# Lesson 3: Individual Health Behavior Theories *Companion-podcast transcript • Sarah & Kiffer* *Office Hours episode to listen to after working through the lesson* --- **Sarah:** In 2005, the addiction researcher Robert West published an editorial calling for the Transtheoretical Model to be put to rest. **Kiffer:** And Carlo DiClemente, one of the model's authors, replied in the same journal that the obituary was premature. **Sarah:** So the people who study it still argue about whether the stages are real. And Kiffer and I argue about it too. **Kiffer:** We do. I'm doubtful about the stages, at least in the strong sense, and you'd defend them as more than labels. **Sarah:** Welcome back to Office Hours. I'm Sarah. **Kiffer:** And I'm Kiffer. We're on Lesson three, Individual Health Behavior Theories, and we'll start with a made-up person who's due for colon screening. Sarah, can you give us the real background first? **Sarah:** Sure. In British Columbia, people aged fifty to seventy-four with no symptoms are advised to screen for colon cancer every two years with a take-home stool test called a FIT. A health care provider gives you an order form, you pick up the free kit at a participating lab, you collect the sample at home, and you bring it back to a lab within seven days. It can't be mailed back. So who's our made-up person? **Kiffer:** His name is Gord. He's fifty-six, he runs a hardware store in a small Interior town, and he feels perfectly healthy. His doctor gave him the order form last month. He's heard that colon cancer found early is usually treatable. He finds the idea of collecting a stool sample unpleasant. The only lab in town closes at four, and he's at the store until six. He isn't sure he'd collect the sample properly. And last week a friend from his curling club told him he'd been diagnosed with colon cancer. **Sarah:** That's six details after the order form. Let me try sorting them. **Kiffer:** Before you do, here's the question for everyone listening. Which construct of the Health Belief Model does each of those six details belong to? Take a few seconds. *(Pause)* **Sarah:** Feeling perfectly healthy is low perceived susceptibility. He doesn't think colon cancer applies to him. Hearing that it's treatable when it's found early is a perceived benefit of screening. Finding the sample unpleasant is a perceived barrier. The lab hours and his doubt about collecting the sample both go under self-efficacy, since both are about whether he feels able to do it. And his friend's diagnosis is a cue to action. **Kiffer:** Most of that is right. Check the lab hours with one question. Is that detail about Gord himself, or about the action and the world around it? Would anyone who works his hours in that town face the same problem? **Sarah:** Anyone who works until six in that town would face it. So it's a fact about the world, and it belongs under perceived barriers. His doubt about collecting the sample properly is about him, so that one stays under self-efficacy. **Kiffer:** That's the check I'd use every time, and it matters because each construct points to a different fix. A short how-to video could help Gord's confidence and would do nothing about the lab hours. Longer hours or an after-hours drop-off would solve the timing and leave his doubt where it was. **Sarah:** And the friend's diagnosis? I said cue to action. **Kiffer:** That's right. It's an event that gives him a push. If it makes Gord think that someone his age who felt fine can still get colon cancer, it can also raise his perceived susceptibility. One event can act on two constructs. When you build a program around something like that, say which one you're counting on, so you know what to measure. **Sarah:** Now take the same two details into the Theory of Planned Behavior. Where do the lab hours go? **Kiffer:** Into perceived behavioural control. His belief that the lab hours make the test hard to do is a constraining control belief, and because it's the only lab in town, that belief carries a lot of perceived power. His doubt about collecting the sample also lowers his sense of control. So one construct in this theory holds two details that the Health Belief Model keeps apart. **Sarah:** Which is the critique that perceived behavioural control is hard to tell apart from self-efficacy. **Kiffer:** Yes, and the Health Belief Model keeps them apart on purpose. When Irwin Rosenstock and his colleagues added self-efficacy in 1988, they proposed it as a separate construct alongside susceptibility, severity, benefits and barriers. My advice, whichever theory you use, is to keep asking that one question. Is the obstacle in the person's confidence, or in the action and their circumstances? The answer usually tells you whether to change a message or change a