Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Health Beliefs, Perceived Risk, and Preventive Decisions

Abstract. Health beliefs, perceived risk, and preventive decisions constitute an interconnected research field investigating how people interpret vulnerability to illness, evaluate protective alternatives, and translate appraisals into preventive action. This monograph reconstructs the historical formation of the Health Belief Model and adjacent theories, analyzes deliberative, affective, and experiential risk representations, and critically examines empirical evidence from vaccination, screening, infectious-disease prevention, and repeated health behavior. It integrates historical theory, experimental risk communication, prospective prediction, measurement research, randomized intervention studies, and major meta-analyses, while distinguishing observed associations from causal claims, decision quality from mere uptake, and individual representations from social and institutional constraints. Principal findings of the synthesis concern the distinct roles of susceptibility, severity, expected benefits, barriers, efficacy, norms, numerical understanding, framing, and planning; the importance of reciprocity between past behavior and current risk perception; and the limits of communication-only interventions where genuine opportunity is restricted. The analysis identifies open problems in longitudinal measurement, cross-cultural validity, mechanistic identification, equitable intervention design, and artificial decision support. It is a critical narrative research synthesis, not an independently conducted systematic review or meta-analysis.

Keywords: health beliefs; perceived risk; preventive decisions; Health Belief Model; perceived susceptibility; perceived severity; self-efficacy; protection motivation; theory of planned behavior; precaution adoption; vaccination psychology; screening decisions; risk communication; affective risk perception; numerical literacy; health behavior; fear appeals; decision aids; intention–behavior gap; health equity; artificial decision support.

Scientific Scope and Research Questions

Health beliefs are representations through which people interpret bodily vulnerability, the seriousness of potential harm, the likely consequences of action, and the practical conditions under which prevention appears worthwhile. Perceived risk is a family of judgments and experiences concerning the possibility and significance of an adverse health outcome. Preventive decisions are choices about taking, delaying, refusing, or seeking further information about an action intended to reduce future harm. The three constructs form a research domain because an individual may recognize a hazard, interpret it through prior experience and social information, anticipate the benefits and burdens of a response, and select an action under incomplete information. Their relationships can be observed, measured, experimentally manipulated, and compared across populations and health contexts.

The central question of this monograph is how health-related beliefs become preventive decisions, and under what conditions that relationship changes. Subsidiary questions concern the formation and calibration of perceived susceptibility; the independent contribution of perceived severity; the evaluation of expected benefits, costs, and barriers; the roles of self-efficacy, affect, norms, and trust; the consequences of presenting risks as probabilities, frequencies, or narratives; and the distance between reported intention and realized action. Vaccination, screening, infectious-disease protection, and sustained prevention provide complementary applications because they differ in timing, reversibility, certainty of benefit, institutional mediation, and the visibility of outcomes.

The analysis is a critical integrative research monograph based on original theoretical writings, observational and experimental investigations, randomized interventions, and relevant quantitative syntheses. It is a narrative scholarly synthesis: it makes no claim to an exhaustive preregistered systematic search or a newly computed meta-analysis. Each class of evidence answers a different kind of question. Cross-sectional surveys can describe belief structures; longitudinal studies can assess temporal prediction; experiments can test the consequences of manipulating a message or representation; randomized trials can estimate intervention effects in specified populations; and syntheses can assess patterns and heterogeneity across studies. The strongest inference about a causal mechanism arises when the proposed mediator is measured, manipulated where feasible, and connected prospectively to behavior.

The scientific boundary is the appraisal-to-decision system. Detailed organization of screening services, clinical diagnostic pathways, medical contraindications, and disease-specific treatment are independent subjects. This boundary allows prevention to be studied as an explanatory problem without substituting the psychology of a decision for the biological efficacy of the preventive intervention. Psychological acceptability and medical effectiveness are distinct variables requiring different evidence. The same logic applies to choices mediated by digital applications or artificial systems: a system's capacity to present a risk estimate is empirically distinguishable from its accuracy, the user's understanding, and the quality of the resulting choice.

Historical Formation of Health Belief Research

Modern health-belief research emerged from attempts to explain why people did not consistently use available preventive services even when those services were thought beneficial. Mid-twentieth-century public-health psychology increasingly treated health behavior as a problem of subjective expectation rather than a direct reaction to epidemiological probability. In his historical reconstruction, Rosenstock, I. M (1974) located the Health Belief Model within this preventive-health tradition. The companion theoretical exposition, Rosenstock, I. M (1974), specified a scheme for understanding why an individual might regard an action as relevant, worthwhile, and feasible. These works helped establish a research vocabulary that could be translated into survey questions and compared with actual behavior.

The theory was subsequently developed and reviewed across a broad literature. Janz, N. K., & Becker, M. H (1984) surveyed the model's applications over a decade and treated the predictive contribution of distinct components as an empirical question. An especially important historical change was the incorporation of perceived self-efficacy into explanations of behavior that requires repeated performance, effort, or skill. A person may accept a threat and value a preventive measure while doubting their capability to undertake it. The model thereby moved from elementary evaluations of threat and utility toward a more complex account of perceived agency and practical constraints.

Parallel traditions transformed the field. Rogers, R. W (1975) analyzed the persuasive effects of fear appeals by asking how threat information motivates protection. Maddux and Rogers (1983)90023-9) developed the role of self-efficacy within that framework. The theory of reasoned action and Ajzen's theory of planned behavior (1991)90020-T) placed intention, normative expectations, and perceived behavioral control at the center of prospective explanation. Weinstein, N. D (1988) foregrounded the possibility that individuals occupy qualitatively different decision states, including unawareness, undecidedness, decisions against action, and decisions to act.

These traditions were shaped by further developments in cognitive psychology and behavioral decision research. Experiments on framing established that equivalent outcome descriptions can yield systematically different preferences, as shown by Tversky, A., & Kahneman, D (1981). Slovic, P (1987) emphasized the psychological dimensions that organize risk judgments beyond statistical probability. The field increasingly asked how feelings, memory, social trust, media representation, numerical understanding, and institutional arrangements change what people perceive as dangerous and what they consider possible. Historical development is therefore best understood as a widening set of testable explanatory programs rather than the progressive replacement of one universal model by another.

The Health Belief Model as a Structure of Preventive Appraisal

The Health Belief Model identifies several interrelated judgments relevant to preventive action: perceived susceptibility, perceived severity, perceived benefits of a proposed action, and perceived barriers to taking that action. Cues to action and self-efficacy feature prominently in extended presentations. These constructs provide a disciplined language for separating the belief that a problem could occur, the belief that its consequences matter, the belief that a specific response would help, and the belief that the response could be executed in real circumstances. The distinctions are analytically valuable because people can exhibit any combination of these orientations.

A person might believe a disease is severe while judging their own exposure negligible; another may perceive substantial susceptibility yet distrust an intervention's benefit. High concern can coexist with low action when financial cost, inconvenient access, embarrassment, anticipated discomfort, or competing obligations dominate the decision. Conversely, a modest threat appraisal may accompany preventive action when the behavior is easy, routinized, socially supported, or institutionally offered as a default. The model is most informative when each variable is measured independently with reference to a concrete action and a defined health outcome.

The model does not specify a single universally validated mathematical equation for combining these variables. Different investigations have used additive scales, multiplicative formulations, interaction terms, and context-specific measures. This matters when interpreting evidence: agreement that barriers predict behavior does not establish that barriers have the same effect across vaccination, cancer screening, oral health, and chronic-disease self-management. Carpenter, C. J (2010) examined predictive relationships across studies and reinforced the need to distinguish the model's named components rather than treating “health beliefs” as one undifferentiated score.

Construct overlap is a recurring concern. A questionnaire item such as “I am likely to be harmed unless I take the recommended action” combines susceptibility, conditional beliefs about intervention benefits, and perhaps an intention-related demand characteristic. Measurement precision requires separating the probability of the untreated outcome from the expected effect of prevention. The model also tends to emphasize individual cognition, which must be embedded in an account of clinicians, institutions, social ties, access, and material resources when interpreting observed health behavior. A conceptual prediction is not a population-wide law until its measurement and boundary conditions have been established.

Perceived Susceptibility and Perceived Severity

Perceived susceptibility is a person's judgment that a specified health outcome could occur to them within an understood time frame. The judgment may be expressed as an estimated probability, a verbal likelihood, a comparison with other people, or a more immediate experiential feeling of vulnerability. These operationalizations are not interchangeable. Someone can assign a low numerical probability while nevertheless feeling particularly vulnerable when symptoms, an illness in the family, or a vivid news story makes the possibility mentally available. A study must state whether it is measuring absolute expectation, comparative expectation, perceived exposure, or affective vulnerability.

Perceived severity concerns the anticipated significance of an outcome. It may include expected pain, disability, mortality, disruption of family life, economic loss, social stigma, and future uncertainty. Medical severity categories and subjective severity may correlate, but the individual's valuation is shaped by social position, previous illness, caregiving obligations, and the anticipated ability to cope. For example, a short period of illness may carry disproportionate consequences for someone with precarious employment or responsibility for dependent family members. The meaningful unit for explanation is often the expected burden under the person's real circumstances.

Research frequently combines susceptibility and severity under the heading of perceived threat. That shorthand is useful if studies also test whether the components behave differently. Brewer, N. T., Chapman, G. B., Gibbons, F. X., Gerrard, M (2007) found associations between risk perceptions and vaccination, with differences across the risk-perception measures studied. These correlations support the predictive relevance of risk beliefs in vaccination contexts, while leaving open questions about causality, confounding, reciprocal effects, and the practical size of the influence. The decision to vaccinate can itself change a participant's subsequent perception of risk, which complicates cross-sectional estimates.

