Cross-cultural adaptation of the Morisky Green Levine medication adherence scale for use in patients with Ischaemic Heart Disease in Sri Lanka

Background  Ischaemic Heart Disease (IHD) is a major constituent of cardiovascular disease in the world at present and has become the leading cause of death in Sri Lanka. Poor adherence to medication in IHD is related to increased re-hospitalisation and mortality. Therefore, continuous assessment of medication adherence is extremely important among these patients. The Morisky Green Levine Medication Adherence Scale (MGLS) is a widely used instrument to evaluate medication adherence in the clinical setting. The study aimed to translate the MGLS into Sinhala and cross-culturally adapt it to the Sri Lankan setting. Methods A Delphi process was conducted with a panel of experts where content and consensual validity of the scale was assessed after translation, back translation and pre-testing of the original scale. The ratings received from the first round of the Delphi process were noted and the amended translation re-sent for a second round. The re-ratings were evaluated for the degree of consensus and Items for the MGLS Sinhala version were chosen based on the following criteria: (i) the item was reworded or removed if 70% or more of the re-ratings were in category 0–3. If reworded, the Delphi Process was repeated for that item, (ii) the item was kept with no change if 70% or more of the re-ratings were in categories of 4–6 and 7–9. Results No items were removed from the original scale. Some items were reworded according to the experts’ suggestions to retain the conceptual meaning when translated into Sinhala. The Sinhala version of the MGLS exhibited adequate content and consensual validity. Conclusion The Sinhala version of the MGLS can be used in research, medication adherence evaluations and IHD prevention programmes in Sri Lanka.


Background
Ischaemic Heart Disease (IHD) is a major constituent of cardiovascular disease in the world at present and has become the leading cause of death in Sri Lanka. Poor adherence to medication in IHD is related to increased re-hospitalisation and mortality. Therefore, continuous assessment of medication adherence is extremely important among these patients. The Morisky Green Levine Medication Adherence Scale (MGLS) is a widely used instrument to evaluate medication adherence in the clinical setting. The study aimed to translate the MGLS into Sinhala and cross-culturally adapt it to the Sri Lankan setting. Methods A Delphi process was conducted with a panel of experts where content and consensual validity of the scale was assessed after translation, back translation and pre-testing of the original scale. The ratings received from the first round of the Delphi process were noted and the amended translation re-sent for a second round. The re-ratings were evaluated for the degree of consensus and Items for the MGLS Sinhala version were chosen based on the following criteria: (i) the item was reworded or removed if 70% or more of the re-ratings were in category 0-3. If reworded, the Delphi Process was repeated for that item, (ii) the item was kept with no change if 70% or more of the re-ratings were in categories of 4-6 and 7-9.

Results
No items were removed from the original scale. Some items were reworded according to the experts' suggestions to retain the conceptual meaning when translated into Sinhala. The Sinhala version of the MGLS exhibited adequate content and consensual validity. Conclusion The Sinhala version of the MGLS can be used in research, medication adherence evaluations and IHD prevention programmes in Sri Lanka.

Introduction
Ischaemic heart disease (IHD) is considered the foremost cause of mortality in terms of overall diseases, including cardiovascular diseases [1]. The prevalence of IHD among individuals aged over 18 years, in 2013, was estimated to be 6.4%, 6.1%, 5.3% and 3.7% in Africans, Caucasians, Latinos and Asians respectively [1]. In Southeast Asia, the number of deaths due to IHD were projected to increase from 5.73 million to 8.14 million between 1990 to 2013 [2]. In Sri Lanka, IHD is recognised as the principal cause of morbidity as well as mortality [3]. Recent studies have noted that the incidence of IHD is increasing in the younger generation in Sri Lanka. Further, it was found that urban populations and those in the higher socioeconomic classes were more at risk [3].
Patients who are diagnosed with IHD are prescribed with secondary prevention medicines (SPMs) including aspirin, beta-blockers, angiotensin-converting enzyme inhibitors, statins, oral antiplatelets and angiotensin receptor blockers. National and international guidelines have endorsed the use of SPMs, and their advantages are well recognised [4,5,6]. However, low adherence to medication, especially to SPMs, constitutes an identified barrier to gaining the expected health benefits and noncompliant patients end-up with catastrophic health consequences, a poor quality of life and enhanced demand for health care [7,8,9]. Studies have identified reduced adherence levels, typically in the range of 33% to 50%, among IHD patients [8][9][10][11][12][13]. Poor adherence to SPMs leads to a 50%-80% increased risk of mortality and a 10%-40% increased risk of cardiac hospitalisation [7,14,15].
Medication adherence is defined as "the extent to which the person's behaviour (including medication-taking) corresponds with agreed recommendations from a healthcare provider" [16]. Medication adherence comprises beginning the medication treatment regime, continuation of a prescribed medication treatment regime and cessation of the prescribed pharmacotherapy [17]. Medication adherence is measured by direct and indirect means [18][19][20][21]. Measuring drug concentrations in blood or urine and measuring added biological markers through medications are considered as direct methods. Assessing accessibility of the medication, assessing clinical response, evaluating polycounts, utilising electronic medication monitors, measurement of the refilling prescriptions, patient questionnaires or medication diaries and evaluating physiological markers are considered as indirect methods [22].
Assessing medication adherence using patient questionnaires is one of the commonest, simplest and most economical methods used in clinical settings (18). The Morisky Green Levine Medication Adherence Scale (MGLS) is a frequently utilized, selfreported assessment of medication adherence [23][24][25]. It consists of four items with a yes/no response option and the scoring system ranges from zero to four. Three levels of medication adherence on MGLS (low, medium and high adherence with 3-4, 1-2 and 0 points respectively) have been suggested by the developers based on this score [25]. A dichotomous explanation of adherence grounded on MGLS is also frequently used with 0 points signifying perfect adherence and 1+ points signifying some level of non-adherence [26]. The concurrent and predictive validity of the original MGLS are excellent, with an internal consistency reliability of alpha=0.61. This scale, when it is integrated into medical clinics can be used to strengthen adherence to medication [27]. The main focus of the present study was to translate the MGLS into Sinhala and cross-culturally adapt it to the Sri Lankan setting.

