The RCAC Scale
The RCAC scale consists of 18 items that constitute six dimensions: Fertility potential, Partner disclosure, Child’s health, Personal health, Acceptance and Becoming pregnant. Each dimension has three items with responses scored on a five-point scale (ranging from 1=Strongly disagree to 5=Strongly agree), with higher scores indicating higher levels of reproductive concerns (7). The psychometric properties of the original scale in English showed that the internal consistency of the total RCAC scale was good (α =0.82), and good or acceptable for the six sub-scales (α=0.78-0.86) (7). Known-groups validity was demonstrated by significantly higher mean scores among women who wanted to have children compared to those who did not, and among those for whom having a biological child was very important compared to those for whom it is was less important (7). Total RCAC scores were positively associated with depression and negatively associated with social support and satisfaction with life.
Swedish translation and cultural adaptation of the Swedish RCAC Scale
For the purpose of use in a study from our research group (12), the original RCAC scale was translated and culturally adapted to be able to measure reproductive concerns in a Swedish context. To achieve this, the scale was forward translated into Swedish based on a dual panel approach and further assessed for cultural adaptation by lay panel, patient/target group assessment and cognitive interviews (13). The process of translation and cultural adaptation was coordinated by the same coordinator to ensure none of the parameters were neglected and to maintain the quality of the adaptation.
Forward translation by bilingual expert panel
The RCAC scale was translated from English to Swedish by two researchers who were native Swedish speakers and well-versed in English. The first translator (T1) had extensive research knowledge in the field of fertility and cancer. The second translator (T2) coordinated the translation and was experienced in instrument translation and adaptation, but was not knowledgeable in the research fields of fertility and cancer. Both researchers had broad experience of clinical work with diverse patient groups. Following individually performed translations of the scale, T1 and T2 discussed discrepancies between the two versions. As a next step, they consulted the principal investigator of the original English version of the RCAC to discuss and clarify the intended conceptual meaning of the scale and specific items. Subsequently, the two translators (T1 and T2) and two additional experts in the field of psychosocial oncology (native Swedish speakers) discussed the translation in a panel meeting. As a result, a consensus version of the Swedish version of the RCAC was created. Cultural adaptation included changing the use of “spouse/partner” in the three items of the dimension Partner disclosure to “partner”, as this was deemed more appropriate for use in the Swedish context where it is common for two adults to live together without being married.
Lay panel assessment
For evaluation of the translated version of the scale, two lay panels were recruited through personal contacts and local advertisements. The panel members were 3 women and 4 men between 18-41 years old; three had secondary education and four had higher education. Two had children and none of the panel members had been diagnosed with cancer. Panel members were compensated with cinema tickets for their participation. The lay panels were only provided with the Swedish version of the scale, as suggested (13). The lay panel members were instructed to go through all parts of the RCAC scale, including instructions, items and response options. Everyone read each item, then discussed how they perceived the issue and whether there were any alternative ways of phrasing the question. Based on the assessment by the lay panels, minor changes in wording were made. The main role of the lay panels was to produce a version that was easy to understand for the average Swedish speaking person. The lay panel assessment was led by the same coordinator as for the expert panel.
Patient/Target group assessment
The patient/target group included 5 women and 3 men (aged 20-41) who had been treated for cancer. The target group members came from different geographical areas in Sweden and all had secondary or higher education. They evaluated the translated RCAC scale for face validity i.e., if the items and response alternatives were relevant and acceptable. Some concerns were expressed regarding the suitability and relevance of the scale for patients in their late teens, which led us to conducting cognitive interviews as described below. The target group members were compensated for their travel costs and time spent.
Cognitive interviews were performed individually with 3 young individuals (1 female aged 18 years and 2 males aged 17 years) currently being treated for cancer. The participants completed the Swedish RCAC scale and were then interviewed on their experience of responding to the items. We used a flexible approach as suggested earlier (14). In brief, respondents were asked to think aloud and share thoughts, perceptions and opinions that came up when answering the questions. The interviewer probed every issue that was mentioned to capture difficulties related to the items of the RCAC. We asked if items were difficult to understand (wording, knowledge requested and information given), included inappropriate assumptions, were too sensitive (content, wording and social acceptability), as well as if the response categories were adequate.