An adequate account of preventive susceptibility also differentiates baseline risk from residual risk following an intervention. Individuals who have already adopted protection may correctly report lower current risk because their preventive behavior changed exposure. If researchers interpret their lower risk perception as evidence that concern never influenced protection, the time ordering has been reversed. Longitudinal measurement and carefully worded counterfactual questions—such as perceived risk in the absence of a specific precaution—can improve interpretation, although hypothetical judgments bring their own measurement problems.

Perceived Benefits, Barriers, and Cues to Action

Perceived benefits concern the expected efficacy and personal value of a proposed preventive response. The belief that an intervention works in a trial population, the belief that it will help in one's own circumstances, and the belief that its benefits justify the required effort are related but separable judgments. Preventive decisions often involve delayed, probabilistic benefits whose success is invisible: when prevention works, the feared event does not occur, and the individual may receive little direct feedback. This creates an attribution problem that makes recommendations, social evidence, and confidence in the source especially important.

Perceived barriers include anticipated harms, opportunity costs, access obstacles, administrative complexity, transportation, language, financial constraints, time, fear of procedures, and uncertainty about eligibility. Some barriers are represented cognitively before action; others arise only during execution. A survey measuring anticipated barriers may therefore predict intention without adequately capturing the actual obstacles encountered later. The importance of this category was already apparent in Janz, N. K., & Becker, M. H (1984), and subsequent work has treated benefits and barriers as central components of preventive appraisal.

Cues to action are events or signals that bring a preventive possibility into consideration: a clinician's recommendation, reminder, official notice, workplace policy, symptoms, a public-health campaign, or a family member's illness. Their effects depend on whether the signal is trusted, understood, timely, and linked to an available response. A reminder about screening may function as a retrieval cue for someone who has already decided to participate; for an undecided person, the same message may become new evidence requiring interpretation. Consequently, intervention design should distinguish awareness, persuasion, scheduling, and execution support instead of assuming that every message acts on the same psychological stage.

The benefit–barrier formulation also invites a crucial empirical comparison: changing beliefs versus changing the environment in which action occurs. Eliminating fees or simplifying appointment systems may increase uptake without producing a large measured change in perceived severity. Conversely, persuasive information may alter reported risk while leaving behavior unchanged when a service remains unavailable. The practical force of the Health Belief Model lies partly in making these possibilities visible and measurable, rather than treating failure to act as an unexplained deficiency of motivation.

Self-Efficacy, Response Efficacy, and Perceived Control

Self-efficacy concerns a person's belief that they can successfully organize and carry out a particular course of action. Response efficacy concerns the belief that the action itself will reduce a specified risk. Perceived behavioral control concerns the experienced feasibility of performing behavior under internal and external conditions. These judgments overlap in ordinary language but identify distinct hypotheses. A person may trust a vaccine's protective value yet believe they cannot reach a clinic, or feel capable of arranging an appointment while doubting the medical benefit of the proposed intervention.

The distinction became prominent in the revised protection motivation framework of Maddux and Rogers (1983)90023-9). In behavioral decision research, Ajzen (1991)90020-T) connected perceived control with intention and the possibility of direct behavioral effects when control judgments reflect actual opportunities and constraints. Neither self-reported confidence nor perceived control automatically measures objective capability. Studies seeking a causal account must assess whether the skill, time, money, service availability, and assistance required for action are genuinely present.

Confidence is often specific to the action, context, and difficulty level. A person may feel able to obtain reliable health information and still lack confidence in discussing it with a clinician; another may be confident about a single appointment but struggle with repeated adherence under shifting work schedules. Measures should name the action and relevant obstacles rather than substitute a global confidence rating. Repeated preventive decisions additionally require coping with setbacks, plan disruption, and motivational changes, which makes the stability of self-efficacy across time an empirical question.

The concept has special importance for interventions using fear. Threat messages can raise concern while reducing perceived control if the proposed response seems inaccessible or the danger unavoidable. An experimentally credible explanation should therefore test threat appraisal and coping appraisal as separate pathways. For practical prevention, behavioral capability, response efficacy, and actual opportunities often need to be developed together. Their alignment is a plausible mechanism, but it must be tested in the population and decision setting under study.

Protection Motivation Theory and the Psychology of Threat

Protection Motivation Theory proposes that decisions about protective responses are influenced by evaluations of a threat and evaluations of coping. Threat appraisal includes beliefs about likelihood and severity and, in some formulations, rewards associated with continuing a risky behavior. Coping appraisal includes the effectiveness of the recommended response, perceived ability to perform it, and expected response costs. Rogers, R. W (1975) offered an early formulation in the context of fear appeals, while Maddux and Rogers (1983)90023-9) expanded the role of perceived self-efficacy.

The theory helps explain why increasing the salience of disease does not reliably increase protective behavior. Information about severity can motivate action when a credible response is available and perceived as effective. It can also produce avoidance, fatalism, or defensive processing when the person feels unable to act. This is an interaction hypothesis, not a promise that a certain emotional intensity will produce a predictable outcome. Risk communication studies must examine message content, the recommended response, prior exposure, individual vulnerability, and context as possible moderators.

Witte, K., & Allen, M (2000) synthesized research on fear appeals and argued for attention to efficacy alongside threat. A later quantitative synthesis by Tannenbaum, M. B., Hepler, J., Zimmerman, R. S., Saul, L., J (2015) found that fear appeals can influence attitudes, intentions, and behaviors, with meaningful variation across conditions. The evidence supports examining persuasive mechanisms and boundary conditions rather than assuming that any frightening message is either inherently effective or inherently counterproductive. The ethical evaluation of a message additionally depends on accuracy, respect, and whether people can realistically take the recommended action.

The framework is particularly useful when contrasted with ordinary preventive decisions that occur without acute fear. Routine vaccination, screening invitations, and recurring protective habits are affected by reminders, trust, norms, availability, and accumulated learning. Threat appraisal may remain part of the background, but the causal pathway can shift to service design or habit. A successful integrative account must accommodate both the immediate response to danger and the ordinary organizational conditions that make prevention sustainable.

Reasoned Action, Planned Behavior, and Intention Formation

The theory of planned behavior locates behavioral intention within a system of attitudes toward an action, perceived social norms, and perceived behavioral control. Ajzen's formulation (1991)90020-T) offers an important complement to the Health Belief Model because it explicitly represents the influence of perceived expectations from others and the individual's appraisal of control. A preventive action can be evaluated favorably while remaining unlikely because the person anticipates disapproval, believes the action is beyond their control, or assigns higher priority to competing responsibilities.

Attitudes concern the evaluated consequences of performing a specific action, rather than a general attitude toward health. Subjective norms concern the perceived social pressure or approval attached to that behavior. Later work often distinguishes injunctive norms—what relevant others approve—from descriptive norms—what they are understood to do. For preventive choices, these can diverge: people may believe clinicians strongly recommend vaccination while thinking that few members of their social network have received it. The psychological consequences depend on the relevant reference group and credibility of the information.

In their Armitage, C. J., & Conner, M (2001) evaluated the theory's performance across varied behaviors, documenting substantial prediction of intentions and more limited prediction of observed behavior. This asymmetry is central to preventive decision research. Theories explaining why an individual says they plan to act do not fully explain whether that action occurs. Action requires environmental opportunities, temporal coordination, sufficient resources, and resistance to competing demands.

For a study of preventive choices, the principle of compatibility is decisive: measures of intention, perceived control, and behavior should refer to the same action, target, context, and time frame. “I intend to protect my health” cannot be expected to predict a specific clinic visit with the precision of “I intend to attend this appointment within the next month.” Even compatible measures are subject to recall, desirability, and unexpected changes in circumstances. The model therefore functions best as an explicit set of measurable propositions that can be examined alongside behavioral records and structural constraints.

Decision Stages and Precaution Adoption

The precaution adoption process model introduces a sequence of decision states to explain transitions from unawareness to action. In Weinstein, N. D (1988), preventive adoption was analyzed through changes in knowledge, concern, and decision status rather than a single continuous motivation score. The model includes people who have never heard of a hazard, those who know about it but are personally unengaged, those actively deciding, those who decide against action, those who decide in favor, those who act, and those who maintain action where relevant. The empirical value of such a scheme depends on whether the proposed stages predict different information needs or intervention responses.

Weinstein, N. D., & Sandman, P. M (1992) applied the model to home radon testing. The case illustrates that awareness of a hazard is not equivalent to perceiving personal relevance, and that deciding to test is not equivalent to completing the test. The process framework encourages researchers to locate the exact transition targeted by a message. For an unaware individual, reliable information about the hazard may be necessary; for a person who has resolved to act, logistical support may be more effective than additional threat explanation.

Stage models face a serious methodological challenge: the classification procedure must yield states that are distinguishable, temporally meaningful, and capable of predicting transitions. If stages are assigned entirely from a single question about intention, the model may simply rename an existing motivational continuum. Longitudinal designs can address this by examining actual movement among states, testing whether purported stage-specific interventions improve transition probabilities, and determining whether trajectories are reversible.

Preventive decisions often recur. A person may adopt a precaution during a particular outbreak, later reassess personal risk, and then alter behavior as conditions change. The distinction between action initiation and maintenance is especially significant for recurring infection prevention, screening schedules, and durable lifestyle practices. A stage description should be used as a hypothesis about decision organization, not as an irreversible developmental sequence imposed on every individual.

