Abstract
Testicular cancer is one of the greatest threats to health and wellness among 15- to 40-year old males. A concerted effort in the literature promoting awareness, risk factors, and preventative measures is warranted. There is limited discussion on the validity of interventions aimed at promoting testicular self-examination (TSE) performance; the existing body of evidence offers little discussion on what specific factors motivate performance. To assist in making Healthy People 2020 an all-inclusive success, a comprehensive assessment of existing evidence is necessary to assist in closing this research gap. A systematic review of interventions promoting TSE performance discovered moderate levels of effectiveness among 10 studies promoting the behavior. Concerning methodological quality, nine were of average quality and one was of high quality. In terms of significant TSE reporting between intervention and control/comparison groups, 3 out of 10 did not achieve the statistical causal threshold. Based on our assessment of TSE intervention quality and outcomes pertaining to behavior adoption, a best-practices guideline is presented for researchers in the field to consult as they design their interventions. This guideline aims to improve on internal and external validity of TSE promotion research in order to make them more effective.
Background and Objectives
Overview
Testicular cancer is the most prevalent form of cancer among males aged 15 to 40 years (Giannandrea et al., 2013). Identified risk factors of the disease include cryptorchidism (Ferguson & Agoulnik, 2013), abnormalities of the genitourinary system (Leone, Maurer-Starks, & Williamson, 2011), and endocrine abnormalities/dysfunction (Hayes-Lattin & Nichols, 2009), among others. Testicular cancer incidence is low compared with other forms of cancer (~1% of all cancers affecting males); however, rates continue to increase worldwide (McCullagh, Lewis, & Warlow, 2005; Powe, Ross, Wilkerson, Brooks, & Cooper, 2007). Whether or not these increases are attributable to an increased effectiveness in screening or an elevated exposure to carcinogenic environmental, behavioral, and/or genetic components is up for debate among apposite academic and clinical circles. According to the author’s knowledge, no uncontested evidence exists at the present time regarding this concern.
The lack of information dissemination on the disease’s incidence, prevalence, etiology, treatment, and prevention measures within the general population contributes to a knowledge deficit regarding awareness and preventative behavior performance (McClenahan, Shevlin, Adamson, Bennett, & O’Neill, 2007; McGilligan, McClenahan, & Adamson, 2009). This is a cause for concern, as with all cases of cancer, early discovery of testicular cancer affects survivorship (Cox, McLaughlin, Rai, Steen, & Hudson, 2005; Gilligan, 2011) as cure rates for testicular cancer are approximately 96% if caught in the early stages (Siegel et al., 2012). In contrast, reports ranging from 20% to 50% of patients are diagnosed with late-stage testicular cancer, implying that early detection did not occur (Moul, 2007; Trumbo, 2004). The cases of late-stage diagnosis are disproportionately affecting minority male populations as a greater percentage of White men typically present with localized testicular cancer at the time of diagnosis, possibly because of better screening methods (McGlynn, Devesa, Graubard, & Castle, 2005).
The American Cancer Society (2014) recommends testicular self-examination (TSE) as an effective and cost-free method to prevent such late-stage diagnoses. TSE is seldom taught, and/or properly taught, and minimally practiced, with varying degrees of efficacy (Rovito, Gordon, Bass, & Ducette, 2011). Despite some concern for TSE’s net benefit because of possible increases in anxiety from false-positives and a lack of clinical trials demonstrating its mortality reduction capabilities (Dearnaley, Huddart, & Horwich, 2001; Hopcroft, 2012; Lin & Sharangpani, 2010), the majority of research over the years calls for regular TSE performance among males and for an expansion of promotion programs (see Brewer, Roy, & Watters, 2011; Rosella, 1994; Steadman & Quine, 2004). Recently, Aberger, Wilson, Holzbeierlein, Griebling, and Nangia (2014) make a case for conducting TSE using a cost-utility analysis stating that a 2.4:1 cost–benefit ratio exists for a case of testicular cancer caught early versus a more advanced discovery.
As previously outlined evidence suggests, knowledge of the TSE is limited at best, which likely serves as a barrier for males to seek information on self-examination techniques. Therefore, the major focus of TSE promotion is directed toward improving communication efficacy between providers and patients/participants to increase knowledge/awareness of testicular cancer and TSE. Our concern is that the majority of existing interventions attempt to solely raise awareness of TSE and/or intention to perform the behavior, and lack a concerted effort to increase actual TSE performance.
