Design and psychometric properties of a questionnaire to assess gender sensitivity of perinatal care services: a sequential exploratory study

Background Providing gender sensitive reproductive health service is recently emphasized by health organizations. This study aims to develop and assess psychometric properties of a questionnaire to assess gender sensitivity of perinatal care services (GS-PNCS) to be used by managers of perinatal services. Methods This study is a mixed sequential (Qualitative-Quantitative) exploratory study. In the qualitative phase, 34 participants were interviewed and the items were generated. To evaluate the validity; face, content and construct validity were assessed. The reliability was assessed by internal consistency and stability calculation. Results The content validity and reliability were demonstrated by S-CVR = 0.92 and S-CVI = 0.98, Cronbach’s α = 0.880 and the ICC = 0.980 to 0.947. Exploratory factor analysis showed 8 factors which explained more than 52.53% of the variance. Conclusion GS-PNCS is a valid and reliable questionnaire, with 49 items to assess gender sensitivity of perinatal care services and helps health care managers and planners to improve the quality of the services.

individuals' reproductive health, especially on maternal health, contraceptive use, the prevention of high-risk sexual relationships, and the transmission of STIs [6]. Gender is one of the most important factors that must be taken into account when deciding about management and providing reproductive health care services [9].
Prenatal Care services (PNCS) are important programs in reproductive health, which aim to provide quality care and counseling for mothers to achieve women's empowerment and rights [10]. However, the focus on women and the tendency to think that pregnancy, childbirth, child health and family planning are woman's job caused men to be excluded from these services. While access to health services is one of the fundamental rights of both, men and women [11].
Development of indicators and tools to assess gender sensitivity of reproductive health including PNCS are essential for monitoring and evaluating of the services and improving the quality of care. Gender sensitive health policies and programs require a thorough analysis of needs to achieve women and men health [4,12]. Reproductive health programs are not succeed if they would not be able to identify needs for gender sensitivity in their policies and implementation, and if they would not be able to response to the needs of women and men [13]. Therefore, a valid and reliable tool is necessary for health policy makers and managers to recognize gender sensitivity of reproductive health care and overcome the barriers and meet the needs to achieve quality reproductive health care services.
In this regard, a few tools were designed to measure gender sensitivity in STI services [14], male participation in PNCS [15] and reproductive health services [16]. The most comprehensive questionnaire to assess gender sensitivity in reproductive health services is available at the level of staff and facilities [17].
Since the causes of insensitivity of reproductive health services are complex and related to many other factors rather than health service providers and managers; such as institutional structure, values, priorities and process of providing services, the characteristics of health services and the culture of each country [17], this study aims to design a comprehensive, valid and reliable questionnaire to assess needs for providing the gender sensitive PNCS respecting all factors related to gender sensitivity of PNCS.

Methods
This study was a mixed sequential exploratory study to develop a valid and reliable questionnaire to assess gender sensitivity of perinatal care services (GS-PNCS). So, the study was performed in two qualitative and quantitative phases, using Waltz steps [18] for tool development.

Design of the study
To generate appropriate preliminary items, an inductivedeductive approach was conduct. Firstly, a qualitative study with the content analysis approach was performed to explain the concept and dimensions of gender sensitive PNCS from the perspectives of the experts and providers. Then, a detailed related literature review was performed [19]. The items were extracted from both studies.

The participants
The participants of the study were policy makers, providers and managers of PNCS including prenatal-, child birth and postpartum care services. The participants had at least 2 years of experience in PNCs and interest to participate in the study.

Sampling
Sampling was started purposive and continued with a snowball sampling method. It was performed with the maximum diversity in gender, work experience and education. Finally, 34 policy makers, managers and service providers in PNCs participated in the study.

Setting
All public and private clinics and hospitals in Shiraz and the headquarters of PNCS of Ministry of Health in Tehran were selected as the research environment.

Tool of the study
The guide questions for the interview and data collection were: What is the concept of "gender sensitive PNCs?; What is your understanding and experience of the specific cultural, social and religious conditions that lead to the creation of different needs of women and men in PNCs?; What are the specific educational needs of the providers for a gender based counseling and care?