Comparing Models and Their Explanatory Domains

The Health Belief Model, protection motivation theory, planned behavior, and precaution adoption frameworks illuminate different portions of the same decision process. The Health Belief Model organizes judgments about susceptibility, severity, benefits, barriers, and cues; protection motivation theory distinguishes threat appraisal from coping appraisal; planned behavior focuses on attitudes, norms, control, and intention; stage models ask where a person is in the formation and execution of a decision. A useful synthesis preserves these explanatory functions instead of collapsing all constructs into a single general “motivation” factor.

Comparison requires common endpoints and equivalent opportunities for action. A model tested against self-reported vaccination intention cannot be declared superior to another tested against administrative uptake records without accounting for measurement differences. Predictive improvement should be evaluated out of sample, prospectively where feasible, and relative to an appropriate baseline of demographic, access, and prior-behavior predictors. The size, stability, and practical usefulness of improvement matter as much as the statistical significance of an individual coefficient.

Theories also differ in how they treat feedback. Taking preventive action may alter felt vulnerability, self-efficacy, trust, and expected benefits, which then influence subsequent choices. Some studies interpret beliefs as one-way causes of action even when a reciprocal model is more plausible. Designs that measure beliefs before, during, and after decisions, and that distinguish actual exposure from protective response, can reveal these dynamic relations. Intervention research should test whether measured belief changes mediate observed behavioral change rather than infer mediation merely because both changed.

The resulting research program is integrative but empirically disciplined. It asks which constructs are necessary for a specified outcome, what their operational definitions are, whether their associations replicate, and whether manipulating a proposed mechanism changes a meaningful endpoint. The goal is a better account of preventive decision formation, including the conditions under which beliefs are influential and the conditions under which opportunities, institutional design, or established habits dominate.

Risk Perception as a Multidimensional Psychological Construct

A health risk can be expressed as an estimated probability, a relative comparison, an emotional impression, an expected loss, or a story about what might happen. These representations may arise from the same external hazard and yet differ substantially in their behavioral implications. Slovic, P (1987) demonstrated the importance of perceived qualities of hazards, including characteristics such as dread and perceived knowledge or controllability, within research on risk perception. The broader implication is methodological: a numerical estimate does not capture every feature of what a person experiences as risky.

Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, N (2001) developed the risk-as-feelings perspective, describing how emotional reactions can influence judgments and actions alongside or independently of deliberate assessments. Ferrer, R. A., & Klein, W. M. P (2015) later organized health behavior research around the complexities of risk perceptions and the importance of distinguishing measures. The Ferrer, R. A., Klein, W. M. P., Persoskie, A., Avishai-Yitsh (2016) articulated deliberative, affective, and experiential dimensions of perceived risk. The value of the framework lies in specifying testable differences among estimates of likelihood, worry or fear, and an intuitive sense of vulnerability.

A deliberative risk item asks an individual to estimate how likely a disease is within a defined period. An affective item assesses how worried or frightened they feel about it. An experiential item asks whether getting the disease feels likely, even when a precise probability is difficult to specify. Those responses may correlate without being identical. Researchers who pool them into one index can conceal theoretically important effects; investigators who analyze them separately must guard against measurement proliferation and post hoc selection of significant outcomes.

These distinctions matter in risk communication. Absolute frequency information may improve statistical comprehension, yet a vivid experience can alter worry more than numerical expectation. A frightening message may increase concern while leaving willingness to act unchanged if the action is burdensome. Conversely, a trustworthy recommendation can guide behavior with little reported fear. The scientific question is therefore which representation is being modified, which downstream variable is expected to change, and whether the observed result supports that causal pathway.

Comparative Risk, Optimistic Bias, and Personal Exceptionalism

People routinely assess their own prospects against those of similar others. Comparative risk judgments can help locate one's perceived standing within a group, but they need not be statistically calibrated. Weinstein, N. D (1982) is a landmark in the study of systematic personal risk comparisons. The basic phenomenon concerns judgments that adverse events are less likely for oneself than for comparable others, even when that aggregate belief cannot hold for everyone in a comparable population.

An important methodological distinction separates comparative optimism at the group level from miscalibration in a particular person. If someone has lower exposure because of behavior, occupation, vaccination, or environmental circumstances, perceiving lower risk may be appropriate. A study claiming unrealistic optimism must specify the comparison group, the available objective risk indicators, the outcome time horizon, and the statistical interpretation of the average response. A mean self–other gap does not identify every individual as misinformed.

Comparative judgments may also be influenced by psychological distance. People often have detailed information about their own precautions and less information about those of an abstract “average person.” This asymmetry can make one's protective actions especially salient while the average other's actions remain invisible. Survey wording can intensify the effect by using ambiguous reference groups or asking for likelihood without a common numerical scale. Such features are not incidental, because they determine what the measured construct actually represents.

The preventive consequences of optimism are conditional. Feeling personally safe might reduce the salience of prevention, but favorable self-beliefs can coexist with action when a recommended precaution is easy or socially normative. High reported vulnerability can likewise reflect prior disease experience or current symptoms rather than an impending choice. Investigators should treat unrealistic optimism as a possible source of risk-appraisal distortion and then test its behavioral consequences with prospective and intervention methods.

Affect, Dread, Worry, and Anticipated Regret

Emotions supply information about significance, urgency, and anticipated outcomes. Worry can focus attention on a threat, organize information seeking, and sustain an intention to prevent harm. At the same time, intense worry may disrupt deliberation, encourage avoidance, or narrow attention to immediate relief. The effects depend on the nature of the emotion, its timing, the perceived availability of a protective response, and the individual's coping resources.

The Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, N (2001) is especially useful where anticipated bodily harm, diagnostic uncertainty, and medical procedures evoke strong affective reactions. Such responses can diverge from explicit probability estimates. Someone may correctly understand that an outcome is statistically uncommon yet remain strongly distressed by the vividness of a possible loss. A model relying solely on stated probability would fail to capture the emotional burden entering that decision.

Anticipated regret adds another dimension. Individuals may consider how they would feel if they acted and experienced an adverse effect, or if they failed to act and later experienced preventable harm. These imagined emotions can be asymmetric because acts and omissions attract different social and moral interpretations. Their measurement requires care: asking people whether they would regret an unfavorable outcome may itself change the decision context by highlighting a previously neglected possibility. Experimental studies should therefore assess whether the elicitation of regret functions as an intervention rather than a neutral observation.

Affective information has legitimate explanatory value because prevention concerns outcomes that people care about, frequently under uncertainty. The challenge is to determine whether a feeling reflects useful experiential knowledge, a response to misleading presentation, a stable preference, or an immediate contextual cue. This requires designs that compare emotional and deliberative measures with actual choices and objective information. Emotional expression should not be equated automatically with a cognitive error, nor should the presence of emotion replace examination of numerical and clinical evidence.

Framing, Reference Points, and Behavioral Choice

Equivalent outcomes can be expressed through gains or losses, survival or mortality, relative reduction or absolute difference, and prevention or exposure. The classical framing research of Tversky, A., & Kahneman, D (1981) showed that the presentation of formally equivalent choices can change preferences. The effect is relevant to health decisions because prevention commonly involves a present action undertaken to avoid a future uncertain loss. The representation of the choice can shift which outcomes become psychologically prominent.

Health communication must distinguish legitimate framing from selective disclosure. Saying that a treatment reduces a relative risk by a particular percentage may be mathematically accurate while omitting the underlying absolute risk. Readers cannot reconstruct the absolute benefit without the relevant baseline or event rates. A comparison of “90% survival” and “10% mortality” may communicate equivalent proportions but evoke different expectations. Whether framing changes a real preventive decision depends on the domain, outcome, time frame, presentation format, and personal experience.

An effective research design isolates framing while holding substantive information constant. If one version of a message includes a dramatic narrative, authoritative endorsement, or different denominators, then the manipulation is more than gain–loss framing. Random assignment can identify the total effect of the message versions, but a mechanism claim requires that the intended component be isolated or independently measured. Researchers should also report whether outcomes concern comprehension, attitudes, intentions, decisions, or completed actions.

A persistent issue is whether a frame improves decision quality or merely changes choice. Increased uptake is not automatically a better outcome when the relevant choice involves tradeoffs that reasonable individuals value differently. Ethical public-health communication presents accurate benefits and harms, identifies uncertainty, and makes informed choice feasible. Where a preventive measure is strongly indicated by high-quality clinical evidence, behavioral interventions can also be assessed against equity and population outcomes, while preserving transparency about the evidence being communicated.

Numerical Literacy and the Interpretation of Health Statistics

Health decisions frequently require reasoning with small probabilities, different denominators, conditional risks, false-positive results, and time-dependent outcomes. Numeracy is therefore a substantive psychological variable rather than a peripheral educational characteristic. Peters, E., Västfjäll, D., Slovic, P., Mertz, C. K., Mazzocc (2006) examined associations between numerical ability and decision processes. Fagerlin, A., Zikmund-Fisher, B. J., Ubel, P. A., Jankovic (2007) developed the Subjective Numeracy Scale as a way to assess perceived numerical competence and preference for numerical information. Such scales can be useful, but they do not demonstrate that someone can correctly solve every probability problem encountered in health care.

Conditional probabilities can be particularly difficult when the relevant outcome is rare. A test's sensitivity and specificity are properties measured relative to disease status, whereas the probability of disease after a positive result also depends on its prevalence in the population being tested. Gigerenzer, G., Gaissmaier, W., Kurz-Milcke, E., Schwartz, L (2007) discussed how physicians and patients can misinterpret health statistics and how communication can be improved. The reasoning problem is not solved by presenting more decimal places: the structure of the numerical information and the clarity of the reference population matter.