Outcomes of Interest/Goals
Existing analytical designs primarily assess knowledge, awareness, intention to perform, and even execution efficacy of TSE (procedural completeness or correctness, but the extent of this is limited in the literature). To these authors’ knowledge, no cohesive summary exists in the literature on TSE promotion research that assesses actual performance of the behavior. This can be problematic to an iterative approach to wellness promotion among males.
This review organizes and assesses the literature on analytical interventions promoting TSE performance for at-risk male populations. This evidence is presented to encourage a larger discussion on the quality of existent research promoting TSE performance. From this analysis, a best-practices guideline is produced for TSE promotional efforts seeking to measure behavioral change.
Method
Data Sources and Search Strategy
The search strategy used the Ovid Medline (1950 to present), CINAHL (1982 to present), PsycInfo (1806 to present), All EBM Reviews (2014), Ovid Healthstar (1966 to present), ERIC, and Google Scholar (2014) databases to locate relevant literature. Furthermore, the review used ancestry and gray literature searches to ensure full capture of relevant behavioral intervention research. The gray literature searches were confined to conferences and dissertation research concerning TSE promotion. All databases except for Google Scholar and ERIC used OVID Gateway. Google Scholar used its own searching catalog while ERIC used EBSCOhost.
Inclusion/Exclusion Criteria
This systematic review of literature includes peer-reviewed English-language analytical design trials (randomized controlled and quasi-experimental) examining interventions measuring TSE performance among males. This review did not include studies using participants who have sought care at a genitourinary medical clinic as these individuals may be more apt to follow recommended pelvic (i.e., testicular) health regimens than others (Kennett, Shaw, & Woolley, 2014). Studies were excluded if it solely discussed the etiology of, and surgical or physical treatments for, testicular cancer. Furthermore, interventions exclusively seeking to increase knowledge, awareness, intention to perform, and/or execution efficacy (the ability to conduct a proper self-examination) of TSE also were excluded from the review. Though the authors agree with Finney, Weist, and Friman’s (1995) conclusion that execution efficacy, intention to perform, knowledge, and awareness are important in performing self-screening, researchers must first determine best practices on how to encourage individuals to practice TSE before they dissect individual components of a comprehensive health promotional intervention. In other words, researchers should determine what causes behaviors to change and then go back and perfect those individual components to make it even more explanatory.
Information on testicular cancer risks, TSE knowledge, behaviors (discussions with health care professionals about testicular cancer and TSE), preferences (intentions to self-screen or not), and behavioral outcomes (exercising of TSE) provided the evidence base necessary to conduct a systematic evaluation of TSE behavioral outcome interventions. To assess the efficacy of specific interventions and produce a best-practices model for TSE promotion among at-risk males, only peer-reviewed studies published in academic journals that experimentally assessed TSE behavioral outcomes were reviewed.
Screening Procedure
Primary Screening
Our first assessment was to screen titles and abstracts for relevance. Articles were excluded that solely discussed the etiology and surgical or physical treatments for testicular cancer.
Secondary Screening
Our second assessment screened full articles for relevance. Descriptive, nonintervention studies were excluded, as well as non–peer-reviewed and/or nonpublished material (inclusive of gray literature). Interventions solely assessing TSE behaviors on cancer survivors also were excluded.
Tertiary Screening
Our third assessment of the articles further eliminated interventions solely assessing testicular cancer and/or TSE knowledge and awareness. Interventions were excluded if they solely considered and appraised intention to perform and/or execution efficacy of TSE and did not directly measure performance of TSE in response or exposure to an intervention.
Review of Methodological Quality
Overview
Three reviewers analyzed the literature and collaborated on potential differences in scoring. A unanimous agreement among all raters was required to finalize the score for each article. The review did not conduct multiple analyses or comparisons of different age groups or gender as testicular cancer affects males, primarily between the ages of 15 and 40 years. No systematic reviews were identified assessing TSE behavioral promotion intervention quality. Therefore, this review did not calculate congruence or grant primacy to previous reviews.