Procedure of the study
Data was collected using a deep face to face individual interview by using the semi-structured interviews and continued until data saturation, when no new code of data was added to the study. The interviews were conducted by second author, Dr. Rahmanian who is PhD in Reproductive Health, and an assistant professor in the Department of Midwifery and Reproductive Health at Shiraz University of Medical Sciences. She has more than 15 years' work experiences in perinatal care services as a midwife, trainer and manager. After introducing the interviewer, the participants were informed about goals of the study and confidentiality of their personal information. Also, field notes were made during and after interview. The interviews were performed after two pilot interviews. The average duration of interviews was 60 to 90 min. All interviews were audio recorded, transcribed, typed and coded on the same day. The transcripts were returned to participants for comments and corrections.

Data analysis
The collected data was analyzed using qualitative conventional content analysis approach based on the Graneheim and Lundman's method [20]. MAXQDA v.10 was used for data management.

Data trustworthiness
Lincoln and Guba's [21] criteria was considered to evaluate trustworthiness, through assessing credibility, transferability, conformability, and dependability of the qualitative data. The codded data were checked by participants, peers and the experts and their feedback were considered.

The review process
Then an extensive review of literature was performed using key words include: "scale", "tool", "instrument", "check list", "questionnaire", "gender", "maternity", "perinatal", through databases, PubMed, Google Scholar, Science Direct, Scopus, and World Health Organization. Findings of this part did not add any items.

Item generation
The first extracted items were generated from qualitative part of the study (72 items). The review of literature did not add any items to the preliminary questionnaire.

Scoring
The scale was scored based on a 3-point Likert scale, scoring 1 to 3 for "not at all", "a little" and "much" options, respectively.

Quantitative part: psychometric assessment of GS-PNCS
In the quantitative part, the psychometric properties of the questionnaire including; quantitative and qualitative face validity, quantitative and qualitative content validity, construct validity and reliability of the tool were assessed.

Face validity assessment
For qualitative assessment of face validity, the preliminary questionnaire was evaluated by 15 perinatal care providers and managers including; 2 reproductive health specialists, 3 perinatal service managers and 8 midwives, and 2 health experts. These participants assessed difficulty, generality and ambiguity of the items. The items impact scores were calculated to assess the face validity quantitatively. At this stage, the above mentioned participants rated each item by the 5-point Likert scale from completely important to not at all important, scoring 5 to 1. To calculate the item impact score, the following formula was applied: Item impact score = Frequency (percentage) × Importance. The items with an impact score of more than 1.5 were appropriate and maintained for later stages [22].

Content validity assessment
Content validity of the questionnaire was examined qualitatively and quantitatively. In the qualitative content validity assessment, 10 experts in reproductive health and midwives were asked to comment on the items regarding the grammar of items, choice of vocabulary, placement of items, and scoring [23]. In quantitative content validity assessment, content validity ratio (CVR) and content validity index (CVI) were calculated.
The content validity ratio was assessed by 13 experts. The participants scored the items based on a 3-point Likert scale (essential, useful but not essential, not essential).
CVR was calculated through the following formula where n E stands for the number of specialists who have chosen the option "essential" and N is the total number of specialists. According to Lawshe's table [24], the CVR higher than 0.54 for 13 individuals indicate the necessity of the item at a statistically significant level (P = 0.05).
Content Validity Index (CVI) was assessed by the same 13 experts who scored items of the questionnaire based on their "simplicity", "relevance" and "clarity" using the 4-level Likert scale (scores 0 to 3 for "not at all" to "completely") based on Waltz & Bausell's content validity index [25].
CVI was calculated according to the following formula: CVI ¼ Number of raters chosing points 3 and 4 Total number of raters Items with a CVI higher than 0.79, between 0.70 and 0.79, and lower than 0.70 were considered suitable, needing modification, and unacceptable, respectively [23]. The scale's content validity ratio (S-CVR) and the scale's content validity index (S-CVI) were obtained through calculating of mean of items' CVR and CVI.

Construct validity assessment
Design of the study Construct validity of GS-PNCS was evaluated through exploratory factor analysis (EFA) through a cross sectional study.
Subjects of the study 285 PNCs' providers including prenatal, child birth and postpartum care providers, with at least 2 years of care or management experience, and willingness to participate were recruited.

Sampling
The number of samples in this study was determined 5 samples for each item of the designed questionnaire. Plichta et al. (2013) states that the required number of responders for EFA is between 3 and 10 persons per item, or a total of 100 to 200 responders [26]. Therefore, all 285 perinatal care providers of health centers and hospitals affiliated to Shiraz University of Medical Sciences in Iran, who had the inclusion criteria of the study recruited for the study using convenience method of sampling.

