the context triggers the impulse to perform the behavior, with minimal cognitive effort or intention and the habit has become more automated and less reliant on our goals (Verplanken, 2018). This definition is depicted graphically in Figure 1. The six elements of habit formation discussed in the main body of this paper were distilled from this theory. Further, this theory depicts the way in which the six key elements interact.

Figure 1. Theory of Habit Formation.

There are two important issues to address before proceeding. First, within this theory and throughout this paper, the terms “behavior” or “action” are used to refer to the target of habit formation. However, these are used as umbrella terms that refer to and encompass a broad array of behavioral, cognitive and emotional processes (e.g., Verplanken, Friborg, Wang, Trafimow, & Woolf, 2007). Second, within EBPTs we are concerned with the development of new habits and the elimination of unwanted habits. For the latter, following Gardner and Lally (2018), we will use the term ‘habit disruption’ in recognition of the possibility that the processes underlying habits may always remain in place, even if a stronger alternative habit is formed.

Distilling the Elements of Habit Formation

In this section, the six key aspects of habits that feature in the habit formation theory just presented are considered along with potential applications to EBPTs and domains for future research.

1. Habits are Independent of Goals.

Habits are typically the result of a goal that an individual had set and pursued in their past (Wood & Neal, 2007). Indeed, the path to habit formation begins when a person repeats a behavior in a specific context in pursuit of a goal. Interestingly, once the habit is established, it tends to endure and be independent of any new goal that we set, even if the consequences of the habit are unwanted and/or have become aversive (de Wit & Dickinson, 2009; Orbell & Verplanken, 2010). In other words, goals and habits can become divergent. A vivid example is substance use. Initially, a substance might be used to manage social anxiety, fit in with peers or for the hedonic value. Once the use of the substance becomes habitual it can be hard to disrupt even if the person desperately wants to quit because of hazardous health consequences or difficulty retaining relationships or employment. The goal has changed to “getting my life back on track” but this new goal does not easily shift the habit.

Indeed, most scholars of habit formation emphasize a shift from goals to habits (Cushman & Morris, 2015; Gardner, 2015; Wood & Rünger, 2016). Further, Pittenger and Taylor (2018) review data showing that the biologic systems underpinning goal pursuit are adapted for new or complex environments whereas the systems for habits are adapted to improve efficiency but are inflexible. The bottom line is that when the goal and habit systems are in balance and used skillfully, a person is well equipped to efficiently navigate the complexities of the various environments they encounter. However, a failure in the balance between goals and habits can contribute to “inappropriately rigid behaviors and a range of psychopathology” (Pittenger & Taylor, 2018; p. 322).

A broad range of approaches within neuroscience, including animal and human studies, optogenetics, lesion studies and the use of chemical inactivation, have identified a separation in the brain circuits and systems that are associated with goal-directed versus habitual behavior. Across these various approaches, there is consensus that habit formation involves a gradual transition from flexible and goal-directed behavior associated with the pre-frontal cortex to brain regions such as the striatum, particularly the dorsal striatal system, the medial prefrontal cortex, the caudate nucleus, and regions of the basal ganglia (Balleine & O’doherty, 2010; Dezfouli & Balleine, 2012; Graybiel & Smith, 2014; Smith & Graybiel, 2016; Yin & Knowlton, 2006). For stronger habits, such as addiction, the anatomy and circuitry implicated are similar but also include midbrain dopamine cell groups and limbic parts of the pallidum (ventral pallidum), the thalamus (mediodorsal nucleus), and amygdala (Everitt & Robbins, 2005; Kalivas & Volkow, 2005).

Application to EBPTs.

EBPTs already encompass several components that are, at least in part, designed to support patients in identifying, clarifying and pursuing their goals. Specifically, goal setting is a feature of most EBPTs, along with progress monitoring to track progress toward goals. Motivational interviewing is used to clarify the goal and build the motivation for change (Miller & Rollnick, 2002). Psychoeducation is typically given to provide information about the benefits of pursuing the chosen goal, such as improving sleep-related health, quitting smoking, reducing anxiety or depression, etc. The weekly home practice, common across EBPTs, also contributes to repeated practice of goal pursuit. However, EBPTs should go further to recognize that these are only the very beginning steps and that the additional specific steps, guided by the elements of habits distilled here, are necessary to truly form new habits and disrupt unwanted habits.

It is likely that the popular understanding of habit formation does not recognize the shift from goal-directed behavior to habit. Given this, within EBPTs we should provide education on the habit formation and disruption process. One EBPT, rumination-focused CBT (RFCBT), incorporates this approach (Watkins, 2018). The provider explains the characteristics of rumination as a habit; namely, that rumination tends to be automatic, is triggered by various cues, will be hard to change and will recur under conditions of stress or tiredness. The approach also prepares patients for inevitable setbacks in breaking the rumination habit and developing new desirable habits.

2. Habits are Cued by Specific Contexts.

Habits are formed via the direct association between a stable contextual cue and a behavior (Figure 1). Cues can take several forms; they can be internal (e.g., a thought or body sensations) or external (e.g., clock time, they can be deliberate (e.g., making a coffee before sitting down to work) or inadvertent (e.g., grabbing yet another mini-donut from the box a co-worker left in the office) (Wood & Rünger, 2016). Importantly, cues that are salient, accessible and perceptible are more likely to become associated with habits (Gardner & Lally, 2018).

