Translation, Cross-cultural Adaptation and Validation of Hindi Version of Polycystic Ovary Syndrome-quality of Life Scale (PCOSQOL-I) - A Cross-sectional study

Abstract

Background:

With specific symptomatology and clinical presentation, the validity of standard quality of life assessment scales is questionable for polycystic ovary syndrome (PCOS) women.

Aim:

The aim of the study was to translate and validate the Hindi version of the PCOS-related Quality of Life (PCOSQOL) scale into culturally adaptable, screening tool for Indian women with PCOS.

Settings and Design:

A cross-sectional study was conducted in the outpatient department of a tertiary care hospital.

Materials and Methods:

After standard translation and linguistic validation, modifications were done as per culturally acceptable items. After the item content validity index and scale content validity ratio testing, the preliminary tool was prepared. After pilot testing and minor modifications, the final PCOSQOL for Indian Women (PCOSQOL-I) tool was applied to 287 PCOS women to check reliability, exploratory factor analysis, uniqueness, and factor loading.

Statistical Analysis Used:

R (version 4.3.0 [2023-04-21]).

Results:

The CVI of the 27-item Hindi tool was 0.73, 0.83, 0.70, and 0.72, respectively, for relevance, clarity, appropriateness, and necessity, thus indicating recommended content validity. A three-factor solution emerged through exploratory factor analysis with proportional variance for factors 1 (impact of PCOS), 2 (hirsutism), and 3 (infertility) as 0.22, 0.12, and 0.10, respectively. The cumulative variance was 0.22, 0.35, and 0.44, respectively. There was high internal consistency (Cronbach Alpha 0.89). On factor loadings, and dropping items with value <0.3, final 25-item PCOSQOL-I scale was developed where each item scored 0–4 with higher score indicating better quality of life.

Conclusion:

The PCOSQOL-I scale has good reliability and may be used to screen the quality of life for PCOS women. Further testing of the tool is warranted to check its competency to differentiate among PCOS women with different phenotypes.

INTRODUCTION

Polycystic ovary syndrome (PCOS) is the most common endocrinological disorder among reproductive-age women.[1] The syndrome is characterized by chronic anovulation, hyperandrogenaemia, and polycystic ovarian morphology which leads to irregular cycles, excessive hair growth, obesity, and infertility. The prevalence of PCOS varies from 6% to 21% worldwide depending upon the ethnic population studied and the diagnostic criteria used.[2] Prevalence in India has been reported to be 3.7%–22.5%.[3]

Due to the multifaceted pathophysiology of the disease and hormonal and metabolic dysfunctions, most women with PCOS experience one or more features including physical (hirsutism and acne), reproductive (irregular menstrual cycles and/or infertility), and metabolic (insulin resistance, metabolic syndrome, prediabetes, and type 2 diabetes) disorders. All these affect the psychological health of the women leading to poor quality of life and symptoms such as body shaming, anxiety, and depression.[4] Few authors have described PCOS condition as a feeling of less feminism among affected females thus affecting their routine quality of life.[5]

Most of the treatment strategies for PCOS usually depend upon her desire to have fertility, irregular menstrual cycles, and/or excessive hair growth. Although the disease negatively affects the quality of life,[4] social, psychological, and mental aspects of the disease are usually missed during routine treatment consultations. Considering psychological health as an important health parameter in PCOS women, the latest European Society of Human Reproduction and Embryology PCOS guideline has recommended to include psychological health screening as a part of routine health assessment and should not be ignored.[1]