Data Extraction and Methodological Quality
Extracted data included source citation, sample size, brief content summary of intervention, theoretical framework (if present), primary outcomes, significance level of TSE performance results behavior (male participants systematically self-checking their testicles for abnormalities; a value of .05 served as alpha), and key weaknesses of the design. This information allowed us to assess the quality and effectiveness of the included trials on TSE promotion and performance among at-risk males.
Specifically, an amended Downs and Black (1998) checklist was used to measure methodological quality of included interventions. The instrument was used to measure validity and reliability of both randomized controlled trials and quasi-experimental trials, which allowed for an assessment of generalizability. Downs and Black’s (1998) checklist has been validated by the Agency for Healthcare Research and Quality (2002) as a system that comprehensively addresses the key domains of methodological quality of individual articles in order for researchers to be able to measure said quality.
The method was fashioned from epidemiological principles, reviews of study designs, and existing checklists for randomized controlled trial assessments. There was a total of 27 items to measure methodological quality, with a total of 32 possible points. However, Question 27 was rescaled in this review from a score range of 0 to 5, to 0 to 1 where methodological power was assessed as 0 = did not mention power and 1 = did mention power. Similar methods were used by Eng et al. (2007).
There was a total of 27 items with 28 possible points on the revised assessment checklist. Interventions with a score ranging from 0 to 12 were rated as poor quality, 13 to 22 as average quality, and 23 to 28 as high quality. These original values were determined a posteriori through unanimous agreement among the three reviewers on what were appropriate upper and lower bounds of the review score interval because of the absence of a criterion method for establishing cut points for this tool (see Brouwers et al., 2005, for an expanded discussion on this issue).
Results
Search Strategy Results
SEARCH: OVID Gateway (includes Ovid Medline, CINAHL, PsycInfo, All EBM Reviews, and Ovid Healthstar databases; see Figure 1)
Flowchart of screening process.
Reviewed Articles
The reviewed literature characterizes moderate levels of effectiveness of interventions promoting actual TSE behavior. In terms of methodological quality scores, there was a range of 13 to 24 out of 28 possible points. There was one high-quality intervention and nine average-quality interventions among the selected studies, which are listed in chronological order of publication date in Table 1.
Methodological Quality Scores.
The aforementioned methodological quality criteria shepherded the review of the 10 identified interventions. The information extracted is presented in Table 2.
Study Summaries.
Note. TC = testicular cancer; TSE, testicular self-examination; TPB = theory of planned behavior; ACS = American Cancer Society.
Discussion
Summary
In terms of significant TSE reporting between intervention and control/comparison groups, 3 out of 10 (Brown, Patrician, & Brosch, 2012; Finney et al., 1995; McCullagh et al., 2005) did not achieve a statistical causal threshold. However, McCullagh et al. (2005) did report a significant increase in pre- and posttest reported TSE among the experimental group. All reviewed articles, save one (Steffen, Sternberg, Teegarden, & Shepherd, 1994), were of average methodological quality; Steffen et al. (1994) was of high methodological quality.
Findings of this review suggest that the reviewed interventions are largely affected by the presence of systematic error, and therefore not completely indicative of the true relationship between the intervention and the outcome (i.e., actual TSE performance). However, there was documented success in increasing TSE behaviors among males using various techniques of information delivery mechanisms (i.e., Wanzer, Foster, Servoss, & LaBelle’s [2014] Check Yo Nutz campaign), communication strategies (i.e., Steffen et al.’s [1994] “message framing”), and health behavioral theories to develop promotional campaigns (i.e., Marty & McDermott’s [1985] use of the health belief model], among other methodological tactics.
Notable weaknesses among the reviewed articles include the ubiquitous use of self-reporting of the TSE performance variable, significant levels of attrition with little or no discussion on its cause or possible effects on results, and a general lack of reporting on psychometric qualities of instruments and theoretical frameworks from which the interventions were designed. Notable strengths include the documented success of using facilitators (either peer or physician) to inform males of the disease and TSE compared with a nonfacilitated treatment, the benefit of using a theoretical framework (documented from those studies who clearly identified a model/theory) in the design and implementation of the intervention, and the use of structured communication strategies within the intervention.