There are certain types of cues that may more powerfully assist in habit formation. Gardner and Lally (2018) propose that event-based cues (e.g. “after breakfast”) may be more suitable relative to time-based cues (e.g., “at 10a.m.”), which require conscious monitoring. These scholars also point out that while any contextual feature can become a habit cue, some contexts may be more suited to supporting habit formation than others. For example, Pimm et al. (2016) found that people who consistently exercised with the same people, in the same part of their routine, or in the same mood, reported stronger physical activity habits. Also, Lally et al. (2010) showed that when people perform a behavior repeatedly in the same context (e.g., taking a walk after dinner), over time the context automatically triggers the behavior. Interestingly, an established habit can serve as a cue to forming a new habit (Judah, Gardner, & Aunger, 2013). Judah et al. (2013) compared habit formation among those who were instructed to floss after teeth brushing relative to before teeth brushing. Habit formation was stronger for those who flossed after teeth brushing. This raises the likely importance of the position of cues within existing bundles of habits (e.g., nighttime routine).

One path to disrupting unwanted habits is to reduce the contact with the cues associated with the unwanted behavior by changing the environmental context (Gardner et al., 2019). However, just as habits are easier to build within stable contexts, they are likely to be harder to break within stable contexts (Carden & Wood, 2018). This leads us to the Habit Discontinuity Hypothesis which proposes that attempts at behavior change will be more effective if they capitalize on moments of change, such as a change of job, birth of a child or a house move (Verplanken & Wood, 2006). These are natural life events that involve less contact with powerful old cues for unwanted habits and that create an opportunity for behavior change. In support, Heatherton and Nichols (1994) asked participants to write about experiences with successful or failed life changes (e.g., quitting smoking). Many of the participants (36%) who had successfully changed had moved to a new location whereas only 13% of reports of unsuccessful attempts involved moving. Also, 13% of successful change reports involved a change in environmental cues, whereas none of the unsuccessful reports involved such a shift.

Application to EBPTs.

There are multiple burning questions for future research. For patient’s seeking an EBPT to help them disrupt an unwanted habit, which types of cues are easier and which are harder to disrupt? For those seeking to build new desired habits, which types of cues promote efficient habit formation (e.g., internal, external, event-based, time-based)? Within EBPTs, how do we position cues optimally within a set of behaviors that are bundled together in the pursuit of building a sequence of new habits?

A relevant distinction has been drawn between habit instigation and habit execution (Gardner, Phillips, & Judah, 2016; Gardner, Rebar, & Lally, 2020). Habit instigation describes the processes involved in the selection and initiation of a behavior. For example, selecting to begin the wind-down from all available options and committing to performing it, and taking the first step in the ensuing “wind-down” sequence before bedtime (e.g., closing the computer and walking over to dim the lights). Habit execution refers to the processes that contribute to enacting the habit itself (e.g., engaging in the full wind-down sequence). Gardner et al. (2016) have reported that measures of habit instigation are more predictive of enacting the desired habitual behavior, relative to measures of habit execution. Future research is needed to determine if this finding replicates in the context of EBPTs.

While we await answers to these domains of future research, consider that most EBPTs include a functional analysis which involves mapping out the sequence of stimuli and responses, which typically includes behaviors, cognitions and emotions, for a situation of interest. However, explicitly framing the functional analysis process as a method to uncover the cue/s to undesirable habits and to identify potential cue/s to developing new desired habits has potential to lay a strong foundation for intervention. For example, in RFCBT, functional analysis is used to determine how, when, with whom and where the rumination habit does and does not occur as well as its antecedents (to spot the warning signs and triggers to the habit) and consequences (Watkins, 2018).

In EBPTs, we often make use of contextual cues by strategically discussing where to place the self-monitoring form (e.g., leave the daily sleep diary and a pen on the breakfast table) and placing cues in the environment to remind the patient of their homework (e.g., post-it note on the bathroom mirror to remind the patient of the time they should start their wind-down routine when they brush their teeth each night). However, within EBPTs there is room for a more explicit emphasis on setting-up environmental cues to help with the formation of new desirable habits that patients wish to develop and a more explicit emphasis on identifying and removing cues for habits that patients wish to disrupt. Once the habit is formed in a stable context, research is needed to determine how to generalize the habit to other contexts

Marteau et al. (2012) draws attention to the relevance of Tolman’s law of least effort which proposes that we can alter cues in the environment to make the least effortful course the most likely. Applying this to EBPTs, we can consider how to position cues to make it EASY to engage in desired habits and HARD to engage in unwanted habits. Relatedly, Rothman et al. (2015) propose introducing behavioral friction to existing contexts that make it harder for people to follow their unhealthy habits. For example, a common cue for smoking is drinking an alcoholic beverage. When UK pubs banned smoking, people could no longer smoke while drinking (Orbell & Verplanken, 2010) and this disrupted the automated connection between the cue (drinking) and the behavior (smoking). While this is a population-level intervention, behavioral friction could also be incorporated at the individual-level within EBPTs. Carden and Wood (2018) highlight environmental reengineering interventions that involve changing the structure of everyday decisions. For certain eating disorders, it may be helpful to change the type of food available. Altering the environment, such as dedicating a prominent place for fruits and vegetables on the kitchen counter, might also guide people into “rip currents,” potentially leading to a cascade of changes that help maintain new behaviors, including shifts in identity (e.g. I am a healthy eater) (Carden & Wood, 2018).

