Questionnaires assessing patient knowledge and behavior. The knowledge assessment consisted of 18 statements. Patients were instructed to indicate whether each statement was correct or incorrect. For each patient, the knowledge assessment was scored by calculating the percentage of correct responses among the answered items. Only patients who responded to at least 75% of the statements received a score. The behavioral assessment consisted of 17 items. Patients were instructed to indicate their level of compliance using a 5-point Likert scale (1 indicates all of the time; 2, most of the time; 3, some of the time; 4, rarely; and 5, never). For each patient, the behavioral assessment was scored by calculating the average response across the answered items. A lower score indicates better compliance with behavior. Only patients who responded to at least 75% of the items received a score.
Patient enrollment and participation. Three-month follow-up indicates the findings of the questionnaires returned at a median of 3.3 months. Ten-month follow-up indicates the findings of the questionnaires returned at a median of 10.2 months. Patients who had died or were affected by mailing error were not included in calculating the 3-month response rate.
Clowers-Webb HE, Christenson LJ, Phillips PK, Roenigk RK, Nguyen TH, Weaver AL, Otley CC. Educational Outcomes Regarding Skin Cancer in Organ Transplant RecipientsRandomized Intervention of Intensive vs Standard Education. Arch Dermatol. 2006;142(6):712-718. doi:10.1001/archderm.142.6.712
To determine whether an intensive educational program focused on the risk of skin cancer in organ transplant recipients, a population at high risk for development of skin cancer because of immunosuppression, produced measurable improvement in patient knowledge and sun-protective behavior.
Patients were randomly assigned to receive standard episode-of-care–based education or intensive repetitive written education about skin cancer after organ transplantation. Preintervention knowledge was assessed and documented through a self-administered educational assessment tool. Retention of knowledge and the effect on sun-protective behavior were assessed with a follow-up questionnaire at 3 and 10 months.
Transplant center of an academic medical center.
Two hundred two patients presenting for transplant dermatologic consultation.
Randomized intensive, repetitive written educational reinforcement.
Main Outcome Measures
Retention of knowledge and the effect on sun-protective behavior were assessed with a follow-up questionnaire at 3 and 10 months.
Both intervention groups had similarly high baseline and 3- and 10-month scores on the knowledge portion of the surveys, and they had similar scores on the behavioral assessment portion of the surveys at baseline. Subjects receiving intensive education scored significantly better on the behavioral assessment at 3 and 10 months, although an improvement in knowledge was not documented.
This cohort of transplant recipients was well educated about skin cancer prevention before educational intervention and retained this knowledge. Patients who received the intensive educational intervention were significantly more compliant with recommendations for sun-protective behavior than those who received standard education, although differences in knowledge were not apparent. Lack of time and hassle were the most commonly cited barriers to behavioral compliance with sun protection.
Melanoma and nonmelanoma skin cancers are the most common cancers, accounting for more than 50% of all cancers. Currently in the United States, more than 1 million nonmelanoma skin cancers are diagnosed annually,1,2 and the incidence of melanoma is increasing faster than that of any other cancer.3 Despite the risk of skin cancer inherent with sun exposure, a large percentage of white adults in the United States do not use sun avoidance and protective measures.4 Considerable resources are devoted to preventive efforts, including education. Education appears to affect patient behavior in some instances.5 However, questions remain about the efficacy of specific educational interventions in increasing knowledge and sun-protective behavior.
A patient population in which preventive education is particularly important is solid-organ transplant recipients. The most common malignancies occurring in immunosuppressed patients are skin cancers. Kidney transplant recipients have a greater number of more aggressive skin cancers than do immunocompetent patients, leading to considerable morbidity and mortality.6,7 Despite the potential effect of skin cancer on the lives of these patients, Cowen and Billingsley8 found that 41% of kidney transplant recipients did not realize they were at increased risk for skin cancer because of immunosuppression.
The objectives of our study were to quantify the effect of an intensive prevention educational program on knowledge and behavior in this high-risk population and to assess whether intensive education produces a measurable improvement compared with standard episode-of-care–based education for these outcomes.