Methodological Commentary
General Concerns
Extant TSE promotion intervention research is reported in short-term analyses and is not very well disseminated to the general population. Furthermore, none of the interventions were longitudinally designed to observe TSE behaviors more than 6 months, which we consider to be insufficient follow-up. However, in their defense, resource issues prevent studies from properly measuring sustainability of TSE over years, versus just weeks or months. We presume most of the reviewed interventions perhaps faced such resource concerns as external granting entities determine funding priority topics year to year. Considering the repetitious issue of a lack of long-term follow-up across the body of review interventions over the span of nearly three decades, it appears that funding and/or priority calls for testicular cancer and TSE-based interventions are limited at best.
Sampling and Attrition Concerns
Among the reviewed studies, the consistency and validity of sample size, and the issues surrounding sample pooling and attrition, were major issues of generalizability and reliability, as well as internal validity, if attrition was selective. Furthermore, all included interventions used a convenience sample of college or high school students (save McCullagh et al., 2005, who used work sites and Brown et al., 2012, who sampled military personnel), which could have varying negative effects on generalizability. Relatedly, McCullagh et al. (2005) may have discovered higher increases in the numbers of individuals practicing TSE than other studies because of the older population of men used in the study. It is quite possible that TSE behavior increases with age even though testicular cancer mainly affects 15- to 40-year old males (Rovito et al., 2011). More evidence is needed to claim that with certainty, however, it is entirely plausible that this is the case.
In terms of ethnicity/race, White males have the highest incidence of testicular cancer among all groups (Rovito et al., 2011). However, the majority of interventions did not mention the ethnic/racial makeup of the participants, and those that did, approximately 80% were White males. This introduces generalizability concerns of reviewed article results.
Loss to follow-up was a major issue with most of the reviewed studies reaching upward of 40% to 50% attrition rates. Higher attrition rates could possibly stem from college semester breaks where students leave campus and/or a change in student contact information. Only Steffen et al. (1994), Wanzer et al. (2014), and Murphy and Brubaker (1990) had acceptable attrition rates.
In the cases of Wanzer et al. (2014) and Murphy and Brubaker (1990), their data collection was primarily conducted during mandatory class periods. For Steffen et al. (1994), the sample’s reported 13% loss to follow-up probably stemmed from the need to complete this study in order to fulfill their introductory psychology course requirements. One could possibly bring up ethical issues of autonomy as participants may not have had the realistic choice of saying no to participating in this study. However, not enough information was given on this consent process in the piece. In fact, the body of reviewed studies had insufficient, or even a complete lack of discussion on attrition, intent to treat, and/or statistical power.
Measurement Concerns
All included interventions contained potential issues of self-reporting systematic error. This, however, may be almost unavoidable given the sensitive nature of measuring TSE behaviors. Furthermore, almost all reviewed studies failed to report on the psychometric qualities of primary intervention and assessment tools, which raises issues of reliability and validity of data.
Walker and Guyton (1989) indicated that individuals may not report TSE because of the lack of information dissemination on the subject, including what it is and how it is performed. Therefore, it is quite possible some participants might have performed TSE but not known it because of unavailable information/lack of knowledge on the subject, or they might be embarrassed to report on their behaviors because of perceived peer judgment in that their peers are privy to their data/responses/actions. All reviews indicate that a deficit exists between ongoing research and the circulation of said information to the appropriate populations.
Concerning measurement error, 3 out of the 10 review interventions mentioned any type of theoretical framework from which to design the program. Out of those 7 who did identify a particular theory to model their intervention, only Marty and McDermott (1985; health belief model), Murphy and Brubaker (1990; theory of planned behavior), Steadman and Quine (2004; implementation intentions), Brown et al. (2012; health belief model), and Wanzer et al. (2014; standard model of health communication) have overt references to a formal, recognized health behavior theory/model. One must call into question the validity of those campaigns that did not use a theoretical framework as they may have left out core explanatory variables from the intervention. However, Steffen et al. (1994) did use message-framing techniques, which may alleviate some of these concerns.
Intervention Effectiveness Commentary: Best-Practices for Future Interventions
General Recommendations
The included research indicates that interventions have a significant positive impact on self-report TSE performance among male populations. Despite the lack of knowledge and awareness in regards to TSE/testicular cancer, participating males seem impressionable to promotion messages about TSE from a well-designed intervention.
