# Lesson 10: High-Risk Populations, Culture, and Health Disparities *Companion-podcast transcript • Sarah & Kiffer* *Office Hours episode to listen to after working through the lesson* --- **Sarah:** In 2016, the Transportation Safety Board of Canada reported that since 2006, roughly seventy percent of fishing deaths in British Columbia had resulted from not wearing a flotation device. **Kiffer:** And since June 2019, WorkSafeBC has required every crew member on the deck of a fishing vessel to wear one. **Sarah:** Transport Canada has said that changing the practice requires a shift in safety culture. **Kiffer:** Which is a good reason to ask what wearing a vest, or leaving it off, means to the people on deck. **Sarah:** Welcome back to Office Hours. I'm Sarah. **Kiffer:** And I'm Kiffer. We're on Lesson ten, High-Risk Populations, Culture, and Health Disparities. Question one: output theories assume the health behaviour is what a person is paying attention to, and with high-risk populations it often isn't, so let's test them on a made-up case. **Sarah:** Go ahead. I'll try the mapping. **Kiffer:** Picture a deckhand, call him Travis, who's twenty-six and works on a small commercial fishing boat on the coast of British Columbia. He's invented. A friend of his on another boat went overboard two years ago and drowned. Travis knows cold water can kill within minutes. He agrees that a flotation device would keep him afloat. He wore one during a safety course and knows how to use it. And the crew had a safety talk at the start of the season. **Sarah:** So far he sounds like someone who'd wear one. What else do we know about him? **Kiffer:** Three more details. The vest on board is bulky and catches on the gear. Travis is paid a share of the catch, like the rest of the crew. And the skipper, who has fished for thirty years, never wears one. Travis doesn't wear one either. **Sarah:** Okay. On the Health Belief Model, the friend who drowned gives him high perceived susceptibility, because it happened to someone like him. Knowing that cold water kills within minutes is perceived severity, and that's high too. Agreeing that a vest would keep him afloat is a perceived benefit. The bulky vest catching on gear is a perceived barrier, a cost of action. The safety talk is a cue to action. And knowing how to use one from the course supports his self-efficacy. **Kiffer:** And where does the skipper go in your mapping? **Sarah:** He doesn't fit neatly into those constructs. In the Theory of Planned Behavior, he shapes the subjective norm, because what the most experienced person aboard does tells Travis what's expected. **Kiffer:** Good. So here's the question for everyone. Most of the Health Belief Model's constructs point toward wearing the vest, and Travis still leaves it off. If leaving it off isn't about his health, what might it be about? Take a few seconds. *(Pause)* **Sarah:** I'd look at the work. He's paid a share of the catch, so anything that slows the work costs the whole crew money. A deckhand who keeps up is one the skipper hires again next season. And wearing a vest that nobody else wears could mark him as new or nervous. So leaving it off is about keeping his place on the crew and earning a living. **Kiffer:** That's how I'd read it too. The pay was in the details all along, and the Health Belief Model could only file it as one more cost in a decision about health. Real data show how common leaving it off is. In 2016 the Transportation Safety Board of Canada reported that since 2006, roughly seventy percent of fishing deaths in British Columbia had resulted from not wearing a flotation device. **Sarah:** And the rules have moved since then. WorkSafeBC has required every crew member on the deck of a fishing vessel to wear one since June 2019, and Transport Canada has said that changing the practice requires a shift in safety culture. So does that make the Health Belief Model useless for someone like Travis? **Kiffer:** I wouldn't go that far. The constructs can hold this content once you fill them in. The perceived barrier becomes "a vest slows me down and costs the crew money," and in the Theory of Planned Behavior the subjective norm comes from the skipper. What the Health Belief Model doesn't do is prompt you to ask what the behaviour is for, so it points you toward more information about drowning, which Travis already has. I'd keep the constructs as a set of questions and ask the generative question first: what need does this behaviour meet? **Sarah:** And then the program looks for ways to stay safe that still meet that need. A slim inflatable vest that doesn't slow the work. The skipper wearing one, which is modelling in the Social Cognitive Theory sense. A crew culture in which wearing one is part of being a reliable deckhand. And the regulation, which works at the policy level of the ecological model. Notice what a label would have done here, too. If a safety report had called Travis non-compliant, that would have become the whole story about him, the way "high-risk population" can become