service. **Sarah:** Question two. If a belief predicts a behaviour, will changing the belief change the behaviour? Here's the reasoning I hear a lot. Carpenter's meta-analysis found that perceived benefits and barriers predicted behaviour, so a campaign that raises perceived benefits should raise screening. **Kiffer:** It's a sensible first thought, and it skips a step. Carpenter pooled eighteen studies that measured people's beliefs first and their behaviour later, and benefits and barriers were the most consistent predictors. Measuring the belief first establishes the order in time, and showing cause takes more evidence than that. People who see big benefits in screening may also be people with a regular doctor, flexible hours, or a family that talks openly about health. **Sarah:** So the real test is to change the belief on purpose, in an experiment, and see whether the behaviour follows. **Kiffer:** Right, and that's been done for intentions. In 2006, Thomas Webb and Paschal Sheeran pooled forty-seven experiments that randomly assigned people to an intervention designed to change their intentions. On average the interventions produced a medium-to-large change in intention and a small-to-medium change in behaviour. On the standard effect size scale, that's about 0.66 for intention and 0.36 for behaviour. **Sarah:** So changing intention does change behaviour, by a bit more than half as much on that scale. That fits the Theory of Planned Behavior's track record, too. Armitage and Conner pooled one hundred eighty-five studies of the theory, and its constructs together accounted for about twenty-seven percent of the variation in behaviour. When behaviour was measured objectively, the figure was about twenty-one percent, and with self-reports it was about thirty-one. **Kiffer:** Which matters for the high school physical activity study from the lesson, since its behaviour gains were self-reported. Self-reports still carry information, and plenty of good studies rely on them. Even so, I'd give a program that measures behaviour only by asking people a second look before anyone scales it up. **Sarah:** Can we put numbers on the gap between intending and doing? **Kiffer:** Let's do it with a made-up campaign about colon screening in a made-up health region. Before the campaign, thirty percent of eligible adults say they intend to do the test in the next three months. Suppose sixty percent of those people actually do it, and ten percent of everyone else does it anyway, maybe because a doctor hands them the form at a visit. What share gets screened? **Sarah:** Sixty percent of thirty percent is eighteen percent. Ten percent of the other seventy percent is seven percent. So twenty-five percent get screened. **Kiffer:** Now the campaign works, and afterward fifty percent say they intend to screen. If those two follow-through rates stay the same, what share gets screened now? Take a few seconds. *(Pause)* **Sarah:** Sixty percent of fifty is thirty, and ten percent of the other fifty is five. So thirty-five percent. Intention went up twenty points, and screening went up ten. **Kiffer:** That's the arithmetic. Here's the trap in it. We assumed the people the campaign just persuaded will follow through as often as the people who intended all along. Is that likely? **Sarah:** Probably not. The people who already intended had their own reasons and maybe a plan. The new intenders were persuaded by a campaign, and their intention may be weaker or may fade before they get to the lab. That's the critique about the time between intention and action. **Kiffer:** So suppose the twenty points of new intenders follow through at thirty-five percent. That's eighteen percent from the original intenders, seven from the new ones, and five from everyone else, for thirty percent. Intention rose twenty points, and screening rose five. **Sarah:** So if the evaluation had stopped at intention, it would have reported a twenty-point gain when screening rose by five. **Kiffer:** And it shows where the effort might go. Every step between intending and doing is a place where people drop out. For Gord, that's the order form, the trip to the lab, the sample, and the trip back within seven days. In my view, making those steps easier is often a better use of effort than another round of persuasion, and that's perceived behavioural control at work. **Sarah:** Question three. How do you use a stage model and a point-in-time theory on the same problem? This is the harder example, and it starts with something real. During British Columbia's heat dome at the end of June 2021, the B.C. Coroners Service later identified six hundred nineteen heat-related deaths. Sixty-seven percent of the people who died were seventy or older, fifty-six percent lived alone, and ninety-eight percent