As PCOS has specific symptomatology and clinical presentation, the standard quality of life assessment scales cannot be used for these women. Among the three disease-specific scales developed for PCOS-specific quality of life, the Polycystic Ovary Syndrome Quality of Life scale (PCOSQOL) scale is the latest one.[6] This is a 35-item self-administered questionnaire with four domains: impact of PCOS (16 items), infertility (7 items), hirsutism (6 items), and mood (6 items). Being in the English language, the use of this scale may pose challenges, especially to women residing in countries with English not as the primary language. Poor understandability of scale items may lead to difficulties in gathering comprehensive and accurate information about the overall understanding of different disease parameters by the patient and the effect of intervention on psychological health. These scales assessing the PCOS-specific quality of life (Polycystic Ovary Syndrome Health-related Quality of Life Questionnaire (PCOSQ), Polycystic Ovary Syndrome-related Quality of Life scale (PCOSQOL) modified Polycystic Ovary Syndrome Health-related Quality of Life Questionnaire (mPCOSQ)) have been translated and validated into a handful of foreign languages,[7,8,9] but none of these scales have been translated into Hindi language for Indian women.

The present was planned to translate the PCOSQOL scale into the Hindi language for Hindi-speaking Indian women and to modify it into a culturally adaptable easy-to-use tool. We chose this scale as this is the latest updated one assessing almost all aspects of the disease-specific psychological morbidity. The final tool was validated for reliability and internal consistency for Hindi-speaking Indian women diagnosed with PCOS.

MATERIALS AND METHODS

This was a cross-sectional study conducted in the outpatient department and infertility clinic of a tertiary care referral center. Women aged 21–38 years and diagnosed with PCOS according to modified Rotterdam criteria were screened for the study. Only those who could read and comprehend Hindi and were aware of their PCOS condition were asked to participate in the study. The study was started after ethical clearance from the Institute Ethics Committee (IEC-561/06.08.2021, RP-56/2021) (November 2021–June 2024). Informed written consent was taken from the participants willing to participate for the questionnaire filling. The study was conducted as per Helsinki Declaration

Study methodology

Forward and backward translation

After availing permission from the developers, the PCOSQOL tool was first forward-translated to Hindi by two independent experts and then backward-translated by two other experts following WHO guidelines. Both the forward and backward translations were done by one expert who was aware of the intended concept of the original questionnaire and another naive translator who was unbiased and was not from a medical or clinical background. Furthermore, the persons who did forward translation were different from those who did backward translation, and all the translators were expert and proficient in both languages (English and Hindi).

• A committee comprising 3 OBGYN clinicians, 2 psychologists, and 2 translators reviewed the items of the Hindi version of 35-item PCOSQOL tool. All 35 items of the translated scale were checked for understandability and feasibility and questions which were difficult to understand were modified according to local cultural adaptations.

• The items having similar meanings were merged and domain naming was done

• It was decided to keep all the items pertaining to a particular domain together

• After rearranging the items as per specific domains and removing the items with similar meanings, a 27-item tool was generated under 3 domains (Impact of PCOS; hirsutism items; Infertility items).

Detailed information about modification in the original scale and development of the new tool is given in Supplementary File 1.

Expert validity

Direct linear transformation of the number of raters agreeing on an item is essential as it helps to determine which items should be revised or removed from the instrument. A Delphi-based methodology was used and the 27-item tool was then tested for content validity. For this, 10 experts with knowledge about PCOS were given the tool. The content validity was tested using relevance, clarity, appropriateness, and necessity of each item. A 4-point Likert scale was used to rate each of these characteristics of every question with score 1 representing “not relevant” and 4 representing “highly relevant.” The responses were analyzed through a modified content validity index (CVI).[10] Item-level Content Validity Index (I-CVI) was calculated as Na/Ne where Na stands for the number of agreements and Ne denotes – “total number of experts.” The average of I-CVIs was calculated as the Scale level CVI. I-CVI was taken as the proportion of content experts giving each item a relevance rating of ≥3. Content validity ratio (CVR) of 0.5 indicates that the item is good as judged by the experts. Items with CVR <0.5 were relooked and reviewed and those with clinical importance were further discussed with the experts and retained back.

