Introduction
Health-related quality of life (HR-QoL) is characterized as multidimensional including physical, social, and emotional well-being perceived by individuals with a medical condition and/or those affected by treatment-related side effects (Kaplan and Ries, Reference Kaplan and Ries2007). Cancer patients’ HR-QoL has become more important especially after the Food and Drug Administration has recommended HR-QoL assessment in the efficiency analysis of the newly developed anti-cancer drugs (Johnson and Temple, Reference Johnson and Temple1985). The assessment of HR-QoL also helps health professionals during the clinical decision-making process (McQuellon et al., Reference McQuellon, Russell and Cella1997). Since different cancer types can cause different conditions, site-specific questionnaires focused on unique problems related to the cancer type have been developed (Aaronson et al., Reference Aaronson, Ahmedzai and Bergman1993; Cella et al., Reference Cella, Tulsky and Gray1993). These problems may be related to the cancer site itself (e.g., defecation problems in patients with colon cancer) or may be specific to the chemotherapy, radiotherapy, immunotherapy, and/or hormonal therapy regimens.
Bone marrow transplantation (BMT) is associated with unique treatment procedures as well as treatment-related complications (Copelan and Edward, Reference Copelan and Edward2006). Since BMT needs high-dose chemotherapies, long hospitalization, and strict isolation rules, BMT-specific HR-QoL measurement can help health professionals understand the physical, psychological, and social well-being of patients and to determine their rehabilitation needs (Pidala et al., Reference Pidala, Anasetti and Jim2009). In addition, besides objective assessments, the measurement of HR-QoL in patients treated with BMT has been described as essential to understand their perception related to their health (Pidala et al., Reference Pidala, Anasetti and Jim2009; Kurosawa et al., Reference Kurosawa, Yamaguchi and Mori2015). Accordingly, it is needed to measure HR-QoL with a BMT-specific scale in patients undergoing BMT.
The Functional Assessment of Cancer Therapy–Bone Marrow Transplant Version 4 (FACT-BMT) is a widely used instrument to assess HR-QoL in individuals treated with BMT. The FACT-BMT is a valid, brief, and self-administered scale and it has been translated into various languages including Chinese (Lau et al., Reference Lau, Chang and Tai2002), Korean (Yoo et al., Reference Yoo, Lee and Lee2006), Arabic (Soudy et al., Reference Soudy, Maghfoor and Elhassan2018), Spanish (Rivera-Fong et al., Reference Rivera-Fong, Benjet and Robles García2020), and Portuguese (Mastropietro et al., Reference Mastropietro, Oliveira and Santos2007). The FACT-BMT versions have demonstrated acceptable reliability and validity in these cultural adaptation studies. The FACT-BMT includes two parts: The Functional Assessment of Cancer Therapies–General Version 4 (FACT-G) and the Bone Marrow Transplantation Subscale (BMTS). The FACT-G assesses the global effects of cancer treatment in terms of physical, social/family, emotional, and functional well-being. The FACT-G Turkish version has been widely used in cancer patients in Turkey (Can and Aydiner, Reference Can and Aydiner2011; Beyhun et al., Reference Beyhun, Can and Tiryaki2016).
The Functional Assessment of Chronic Illness Therapy (FACIT) Organization translated the FACT-BMT into Turkish using a universal language approach that included individual translators from multiple Turkish-speaking nations creating a single Turkish language translation. Since cultural and linguistic differences may have an impact on the understanding and interpretation of the items, we aimed to investigate the reliability, validity, and sensitivity of the FACT-BMT Turkish version in individuals treated with BMT.
Methods
Patients and setting
The present study was conducted at Hacettepe University Department of Hematology in Turkey. Patients were interviewed during their follow-up hospital visits. Inclusion criteria were (1) having received BMT at least 3 months prior to the study, (2) being in an age range of 20–65 years, (3) being able to cooperate, and (4) willing to participate. Exclusion criteria were (1) being within the first 3 months after BMT, (2) being unable to cooperate, and (3) not willing to participate. The participants were asked to self-administer the questionnaires. If the patient was illiterate/requested, the questionnaire was administered face-to-face by the first author. The flow diagram of the study is shown in Figure 1. For test–retest reliability, the FACT-BMT was applied to 47 participants after 1- to 7-day intervals from the first assessment. Ethical approval was obtained from the Hacettepe University Non-Interventional Ethical Committee (GO 20/307). Informed consent was obtained from all patients.