Natural frequencies can make relationships easier to understand because they retain the reference classes from which conditional probabilities are derived. A carefully explained display might state, for example, how many of 1,000 people in a specified screening population are expected to have a condition, how many of those would receive a positive result, and how many without the condition would also test positive. Such numbers must come from the actual condition, test, and prevalence context under consideration; an illustrative calculation is not clinical evidence. Hoffrage, U., Lindsey, S., Hertwig, R., & Gigerenzer, G (2000) advanced this research into the communication of statistical medical information.

The challenge is also social and institutional. Numeracy differs across educational experiences, languages, age groups, and access to explanations, while highly numerate people can also be influenced by framing or motivated reasoning. A risk communication system designed only for a mathematically confident reader can systematically exclude others. Effective materials use consistent denominators, explicit time horizons, meaningful comparisons, tested visual formats, and opportunities to ask questions. Better communication quality should be demonstrated through comprehension, calibrated expectations, and decisions aligned with informed values rather than inferred from the sophistication of the numbers displayed.

Absolute Risk, Relative Risk, and Screening Probabilities

The distinction between absolute and relative measures is essential to an accurate preventive decision. Relative risk describes a ratio between event probabilities under different conditions; absolute risk describes an event probability in a defined population and period. An intervention can produce a large relative reduction while yielding a small absolute difference when baseline risk is low. Conversely, a moderate relative reduction can have considerable absolute importance in a population with high baseline risk. Any psychological study of perceived benefit should clarify which quantity participants saw and what they inferred.

Schwartz, L. M., Woloshin, S., Black, W. C., & Welch, H. G (1997) examined the role of numeracy in understanding information about the benefits of screening mammography. The research demonstrates why measuring recall of a percentage is different from measuring comprehension of the clinical tradeoff. The evaluation of screening also involves false positives, overdiagnosis in relevant contexts, follow-up procedures, and opportunity costs. These issues should be represented with evidence appropriate to the particular screening program rather than imported from one disease into another.

In a screening decision, a person may estimate the personal probability of disease, the probability that a test will detect it, and the probability that detection will improve a health outcome. Those quantities describe different stages of the causal pathway. Detecting more abnormalities is not equivalent to reducing mortality or clinically significant disease burden. The relevant evidence may involve randomized screening trials, observational follow-up, and modeling, each with its own assumptions and biases. Psychological research should avoid treating increased screening intention as proof of medical benefit.

Hoffmann, T. C., & Del Mar, C (2015) have been documented across medical decisions in a systematic review by Hoffmann and Del Mar (2015). Such findings motivate careful communication but do not justify assuming that every respondent holds the same beliefs or seeks the same outcome. Some individuals will reasonably assign different importance to convenience, uncertainty, reassurance, possible adverse effects, and long-term benefit. A high-quality decision process combines accurate probabilities with an explicit account of these values.

Information Sources, Credibility, and Health Communication

Health beliefs are formed through encounters with clinicians, health institutions, family members, peers, news media, commercial communication, and digital systems. The informational environment determines which hazards are made salient, which probabilities are accessible, and which authorities are considered credible. Trust is therefore part of the decision context. It can influence whether a recommendation is believed, whether uncertainty is interpreted as scientific transparency, and whether the person expects an institution to act competently and fairly.

A research design should separate source credibility from message content. The same numerical information may evoke different reactions when attributed to a public-health agency, an individual physician, a commercial sponsor, or an anonymous online account. Yet source effects depend on previous institutional experience and social position. A person whose earlier encounters with health care involved barriers or discrimination may interpret assurances differently from someone with consistent access to trusted care. Attributing these differences to individual irrationality would overlook causal features of the environment.

Tailored communication attempts to increase relevance by adapting material to measured characteristics, concerns, or behavior. Kreuter, M. W., & Wray, R. J (2003) articulated strategies for increasing relevance through targeted and tailored messages. A Noar, S. M., Benac, C. N., & Harris, M. S (2007) evaluated tailored print interventions, supporting the study of measurable but heterogeneous effects. Personalization can improve comprehension or attention, but it may also reinforce stereotypes or use weak individual data. Responsible implementation requires validated variables, privacy safeguards, and outcome assessment.

Communication interventions should specify the intervention's objective. Increasing the accuracy of risk understanding, raising knowledge, reducing uncertainty, increasing use of an effective preventive service, and strengthening autonomy are related but independent endpoints. An alarming message may increase recall without changing action. A simple invitation with a booking link may improve attendance while barely changing the person's beliefs about disease. The intervention should be evaluated against its stated mechanism and outcome, including possible harms such as confusion, heightened distress, misinformation, or unequal reach.

Measuring Health Beliefs and Risk Perception

A construct becomes useful to science through a reproducible operational definition. Susceptibility can be measured as numerical probability, verbal likelihood, comparison with peers, or likelihood in the absence of protection. Severity can be assessed through expected clinical consequences or perceived disruption. Barriers can be measured as anticipated difficulty, self-reported obstacles, actual out-of-pocket cost, or documented service availability. Self-efficacy can be measured through confidence ratings tied to specific tasks. Each choice defines a particular observed variable and limits what can be concluded from it.

Validating a measure requires evidence about content, response processes, internal structure, reliability, and relations with relevant outcomes. Internal consistency cannot establish that the items reflect a single stable psychological construct, especially where an index combines conceptually separate beliefs. Test–retest stability must be interpreted against the possibility that real beliefs change as information and exposure change. Cross-cultural comparisons require attention to language, reference groups, measurement invariance, and the social meaning of medical action.

An important methodological danger is common-method dependence. If beliefs and intentions are assessed in the same questionnaire, their association may reflect shared response tendencies, the proximity of similar questions, and social desirability. Self-reported behavior collected in the same interview can increase that problem. More persuasive tests combine questionnaires with prospective follow-up, observed choices, electronic records where ethically and legally appropriate, and independent measures of access and exposure. Consent, privacy, and the incompleteness of administrative records remain essential considerations.

Researchers should prespecify how they will handle missing data, multiple outcomes, subgroup comparisons, and time-varying beliefs. Many self-report scales permit convenient statistical modeling, but small shifts in a Likert average may have little practical meaning. Calibration can be evaluated when a validated benchmark probability exists; discrimination can be evaluated when the research question concerns ranking people by future outcomes; and decision relevance can be evaluated when the question concerns what people actually do. These are distinct measurement goals, and no single score can substitute for all three.

Feedback, Time, and Reciprocal Causation

The association between perceived risk and preventive behavior is dynamic. Taking an action can reduce actual exposure, produce a sense of security, and change the reported probability of future illness. A successful prevention program may thus create an inverse cross-sectional relationship between risk perception and uptake among participants whose protection has already lowered risk. Conversely, someone who remains exposed and concerned may report both high perceived risk and a failure to act because of barriers. Neither observed pattern alone identifies the effect of the initial risk judgment.

Prospective studies should measure beliefs before the decision, document the timing of action, and reassess beliefs and external circumstances afterward. Repeated-measures designs can separate stable between-person differences from within-person change, although causal conclusions still require careful assumptions. Natural experiments may help when an external policy changes access or message exposure. Randomized experiments provide stronger leverage on the total effect of a controlled communication intervention, but they do not automatically demonstrate that a specific psychological mediator caused the behavioral result.

Feedback also complicates the interpretation of perceived efficacy. A person who has never experienced a preventable illness may conclude that the precaution worked, that the baseline threat was exaggerated, or that the outcome was unlikely regardless of action. Each interpretation is compatible with the same observed absence of illness. Theories of preventive belief must therefore distinguish learning from experienced outcomes, learning from scientific information, and learning from social communication. Their relative importance is an empirical matter likely to vary by hazard.

Longitudinal research can clarify the temporal structure of intention formation, action, maintenance, and disengagement. A single preventive decision may occur within hours, whereas screening participation may unfold through months of consultation and scheduling. The correct period for analysis follows the decision's actual structure. Apparent prediction failures sometimes arise because a survey predicts a behavior on a time scale that was never plausible for the person studied.

Vaccination Decisions as a Test Case for Health Beliefs

Vaccination is a useful scientific setting for studying preventive decisions because the relevant action can often be observed, its medical benefits depend on vaccine and population, and the decision involves both personal and collective expectations. People evaluate the likelihood and seriousness of infection, the effectiveness and safety of vaccination, expected adverse effects, recommendations from trusted sources, access arrangements, and the behavior of others. The relationship between any single belief and vaccination is therefore embedded within a larger causal system.

Brewer, N. T., Chapman, G. B., Gibbons, F. X., Gerrard, M (2007) synthesized studies linking risk perception to vaccination behavior. Their findings support risk perception as a relevant predictor, while the limitations of correlational research remain important. People who intend to vaccinate may seek different information, while those already vaccinated may subsequently report reduced vulnerability. Vaccine uptake can also depend directly on supply, workplace requirements, paid leave, transportation, and whether an appointment is offered. A theory limited to perceived personal risk would not explain all of these influences.

A broader scientific synthesis by Brewer, N. T., Chapman, G. B., Rothman, A. J., Leask, J., & (2017) evaluated psychological strategies for increasing vaccination. A central implication is that interventions helping people act—through convenient access, reminders, recommendations, and structured opportunities—can be important alongside interventions aimed at changing thoughts and feelings. The evidence should be analyzed by intervention type and the actual uptake outcome rather than by the persuasive attractiveness of a communication campaign.

The Betsch, C., Schmid, P., Heinemeier, D., Korn, L., Holtmann (2018) offers a multidimensional instrument for examining vaccination-related psychological antecedents: confidence, complacency, constraints, calculation, and collective responsibility. These categories broaden explanation beyond a single perceived-risk scale. They are not identical to the Health Belief Model's constructs, and measurement comparisons should examine their convergence and discriminant validity rather than assume conceptual equivalence. Cultural adaptations and disease-specific applications require empirical assessment.