Face validity and pilot testing

The face validity was done with 10 experts and 5 women with the diagnosis of PCOS at least for 1 year, well aware about their condition, and were fluent in both Hindi and English. After this, pilot testing was done with 30 women with the diagnosis of PCOS. The women were asked whether the language of the tool was easily understandable, whether they were able to comprehend the questions, any modifications if needed. Minor modifications were done and a final tool with 27 questions was prepared.

Dimensionality assessment using factor analysis

The sample size for this step was based on the recommendation of 10 respondents per item given by Nunnally.[11] Since there were 27 items, and considering missing data in 5% of respondents, a sample size of 284 was calculated. The eligible screened PCOS women were given brief information about the tool and written informed consent was taken before participation in the study. The study was conducted as per the Helsinki Declaration. All the responses were collated in the Excel sheet and were subjected to statistical analysis.

Exploratory factor analysis through Promax rotation was performed to identify the degree of loading of items on the factors which enables to determination of the factor structure of the relationship between the items and their groupings with the respective factor. An item was retained if there was a loading >0.3 in any of the factors. Failure to load above 0.3, the items were deleted after examining their clinical relevance. The variance explained by a component or factor, which represents the grouping of items, was determined through the calculation of Eigenvalue. Calculation of internal consistency represented by calculation of Cronbach’s alpha, was performed, and the value of alpha was considered satisfactory if it was equal to or >0.7. Reliability was measured by of split half method.

RESULTS

After the translation, modification, and adaptation according to the local language, the 27-item tool was developed. The item-level content validity index of the tool was 0.73, 0.83, 0.70, and 0.72, respectively, for relevance, clarity, appropriateness, and necessity [Table 1]. The scale CVI of 0.75 for the modified PCOS scale was similar to and within the acceptable range of recommended content validity (0.78) as per Lynn.[12,13] In fact, the item content validity for each index was within the acceptable and recommended range as per guidelines. This indicated that the Hindi-translated and modified PCOS scale measured the constructs what it was supposed to measure.

Table 1.

Content validity index of the items of the scale items

Item		CVI_relevance		CVI_clarity		CVI_appropriateness		CVI_necessity
1		0.60		0.90		0.90		0.90
2		1.00		0.90		0.90		0.90
3		0.60		0.90		0.60		0.70
4		1.00		1.00		1.00		1.00
5		1.00		1.00		0.90		0.80
6		0.90		1.00		1.00		1.00
7		0.90		0.90		0.80		0.90
8		0.80		0.60		0.80		0.60
9		0.50		0.60		0.50		0.50
10		0.80		1.00		0.60		0.70
11		0.90		1.00		0.80		1.00
12		0.90		0.90		0.80		0.80
13		0.80		0.90		0.70		0.70
14		0.80		0.80		0.80		0.70
15		0.70		0.80		0.70		0.80
16		0.70		0.80		0.70		0.70
17		1.00		1.00		1.00		1.00
18		1.00		1.00		1.00		1.00
19		1.00		1.00		0.90		0.90
20		1.00		1.00		1.00		1.00
21		0.90		1.00		0.90		1.00
22		1.00		1.00		1.00		1.00
23		1.00		0.90		0.90		0.90
24		1.00		1.00		1.00		1.00
25		0.80		0.90		0.90		0.80
26		0.90		1.00		0.90		1.00
27		0.90		0.90		0.90		0.90

CVI: Content validity index

Table 2 describes the baseline characteristics of 287 study participants. The mean age of the study participants was 28.5 ± 4.3 years and the median duration of infertility was 4 years.

Table 2.