Another related approach to altering the cues in the environment is stimulus control in which a behavior is triggered by the presence or absence of a specific cue or set of cues. For example, stimulus control is a powerful EBPT for the treatment of insomnia (Morin et al., 2006). Stimulus control for insomnia involves changing the behavior of the sleeper so that the bed is associated with sleeping and not with wakefulness, anxiety or tossing and turning. For example, insomnia patients are asked to go to bed only when sleepy and if they do not fall to sleep in 15–20 minutes then go to another room until they are really sleepy, then return to bed. They repeat this process until they fall to sleep. The patient is also instructed to use the bed only for sleep and for sex (Bootzin, 1972). Other changes made to the environment to improve the association between bed and sleeping include moving work, study and meals to a location that is not the bed.

Furthering our prior discussion on trying to develop multiple habits simultaneously, Wood and Neal (2007) note that performing multiple behaviors in response to a single cue dilutes the mental association between that cue and any one behavior, limiting the likelihood that a behavior will become habitual. Unfortunately, EBPTs typically involve multiple complex habits. One potential solution is to test if it is possible to achieve habit formation involving multiple behaviors by devising a cue that triggers a tightly coupled bundle of behaviors. Consider a teenager whose parent provides the cue “It’s time to get up for school.” An example of a desired bundle of behaviors triggered by this cue would be: (a) getting out of bed, (b) opening the curtains to let sunlight in, (c) making the bed so it’s hard to get back in and (d) taking a shower to promote wakefulness.

The discussion thus far has focused on removing the cue. As highlighted earlier, several scholars (Wood & Neal, 2007; Wood & Rünger, 2016) draw attention to another possibility; namely, inhibit the habitual response once the cue has activated the habitual behavior. In one application of this with relevance to EBPTs, Quinn, Pascoe, Wood and Neal (2010) demonstrated that “vigilant monitoring” in the form of thinking “don’t do it” and watching carefully for slipups was useful to control strong habits. However, it is possible that some types of suppression may trigger a rebound in the inhibited unwanted behavior or thought (Wenzlaff & Wegner, 2000), a concern that was raised about “thought stopping” approaches for obsessions in obsessive compulsive disorder (Purdon, 1999).

3. Habits are Learned via Repetition.

Repetition is a key component in forming a new habit. As we repeat or practice a behavior, in a stable context, the habit starts to form and our intentions and goals related to that behavior gradually become less influential (Carden & Wood, 2018). Repetition exerts its effects via increased pairings between the stimulus and the response (Hull, 1943; Skinner, 1938; Thorndike, 1998). Repetition has been proposed to (1) improve skills, (2) ensure the behavior is selected with less effort and (3) automate behavior selection (Haith & Krakauer, 2018).

Although various creative designs have been developed to study habit formation in a laboratory setting with humans (e.g., Daw, Gershman, Seymour, Dayan, & Dolan, 2011; de Wit et al., 2018; de Wit et al., 2012; Luque, Molinero, Watson, López, & Le Pelley, 2019) and animals (for review see Lerner, 2020), we focus on naturalistic longitudinal studies as they provide useful insights into the amount of repetition or practice that is likely to be needed to form a habit within EBPTs. For these studies, the outcome measure tends to be the Self-Report Habit Index (SRHI) (Verplanken & Orbell, 2003). This measure starts with this stem: “Behavior X is something …”. This is followed by items that assess facets of habit formation such as “I do frequently,” “I do automatically” and “I do without having to consciously remember.” A higher score denotes greater habit strength. The respondent is asked to agree or disagree with each item. There are 12-item (Verplanken & Orbell, 2003) and 7-item (Lally, Van Jaarsveld, Potts, & Wardle, 2010) versions. In addition, the Self-Report Behavioral Automaticity Index (SRBAI) is a 4-item measure of automaticity (Gardner et al., 2012).

The handful of longitudinal studies conducted thus far, all using variants of the SRHI or SRBAI, highlight that: (1) repetition is an important part of the habit formation process, (2) the amount of repetition needed varies across types of habits and (3) there are individual differences in the time it takes to form a habit. Regarding the importance of repetition for habit formation, Kaushal, Rhodes, Spence and Meldrum (2017) allocated 94 new gym goers to a habit formation group versus a control group. By 8 weeks, the members of the habit formation group were 1.67 times more likely to engage in moderate-vigorous exercise measured with an accelerometer and self-report and they engaged in exercise with more consistency (i.e., more repetition), relative to the control group. This study is unique as it included a measure of the extent of repetition of gym attendance across the habit formation process. Finally, Van der Weiden, Benjamins, Gillebaart, Ybema and de Ridder (2020) recruited people who wanted to form a new habit in one of four domains: health, relationships, spending of money or engagement in an environmentally friendly practice. Over a period of 3 months, there was an increase in habit strength measured by the SHRI and the effect was strongest for those who consistently performed the behavior.

Regarding differences in the amount of repetition needed across habits and individuals, Kaushal and Rhodes (2015) studied the process of developing the habit of using a new gym membership over 12 weeks. A minimum of six weeks was required to form the habit. Fournier, d’Arripe‐Longueville and Radel (2017) monitored the process of building physical activity habits over 28 weeks. Habit formation, measured by the SRHI, reached an asymptote at around 19 weeks. Another study by Fournier et al. (2017) investigated the formation of the habit of performing a psoas iliac stretch daily. This stretch maintains flexibility and prevents low back pain. Habit formation occurred within 22 weeks. Lally, Van Jaarsveld, Potts and Wardle (2010) asked participants to choose a habit they would like to form, such as eating healthy foods, drinking healthy drinks or engaging in more physical activity. Habit formation varied from 18 days to 36 weeks.