Flow diagram of the study participants.
Demographic and medical variables
Age, gender, body weight, height, educational level, current occupational status, and marital status of the participants were recorded. The medical variables were recorded including diagnosis, duration after diagnosis, comorbidities, BMT type, duration after BMT, and complications related to BMT.
Instruments
The Functional Assessment of Cancer Therapies–Bone Marrow Transplantation Version 4 questionnaire
The FACT-BMT was designed by McQuellon et al. to assess BMT-related concerns and its’ effects on HR-QoL (McQuellon et al., Reference McQuellon, Russell and Cella1997). The 37-item questionnaire includes two parts: the FACT-G and the BMTS. The FACT-G includes four domains: physical well-being (7 items), social/family well-being (7 items), emotional well-being (6 items), and functional well-being (7 items). A total of 27 items are scored between 0 and 4 points. The total score ranges from 0 to 108 points. Higher scores represent higher HR-QoL (Cella et al., Reference Cella, Tulsky and Gray1993). The FACT-G was translated into and validated in various languages including Turkish (Dapueto et al., Reference Dapueto, Francolino and Servente2003; Conroy et al., Reference Conroy, Mercier and Bonneterre2004; Sánchez et al., Reference Sánchez, Ballesteros and Arnold2011; Beyhun et al., Reference Beyhun, Can and Tiryaki2016). The BMTS includes 23 items specific to BMT-related problems. Among the 23 items of the BMTS, only 10 items (BMT1, BMT2, BMT3, BMT4, BMT5, BMT6, BMT8, C6, C7, and BL4) were recommended to be scored in the original study (McQuellon et al., Reference McQuellon, Russell and Cella1997). As part of the scoring of the FACT-BMT, the Trial Outcome Index (TOI) score is also calculated. The TOI is accepted as an indicator of physical/functional outcomes. The TOI score includes physical well-being and functional well-being scores of the FACT-G and BMT-specific items. Because the TOI score is responsive to change in physical and functional status, it is also widely used in studies as an endpoint (Soudy et al., Reference Soudy, Maghfoor and Elhassan2018). The FACT-BMT total score is the sum of the FACT-G and the BMTS, and ranges from 0 to 196 points. Higher scores indicate higher HR-QoL.
The European Cancer Research and Treatment Organization Quality of Life Questionnaire–Cancer30 (EORTC QLQ-C30)
To measure the convergent validity of the FACT-BMT, The European Cancer Research and Treatment Organization Quality of Life Questionnaire-Cancer30 (EORTC QLQ-C30) was used. The EORTC QLQ-C30 is a widely used HR-QoL measurement instrument for cancer patients (Aaronson et al., Reference Aaronson, Ahmedzai and Bergman1993). The questionnaire includes a total of 30 items scored between 1 (not at all) and 4 (too much) points. Three different scores are calculated as a result of the questionnaire: general health, functionality, and symptom score. Higher scores indicate higher general health and functional status. On the other hand, higher scores represent higher symptom severity. In the EORTC QLQ-C30 Turkish version study, the instrument was found valid and reliable (Guzelant et al., Reference Guzelant, Goksel and Ozkok2004).
Performance level
The Eastern Cooperative Oncology Group Performance Score (ECOG) was used to evaluate the overall performance level of the participants. Since the performance level was found closely associated with HR-QoL in cancer patients, we used the ECOG to measure the convergent validity of the FACT-BMT (Laird et al., Reference Laird, Fallon and Hjermstad2016). The ECOG is scored between 0 and 4 points. Higher scores represent a lower performance level (Azam et al., Reference Azam, Latif and Farooq2019).