died indoors, mostly in homes without adequate cooling. **Kiffer:** So the precaution that matters is having a way to keep cool at home. Now the made-up part. A health authority surveys four hundred people aged seventy or older who live alone in an Interior town. The precaution is having a way to keep at least one room of their home cool during extreme heat. Because it's a precaution, the survey sorts people into the stages of the Precaution Adoption Process Model. **Sarah:** What does it find? **Kiffer:** Five percent have never heard that extreme heat is a danger to health, so they're unaware. Forty-five percent know heat waves can kill but have never thought of it as something to do with their own home, so they're unengaged. Fifteen percent are undecided. Twenty percent have decided not to act. Ten percent have decided to act and haven't done it yet. And the last five percent already have a cool room. **Sarah:** And the plan? **Kiffer:** A made-up planner drafts one brochure about the dangers of extreme heat and mails it to all four hundred. Remember the radon experiment: risk information moved the undecided toward deciding, and an easy how-to moved the people who had already decided. Which groups is a risk brochure matched to? Take a few seconds. *(Pause)* **Sarah:** The undecided, and the unaware, since they need the basic information first. That's fifteen percent plus five percent, so twenty percent, or eighty of the four hundred people. **Kiffer:** Right. The ten percent who've decided to act need the how-to: where to get an air conditioner, how to set up one cool room, and who will check on them during a heat warning. The biggest group is the unengaged. They already know heat can kill. What's missing for them? **Sarah:** This is where the second theory comes in. The stage tells me where they are, and the Health Belief Model tells me what's missing. They already believe the danger is severe. What they lack is perceived susceptibility. They think heat waves are dangerous for someone else. So the message has to be about people like them: older, living alone, at home indoors, which is where most of the deaths in 2021 happened. **Kiffer:** And the Theory of Planned Behavior adds a question about who delivers that message. For someone who lives alone, the opinion that counts might belong to an adult child, a family doctor, or a home support worker. That's the reference group question, and the survey can ask it directly. **Sarah:** That leaves the twenty percent who decided not to act. **Kiffer:** Here's the trap. The planner's notes say that group should get a stronger warning, since they've heard about the risk and said no. What would you want to know before agreeing? **Sarah:** Why they said no. The stage label tells me they decided, and it tells me nothing about the reasons. **Kiffer:** In our made-up survey, three-quarters of that group rated the danger to themselves as high. Their reasons were cost, a building rule about window units, or having no one to install a unit for them. **Sarah:** Then their risk beliefs are already in place. Their susceptibility and severity are high. What's stopping them is a barrier, in Health Belief Model terms, or a constraining control belief with a lot of perceived power, in Theory of Planned Behavior terms. And the belief is accurate. A stronger warning would tell worried people about a danger they already see and can't fix. **Kiffer:** That's sixty people, three-quarters of the eighty in that group, whose obstacle no brochure can remove. For them, the program has to change their circumstances. **Sarah:** And British Columbia has done a version of that. In 2023 the province gave BC Hydro ten million dollars to install about eight thousand free portable air conditioners over three years, for people with low incomes who are medically vulnerable to heat, and home health care workers could refer clients. So the single brochure was well matched to eighty people, and for sixty others the real obstacle was something only a change in their circumstances could fix. **Kiffer:** And that's what makes this one harder than sorting people into stages. Two groups, the unengaged and the people who decided not to act, both look like people who haven't acted. The stage model separated them, the point-in-time theories told us what each one lacks, and one group turned out to need a change in the world around them. That's the transtheoretical idea at work, a different theory for each stage. To finish the plan, you'd set a measure of movement for each group, such as the share with a cool room before the next heat warning. **Sarah:** Which brings us to the part where we disagree. Are the stages real? **Kiffer:** I'm doubtful, at least in the strong sense. My view is that readiness is better described by a few things