We concluded that no method was substantially more explanatory or predictive than the others as only 10 interventions were assessed and most produce significant results in promoting TSE behaviors among the treatment group. Furthermore, we could not confidently suggest using one campaign in its entirety over another because of the amount of systematic error present in each design and the varying number of outcomes assessed by each individual study, although all did measure TSE performance. However, although the aforementioned made a true comparison of overall campaign success difficult, we identified key components from individual campaigns that were successful in promoting TSE performance to produce a best-practices recommendation for future intervention designs.
Sampling and Attrition Recommendations
Generally, we suggest that researchers must focus on developing a valid methodology for sampling (sample size, population pools) and sustainability of sample sizes to assess the significance of interventions promoting TSE. The convenience of sampling students in health science classes and other educational settings presents generalizability issues with research findings. If at all possible, researchers should strive to gather a random sample of individual participants and then randomly allocate individuals to treatment groups. Also, there needs to be more focus on reducing the amount of attrition. Using strategies like Steffen et al. (1994), Wanzer et al. (2014), and Murphy and Brubaker (1990) could reduce loss to follow-up, but could then bring in either ethical issues or concerns of generalizability. We suggest providing appropriate incentives that are topic-related, such as prepaid pharmacy or health care access cards, or academic-related that are not required by the course for grading purposes, such as service-learning credit.
There needs to be a study design that is more representative of the at-risk population. High schools and larger nonstudent populations need to be sampled more. Although sampling is an issue because of limited current resources to conduct broader research, researchers should also use the method chosen by Steadman and Quine (2004): random assignment from one population, despite it being drawn from a single student population. Researchers should refrain from using two or more distinct, separate populations to draw samples for one study because of possible differences between groups. There needs to be a means of comparison across groups, not collectively analyzing TSE performance as one outcome variable for all groups, as demonstrated by McCullagh et al. (2005).
Despite 15- to 40-year old White males having the highest incidence and prevalence of testicular cancer among all other groups (Giannandrea et al., 2013), interventions need to include other groups in the experiments, at least in proportion to national incidence rates. The sustainability issues primarily involve resource allocation, which stems from national research institute funding shortages supporting testicular cancer/TSE research. Considering Healthy People 2020’s goals, including eliminating disparities, it is crucial that researchers address the issue of gender equity as it pertains to health resources concerning these types of cancers, among other wellness issues in males. Future interventions are encouraged to factor in a more balanced demographic sample regarding race and ethnicity in order to increase generalizability.
Measurement Recommendations
Self-reporting TSE performance is, again, an almost unavoidable concern. Researchers should use self-reporting methods and couple it with an execution efficacy test where those men who report moderate to high TSE performance posttest (assuming the intervention demonstrated how to properly perform TSE) should be able to perform the exam properly (within varying degrees of success) on a lifelike model, as discussed most clearly in Finney et al. (1995). Perhaps researchers can use a system of primary care professional reporting measures in their interventions to provide alternative courses of behavior reporting, or at the very least, the participant spurring a conversation/asking questions relating to the behavior/disease.
In regards to psychometric analyses, the instruments used to collect data on TSE performance (even self-report, as discussed prior) must be psychometrically tested to ensure reliability and validity of data. Pilot testing to calculate a test–rest reliability coefficient or organizing questionnaires according to composite themes to calculate internal consistency can easily relieve these methodological concerns.
Summary Best-Practices Recommendations
The differing intervention approaches presented the aforementioned foundations of a best-practices guideline for TSE promotion. In summary, these authors specifically suggest the following:
Researchers should make reference to Steadman and Quine’s (2004) “cueing” strategy in their intervention design as it produces significant findings between groups and is a very simple, yet effective method in increasing TSE. The implementation intentions approach is also effective with other secondary screening procedures, including cervical cancer (see, Sheeran & Orbell, 2000). Furthermore, Walker and Guyton (1989) stated that viewing a slide presentation demonstrating TSE with accompanying written information did not reduce fear. However, they suggested demonstration and practice of TSE on a lifelike model could significantly decrease a person’s fear about testicular cancer, thus increasing behavior. Coupled with Cox et al.’s (2005) multimedia approach to TSE training and reminding system, cuing and active modeling can present a comprehensive approach to increasing TSE performance among at-risk males. However, it is essential to provide a well-rounded knowledge/awareness portion of the intervention.