a person's primary identity. **Kiffer:** Question two. This one often comes up when harm reduction is in the news. Doesn't harm reduction mean giving up on treatment? **Sarah:** It's one of the most common objections, and students who can recite the answer aren't always convinced by it. What's the logic, and what's the evidence? **Kiffer:** The logic is a sequence. You deal with the most pressing threat first, and the trust that builds makes the harder conversations possible later. Researchers in Vancouver have looked at both steps. A 2011 study in The Lancet compared fatal overdoses before and after Insite opened in 2003. Within five hundred metres of the site, the overdose death rate fell by thirty-five percent. In the rest of the city it fell by about nine percent. **Sarah:** That's keeping people alive. What about treatment? **Kiffer:** The same research group looked at that. In a cohort of about a thousand people who used Insite, use of detox rose by about thirty percent in the year after it opened, compared with the year before, and people who went through detox were more likely to start methadone or other addiction treatment afterward. An earlier report found that people who used Insite more regularly, and people who had contact with its addictions counsellor, started detox sooner. Since 2007 there has also been a program called Onsite, directly above Insite, with stabilization and recovery beds. **Sarah:** So here's a question for everyone. People who used Insite more regularly started detox sooner. Does that show that using Insite caused people to start treatment? Take a few seconds. *(Pause)* **Kiffer:** On its own, no. That finding is an association from an observational study. People who came often might differ from people who came rarely, for example in how settled their lives were or how ready they already were to stop. Or someone already thinking about detox might have started visiting more often to see the counsellor. And a rise in detox use from one year to the next could reflect other changes that year. What the evidence does do is undercut the claim that harm reduction keeps people away from treatment, and it fits the idea that trust precedes movement toward change. **Sarah:** I'd add a critique. The sequence only works if the next step exists. If someone builds trust at a supervised consumption site and decides they want treatment, and the wait for a treatment bed is several weeks, that trust may lead nowhere. Harm reduction can end up funded as the first stop on the continuum with nothing behind it. **Kiffer:** I think that's the fair version of the critique, and it's aimed at funding decisions more than at harm reduction itself. I'd judge a service that keeps people alive on its own terms, because a person who dies can't move toward anything. The claim that harm reduction leads to treatment should be judged by whether the later stops on the continuum exist. **Sarah:** Question three. A generative schema is described as generative of behaviour but not necessarily predictive. Students find that hard to accept, and so do I, a little. If a schema can't tell you what someone will do, what can a program do with it? **Kiffer:** In an output theory, prediction means that given these inputs, you expect this behaviour. A schema works differently. It tells you what a whole range of behaviours is for, and a person can perform it in different ways on different days. What a program needs to know is what any alternative has to satisfy, whatever form the behaviour takes. That can still be tested: do people take up the alternative, and does the risky behaviour fall among the people who do? **Sarah:** The weak point is that someone has to identify the schema in the first place. **Kiffer:** Yes, and that takes the kind of fieldwork that doesn't fit easily into a three-to-five-year funding cycle. An outsider can also get it wrong. Let's try one. Here's a made-up man, call him Ray, who injects drugs and lives in a mid-sized city. Outreach workers know him as the most generous person in his circle. When he has money, he buys food for everyone. When a friend is sick, he's the one who checks on them. When he has drugs, he shares, and when a friend has no sterile equipment, he shares his own. **Sarah:** So the health risk is the shared equipment. **Kiffer:** Right. Here's the question for everyone. Using the generative approach, what would you say Ray's sharing is about, and what could a program offer him? Take a few seconds. *(Pause)* **Sarah:** Okay. I'd start with the risk and protective factors model. His friends use drugs and share equipment, so that's a peer-domain risk factor, and his attachment to them is strong. So the program should weaken that attachment, maybe by connecting him with a group of people who don't use drugs. **Kiffer:** That's how the exposure model would read it, and the peer risk factor is real. Now compare what you did with the generative move from the lesson. What's the first step, and did you take it? **Sarah:** Name the need. And I skipped it. I treated his friendships