The study was approved by the Mayo Foundation Institutional Review Board, Rochester, Minn, and verbal consent was obtained from all patients. All transplant recipients presenting for dermatologic consultation in the von Liebig Transplant Center at Mayo Clinic were eligible and given a standard cover letter requesting their participation and assuring confidentiality. After verbal consent was obtained, all patients were given a comprehensive baseline questionnaire assessing knowledge, behavior, intent to change behavior, and perceived barriers to changing behavior (eFigure). Baseline information, including demographic data, transplantation history, Fitzpatrick skin type, and history of skin cancer, was collected by the recruiting physicians or nurses.
A laminated, pocket-sized copy of standardized verbal education guidelines was given to all recruiting physicians to ensure coverage of essential points. Patients were informed of their increased risk for and potential morbidity owing to skin cancer. An individualized risk assessment was performed, and patient-specific risk factors were discussed. Patients were instructed to use sunscreen with a sun protection factor of 15 or greater on all exposed skin daily for all activities and even for short periods of sun exposure regardless of the weather. They were instructed to wear protective clothing and hats and to avoid times of peak UV light year-round. They were asked to stop intentional tanning outdoors or indoors by means of tanning lamps or at salons. Monthly skin self-examinations were recommended, with any changes reported promptly to their physician. They were given a copy of the Mayo Clinic pamphlet Skin Cancer and Organ Transplant Recipients, which includes this information in more detail, reviews additional risk factors for skin cancer (ie, fair skin, personal or family history of skin cancer, and past exposure), stresses the need for general sun protection (especially sunscreen use and reapplication), and describes and illustrates the appearance of skin cancers. Follow-up examination was based on standard practice and was not different between the intervention and control groups.
As patients were recruited, they were assigned randomly to receive routine episode-of-care–based standardized verbal and written education about skin cancer (standard group) or intensive repetitive written education in addition to the routine education (intensive group). The randomization schedule was generated using a block randomization scheme and stratified according to history of skin cancer. No patient was denied education.
At 2, 6, and 9 months after recruitment, patients in the intensive group were sent a cover letter encouraging careful review of the enclosed pamphlets (at 2 months, the Skin Cancer Foundation pamphlets Simple Steps to Sun Safety and Skin Cancer: If You Can Spot It, You Can Stop It; at 6 months, the American Academy of Dermatology pamphlet Skin Cancer—An Undeclared Epidemic and leaflet Stop—Look for Danger Signs in Pigmented Lesions of the Skin; and at 9 months, the Mayo Clinic pamphlet Skin Cancer and Organ Transplant Recipients). Three months after recruitment, the intensive and standard groups received a standardized letter from the dermatologist (C.C.O.) and a follow-up questionnaire (online-only eFigure, assessing knowledge, behavior, the intent to change behavior, and perceived barriers to behavioral change. Ten months after recruitment, patients in both groups were again sent the follow-up questionnaire, which completed the study. The follow-up questionnaires were mailed at the scheduled times and, if no response was received, 1 month later.
The knowledge assessment consisted of the 18 statements listed in the eFigure. Patients were asked to indicate whether each statement was correct or incorrect. For each patient, the knowledge assessment was scored by calculating the percentage of correct responses to the answered items. Only patients who responded to at least 75% of the statements received a score. In addition, as a secondary analysis, the knowledge assessment was recorded by assuming that the unanswered items were incorrect.
The behavioral assessment consisted of the first 17 behavior items listed in the eFigure. Patients were asked to use a 5-point scale (1 indicates all of the time; 2, most of the time; 3, some of the time; 4, rarely; and 5, never) to indicate their level of compliance. The Cronbach α coefficient of reliability was acceptable at 0.85, and ranged from 0.821 to 0.854 after deleting each of the 17 items in turn one at a time. There also were items to collect additional detail on the level of behavior. The behavioral assessment was scored by calculating the average response across the answered items for each patient (item 16 was reverse coded before the average was calculated). A lower score indicated better compliance with behavior. Only patients who responded to at least 75% of the items received a score.
