Cross-cultural adaptation of the Type D Personality Scale for use with patients diagnosed with Ischemic Heart Disease in Sri Lanka

Introduction Ischemic Heart Disease (IHD), a major cardiovascular disease globally, has become the primary cause of death in Sri Lanka. Negative affectivity (NA) and social inhibition (SI) are two personality traits which increase the risk of IHD. The Type D Scale (DS-14) evaluates a person’s general level of distress on NA and SI. However, DS-14 has not been translated and validated into Sinhala in Sri Lanka.


Introduction
Ischemic heart disease (IHD) is considered the foremost cause of death [1], and disability around the world and its prevalence are approximately one in every six deaths in Western countries [2]. Although the mortality of IHD has declined gradually over the last decades in Western countries, it is still responsible for about onethird of all deaths in individuals older than 35 years of age [1]. Further, a study conducted among individuals over 18 years revealed that the IHD prevalence was 6.1, 6.4, 5.3 and 3.7% in Caucasian, African, Latino and Asian populations, respectively [3]. Moreover, it is the prevalence increases with age, and males are at a higher risk [4]. A study has indicated that IHD deaths in Southeast Asia have amplified from 5.73 to 8.14 million from 1990 to 2013 [5]. Ischemic Heart Disease has been identified as the foremost cause of morbidity and mortality in Sri Lanka [6] where risk factors are more significant in urban areas, higher socioeconomic classes and among younger people [7,8].
Psychological distress is closely related to the advancement of IHD and directly associated with the medical consequences of IHD [9,10]. Type D (distressed) personality is an emerging risk factor in patients with cardiovascular diseases (CVDs) and closely connected with poor prognosis, reduced health status, and mental health issues such as depression, anxiety and Post Traumatic Stress Disorder [12][13][14].  is an instrument which is used to evaluate a personality type commonly associated with IHD, and it was developed from the 16-item, DS-16 [15] scale. The impact of Type D personality on the prognosis of CVDs has been widely discussed with the use of both instruments [16,17]. Type D personality is the person's consistent inclination to experience negative affectivity (NA) and social inhibition (SI) [18]. While NA signifies the inclination to experience negative emotions (i.e. depression, anger, anxiety), SI signifies restraining from disclosing feelings and opinions in social interactions to avoid others' disapproval [18].
Although several studies have been conducted in Sri Lanka on diverse aspects of IHD, no instrument has been found in Sinhalese with proven reliability and validity which assesses Type D personality in IHD patients. The majority (>75%) of Sri Lankans speak the Sinhala language as the mother tongue, and having an instrument in the Sinhala language to evaluate type D personality will be valuable in the treatment and prevention of IHD which has become a crisis in the country. Therefore, the present study was carried out to cross-culturally adapt and evaluate the validity and reliability of the DS-14 for use in patients diagnosed with IHD in Sri Lanka.

Methods
The DS14 is consisted of NA and SI where the subjects are rated on their personality on a 5-point Likert scale ranging from 0 = false to 4 = true. The NA and SI scales can be scored as continuous variables (ranging from 0-28) to assess these personality traits in their own right. A cutoff of 10 on both scales is used to classify subjects as Type D (i.e., NA≥10 and SI≥10).

Study population
Participants were chosen using the systematic sampling method, from patients who are diagnosed with IHD attending medical clinics at a Base Hospital in the Southern Province of Sri Lanka. Patients between 18-60 years and diagnosed with IHD within the last three months of the data collection were selected for the study. The selection of IHD patients were done by going through their clinic records and by further confirming the diagnosis of IHD with the physicians working in the medical clinics at the Base Hospital. Among them, patients of various ethnic groups and competent in the Sinhala language were selected. Patients who had significant health problems, physical disabilities, previously diagnosed with significant health problems (data were gathered through previous clinic reports), critically ill and pregnant women were omitted. The DS-14 is consisted of 14 items and the participants per item were kept at a ratio of 5:1 as per the literature [31]. Accordingly, the calculated total sample size of the validation study was 140 by allocating equal subsamples for CFA (n=70) and EFA (n=70).

Translation and pre-test
Initially, the instrument was translated and then it was back-translated, the translated version back into English by two subject specialized professional translators. Afterwards, to determine the conceptual equivalency of the translated and back-translated versions, they were scrutinised by a bilingual professional in the field of behavioural sciences, and the corrections were made accordingly.
A structured interview with four patients diagnosed with IHD was carried out as the pre-test of the DS-14 Sinhala version. Minor revisions were made according to the comments made by these patients on aspects such as the ease of comprehending the instrument, difficulty level when responding, and the length of the DS-14 instrument's items.

Content and consensual validity
The content and consensual validity were determined by an expert panel of two psychiatrists and four clinical psychologists through a Delphi process. In terms of consensual validity, the panel of experts rated each item, on a scale of 0 (total disagreement) to 9 (full agreement): (i) if an item's conceptual meaning was maintained after it was translated to Sinhala; (ii) if the items were suitable to be used with patients diagnosed with IHD, and (iii) if the instrument items were culturally relevant to Sri Lanka. Concerning content validity, the panel of experts again rated the instrument from 0 to 9: (i) if each item was a suitable indicator of its scale/sub-scale, and (ii) if the combination of items in the scale/sub-scale was adequate to measure the concept assessed by the scale/sub-scale. The ratings received from stage one was summarised and presented for a second round of the Delphi Process. The repeat ratings were evaluated for their degree of consensus.

