Evaluation of Factors Associated with Skin Self-Examination¹

The incidence of melanoma has increased dramatically over the past four decades (1).³ It is estimated in 1999 that there will be 44,200 new cases of melanoma diagnosed and 7,300 deaths attributable to this cancer (2). Although the most recent cancer statistics indicate a trend of decreasing incidence and mortality for many types of cancer, this has not been observed for melanoma, and it remains a potentially fatal disease (3). Lesion thickness (Breslow depth) has been identified as the most important prognostic factor for melanoma, with survival inversely related to lesion thickness (4, 5). Early detection and excision of thin lesions may be important in reducing mortality from this cancer. Periodic skin self-examination may be beneficial in identifying thin lesions. In a recent study, skin self-examination was associated with a reduction in the incidence of melanoma as well as the development of advanced disease among melanoma patients (6).

The American Academy of Dermatology has recommended that individuals practice skin self-examinations to detect new and/or changing lesions (7). As well, public awareness and educational programs have been implemented but focus mostly on awareness and warning signs (7, 8, 9). The purpose of this study was to describe characteristics of individuals who perform skin self-examination and to further evaluate factors associated with skin self-examination.

The study population was comprised of 549 Caucasian residents of Connecticut, 18 years of age or older, who were selected as controls as part of a population-based case control study on skin self-examination and melanoma conducted during 1987–1989 (6). We limited our study population to controls to evaluate factors associated with skin self-examination in a group of individuals without melanoma likely to represent the general population. Controls were selected through random-digit dialing and frequency matched by age (20-year categories) and sex to the expected number of melanoma cases. Initially, 814 controls were identified and eligible for the study; 70% (n = 573) agreed to be interviewed. Of these, an additional 24 were not Caucasian and subsequently excluded; thus, 549 controls were interviewed.

Personal interviews were conducted to obtain information on skin self-examination practices, demographics (age, gender, education, and marital status), personal and family history of cancer, phenotypic characteristics (including the ability to tan, freckling, hair, eye and skin color, and number of nevi), sun exposure habits, screening and health behaviors, including knowledge and awareness of melanoma signs and symptoms, and sunscreen use. Melanoma awareness was assessed by asking if the subject thought about their skin, how it looked, or whether there were any changes or abnormal marks. Knowledge was evaluated by asking about signs and symptoms of melanoma characteristics: color, size, shape, bleeding, itching, and scab that does not heal. Nevi counts were performed on the backs and arms of subjects by trained nurse interviewers.

Skin self-examination practices were obtained by asking, “Did you ever (in your life) carefully examine your own skin? By this I mean actually check surfaces of your skin deliberately and purposely?” Information about spouse or other nonphysician skin examination was obtained by asking, “Has someone other than a physician ever carefully examined areas of your skin on purpose?” The subjects who responded positively to either question (self or spouse/other nonphysician examination) were classified as performing skin self-examination. We also specifically asked about the areas of the skin that were examined: chest, stomach, arms, side of body, upper and lower back/shoulders, legs (front, back, or sides), and bottom of feet. This enabled us to further classify individuals as rigorous skin self-examiners if self-examination was performed on all of the sites indicated.

Descriptive statistics were performed to describe characteristics of individuals and their skin self-examination practices. Univariate logistic regression analyses were conducted to examine the relationship between each factor of interest and skin self-examination. Indicator variables were created and entered into the models for both categorical and dichotomous variables. For each variable, crude ORs⁴ and corresponding 95% CIs were calculated. Multivariate logistic regression (10) was used to model the relationship between the variables identified as important by the univariate modeling and skin self-examination. We included in the final model all variables with an OR of at least 2.0 (or 0.50) and/or those for which the CI excluded 1 (P < 0.05). This technique quantifies large effects that may be imprecise, while identifying the most parsimonious list of variables (11). The results describe the independent effects of factors associated with skin self-examination, taking into account both the strength of the association and the precision of the confidence interval (12).

There were 130 study subjects who responded positively to the skin self-examination question (includes self, spouse, or other nonphysician examination; 23.1 and 24.5% for males and females, respectively). Epidemiological characteristics by skin self-examination practices are presented in Table 1. These variables have been grouped according to demographic and phenotypic characteristics, personal and family cancer history, sun exposure habits, and screening and health behaviors.

Age, education, and marital status were identified a priori as important variables and were selected for inclusion in the final multivariate models, regardless of whether they met the specified criteria. We performed all analyses separately for females and males.

We also performed analyses excluding spouse/other nonphysician examination as a form of skin self-examination. In these analyses, we classified subjects as having performed skin self-examination if they responded positively to the question about examining their own skin. The results and final models were similar to the initial classification scheme, and thus we present the results that include self/spouse/other nonphysician examination to increase the precision of the estimates.