Responder bias for the 3- and 10-month questionnaires was assessed by comparing baseline characteristics (sex, education, history of skin cancer, age, and baseline questionnaire responses) between patients who did and did not return the questionnaires. Comparisons between the groups (responders vs nonresponders) were made with the χ2 test for nominal variables and the Wilcoxon rank sum test or the 2-sample t test for continuously scaled variables.
One objective of the study was to compare the intervention groups at each of the 2 assessment periods (3 and 10 months). Changes over time within each treatment group were not part of the initial analysis plan because of the seasonal variation that would occur after the randomization phase (a 5-month period) and the effect that season might have on compliance with protective behaviors. The knowledge assessment scores were compared between the intensive and standard groups using the Wilcoxon rank sum test, and the behavioral assessment scores were compared using a 2-sample t test. Although not specified in the trial protocol, analysis of covariance models were also fitted to assess for an intervention effect at 3 and 10 months; this method has been reported to better account for imbalances at baseline because of regression to the mean. However, only the results of the aforementioned analyses are reported because the results from the analysis of covariance models were similar. The analysis was based on an intent-to-treat principle; thus, patients randomly assigned to the intensive group were assumed to have read the material mailed to them. All calculated values were 2-sided, and P<.05 was considered statistically significant.
The Figure summarizes patient enrollment and participation. During the 4-month enrollment period, 202 patients were enrolled. The baseline characteristics of the patients randomly assigned to the intensive group (n = 101) and the standard intervention group (n = 101) are compared in Table 1; there were no significant differences between the 2 groups. Liver transplant recipients were overrepresented relative to their absolute frequency among transplant recipients because the liver transplant recipients are routinely scheduled for follow-up in dermatology at Mayo Clinic.
At 3 months, the response rate was 70% and 69% for the intensive and standard groups, respectively. Responder bias at 3 months was assessed by comparing the baseline characteristics of sex, education, history of skin cancer, and age and baseline questionnaire responses between patients who did and did not return the questionnaires. Within both intervention groups, older patients (the intensive group had a median difference of 5 years in age [P = .08]; the standard group, 7 years [P = .08]) and those who scored higher on the knowledge assessment at baseline (the intensive group had a median difference of 5.5% [P = .007]; the standard group, 0.7% [P = .02]) were more likely to participate at 3 months. The response rate was slightly higher at 10 months, with 74% for both groups. At 10 months, in both intervention groups, older patients again were significantly more likely to participate (median differences in age, 6 years [P = .03] and 13 years [P = .007] for the intensive and standard groups, respectively). In the intensive group, patients who scored better on the behavioral assessment at baseline were more likely to participate at 3 months (mean difference, 0.3 points [P = .02]) and 10 months (mean difference, 0.3 points [P = .03]).
The results of the knowledge assessment are summarized in Table 2. Both intervention groups had similarly high scores on the knowledge assessment portion of the surveys at baseline and at 3 and 10 months. These scores may reflect greater awareness of the risk of skin cancer by transplant recipients at our institution. The distribution of the knowledge assessment scores was highly skewed. In particular, most of the patients had 90% to 100% correct by having all of the statements correct or only 1 incorrect. Therefore, to provide a more meaningful description of the scores, the scores were categorized and summarized as 90% to 100%, 80% to 89%, 70% to 79%, and less than 70% correct. At baseline, 114 (63%) of the 181 patients (62% [58/93] of the intensive group and 64% [56/88] of the standard group) had 90% to 100% correct answers. The scores for the 3-month survey were slightly higher but still similar for the 2 groups (P = .66 based on scores derived from the answered items only; P = .66 based on scores derived from assuming that unanswered items were incorrect). Similarly, the scores were not significantly different between the 2 groups for the 10-month survey (P = .50 based on scores derived from the answered items only; P = .30 based on scores derived from assuming that unanswered items were incorrect). Although 58 patients (83%) of the intensive group had 90% to 100% correct answers compared with 53 (75%) of the standard group, 7 (10%) of the intensive group had 80% to 89% correct answers compared with 15 (21%) of the standard group.
