Evaluation of an Amharic-Language translation of continuity of care satisfaction tool among postnatal mothers in Ethiopia.

Background: Beginning in the 1990s, women’s dissatisfaction with maternity services has been widely reported in the literature. However, there is a lack of consistency in the studies published over the three decades since then. The nature and availability of maternity services vary widely from country to country and even in single countries, the United States for example, there is no such thing as nationally uniform maternal care. We need cross cultural validated tools to measure women’s satisfaction with maternity care. The study reported here aimed to develop a valid and reliable continuity of care satisfaction scale for use in Ethiopia, first by finding an appropriate English-language model and then translating that into an Amharic language version that in an appropriately modified form could be used for studies of postnatal mothers in Ethiopia. Methods: An Amharic-language translation of a satisfaction with continuity of care tool was prepared based on an English-language questionnaire and this was then back translated. A team of experts analyzed the Amharic-language questionnaire as concerns the following elements: and content validity; quality of the forward and back translations; expert panel and pre-testing of the tool. A facility-based cross-sectional study was conducted among 329 postnatal mothers in Debre Berhan town health facilities, Ethiopia using the Amharic version of the questionnaire. Internal consistency of the tool (reliability) was checked by using Cronbach’s alpha. We extracted a new factor structure by carrying out exploratory factor analysis (EFA). For the extracted factor structure, confirmatory factor analysis (CFA) was conducted, and the model fit was assessed. Results: The translated tool was found to be acceptable by the experts and target groups. yielded two subscales for each component of maternity care: “information provision and relationship with care providers” and “women’s self-assessment on quality of care”. The extracted factor structure had good convergent and discriminant validity. The tool has overall Cronbach’s α value of 0.94. All three domains have acceptable internal consistency with α value of >0.70. CFA revealed a satisfactory fit between the questionnaire data and the model which provides support for the suggested factor structure identified by EFA. Conclusion: This study confirmed that the Amharic-language continuity of care satisfaction evaluation tool is a valid tool for in future studies evaluating Ethiopian women’s satisfaction with maternity care. The results from such studies could be used to aid in a trial of the continuity of care model in Ethiopia.


Background
In maternity care, women's satisfaction is an indicator of the quality of the care they receive .
Women who express satisfaction with the maternity care they have received very likely have benefited both physically and psychologically [1,2]. Some of the reasons for their satisfaction are being able to get consistent and useful information from the provider, receiving care from skilled providers who display a positive and respectful approach, and for even the mothers themselves being able to actively participate in decision making [2][3][4].
Maternity care is continuous in nature and continuity of care models are becoming more common with some studies showing that these models are improving the quality of care and satisfaction of mothers. The comparisons made in most satisfaction surveys are with older models, often seen as providing all too fragmented care, which later causes dissatisfaction with care [5].
Researchers in a number of countries have recommended dealing with these problems by establishing approaches that provide higher-level continuity of care in the belief that this will improve women's satisfaction as some studies indicate (8,9). These studies (8,9) and many others have found that the most important determinants of maternal satisfaction are high quality information [1,6]; experiencing a sense of control over the entire childbirth process [7][8][9][10][11][12], and the existence of a trusting relationship with midwives [13]. How a single pregnant mother will experience pregnancy, labour, delivery and postnatal care will depend in part on the care model followed with two extremes, one being the so-called traditiona l model [14], the other the continuity of care model. [8,10,15].
Many studies report on the availability of a wide variety of tools for measuring maternal satisfaction, each, at least in the early stages often focusing on only a select few aspects of maternity care such as antenatal care, labour and birth, an postnatal care.The literature is now so broad that attention is being paid to evaluating quality [16] and two Cochrane Reviews provide 5 comparisons of a variety of models [16] . These reviews, which focused on midwife-led continuity models of maternity care found that there is little consistency in the approaches used and that there are many difficulties in measuring maternal satisfaction across different models of maternity care [14,17]. They make clear the need to develop robust tools that are culturally adapted, translated into the appropriate languages, and validated by using appropriate validation methods.
In Ethiopia, no validated tool has been available in the Amharic language for assessing the mother's birth experience and satisfaction. Therefore, this study was set up with the aim of developing a questionnaire in the Amharic language that could be used determine how a group of mothers who had given birth in the Debre Berhan town health facilities, Amhara Regional State, Ethiopia viewed their experience and to determine if they saw advantages that might result if the continuity of care approach were to be introduced in Ethiopia.

