Reasons for and against participation in studies of medicinal therapies for women with breast cancer: a debate

Background A special challenge for research studies of breast cancer among females is low patient participation rates. We compiled this systematic review to identify reasons why women with, or at high risk of, breast cancer do or do not participate in medicinal studies of breast cancer. Method A systematic literature search in the databases Cochrane Library, EMBASE, Medline, Pascal Biomed, ACP Journal Club and CINAHL, as well as searches through reference lists of relevant texts, was performed. Results Of 39 relevant full texts, ultimately, nine studies (1 qualitative, 8 quantitative) were included after applying the inclusion criteria. Despite a lack of data material, it was possible to identify various factors influencing women's willingness to participate in medicinal studies and group them into three categories: person-related, study-related, and physician-related. Conclusion Reasons for or against participation in studies of medicinal therapies by women with, or at high risk of, breast cancer are multi-dimensional, and should be considered when planning such studies to garner higher participation rates. For a more comprehensive picture of factors that affect participation, further studies in this field are recommended.


Background
Recruitment of participants in breast cancer trials relative to the incidence of breast cancer patients is somewhat higher than in other cancer types, such as colorectal or lung cancer [1]. However, the number of participants in breast cancer trials is still rather low. Estimates of participation of adult persons in cancer studies is about 1.5-11% of the total number of newly diagnosed or incidental cases [1,2]. Low participation in studies can lead to under-representation, which, in turn, can result in an effect or a clinical effectiveness being shown as not significant [3] or in a failure to obtain theoretical data saturation [4]. Low participation rates in a study may also induce bias, whereby those enrolled do not represent the target population very well. The reason why a planned sample size is not reached within the time frame can, among other things, depend on problems in recruiting participants [5]. Research studies, grouped in reviews, have investigated possible barriers to cancer patients' participation in clinical studies [6][7][8]. Among these barriers are, for example, randomization, preferences for a certain therapy, degree of knowledge or additional effort in travel [6,7].
In all of these studies however, patients with different cancer types were combined; for example, breast cancer, lung cancer and intestinal cancer, as well as various therapies.
To gain a better understanding of why especially female breast cancer patients frequently do not participate in clinical medicinal studies, a systematic review regarding this sensitive female patient group was conducted.

Study type
This review includes both qualitative and quantitative studies that discuss reasons and barriers or influencing factors for the participation of female breast cancer patients in medicinal studies. Clinical medicinal studies in this paper, are defined as studies with a variety of medical therapies, which may include, for example, chemotherapies, endocrine therapies and immune therapies within neo-adjuvant and adjuvant therapy [9].
Articles were limited to those published in the last 10 years in German or English. When uncertainties regarding whether the content of the studies was suitable for this review occurred while screening study abstracts (or when articles did not have abstracts), the respective full texts were procured for further evaluation. To find additional relevant literature, the reference lists of the full texts were searched for possible articles.

Selection criteria
Inclusion criteria: studies that focused on participation or non-participation in clinical medicinal studies of adult women with breast cancer.
Exclusion criteria: studies that did not examine consent-competent female patients, did not focus on medications, investigated the effectiveness of medications, or did not refer to primary sources, study protocols and commentaries.

Data evaluation
Full texts found in the literature research were evaluated regarding methodical or theoretical rigor by means of the applicable quality checklist for quantitative or qualitative studies of the Alberta Heritage Foundation for Medical Research [10]; texts with values < .75 by these criteria were not used.

Data analysis
The following data were extracted from the respective full texts: objective, study type/design, method, number and characteristics of the participants, definition of the term "study participation," setting, and results. These extracted data were compared and categorized with regard to factors influencing participation (for and against see Table 1 and 2) in clinical medicinal studies of women, regardless of which research area they originally came from. This categorization was performed according to Mills et al., 2006 [7].

