If you're trying to make a change in areas of your life like diet or exercise, you might come across the idea of "stages of change" that everyone passes through in trying to change their behavior. The stages of change were one of three original components in psychologist James Prochaska's (1983) transtheoretical model of psychotherapy (TTM), although many people have heard about the stages without ever knowing the TTM. The stages of change construct has by now been widely accepted in fields that are increasingly far from psychology, including medical case management, nursing, medicine, and healthcare business administration. Although the stages of change are among the most widely known and cited explanations of health behavior, this success arrived at exactly the same time that the TTM was being questioned and increasingly disputed in the scientific literature.
An interesting case study of the rise and fall of the stages of change can be seen in the connection between the TTM and motivational interviewing. At one time these two ideas were so closely entwined that many people talked about them as though they were the same thing. People used to talk about the stages as a theory underlying motivational interviewing, although the creators of motivational interviewing always maintained that their approach was developed atheoretically. They did, however, include chapters about the stages of change and the TTM in their books and talked about the applicability of motivational interviewing to patients in the earliest stage of change, called precontemplation. But in the latest version of the standard textbook on motivational interviewing, there is no mention of the stages of change.
In light of the scientific ambiguity, should we continue to teach this model, or should we abandon it as one more failed idea in science's theory graveyard? If it's not a good model, why do clinicians still find it so useful, and workshop presenters still cite it so readily? What, exactly, is the matter with the stages of change?
Weaknesses of the Stages of Change
The heyday of the TTM was in the late 1990s and early 2000s, when the stages model gained wide traction among clinicians as well as health behavior researchers. Over time fewer new studies have been published using its concepts, although the model is still widely cited by clinicians. Problems identified include difficulties specifying the stage construct itself, problems measuring it reliably, and a lack of strong evidence for benefit when matching interventions to people's stage of change.
Problems with the Stage Sequence. In the original TTM there were four stages of change: precontemplation, contemplation, action, and maintenance. Other stages have been added or subtracted at different times including preparation (between contemplation and action, as a kind of "late contemplation" where people were thinking about how do change instead of whether to change), decision (in the same place between contemplation and action, but with an "early action" feel instead of a "late contemplation" one), and relapse (a return to precontemplation, which might occur at any stage of the model but is usually shown in the diagram after maintenance and before precontemplation). The stages have been useful for categorizing people across many different health behaviors, although the percentage of people who are found to be in each stage varies widely depending on whether the behavior in question is using seatbelts (mostly maintenance) or losing weight (mostly precontemplation). A more serious concern is that people don't go through the stages in a particular sequence as the model would suggest, with some people jumping straight from precontemplation to action with no cognitive step in the middle, others progressing from precontemplation to contemplation but then going right back again, and people relapsing from any of the stages along the way. Sometimes people also assume that each stage is of fixed duration (e.g., 3 months in each), although this was never Prochaska's argument. Still, a clear temporal sequence was initially proposed as one of the key strengths of the stage model, and without that the stages start to look like discrete categories of people rather than a process through which an individual person progresses.
Measurement Problems. One other persistent difficulty in interpreting the descriptive literature on stages of change is the fact that investigators have used many different measures to operationalize the stage of change construct. Stage-of-change measurement began with a long-form survey instrument called the URICA (University of Rhode Island Change Assessment) that produced multidimensional results (one scale for each stage) and did not cleanly classify people into groups. Initially, researchers used cluster analysis to overcome this problem and create subgroups of patients, but this method is not feasible for routine clinical practice because it depends on looking at data from a large sample of people rather than an individual and the weighting of individual items varies from one sample to another. Later studies used a simple staging algorithm that had just 4 questions, and based people's stage classification on whether they were thinking about and/or doing something about a change in their behavior. A problem with this method is its poor psychometric validation, and the tendency for people who don't want to do anything different to report that they have "already made a change," making it quite hard to differentiate precontemplation from maintenance at the far end of the scale. Finally, researchers have challenged the very idea of discrete stages, and found that a simple continuum from low to high motivation or readiness might predict behavior better than the 4-stage sequence does.
Lack of Evidence for Matching Effects. Ultimately, the power of the stages model should be derived from practitioners' ability to use it to prescribe one type of intervention versus another to help their patients. Although some early reviews cited the evidence in favor of motivational interviewing as evidence for the stages of change, the fact that an intervention works with precontemplators doesn't necessarily mean that the idea of precontemplation itself is valid. Prochaska's first attempt at matching interventions to stages used 10 processes of change that were derived from different competing schools of psychotherapy. More generally, Prochaska suggested that cognitive interventions would be more useful with precontemplators while behavioral ones would be more useful in the action or maintenance stages. And eventually he identified "strong and weak principles of change" associated with the different stages, a concept that I often emphasize in training: We start doing something different because it is important to us regardless of how difficult, but if we are to maintain the new behavior over time it needs to be made easy. The problem with stage-matching approaches is that they don't actually seem to make a large difference in tailored interventions. Across trials, stage-matched interventions generally fail to outperform non-matched ones. Additionally, some of the early stages-of-change research didn't use actual behavior change as the outcome variable, but only "stage progression," in which a change from precontemplation to contemplation, for example, was seen as progress. This has generally been viewed as insufficient when researchers are looking for interventions that actually change behavior. Finally, the idea of matching treatments to stages of change is vulnerable to a problem known as "pseudotailoring," in which people do better simply when a message is presented as having been tailored "just for them." Different people can be given the exact same message with that meta-description of tailoring attached to it, and they respond as though the message were really customized when in fact it was not. To overcome this problem, Weinstein et al. suggested that research designs need to systematically mismatch interventions to the stages they are not supposed to help, and compare that mismatch to the situations in which people's stage and their selected intervention are supposed to align. This approach allows researchers to examine the effect of the specific matching approach, rather than just whether any type of matching is better than a one-size-fits-all kind of treatment.