as an exposure to reduce, when his generosity is what his life is organized around. The schema might be something like, I'm the one people can count on. The need is to be valued and to have a role. A program that pulls him away from his friends takes that role away, so he'd probably refuse it, and the outreach workers could lose his trust. **Kiffer:** So what could satisfy that same need with less risk to Ray and his friends? **Sarah:** Make him the person who brings sterile equipment and naloxone to his circle. He keeps the role, and his generosity now spreads safer supplies. That's also where harm reduction meets the generative approach. Ray sits at a hub in his network, so supplies move through him to people a health unit might never reach, and he becomes a model in the Social Cognitive Theory sense. **Kiffer:** Programs along these lines exist. A 2024 study looked at programs in two Massachusetts cities that paid people who use drugs to hand out naloxone within their own networks, and found that the cash payments supported their dignity and recognized that work. **Sarah:** What I'd take from my mistake is that the same detail can be read as an exposure or as a schema, and the reading decides the program. **Kiffer:** I'd add one caution. Ray's schema is a guess we built from four details. In a real program you'd check it with him and the people around him and plan the role with them, which is the willingness to learn and the shared benefits from the seven working issues. A schema guessed from outside and never checked can turn into the kind of trait list that the critique of cultural competence warns about. **Sarah:** Question four. When one community has worse health than another, how can you tell whether culture has anything to do with it? **Kiffer:** Culture is often the first explanation people reach for, especially when a community looks different from the people running the program. The causes of disparities run across every ecological level, though, and socioeconomic status is a key factor among them. This is the harder worked example of the episode, and the numbers are made up. A health region reports that last year, eighty-eight of the one thousand children in a newcomer community had at least one emergency visit for asthma, or 8.8 percent. In the rest of the region, it was five hundred twenty of ten thousand children, or 5.2 percent. **Sarah:** Let me guess the plan. A translated pamphlet about using controller inhalers every day, because someone has assumed the families don't trust daily medicine. **Kiffer:** That's the plan in this made-up region. Before accepting it, the region splits the children by housing. In the newcomer community, six hundred of the one thousand children live in older rental buildings with damp and mould, and seventy-two of them had an emergency visit. The other four hundred live in other housing, and sixteen of them had a visit. In the rest of the region, fifteen hundred children live in damp housing, and one hundred eighty of them had a visit. The other eighty-five hundred live in other housing, and three hundred forty had a visit. **Sarah:** So here's the question for everyone. For the newcomer children, that's seventy-two of six hundred in damp housing and sixteen of four hundred in other housing. For the rest of the region, it's one hundred eighty of fifteen hundred in damp housing, and three hundred forty of eighty-five hundred in other housing. Work out each rate. Within the same kind of housing, do children in the newcomer community do worse? Take a few seconds, and pause the audio if you'd like to write the numbers down. *(Pause)* **Kiffer:** Let's take damp housing first. Seventy-two out of six hundred is twelve percent, and one hundred eighty out of fifteen hundred is also twelve percent. **Sarah:** And in other housing, sixteen out of four hundred is four percent, and three hundred forty out of eighty-five hundred is four percent as well. So within each kind of housing, the two groups have exactly the same rate, and the whole gap is accounted for by where the children live. Sixty percent of the newcomer children are in damp housing, compared with fifteen percent of the children in the rest of the region. The pamphlet is aimed at a cause that doesn't show up in these data at all. **Kiffer:** That's the first trap, and you've caught it. The second trap is to conclude that the community's history and circumstances have nothing to do with the gap. Ask why sixty percent of these children live in damp buildings. In a real region, you'd want to know whether these families arrived with little savings, whether their credentials were recognized, and whether landlords would rent to them. Housing segregation by race and ethnicity is also linked to health risks regardless of income. That's the trajectory of health, showing up in where families can find a home. **Sarah:** So the gap is an inequity, because it's unfair and avoidable. Its immediate cause sits in the physical environment and in neighbourhood living conditions on the sociocultural framework. Picking