Results
After applying the selection criteria according to the title and abstract screening, 39 publications (of initially 3080 references) remained for further evaluation. Twelve publications were excluded as secondary literature, comments or protocols; 17 publications were excluded because their focus was not of interest (e.g., medicinal studies). One additional article was included from the reference lists. Two articles had to be excluded as they each had a quality score of < 0.75. Ultimately, 9 publications met the inclusion criteria and were subjected to further analysis and evaluation (see Figure 1).
Eight of these nine included studies were quantitative [11][12][13][14][15][16][17][18]; only one was qualitative [19]. The summarized data extraction of the individual studies is found in Table 3. The quality assessments performed, including comments, are depicted in Tables 4 and 5, according to research areas. Four of the studies were conducted in the United States [12,13,17,19], two in Canada [11,14], and the rest in Europe; in Germany [15], France [16] and Italy [18]. All of these articles have been published in English.
In the articles, a great variety of study designs were utilized as case-control [12], cohort [16] or randomized studies [17] (among others), where the design was not explicitly identified in most of them [11,[13][14][15]18,19]. The prevalent method used to elicit women's reasons for or against participation in studies with medicinal therapies were questionnaires [11][12][13][15][16][17][18]. Most of the articles referred to investigations already performed, to identify the relevant reasons for consent or refusal to participate [12,14,15,18,19]. Three studies, though, used theoretical scenarios for the solicitation of reasons [11,13,16]. In one article, on the other hand, women were asked regarding their intent to participate in a currently active study [17].