Table 2 presents the results from the univariate regression models. For males, family history of cancer, total sun exposure, skin awareness, physician skin examination, change in diet, and sunlamp use were associated with skin self-examination according to our criteria. Phenotypic characteristics were not important predictors of whether a male performed skin self-examination. Use of sunscreen was not associated with a greater likelihood of skin self-examination. Important factors associated with skin self-examination in females included: freckling, hair color, total sun exposure, skin awareness, previous benign biopsy or presence of abnormal mole, gynecological examination and/or physician breast examination, and breast self-examination.

Table 3 presents the adjusted parameter estimates for the final regression models; important variables have been included in the models using the established criteria. For males, family history of cancer, skin awareness, physician examination, and change in diet were associated with skin self-examination. Important variables in females included: skin awareness and previous benign biopsy or presence of abnormal mole. It should be noted that for both males and females, the CIs associated with the adjusted ORs are wide, indicating the reduced precision of these estimates. In women, light hair color may increase the likelihood of performing skin self-examination. Older age and college or postgraduate education were associated with a decreased likelihood of performing skin self-examination in both males and females.

To further assess the relationship between skin self-examination and the variables of interest, we examined a subset of individuals who performed rigorous skin self-examination (defined as those subjects who performed rigorous skin self-examination or who had a rigorous self-examination by a spouse or other nonphysician); 6.0% of males and 7.0% of females met our criteria. The small numbers precluded us from modeling the relationship between the factors of interest and rigorous skin self-examination.

In this study, we evaluated factors associated with skin self-examination. It is important to identify those factors that increase the likelihood of an individual performing skin self-examination because early detection and excision of lesions has the potential to reduce morbidity and mortality from melanoma. It has been suggested in a previous study using these data that skin self-examination reduces mortality from melanoma (6). These data are important for researchers conducting future studies on skin self-examination as well as for clinicians in the practice setting.

Skin awareness was strongly associated with skin self-examination for both females and males. For females, previous benign biopsy or the presence of an abnormal mole was identified as an important predictor of future skin self-examination using our criteria. A family history of cancer, physician examination, and change in diet to reduce cancer risk increased the likelihood of skin self-examination in males but not females. In women, light hair color may increase the likelihood of performing skin self-examination. Older age and college or postgraduate education were associated with a decreased likelihood of performing skin self-examination in both males and females.

This study adds to the existing literature on skin self-examination. The novel aspects of our study include the analytical study design, selection of a population-based study population, in-person interview for obtaining data, and statistical approach for evaluating factors associated with skin self-examination.

A comparison of the studies assessing predictors of skin self-examination is presented in Table 4. Girgis et al. (13) examined the prevalence and predictors of self-screening in 1344 study subjects. Participants were asked whether they regularly checked the nevi on their body for changes and whether they were regularly screened by another person for changes in their nevi. Participants were also asked when a doctor had checked their nevi last, and respondents who were checked by a physician within the last year were considered to have been screened by a general practitioner. Forty-eight % reported self-screening (defined as screening by self or other person). Only 15% of participants reported performing screening by self/other and screening by a general practitioner. Predictors of any type of screening (e.g., self/other/general practitioner) were evaluated. Males were less likely to have been screened for melanoma. Lower education level, unemployment, and having only a basic medical insurance were associated with a reduced likelihood of being screened. Individuals with at least one risk factor for melanoma (defined as personal history of melanoma or skin cancer, family history of melanoma, 11 or more nevi, or changes in nevi during past 12 months) were also more likely to practice screening or be screened by another person or general practitioner.

Miller et al. (14) in a 1995 survey (n = 1001) conducted by the American Academy of Dermatology reported 46% of a nationally representative United States sample conducted skin self-examination (“Do you ever closely examine yourself for signs of skin cancer or melanoma?”). Women, white persons, and the elderly were more likely to conduct skin self-examination as well as those who were knowledgeable of melanoma risk factors and early signs and symptoms. Awareness was also a predictor of skin self-examination.

In a telephone survey (n = 1000) conducted by the American Academy of Dermatology in 1996 (15), 46% of the population indicated they practiced skin self-examination. (The question was asked: “In the past year, have you closely examined yourself for signs of melanoma or skin cancer?”.) Older age, white ethnicity, female gender, and higher education levels were identified as predictors of skin self-examination. A self-perceived risk of developing melanoma and discussions with nurses or physicians about skin cancer prevention were also indicative of an increased likelihood of practicing self-examination.

In a study on cancer beliefs and behaviors conducted in Australia (n = 3527), Hill et al. (16) observed that women were more likely to practice skin self-examination as compared with men. Fifty-one % of women versus 43% of men reported that they had ever done a self-examination when asked, “Have you ever deliberately checked your skin or had a friend or relative check your skin for changes that could mean cancer?” No differences were apparent across age groups. With respect to frequent skin self-examination, 13% indicated that they had examined their skin six or more times in the past year.

