Cross-cultural adaptation and validation of the Indonesian version of AQUAREL on patients with permanent pacemaker: a cross-sectional study

The assessment of quality of life has significant impact in device therapy. This research was aimed to translate and evaluate the validity and reliability of the Indonesian version of the AQUAREL questionnaire. We evaluated 32 patients during the cross-cultural adaptation stage and 20 patients during validity and reliability evaluation stages. Indonesian version of AQUAREL showed positive correlation between 6-min walking test and dyspnea domain (r = 0.228; p = 0.048), and showed negative correlation between NT pro-BNP and chest discomfort (r = − 0.231; p = 0.043) and dyspnea domain (r = − 0.268; p = 0.020). The total AQUAREL also showed positive moderate correlation toward total SF-36 (r = 0.543; p = 0.000). The internal consistency was good (Cronbach α = 0.728) and the repeatability between day 1 and day 8 was good, with moderate positive correlation (r = 0.581; p = 0.007).


Introduction
Pacemaker therapy is increasingly done worldwide [1]. Initially to add years in patient's life, its aim is now combined to add quality in life/Quality Adjusted Life Year (QALY) [2]. Although mortality benefit of this treatment is undoubtful [3,4], patient's perspective will influence the benefit of improving quality of life.
Health Related Quality of life (HRQOL) can be assessed by means of generic and specific questionnaires. Short Form-36 (SF-36) as one of the generic questionnaires for HRQOL has already been successfully translated into Indonesian language [5]. There's no specific questionnaire for pacemaker patient's in Indonesian Language, even though it's best to use generic and specific questionnaire together in evaluating patients [6]. Assessment of Quality of Life and Related Events (AQUAREL) (Julius Center for Health Sciences and Primary Care, Utrecht University and Heart Lung Center Utrecht, The Netherlands) was developed specifically for pacemaker patients [7,8]. Thus, we aim to adapt AQUAREL into Bahasa Indonesia and evaluate its validity and reliability.

Main text
Methods A cross-sectional study was performed between September 2014-August 2016 in outpatient cardiology clinic, Cipto Mangunkusumo National General Hospital, Jakarta, Indonesia. Patients aged > 18 years with > 3 months pacemaker implantation were enrolled. Patients with congestive heart failure NYHA III-IV, cognitive impairment, physical disability, non-optimal echo window, and other comorbidities such as pericardial and pleural effusion were excluded. Our translation and cultural adaptation of AQUAREL into Indonesian language follow the method outlined by Guillemin and Beaton [9,10] which have been reviewed by others [11,12] and has been described in detail in our group previous publication [5]. In short, it was divided into 2 phases. Phase 1 (Translation and Cross-Cultural) consisted of 6 stages: initial translation, translation synthesis, back translation, committee review, pretesting, and submission and appraisal of written reports to committee. The initial translation was performed by native Indonesian-speaker (Additional file 1: Appendixes S19, S20) and back translation completed by native English-speaker translators (Additional file 1: Appendixes S22, S23).

BMC Research Notes
Phase 2 (Validity and Reliability) consisted of two procedures. The validity test was done by comparing AQUAREL against Indonesian SF-36, 6 min-walking test (6MWT) and NT-ProBNP with Kendall Tau's correlation. The reliability of the Indonesian version of AQUAREL (Indonesian-AQUAREL) was assessed using Cronbach-α for internal consistency, Kendall's Tau, Wilcoxon Signed Rank and intraclass correlation coefficient (ICC) were used for test-retest reliability comparing results of day 1 and 8.
Sample size for phase 1 was estimated to 32 patients [13] and 20 patients for phase 2 [14]. The statistical analyses were conducted using SPSS statistics software v.23.0.

Results
There were 32 patients participated in translationcultural adaptation process and 20 patients included in validity-reliability testing of Indonesian version of AQUAREL. The characteristics of the patients can be seen in Table 1.

Correlation between AQUAREL and 6MWT and NT-proBNP
Correlation between AQUAREL and 6MWT and NT-proBNP was assessed by Kendall's Tau non-parametric test (Additional file 1: Appendix S10). Domain DYS showed weak positive correlation with 6MWT (r = 0.228; p = 0.048), while two others domains did not show any correlation with 6MWT. There was weak negative correlation between CHS and DYS with NT-proBNP. However, total AQUAREL was not statistically correlated with 6MWT (r = 0.122) and NT-proBNP (r = − 0.220) ( Table 2).
Inter-item correlations were calculated in each domain. Within CHS (number 1-6, 11, and 12), item number 1-5 showed statistically significant moderate to strong correlation to each other (Additional file 1: Appendix S2). However, item number 11 was not statistically correlated  Inter-domain correlation was tested by using Kendall Tau nonparametric test (Additional file 1: Appendix S6). Total AQUAREL showed very strong and strong positive correlation with CHS (r = 0.863) and DYS (r = 0.684), respectively. CHS also showed strong positive correlation with DYS (r = 0.684). Meanwhile, correlation between ARR and DYS only showed weak positive correlation (r = 0.331).