Setting
The subjects of the study were recruited from all 37 health centers and all 9 public and private hospitals in Shiraz.

Tool for data collection
Tool for data collection was GS-PNCS following face and content validity assessment.

Data analysis
To confirm the adequacy of the sample size for EFA, two criteria of Kaiser-Meyer-Alekin (KMO) and Bartlett Sphericity Test were measured. Adequacy of the sample size for EFA could be shown while the calculated KMO index is more than 0.8 and the p value of the Bartlett Sphericity test is less than 0.05 [26].
Then, the items were examined regarding suitability to enter factor analysis by calculating of commonalities. Next, items with commonalities of higher than 0.4 were selected for the analysis.
Quartimax rotation was used for factor analysis in this study. Quartimax minimizes the number of factors needed to explain each variable [27]. The factors of the tool were extracted using the Kaiser (1960) criterion, with the acceptance of factors having an Eigen value of more than one; and drawing the Screeplot.

Reliability assessment
To confirm the GS-PNCS's reliability, internal consistency was assessed through Cronbach's alpha calculation, and the questionnaire's stability was evaluated through the calculation of the correlation coefficient of the test-retest and the intraclass correlation coefficient.

The internal consistency assessment
The internal consistency of GS-PNCS was calculated by calculating the Cronbach's alpha coefficient, and the values above 0.7 were considered acceptable [28,29]. To assess internal consistency, 30 eligible providers of perinatal care filled up the questionnaire

Stability assessment
Stability of GS-PNCS was assessed through the test-retest method. The questionnaires were filled up by 30 eligible providers with a 2-week interval and then Pearson correlation and intra-class correlation coefficients of scores of the two tests were calculated.
The Pearson correlation coefficient more than 0.7 [23] and ICC higher than 0.4 were considered as the acceptable levels for stability [30].
The SPSS-V.21 was used to perform all statistical analyses. A summary of steps for designing and assessment of psychometric properties of GS-PNCS is presented in Fig. 1.

Results
The Findings are presented in two parts: 1) designing of GS-PNCS; and 2) assessing the psychometric properties of the GS-PNCS.
Qualitative phase: designing of the questionnaire In the qualitative section, 34 interviews with 34 perinatal care providers and manages were performed in their office or PNC clinics. Nobody refused or dropped out the interviews. Then an extensive review was performed on the related literature. These led to the explanation of the concept of gender sensitive PNCS. Then, using the extracted concept, the practical definitions of the dimentions of gender sensitive PNCS were extracted. Accordingly, the questionnaire of gender sensitive perinatal care services (GS-PNCS) is a tool that measures the responsiveness of PNCS to the needs of men and women based on their gender roles. The needs are in all dimensions of the services including structure of the services consisting of human resources, facilities and managers; processes of the services including care and educational procedures; supporting policies, consist of intersectoral co-operation and community empowerment.
The extracted items from the qualitative part and the literature review made the primary pool of GS-PNCS. These 72 items were classified in 8 subcategories and 3 categories/themes (Fig. 2).

Quantitative phase: assessment of psychometric properties of GS-PNCS
In the quantitative part, face-, content-and construct validity and then the reliability of GS-PNCS were examined.

Face validity assessment
In the qualitative face validity assessment, 8 items were omitted for ambiguity and generality. In the quantitative face validity assessment, the importance of each item was measured and the items with impact score of more than 1.5 were maintained. In this stage, all phrases received a score of more than 1.5.

Content validity assessment
In the Qualitative Content validity assessment, 7 items were deleted. In quantitative content validity assessment, no item was deleted as they obtained acceptable CVI and CVR level. Finally, content validity of GS-PNCS was Then, the questionnaire with 57 items entered the stage of construct validity assessment. Figure 1 shows the process of designing and assessing psychometric assessment of GS-PNCS and the related changes of the questionnaire.

