Leading health organizations debate the efficacy of testicular self-examination (TSE) in reducing TC related mortality [8], particularly with the advent of chemotherapeutic regimens that are highly-effective even for men with advanced disease during the last several decades. However, this decrease in mortality is accompanied by an increase in treatment-related morbidity, particularly for those cancers that are diagnosed at later stages [6,9]. A recent review found little shift of the approximate 5-month delay from onset of TC symptoms to surgical diagnosis over the past seventy-five years [10]. While TSE may not markedly improve the already high survival rates related to TC, early diagnosis does bring significant potential to reduce treatment-related morbidity, since surgery plus close surveillance is a viable management option for many early-stage patients [10]. Education and instruction for men on the normal shape and texture of the testicles, plus information regarding signs and symptoms associated with TC could be a critical component in reducing treatment delay [11,12].

Additionally, as epidemiological research identifies specific TC risk factors, some organizations have refined their screening recommendations to focus on high-risk groups. Established TC risk factors include cryptorchidism (undescended testes), prior contralateral TC and testicular intra-epithelial neoplasia. Testicular cancer risk is also increased in certain genetic conditions, such as Klinefelter and androgen insensitivity syndromes. Other conditions that may increase TC risk include urogenital malformations, testicular dysgenesis, testicular atrophy, history of infertility, and, possibly, exposures to a variety of etiologic agents during key developmental periods [13]. Perhaps most importantly, family history is known to be an important independent risk factor. Approximately 1–3% of men with testicular cancer report the presence of another affected male in their family. Sons of men with TC have a four- to six-fold increase in risk of TC, and brothers carry an eight- to ten-fold increase in the risk when compared with the general population [14-16].

The American Cancer Society recommends that men with known risk factors such as family history seriously consider performing TSE regularly. This approach to modifying general population early cancer detection practices for high-risk populations is not unique to men with a familial risk for TC. For example, the National Comprehensive Cancer Network recommends routine breast self-examination (BSE) for women at increased risk of hereditary breast and ovarian cancer (HBOC) [17], despite randomized controlled trial data refuting the efficacy of BSE in reducing breast cancer mortality [18]. In addition to the possibility that TSE recommendations for self-examination may reduce chemotherapy-related morbidity by increasing the proportion of early-stage TC diagnosed in high-risk men, there may also be further benefit in providing at-risk individuals with a management strategy that involves them directly in their own care, and gives them an enhanced sense of control over their lives [10].

While TSE practices among general population young males have been extensively studied [2,19-26], there are no published data specific to TSE practices among multiple-case family members.

In a 2005 review of TSE practices that included 6 studies of U.S. men ≥ age 18 [39], the percent of men practicing TSE on a monthly basis ranged from 6 to 36% Our TSE rates are not directly comparable to those data, which defined regular performance of TSE as once a month, since we used a broader inclusion for our regular performer category of 6 times a year or more. This may explain the higher proportion of men in our sample that were defined as regular performers of TSE (46%).

With respect to the HBM variables assessed in the current study, only Cues to Action (i.e. physician recommendation for TSE), remained a statistically significant predictor of TSE behavior in multivariate analysis. The importance of physician recommendation on TSE practice was striking, with men who reported a physician recommendation for TSE having at least six times higher odds of performing regular TSE compared with those who reported no such recommendation. This provides yet another example [33,40,41] of how potent an influence the physician can be with regard to cancer prevention behavior in patients, and underscores the importance of physicians including such recommendations in the course of ongoing health care discussions. Due to the lack of a strong evidence base to support the belief that TSE is associated with improved TC-related survival among those who practice it, and the remarkable therapeutic successes in managing TC patients even with far-advanced disease, it is unlikely that there ever will be a clinical trial testing this question. In this era of evidence-based medical practice, it is difficult for health care providers to insist that at-risk men adopt TSE practice, although common sense and logic suggest that a TC detected by TSE rather than symptoms should be more amenable to treatment regimens that do not include chemotherapy, or permit less intense or shorter chemotherapy regimens. This rationale is not unlike recommendations for both BSE, CA-125 testing, and transvaginal ultrasound recommendations for women at increased risk for hereditary breast and ovarian cancer [42]. Given that TC is a disease which affects young men during the most productive period of their lives, there is potential for real economic and psychosocial benefit in attempting to minimize treatment-associated costs and morbidity.

Two previous studies of the HBM and TSE showed that perceived benefits and barriers were associated with TSE, as did our data [19,29]. In our study, perceived cancer susceptibility was not associated with TSE. Although one previous study examining the relationship between TSE and perceived susceptibility found a significant association, there were only 12 men in the entire study who were considered to be "practicers" of TSE (i.e. reported practicing TSE > 4 times a year) [19,29]. A larger study of TSE behaviors using multivariate analyses failed to replicate this finding [29]. Perceived severity of TGCT was not a significant predictor of PSA testing. This finding is similar to two previous studies of TSE that used the HBM as a theoretical framework [19,29]. Previous cancer screening research for a variety of sites has also found perceived susceptibility of little predictive value with regard to screening behaviors, possibly due to the almost universally held belief that cancer is a severe disease, regardless of type [43-48]. In our study, none of the HBM factors significant in bivariate analyses remained so when they were considered in a multivariate model which included physician recommendation. This result is consistent with many previous studies of cancer screening behaviors.