The Brief Fatigue Inventory
To measure the convergent validity of the FACT-BMT, the Brief Fatigue Inventory (BFI) was used to determine the patients’ fatigue level. Since fatigue is one of the most prevalent symptoms and has a profound effect on patients’ HR-QoL (Curt, Reference Curt2000), we used the BFI. The instrument has 10 items that are scored between 0 and 10 points. One item asks whether the individual experienced fatigue in the previous week. Three items are about fatigue severity, and the last 6 items inquire about the impact of fatigue on daily activities. The instrument is scored separately as a fatigue severity score and a fatigue impact score. Higher scores indicate higher fatigue severity and higher impact of fatigue on daily activities (Mendoza et al., Reference Mendoza, Wang and Cleeland1999).
Developing the Turkish version of the FACT-BMT
The approval for Turkish adaptation of the FACT-BMT was obtained by the FACIT Organization. When we first contacted them, there has already been existed a Turkish version of the FACT-BMT translated according to the standard translation methodology of the organization. Yet, only the appropriateness of the translation was assessed in their methodology. There was a need for psychometric analysis for the Turkish version of the FACT-BMT as well as test–retest reliability and detailed validity analysis (Bonomi et al., Reference Bonomi, Cella and Hahn1996). For this reason, two researchers in this study re-evaluated the existing Turkish translation. One independent Turkish speaker fluent in English retranslated the existing translation into English. Since there was no problem regarding the translation, we decided to use this version during the pilot testing. During the pilot testing phase, participants (n = 20) were asked whether they had any difficulty regarding clarity and understanding of the items. All of the participants reported no difficulty in understanding. The results of the pilot testing were reviewed by the researchers. Since no change was required, the existing version of the FACT-BMT Turkish translation was decided to be used as the final version. In addition, our results were reviewed by the FACIT team and they approved the use of the existing Turkish translation. The Turkish translated and validated version of the FACT-BMT is available on www.FACIT.org. The administration and scoring guidelines are also available on www.FACIT.org.
Statistics
Statistical analysis was performed using the Statistics Program for Social Sciences (SPSS) Version 23 and IBM AMOS Version 22. Descriptive analysis is presented using number (n) and percentage (%) values for qualitative data, and mean (x̄), standard deviation (SD), and minimum–maximum values for quantitative data. The p-value was set at 0.05 for statistical significance in all statistical analyses. The Cronbach's alpha and the item-total correlation (ITC) values were calculated to determine internal consistency. To determine test–retest reliability, the intra-class correlation coefficients (ICCs) were calculated. Values of 0.7 for Cronbach's alpha, 0.5 for ICC, 0.20 for ITC are considered as minimum acceptable values (Kline, Reference Kline1986; Koo and Li, Reference Koo and Li2016). Normality distribution was tested with the Kolmogorov–Smirnov test and the data showed a normal distribution. Pearson's correlations between the FACT-BMT, EORTC QLQ-C30, BFI, and the ECOG performance level were calculated to determine the convergent validity. The construct validity of the FACT-BMT provided by confirmatory factor analysis (CFA) was used to assess construct validity. The Relative Chi-square Index (CMIN/DF) was calculated to assess the model fit. In addition, the root mean square error of approximation (RMSEA), comparative fit index (CFI), Tucker–Lewis Index (TLI), and standardized root mean square residual (SRMR) were calculated, and these values were accepted as acceptable if the scores were between 0.05 and 1.0 for RMSEA and higher than 0.90 for fit indices (Arli and Gurkan, Reference Arli and Gurkan2017). The exploratory factor analysis (EFA) was used to determine the construct validity of the BMTS. The Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy and Bartlett's Test of Sphericity were used to determine the assumption check of the EFA. The minimum acceptable KMO result was considered as 0.5 and the results under 0.5 were accepted as nonfactorable dataset (Field, Reference Field2000). Principal component analysis was used as an extraction method. Varimax rotation was performed to maintain proper factorization. The model fitting indices including CMIN/DF, Root Mean Square Residual (RMR), Normed Fit Index (NFI), Goodness-of-Fit Index (GFI), Adjusted Goodness-of-Fit Index (AGFI), Incremental Fit Index (IFI), TLI, and RMSEA statistics were used. In order to test the sensitivity of the questionnaire, the Student's t-test was used to compare the FACT-BMT scores according to the type of BMT (autologous versus allogeneic).