Aim
The aim of this validation study was to determine if an existing English language questionna ire could not only be translated into a satisfactory Amharic language version but also be modified where necessary in order to be used in a survey in an Ethiopian setting.

Research questions
1. Are the scales in the Amharic-language questionnaire suitable for evaluating if maternal satisfaction could be measured with this version if a continuity of care model were to be employed? Ae these scales reliable and valid tools for the evaluation of women's experiences during the antenatal, labour, and postpartum periods of care? 6 2. Which factors (as measured by subscales for the antenatal, labour and postpartum periods) might influence women's satisfaction and experiences if the women were to be offered the continuity of care model?

Study setting
The study was conducted among mothers who were interviewed 42 days after they had given birth at the Debre Berhan town health facilities, North Shoa Zone, Amhara Regional state, Ethiopia.

Study design
A facility based cross sectional study design was conducted from November 1, 2019 to January 30, 2020. The study describes the development and psychometric assessment of the Amharic continuity of care scale or questionnaire.

Sample size determination
The sample size required for the study was determined based on the number of questions present in the questionnaire. To get the largest possible sample size, we used 10 respondents for one variable ratio which was suggested by scale development studies in which five to ten subjects were found to be needed for each item [18]. Considering the number of items in the draft scale the final calculated sample size for a 33-question questionnaire was 330 mothers.

Sampling technique
Consecutive postnatal mothers who came to the health facilities 42 days after giving birth during the study period who fulfilled the inclusion criteria were added to the study group until the required sample size achieved.

Inclusion and exclusion criteria
Postnatal mothers who visited the health facilities 42 days after delivery and were willing to participate in the study were included in the study. Mothers who had visited the health facilities before the end of the postnatal period were excluded.

Procedures followed for scale development
Five steps were taken in development of the Amharic version of the continuity of care satisfactio n scale: literature review and selection of a valid and reliable English version satisfaction scale for the continuity of care model, translation of the selected scale to the Amharic language, content validity testing, administration of a draft scale and psychometric testing.

Literature review and selection of an appropriate English version tool
In the first stage, factors associated with satisfaction with maternity care as practiced in the continuity of care model were explored by making a comprehensive review of the literature. The search focused on finding an English-language tool for use in evaluating the three aspects of the continuity of care model: antenatal, labour and delivery care, and postnatal care. The systematic review by Perlman and Davis [19] helped us to identify a tool well suited for translation to Amharic that we could use. Four tools were found to be appropriate for evaluating maternal satisfactio n.
We found the three key elements to be studied were: level of satisfaction with antenatal care, intrapartum care and postnatal care.All three of these elements are present no matter what model guides maternity care but we were only interested in examining them in the context of a continuity of care model. The tool we chose has 33 Likert questions designed to aid in evaluating the specific effect of employing a midwife-led team on maternal satisfaction [5]. Of these 33, 10 measured satisfaction with antenatal care, 10 to assess satisfaction with labour and delivery care, and 13 to measure satisfaction with postnatal care. Responses to 33 questions were to be registered on a five-8 point Likert scale with 5 indicating strongly agree and 1 strongly disagree. In all the three domains of care, the questions were framed to evaluate the quality of communication and of the relations hip between the mother and the health-care person, Permission was obtained from the original tool developer to translate the tool into the Amharic language and to then validate this version.
Content/face validity and pilot testing were undertaken to achieve these ends. We found that measuring satisfaction with each of the three elements was appropriate for measuring overall satisfaction with care offered in a continuity of care system.