Baseline sociodemographic characteristics of the study participants

Characteristics		Results (n=287)
Age (years)*		28.5±4.3
Socio-economic status@
Upper class		49 (17.1)
Upper middle		113 (39.3)
Lower middle		83 (28.9)
Upper lower		42 (14.7)
Patient education@
Postschool		178 (62.03)
School		109 (37.97)
Patient profession@
Employed		85 (29.6)
Housewives		202 (70.4)
Weight (kg)*		60.12±9.9
BMI (kg/m2)*		24.10±4.1
Type of infertility@
Primary		219 (76.3)
Secondary		68 (23.7)
Duration of infertility (years)#		4 (2–6)

Data presented as *Mean±SD, ^@n (%), ^#Median (range). BMI=Body mass index

Table 3 describes the descriptive analysis of the items. After pilot testing and minor modifications, the tool was finalized and named as PCOSQOL for Indian Women (PCOSQOL-I) scale.

Table 3.

Descriptive analysis of the scale items

		n		Mean		SD		Median		Trimmed		Mad		Minimum		Maximum		Range		Skew		Kurtosis		SE
Item 1		287		2.82		0.78		3		2.76		1.48		1		5		4		0.49		0.22		0.05
Item 2		287		2.72		0.81		3		2.62		1.48		1		5		4		0.85		0.71		0.05
Item 3		287		2.25		0.59		2		2.18		0		1		5		4		1.61		4.54		0.03
Item 4		287		3.66		0.82		4		3.67		1.48		2		5		3		−0.14		−0.51		0.05
Item 5		287		3.02		1.07		3		2.91		1.48		1		5		4		0.47		−1.02		0.06
Item 6		287		2.67		0.86		2		2.55		1.48		1		5		4		0.95		0.52		0.05
Item 7		287		3.46		0.88		3		3.45		1.48		1		5		4		0.18		−0.54		0.05
Item 8		287		3.4		0.88		3		3.38		1.48		1		5		4		0.11		−0.44		0.05
Item 9		287		2.91		0.9		3		2.83		1.48		1		5		4		0.57		0.08		0.05
Item 10		287		2.45		0.77		2		2.33		0		1		5		4		1.2		1.28		0.05
Item 11		287		3.08		0.89		3		3.03		1.48		1		5		4		0.29		−0.36		0.05
Item 12		287		3.17		0.82		3		3.13		0		1		5		4		0.42		0.09		0.05
Item 13		287		2.87		0.79		3		2.82		1.48		1		5		4		0.4		−0.43		0.05
Item 14		287		3.09		0.82		3		3.04		0		1		5		4		0.42		−0.14		0.05
Item 15		287		2.82		0.85		3		2.73		1.48		1		5		4		0.73		−0.13		0.05
Item 16		287		2.52		0.79		2		2.39		0		1		5		4		1.15		1.17		0.05
Item 17		287		2.53		1.18		2		2.42		1.48		1		5		4		0.75		−0.34		0.07
Item 18		287		2.41		1.1		2		2.3		0		1		5		4		0.91		0.16		0.06
Item 19		287		2.34		1.11		2		2.23		1.48		1		5		4		0.77		−0.08		0.07
Item 20		287		2.21		1.08		2		2.09		1.48		1		5		4		0.65		−0.17		0.06
Item 21		287		2.85		1.59		3		2.82		2.97		1		5		4		0.02		−1.6		0.09
Item 22		287		3.63		0.86		4		3.67		1.48		1		5		4		−0.31		−0.24		0.05
Item 23		287		3.18		0.94		3		3.12		1.48		1		5		4		0.24		−0.52		0.06
Item 24		287		3.53		0.86		4		3.55		1.48		1		5		4		−0.21		−0.03		0.05
Item 25		287		3.37		0.96		3		3.36		1.48		1		5		4		−0.13		−0.49		0.06
Item 26		287		3.24		0.91		3		3.19		0		1		5		4		0.26		−0.17		0.05
Item 27		287		3.3		0.97		3		3.27		1.48		1		5		4		0.11		−0.6		0.06

SD=Standard deviation, SE=Standard error

Figure 1 describes the correlation plot of the scale items. The scree plot of the principal component analysis and common factor analysis suggested a three-factor solution as indicated by the variables with Eigenvalues above 1 [Figure 2]. Figure 3 describes the uniqueness of the scale items.