Results
A total of 114 individuals treated with BMT at Hacettepe University were included in the present study. The demographic and medical characteristics of the participants are presented in Table 1. The mean duration after diagnosis and BMT were about 2.5 and 1.5 years, respectively. The majority of the participants received autologous BMT. There was no comorbidity and BMT-associated complication in the majority of the participants. According to the ECOG performance score, the majority of the patients had limited vigorous physical activity. In the FACT-BMT, the questions about sexual life (Item Code: GS7 and BL4) were not answered by the majority of the participants (71.05%). For this reason, the items (GS7 and BL4) were excluded from the calculation of the total score and also from the reliability and validity analyses.
Demographic and medical characteristics of the participants (n = 114)
BMI, body mass index; BMT, bone marrow transplantation; ECOG, The Eastern Cooperative Oncology Group Performance Score; ADL, activities of daily living.
Validity
Construct validity
The construct validity of the Turkish version of the FACT-BMT was tested with the CFA. The calculated fit indices values were: X 2/df = 1.912, CFI = 0.880, TLI = 0.855, SRMR = 0.087, and RMSEA = 0.090. According to the CFA statistics, the Turkish version of the FACT-BMT demonstrated an acceptable fit.
The results of the assumption check for the EFA were satisfactory for the BMTS. The KMO measure of sampling adequacy value was 0.749 (moderate level), and Bartlett's Test of Sphericity Chi-square value was acceptable (X 2 = 410.123, df = 36; p < 0.001). According to the EFA results, the structure of the BMTS showed two factors. The explaining variance was 62.946% for the two-factored structure of the BMTS. The item codes of the first component were C6, C7, BMT5, and BMT8. The second component included items coded as BMT1, BMT2, BMT3, BMT4, and BMT6. The rotated component matrix is presented in Table 2. Factor loadings of the items ranged from 0.574 to 0.924. Model fitting indices (CMIN/DF = 1.293, RMR=0.035, NFI = 0.921, GFI = 0.941, AGFI = 0.898, IFI = 0.981, TLI = 0.973, and RMSEA = 0.052) revealed that the two-factored model structure had sufficient fit. The EFA results of the BMTS are presented in Figure 2.
Exploratory factor analysis of the BMTS.
Rotated component matrix of the BMTS
BMTS, Bone Marrow Transplantation Subscale; Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Bold values indicates factor loading belonging to each factor.
Convergent validity
The correlation analysis is presented in Table 3. There were strong correlations between the FACT-BMT total score and the EORTC QLQ-C30 subscores (p < 0.001). Additionally, the FACT-BMT moderately correlated with the BFI and the ECOG scores (p < 0.001). The BMTS scores were moderately correlated with the EORTC QLQ-C30 subscores, the BFI, and the ECOG scores (p < 0.001; Table 3).
Correlation analysis of the FACT-BMT and the BMTS
FACT-BMT, Functional Assessment of Cancer Therapy–Bone Marrow Transplant; BMTS, Bone Marrow Transplantation Subscale; EORTC QLQ-C30, The European Cancer Research and Treatment Organization Quality of Life Questionnaire–Cancer30; ECOG, The Eastern Cooperative Oncology Group; Pearson Correlation Coefficient Test.
* p < 0.001.
Reliability
Internal consistency
The Cronbach's alpha value of the FACT-BMT and the BMTS total scores were 0.925 and 0.744, respectively. The Cronbach's alpha value was 0.860 and 0.772 for the first and second components of the BMTS, respectively. The Cronbach's alpha values of the sub-titles of the FACT-BMT are presented in Table 4. The item analysis of the FACT-BMT is presented in Table 5. The ITC values of the FACT-BMT ranged from 0.296 to 0.737 (Table 5).
The mean scores and Cronbach's alpha coefficients of the FACT-BMT
FACT-G, The Functional Assessment of Cancer Therapies–General; BMTS, Bone Marrow Transplantation Subscale; FACT-BMT, The Functional Assessment of Cancer Therapies–Bone Marrow Transplantation; TOI, Trial Outcome Index; Auto-BMT, Autologous Bone Marrow Transplantation; Allo-BMT, Allogeneic Bone Marrow Transplantation; Student's t-test.
* p < 0.05.
Item analysis of the FACT-BMT
BMT, Bone Marrow Transplantation; ITC, Item-Total Correlation.