There are likely multiple timing and other characteristics that influence the amount of repetition needed. For example, Fournier et al. (2017) reported that habits practiced in the morning were formed more quickly relative to habits practiced in the evening. This finding was mediated by higher morning cortisol levels which has been implicated in the development of habits. Also, the impact of stress and fatigue on repetition and habit formation is not yet clear, although it is clear that stress and fatigue have strong effects on habit performance (Neal, Wood, & Drolet, 2013; Schwabe & Wolf, 2013). As stress and fatigue may be a characteristic of those seeking an EBPT, the impact on habit formation is an important domain for future research. Another issue relevant to timing is the level of consistency of repetition that is required for habit formation. Lally et al. (2010) reported that missing one opportunity for repetition did not impact habit formation. However, Gardner and Lally (2018) have suggested that inconsistent performance may nonetheless hinder habit formation, because “failing to act reduces the likelihood of subsequent performance, derailing maintenance” (p. 215).

Application to EBPTs.

Within an EBPT there are typically multiple, multi-step habit formation and habit disruption targets and these targets are arguably more complex than the discrete habits typically studied. Hence, there is a need to study how the process of habit formation unfolds while simultaneously building and disrupting multiple complex habits. Relatedly, habits are often studied in relation to actions performed at least once per day. Yet, in EBPTs we are often seeking to build habits for less frequent and often intermittent behaviors. For example, people with insomnia may experience a poor night of sleep intermittently (e.g., 4 nights per week). Hence, in treatments designed to improve sleep, the skills that promote coping the day after a poor night of sleep—such as energy generating—are needed only intermittently. Will it therefore be harder for these daytime coping skills to become habits? Gardner and Lally (2018), citing the example of weekly recycling (Klöckner & Oppedal, 2011), suggest that habit formation requires consistent, but not necessarily frequent, action as long as other components of habit formation (discussed below) are in place.

EBPTs typically strongly encourage home practice of the topics covered within each treatment session. This often takes the form of practicing a new skill. An implicit goal of home practice is to develop new habits or disrupt unwanted habits, via repeated experiences. However, given the importance of repetition for habit formation and the time course for the development of habits, it seems unlikely that a week or two of home practice will actually result in the formation of a new habit or the disruption of an unwanted habit. This might, at least partially, account for the partial or full relapse often observed following the completion of an EBPT. To illustrate this point more fully, a goal of CBT-I is to develop a habit of engaging in a “wind-down” before bedtime during which the lights are dimmed and electronics are turned off at a regular time each night of the week. The process of collaboratively devising the individualized wind-down might take about half of a 50-minute treatment session. The setup includes providing a rationale and education about the environmental influences acting on the circadian system (e.g., light) and the benefits of establishing a consistent wind-down and bedtime routine. Then the wind-down is collaboratively devised with a focus on activities that the patient would enjoy and (hopefully) be intrinsically motivated to try. The homework for the coming week involves practicing the wind-down and a portion of the subsequent session, typically one week later, would be used to review the homework and problem solve any obstacles that were encountered when trying to implement the wind-down. In one scenario, the provider might assume that the wind-down habit has been developed after 1 week and the treatment would move on to other topics. However, given the studies reviewed above indicating that the amount of repetition required is between 18 days to 36 weeks for relatively discrete behaviors, a mere 7 days of practice seems entirely insufficient for developing a habit of engaging in a nightly wind-down. In an alternate scenario, the wind-down would become a “rolling intervention” that is discussed week-by-week. Eventually, probably guided by clinical intuition, the provider would assume that the habit has been developed and they would cease weekly monitoring of the wind-down. While this is a somewhat better approach, there is an opportunity to reduce reliance on clinical intuition by adding a measure of habit formation, such as the SRHI or the SRBAI, as part of progress monitoring. Moreover, if the habit has not been formed by the end of the course of treatment then collaboratively developing a plan to encourage repetition after the final session will be essential.

An essential point is that, in the context of EBPTs, we do not know the “dose” of repetition—in terms of the number of days/weeks—that is needed for the formation of new habits and the disruption of unwanted habits. The dose may be higher than indicated in the prior research, given that certain habits in EBPTs are not practiced every day. This will be a challenging research endeavor due to the number of EBPTs and that each EBPT tackles multiple complex habits.

One important aspect of repetition is to ensure the patient develops skills in knowing when and how to modify the habit. For example, as part of the “energy generating” intervention within some sleep treatments (Harvey & Buysse, 2017), a patient may plan to develop a habit of walking in a local park every day instead of napping. However, what if after three weeks the patient develops an injury from daily walking (e.g., plantar fasciitis)? Consider one scenario where the course of the treatment has finished, and the patient does not have the knowledge nor skills to modify the habit of walking daily so that the injury can heal. In this case, repetition would cease, and the process of habit formation would be adversely impacted. Hence, providers should prepare patients by providing a strong rationale, including how to wisely modify the plan, so that repetition and habit formation can continue.

RFCBT (Watkins, 2018) incorporates implementation intentions to promote repetition (Gollwitzer, 1999). Implementation intentions are simple and quick techniques that take advantage of mental imagery and pre-deciding how to implement one’s goals. This approach can be harnessed to build new habits by helping to move goals into action (Verhoeven & de Wit, 2018) and to promote repetition particularly for less frequent or intermittent behaviors. The general format of an implementation intention is that, after identifying a habit one would like to form, they make an “If-Then” plan structured as: “If/When I encounter this situation ________________ I intend to __________________________ at this time ____________ in this_____________ place.” The person is then asked to write down this commitment, visualize it as vividly as possible and then repeat this process a few times. It is thought that the mental representation established with this procedure becomes “highly activated and thus more easily accessible” (Gollwitzer, 1999; p. 495). In a meta-analysis of published findings from 94 articles, implementation intentions had a positive effect on goal attainment, of medium-to-large magnitude (Gollwitzer & Sheeran, 2006). Since this meta-analysis, there have been many demonstrations of the usefulness of implementation intentions, including with people who experience mental health problems (Toli, Webb, & Hardy, 2016). However, it is important to note that other studies have raised concerns, including that the longer-term impact of implementation intentions on habit formation is less clear (e.g., Turton, Bruidegom, Cardi, Hirsch, & Treasure, 2016). Also, implementation intentions to perform alternative behaviors in response to cues associated with unwanted habits may be more helpful for weak or not-yet-developed habits, relative to strong well-established habits (Webb, Sheeran, & Luszczynska, 2009).