Translation of the selected tool to Amharic language
In translating the English version [5] into the Amharic language, the five-step procedure outlined by Wild [20] was followed: The five steps are: 1) evaluate face and content validity 2) forward translation and expert panel discussion, 3) back translation and, comparison of the back translatio n with the original by expert panel, 4) pre-testing and 5) psychometric testing and constructive and criterion validity

Step 1: Face and Content validity
The final questionnaire in the Amharic language was presented to six experts all of whom have Amharic as their first language (two MSc midwife, Reproductive health expert, Medical doctor and two maternity nurse specialist) to review and evaluate the relevance of items to the domain of contents represented in the tool using the content validity index (CVI). The CVI is a widely used index that provides evidence for content validity by using ratings of item relevance by a panel of content experts. Experts rate each item as: 1, not relevant; 2, somewhat relevant; 3, quite relevant; 4, highly relevant. Items with ratings of either 3 or 4 are considered to be relevant. The average item CVI score is the average CVI score of the scale [21,22] and agreement for relevance at the item level should be at least 80%. Accordingly, this tool has a CVI score of 1.0 for each item. Step 2: Forward translation and expert panel discussion Translation was done by two independent translators (masters in maternal and reproductive health with background in midwifery and an English language instructor with a minor in the Amharic language). The first translator was knowledgeable about the content of the tool and the second translator was a language expert. Both were independent native Amharic speakers and fluent in the English language. The two translators sat together with the principal investigator and a discussion was held among them. Finally, a common Amharic language document was created using the two translated versions to develop a consensus. In line with this expert panel (6 in number) by involving different professionals (midwife, reproductive health, medical doctor and nurse) it was possible to evaluate the final forward and back translations and pretesting results.
Moreover, the expert panel was asked to evaluate the importance of each item in the tool with regard to each person's subdomain.

Step 3: back translation and, comparison of the back translation version by the expert panel
The Amharic language version was translated back to the original English language. This was done by using two translators different from the first translator and blinded for both the original and translated documents. These two translators had Amharic as first languageborn to mothers who spoke Amharicbut had also become fluent in English. The back translation was done independently by each. Each back translation was compared with the original English-langua ge document. Then the two translators compared the two Amharic language documents and reconciled any differences by mutual agreement to produce a final version ready for pilot testing.

Step 4: Pre -testing of the pre final tool
Pre-testing of the Amharic-language questionnaire was carried out by interviewing groups of 10 postnatal mothers for each section to be tested who were native speakers of the Amharic language.
Each was asked to evaluate the instruction, the item and possible misunderstanding of words, clarity of the response options and content format clarity. Discussion were made with the research team on the response of the participants to make changes in the tool if necessary.
Step 5: Full psychometric testing of the Amharic tool

Construct validity
In the literature, it is recommended that factor analysis be done to identify potential underlying dimensions/subscales in a scale [23][24][25][26]. In this study, to explore the factor structure of the dataset, exploratory factor analysis with Principal Axis Factoring and Promax with Kaiser Normaliza tio n rotation were used. Moreover, to justify the compliance between exploratory factors and to confirm the protection of factor structure, confirmatory factor analysis was done [27]. For the confirmator y factor analysis, maximum likelihood estimation was used for validation of the model. The goodness of fit indices were assessed by using Tuker Lewis Index (TLI; > 0.90), Comparative Fit Index (CFI; > 0.90), Root Mean Square of Approximation (RMSEA; < 0.08) and Standardized Root Mean Residual (SRMR; < 0.08) [28].

Subscale analysis
For evaluating the internal consistency of the tool, the subscales presented in the tool (antenatal care, labour and delivery care and postnatal care) were evaluated in terms of their correlation with the total satisfaction scale as well as the item-subscale correlation.

Item analysis
According to the literature, in order to have a stronger relationship with the item and the nature of content intended to be measured there must be a higher correlation coefficient. The correlation 11 coefficients < 0.25-0.30 and > 0.70 are not preferred [23][24][25]. In this study, 0.25 was taken as the lower limit for item-total correlations.