that vary smoothly, like how much someone wants to change and how soon, and that the stages are cut points we draw on those. The measurement critique from the lesson points that way. If the line between contemplation and preparation has to be set by an arbitrary rule, there may be no natural line there. **Sarah:** And that's still argued in the research, isn't it? **Kiffer:** It is. In 2005 the addiction researcher Robert West published an editorial calling for the Transtheoretical Model to be put to rest, and Carlo DiClemente, one of the model's authors, replied in the same journal that the obituary was premature. **Sarah:** I'd defend the stages as more than labels. The radon experiment matched and mismatched messages to stages for almost nineteen hundred homeowners, and the results came out the way the model predicted. The risk message worked better for the undecided, and the how-to worked better for those who'd decided. If readiness were one smooth scale, it's harder to explain why a different message works best at different points. **Kiffer:** That's the strongest evidence for stages, and I take it seriously. Here's something that pushes the other way. If smokers really move from contemplation to preparation to action in order, what would you expect to find about quit attempts that start with no planning at all? Take a few seconds. *(Pause)* **Sarah:** They should be rare, and they should fail more often, since the person skipped preparation. **Kiffer:** That's the prediction. In 2006, Robert West and Taj Sohal reported a national survey of almost two thousand smokers and ex-smokers. Nearly half said their most recent quit attempt began the moment they decided to quit, with no planning. And the unplanned attempts were more likely to last at least six months. For attempts made between six months and five years earlier, the odds of success were about 2.6 times higher. **Sarah:** That's hard on the strict sequence, though it rests on what people remembered about attempts made months or years earlier, so I'd hold it a little loosely. I still think it leaves the practical idea standing. People differ in readiness, and a program that treats everyone as ready to act will waste a lot of effort and lose a lot of people. **Kiffer:** I agree with that part. The 2010 Cochrane review of stage-based programs for quitting smoking is a fair summary of the evidence. It found forty-one trials with more than thirty-three thousand participants. Staged programs helped people quit about as much as comparable programs without staging, and the four trials that compared the same program with and without staging found no clear advantage for the staging. The reviewers also wrote that the evidence does not support limiting quitting advice to smokers judged to be in the preparation or action stages. **Sarah:** So where do you land? **Kiffer:** I'd use stages as a planning shorthand for readiness, and for any behaviour I'd keep checking whether matching messages to stages actually adds anything. **Sarah:** And I'd keep them as more than shorthand, because the radon results suggest the barriers really do change from stage to stage. We agree on the practical rule, though. Use the stage to choose the message, and never use it to decide who gets offered help. Offer the easy action to everyone, whatever stage they're in. **Kiffer:** Let's pull it together with three things to take away. **Sarah:** First, when you sort a detail into a construct, ask whether it's about the person's confidence in their own ability or about the action and their circumstances. That one check separates self-efficacy from barriers, and it usually tells you whether to change a message or change a service. **Kiffer:** Second, a belief that predicts behaviour is a candidate for change, and experiments are what show whether changing it works. Where you can, prefer behaviour measured objectively, and expect behaviour to move less than intention does. **Sarah:** And third, a stage tells you where people are, and a point-in-time theory tells you what they're missing. In the heat example, two groups who hadn't acted needed different things, and one of them needed a change in their circumstances. **Kiffer:** If you'd like more practice, try the heat problem again with a different precaution, such as keeping indoor air clean during wildfire smoke. Make up a survey, sort it into the stages, and for each group ask which construct is missing and whether a message can supply it. **Sarah:** Next time, it's Lesson four, Social, Cultural, and Environmental Theories I, which moves the theory outside the person, beginning with Albert Bandura's Social Cognitive Theory, the source of the self-efficacy construct we kept running into here. **Kiffer:** Take care, everyone. **Sarah:** See you in Lesson four.