EBPTs could also incorporate reminders to engage in repetition. Tobias (2009) highlights a critical role for memory and proposes the use of reminders, sent at a critical moment, to perform the behavior. Reminders have potential to scaffold the habit formation process. Reminders to engage in repetition could take many forms. They might be delivered via phone, text, email or perhaps post-it notes left in an obvious location. While reminders were helpful in prompting initial repetition and habit formation, their salience diminished over time, along with their effectiveness (Tobias, 2009). Indeed, Carden (2018) raised the possibility that reminders may be effective in the short term but may inhibit habit formation in the longer term because reminders engage deliberate decision making which may impair learning. Reminders delivered via text or mobile app may also promote app dependence, leading to a behavior that relies on the app rather than an appropriate contextual cue (Carden & Wood, 2018). Future research, in the context of EBPTs, is needed to determine the types of reminders that are effective and to establish the timeframe over which they are effective. Also, the schedule for the reminders or working out if there is a “critical moment” in which to deliver the reminder is another domain for future research.

4. Habits are Automatic.

Automaticity arises as a consequence of repeating a desired behavior in response to a stable contextual cue (Figure 1). It is present when a habit is performed with minimal effort or deliberation (Bargh, 1994; Bouton, Todd, Vurbic, & Winterbauer, 2011; Verplanken & Orbell, 2003; Walker, Thomas, & Verplanken, 2015; Wood & Rünger, 2016). While the centrality of automaticity to habit formation has attracted debate (Trafimow, 2018), habits are typically conceptualized as an “integration of sequences of responses that are automatically executed as a unit” (Wood & Rünger, 2016; p. 292). That is, within many definitions of a habit, performance without conscious oversight is a critical component. This aspect of habits confers multiple advantages because, if our day-to-day routine is engaged in automatically, we are free to devote our attention and energy to more critical aspects of our lives.

Automaticity is widely understood to be comprised of one or more of these four features (Bargh, 1994; Verplanken & Orbell, 2003): efficiency because few or no attentional resources are needed to engage in the habit, non-intentionality such that our goals may not override engagement in the habit, lack of conscious awareness such that we reach for the next piece of chocolate without thinking about it and initiation outside of volitional control.

We also need to consider that habits not only involve automatic responding upon the perception of contextual cues, they also involve automatic attentional biases towards contextual cues that trigger the habit (Carden & Wood, 2018). This, of course, makes it difficult to disrupt unwanted habits. However, as will become evident, it may be possible to inhibit the habitual response once a contextual cue has activated the habit (Wood & Neal, 2007; Wood & Rünger, 2016) or substitute the automatic habitual response with a new response.

Application to EBPTs.

Within the context of EBPTs, there is an interesting opportunity to study and apply the concept of automaticity. Indeed, the empirical evaluation of the process of developing automaticity for specific components of EBPTs is rarely studied. For example, negative automatic thought forms are commonly included as part of cognitive behavior therapy (Beck, 1979; Beck, 2005). They involve the patient learning to become aware of one’s negative automatic thoughts or “hot thoughts.” Then, guided by a series of questions, the patient learns to evaluate their thoughts. Examples of common negative automatic thoughts are “I am useless,” “I can’t cope” or “They don’t like me.” The evaluation questions include: “What’s the evidence for the thought? What’s the evidence against the thought” and “What effect does thinking this thought have on me?”. Anecdotally, with close to daily practice over several weeks, negative automatic thinking habits can be fundamentally changed; the patient finds themselves asking the evaluation questions automatically. This transition to automaticity is highly clinically relevant and has not, to the best of our knowledge, been subject to empirical investigation.

EBPTs cannot rely solely on attempts to disrupt poor health habits via psychoeducation and persuasive appeals (Marteau, Hollands, & Fletcher, 2012). Indeed, a meta-analysis by Webb and Sheeran (2006) reported that persuasion consistently yielded a medium to large effect size change on intention but only a small to medium effect size change on behavior change. For example, a common practice in CBT-I is to provide education to tackle unhelpful beliefs about sleep. However, greater change in these habitual beliefs might be possible by testing the unhelpful beliefs with one or a series of behavioral experiments (Ree & Harvey, 2004). Behavioral experiments are “planned experiential activities, based on experimentation or observation, which are undertaken by patients in or between … therapy sessions” (p. 8; Bennett-Levy, et al., 2004). As behavioral experiments are multi-sensory experiential exercises they are thought to be processed at a deeper level, relative to verbal methods such as education or Socratic questioning (Bennett-Levy et al., 2004) and this may be more effective in building automaticity in habit formation and reducing automaticity in habit disruption.

It will also be important to empirically establish if some habits are easier to automate than others. For example, are relatively discrete and well-defined behavioral habits (e.g., establishing a rise-up routine) easier to establish relative to more complex changes to thought (e.g., rumination) and attention (e.g., attention bias to threat) patterns?