Internal reliability
Cronbach's α coefficient is concerned with the degree of interrelatedness between a set of items designed to measure a single construct. The internal consistency of a midwife-led continuity of care satisfaction tool was measured by determining Cronbach's α coefficient for each domain. A Cronbach's α value higher than 0.7 is considered as indicating that the tool is satisfactory [23,24].
The extracted tool was also evaluated for convergent and discriminant validity. Convergent validity is the extent to which scale correlates positively with other measures of the same construct as measured by composite reliability (CR ≥ 0.7), and the average variance extracted (AVE ≥ 0.5).
AVE < CR was used to establish convergent validity [29].

Statistical analysis
The data were checked for completeness and consistency and partially completed questiond were removed from the analysis. Then the cleaned and coded data were entered into SPSS version 20 windows for exploratory factor analysis. Confirmatory factor analysis (CFA) was done by using Analysis of Moment Structures (AMOS; version 25, Chicago, IL). Descriptive statistics and independent t-test were done using STATA version 14 software. Sociodemograp hic characteristics, continuity of care, past and current obstetrics and gynecologic information were described with descriptive statistics.

Ethical approval
Ethical approval was obtained from the Institutional Review Board of University of Gondar.
Permission letter was obtained from regional health bureau, zonal health department and hospital administration. An informed and signed consent was obtained from each participant. Participants were informed that their participation in this study was voluntary, the information they gave remains confidential and is used only for research purpose. Participants were told that they could also withdraw from participating in the study at any time.

Sociodemographic characteristics of study participants
A total of 330 postnatal mothers were invited to participate in this study. One was excluded from the final analysis due to incomplete response. A total of 329 postnatal mothers participated giving a response rate of 99.6%. Of these 329 mothers, 145 (44.07%) mothers were in the midwife-led group and 184 (55.93%) were in the shared-care group. There was no significant age differe nce between the two groups (p value 0.97). The mean age in the in midwife-led group was 28.

Findings of face and content validity
There was general consensus among the six experts that all the items in the questionnaire were relevant. The content validity index was 1.0. This indicates good validity. None of the items was rejected by the experts, but they recommended changing the wording of a few items to enhance face validity and content validity of the tool. Rewording of such items was carried out to remove any ambiguous phrasing and to promote easier understanding of the items. The expert panel also suggested three items that were added to further enhance domain coverage. All the items in the continuity of care satisfaction scale were required for domain coverage. After modifying the questionnaire based on the feedback from the expert panel, the questionnaire was then further tested in a pilot study that involved cognitive testing and test-retest. According to participant's response, we found the questions to be clear and understandable by the community and that they could be used to achieve face validity.

Reliability and item analysis
All the domains had positive and statistically significant item-total correlation coefficients > 0.25 (r = 0.27-0.67). Accordingly, items with low item total correlations (< 0.25) were supposed to be deleted. But in this study, there was no item with correlation value of < 0.25.  alpha was almost uniform across items, and in no instances did removal of an item from the scale result in an increase in the value of Cronbach's alpha.   Similarly, for labour and delivery domain the exploratory factor analysis provides two factors.
Factor one includes item # 1 to #3 which is represented as provision of information and relations hip 19 during labour and delivery. Factor two incudes item number 4 to 10 and designated as women assessment of quality of care during labour and delivery. For the postnatal care domain, the factor analysis provides two factors. Factor one includes item # 1 to # 7 named as provision of information and relationship during postnatal period. Factor two includes item # 8 to 13 and designated as women assessment of quality of care during postnatal care.

Known-groups validity
Known-groups validity was evaluated based on different models of maternal health care types.
This study assumed that mothers receiving midwife-led continuity of care would report greater satisfaction than mothers receiving shared model of care. Thus, we selected two subgroups among postnatal mothers, a midwife-led group (145 cases) and a shared model of care group (184 controls) and we then compared the mean satisfaction scores between the two groups by t-tests.
There were significant differences for antenatal care labour and delivery care and postnatal care domains (Mann -Whitney U test; p-value <0.001) (table 4) 25