Participants and procedure
A detailed description of the study participants and procedure is presented elsewhere (12). Briefly, a sample of 301 women consecutively diagnosed with invasive breast cancer (all stages except in-situ and very few with distant metastasis) at age 18-39 years was identified from the Swedish National Quality Register for Breast Cancer. Data collection was conducted by means of a comprehensive postal survey approximately 1.5 years post-diagnosis. Ethical approval for the study was obtained from the Regional Ethical Review Board in Stockholm, Sweden (Ref No: 20131746‐31/4).
The survey completed by the participants comprised several patient-reported outcome measures. For the purpose of the psychometric evaluation of the Swedish version of the RCAC, the Emotional Function (EF) scale of the EORTC QLQ-C30 version 3.0 (15) and a study-specific item regarding current wish to have (additional) children (response alternatives: Yes, Uncertain, No) were used from the survey.
Statistical analysis was performed using SPSS statistics for Windows, version 24 (IBM Corp., Armonk, N.Y., USA) and IBM® SPSS® Amos, version 25 and R version 4.0.0 (The R Foundation for Statistical Computing) with the additional R package ’userfriendlyscience’ (16). For all statistical tests, the level of statistical significance was defined as p <0.05.
Construct validity was ascertained using Confirmatory factor analysis (CFA) and convergent validity. CFA with maximum likelihood estimation was performed to determine the adequacy of the original six-factors of the RCAC on our sample data (7). Four participants were excluded from the CFA due to not having completed the RCAC scale (n=3) or missing values for 2 items in one dimension (n=1). Missing values for single items (n=15) were handled by imputing the mean of the other two items of the same individual in the same dimension. Standardized factor loadings and model fit were determined from the CFA. Standardized factor loadings of >0.4 were considered as acceptable as determined from the CFA (17). Model fit was estimated by two absolute indices of overall model fit, Root mean square error of approximation (RMSEA) and Standardized root mean residual (SRMR), and one relative index of model fit compared to the null model, Comparative fit index (CFI). The acceptable thresholds for these indices were defined as RMSEA: 0.05-0.08, SRMR: <0.10 and CFI: >0.90 according to Kline’s (18) guidelines. The model chi-square (χ2), degrees of freedom (d.f.) and associated p value were reported, but were not considered as an indicator of model fit owing to their restrictiveness by being sensitive to sample size (19). Additionally, the relative/normed χ2 ratio (χ2/d.f.) was also reported as recommended, and a cut-off value of <5 was regarded as good (20). Convergent validity was assessed by calculating the Pearson correlation coefficient for the mean scores of the Swedish RCAC scale and the EF scale of the EORTC QLQ-C30. The EF scale covers both symptoms of anxiety and depression and is a robust measure of emotional distress in patients with cancer (21). The construct represented by the EF was thus expected to be related to RCAC scores although it is distinct from that of reproductive concerns. We hypothesized that RCAC total scores would show a negative association of moderate size with EF (higher scores of the scale indicate better EF). Correlation coefficients of 0.10-0.29, 0.30-0.49 and 0.50 and above were interpreted as small, moderate and large, respectively (22, 23).
Data quality was assessed by examination of missing values, mean scores, standard deviations (SD), and floor and ceiling effects. Floor and ceiling effects were considered present if >15% rated at the lowest (floor) and highest (ceiling) scores (24).
Reliability was assessed by means of the internal consistency of the six dimensions using Cronbach’s α as well as Ω total and Revelle Ω total (25); values of >0.70 were considered acceptable (26).
Known-groups validity was assessed by comparing groups of women who differed in their reported wish for children or additional children. It was hypothesised that women with a wish for children would report higher levels of reproductive concerns than those without such a wish, in line with previous reports (7, 27). A one-way ANOVA with post-hoc comparisons using the Tukey test was conducted to compare the mean scores of the Swedish RCAC scale of the three groups of women who reported a certain, uncertain or no wish for (additional) children respectively.