Figure 1.

Correlation plot

Figure 2.

Scree plot obtained following factor analysis – Circles above Eigenvalue 1 represent the factors which can be extracted for the scale (Eigenvalue of principal component analysis represented as bold circles and eigenvalue of common factor analysis represented as plain circles)

Figure 3.

Uniqueness of the items

Table 4 describes the factor loading of the tool items. After removing the coefficients below 0.3, the factor loading of 25 items ranged from 0.34 to 0.97. Coefficients below 0.3 were suppressed and removed. The proportional variance for factors 1, 2, and 3 were 0.22, 0.12, and 0.1, respectively. The cumulative variance was 0.22, 0.35, and 0.44, respectively. Although the variance of the scale is low, future studies will look into it with a larger sample size and with varimax rotation.

Table 4.

Factor loading of the scale items

Loadings, unrotated								After promax rotation and removing factor loadings <0.3
Item		Factor 1		Factor 2		Factor 3		Item		Factor 1		Factor 2		Factor 3
Item 1		0.41				0.22		Item 1		0.38
Item 2		0.50		0.27		0.25		Item 2		0.45
Item 3		0.24		0.13		0.16		Item 3
Item 4		0.44		0.11		0.33		Item 4		0.35
Item 5		0.15		0.28		0.23		Item 5
Item 6		0.43		0.21		0.29		Item 6		0.34
Item 7		0.34		0.16		0.83		Item 7						0.97
Item 8		0.27		0.11		0.90		Item 8						1.09
Item 9		0.39				0.39		Item 9						0.36
Item 10		0.61		0.11				Item 10		0.70
Item 11		0.55		0.24		0.33		Item 11		0.46
Item 12		0.62				0.18		Item 12		0.69
Item 13		0.51				0.15		Item 13		0.54
Item 14		0.75				0.11		Item 14		0.87
Item 15		0.65						Item 15		0.75
Item 16		0.63				0.12		Item 16		0.70
Item 17		0.20		0.91		0.12		Item 17				0.94
Item 18		0.21		0.91		0.11		Item 18				0.94
Item 19		0.16		0.86		0.11		Item 19				0.88
Item 20				0.58				Item 20				0.61
Item 21		−0.12		0.36				Item 21				0.39
Item 22		0.47				0.18		Item 22		0.47
Item 23		0.58		0.16		0.16		Item 23		0.61
Item 24		0.56				0.25		Item 24		0.56
Item 25		0.48				0.22		Item 25		0.47
Item 26		0.57						Item 26		0.67
Item 27		0.58				0.25		Item 27		0.58
												Factor 1		Factor 2		Factor 3
								SS loadings				5.98		3.33		2.61
								Proportion Var				0.22		0.12		0.10
								Cumulative Var				0.22		0.35		0.44

Cronbach’s alpha was calculated to be 0.89, indicative of high internal consistency of the final scale. Reliability, as calculated by the split-half method, had a maximum split-half reliability (lambda 4) of 0.95 (indicating that the scale has excellent reliability) and a minimum split-half reliability (beta) of 0.76; Guttman Lambda 6 of 0.94, and Guttman Lambda 2 was 0.91; average interitem correlation had r = 0.26, with median = 0.2.

Thus, we developed a 25-item polycystic ovary syndrome quality of life scale-India (PCOSQOL-I) scale (3 domains: Impact of PCOS (item 1–14), hirsutism (item 15–19), and infertility (item 20–25) where the score of each item ranges from 1-4; with maximum score 100 [Figure 4]. Higher score means better quality of life.

Figure 4.

Final Polycystic Ovary Syndrome-related Quality of Life for Indian Women tool with 25 items

DISCUSSION

The present study was designed to translate, modify, and validate the PCOSQOL scale for the Hindi language. The new scale PCOSQOL-I (25-item scale) has shown good internal consistency and reliability to assess the PCOS-specific quality of life for Hindi-speaking Indian women suffering from PCOS.