that battle would mean housing inspections and mould repairs, and maybe a health broker who helps families report problems to a landlord, since fear of losing their home would be a perceived barrier. **Kiffer:** That combines an individual approach, lowering a perceived barrier, with community mobilization and advocacy to remove the environmental risk in the housing itself. Real data rarely split as cleanly as our made-up numbers. Often some gap remains within each kind of housing, and that's where you'd look for differential treatment, ethnomedical views, or something nobody has measured. You can only see any of this if the data are collected finely enough. British Columbia's Anti-Racism Data Act, which became law in 2022, is meant to make that kind of data collection possible. **Sarah:** Okay. Now the part where we disagree. How far can you adapt a program for a culture before it stops being the program that was evaluated? **Kiffer:** Let me start with why fidelity matters. An evidence-based program was tested in a trial. If you change its core components, the parts that actually produce the effect, you can't borrow the trial's evidence any more. And adaptation can drift. Sessions get dropped, the harder content gets cut, and it's all described as cultural fit. In 2004, Felipe Castro and colleagues described this as a tension between fidelity and fit. **Sarah:** I take the point about drift, but I think the bigger risk runs the other way. Programs are usually adapted for populations they were never tested in, so the evidence never covered those communities in the first place. A 2006 review that pooled seventy-six studies of culturally adapted mental health interventions found that programs aimed at one specific cultural group had effects about four times as large as programs for mixed groups, and programs delivered in the clients' own language, where that wasn't English, had effects about twice as large as programs delivered in English. **Kiffer:** That's a strong finding, and I'd read it carefully. Those comparisons were made across studies, so the single-group programs may have differed in other ways too. And the review covered mental health interventions, which may not carry over to every kind of health promotion program. It's still good evidence that fit matters. What it can't tell you is which changes keep a program's effect and which ones lose it. **Sarah:** Then there's the case where a community doesn't want an adapted program at all. A culturally grounded program starts from the community's own knowledge and governance. And when the system itself carries a history, fidelity to a program built elsewhere can repeat that history. British Columbia's In Plain Sight review reported in 2020 that eighty-four percent of the Indigenous people who responded to its survey had experienced some form of discrimination in health care. A program delivered faithfully inside that system may still not feel safe to the people it's meant for. **Kiffer:** I agree that a grounded program should be judged as its own program, with outcomes the community helps set. Where I'd hold my ground is on honesty about the evidence. If a team changes the deep structure, it should say so and evaluate what it built, and it shouldn't cite the original trial as proof that the new version works. **Sarah:** I can live with that, and I think we agree on a practical rule. Before changing anything, name the program's core components and how they're supposed to work. Adapt the surface structure freely. Change the deep structure on purpose and with the community. And if the core changes, treat the result as a new program, evaluate it, and report what was changed. **Kiffer:** Let's pull it together with three things to take away from this episode. **Sarah:** First, high-risk behaviour is usually about something besides health, so before you apply an output theory, ask what the behaviour is for. Travis was protecting his place on the crew, and Ray was being the person his friends count on. **Kiffer:** Second, harm reduction is a sequence. The Vancouver evidence supports the first step, keeping people alive, and it fits the idea that trust precedes change, although an association with later treatment can't prove cause. The sequence only works when the later steps exist. **Sarah:** And third, before you explain a disparity by culture, look at how each population is spread across conditions like housing and income, and then ask how that spread came about. That second question is where the trajectory of health comes in. **Kiffer:** If you'd like more practice, rework the asthma example from this episode with your own numbers. Try a version where a gap remains within each kind of housing, and decide which explanations from the lesson could account for what's left over. **Sarah:** Next time, it's Lesson eleven, Evaluation and Careers in Health Promotion, where we take up how you'd know whether any of these programs worked, and where this theory gets put to use. **Kiffer:** Take care, everyone. **Sarah:** See you in Lesson eleven.