Vaccination choices also raise a question about public-health ethics. Information should distinguish established benefits, recognized risks, scientific uncertainty, and the implications of disease exposure without manufacturing individual clinical claims. In settings where people face competing obligations or limited access, effective policy may involve reducing barriers more directly than escalating perceived threat. The psychological literature is most useful when connected to verified clinical recommendations and the epidemiological context of the particular vaccine.

Screening, Testing, and Decisions Under Diagnostic Uncertainty

Screening decisions are shaped by the expectation that early detection could produce a meaningful benefit, together with concerns about false-positive results, overdiagnosis in applicable programs, subsequent procedures, and the burden of uncertainty. The psychological task differs from deciding whether to receive a vaccine: screening offers information and a possible pathway to additional intervention, rather than protection through the test itself. A person may value reassurance and yet fear the immediate emotional experience of waiting for a result.

Perceived susceptibility and severity influence whether screening feels relevant, while barriers and self-efficacy influence whether someone can complete the process. The invitation format, consultation time, recommendation source, availability of accessible explanations, and ease of scheduling may also matter. Weinstein, N. D., & Sandman, P. M (1992) illustrates the distinction between deciding to obtain information about exposure and actually carrying out a test. The broader lesson applies to screening pathways, even though disease-specific clinical outcomes must be evaluated independently.

Shared decision-making becomes especially salient when choices involve legitimate tradeoffs. The Stacey, D., Lewis, K. B., Smith, M., Carley, M., Volk, R., D (2024) assessed patient decision aids for treatment and screening decisions. Decision aids make options explicit, present evidence about benefits and harms, and help people clarify values. Their success should be evaluated in terms of knowledge, realistic expectations, decisional conflict, participation, value congruence, and the decision process—not merely the proportion who choose the intervention.

Psychological research on screening must also distinguish preventive intention from clinical effectiveness. An increase in test use can have varying medical implications depending on the condition, test properties, population, and follow-up care. Differences in clinical guidelines across countries or age groups may be scientifically justified by differing risk distributions and health-system circumstances. Accordingly, no general claim that “more screening is better” follows from the psychology of willingness or perceived benefit alone.

Infectious-Disease Prevention and Changing Risk Environments

Infectious-disease prevention presents a dynamic decision setting: exposure varies over time, other people's actions affect risk, information about transmission can evolve, and several protective behaviors may be combined. Risk perception may respond to local case trends, personal illness experience, institutional messages, or the perceived behavior of peers. A single baseline appraisal may therefore have limited predictive value for decisions extending over many months.

The distinction between initiating and maintaining protection is essential. A person may start using a precaution during an acute increase in concern, then discontinue it as perceived risk changes, inconvenience accumulates, or recommendations are revised. This trajectory can reflect rational adaptation to changing evidence as well as fatigue or misinformation; analysis requires actual exposure and policy context. Sorrentino, M., Fiorilla, C., Rubba, F., Montuori, P., & Pal (2025) examined risk perception and preventive behavior in relation to COVID-19 vaccination status, illustrating the empirical difficulties of interpreting maintenance, vaccination, and perceived risk simultaneously.

A study should define whether the intended outcome is mask use under specified circumstances, hand hygiene, staying home during symptomatic illness, testing, vaccination, or another behavior. Each involves different costs, benefits, social norms, and levels of available infrastructure. Aggregating them as a single “compliance” index can obscure mechanisms. It can also inadvertently treat a change in formal policy or access as a change in individual motivation.

Public-health communication in a changing outbreak must address uncertainty without suggesting that new evidence invalidates every earlier recommendation. Clear dates, sources, definitions, and explanations of why recommendations change can help preserve understanding. Behavioral research can test how uncertainty statements influence credibility, concern, and action, but the scientific adequacy of disease-specific advice rests on current epidemiological and clinical evidence.

Lifestyle Prevention and Repeated Choices

Lifestyle-related prevention often consists of recurrent decisions rather than a single discrete action. Physical activity, dietary change, smoking cessation, sleep practices, and other behaviors involve repeated opportunities, habits, rewards, and environmental constraints. A person may believe that an action will reduce long-term risk and nevertheless fail to implement it because the immediate costs are salient and the benefits are delayed. The explanatory system must therefore extend beyond initial willingness.

Risk appraisal can initiate deliberation, especially following a clinical consultation or family illness. Sustained change, however, may depend on planning, reliable access to resources, feedback, social support, routines, and contextual stability. The Sniehotta, F. F., Schwarzer, R., Scholz, U., & Schüz, B (2005) distinguished action planning from coping planning and followed cardiac rehabilitation patients as they returned to everyday life. Its practical significance lies in identifying processes between endorsing a goal and handling the barriers that emerge during action.

Different preventive behaviors require different models of efficacy. Smoking cessation includes dependence-related processes and often benefits from clinical supports that cannot be reduced to belief change. Physical activity is affected by transport systems, work patterns, safety, disability accommodations, and available space. Food choices depend on affordability, time, cultural practices, and access. An intervention that increases perceived risk may leave these constraints entirely unchanged.

The scientific endpoint should therefore match the level of explanation: a belief-focused intervention might be evaluated on appraisal and intention, while a program intended to improve physical activity should measure movement or a validated behavior outcome. Disease prevention claims require the relevant clinical or epidemiological evidence and appropriate follow-up. Psychological mechanisms are part of the causal pathway, but their existence does not establish the effect of an unspecified intervention on morbidity or mortality.

Fear Appeals, Defensive Responses, and Protective Action

Fear appeals attempt to change decisions by presenting an aversive outcome and, in many cases, a response intended to prevent it. Their study requires separating the communicated threat, the emotional reaction, the perceived effectiveness of the response, and the perceived ability to act. Witte, K., & Allen, M (2000) and the broader synthesis by Tannenbaum, M. B., Hepler, J., Zimmerman, R. S., Saul, L., J (2015) provide an evidence base for analyzing when fear-based communication changes attitudes, intentions, or behavior.

The critical mechanism question is whether elevated threat appraisal is accompanied by actionable coping information. A message that makes severe harm salient while withholding a feasible protective pathway may increase distress or encourage psychological distancing. In contrast, a message that presents a credible, accessible response can engage concern within a more constructive decision frame. Empirical studies must establish these effects directly; rhetorical assumptions about what “should frighten people into action” have no evidential standing.

Defensive responses are difficult to identify from self-report alone. A participant who questions a message may be engaging in motivated avoidance, but may also be identifying a genuine flaw, disagreeing with an unsupported probability, or drawing on valid clinical information. Researchers need measures of message accuracy, beliefs before exposure, comprehension, available coping options, and eventual behavior. Labelling skepticism as defense without this evidence risks circular explanation.

Ethically appropriate interventions use calibrated claims, actionable information, and transparent uncertainties. They also monitor unintended effects, including excessive fear, shame, stigmatization, avoidance of health care, and loss of trust. Where an intervention is delivered at scale, evaluation should examine differential responses across groups and whether the message amplifies existing health inequities.

Planning and the Intention–Behavior Gap

A preventive decision becomes practically meaningful when an intention is translated into an action at the relevant time and place. This transition is often imperfect. Armitage, C. J., & Conner, M (2001) identified a gap between the prediction of intentions and the prediction of behavior across applications of the theory of planned behavior. The gap arises from a combination of competing goals, imperfect control, delayed opportunities, forgetting, changes in circumstances, and the differences between reported intentions and observed choices.

Action planning specifies the where, when, and how of behavior. Coping planning anticipates likely obstacles and identifies feasible responses to them. Sniehotta, F. F., Schwarzer, R., Scholz, U., & Schüz, B (2005) investigated these planning dimensions longitudinally. Gollwitzer and Sheeran (2006)38002-1) synthesized research on implementation intentions, which connect a situational cue with a prepared response. These processes offer testable ways of connecting motivation with execution.

Planning interventions should be evaluated with behavior-specific outcomes. An intention to attend a screening visit may be made more actionable by selecting a date, arranging transportation, and preparing questions. Yet a plan cannot substitute for a service that is unavailable or unaffordable. A failed implementation intention under such conditions is evidence about environmental constraint as well as psychological prediction.

The gap also creates opportunities for measurement bias. People with greater access may report both stronger confidence and higher behavioral completion, making confidence appear to be the entire mechanism. Studies should distinguish the total influence of access, the direct effect of a planning intervention, and any mediation through perceived control. Ecological momentary assessment, real-world scheduling records, and time-stamped behavioral measures can strengthen this differentiation when privacy and consent are managed appropriately.

Norms, Relationships, Identity, and Collective Responsibility

Preventive choices are often relational. Family members exchange information, employers establish expectations, health professionals recommend actions, and communities generate shared interpretations of disease. Perceived norms can make a preventive behavior appear ordinary, burdensome, stigmatizing, or socially meaningful. The theory of planned behavior90020-T) places normative expectations within intention formation, but contemporary research also distinguishes descriptive beliefs about prevalence from injunctive beliefs about approval.

Identity can change the interpretation of the same message. A person may understand vaccination as an expression of responsibility, a source of anxiety, a condition of employment, or a symbol of group membership. These meanings are empirically consequential when they influence trust, perceived legitimacy, and the anticipated social costs of action. Researchers should measure the relevant reference group and avoid assuming that a norm observed in one context transfers across cultural or political settings.

Prevention may create benefits for others as well as for the individual, making collective responsibility an important variable. The Betsch, C., Schmid, P., Heinemeier, D., Korn, L., Holtmann (2018) includes a collective responsibility dimension among its proposed antecedents. The existence of this construct does not eliminate legitimate questions about personal benefit, risk, or autonomy; it adds a further motivational consideration that can be studied alongside confidence and constraints.