Functional analysis and self-monitoring are likely to be useful to “unpack” the automatic process of engaging in an unwanted habit. This would involve identifying contextual cues that trigger unwanted habits and that increase awareness of when contextual cues are occurring. Increased awareness of cues that trigger habits is a critical first step to tackling them. Habit reversal includes explicit awareness training, during which patients are taught to identify early warning signals and the environments associated with unwanted habits (Azrin & Nunn, 1973). Interestingly, Ladouceur (1979) conducted a component analysis of habit reversal and found that awareness training alone is an effective intervention on its own.

A second path to addressing the automaticity of unwanted habits is to block the enactment of the habit (Gardner et al., 2019). Indeed, substituting a new response to the same cue as an unwanted habit is a common solution discussed in the literature (Hertel, 2004; Watkins, Owens, & Cook, 2018; Wood & Neal, 2007). Moreover, ‘behavior substitution’ is included in the Behavior Change Taxonomy, a classification system developed to provide a common vocabulary for evidence-based behavior change techniques (Michie, Hyder, Walia, & West, 2011; Michie, Johnston, Abraham, Francis, & Eccles, 2013). Relatedly, Gardner and Lally (2018) note that unhelpful habits provide a strong cue that may even promote learning of a desired new substitute habit. However, there is reason to be cautious. Patey, Hurt, Grimshaw and Francis (2018) note that studies that use behavior substitution typically do not offer a theoretical rationale for the substitution strategy nor for the behaviors selected as substitutes. Hence, going forward we will need to improve our conceptualization of behavioral substitutions to address why a specific substitution was selected and how it will disrupt the unwanted habit.

Within EBPTs, RFCBT includes a thoughtful implementation of behavior substitution. The alternative behaviors developed as substitutes for rumination include thinking in a more concrete and specific manner (Watkins, Baeyens, & Read, 2009; Watkins et al., 2012) and guided imagery to recreate states of mind that are inconsistent with rumination (e.g., “flow” states). Great care is taken to select a substitute behavior that has already been used by the patient and is likely to be helpful and reinforcing. In other words, the substitute behavior is chosen based on its likelihood of successfully becoming instantiated as a new habit. After selecting a substitute behavior, RFCBT includes multiple practice sessions whereby the substitute behavior is enacted in response to cues that usually trigger rumination in order to strengthen the more helpful response.

Habit reversal (Azrin and Nunn, 1973), developed for tics and other nervous habits, also involves behavior substitution. Specifically, ‘competing response practice’ involves learning a response that is incompatible with the habit. Then, the competing response is enacted for an extended period of time following each occurrence of the unwanted habit. For example, following each occurrence of a tic, patients practice engaging in the competing response for 3 minutes (e.g., strengthening opposing muscles).

Anshel et al.’s values-based approach may be helpful in guiding patients to select substitute behaviors (Anshel, Brinthaupt, & Kang, 2010; Anshel & Kang, 2007). This involves collaboratively determining how the patient’s values compare with the costs and long-term consequences of the unwanted habits and clearly articulating the disconnect. Next, patients are guided to conceptualize and select substitute behaviors that are more consistent with their values.

Implementation intentions can also be used to help consumers of EBPTs to break unwanted habits. Specifically, “counter habitual implementation intentions” involve substitution. In one study, participants engaged in implementation intentions in which they replaced one snack or drink with an alternative healthier snack or drink (Adriaanse, Gollwitzer, De Ridder, De Wit, & Kroese, 2011). The structure of this implementation intention was: “If I am at home/in a bar and I want a snack/drink then I will take [alternative].” The proposed mechanism is that a new response is built on encountering the cue to the unwanted habit and this creates an association that directly competes with the unwanted habit (Adriaanse & Verhoeven, 2018). Interestingly, “negation implementation intentions,” which take the form “if [situation, e.g., feeling bored], then don’t [habit, e.g., eat chocolate],” appear to result in cognitive and behavioral rebound effects (Adriaanse, van Oosten, de Ridder, de Wit, & Evers, 2011). However, “ignoring the critical cue” in the form of “if [situation, e.g., feeling bored while working], then ignore [habit, e.g., the urge to snack]” does appear to help reduce the performance of the unwanted habit (Adriaanse & Verhoeven, 2018).

In the context of this section on automaticity, when new preferred habits are substituted to replace non-preferred habits, there is evidence that the non-preferred habits continue to be readily accessible (Bouton et al., 2011; Walker et al., 2015). In EBPTs, it will be important to determine the extent to which unwanted and preferred new habits compete and to delineate the conditions under which preferred habits “win.” Obviously, providing knowledge and skills to patients to promote stronger automatic connections to the preferred habits will be key (Brewin, 2006).