Ethical considerations
Initially, the permission was obtained from the original author to translate, cross-culturally adapt and validate the DS-14 into the Sinhala language. Ethics approval was obtained from the Ethics Review Committee, Faculty of Medicine, University of Colombo, Sri Lanka (EC-18-011) and institutional permission was obtained from the Medical Superintendent of the Base Hospital in which the patient data was collected. Informed written consent was obtained from all the study participants.

Statistical analysis
Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) were performed by the R (3.6) and R Studio (1.2) statistical software to evaluate the invariance of the instrument, and model-data fit. The data were divided randomly into two equal subsamples; one subsample was used to analyse EFA (n=70), and the other was used to do CFA (n=70) to confirm the factorial structure of the original study.
To evaluate the sample adequacy subsample for the exploratory factor analysis, the Kaiser-Meyer-Olkin (KMO) measure and Bartlett test of sphericity were performed, and KMO values of ≥0.70 were considered average [32].
A scree plot was shown to determine the number of components to extract. EFA by principal components was performed with Varimax rotation. Reliability was verified with internal consistency, using Cronbach's alpha, and coefficients of ≥0.70 were considered to possess a satisfactory internal consistency [33]. In CFA, following values were calculated and compared: the comparative fit index (CFI), the Tucker-Lewis index (TLI), the incremental fit index (IFI), the root-mean-square error of approximation (RMSEA), the ratio of chi-squared test χ 2 to the degree of freedom (χ 2 /df). The two-factor model of the original study was tested. The model was considered as consistent with the experimental data at the values of CFI [34], TLI [35] and IFI≥0.90, RMSEA [36] less than 0.08 and χ 2 /df less than five.

Delphi process
The following principles were used in the Delphi process to select the items in DS-14 Sinhala version; (i) if 70% or more of the re-ratings were in category 0-3, that item/sub-scale was omitted or reworded to make it acceptable. If reworded, the Delphi Process was repeated for that item/sub-scale, and (ii) if 70% or more of the reratings were in categories 4-6 and 7-9 (summatively), that item/subscale was retained. None of the items was removed from the original instrument. Some items were reworded according to suggestions made at stage one, by the experts, to retain the conceptual meaning when translated to Sinhala.

Factor structure Exploratory factor analysis (EFA)
The KMO-index (0.83) and Bartlett's test of sphericity (χ 2 (91)=559.68, p<0.05) showed appropriate sample adequacy on the data to carry out the factor analysis. EFA (Principal components, Varimax rotation) of the DS-14 items showed two components (Eigenvalue > 1) explaining 62.90% of the variance (34.69% of the first factor and 28.21% of the second factor ( Table 2). The scree plot showed a significant break after the 2 nd component. Using Cattell's scree test, and it was confirmed to retain two components ( Figure 1).

Confirmatory factor analysis (CFA)
CFA was performed to test the predicted two-factor structure of the DS14 directly. The second sample (n=70) was used to check the factorial structure of the original study. A CFA of the two-factor structure of the DS-14 Sinhala version indicated a good model fit. The two-factor model gave a chi-squared value of 559.68 (df=91) (p <0.001), and for the 2-factor solution we found, CFI=0.919, TLI=0.903 and RMSEA=0.104 (90% CI 0.074-0.133), confirming full conformity of the theoretical two-factor model with the obtained data.

Reliability analysis
Cronbach's alpha values on the subscale of NA and SI were 0.93 and 0.88, respectively. Mean inter-item total correlations were identified as 0.645 for the NA and 0.522 for SI items. The values for "alpha if item deleted" in both subscales NA and SI were lower than corresponding subscale values in the alpha of 0.001. Therefore, the Cronbach's alpha coefficient proved an acceptable internal consistency for the two subscales of the DS-14 Sinhala version.

Discussion
Negative affectivity (NA) and social inhibition (SI) are two traits which increase the risk of IHD(11) and impacts its treatment and prevention. The DS-14 is a widely used instrument, with excellent psychometric properties, which evaluates an individual's general level of distress, based on NA and SI. The present study aimed to translate, cross-culturally adapt, and assess the psychometric properties of the Sinhala version of DS-14, as an assessment tool to detect type D personality among the Sinhala speaking community.
Forward translation, back translation, bilingual scrutinization, and pre-tests were undertaken to obtain acceptable content and consensual validity of the instrument. Two rounds of the Delphi process tested content and consensual validity. The Delphi panel consisted of experts in mental health -two psychiatrists and four clinical psychologists. None of the items was removed from the original instrument, and rewording of some items was done according to suggestions made by some experts, such that the conceptual meaning was retained when translated to Sinhala.
EFA was performed to find out the probable underlying factors, and CFA was utilized to confirm the factor structure. Exploration of factor structure was led with the first subsample data, and verification of factor, with other subsample data, considering the recommendations [37,38]. Factor analyses of the DS-14 in several different languages had identified NA and SI items loading on the two factors [22,24,30,39,40]. Considering the CFA data, an acceptable model fit with construct validity was detected, and the values of the fit guides for the two-factor model were acceptable and closer to the desired level. Therefore, the original two-factor model best fitted and clarified the items of the DS-14 Sinhala version in the tested sample.
Reliability analysis was performed through evaluating the internal consistency using Cronbach's alpha. Cronbach's alpha coefficients were calculated for each subscale were NI=0.93 and SI= 0.88; all of which lie above the commonly accepted thresholds for internal consistency. Results are equivalent to those of the original study and other language versions [22,24,30,39,40] as illustrated in table 03.

Conclusion
The validated DS-14 Sinhala version has indicated acceptable psychometric properties, in par with the original instrument. It can confidently be applied in research, investigation of Type D personality and IHD preventive programmes.

Data availability
The data used to support the findings of this study are available with the corresponding author upon request.

Conflicts of interest
All authors declare that they have no conflicts of interest.