Test–retest reliability
The ICC values of the FACT-BMT and the BMTS total score were 0.956 and 0.948, respectively. The ICC was 0.854 for the first component and 0.959 for the second component of the BMTS.
Sensitivity
The mean FACT-BMT scores are presented in Table 4. We compared the HR-QoL scores and the BMT type to determine the questionnaire's sensitivity. There was a significant difference only in terms of the BMTS total score between the groups. Accordingly, allogeneic BMT patients had significantly lower scores than autologous BMT patients (p < 0.05; Table 4).
Discussion
To the best of our knowledge, there exists no valid and reliable questionnaire used in Turkey to assess HR-QoL specifically for patients undergoing BMT. Our results indicate that the FACT-BMT is a feasible, valid, reliable, and sensitive questionnaire to detect BMT-related health problems in Turkish patients treated with BMT. This study met the need for a culturally adapted BMT-specific questionnaire in Turkey.
The measurement of HR-QoL has been established as essential to understand patients’ perception of disease- and treatment-related side effects (Garratt et al., Reference Garratt, Schmidt and Mackintosh2002). Along with the survival rates of BMT has increased in recent years, there is a growing interest in patients’ HR-QoL level after BMT (Kopp et al., Reference Kopp, Schweigkofler and Holzner2000). Our study population consisted of both autologous and allogeneic BMT patients. While some studies in the literature included only autologous or allogeneic BMT patients (Yoo et al., Reference Yoo, Lee and Lee2006; Soudy et al., Reference Soudy, Maghfoor and Elhassan2018), some validation studies and the original version included both autologous and allogeneic patients (McQuellon et al., Reference McQuellon, Russell and Cella1997; Rivera-Fong et al., Reference Rivera-Fong, Benjet and Robles García2020). Since transplant types may affect patients in different aspects (Hjermstad et al., Reference Hjermstad, Knobel and Brinch2004; Khera et al., Reference Khera, Storer and Flowers2012), our results revealed that the Turkish version of the FACT-BMT is feasible to assess HR-QoL in both transplant types. There was a low number of patients with comorbidity or complication related to BMT in our study population. On the other hand, more than 50% of the patients had limited physical activity level according to the ECOG.
During the adaptation process, there was no need to change or modify any item of the FACT-BMT to increase patients’ understanding. When we asked participants, they reported no difficulty in understanding or interpreting the items. The only problematic items were “GS7” and “BL4” asking about sexual life and only a small number of the participants answered these items. For this reason, the “GS7” and “BL4” were not included in reliability and validity analysis. As reported in previous studies, talking about sexuality is still taboo in Turkey (Duyan and Duyan, Reference Duyan and Duyan2005; Oksuz and Malhan, Reference Oksuz and Malhan2006). For this reason, it was thought that the majority of the participants were not willing to answer these items.
Validity
The statistical analyses revealed adequate convergent and construct validity for the FACT-BMT in Turkish patients treated with BMT. These results were similar to the original and previous validation studies (Lau et al., Reference Lau, Chang and Tai2002; Yoo et al., Reference Yoo, Lee and Lee2006; Soudy et al., Reference Soudy, Maghfoor and Elhassan2018). The CFA statistics demonstrated sufficient fit indices, which were close to the acceptable level. Similar fit indices were found in the Turkish validation study of the Functional Assessment of Cancer Therapy–Brain (Arli and Gurkan, Reference Arli and Gurkan2017). The authors suggest that these values can be affected by different factors such as sample size and/or the structure of the questionnaire (Arli and Gurkan, Reference Arli and Gurkan2017). The FACT-BMT includes two parts: the first part (FACT-G) is related to general HR-QoL for cancer patients and the second part is specific to the BMT-related issues. For this reason, we suggest that the low level of fit indices for the total FACT-BMT may be derived from the structure of the FACT-BMT. According to the EFA results, the BMTS had sufficient fit and the results revealed that the BMTS had a two-factored structure. In the original version, it was indicated that the BMTS had a heterogeneous structure, so it falls into a miscellaneous category (McQuellon et al., Reference McQuellon, Russell and Cella1997). There is only one study in which factor analysis was performed for the FACT-BMT. Similar to our findings, the EFA results revealed a similar structure for the BMTS including two factors in the validation study for the Mexican