Research ethics approval
Institutional Ethics Committee of the Pravara Institute of Medical Sciences, Loni, Maharashtra, India provided the ethics approval.
Study setting
The study was undertaken in the Pravara Institute of Medical Sciences (PIMS), a medical university in the rural western part of Maharashtra state in India. The university has a multi-specialty 1275 bedded tertiary health care center catering to a population of approximately 500,000. Present translation and adaptation of the PSS and the MSPSS were part of the larger study titled, ‘Yoga-based Lifestyle Intervention for Antenatal Depression (YOGA-D)’. Pregnant women attending the antenatal clinic of the Department of Obstetrics and Gynecology participated in the pre-testing of the tools and the assessment for operational equivalence while the assessment of the conceptual equivalence was completed with the teaching and technical staff members of the PIMS.
Approach
We followed the guidelines proposed by the World Health Organization (WHO) to adapt and translate both the scales (12). Some changes were made in the order of the proposed steps. The WHO guidelines recommend forward translation of the English version of the tool into the local language followed by the expert panel discussion and feedback. The revised tool is then back translated into English. The fourth step is pre-testing and cognitive interviewing and then finalizing the tool. We did the pre-testing and cognitive interviewing immediately after the feedback from the expert panel and then back translated the revised Marathi versions of the tools into English. Linguistic equivalence was assessed by a native English language researcher by comparing the original English language version of the tools with the back translated English versions. This step is not explicitly mentioned in the WHO guidelines. In addition to this, we also completed a quantitative assessment of the conceptual and operational equivalence using the Bland-Altman approach (13). This is in alignment with the process-based criteria for translation of tools proposed by Shrestha et.al. (14). The work described in this paper was guided by an approach used earlier to translate and validate the Marathi version of the Edinburgh Post-natal Depression Scale (EPDS) in the same study setting (15).
Step 1: Forward Translation
First author (RS) translated the original English version of the PSS and the MSPSS into Marathi. He is a psychiatrist by training with previous experience in translation and validation of screening tools to assess depression and anxiety. RS received graduate and post-graduate education in English while he is a native Marathi speaker familiar with the local culture and terminology used by the people in the area to express distress and social support. English language version of the PSS was downloaded from the following website: https://www.das.nh.gov/wellness/docs/percieved%20stress%20scale.pdf while the MSPSS was downloaded from https://gzimet.wixsite.com/mspss. During the forward translation, care was taken to identify conceptually equivalent words/phrases instead of literal/word-to-word translation. Step 1 resulted in version 1 of the translation (PSS_Marathi_Ver1 and MSPSS_Marathi_Ver1).
Step 2: Feedback From The Bilingual Committee
A bilingual (English and Marathi) committee of experts reviewed version 1 of the PSS and MSPSS and provided detailed feedback. The committee consisted of two female psychiatrists, one of them working in the field of public mental health, three public health researchers (two females and one male), and one male mental health service user with rich experience in the field of mental health stigma and mental health literacy interventions. The feedback provided by the committee members was collated and synthesized by RS and version 2 of the PSS and MSPSS was drafted. Several disagreements were identified, and the members were requested to provide their preference for the suggested alternatives (including the translation of the title and response options) using a Google Forms link. Based on the feedback from the committee members, version 3 of the PSS and MSPSS was finalized (PSS_Marathi_Ver3 and MSPSS_Marathi_Ver3).
Step 3: Pre-testing And Cognitive Interviewing
Two research assistants (CT and ST) with previous experience of interviewing patients and administering research tools did the pre-testing and cognitive interviewing. They were trained by the first author (RS) on the administration of the PSS and the MSPSS and the in-depth interview procedure. The interviews were conducted in an ante-natal clinic of the department of obstetrics and gynecology. Pregnant women were explained about the study procedure and those who expressed interest and willingness to complete the tools (either interviewer administered or self-administered) were included in the study. An informed consent was obtained from all the participants. Pregnant women (n = 10) were administered the Marathi version of the PSS and the MSPSS by one of the research assistants (CT). The interaction between the participants and the interviewer was observed by another research assistant (ST). She did not interfere in the interviews and made her observations on the following: 1) how easily the interviewer was interacting with the respondents? 2) whether and how many times any question had to be repeated? 3) What was the verbal/non-verbal reaction of the respondents after listening to the question? 4) Did the respondents need clarification on any question? 5) Whether the interviewer used a different word from that on the printed version?
Later, all participants (n = 10) were interviewed to further understand their experience of completing the tool. Participants were asked to repeat questions in their own words, what according to them was asked in the questions, and if they found any questions confusing or difficult to understand. They were specifically asked about the key phrases used in the questions and what came to their mind on hearing/reading these phrases and the reason for choosing a particular response option for the questions. Based on these interviews, amendments were made, and version 4 was drafted (PSS_Marathi_Ver4 and MSPSS_Marathi_Ver4).
Step 4: Backward Translation
The versions finalized after step 3 were sent for backward translation to a psychologist. She was neither aware of the original English version of the tools nor the versions used in the forward translation process earlier. The translator is a native Marathi speaker with excellent English language skills and a doctoral degree in the field of psychology. She did an independent translation of the Marathi tools into English. The back-translated English versions were PSS_English_Ver5 and MSPSS_English_Ver5.
Step 5: Linguistic Equivalence
The two versions of the tools (the original English version and the back-translated English version) were linguistically compared by a native English speaker who has led several studies in the field of global mental health and has developed and translated psychometric tools in Amharic in Ethiopia (16). Several changes were made in the tools based on the suggestions from this reviewer which resulted in the final version of the tools (version 6). These versions are in the supplementary material (Perceived Stress Scale-Marathi, Multidimensional Scale of Perceived Social Support-Marathi).
Step 6: Testing For Conceptual Equivalence
Staff/faculty members working in the PIMS (n = 10) and fluent in both English and Marathi were approached for assessing conceptual equivalence. Those who expressed interest were explained about the study and informed consent was obtained from them. Participants were not related to the study team in any form. One set of participants completed the English and the Marathi version of the PSS while the other completed two language versions of the MSPSS. For both the tools, English version was completed in the morning and Marathi version in the evening by half of the participants and other half of the participants completed Marathi version in the morning and English version in the evening. Scatter plots, Pearson’s correlation coefficient, and Bland-Altman plots (13) were used to assess the agreement between the English and the Marathi versions of both the tools.
Step 7: Testing For Operational Equivalence
Operational equivalence was assessed by comparing the self-administered and the interviewer administered versions of both the tools. Pregnant women attending the antenatal clinic of the PIMS (n = 10) and who could read, speak, and understand Marathi were invited to participate in the study. Informed consent was obtained from all the participants. Different sets of participants were involved in the PSS and the MSPSS assessments. A trained female interviewer with previous experience in field data collection first administered the tool to half of the participants. They then completed the tool on their own (self-administered version). Other participants first completed the self-administered version and then the interviewer administered version. The association between the self and interviewer administered version was assessed by estimating the Pearson’s correlation coefficient, scatter plots and using the Bland-Altman approach.