Abstract

This study characterized and compared measures of cervical cancer risk perception among women preparing for colposcopy due to suspected cervical dysplasia. Participants included 110 women recruited from a university-based colposcopy clinic. All were first-time colposcopy patients with no cancer history and mean age of 30. Most were Caucasian, single, employed, uninsured, and high schooleducated. Prior to colposcopy, participants completed a demographic questionnaire, health history, and multiple measures of perceived cervical cancer risk. Most perceived risk measures were positively correlated (p's<.01). Most participants reported believing that they were at risk for cervical cancer (81%), but more than half (59%) reported never having previously considered their risk. Personal risk was grossly overestimated: mean personal risk and average woman's risk estimates were 50%, as compared to national average lifetime risk of less than 1%. Several variables were positively related to ratings of personal risk, including being Caucasian, employed, and having more education (p's<.05). No significant relationships were found between perceived risk and objective cervical cancer risk factors (e.g., number of lifetime sexual partners). In conclusion, reports of perceived cancer risk were remarkably consistent across measures. As in previous research, perceived risk was dramatically overestimated in comparison to known cervical cancer rates. Notably, perceived risk measures were unrelated to presence of objective cervical cancer risk factors. Future research will examine relationships between perceived risk and adherence to post-colposcopy recommendations.

Introduction

Background

• Perceived risk is a construct essential to several theories of health behavior (e.g., the Health Belief Model, Precaution Adoption Process Model, Social Cognitive Theory), and is a commonly used construct in cancer prevention and control.
  
• While there are several methods available for measuring perceived risk, these measures have not been systematically compared.
  
• It is unclear whether the typically used measures of perceived risk actually measure the same aspects of the construct , and how any differences in measures may influence findings.
  
• Perceived risk of cervical cancer is important to understand in women undergoing colposcopy, as it may be related to compliance with follow-up procedures and treatment.
  

Purpose

The primary aim of this study was to (1) compare and contrast the information obtained from several measures of cervical cancer risk perception among women preparing to undergo colposcopy.

In addition, further analyses were planned to:
(2) describe the levels of perceived risk reported by participants;
(3) assess whether participants had previously considered their level of risk for cervical cancer; and
(4) identify factors associated with level of risk perception in this population.

Procedure

Inclusion CriteriaAge ≥ 18 years
First-time colposcopy patient following abnormal Pap smear
English-literate
Able to provide informed consent

Exclusion CriteriaHistory of pregnancy or childbirth in prior 12 months
History of cancer (except non-melanotic skin cancer)

Steps
128 eligible patients approached in University-based colposcopy clinic waiting room
111 (87%) agreed to participate and provided informed consent
1 later requested to be disenrolled:
Total N = 110
Each participant completed a questionnaire packet in private room (RA available to assist if needed)
RA escorted participant to exam room for colposcopy procedure
After colposcopy, participant returned to RA in private room and completed follow-up questionnaire

Measures

Items measuring sociodemographic characteristics

Four measures of risk perception:
(1) Percentage Method (for self and average woman) (PM; Lerman et al., 1995)
(2) Champion’s Susceptibility Scale (CSS; Champion, 1993)
(3) Risk Behavior Diagnosis Scale – Susceptibility Scale (RBDS-SS; Witte et al., 1996)
(4) Relative Risk Item (RRI; e.g., Audrain et al., 1997)

Items measuring objective cervical cancer risk factors, including (a) age at first intercourse, (b) number of lifetime sexual partners, and (c) personal history of sexually transmitted disease

Sample Description (N = 110)
Mean Age (SD) & Range: 30 (10) / 18 – 62

Race:
56% Caucasian
41% African American

Highest grade completed
29% < 12th grade
48% = 12th grade

Marital Status
83% unpartnered

Employment
43% full time
19% part time

Household Income
49% ≤ $10,000
26% $10,001 - $20,000
13% $20,001 - $30,000

Health Insurance
69% none
15% private insurance
12% Medicaid
4% Medicare

Objective Risk Factors
Mean age of first sexual intercourse: 15.5
Median number of lifetime sexual partners: 5
53% positive history of STD

Results

Aim #1: Comparison of Measures

Correlations between measures of perceived cervical cancer risk
PM (average woman) with PM (self): 0.65; p<.01
CSS with PM (self): 0.46; p<.01
CSS with PM (average woman): 0.37; p<.01
CSS with CSS: alpha = .87
RBDS-SS with PM (self): 0.36; p<.01
RBDS-SS with PM (average woman): 0.17
RBDS-SS with CSS: 0.47; p<.01
RBDS-SS with RBDS-SS: alpha = .54
RRI with PM (self): 0.54; p<.01
RRI with PM (average woman): 0.43; p<.01
RRI with CSS: 0.50; p<.01
RRI with RBDS-SS: 0.18

Aim #2: Description of Levels of Perceived Risk

Aim #3: Previous Consideration of Risk

• 59% of the sample reported never having thought about their risk for cervical cancer before
  
• Those who reported having thought about risk before had significantly higher RBDS-SS scores (M = 11.74, SD = 2.39) than those who had not previously considered their risk (M = 10.78, SD = 2.34; t104 = 2.07, p<.05)
  
• Previous consideration of risk was not related to other demographic variables, objective risk factors, or measures of perceived risk
  

Aim #4: Factors Related to Measures of Perceived Risk

Associations with measures of perceived cervical cancer risk

• Uninsured respondents scored higher on the RBDS-SS than insured respondents

• Caucasian respondents scored higher on the CSS and RBDS-SS and were more likely than minority respondents to rate themselves at higher or much higher than average risk on the RRI
  
• Employed respondents scored higher on the CSS and were more likely than unemployed respondents to rate themselves at higher or much higher than average risk on the RRI
  
• No associations were found between other sociodemographic variables or objective risk factors (age at first intercourse, # lifetime sexual partners, STD) and measures of perceived risk

Conclusions

• While most of the measures of perceived cervical cancer risk were significantly positively correlated, they do not appear to be interchangeable and may measure different aspects of this construct.
  
• Risk estimates obtained via the Percentage Method, in particular, reflected gross overestimations of risk. This may be due to innumeracy and/or availability heuristic bias.
  
• Caucasians consistently reported higher risk estimates than minority respondents, across all four measures. Minority respondents were more likely to rate themselves at average or lower than average risk, despite having abnormal Pap smear results.
  
• The lack of association between reported objective risk factors and estimates of cervical cancer risk suggests that respondents may have limited knowledge regarding cervical cancer.
  
• Most respondents had never considered their cervical cancer risk before. Previous consideration of risk was related to only the RBDSSS risk estimate, in which prior thought was related to lower scores.
  

Limitations

• The cross-sectional nature of the data precludes causal inferences.
  
• Possible order effects were not analyzed, as all questionnaires were administered in the same order to all respondents.
  
• Generalizability of results may be limited due to the use of a clinic sample of women presenting for colposcopy.