Patients also were asked whether they had ever received verbal or written information about skin cancer after organ transplantation and before enrollment in our study. On the baseline survey, 26% of the patients (23 [25%] of 92 in the intensive group and 25 [27%] of 91 in the standard group) indicated that they had not received information previously, although knowledge scores were high. At 3 months, 16% of patients (6 [10%] of 62 in the intensive group and 14 [21%] of 66 in the standard group) still indicated that they had not received this information, despite the educational interventions of the study. At 10 months, 10% of the patients (7 [11%] of 66 in the intensive group and 7 [10%] of 71 in the standard group) still indicated that they had never received information about skin cancer after transplantation.
The results of the behavioral assessment are summarized in Table 3. Both groups had similar scores on the behavioral assessment portion of the surveys at baseline. However, for the 3-month survey, the behavior score was significantly better (ie, a lower score) for patients in the intensive group (median, 2.4 for the intensive group vs 2.8 for the standard group [P = .006]). This difference was noted also in the 10-month survey (median, 2.4 for the intensive group vs 2.8 for the standard group [P = .007]).
At baseline, 47 (26%) of the 182 patients (27 [29%] of 93 in the intensive group and 20 [22%] of 89 in the standard group) reported that they were already practicing sun avoidance and protection as recommended. In addition, 98 (54%) (52 [56%] in the intensive group and 46 [52%] in the standard group) reported that they intended to implement additional sun-protective behaviors in the next 6 months. At 3 months, 27 (43%) of 63 patients in the intensive group reported current compliance compared with 22 (32%) of 68 in the standard group. In both groups, 64 (49%) of the 131 patients reported that they intended to increase compliance. At 10 months, 25 (39%) of 64 patients in the intensive group and 31 (44%) of 71 in the standard group reported current compliance. This may reflect the fact that patients who returned questionnaires at 10 months were not identical to those who returned them at 3 months. Thirty-two (50%) of 64 patients in the intensive group and 28 (39%) of 71 in the standard group reported that they intended to increase compliance.
At baseline, 71 (39%) of 181 patients (34 [38%] of 90 in the intensive group and 37 [41%] of 91 in the standard group) reported that they already performed skin self-examinations monthly, and 90 (50%) (46 [51%] in the intensive group and 44 [48%] in the standard group) indicated that they intended to start performing skin self-examinations in the next 6 months. At 3 months, 30 (48%) of 62 patients in the intensive group and 43 (63%) of 68 in the standard group reported current compliance, and 25 (40%) and 20 (29%), respectively, intended to start skin self-examinations. At 10 months, 37 (58%) of 64 patients in the intensive group and 46 (68%) of 68 in the standard group reported current compliance, and 24 (37%) and 18 (26%), respectively, intended to start.
Patients cited “hassle” and “lack of time” as the most common barriers that discouraged them from practicing sun avoidance, protective measures, and skin self-examinations. The barriers that were selected at baseline are listed in Table 1; some patients indicated more than 1 barrier and some patients did not indicate any.