Construct validity assessment
The EFA method was used to assess the construct validity of GS-PNCS. Sample size for this section of the study was considered 5 sample for each items. Thus, for 57 items, 285 perinatal care providers recruited for the study ( Table 1). The calculated KMO index was 0.822 and the Bartlett Sphericity test showed the correlation matrix 7715.23 with P < 0.0001 which both showed sample adequacy for EFA.
Then, Commonalities calculations for items led to omission of an item "training medical students about health education of men and families" was omitted with commonality of< 0.3.
Scree plot was used to predict the number of factors. The scree plot suggested 9 factors that became the default for factor analysis (Fig. 3) Nine factors that explained 52.53% of cumulative variance of GS-PNCS were identified using the minimum eigenvalues of 1. After Quartimax rotation and considering the factor loading of at least 0.4, the items forming each factor were identified. Then, factor 8 was merged with factor 7 due to the fact that it contained only one item. 8 items with factor loading of 0.4 were omitted. Table 2 reports the rotated factor matrix of the GS-PNCS.
The factors 1 to 8 named as "Supportive policies to promote the gender sensitive services"; with 16 terms (explaining 14.18% of variance); "Structural reforms" with 6 items (explaining 6.57% of variance); "management considerations" with 6 items (explaining 5.83% of variance); "Women's rights promotion" with 5 items (explaining 4.99% of variance); "educational considerations" with 3   items (explaining 4.78% of variance); "care considerations" with 5 items (explaining 4.75% of variance); "facilitating participation" with 5 items (Factor 7 and 8 totally explaining 7.74% of variance); "Sexual Health education" with 3 items (explaining 3.94% of variance); respectively. Table 3 demonstrates the comparison of dimensions and sub-scales in qualitative and quantitative findings in the mix study of gender sensitive perinatal care services.

Reliability
To ensure reliability, both internal consistency and stability of GS-PNCS were assessed. Internal consistency of GS-PNCS was demonstrated by Cronbach α at 0.880 for whole instrument. To investigate stability, using the test-retest method, the correlation between the two testing occasions was computed. Correlation coefficient and intraclass correlation coefficient of the whole questionnaire were reported 0.980 and 0.973, respectively. Table 4 Displays the results of questionnaire's reliability assessment. After confirming validity and reliability of the GS-PNCS, the questionnaire was finalized.

Scoring procedure by GS-PNCS
GS-PNCS was scored by a rating scale 1 to 3. The range of scores for the whole questionnaire and its subscales are presented in Table 5. The total score of the GS-PNCS and its subscales are calculated and presented as percentages. The range of scores is from 49 (0%) to 147 (100%) describing adequate gender sensitive PNC services to completely non adequate PNC services.

Description of GS-PNC
GS-PNCS is a valid scale with 49 items and 8 subscales that can be scored fron 49 to 147 (o to 100%) and measures needs for a gender sensitive perinatal care service and its higher scores shows higher needs for the gender based PNCS.