The literature demonstrates that in order to have significant increases in TSE behavior, knowledge and awareness must increase. We recommend that researchers not overextend their approach in achieving significant TSE performance from the intervention group by spending time and resources on variables that are not explanatory for increasing the frequency of TSE, most notably, attitudes. Although attitudinal measures indicated baseline score differences, and little, if any, score changes postintervention, this does not indicate a male’s possible noncompliance with performing TSE. Although attitude may not change from the intervention method, knowledge and awareness increases, thus altering behaviors. It is quite possible that fear can predict TSE performance despite a present negative attitude toward self-screening.
Increasing knowledge and awareness, demonstra-ting TSE via active modeling, and incorporating cuing mechanisms to remind males to perform TSE may potentially produce positive, sustainable health behavior choices. This approach does not follow the exact model of the theory of planned behavior (Murphy & Brubaker, 1990) as attitudinal shift is not an essential component to spur change. However, properly trained staff (i.e., nurses or general practitioners) are needed to implement the best-practices model (see Peate, 1999, and Rosella, 1994). Using Marty and McDermott’s (1985) message implementation method with limited training offered to intervention instructors/facilitators, is risky and not the most valid of approaches. The use of trained, professional staff (i.e., nurses, general practitioners, health communication specialists) is the gold standard for improving knowledge and awareness sessions, as well as message design strategies/tools for dissemination (pamphlets, television, billboards, etc.).
Related to the previous recommendation, we urge researchers to use a formal model to design their interventions, as seen in Marty and McDermott (1985; health belief model, Murphy and Brubaker (1990; theory of planned behavior), Steadman and Quine (2004; implementation intentions), Brown et al. (2012; health belief model), and Wanzer et al. (2014; standard model of health communica-tion). Also, Wanzer et al.’s (2014) use of themed interventions and/or Steffen et al.’s (1994) structured message framing appears to be an effective strategy to promote TSE among male participants.
Improve on the following (as discussed prior): sample selection through randomization and allocation techniques from mixed populations, sample retention by providing unique incentives, and outcome reporting through the use of mixed-reporting measures (self- and primary care professional) and/or using execution efficacy of TSE behaviors among those who report performing the behavior. These recommendations will provide for more methodologically sound results.
Limitations
Several limitations exist within this review. Results are weakened because of the unreliability of self-report TSE performance. Because of the dichotomous nature of reporting the primary outcome (TSE), the analytical methods between treatments on the effect on TSE promotion were limited. Although three raters independently reviewed and scored the selected studies, there was no testing of data reliability between individual scores. There was a consultation among all three raters to address scoring differences and come to a conclusion on a final score.
This review was specifically focused on TSE behavioral outcomes, which were self-reported by participants. At least one of the reviewed studies did not focus on TSE performance as its primary outcome measure. Finally, the inclusion of studies with small sample sizes, as well as studies with substantial levels of selection bias, were included in the review. This limited the ability to generalize statistically significant findings of increased TSE performance to larger populations.
Conclusion
As there has been an exhausting amount of descriptive, cross-sectional research conducted on testicular cancer during the past 30 years, the field needs to switch course and focus on experimental interventions to increase TSE performance. The significance demonstrated by the available intervention research suggests that increasing TSE behaviors among at-risk males is not as difficult to execute as was once believed.
TSE is a viable and useful method to detect testicular cancer and may contribute to healthier lifestyles for at-risk males, including learning the value of self-awareness in terms of their overall wellness. More research and advertisement is needed to expand the field and gain exposure. But first researchers need to greatly improve the methodological quality and education effectiveness of their designs in order to conclusively state whether or not an intervention promotion TSE is successful. Therefore, researchers are encouraged to include assessments, however brief or lengthy, of proper TSE execution conducted by participants in their interventions. Finally, researchers are encouraged to include long-term follow-up (>6 months) assessments to determine adherence to the behavior.
This review provides a framework from which researchers can investigate the existing intervention literature on TSE promotion, as it is (to the authors’ knowledge) the only existing systematic review attempting to organize the available studies. Continuing efforts will be made to include future research into this assessment.
Footnotes
Acknowledgements
We would like to thank everyone at Men’s Health Initiative for their tireless efforts in executing this study as well as their devotion to promoting men’s health.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.