Network effects also complicate causal inference. People share sources and environments with their peers, so similarity in beliefs or behavior may result from influence, selection, or common exposure. Longitudinal network analysis, experiments in message delivery, and designs accounting for clustered environments can help distinguish these explanations. Responsible scientific communication states which effects are observed and which are inferred from a model.

Structural Barriers, Equity, and the Ecology of Choice

The apparent decision to “decline” a preventive service may reflect a complex sequence of constraints. Travel time, income insecurity, disability access, insurance arrangements, language barriers, clinic hours, digital literacy, administrative burden, and fear of stigmatization can all affect whether a preventive option is actually available. Psychological models that measure perceived barriers should be paired with evidence about objective access where the research question concerns realized action.

An opportunity structure can alter decisions without altering stated beliefs. Expanding clinic hours, reducing costs, providing interpreters, or simplifying scheduling can increase the feasibility of a preventive action, even if participants' perceived disease severity remains unchanged. A purely informational intervention may have little effect under the same conditions because the binding constraint lies elsewhere. This is a causal hypothesis that should be assessed through service records, randomized or quasi-experimental policy evaluation, and population-specific analysis.

Equity is also a problem of measurement. Surveys administered only in dominant languages, online-only recruitment, inaccessible questionnaire formats, and reliance on people already engaged with health care can systematically exclude populations facing the greatest barriers. Findings about “low motivation” from such samples can become misleading when generalized to entire communities. Studies should report sampling frames, missingness, language adaptation, and the relation between demographic groups and service opportunity.

The practical conclusion is that prevention emerges from relationships between cognition and environment. Belief changes can matter, especially when inaccurate risk estimates or poorly understood benefits obstruct informed choices. Institutional design can matter, especially when the person already wishes to act. Strong prevention programs determine which mechanism is actually limiting progress and evaluate the intervention accordingly.

Cross-Cultural Differences and Health Decision Contexts

The social meanings of illness, risk, prevention, and medical authority differ across cultural and institutional environments. Perceptions of severity may include effects on family obligations, employment, marriage prospects, or social participation that are absent from a narrow symptom-focused scale. Norms about whose advice is authoritative and how health decisions are discussed can vary across households and communities. Consequently, transferring a survey instrument or communication intervention across settings requires empirical validation.

A language translation can preserve literal wording while altering the implied reference group, level of certainty, or meaning of an action. The word “risk” itself can evoke statistical probability, danger, or moral responsibility depending on context. Cognitive interviewing, locally grounded qualitative research, and measurement-invariance analyses are valuable for interpreting differences. Without such checks, an apparent cultural effect may reflect differences in how a question is understood.

Cross-cultural comparison also involves health-system design. A person invited to free, conveniently located preventive care faces a different decision problem from one who must pay, travel extensively, or navigate fragmented services. Comparing belief–behavior correlations without modeling these differences can attribute structural effects to culture. Cultural explanation should be built from documented meanings and practices, and structural explanation from observed opportunities and constraints.

The most generalizable scientific conclusions concern relationships that have been tested across diverse conditions with sufficiently comparable measurement. Where evidence remains concentrated in a few high-income countries or convenience samples, the scope of inference should be stated. The purpose of comparative research is to identify mechanisms, moderators, and context dependence with increasing precision.

Research Designs and the Identification of Causal Effects

Research on health beliefs uses several designs, each capable of supporting different conclusions. Cross-sectional studies estimate associations within a population at one point in time. Prospective cohorts examine whether beliefs measured before an outcome help predict subsequent behavior. Laboratory experiments manipulate the content, frame, source, or timing of information under controlled conditions. Field experiments and randomized trials test interventions in ordinary service settings. Natural experiments and quasi-experimental studies examine changes in institutions or policies when random assignment is unavailable. These designs offer complementary evidence only when their assumptions and outcomes are made explicit.

In an observational study, perceived risk may be associated with vaccination uptake because perceived risk influenced the decision, because vaccination changed current risk perceptions, or because both are affected by prior health experience, socioeconomic opportunity, or trust in clinicians. Statistical adjustment can address measured covariates under assumptions but does not remove unmeasured confounding by declaration. Temporal ordering is helpful, yet it does not by itself establish causal identification. Repeated measures and clear causal diagrams can sharpen the questions that statistical models are asked to answer.

Randomizing a message creates a stronger test of the effect of that message, but its internal validity applies to the complete manipulation, including its imagery, wording, source, and call to action. If the intervention increases perceived susceptibility and behavioral uptake, mediation by susceptibility remains a separate empirical hypothesis. Mediator–outcome confounding, measurement error, intervention-induced changes in other beliefs, and differential dropout can undermine a simple mediation interpretation. Mechanism claims become more credible when experiments isolate relevant constructs or combine independent interventions with appropriate longitudinal measurement.

Real-world interventions may face interference: one participant's behavior changes another's exposure or decisions. Vaccination and infectious-disease prevention are examples where individual outcomes can depend on community patterns. Cluster-randomized designs and carefully specified units of assignment may be appropriate. Analysts should distinguish individual uptake from disease incidence, knowledge scores from clinical effects, and short-term compliance from long-term protection. Every causal conclusion must specify the population, time frame, treatment contrast, and endpoint it actually concerns.

Evidence from Meta-Analysis and Experimental Synthesis

Quantitative reviews supply a broader perspective than individual positive experiments, but they are only as informative as the studies, measures, and comparisons combined. Janz, N. K., & Becker, M. H (1984) reconstructed the early Health Belief Model literature across prospective and retrospective investigations. Their synthesis highlighted a substantial but heterogeneous tradition rather than validating every predicted relationship. Carpenter, C. J (2010) later meta-analyzed relationships between model constructs and health behavior, demonstrating the continuing importance of construct-specific estimates.

Risk perception research offers an instructive contrast between observational and experimental evidence. Brewer, N. T., Chapman, G. B., Gibbons, F. X., Gerrard, M (2007) synthesized the relationship between risk perceptions and vaccination behavior. Sheeran, P., Harris, P. R., & Epton, T (2014) examined experiments that heightened risk appraisals and assessed resulting changes in intentions and behavior. Their joint relevance is conceptual: an association between concern and uptake is not equivalent to the effect of deliberately increasing concern. The studies differ in research design, outcome definition, intervention content, and causal interpretation.

Meta-analyses of Witte, K., & Allen, M (2000) and Tannenbaum, M. B., Hepler, J., Zimmerman, R. S., Saul, L., J (2015) show that communication can change responses, while substantial variation remains across studies. Brewer, N. T., Chapman, G. B., Rothman, A. J., Leask, J., & (2017) integrates psychological evidence with the practical task of increasing actual vaccine uptake. The Stacey, D., Lewis, K. B., Smith, M., Carley, M., Volk, R., D (2024) addresses a different objective: supporting knowledgeable, informed, value-consistent choices in screening and treatment contexts.

A careful synthesis should report the number of included comparisons, population characteristics, risk of bias, study heterogeneity, and the distinction between proximal and distal outcomes when those figures have been verified from the source. This monograph cites identifiable syntheses and their general conclusions; it does not pool new effect estimates. Published averages should not be treated as universal parameters for every preventive decision. A meta-analytic result generated from specific vaccines or communication formats requires reexamination before it can support an intervention for another disease, age group, or healthcare system.

Distinguishing Correlation, Mediation, and Decision Quality

Three questions must be separated in preventive decision science. The descriptive question asks whether people who perceive greater risk are more likely to intend or perform an action. The causal question asks whether changing perceived risk alters that outcome under a defined intervention. The normative question asks whether the resulting decision is well informed, medically appropriate to the person's circumstances, and consistent with their values. Each question requires different evidence and may have a different answer.

A positive correlation can arise from common causes or prior behavior. A randomized message can establish a total effect without isolating the psychological mechanism. A behavioral increase can occur without an improvement in comprehension. A decision aid can improve understanding and value clarity while producing heterogeneous choices, which may be appropriate where benefits and burdens vary across individuals. Confusing these levels leads to inaccurate claims about what the research demonstrates.

For mechanism evaluation, direct measurement of beliefs before and after exposure is useful but insufficient on its own. A proposed mediation path needs temporal plausibility, reliable measures, appropriate adjustment for mediator–outcome confounding, and evidence against simple alternative explanations. When several beliefs change together, the investigator should avoid declaring one construct the exclusive mediator without stronger identification. Factorial designs or separate manipulations of source, efficacy information, and risk format may make competing explanations more distinguishable.

Decision quality itself requires an explicit definition. For some questions, it is the accurate understanding of absolute benefits and harms; for others, it is value-congruent choice, appropriately calibrated confidence, or completion of an evidence-based preventive action that is acceptable to the person. The relevant standard can change with the decision context. Quality assessment should therefore be specified before analysis and should not be retrospectively defined as whatever option the investigator preferred.

Bias, Reproducibility, and Generalizability

The health-belief literature is susceptible to selective measurement, outcome switching, limited statistical power, and publication bias. Investigators may measure multiple dimensions of risk, select among several behavior indicators, or examine many subgroups. A finding can appear robust when the underlying analysis has not accounted for these researcher degrees of freedom. Preregistration, transparent analysis plans, access to instruments and code where ethically possible, and replication across independent samples can help establish which effects are stable.

Self-reported preventive behavior is particularly vulnerable to recall error and social desirability. The problem can intensify during highly visible public-health campaigns, when respondents know the action socially expected of them. Administrative records and validated behavioral observations can improve outcome measurement but introduce other challenges: inaccessible services, incomplete data capture, cross-institution records, consent restrictions, and differences between services received and actual adherence. Triangulation across sources is preferable when feasible.