Repeatability of AQUAREL
Repeatability test was conducted by asking the patients to fill the questionnaire twice. Patients were asked to visit the hospital 7 days after the first meeting.
Item 13 showed a very strong positive correlation between day 1 and 8 (Additional file 1: Appendix S11). On the contrary, items 1,2,4,7,8,9,11,14, and 19 did not show any significant correlation between day 1 and 8, and item 17 could not have correlation because the data result was constant for all patients in both day 1 and 8. However, when combined as per domain, the ARR, CHS, and DYS showed strong (r = 0.632), moderate (r = 0.493), and weak (r = 0.393) positive correlation for day 1 and 8, respectively (Additional file 1: Appendix S12). For total questionnaire, the correlation between day 1 and 8 showed moderate positive correlation (r = 0.581).
There was no significant difference of Wilcoxon test for repeatability test in 20 patients between day 1 and 8 for CHS, DYS, and ARR (p = 0.826; 0.682; 0.717 respectively) (Additional file 1: Appendixes S13, S14). The t-test analysis of total AQUAREL also showed no difference (p = 0.834) (Additional file 1: Appendix S15). The results of ICC per domain of CHS, DYS, ARR, and total AQUAREL were 0.698, 0.776, 0.859, and 0.779 respectively (Additional file 1: Appendix S17). The summary of validity-reliability test of Indonesia version of AQUAREL can be seen in Table 3.

Discussion
Indonesian-AQUAREL consists of 20 item questions, which were derived from 3 domains: Chest Discomfort, Dyspnea and Exertion, and Arrhythmia. The structure of Indonesian version followed the English version of AQUAREL. Each question was followed by 5 multiple choices. Up to our knowledge, this is the second formal publication of AQUAREL questionnaire adaptation after Portuguese [15].
Domains in Indonesian-AQUAREL have significant weak to strong correlation with domains in SF-36,   [16], only showed weak positive correlation with DYS. Weak negative correlation was found between NT-proBNP as the quantitative markers of heart failure [17], with DYS and CHS. This might happen as dyspnea and chest discomfort are neither sensitive nor specific for cardiac diseases [18,19]. Indonesian-AQUAREL did not show any correlation with 6MWT and NT-proBNP. Oliveira et al. also did not found any significant correlation between AQUAREL and 6MWT in patients with permanent pacemaker [15]. The whole Indonesian-AQUAREL has good internal consistency (Cronbach-α > 0.7). However, Cronbach-α coefficient for both DYS and ARR were < 0.7. This finding was in accordance with study in Portugal which also showed similar results [15].
In CHS domain, inter-item correlation varied from moderate to strong, except item number 6, 11, and 12 which showed no correlation with most of other items in same domain. We assumed that many patients have difficulty to differentiate questions in item number 6, 11, and 12. In DYS domain, most items have weak to strong correlation, except item number 18 and 20 which show no correlation with other items. In ARR domain, only item number 13 and 14 showed correlation. We assumed patients did not considered arrhythmia as a problem that interferes their quality of life.
Inter-domain correlation showed weak to strong correlation. There were no other studies which assessed either inter-items or inter-domains correlation in AQUAREL questionnaire. Most of other studies assessed internal consistency based on Cronbach-α coefficient in each domain.
Even though Indonesian-AQUAREL has good internal consistency, some items were not correlated to each other. This might affect the ability of the questionnaire to discriminate the quality of life in some sub-populations of patients. Combination with Indonesian version of SF-36 is needed if the questionnaire wants to be used in Indonesian culture setting. The difference of correlation coefficient between inter-item, inter-domain, and total questionnaire was assumed because of the statistical method difference used to analyze the correlation (parametric and non-parametric).
Repeatability of Indonesian-AQUAREL was considered good with no significant difference between day 1 and 8. This finding was in accordance with other study which showed moderate to strong repeatability coefficient, assessed by t-test. Oliveira et al. also did not get any significant difference result between domains, which was repeated in 1-2 weeks' period [15].
The ICC of total AQUAREL both day 1 and 8 showed that CHS has moderate reliability and DYS, ARR, and total AQUAREL has good reliability [20].

Conclusion
In conclusion, the Indonesian version of AQUAREL has good convergent and discriminant validity, as well as good reliability. It's easily applicable in clinical setting due to its simple nature and comprehensible test. Thus, it could be used as a specific questionnaire to assess quality of life in permanent pacemaker patients.

Limitation
This study has few limitations. First, we did not evaluate the correlation between the receptiveness of the patients with their education level and ethnical background. Our samples were also very homogenous and from one center only, prompting the need to validate our questionnaire in other patient's groups.