Considering the higher prevalence in Southeast Asian countries, it is important to have country or language-specific tools to assess the psychological health of PCOS women. The tool should focus on their perception of PCOS-related symptoms, impact on quality of life, key concerns, priorities for management, and effect of different disease-specific treatment interventions.

The first disease-specific scale for PCOS was PCOSQ (26-item questionnaire) developed in the United States by Cronin et al.[14] and has been translated into other languages also[8,9,15,16,17,18] including Pakistan and Chinese versions. But later, this scale was modified in 2007 by Barnard et al., who modified this scale into mPCOS scale with 30 items under 6 domains to overcome the limitation of inability to assess all the aspects of psychological health.[19] This scale was also validated in foreign languages.[19,20]

Considering the criticism of the PCOSQ scale regarding the assessment of all aspects of health, the PCOSQOL was developed by Williams et al. in 2018 which covered the impact of disease on health to represent different phenotypes of PCOS. The authors of this article first explored PCOS-related morbidities[21,22] and then developed a new 35-item scale with four domains including the impact of PCOS, mood items, hirsutism items, and infertility items. The authors claimed to include the psychological, environmental, and social domains in addition to items reflecting the impact of symptoms.

Jiskoot et al. further translated and compared the Dutch versions of the mPCOS scale and PCOSQOL scale and concluded that both scales are comparable and equal to assess the quality of life among PCOS women.[7] Most of the women had no preference of one scale over the other so both scales were found similar. However, no study has translated the PCOSQOL scale into any of the foreign languages.

Ours is the first study to translate the PCOSQOL scale into an Asian language and validation of the Hindi-translated version. Although the scale has shown good reliability and internal consistency, further data should be generated to study the competency of the scale to identify the severity of diseases and its correlation with different phenotypes.

The limitations of the study are that the study participants were infertile PCOS women attending a tertiary care referral hospital for infertility. As most of these patients have already received multiple fertility treatments, the generalizability of the tool to all PCOS women should further be tested with its application to the broader population of PCOS. Besides, we could not do test–retest reliability and consistency testing with the original questionnaire.

CONCLUSION

PCOSQOL-I scale may be used for quick assessment of the quality of life among Hindi-speaking Indian women suffering with PCOS.

Author contributions

RM and SS developed the idea of translation after screening all the available tools. SK, MG, AK, RC, RR, and AN helped to conduct different steps of translation and validation, screening, and recruiting patients. AKJ and SK collected all the data. PH and SS did the formal analysis of the study.

Conflicts of interest

There are no conflicts of interest.

Data availability statement

Data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

SUPPLEMENTARY FILE

Supplementary File 1: Following modification were done in the original tool to develop PCOSQOL-I tool

Impact of PCOS

Of 16 impact items, modification was done by removing duplications and difficult questions, and final 11 items were retained:

• Items 4 and 10 were clubbed to one item (2)

• Items 15,20 and 29 were clubbed to one (item 6)

• 18, 19, 27, and 33 were removed, and Items 4 and 5 were added in easily understandable and culturally acceptable items.

Mood items

• Out of 6 items, 5 items were retained (item 2 was removed).

Impact of hirsutism

• Items 22 and 26 were clubbed to one (item 18)

• Items 7 and 21 were removed and replaced by new items 20 and 21. These two items were added as treatment modalities for hirsutism also causes significant psychological stress to PCOS women and the frequency of these treatments also indicates the severity of the hirsutism.

Impact of infertility

In the original scale, there were 7 items pertaining to infertility (1, 8, 9, 23, 25, 31, and 34). All the items were clubbed and due to overlapping items with similar meaning, 23 and 25 were clubbed to one item 9 was modified to item 22 in the new scale.

So the scale prepared in Hindi had total 27 items which was ready for further testing.

Funding Statement

Nil.