The measurement of automaticity, particularly in the context of EBPTs, requires further research. Recall that the SRBAI is comprised of four automaticity response items to the stem “Behavior X is something ….” The four items are: “… I do automatically,” “… I do without having to consciously remember,” “… I do without thinking,” “… I start doing before I realize I’m doing it.” There is a need to adapt or further develop the SRBAI for EBPTs for several reasons. First, EBPTs typically involve working on the formation of multiple new habits. As progress on the formation of each habit may be uneven, there is a need to administer the SRBAI for each specific habit (e.g., one for the completion of a negative automatic thoughts form each day and another for participating in a wind-down routine each evening). This would place excessive burden on patients. Second, the SRBAI is offered as a measure of automaticity. However, the extent to which it measures the formation of automaticity as well as the disruption of automaticity—both of which are common goals of EBPTs—is not clear. Third, although the extent to which an action is automatic is not dichotomous, but rather falls across a continuum, empirically defining a cut-point on the SRBAI may be helpful as it would establish a bench-mark (Kaushal & Rhodes, 2015; Lally et al., 2010). Relatedly, for the types of habits targeted in EBPTs, it is unrealistic to achieve a perfect score for an item like “I start doing [the desired habit] before I realize I’m doing it.” For example, consider the process of building the habit of getting out of bed if you can’t sleep for 15–20 minutes, which is part of stimulus control for insomnia. A patient is unlikely to ever get to the point where they start doing this before they realize they are doing it. Instead, they are likely to think “I’m awake” and then think “I guess it’s been 15–20 minutes, I should get up” and then “I don’t want to get up” and then “but I should get up.” More realistically, a cue may come to automatically trigger a decision-making process that, in turn, may result in engagement in the desired action (Maddux, 1997). These realities need to be taken into account for future measures of automaticity to be used within EBPTs. Fourth, the SRHI is worded to index the formation of habits relating to behavior. For use within EBPTs, a measure is needed that also encompasses the formation of cognitive (e.g., attentional bias to threat, worry) and emotional (e.g., use of new emotion regulation skills) habits. Hence, wording changes to incorporate this broader range of goals are needed.

5. Reinforcers Promote Habits.

Thorndike’s ‘Law of Effect’ (1927) states that if a behavior produces a reinforcing outcome, that behavior will strengthen. Since Thorndike’s initial conceptualization, a long and rich tradition of animal and human research shows that reinforcers have a profound impact on the frequency and the longevity of behavior (Ferster & Skinner, 1957; Lerner, 2020). Thus, careful analysis and strategic use of reinforcers will facilitate habit formation by reinforcing the repetition of the new habit in a stable context (Figure 1).

A critical element to consider is the schedule for delivering reinforcers. Continuous reinforcement schedules occur when a reinforcer follows a specific behavior after every instance of the behavior. In contrast, partial reinforcement schedules occur when a reinforcer follows a specific behavior after only some instances of the behavior. A particularly robust finding is that behaviors are more resilient to extinction under partial reinforcement schedules relative to continuous reinforcement schedules. This is known as the partial-reinforcement extinction effect (Capaldi, 1966; Mowrer, 1956). There are also several types of partial reinforcement schedules that show differential effectiveness. Interval reinforcement schedules, in which reinforcers follow specific behaviors only after a certain time interval has elapsed, are more resistant to extinction relative to ratio reinforcement schedules, in which a behavior results in a certain probability of a reinforcer (Dickinson, Nicholas, & Adams, 1983; Yin & Knowlton, 2006). A combination of reinforcement schedules may be most desirable. For example, starting with continuous reinforcement rapidly establishes a causal relationship between behaviors and reinforcers, then switching to partial reinforcement promotes resistance to extinction (Boyagian & Nation, 1981; Nation, Cooney, & Gartrell, 1979; Nation & Massad, 1978; Nation & Woods, 1980).

Of course, once a habit if formed it is relatively automatically triggered by the specific context. By this point, reinforcement is not so relevant.

Application to EBPTs.

In translating knowledge of reinforcers to EBPTs, providers need to carefully consider how to use reinforcement for habits patient’s wish to disrupt and for habits patient’s wish to build.

For habits patient’s wish to disrupt, functional analysis will be helpful for discovering how the unwanted habitual behavior is being reinforced. For example, the habit of problem substance-use possibly arose initially for the ‘high’ or hedonic value. However, the patient may present for treatment because the ‘high’ has been replaced with tolerance and unpleasant withdrawal. A contrasting example is problematic technology use (e.g., gaming) late into the night. Connection with friends and the thrill of competition are reinforcers and contribute to the difficulty disrupting this habit. In this case, providers could work with their patients to try to obtain these reinforcers in nongaming parts of their life. Also, a harm reduction approach (Marlatt, 1996) may be helpful in which there is a shift in focus away from reducing gaming itself and toward reducing the adverse consequences of gaming. For instance, gaming at a different time of day and establishing the habit of turning off technology at a set time each day would increase the opportunity for getting sufficient sleep during the nighttime hours.

For habits patient’s wish to build, there are several considerations. Specific reinforcers may be helpful at the beginning of the habit formation process to assist patients in starting new behaviors. Example include: praise from the therapist; between session ‘cheer leading’ by the therapist via phone calls, texts or email; reinforcers from a parent or caregiver (e.g., sticker chart for children); reinforcement that a patient can give to themselves; and reinforcers that are delivered by significant others (e.g., celebration dinner cooked by a friend). Also, providers can optimize the habit formation process by adopting partial instead of continuous reinforcement schedules (Capaldi, 1966; Mowrer, 1956). Moreover, we should consider how reinforcers may be delivered following the completion of a course of treatment. Reinforcers could continue to support habit formation after treatment has ended if the patient can self-deliver them. Thoughtfully spaced booster sessions with the therapist also have great potential. Additionally, the appropriateness of a reinforcer may change over time according to the individual’s subjective evaluation of the reinforcer and their preferences. For example, five minutes of ‘freeze dancing’ maybe a highly valued reinforcer for a 4-year learning to stay in their own bed all night initially. However, after 3 or 4 dance episodes, it will probably become clear that a new reinforcer is needed.

Finally, Marteau et al. (2012) highlight strategies that increase positive associations with the actions we wish to build into habits. This can be applied to EBPTs. People with sleep problems could explicitly build new associations with getting into bed at the beginning of the night and waking-up in the morning. For many people, a negative association has been established between sleep onset with floods of worry and rumination. Building new habits whereby we re-associate the head hitting the pillow with gratitude practice (Wood, Joseph, Lloyd, & Atkins, 2009) or with savoring (McMakin, Siegle, & Shirk, 2011) has potential, over time, to automatically evoke positive sensations and calm that will facilitate sleep.

