Perceived severity refers to the degree people deem a particular disease or condition is serious. Perceived severity includes how people perceive the deleterious consequences of a serious health event or outcome, such as a diagnosis of cancer. Witte et al. (1992, p. 10) defines perceived severity as a person's beliefs about the significance or magnitude of a health threat (e.g., "skin cancer leads to death").
Perceived severity is a central concept in several fear appeals health information processing models, including the Health Belief Model (Becker, 1974; Rosenstock, 1974; Rosenstock, Strecher, & Becker, 1994), parallel response model (Leventhal, 1970), Protection Motivation Theory (Rogers, 1975, 1983) and the Extended Parallel Process Model (EPPM, Witte, 1992, 1998).
Perceived severity combined with perceived susceptibility form a perceived threat (Sturges & Rogers, 1996; Witte, 1992), which can influence how people process health information and how motivated they may be to engage in a particular behavior. The perceived severity of consequences may relate to an anticipated event that may occur in the future, or to an operant state such as a pre-existing health problem. In essence, perceived severity is the perception of how serious a health problem or disease is if one were to contract it.
In Protection Motivation Theory (PMT, Rogers, 1983) severity is part of a first appraisal (threat) upon exposure to a fear appeal message. If perceived severity and perceived susceptibility are judged to be high, then an individual will engage in a second appraisal (coping) by appraising both self and response efficacy. If both efficacy appraisals are judged to be sufficiently high, then an individual will be motivated to reduce the threat (by adopting message recommendations). If one or both of the efficacy appraisals are insufficiently high, then the individual will be motivated to reduce the fear by a number of different strategies (e.g., message derogation, message avoidance).
Perceived severity also has been evaluated in assessments of the non-health consequences of a health threat, such as the impact of disease on financial, social, or psychological outcomes (e.g., Champion, 1984). In this case, perceived severity has a conceptual overlap with outcome expectancies in models such as the Theory of Reasoned Action (Fishbein & Ajzen, 1975) and the Theory of Planned Behavior (Ajzen, 1985). Perceptions of severity are likely to vary widely between individuals (Janz & Becker 1984; Rosenstock, 1974).
A simple, adaptable, and short index was created by Witte, Cameron, McKeon, & Berkowitz (1996). The generic nature of the items in the index may reduce its predictive utility in a particular disease or behavioral context. In the aforementioned study, Witte et al. administered the scale as part of a Risk Behavior Diagnosis Scale to 179 college students regarding their attitudes about genital warts. Witte et.al. obtained a Cronbach's α = .90.
- I believe that (the name of the health threat) is severe.
- I believe that (the name of the health threat) is serious.
- I believe that (the name of the health threat) is significant.
5-point scales were anchored by strongly agree to strongly disagree.
Moss-Morris et al. (2002) and Figueiras and Alves (2007) tested versions of the Illness Perceptions Questionnaire—Moss-Morris the IPQ-R (Illness Perceptions Questionnaire- Revised) and Figueiras and Alves the IPQ-RH (Illness Perceptions Questionnaire-Revised healthy people). The Illness Perceptions Questionnaire was developed to measure patients' representations of their illness. One of the strengths of the latter measure is its use of self-referencing. There are a number of disease-specific versions of the IPQ (e.g. for asthma, diabetes, chronic pain). While their entire scale is multidimensional, the section on severity (labeled "consequences" in the IPQ-R) is the same across all the different disease-specific measures. The only variation is the insertion of the specific illness within the question.
One limitation with some of the items within the various versions of the IPQ is they are ambiguous with respect to valence. Items such as 'My illness strongly affects the ways others see me' do not clarify whether these changes are positive (e.g. causes people to be sympathetic or view the individual as strong) or negative.
- My [name of illness] is a serious condition
- My [name of illness] has major consequences on my life
- My [name of illness] does not have much effect on my life (reverse scored)
- My [name of illness] strongly affects the way others see me
- My [name of illness] has serious financial consequences
- My [name of illness] causes difficulties for those who are close to me
Each item uses 5-point scales anchored by strongly agree to strongly disagree. Cronbach's alpha = 0.84. This value was observed using a total of 711 people from one of eight different illness groups (e.g. asthma and diabetes sufferers). The majority of respondents were recruited from clinics within hospitals.