In a survey of 995 residents of Queensland, Australia, >60% of the subjects reported that they presently check their skin for early signs of skin cancer (17, 18). Over 76% of respondents reported some form of skin examination (self-examination, examination by other nonmedical persons, or examination by a medical practitioner or specialist). Females and those who were married or living together reported higher rates of skin self-examination behavior as compared with those who were male or single. Age was also positively associated with self-examination. There was an increased likelihood of skin self-examination for those with a higher education.

Cody and Lee (19) observed differences in skin self-examination for university students (n = 312) based on their previous personal or vicarious experience with cancer and skin type. In a cross-sectional study of individuals (n = 384) at increased risk for skin cancer, Friedman et al. (20) evaluated predictors of skin self-examination. Sixty-one % reported that they performed skin self-examination during the past year. Knowledge and other psychological factors (such as optimism and perceived self-efficacy) were positively associated with an increased frequency of skin self-examination. In a recent telephone survey of 200 individuals from Rhode Island (21), a measure of complete or thorough skin self-examination of the body was obtained. Female gender and whether a health care provider asked the participant about skin examination were predictive of a thorough skin self-examination.

We presented the results of previous studies that have examined correlates of skin self-examination. In our data, we observed that younger persons and those with lower educational levels were more likely to perform skin self-examination. This is in contrast to prior studies that reported older age and higher education to be associated with increased skin self-examination. However, not all these studies observed this relationship between age and education. Possible explanations for these differences include the fact that our study was population-based and in a different geographical region. Furthermore, the definition of skin self-examination, method of data collection, and time period when the study was conducted varied among the studies. Response rates also varied significantly between studies, potentially affecting the results. Perhaps most importantly, the method we used to evaluate factors associated with skin self-examination takes into account both the strength of the association and the precision of the CI compared with the majority of other studies that simply relied on statistical significance assessed by Ps.

The major limitation of our study is the potential for misclassification in the reporting of skin self-examination practices and other factors potentially related to skin self-examination (22). We acknowledge that the data used for this study were originally collected as part of a population-based case control study evaluating the relationship between skin self-examination and risk of melanoma (6), and thus, there are inherent limitations. We limited our analyses to include control subjects only, thereby minimizing the potential bias associated with the recall of information by diseased cases. Furthermore, the controls were originally selected to reflect the age and sex distribution of the expected number of melanoma cases; thus, this study population may not accurately represent the general population.

A priori, we identified criteria to identify those variables that would be included in the multivariate models. We were interested in the strength and precision of the parameter estimates and thus included those variables we deemed important based on our criteria. The univariate estimates are intended to be a general list of correlates of skin self-examination that can be used in general practice, whereas the multivariate estimates will allow us to identify those key factors that are associated with skin self-examination. Identifying predictors of skin self-examination will enable health care providers to target individuals who may not be performing skin self-examination but who are at increased risk for developing melanoma.

We thank Drs. W. Douglas Thompson, Carol Portlock, and Jean Bolognia for assistance in carrying out this study; Dorothy Clow, Michelle Miclette; and Sheila Griffin for assistance with data collection; and the registered nurse interviewers: Marjorie Jasmin, Annette Hopkins, Susan Anderson, Karen Wu, Joanne Backes, Alyson Cohen, and Paula Clark for their dedication to this study. We thank the following institutions in Connecticut, without whose assistance this study would not have been possible: Connecticut Dermatopathology Laboratory, Inc., Laboratory of Hope-Ross and Portnoy, University of Connecticut Dermatopathology Laboratory, Yale Dermatopathology Laboratory, Hartford Hospital, Yale-New Haven Hospital, St. Francis Hospital and Medical Center, Bridgeport Hospital, Waterbury Hospital, Hospital of St. Raphael, Danbury Hospital, New Britain General Hospital, Norwalk Hospital, St. Vincent’s Medical Center, The Stamford Hospital, Middlesex Memorial Hospital, Mt. Sinai Hospital, St. Mary’s Hospital, Lawrence and Memorial Hospital, Manchester Memorial Hospital, Greenwich Hospital Association, Veterans Memorial Medical Center, Griffin Hospital, Bristol Hospital, St. Joseph Medical Center, University of Connecticut Health Center/John Dempsey Hospital, William W. Backus Hospital, Park City Hospital, Charlotte Hungerford Hospital, Windham Community Memorial Hospital, Milford Hospital, Day Kimball Hospital, Rockville General Hospital, Bradley Memorial Hospital, The Sharon Hospital, New Milford Hospital, Johnson Memorial Hospital, Winsted Hospital, and Westerly Hospital (Rhode Island). We thank Dr. Martin Weinstock and Dr. Marianne Ulcickas Yood for review and helpful comments on the manuscript.