Most importantly, construct validity is not secured by repeated statistical significance. A short questionnaire about “risk” may measure current worry in one sample and probability estimates in another. Translation, age-related interpretation, trust, and reference-group ambiguity can further alter responses. Multigroup measurement invariance, cognitive interviews, and longitudinal assessment should be considered when comparing groups and periods. Positive findings with different instruments should not be pooled casually under a common theoretical label.

External validity is a substantive question. Samples of university students responding to hypothetical scenarios may reveal cognitive mechanisms under controlled conditions, but they cannot by themselves establish effects on clinic attendance among older adults or on vaccination in underserved rural communities. Intervention studies should describe recruitment, exclusion criteria, baseline risk, healthcare setting, competing demands, and implementation fidelity. Null findings warrant as much scrutiny as positive results because they reveal where the proposed mechanism may lack practical influence.

Decision Aids, Shared Deliberation, and Value Clarification

A preventive decision often requires understanding several options rather than accepting or rejecting one proposed behavior. Patient decision aids are designed to make a choice explicit, present evidence on benefits and harms, and help people articulate which outcomes matter to them. The Stacey, D., Lewis, K. B., Smith, M., Carley, M., Volk, R., D (2024) is a major evidence source for studying these tools across treatment and screening contexts. Its relevance extends beyond a single psychological model because decision aids integrate comprehension, preferences, and conversation.

A shared-decision framework must address the difference between scientific uncertainty and individual valuation. Evidence may estimate the frequency of outcomes under specified alternatives, while the individual assigns importance to pain, time, uncertainty, possible side effects, future benefits, or family needs. A researcher can evaluate knowledge and decisional conflict without assuming that every informed participant will choose the same option. The endpoint should reflect the purpose of shared deliberation.

Decision aids can take printed, digital, graphical, or conversational forms. Their complexity can become a barrier if they demand excessive numeracy, internet access, language proficiency, or time. Formats should be tested with intended users, including people with disabilities and people unfamiliar with medical vocabulary. The research question includes whether a tool is accessible, comprehensible, and usable within a genuine clinical conversation, rather than merely whether it contains all relevant statistics.

A system that provides technically correct numbers can still mislead through selective comparison, an inappropriate reference population, unsupported personalization, or omission of uncertainty. Quality therefore depends on the provenance of the evidence, the clarity of the options, and the absence of manipulative framing. Future research should study whether decision aids produce durable benefits in understanding and decisions across diverse settings, and how their effects vary when combined with clinician communication or altered service access.

Digital Health, Algorithmic Risk Estimates, and Artificial Decision Support

Digital platforms increasingly mediate how health information is delivered and how preventive choices are organized. Risk calculators, patient portals, reminders, conversational interfaces, and automated scheduling can transform the timing and salience of decisions. The immediate psychological questions concern understanding, confidence, trust, information overload, anxiety, and the perceived legitimacy of a recommendation. The technical questions concern model accuracy, data provenance, calibration, representativeness, privacy, and reliability. The two domains intersect but require distinct methods of evaluation.

A personalized score can look authoritative even when it is based on incomplete or inappropriate data. Calibration concerns whether predicted probabilities correspond to observed event rates in the relevant population; discrimination concerns whether a model ranks higher-risk cases above lower-risk cases. Neither property guarantees that the risk information will improve a person's decision. A score could be statistically useful yet misunderstood, misapplied to the wrong population, or presented without the relevant action's benefits and harms.

Conversational artificial systems add a relational dimension because users may treat confident, fluent explanations as indicators of understanding or expertise. That perception is a psychological variable requiring direct investigation. The scientific evaluation should distinguish the content's factual quality, the user's comprehension, the inferred reliability of the system, and any change in actual preventive behavior. Claims about artificial agency, intentionality, or psychological organization belong to explicitly specified theoretical models and cannot be inferred from fluency or user trust alone.

In the historical horizon of the Artificial Era, preventive decision-making can be studied as a distributed informational process involving people, clinicians, institutional rules, and artificial systems. Testable hypotheses include whether an interface improves comprehension of absolute risk, whether automated reminders reduce missed actions, whether explanation style produces overconfidence, and whether performance differs across demographic groups. Such studies should use predeclared outcomes, established clinical standards, and transparent oversight. Medical decisions require verified clinical evidence, appropriate professional responsibility, and an accountable path for correcting errors.

Contemporary Evidence and the Durability of Preventive Behavior

Recent work continues to examine the stability of preventive behavior once the initial threat becomes less salient. Sorrentino, M., Fiorilla, C., Rubba, F., Montuori, P., & Pal (2025) reviewed studies of COVID-19 vaccination, perceived risk, and protective behaviors and conducted a meta-analysis of relevant subgroups. The review demonstrates the importance of studying risk judgments and behavior together over time, while also exposing the interpretive difficulty of separating vaccination status, prior protection, changing circumstances, and individual motivation.

Research in this area should avoid treating observed changes in protective behavior as evidence for a single psychological explanation. Revisions to medical guidance, changes in epidemiological risk, fatigue, institutional restrictions, and perceived efficacy can all affect behavior. A review of heterogeneous studies can identify patterns but cannot resolve all sources of confounding. Subgroup comparisons should be interpreted in light of their design, sampling, and measured covariates.

At a broader level, contemporary Stacey, D., Lewis, K. B., Smith, M., Carley, M., Volk, R., D (2024) and the synthesis of Brewer, N. T., Chapman, G. B., Rothman, A. J., Leask, J., & (2017) show why the field benefits from outcome-specific research. Better decisions, greater knowledge, more accurate beliefs, and greater uptake are all legitimate objects of study, but their equivalence cannot be assumed. There is also a continuing need to evaluate effects across languages, clinical settings, information systems, and forms of disadvantage.

The most important advance is methodological: prevention can now be investigated through combinations of behavioral theory, field experimentation, implementation data, psychometrics, health communication, and population-level analysis. The presence of more sophisticated tools does not settle the underlying conceptual problem. The researcher must still define the object, the proposed causal pathway, the observed outcome, and the population to which a claim is intended to apply.

Critical Scientific Discussion

The central theoretical tension in this field concerns the relation between subjective evaluation and feasible action. Models of health beliefs often explain how a person represents danger and expected benefit; studies of behavior reveal that choice is also conditioned by norms, habits, service design, and material resources. These levels are causally related. A person's belief that screening is impossible may reflect an objectively inaccessible service; an institution's redesign may alter behavior by making a previously desired action feasible. Scientific synthesis requires measuring both the representation and the condition it represents.

The second tension concerns the relation between persuasion and understanding. Raising concern may promote a selected preventive behavior, but a decision can become more informed without moving toward the same action in every case. This is particularly relevant to preference-sensitive medical choices. The proper standard of success must follow the clinical evidence, the decision's structure, and the individual's values. Evaluation should distinguish knowledge, calibrated belief, decisional quality, intention, actual uptake, and long-term health outcomes.

A third tension is the stability of risk perception. The phenomenon is a moving relation among objective exposure, personal experience, communication, action, and feedback. Cross-sectional associations can be misleading if the preventive action has already changed the person's risk or sense of vulnerability. Temporal models, counterfactual clarity, and naturalistic follow-up are necessary to avoid interpreting effects as causes. Researchers also need comparable measures across time so that apparent change does not reflect a change in question meaning.

A fourth problem concerns the role of affect. Emotional responses are legitimate psychological components of decisions involving bodily vulnerability and uncertain outcomes. Their relationship with probability judgments depends on context. An explanation of affect should test its predictive and causal role, assess when it supports or disrupts comprehension, and examine whether risk communication produces avoidable distress. The goal is an empirically specified account of psychological organization under uncertainty.

Finally, the field must integrate individual and collective evidence. Public-health interventions can benefit some populations and burden others. An average effect may conceal unequal access, exclusion from research, or disproportionate psychological harm. Evaluating preventive decisions therefore requires careful population description, ethics, equity indicators, and implementation conditions. Theory gains explanatory strength when its constructs are connected to real constraints and independently observed outcomes.

Open Research Problems and Future Directions

One priority is the development of risk-perception measures that can distinguish deliberative probability estimates, affective reactions, experiential vulnerability, perceived severity, response efficacy, and structural feasibility across languages and cultures. Validation should cover reliability, content and construct validity, interpretation by participants, and longitudinal comparability. These efforts can clarify which dimensions predict choice in specific settings and which apparent relationships arise from measurement artifacts.

A second priority is the experimental analysis of mechanisms. Multi-arm studies can vary probability format, source credibility, efficacy information, reminders, and ease of action while measuring intermediate constructs and meaningful behavior. Factorial experiments may test interactions between risk appraisal and coping opportunity. Predeclared causal models, longer follow-up, and independent replication can help determine whether a psychological change truly mediates an outcome.

A third priority is the study of decision trajectories. Rather than assuming a one-time choice, longitudinal research can examine how uncertainty, new information, prior action, disease incidence, and social norms alter beliefs over months or years. Relevant endpoints include initiating action, maintaining it, appropriately revising it, and disengaging when risk or medical recommendations change. The temporal organization of prevention remains a substantial explanatory problem.

A fourth priority is intervention equity. Studies should compare communication-only programs with changes to cost, scheduling, travel, language access, or institutional outreach; they should also test combinations. The benefit of a theoretically compelling message can be limited when structural barriers dominate. Outcomes should be reported for groups whose circumstances are substantively different, with sufficient power to examine differential effects where possible.