To our knowledge only one other study has examined HBM with respect to TSE using multivariate analysis. Overall, it appears that the HBM does not adequately explain TSE behavior. Similar to our study, HBM variables accounted for only a small proportion of the variance in TSE behavior (21%) [29]. As discussed below, cancer-related worry was the main psychosocial predictor of TSE behavior in our study and may lend some support to the commonly-raised concern regarding the HBM and other cognitive models of health behavior, namely that they do not include affective variables as behavior predictors [30].

Cancer-related worry was the other major factor associated with TSE practice patterns in our study. From a theoretical perspective, several relationships between cancer worry and cancer prevention behaviors have been proposed. Some models posit that worry may serve as a facilitator to cancer screening behavior, others postulate worry acts a barrier, and some propose a curvilinear relationship between cancer worry and preventive behaviors, whereby worry is a facilitator to a certain level after which it becomes a barrier [49]. The HBM, which was the theoretical basis of the present study, has been used to support the role of worry as both a barrier and facilitator for cancer screening behavior [49]. In our sample, it appears that cancer worry motivates individuals to practice TSE, since those with higher levels of cancer worry were also more likely to perform TSE. However, the directionality of this relationship can only be established in the context of a prospective study; our data are cross-sectional. Studies in other high-risk populations provide some evidence that cancer worry may in fact precede screening behavior. In a review of the relationship between cancer worry and mammography screening among high-risk women, four of five prospective studies indicated that higher levels of cancer worry were associated with greater rates mammography screening and breast self-examination [49]. However, no similar prospective studies of high-risk men are available. One cross-sectional study of prostate cancer screening among a sample of men at increased risk of hereditary prostate cancer found a negative relationship between cancer worry and screening behavior [50].

It is noteworthy that we observed no statistically significant differences in TSE practices between men with and without a prior history of a testicular abnormality. A personal history of cryptorchidism can increase the risk of TC up to 11-fold [51]. Of the 13 men reporting a testicular abnormality in our series, 29% (n = 4) reported a history of cryptorchidism (data not shown). In a study by Blesch and colleagues, none of the 5 men with a history of cryptorchidism, reported regularly practicing TSE [19]. While this finding may reflect the small number of men with this particular risk factor and the associated lack of statistical power to detect a difference in that study, it is possible that men may be unaware that testicular maldescent is among the most strongly-established TC risk factors. However, no study to date has focused specifically on the levels of awareness/knowledge of cryptorchidism as a risk factor for TC or as a multivariate predictor of TSE behavior.

In multivariate analyses, men age = 35 and those who were first-degree relatives of a TC case were no more likely to perform TSE than were older men or those whose affected family member was a more distant relative. In the 2005 review mentioned above [27], only 12–25% of participants in U.S. studies were aware of the age group most affected by TC. In the study by Blesch and colleagues [19], men with a family history of TC were significantly more likely to perceive greater benefits to TSE. However, the study did not assess the impact on actual TSE behavior. Taken together, the non-differential rates of TSE between men in younger and older age groups, men who are first versus second-degree relatives of a TC case, and men with a prior testicular abnormality compared with those without, suggest a lack of awareness on the part of unaffected family members regarding the additional risk conferred by having these characteristics. It is also possible that men recognize the implications of these characteristics, but they are not sufficiently concerned by this information to motivate screening behavior.

To our knowledge, this is the first study to examine TSE behaviors among unaffected men from multiple-case testicular cancer families. While the results provide important information about current rates and correlates of TSE in this high-risk group, they should be interpreted with caution due to certain limitations of our study. A goal of our study was to evaluate the utility of HBM in understanding TSE behavior. Ideally, use of an analytic technique such as structural equation modeling would allow us to definitively test the utility of HBM in predicting TSE behavior. However, our study was designed as a pilot study of the utility of HBM and it did not have a sample size sufficient to reach definitive conclusions. It is possible that the HBM did not adequately capture those factors that were important in TSE behavior. For example, one of the major criticism of HBM and other cognitive models of health behavior is the lack of inclusion of affective variables (e.g. distress) [30]. To address this issue, we included scales for both cancer worry and cancer distress. Additionally, we included demographic and SES variables that may also impact screening behavior. Finally, in a study that compared the HBM to the Theory of Planned Behavior, the quality of the models was very similar. The TPB explained 22%of the variance in behavior versus 21% by the HBM, suggesting that applying another theoretical framework may have yielded similar results [29]. Thus, based on our review of the literature related to theories of health behavior and predictors of screening, we feel that we adequately expanded the HBM to include other factors associated with TSE behavior. In addition, this was a cross-sectional study, making it difficult to establish that the beliefs and psychological factors we identified preceded the behavior, as hypothesized in the HBM. Additionally, men volunteering to participate in a study of familial TC may be more aware of and concerned about TC. Thus, the rates of TSE may be higher in this study population than men with a similar history not participating in such a study. Similarly, these men may be more knowledgeable about TC and have differing beliefs and psychological responses to TC.

RPM conducted all statistical analyses and participated in manuscript preparation

JL participated in design of original sub study protocol and manuscript preparation

JAP participated in design of original sub study protocol and manuscript preparation

MHG conceived and designed original main research protocol and participated in manuscript preparation

LK oversaw data collection and participated in manuscript preparation