Discussion and conclusions
The various factors influencing participation in breast cancer medicinal research identified in the nine studies were placed by the authors into three categories: personrelated, physician-related and study-related.
The person-related category comprised health, psychosocial and demographic reasons. A younger age (demographic reason) was identified as a factor influencing willingness to participate [15,17,18].
Regardless of whether study participants were younger or older, they frequently had high subjective perceptions of their risk of breast cancer [16,18], although their objective risks, assessed according to Gail scores, in one of the two studies, were relatively low on average [17,19,20]. Another study also showed women who had participated in medication studies assessed their breast cancer risk subjectively much higher than women who had not participated, though both groups did not differ according to their Gail scores [19]. This suggests that, in future studies, subjective perceptions of risk should be addressed. Also, in ovarian cancer investigations, potential subjects' higher personal risk perception and concern raised the probability of making use of screening [21,22]. These results suggest that women's participation in such studies depends more on subjective risk than objective risk. Though subjective breast cancer risk in two studies in this review correlated with the women's willingness to participate [17,19], in another medication study, women's participation was more likely when they were less personally concerned about breast cancer [18]; however, this last-mentioned study does not reveal whether relatives of the study participants had suffered from breast cancer or not, which could increase the subjects' concern, and thus interest, in participating in a medication study. One study showed women who had first and second degree relatives with breast cancer requested Altruism (support of medical research and/or helping other patients) Support of medical research or helping other people (33%, n = 8) [11]; contribution to therapeutic knowledge (44%, n = 116) [18]; altruism in general (50%, n = 14) [19]; non-participants felt bad due to altruistic reasons as they themselves did not contribute to the study efforts (39%, n = 9) [19] Wish for improvement (regarding own chances or regular medical care) Improve the chances for breast cancer prevention (P = 0.033; OR = 3.16; 95% CI: [1. .06]) [12]; a regular medical care (58%, n = 154) [18], wish for helping themselves (50%, n = 14) [19] Concern of breast cancer or often thinking about the topic of breast cancer Memory of close person(s) with breast cancer 21% (n = 6) valued their own breast cancer risk as low, but nevertheless participated in the study out of respect or a memory of a close person who suffered from breast cancer [19] Readiness for treatment in a study Nothing to lose by participating (33%; n = 8) [11]; general willingness (60%; n = 160) [18] Acceptance of randomization Acceptance of randomization (P < 0.001; OR = 4.6; 95% CI: [2.7-7.7]) [16] No placebo Placebo vs. no placebo (RR = 0.80; P = 0.05) [14] Longer interval from diagnosis/surgery/end of therapy until enrollment For non-metastatic trials: if the study protocol planned an interval of 12 weeks or longer between diagnosis/op/end of therapy and recruiting in comparison to a shorter interval (RR = 1.36; P < 0.01) [14] Already decided once to participate in a medicinal study Already having decided to participate in a study (P < 0.001; OR = 5.0; 95% CI: [2.9-8.7]) [16] Predominantly advantages through the study Perceived value of the trial (P = 0.020; OR = 2.92; 95% CI: [1.18-7.21]) [12], predominant advantages of the study (50%, n = 14) [19] Each treatment strategy seems helpful Each of the treatments seems to be beneficial (42%; n = 10) [11] Idea of receiving a better treatment Idea to receive better treatment (25%; n = 6) [11] Feeling, physicians must make decisions Medical decision-making preferences (P = 0.045; OR = 2.2; 95% CI: [1.0-4.9]) [16] Feeling of not being able to reject physician's suggestion Unable to refuse the physician's suggestion (P = 0.031; OR = 1.8; 95% CI: [1.1-3.2]) [16] Satisfaction with receiving information during consultation with physician Satisfaction with communication processes (P < 0.001; or = 3.1; 95% CI: [1.5-7.8]) [16], satisfaction with the physician's explanations (P < 0.001; OR = 9.33; 95% CI: [4.04-21.55]) [18] Receiving information regarding financial conflicts of interest 61-72% (n = 614-724) wish to receive information about financial conflicts of interest [13]. 61-84% (n = 614-845) would participate in a medication study in spite of financial conflicts of interest [13] Adequate medical expert knowledge or qualification of the physician information about the medication study twice as often as women who did not have this diagnosis among their relatives (OR 2.35, 95% CI, 0.99-5.57) [15].
Studies show a negative correlation between the concern regarding breast cancer and satisfaction with the physician's consultation [23]. Nevertheless, satisfaction with physicians' consultation and communication processes (physicianrelated reason) was mentioned as reason for participation/ non-participation in medication studies [16,18], indicating that patient-clinician relationships play a decisive role in patients' willingness to participate in a study.
Identification of potential differences in consultations in the medication studies was not possible [11,13,16] because only one study reported the content of the conversations [17]. Therefore, not all women in the studies might have been informed about the same things. To avoid this distortion in future studies, the use and documentation of conversation manuals in these consultations seems advisable.
The main study protocol-related reason against participation was additional time needed [12,18,19].
The randomization procedure was mentioned in two studies as reason against participation [11,19]; willingness for randomization was mentioned in one study as an influencing factor for participation [16]. In cancer research, lack of understanding of the principle of randomization has been researched as barrier to subject participation [24,25]. Another study showed that, among those who initially decided against participation in randomized studies, more than half ultimately consented to participation after they had received more detailed information regarding the randomization process [26]. Though this connection was not identifiable from studies in this review, our results showed that randomization could influence participation in medication studies. Informing potential study subjects of the reasons for randomization could therefore promote their participation.
A limitation of this review concerns its ten-year time frame. Although more full texts might have been included if our criteria allowed older investigations, the primary goal of this review was to identify current studies; we therefore restricted this study to the past decade.  (1) Full texts in review (9) Exclusion due to low quality (2) Figure 1 Flow chart of study selection process. The strengths of this review include its use of an extensive assessment scheme, allowing comprehensive quality evaluation of the respective articles, using consistent criteria. This scheme also could function as a kind of checklist, thus reducing the probability of forgetting any items in the assessment. Another strong point was that six databases were searched, allowing wide coverage of possible publications, as a result, of the articles gleaned from the references of all the full texts, only one was found that had not been part of the original database literature research. Inclusion of only high-quality studies is a further strength, as all studies had to show a high quality, of at least 75%, to be included in this review. Two studies did not meet this requirement and were therefore not included in the results [29,30].
Physicians' viewpoints as to why women with breast cancer or breast cancer risk choose or decline to participate in medication studies is being researched, both with regard to general cancers [31] and breast cancer [24,32], but not specifically with medication studies.
In sum, this review shows that the reasons for participation/non-participation in medication studies are multifactorial. Moreover, while factors affecting patient participation in medication studies are obviously useful to know in planning and realizing future investigations, few such insights are currently available, apparently due     Control at analysis level with multivariate model; but in asking the control group, the same standardized questionnaire as in the intervention group was used Result depiction Partially Secondary results depicted in the text and tables; but the primary result is not mentioned in the text, it is only listed in the table; recording of influencing factors not conducted for both interventions. Differences regarding the objective breast cancer risk between those women who consented to participation in the medication study and those who refused participation were not pointed out. The text only lists percentages and no absolute numbers; consequently, readers have to infer from the tables how many women in total participated in the data collection, and subsequently in the medication study. to the small number of relevant studies; further quantitative and qualitative research is needed. Control at analysis level with multivariate model; also, control regarding age distribution based on low response rate for questionnaires; furthermore, an attempt was made to standardize the preceding consultation session

Conclusions
Result depiction Partially Results depicted in the text and tables and correspond with each other; but results of the regression analysis are interpreted in the text as "in connection with" and not as "influence" -the interpretation in the framework of the remark in the table is again depicted correctly