A fifth priority concerns the relation between artificial information systems and informed preventive decisions. Experiments should test whether automated explanations improve knowledge or merely confidence, whether risk estimates are appropriately calibrated, how users detect errors, and whether interfaces widen or reduce inequalities. Privacy, clinical safety, bias evaluation, provenance, and accountable correction procedures must be integral to the design. The long-term question is how preventive reasoning changes as people increasingly encounter health information through a mixture of human professional guidance and artificial systems.

Conclusion

Health beliefs, perceived risk, and preventive decisions form a coherent scientific object because they connect interpretations of future harm with evaluations of available action. The Health Belief Model, protection motivation theory, planned behavior, and precaution adoption framework each contribute a distinct account of this connection. Research on cognitive framing, numerical understanding, affect, and social influence explains why formally similar information can be interpreted differently. Intervention studies show that communication and risk appraisal can affect behavior, while planning, service access, trust, and structural opportunity determine whether an intention becomes an action.

The major empirical lesson is the necessity of matching scientific claims to measured outcomes and research designs. Cross-sectional prediction, experimental change in a message response, completed preventive behavior, informed decision quality, and clinical benefit are analytically separate achievements. Their relations can be tested, but none can be substituted for another. The strongest research combines validated constructs, causal identification, prospective measurement, clinically meaningful outcomes, and explicit consideration of the population and environment.

The resulting synthesis places preventive choice within a dynamic organization of cognition, affect, action, institutions, and communication. Its future development depends on independently verified evidence, transparent risk representation, and an increasing ability to explain where beliefs matter, where practical barriers dominate, and how informed choices can be sustained as personal and social conditions change.

References

  • Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. https://doi.org/10.1016/0749-5978(91)90020-T
  • Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: A meta-analytic review. British Journal of Social Psychology, 40(4), 471–499. https://doi.org/10.1348/014466601164939
  • Betsch, C., Schmid, P., Heinemeier, D., Korn, L., Holtmann, C., & Böhm, R. (2018). Beyond confidence: Development of a measure assessing the 5C psychological antecedents of vaccination. PLOS ONE, 13(12), e0208601. https://doi.org/10.1371/journal.pone.0208601
  • Brewer, N. T., Chapman, G. B., Gibbons, F. X., Gerrard, M., McCaul, K. D., & Weinstein, N. D. (2007). Meta-analysis of the relationship between risk perception and health behavior: The example of vaccination. Health Psychology, 26(2), 136–145. https://doi.org/10.1037/0278-6133.26.2.136
  • Brewer, N. T., Chapman, G. B., Rothman, A. J., Leask, J., & Kempe, A. (2017). Increasing vaccination: Putting psychological science into action. Psychological Science in the Public Interest, 18(3), 149–207. https://doi.org/10.1177/1529100618760521
  • Carpenter, C. J. (2010). A meta-analysis of the effectiveness of health belief model variables in predicting behavior. Health Communication, 25(8), 661–669. https://doi.org/10.1080/10410236.2010.521906
  • Fagerlin, A., Zikmund-Fisher, B. J., Ubel, P. A., Jankovic, A., Derry, H. A., & Smith, D. M. (2007). Measuring numeracy without a math test: Development of the Subjective Numeracy Scale. Medical Decision Making, 27(5), 672–680. https://doi.org/10.1177/0272989X07304449
  • Ferrer, R. A., & Klein, W. M. P. (2015). Risk perceptions and health behavior. Current Opinion in Psychology, 5, 85–89. https://doi.org/10.1016/j.copsyc.2015.03.012
  • Ferrer, R. A., Klein, W. M. P., Persoskie, A., Avishai-Yitshak, A., & Sheeran, P. (2016). The tripartite model of risk perception (TRIRISK): Distinguishing deliberative, affective, and experiential components of perceived risk. Annals of Behavioral Medicine, 50(5), 653–663. https://doi.org/10.1007/s12160-016-9790-z
  • Gigerenzer, G., Gaissmaier, W., Kurz-Milcke, E., Schwartz, L. M., & Woloshin, S. (2007). Helping doctors and patients make sense of health statistics. Psychological Science in the Public Interest, 8(2), 53–96. https://doi.org/10.1111/j.1539-6053.2008.00033.x
  • Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. Advances in Experimental Social Psychology, 38, 69–119. https://doi.org/10.1016/S0065-2601(06)38002-1
  • Hoffmann, T. C., & Del Mar, C. (2015). Patients' expectations of the benefits and harms of treatments, screening, and tests: A systematic review. JAMA Internal Medicine, 175(2), 274–286. https://doi.org/10.1001/jamainternmed.2014.6016
  • Hoffrage, U., Lindsey, S., Hertwig, R., & Gigerenzer, G. (2000). Medicine: Communicating statistical information. Science, 290(5500), 2261–2262. https://doi.org/10.1126/science.290.5500.2261
  • Janz, N. K., & Becker, M. H. (1984). The health belief model: A decade later. Health Education Quarterly, 11(1), 1–47. https://doi.org/10.1177/109019818401100101
  • Kreuter, M. W., & Wray, R. J. (2003). Tailored and targeted health communication: Strategies for enhancing information relevance. American Journal of Health Behavior, 27(Suppl. 3), S227–S232. https://doi.org/10.5993/ajhb.27.1.s3.6
  • Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, N. (2001). Risk as feelings. Psychological Bulletin, 127(2), 267–286. https://doi.org/10.1037/0033-2909.127.2.267
  • Maddux, J. E., & Rogers, R. W. (1983). Protection motivation and self-efficacy: A revised theory of fear appeals and attitude change. Journal of Experimental Social Psychology, 19(5), 469–479. https://doi.org/10.1016/0022-1031(83)90023-9
  • Noar, S. M., Benac, C. N., & Harris, M. S. (2007). Does tailoring matter? Meta-analytic review of tailored print health behavior change interventions. Psychological Bulletin, 133(4), 673–693. https://doi.org/10.1037/0033-2909.133.4.673
  • Peters, E., Västfjäll, D., Slovic, P., Mertz, C. K., Mazzocco, K., & Dickert, S. (2006). Numeracy and decision making. Psychological Science, 17(5), 407–413. https://doi.org/10.1111/j.1467-9280.2006.01720.x
  • Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude change. The Journal of Psychology, 91(1), 93–114. https://doi.org/10.1080/00223980.1975.9915803
  • Rosenstock, I. M. (1974a). Historical origins of the health belief model. Health Education Monographs, 2(4), 328–335. https://doi.org/10.1177/109019817400200403
  • Rosenstock, I. M. (1974b). The health belief model and preventive health behavior. Health Education Monographs, 2(4), 354–386. https://doi.org/10.1177/109019817400200405
  • Schwartz, L. M., Woloshin, S., Black, W. C., & Welch, H. G. (1997). The role of numeracy in understanding the benefit of screening mammography. Annals of Internal Medicine, 127(11), 966–972. https://doi.org/10.7326/0003-4819-127-11-199712010-00003
  • Sheeran, P., Harris, P. R., & Epton, T. (2014). Does heightening risk appraisals change people's intentions and behavior? A meta-analysis of experimental studies. Psychological Bulletin, 140(2), 511–543. https://doi.org/10.1037/a0033065
  • Slovic, P. (1987). Perception of risk. Science, 236(4799), 280–285. https://doi.org/10.1126/science.3563507
  • Sniehotta, F. F., Schwarzer, R., Scholz, U., & Schüz, B. (2005). Action planning and coping planning for long-term lifestyle change: Theory and assessment. European Journal of Social Psychology, 35(4), 565–576. https://doi.org/10.1002/ejsp.258
  • Sorrentino, M., Fiorilla, C., Rubba, F., Montuori, P., & Palladino, R. (2025). Does COVID-19 vaccination affect risk perception and adherence to preventive behaviors? A systematic review and meta-analysis. Frontiers in Public Health, 13, Article 1661015. https://doi.org/10.3389/fpubh.2025.1661015
  • Stacey, D., Lewis, K. B., Smith, M., Carley, M., Volk, R., Douglas, E. E., Pacheco-Brousseau, L., Finderup, J., Gunderson, J., Barry, M. J., Bennett, C. L., Bravo, P., Steffensen, K., Gogovor, A., Graham, I. D., Kelly, S. E., Légaré, F., Sondergaard, H., Thomson, R.,. Trevena, L. (2024). Decision aids for people facing health treatment or screening decisions. Cochrane Database of Systematic Reviews, 2024(1), Article CD001431. https://doi.org/10.1002/14651858.CD001431.pub6
  • Tannenbaum, M. B., Hepler, J., Zimmerman, R. S., Saul, L., Jacobs, S., Wilson, K., & Albarracín, D. (2015). Appealing to fear: A meta-analysis of fear appeal effectiveness and theories. Psychological Bulletin, 141(6), 1178–1204. https://doi.org/10.1037/a0039729
  • Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453–458. https://doi.org/10.1126/science.7455683
  • Weinstein, N. D. (1982). Unrealistic optimism about susceptibility to health problems. Journal of Behavioral Medicine, 5(4), 441–460. https://doi.org/10.1007/BF00845372
  • Weinstein, N. D. (1988). The precaution adoption process. Health Psychology, 7(4), 355–386. https://doi.org/10.1037/0278-6133.7.4.355
  • Weinstein, N. D., & Sandman, P. M. (1992). A model of the precaution adoption process: Evidence from home radon testing. Health Psychology, 11(3), 170–180. https://doi.org/10.1037/0278-6133.11.3.170
  • Witte, K., & Allen, M. (2000). A meta-analysis of fear appeals: Implications for effective public health campaigns. Health Education & Behavior, 27(5), 591–615. https://doi.org/10.1177/109019810002700506