population (Rivera-Fong et al., Reference Rivera-Fong, Benjet and Robles García2020). In their study, items were divided into two categories and named as “transplant implications (items: BMT1 to 4 and BMT6)” and “patient's health status (items: C6, C7, BMT5, BMT8).” They also examined one item (BMT8) separately and named it as “nursing care” (Rivera-Fong et al., Reference Rivera-Fong, Benjet and Robles García2020). In our study, this item was included in the second factor. Similar to this study, we suggest that the BMTS had a two-factored structure. The factors could be named as “transplant implications” and “patient's health status” as Rivera-Fong et al. suggested (Reference Rivera-Fong, Benjet and Robles García2020). The other interesting finding in our study was that the items with “positive statements” (BMT5, BMT8, C6, and C7) were included in one factor and the items with “negative statements” (BMT1, BMT2, BMT3, BMT4, and BMT6) were included in the other factor. The impact of statement polarity in answering/interpreting the questions/items has been a conflicting issue in the literature (Kamoen et al., Reference Kamoen, Holleman and van den Bergh2013; Holleman et al., Reference Holleman, Kamoen and Krouwel2016). It was reported that positive, negative, and bipolar questions have different effects on respondents’ answers, especially when the question is worded negatively, respondents express their opinions more positively (Kamoen et al., Reference Kamoen, Holleman and van den Bergh2013). The item polarity may have affected the participants’ responses to the items in the present study. In the Arabic version, the EFA results revealed a heterogeneous structure for the BMTS (Soudy et al., Reference Soudy, Maghfoor and Elhassan2018). McQuellon et al. (Reference McQuellon, Russell and Cella1997) stated in the original study that the BMTS should be added to the FACT-G. Therefore, it is reasonable to suggest that more information is still needed regarding the factor structure of the BMTS.
Reliability
In the present study, the alpha coefficient and the ICC values were adequate for the BMTS and the FACT-BMT. Similar to our finding, the Cronbach's alpha value of the FACT-BMT was higher than 0.80 in the original version and the other versions (Lau et al., Reference Lau, Chang and Tai2002; Yoo et al., Reference Yoo, Lee and Lee2006; Soudy et al., Reference Soudy, Maghfoor and Elhassan2018). Since the EFA results revealed that the BMTS had a two-component structure, the Cronbach's alpha values were also calculated separately in the present study. As a result, the internal consistency values were acceptable in the Turkish version of the FACT-BMT. There exists no study in the literature in which test–retest reliability of the FACT-BMT was analyzed. Our results suggest that the Turkish version of the FACT-BMT had adequate test–retest reliability.
The comparison analysis across the BMT types revealed that the patients undergoing autologous BMT had better HR-QoL than those undergoing allogeneic BMT in terms of the BMTS total score. A similar analysis was performed in the Mexican validation study, and they also found a significant difference in the FACT-BMT total score according to the BMT type (Rivera-Fong et al., Reference Rivera-Fong, Benjet and Robles García2020). Our results suggest that the BMTS has an adequate sensitivity to show differences in BMT-related quality of life.
The present study had some limitations. Firstly, we included patients undergoing BMT from a single center in Turkey. Also, the study was performed as cross-sectional, and the sensitivity to the clinically significant change of the FACT-BMT could not be measured. The strength of the present study was that we included patients living in different cities in Turkey. Since BMT could not be performed in every hospital, patients from different cities all around Turkey are referred to Hacettepe University to receive BMT treatment. For this reason, we believe that our results can be generalized to the Turkish population.
In conclusion, the present study met the need for a culturally adapted BMT-specific HR-QoL questionnaire in Turkey. Our results suggest that the Turkish version of the FACT-BMT is a feasible, valid, and reliable questionnaire for patients undergoing autologous or allogeneic BMT. In addition, it is adequately sensitive to detect changes in HR-QoL. The BMTS has a two-factored structure and we suggest that the components could be named as “transplant implications” and “patient's health status.”
Acknowledgments
The authors are grateful to the Functional Assessment of Chronic Illness Therapy agency for allowing translation and their support in all stages during the present study.
Funding
The research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Conflict of interest
None declared.