Discussion
In this study we identified the multidimensionality of measurements of women's satisfaction with maternity care. The Amharic version continuity of care satisfaction scale was found to be content valid and translated to Amharic language successfully. Exploratory factor analysis identified two factor structures ("information provision and relationship with care providers" and "women's selfassessment on quality of care") for each domain of the scale. The extracted factor structures were found to be internally consistent with Cronbach's alpha value > 70. The appropriateness of the extracted dataset for the model was evaluated by confirmatory factor analysis and found to have good model fit with good known group and convergent validity.
Studies designed to investigate women's childbirth experience and satisfaction with care have identified the multidimensional structure of factors affecting maternal satisfaction [30,31].
Therefore, measurement tools that evaluate maternal level of satisfaction with continuity of care need to be evaluated for their reliability and validity [32]. The support that women received during  In this study we conducted exploratory factor analysis first to identify the possible underlying constructs under each domain of the continuity of care satisfaction tool. Then we did confirmator y factor analysis for assessing the appropriateness of the extracted dataset for the model. The women's assessment regarding the quality of care she received explained the largest percentage of variance in the total scale in each domain. This is not surprising given that pregnant mothers are focused on the quality of care they received in relation to improving their health and wellbeing of their baby [41]. This finding is further supported by different studies that have reported that midwifery support for women in continuity of care increases maternal satisfactio n [34,42,43]. Besides, meeting expectations during care provision and women's experience for labour and birth are well known significant predictors of maternal satisfaction with care [34,39].
In general several important factors influence the overall birth experience and satisfaction of women with maternity care [44,45]. These include: competency of the provider, respect for privacy, dignity, and respect (addressed in the subscale women's self-assessment on the quality of care); compassion and understanding, receiving information and involvement in decision-mak ing (addressed in the subscale information and communication).
All the factors identified through the EFA in the current study supported the concept of satisfactio n with maternity care as a multidimensional construct, consistent with the literature reviewed indicating that the need for developing a multidimensional tool. The factors extracted in each domain have met the needs of at least three items in each extracted factor as suggested by other scholars [46].
All of the extracted factors for each domain of maternal health care showed acceptable Cronbach alpha and demonstrated good convergent and known group validity. The satisfaction level of women in midwife-led continuity of care group were higher than women in the shared model of care. This indicates that the tool is important for measuring level of maternal satisfaction in maternity care.

Strength and limitation of the study
In this study, the tool was able to evaluate satisfaction for each component of maternity care (antenatal care, labor and delivery care and postnatal care). Besides, the tool has a communicatio n, relationship, women's involvement in decision making and women's self-assessment of quality of care components as measurement variable for satisfaction. The translated tool was tested for recommended psychometric properties and demonstrated high reliability and validity.
Although the sample size was adequate for this study, the fact that this research was impleme nted in Debre Berhan town health facilities can be considered as a limitation of the study. Further studies are needed to test the reliability and validity of the tool in rural communities.

Conclusion
We conclude that the translated Amharic version continuity of care satisfaction scale is a valid, reliable and acceptable instrument for measuring women's experiences and satisfaction in the entire course of maternity care in the health care system. In the future, the tool may contribute to assessment of the quality of care and service developments given by health care providers in continuity of care model. It can be also used as a reference for other researchers used to measure maternal satisfaction and experience related with continuity of care model.

Acknowledgments
We would like to thank the subject matter experts and language translators for their professiona l contribution in the development of the scale. We would like to thank the original English language developers for their willingness to translate and use the tool in Amharic language. We are also thankful for postnatal mothers who participated and providing us their response in this study, as well as the participating health facilities for allowing us to conduct the data collection.

Author contributions
SHB facilitated the scale development and language translation. SHB drafted the initia l manuscript. SHB, KAG, KC, ETA and HL participated in the study design and data analysis and reviewed the final manuscript. All authors contributed toward analyzing data, drafting and revising the paper, giving final approval of the version to be published and we agreed to be accountable for all aspects of the work.

Ethics approval and consent to participate
The study was approved by the Institutional Review Board at the University of Gondar (O/V/P/RCS/05/1050/2019 on March 2019) and all methods were carried out in accordance with relevant guidelines and regulations. An informed and signed consent was obtained from each participant. Participants were informed that their participation in this study was voluntary, the information they gave remains confidential and is used only for research purpose. Participants were told that they could also withdraw from participating in the study at any time.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.