Champion (1984) developed a 12-item index to assess perceived severity (seriousness) of breast cancer (Cronbach's α = .78). Although the following scale is multidimensional, it is moderately reliable as a single index. Please note the scale below includes more multidimensional measures than threats to one's health, such as breast cancer's perceived impact on financial, social, and psychological outcomes.
- The thought of breast cancer scares me.
- When I think about breast cancer I feel nauseous.
- If I had breast cancer my career would be endangered.
- When I think about breast cancer my heart beats faster.
- Breast cancer would endanger my marriage (or a significant relationship).
- Breast cancer is a hopeless disease.
- My feelings about myself would change if I got breast cancer.
- I am afraid to even think about breast cancer.
- My financial security would be endangered if I got breast cancer.
- Problems I would experience from breast cancer would last a long time.
- If I got breast cancer, it would be more serious than other diseases.
- If I had breast cancer, my whole life would change.
Each item used 5-point scales anchored by strongly agree to strongly disagree.
Rationale for Selection
The aforementioned items were chosen because they represent a wide range of contexts in which perceived severity has been measured reliably. Each provides a flavor of typically used items, and each can be adapted to a particular health context under study.
The alphas for perceived severity tend to have a broad range within the literature depending on the unique context of a study and the particular health threat considered. For example, the items selected as examples herein have a range (e.g., .78 to .90).
Use of Measure (with examples)
Figueiras & Alves, (2007) tested a revised Illness Perceptions Questionnaire for healthy people (IPQ-RH). The items are self-referenced -- and it is possible this measure assesses illness stereotypes instead of individual severity perceptions. As with the Moss-Morris et al, 2002 scale, perceived severity is represented by the dimension labeled "consequences." Healthy individuals were obtained in a sample of 1113 members of the general population recruited from different working environments. They were aged 18-65 and the majority was female (64%). Cronbach's α = .66.
Rimal & Morrison (2006) measured perceived severity by asking participants to estimate how serious each of 15 risk events would be (if it happened to them and if it happened to a similar, moderately dissimilar, and a dissimilar referent). Responses were recorded on 5-point Likert-type scales ranging from 1 (not at all serious) to 5 (extremely serious or fatal). Severity indexes were calculated as the average of responses to the 15 items. Reliability coefficients ranged from α =.81- .90. Rimal & Morrison (2006) found severity ratings were highest for a similar referent. This suggests individuals strategically use severity (and susceptibility) ratings to moderate their assessment of overall risk and personal identity.
Range of items Used
The range of items varies significantly in the literature. Some studies use a single item to measure perceived severity. For example, van den Berg, et al. (2008) used a single item to measure the perceived severity of the birth of a child with Down's syndrome. To illustrate two examples at the other extreme, Weinstein (2000) measured the perceived severity of 201 events, each with a 10-point rating scale anchored by "innocuous, no harm at all" to "extremely devastating." The diverse 'events' measured within the study included allergies, bananas, and skin cancer. Champion's (1984) used a 12-item index to assess perceived severity (seriousness) of breast cancer.
The relationship between perceived severity and perceived susceptibility, and their impact on perceived threat has been tested both additively and multiplicatively. Currently, there seems to be little agreement how severity and susceptibility combine to predict behavior. According to several health behavior theories, if either susceptibility, severity, efficacy (either response - or self -) is zero, then, an individual's motivation to act also should be nil (e.g. Feather, 1982; Weinstein, 2000).
Rogers' (1975) original version of PMT indicated severity, vulnerability and response efficacy combined multiplicatively. Rogers (1983) later revised the model in which severity and vulnerability were summed.
Witte, in the EPPM, also proposed an additive model, suggesting that vulnerability and severity should be summed. However, Witte argued threat should be subtracted from efficacy. In addition, Witte (1998) proposed threat perceptions need to reach a certain threshold before one's motivation increases to consider specific health behaviors.
In a meta-analysis, Brewer, Chapman, Gibbons, Gerrard, McCaul, & Weinstein (2007) examined the relationship between perceived severity and vaccination in 32 studies with 13,945 participants. The pooled effect size was small to moderate (r = .16) and significantly different from zero (p < .001). They concluded participants who perceived the severity of illness to be higher were more likely to be vaccinated.