6. Habits Take Time to Develop.

As highlighted earlier, there are a range of laboratory-based tasks that have revealed important aspects of the habit formation process. However, as “real-life” habits take time to develop, it is difficult to measure and experimentally manipulate them in the laboratory. The best insight into the length of time it takes to develop a habit comes from the handful of naturalistic longitudinal studies that we reviewed earlier. Together, these suggest that the timeframe for habit formation varies enormously depending on the individual and the type of habit to be formed. For the specific and discrete habits that have been studied in health psychology, the range was 18 days through to 36 weeks and the pattern for habit formation is typically asymptotic.

Application to EBPTs.

It is alarming to consider that most EBPTs fall into the range of 6 to 16 sessions conducted over 6 to 16 weeks. This would be sufficient time if the goal of the treatment is to build one discrete habit. However, EBPTs typically seek to build and disrupt a number of complex habits. Given the wealth of evidence suggesting that habits take time to develop and change, it seems highly unlikely that this is sufficient time for habits to form. If under-funded health systems are not likely to provide more sessions for EBPTs, we need to consider how to maximize habit formation in the sessions that are available. Shorter carefully spaced booster sessions would be one possibility.

More generally, in CBT we often prepare patients for the time it will take to replace an old habit with a new habit by saying “you’ve had many years practice of thinking this way so developing these new skills will take practice and time.”

EBPTs likely intervene on complicated networks of behaviors and multiple targets for habits to disrupt and habits to form. Hence, it may be possible to “speed up” progress if we develop methodology to assess and conceptualize these networks to guide treatment planning. For example, modeling approaches (Lally et al., 2010) might be adapted to study bi-directional relationships between an individual’s habits and their symptomatology in order to identify which habits have the most negative influence on an individual’s functioning. This personalized assessment of habits and related symptoms could then be targeted in a personalized version of the appropriate EBPT (Fisher, 2015). Also, there will be a need to determine the optimal number of habits to intervene upon and how to best select them considering issues like the habits with the greatest health impact or individual preference (Spring, Moller, & Coons, 2012).

As already discussed, another pathway that may help therapists effectively help patients with the multiple habits that are typically tackled within EBPTs may be to develop “bundles” of desired behaviors that can be chunked together and activated as one unit (Spring et al., 2012; Wood & Rünger, 2016). In a similar vein, Rothman et al. (2015) discusses that piggy-backing a new habit onto an existing habit can be helpful (recall the teeth flossing example).

Of course, the ever-important flipside is that habits patient’s wish to disrupt will likely to also take time, although this issue has been minimally studied.

Concluding Thoughts and Recommendations.

Following the receipt of an EBPT in a routine practice setting, approximately 1/2 of patients recover and 2/3rds show worthwhile benefits (Clark, 2018). This is highly impressive; yet there is room for improvement. The central empirical question posed in this paper is: Would outcomes following the receipt of an EBPT improve if the basic science of habit formation were fully leveraged? The six elements of habits distilled are a first set of principles that could be infused into EBPTs. However, an enormous amount has yet to be discovered about how these processes apply to EBPTs. This pursuit will clearly be fascinating and require immense creativity in research design given that patients typically pursue an EBPT because they wish to get help with multiple habit formation and habit disruption targets.

It is noteworthy that there are at least three discernable patterns of the outcome from long-term follow-ups of EBPTs: a wash out of treatment effects, such that improvements at post intervention are not maintained at long-term follow-up (Halldorsdottir & Ollendick, 2016; Rose, Hawes, & Hunt, 2014); an upward spiral effect, such that improvements at post intervention become more pronounced at long-term follow-up (Compas et al., 2009; Ginsburg, 2009); and an occasional delayed treatment effect, such that improvements are only evident at follow-up and not immediately post intervention (Bell, Marcus, & Goodlad, 2013; Carroll et al., 1994). As the most common pattern of findings is the first—at least some intervention effects are lost by follow-up—another central empirical question is: Would outcomes from EBPTs show more durability if habit formation principles were rigorously intertwined with treatment-as-usual?

Table 1 offers a range of ideas for providers of EBPTs to incorporate the science of habit formation. Given that the current dose of EBPTs (50 minutes once per week) is likely not optimal to promote the formation of multiple new habits and disruption of unwanted habits, providers of EBPTs must ensure that the patient has the knowledge and skills on habit formation to continue the process once the course of treatment has finished. Of course, this yields another area of investigation: adherence to the habit formation process once provider oversight is removed. Additionally, we need to consider how and when to introduce new treatment content. Providers of EBPTs often present new material in each session, set practice of the new material for homework, briefly check on the homework in the subsequent session, and then move on to new content. This is very far from harnessing the science of habit formation. Instead, we envisage that an intensive and sustained focus on individual habits will be the fundamental building blocks to more successful EBPTs. Optimal support for habit formation is likely to include: a longer time span over which the sessions are delivered, the flexibility to deliver reinforcement between sessions (e.g., through phone or other electronic contact with the patient), and booster sessions after the main course of treatment is complete.

Finally, there are a plethora of questions and problems about habit formation that await probing by current and future generations of psychological scientists from a broad range of subfields. Some of these questions are articulated here and many more are yet to be defined, operationalized and tested with the creativity, precision and rigor that characterizes our field. The challenge ahead is to identify the science questions and research designs that directly illuminate the habits that are tackled within EBPTs.