Strengths of the Stages Model
Despite these fairly severe research-based critiques, the stages of change remain one of the most widely known and popular models of health behavior change. Here are some of the reasons that they are still so widely used in clinical practice even though researchers have largely moved on:
Widespread Applicability. The stages of change are designed to measure a client's "readiness" or "willingness" to make a change in behavior. The specific behavior to be changed could be anything, which is a strength of the model. Psychology as a discipline aims to be both a healthcare profession and a science of human behavior, and the stages of change are an excellent example of translation from science into practice. The TTM's original empirical base, in smoking cessation, had health-promotion aspects as well as addiction components, and the model spread from there into many other aspects of addiction treatment, health promotion, and chronic disease management. In fact, the TTM's success in health promotion areas probably facilitated the widespread acceptance of motivational interviewing, which is now well-known on its own but at the time was only being described as a treatment approach for alcohol use disorders. At this point the stages of change and MI are both widely disseminated in training programs for many different health professions.
Positive Potential for Change. Prochaska's TTM makes a crucial distinction between a counseling approach's theory of change -- how it helps people -- and its theory of problem causation -- i.e., its explanation for why people have problems in the first place. Many other theories of health behavior change are deficiency-based, emphasizing problem causation and offering just one way to fix the problem. Other theories also tend to be static, viewing a person as unchanging over time in the absence of interventions. The TTM, by contrast, describes change as a dynamic process with a time or transition element. Other theories lack this complexity and therefore seem less able to capture people's rich experiences when trying to change behavior over a period of months or years. An interesting early study of Prochaska's showed that people passed through the same stages whether they were working with a professional counselor or trying to make changes on their own, which supports the stages as a naturalistic description of the normal process people go through in changing their behavior. The stages are fundamentally a hopeful model, focused on people's own strengths, the possibility of improvement over time, and even the possibility of recovery after a relapse.
Predictive Power of Stages. Perhaps most importantly, a person's stage of change at the start of treatment does seem to predict whether they will eventually be successful at the end of treatment. Initial stage of change can also be used to predict important treatment process variables like the therapeutic alliance between a client and a counselor. In particular, people in the precontemplation stage (assessed as "not yet ready for change") have a harder time developing good working relationships in counseling and tend to have worse outcomes overall. This finding could still be true even if the stages identify certain groups of patients rather than patients at a particular point in the process, even though that would contradict the stages' hopeful aspect: Maybe at the simplest level, the stages are just a screening tool that identifies the "hard to help" patient, who is a very familiar character in clinicians' experiences. This is borne out by clinicians' common misuse of the term "precontemplator" for a patient who isn't willing to cooperate, or in some groups I have trained the idea that a very resistant patient "isn't even in precontemplation yet"! The idea of precontemplation just means "not yet ready" and implies that readiness could later arrive. Clinicians' use of the term as a pejorative implies that the precontemplator is intransigent and difficult to help.
Intervention Options for Non-Motivated Patients. The psychotherapy literature as a whole provides many different approaches that can be used effectively to help motivated patients, including a whole toolbox of well-supported methods under the "cognitive-behavioral" label, as well as evidence for psychodynamic treatments or mindfulness interventions. But for non-motivated patients the options are much slimmer, and many intuitively appealing approaches such as inducing fear of negative consequences are actually counterproductive. People who are less motivated are less likely to seek out new information about their health and less receptive when it is presented, so education works less well for them as a behavior-change strategy. They are, in short, a frustrating group for health care providers who want to help, and the stages of change's appeal may simply be that of any port in a storm. The idea of differential treatment for non-motivated patients also fits well within a medical model, where practitioners assess a patient's needs and then prescribe different interventions according to the results. The fact that matching works better than one-size-fits-all care also supports practitioners' inclination to use the stages algorithm, even if the specific matchup of stages with interventions isn't well supported. As described above the benefit could be a result of pseudotailoring, but it's a benefit nonetheless.
Future Ideas for Practice and Research
Based on the ongoing clinical popularity of the stages of change, and the partial strengths identified in some areas even though the model is weak in others, it seems that we are not yet ready as a field to abandon this theory.
Clinically we can suggest using the stages to identify people who need a different approach than education and persuasion, and as a reminder to focus on people's strengths and potential for making changes in their behavior over time. Simply prompting clinicians to avoid heavy-duty education and problem-solving advice with patients who are not ready for that might be the most important single contribution that the stages of change can make to practice. And even if specific matching isn't possible, we can use the stages to remind us to provide person-centered care that is adapted to the needs of a particular patient at a particular point in time.
From both a clinical and a research perspective, we should remember that motivational interviewing is different from the stage model and can be used independently of it. For researchers, we need more consideration of readiness for change as a continuum instead of a set of stages, or perhaps more studies that directly compare different conceptualizations of readiness for change. Further tests of stage-matched versus stage-mismatched interventions would provide better control for pseudotailoring effects, but tests of stage-matched versus non-matched interventions may be sufficient to meet clinicians' needs when they are simply trying to find any way to help their patients more effectively.
Comments
Post a Comment