Discussion
GS-PNCS is the first tool to assess gender sensitivity and appropriateness of PNCS for men and women. This valid and reliable tool is able to measure the responsiveness of perinatal services to the gender specific needs and so helps health care managers and planners to improve the quality of PNCS. Gender equity is mentioned as the characteristics of quality of maternal care services [31] and GS-PNCS is able to evaluate the adequacy of perinatal care responsiveness to clients' needs based on their gender roles.
GS-PNCS is a valid and important assessment tool to measure quality of PNCS regarding their gender sensitivity and so useful to improve the quality. A few tools are developed and applied to assess quality of PNCS [32][33][34][35], however, they are not able to measure the gender sensitivity of the services and moreover these tools were not assessed regarding some aspects of psychometric properties especially construct validity.
GS-PNCS was designed by inductive-deductive approach [19]. The qualitative part and the literature review demonstrated 8 dimentions for gender sensitive perinatal services including; gender sensitive care and educarional process; gendersensitive facilities, human resources and mangement; and community empowerment, sopurtive policies and intersectoral cooperation for the gender sensitive services. Then, the EFA indicated eigth factors which six of them were similar to the dimentions of the qualitative part. Finding showed PNCS need managerial, structural, facilities, educational and care procedures reforms for respoding to the specific needs of genders especially men which should be supported by the supportive policies [36,37].
Face and content validity of GS-PNCS was confirmed qualitatively and quantitatively. Proper validity of a questionnaire usually refers to the vision of the target group about face validity, suitability, attractiveness, comprehensibility, culturally and socially appropriateness, logically sequence of the elements and the completeness of the instrument [19]. In qualitative face validity 8 items were deleted due to vague and duplication. In the quantitative face validity assessment,  Note.GS-PNCS = Questionaaire to assess Gender Sensituve Pernatal Care Services; a The score of total and the subscales are calculated and presented as percentage impact score of all items were higher than 1.5 and shown to be acceptable. Content validity of GS-PNCS was also confirmed by S-CVR and S-CVI 0.92 and 0.98, respectively. It shows that GS-PNCS has an appropriate sample of items for measuring gender sensitivity of the services [38].
Results of EFA showed "Supportive policies to promote the gender sensitive services"; with 16 terms explains highest variance and predictability for the sensitivity of the services "Structural reforms" and "management considerations" were second and third factors regarding their predictability for sensitivity of the services. Reproductive health policies and program formulation, has generally relied on data collected from women, while adequate policies and strategies are necessary to both men and women in their fertility control and STIs prevention care services [39]. Therefore, supportive policies are necessary as studies demonstrated that policy makers can increase implementation and effectiveness of an innovation such as making the services gender sensitive, by concentration on creating an environment that the providers perceive importance of the providing gender based services. In addition, managers should consider specific structural changes to increase positive perceptions and condition for implementation [40]. Gender roles are influenced by cultural characteristic of different communities. Besides, optimal patient care is affected by both scientific and social characteristics [41]. Therefore, community supportive policies and then structural changes by the gender sensitive management are necessary to improve Health services [42].
"Women's rights promotion" and "Facilitating male participation" were also extracted as the subscales of GS-PNCS. Studies shows that Reaching men to end gender-based violence and promote sexual and reproductive health rights of women are necessary [43]. Because of unequal gender-power relations, women are especially vulnerable but are often unable to negotiate changes in sexual behavior or to practice safe sex without the cooperation of their sexual partners. Therefore, men participation in reproductive health can span several themes. For example, men can be sources of transmission of STIs to women. When women get pregnant, their partners participate in making decisions which affects on their pregnancy such as seeking health care and place of delivery [44]. Therefore, special efforts should be made to emphasize men's shared responsibility and promote their active involvement in responsible sexual and reproductive behavior. It seems reasonable that if men are brought into a wide range of reproductive health services in such a way that they are supported as equal partners and responsible parents, as well as clients in their own right, better outcomes will be observed among both women and men [45].
Educational and care considerations based on gender specific needs and providing the sexual health education were other predictors of gender sensitivity of the services. The items were mainly related to providing care and education and sexual health services for men. Male involvement in perinatal care led to better birth outcomes. However, men are usually unavailable to attend perinatal programs because of work or feeling unwelcome at programs deemed "only for women" [46]. While they need care and education regarding their fatherhood adaptation process and roles during perinatal period [36]. Appropriate preparation for fatherhood has the potential to enhance maternal, child, and family health and even educational media such as e-health provide opportunities for men to prepare for fatherhood [37,47,48]. Internal consistency and stabilityof GS-PNCS suggest high reliability of the questionnaire for assessment of gendersisivity of the questionnaire.
The study defined the concept of gender sensitivity of PNCS as a variable that can be measured by GS-PNCS containing 8 subscales that predict 53% of variance. Finding demonstrated gender sensitive PNCS needs supportive policies at the first steps and then requires the structural reforms by some management actions and considering some reforms in care and educational procedures and providing sexual health services. Meanwhile promotion of women right and male participation both in community and in the services are necessary. GS-PNCS provides the criteria for making PNCS gender sensitive and can be a base for the reform of the services.
GS-PNCS is developed and validated to show gaps in PNCS in the health care system of Shiraz. However, regarding to high validity and reliability of the questionnaire, it can be utilized not only in Shiraz but also for similar health care system of other provinces in Iran. Also it can be used to evaluate quality of PNCS in health systems of other countries. However, its validity and reliability is recommended to be assessed after translation to other languages.
GS-PNCS is a valid and reliable tool to show gaps in structure and procedures of PNCS and so helps to show priorities for the necessary interventions for planning a comprehensive gender based quality PNCS.

Conclusion
GS-PNCS is a valid (S-CVR = 0.92 and S-CVI = 0.98) and reliable (Cronbach's α = 0.880 and the test-retest and Pearson Correlation = 0.947 and ICC = 0.980) questionnaire with 49 items to assess gender sensitivity of PNCS by a three level Likert scale. It include with 8 subscales including; "Supportive policies to promote the gender sensitive services"; "Structural reforms" "management considerations"; "Women's rights promotion"; "educational considerations"; "care considerations"; "Facilitating male participation"; "Sexual Health education" which predict 52.53% of variance.