National Cancer Institute   National Cancer Institute

Table of Contents
1 General Description and Theoretical Background
2 Role of Perceived Severity in Health Behaviour Theories
3 Protection Motivation Theory and Extended Parallel Process Model.
4 Measurement Issues
5 Similar Constructs
6 References

Appendices: Severity Measures

8 Published Examples

Download Full Text (PDF)

Other Constructs



Dispositional Optimism




Illness Representations

  Implementation Intentions
  Intention, Expectation, and Willingness
  Normative Beliefs
  Optimistic Bias
  Perceived Benefits
  Perceived Control
  Perceived Severity
  Perceived Vulnerability
  Self-Reported Behavior
  Social Influence
  Social Support

Perceived Severity
Anne Miles

<< Previous
Next >>


Appendices: Severity Measures

Although there are reliability indices for the following measures, researchers need to consider their individual research questions before selecting a measure. The strengths and weakness of the various measures listed below are noted to help direct people to make an appropriate choice.

Health Belief Model
Champion (1984) developed measures of the HBM constructs in the context of breast self-examination. The strengths of the scale developed to measure severity (called 'seriousness' in the paper) are firstly that the items are largely self-referenced, and hence measure the impact the illness would have on the individual, and secondly that a number of factors have been measured, such as financial security and personal relationships as well as the severity of the disease itself. However, the scale also contains items that measure emotional response to the disease, and as noted earlier, these are not usually included in measures of severity. The items with an ‘*’ are the items more commonly understood to measure severity (the remainder measure fear).

One potential limitation with this measure is that the questions do not specify a behavioural context. Although breast self-examination has not been associated with an improvement in prognosis, if this measure were applied to other behaviours, such as mammography uptake, this issue might be worth considering (e.g. by rephrasing the items as follows: 'If I had a mammogram and was found to have breast cancer, it would be more serious than other diseases').

  1. The thought of breast cancer scares me.
  2. When I think about breast cancer I feel nauseous.
  3. If I had breast cancer my career would be endangered. *
  4. When I think about breast cancer my heart beats faster.
  5. Breast cancer would endanger my marriage (or a significant relationship).*
  6. Breast cancer is a hopeless disease.*
  7. My feelings about myself would change if I got breast cancer. *
  8. I am afraid to even think about breast cancer.
  9. My financial security would be endangered if I got breast cancer.*
  10. Problems I would experience from breast cancer would last a long time.*
  11. If I got breast cancer, it would be more serious than other diseases. *
  12. If I had breast cancer, my whole life would change.*

Five item response scale: strongly disagree to strongly agree.

Cronbach's alpha = 0.78; test-retest reliability across an interval of 2 weeks = 0.76. These values were observed using a convenience sample of 301 women, aged 17 to 82. The majority were white, married and had a high school education and were of higher SES than the average for the general population. Test-retest reliability was performed on a sub-sample of 57 women.

Extended Parallel Process Model
General measure. Witte and colleagues developed the Risk Behaviour Diagnosis Scale (Witte et al., 1996) which was designed to measure the components of the EPPM. The perceived severity scale has three items which can be used to measure the severity associated with any health threat. One limitation with this measure is that the items do not assess how severe the health threat would be for the individual. Another potential limitation is that the items are generic rather than disease-specific, and the lack of content specificity may reduce the predictive utility of the scale. As with all the other measures listed here, no behavioural context is specified.

  1. ‘I believe that [health threat] is severe’
  2. ‘I believe that [health threat] is serious’
  3. ‘I believe that [health threat] is significant’

Five item response scale: strongly disagree to strongly agree.

Cronbach's alpha= 0.90. This value was observed using a random sample of 179 women at a large university in the Midwestern United States. The majority were aged 17 to 22 (91%) and white (78.5%).

Witte's guidelines as to how to develop a measure of severity using the example of HIV.
Question 1. What happens when you get infected with HIV?
Probe 1. Are you concerned about getting HIV? Why or why not?
Probe 2. Is there anything about your getting HIV that would scare you? (dying, children as orphans, pain, etc.)

Question 2. Do you think HIV eventually leads to AIDS? Why or why not? (People in the country seem to think that no one dies from AIDS and/or it's not a serious threat.)
Probe 1. What do your friends or family think?

Question 3. What happens when you get the disease AIDS?
Probe 1. Are you concerned about getting AIDS? Why or why not?
Probe 2. Is there anything about your getting AIDS that would scare you? (dying, children as orphans, pain, etc.)

Self-Regulation Model - The Illness Perceptions Questionnaire
Revised Illness Perceptions Questionnaire (IPQ-R) (Moss-Morris et al., 2002). The Illness Perceptions Questionnaire was developed to measure patients' representations of their illness. A strength of this measure is the use of self-referencing. Although there are a number of disease-specific versions of the Illness Perceptions Questionnaire (e.g. for asthma, diabetes, chronic pain) the section on severity (labelled 'consequences' in the IPQ-R) is actually the same across all the different disease-specific measures. What varies is simply the insertion of the specific illness in the question. Hence the items are generic and may lack the content specificity researchers may desire.

Another limitation with some of the items in the IPQ-R is that they are ambiguous with respect to valence. Items such as ‘My illness strongly affects the ways others see me’ does not make it clear whether these changes are positive (e.g. causes people to be sympathetic or view the individual as strong) or negative.

In addition, no behavioural context is specified and this may be important in situations where this scale is used to understand adherence to treatment. In other words, beliefs about how severe the illness would be if individuals did not adhere to their treatment may be more informative than general severity items.

  1. My [name of illness] is a serious condition
  2. My [name of illness] has major consequences on my life
  3. My [name of illness] does not have much effect on my life (reverse scored)
  4. My [name of illness] strongly affects the way others see me
  5. My [name of illness] has serious financial consequences
  6. My [name of illness] causes difficulties for those who are close to me

Five item response scale: strongly disagree, disagree, neither agree nor disagree, agree, strongly agree.

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 were recruited from hospital clinics.

Brief illness perceptions questionnaire (Broadbent et al., 2006):

  1. How much does your illness affect your life?

Eleven point response scale: 0-10 anchored ‘no affect at all’ and ‘severely affects my life’.

Revised Illness Perceptions Questionnaire for healthy people (IPQ-RH) (Figueiras and Alves, 2007). The IPQ was initially designed to assess the illness perceptions of patients rather than healthy individuals but recently a version of the IPQ has been developed to assess illness perceptions in healthy individuals.

Unlike the IPQ-R, however, the items are not self-referenced and it may be that this measure assesses illness stereotypes rather than individual severity perceptions. In addition, item 2) is ambiguous with respect to valence.

  1. This illness has serious financial consequences
  2. This illness strongly affects the way the patient sees himself as a person
  3. This illness causes difficulties to those close to the patient
  4. This illness is very serious

Five item response scale: strongly disagree, disagree, neither agree nor disagree, agree, strongly agree.

Cronbach's alpha = 0.66. This value was observed on a sample of 1113 members of the general population recruited from different working environments. They were aged 18-65 and the majority were female (64%).

Test-retest reliability across an interval of 3 weeks = 0.54 and this value was observed using a convenience sample of 157 undergraduate students as part of the initial pilot phase of developing the questionnaire.

<< Previous
Next >>

Search | Help | Contact Us | Accessibility | Privacy Policy

DCCPSNational Cancer Institute Department of Health and Human Services National Institutes of Health

Health Behavior Constructs: Theory, Measurement, & Research