Methods
Ethics approval to conduct the study was obtained from the Ethics Review Committee, Faculty of Medicine, University of Colombo and authorisation to cross-culturally adapt the MGLS into Sinhala was obtained from the original authors. The translation process and the cross-cultural adaptation was based on universal guidelines [28]. While one professional translator translated the scale into Sinhala, a second professional translator back-translated it to English. The conceptual equivalence of the two translations was determined by a bilingual professional who determined whether both translations were congruent.
Once the translation process was finished, the first pre-test of the MGLS Sinhala version was conducted on four patients diagnosed with IHD at a Base Hospital. This was carried out in the form of a structured interview and the participants were questioned on the level of difficulty of the items, conceptual understanding of the items, suitable length for the instrument and their comfort level with the items. With feedback from the participants, the scale was revised and pre-tested on another group of patients diagnosed with IHD. Feedback was obtained on the instrument's formatting style, clarity of instructions, response alternatives and the level of difficulty of the items. Feedback obtained from both pre-tests were used to revise the scale.
After amendments based on the pre-tests was completed, a Delphi process was initiated to assess the content and consensual validity of the Sinhala version. The expert panel included two physicians, a cardiologist (clinical and interventional), a clinical psychologist and a registered pharmacist. Consensual validity was conducted to evaluate the relevance of the words used in the Sinhala version. In assessing consensual validity, each item was rated on a scale of 0 (total disagreement) to 9 (total agreement) (i) if its conceptual meaning was retained after translation (ii) if it was appropriate to be used with IHD patients (iii) if it was culturally relevant to Sri Lanka. For content validity, each item was rated on a scale of 0 to 9, as above, considering (i) if each item was an appropriate indicator of its scale and (ii) if the composite of items in the scale was adequate to measure medication adherence. Both the original English and the Sinhala versions were presented for validation. The scale was revised after the first round of the Delphi process and the ratings were repeated using the same panel after presenting the summarised ratings of the first round. The repeat ratings were further assessed for the degree of consensus. Items to be included in the Sinhala version of the MGLS were chosen through a content and consensual validation process, based on the following criteria: the item was reworded or removed if 70% or more of the re-ratings were in category 0-3. If reworded, the Delphi Process was repeated for that item. The item was kept with no change if 70% or more of the reratings were in categories 4-6 or 7-9.

Results
A few discrepancies in the wording, between the Sinhala translation and the original scale were identified during the translation process. For instance, the item "Do you ever forget to take your medicine?" was changed to "Do you forget to take your medicine regularly?" Such discrepancies are common when translating cross-culturally. Pretesting the questionnaire helped to overcome difficulties and discomfort during administration. The proper length and difficulty of the items, any discomfort when responding and ease of understanding of concepts in the items were resolved at this stage.
The Sinhala version of the MGLS showed good content and consensual validity. No items were removed from the original scale but some items were reworded according to suggestions made by the experts so as to retain the conceptual meaning of the original scale.

Discussion
The aim of the study was to cross-culturally adapt the MGLS for a Sinhala speaking population in Sri Lanka. Cultural modifications are essential when adapting a scale to a culture different from that of its origin. Modification of the MGLS was first addressed during the translation process [29] where retaining the conceptual meaning between the original and the Sinhala version was endeavored [30]. Subsequent to the translation, a Delphi process was undertaken, which further enhanced the conceptual equivalence of the original scale and its Sinhala version [31].
A Delphi process was used for content and consensual validation of the Sinhala version of the MGLS. A Delphi process is based on the evaluation of an expert panel where items of a psychometric scale, such as the MGLS, is sent individually to each expert, who then responds in an anonymous and confidential manner [32] on the items' appropriateness to assess the construct under study. The study's expert panel included two physicians, a cardiologist, a clinical psychologist, and a registered pharmacist who provided discipline-specific, subject knowledge regarding the MGLS, while comparing the original with the Sinhala version. Their content and consensual validity ratings were collated into a comprehensive document and shared with the panel. The experts were then asked to reconsider and, if required, modify their initial rating in the light of the ratings given by the other experts. The re-ratings indicated that all the experts agreed on the validity of the Sinhala version of the MGLS. Therefore, the Sinhala version of the MGLS can be used confidently in research, medication adherence evaluations and IHD preventive programmes in the Sinhala speaking population in Sri Lanka. A similar validation study of the MGLS in another Asian country, Indonesia, yielded similar results [27]. The Indonesian study was on a diabetic population, which indicates that the MGLS could be used in patient populations with diseases other than IHD, in Sri Lanka. Further research is need to validate the Sinhala version of the MGLS in various other treatment groups in Sri Lanka. This study is not without some limitations. For instance, the MGLS was not validated in Tamil nor in niche populations such as the elderly. Future studies may consider such validation.