Several skin cancer education studies that focused on schoolchildren have shown an increase in knowledge about skin cancer prevention but little or no increase in sun-protective behavior.9- 11 The SunSafe Project, a randomized, controlled study conducted in New Hampshire, found an increase in knowledge and behavioral change among children exposed to a multifaceted educational intervention, as well as a noticeable change in the sun-protective policy of the participating institutions.12,13 Many past studies involving children have lacked control groups. Of the randomized, controlled studies, most have compared groups that received no educational intervention with groups that received multiple forms of educational intervention. The Kidskin intervention, conducted in Australia, compared a control group with moderate- and high-intervention groups that received different levels and types of educational interventions. Knowledge was not assessed, but a modest increase in sun-protective behavior, including increased time spent in the shade, decreased time spent outdoors during peak hours, and use of swimsuits that covered the trunk, was found in the intervention groups, especially the high-intervention group.14,15
Several studies that focused on skin cancer prevention education for adults have targeted outdoor workers, a high-risk population with increased exposure to UV light. Girgis et al5 conducted a randomized, controlled 1-month study in outdoor workers. Educational intervention included a lecture and pamphlets. A significant (P<.05) increase in knowledge, measured by preintervention and postintervention questionnaires, was demonstrated for the intervention group. The level of sun-protective behavior was assessed through self-reporting entries in a diary. Researchers compared their observations of the participants' behavior with the self-reported behavior and found the self-reports to be accurate. Girgis et al5 also noted that surveys can measure intent to change rather than actual change. However, those researchers did not address the likelihood that observation or self-report influenced behavior.
In addition to the method of delivery of educational interventions, the style of delivery may also be important. A prevention campaign in Hawaii compared a survey of behaviors before and 4 months after a skin cancer comic book was distributed.16 The study showed an increase in sun avoidance and protective measures that was more substantial than that shown by other studies.16
Several skin cancer education studies in adults have focused on patients with a history of skin cancer. Robinson and Rademaker17 found an increase in knowledge and a change in behavior among helpers of patients who needed postoperative wound care after skin cancer surgery. This knowledge was assessed by questionnaire before the surgical procedure and 1 year after the procedure. Verbal and written educational interventions were directed to the patient 2 and 6 months postoperatively. The patients were instructed to share this information with their helpers. The increase in sun-protective behavior was more substantial than in earlier studies. However, a history of skin cancer among helpers was not reported, and the patients' history of skin cancer was not considered in the statistical analysis.
In another study by Robinson,18 patients who presented for surgical treatment of skin cancer were surveyed about sun-protective behavior. This was followed by a focused written and verbal educational intervention that was repeated 2 weeks and then 6 months postoperatively. Protective behaviors were reassessed annually thereafter for 2 to 6 years. At the annual follow-up visits, patients received written recommendations for preventive behavior. The frequency of physician visits and the development of subsequent nonmelanoma skin cancers did not influence behavioral change. Although many patients stopped tanning, an increasing compensatory behavior of using sunscreen to obtain a “safe” tan was noted, and several patients switched sun protection strategies to maintain their lifestyle. Patients who had restricted their outdoor activities earlier in the study increasingly used sunscreen and protective clothing to resume outdoor activities. It was hypothesized that giving patients a choice of several sun-protective behaviors increased their self-efficacy and supported their ability to achieve and maintain a change in behavior. The greatest change in behavior occurred immediately postoperatively and decreased afterward; thus, the procedure may have served as a negative reinforcement leading to improved sun-protective behavior rather than the educational intervention causing a behavioral change.
In yet another study of sun-protective behavior after surgical treatment of nonmelanoma skin cancer, Robinson19 noted that a small increase in protective behavior still constituted a high noncompliance rate, but when the results were analyzed according to the transtheoretical model of behavioral change,20,21 the results were more positive. According to the transtheoretical model of behavioral change, evaluating the process of change based on the presence or absence of a behavior is overly simplistic, although easier to measure. Change occurs in stages referred to as precontemplation, contemplation, preparation, action, and maintenance. In the precontemplation stage, there is no intention to change behavior. In the contemplation stage, patients seriously consider a change in behavior in the next 6 months. Progress through the stages is influenced by independent variables such as personal life events, monetary resources, and social support and by dependent cognitive variables. Self-efficacy and decisional balance (pros and cons) also have an effect on progress through the stages. Change is defined as a dynamic process, with subjects moving between stages in a nonlinear fashion with expected relapses or failure to maintain changed behavior. Thus, meaningful progress can be made without sustaining a change in the behavior. Although this approach is not commonly reported, physicians intuitively operate within this theoretical framework when they individually tailor patients' education and counseling. This theory of change also supports the idea that continued educational interventions at multiple times increase the likelihood of behavioral change and maintenance. Although preventive education efforts may produce little change in behavior, physicians are obliged to some extent to provide this service regardless of efficacy.
A substantial portion of patients in the intensive and standard groups intended to increase sun-protective behavior and perform skin self-examinations within 6 months after the survey (contemplation stage). Many already claimed to be compliant. Throughout the study, a substantial proportion of patients did not intend to improve compliance (precontemplation stage). There was a shift from contemplation to compliance over time. However, because of the complex and nonlinear nature of change, it is difficult to know whether any important change occurred. There may have been evidence of “compliance fatigue,” as manifested by the decrease in compliance with skin self-examination from the 3- to 10-month assessment.
In summary, this study is, to our knowledge, the first carefully designed, randomized, controlled comparison of the effects of different levels and types of educational interventions during a 10-month period in a high-risk adult population without recent negative reinforcement. This cohort is particularly well suited for long-term follow-up. Throughout the study, patients in the intensive and standard intervention groups were found to be highly knowledgeable. All transplant recipients treated at Mayo Clinic currently receive treatment in a comprehensive multidisciplinary transplant clinic where emphasis is placed on prevention and treatment of all transplant-related complications. Transplant coordinators, who have a pivotal role in the care of these patients, have been given information about skin cancer. Educational efforts have also been directed to staff physicians and nurses. A small number of patients claimed that they were unaware of their increased risk for skin cancer despite educational efforts.
The behavior of patients in both the intensive and standard groups changed little during the study; patients protected themselves adequately some of the time. This was disappointing, particularly because of the high probability of skin cancer in this high-risk population. The purpose of preventive education is to change behavior. Our patients showed less compliance than knowledge, and this did not change significantly during the study. Perhaps if the patients had not been so well informed before the educational intervention, there would have been more opportunity for improvement. However, if knowledge does not translate into appropriate behavior, prohibitive barriers surely exist. For the most part, these barriers are not variables that physicians can influence, and changing barriers was not a goal of our study. The barriers most frequently reported by the patients were the discomfort, unpleasantness, and hassle of compliance and the lack of time. Although our findings may not apply to the general population, patients similar to our high-risk patients require special study and consideration so that they may receive the best care. Because of the high-risk status of these patients, follow-up over a relatively short time will allow outcome to be measured in terms of the actual number of skin cancers in the future. Follow-up studies can also address long-term behavioral change. Because of the considerable time and resources allotted to preventive medicine and the potential effect on patients' health, determining which efforts are efficacious is important for health policy planning.22 We acknowledge several shortcomings of our study, including the lack of formal reliability testing to validate the study instrument and the lack of long-term follow-up to correlate changes in behavior with clinically significant end points such as skin cancer development. Long-term follow-up of this population is planned for that reason.
Correspondence: Clark C. Otley, MD, Division of Dermatologic Surgery, Mayo Clinic, 200 First Street SW, Rochester, MN 55905 (firstname.lastname@example.org).
Financial Disclosure: None.
Accepted for Publication: September 26, 2005.
Author Contributions:Study concept and design: Clowers-Webb, Weaver, and Otley. Acquisition of data: Clowers-Webb, Christenson, Phillips, Roenigk, Nguyen, and Otley. Analysis and interpretation of data: Clowers-Webb, Christenson, Roenigk, Weaver, and Otley. Drafting of the manuscript: Clowers-Webb, Nguyen, and Otley. Critical revision of the manuscript for important intellectual content: Clowers-Webb, Christenson, Phillips, Roenigk, Weaver, and Otley. Statistical analysis: Clowers-Webb and Weaver. Obtained funding: Otley. Administrative, technical, and material support: Clowers-Webb, Roenigk, and Otley. Study supervision: Clowers-Webb, Roenigk, Nguyen, and Otley.
Funding/Support: This study was supported by a Clinical Practice Innovation Grant from Mayo Foundation, Rochester, Minn.
Additional Resources: The online-only eFigure is available.