Indonesian Self-Risk Assessment for Cervical Cancer (SiNara): Instrument Development and Validation

Corresponding Author: Devita Madiuw Faculty of Health, Universitas Kristen Indonesia Maluku, Indonesia Email: imasulydevita@gmail.com Background: Women’s ignorance of cervical cancer risk factors has caused low participation in the screening of the disease. Women can independently assess cervical cancer risk factors in themselves through a risk assessment instrument. However, no instruments were found that assessed cervical cancer risk based on the characteristics of Indonesian women. Purpose: This study aimed to develop and validate an Indonesian self-risk instrument for cervical cancer. Methods: The instrument was developed based on scale development following the guidelines by DeVellis (2017). A cross-sectional study was implemented to validate the instrument. A total of 20 women were involved in the pre-testing, and as many as 200 women were included in the instrument testing. Based on the scale development guidelines, data collection was initiated with a literature review to determine the instrument construct and item pool. A total of 29 articles were used in the formulation of the item pool and resulted in 38 items for validity and reliability testing. Content validity ratio (CVR) and content validity index (CVI) were used to test the content validity of the instrument, which was reviewed by three experts. Exploratory factor analysis (EFA) with principal component analysis (PCA) method and Kuder-Richardson 20 (KR-20) were used to evaluate construct validity and internal consistency reliability, respectively. Results: A total of 26 items met the content validity and 21 items met the construct validity evaluation, with five items being removed because they had a loading factor value of <0.4. After the validity were evaluated, the instrument was reduced from 38 items to 21 items. The internal consistency reliability with Kuder-Richardson 20 (KR-20) was 0.807. Conclusion: The Indonesian self-risk assessment for cervical cancer (SiNara) instrument is of good validity and reliability. However, it needs to be tested in other settings using larger samples to measure its psychometric properties as well as its applicability and acceptability.

formulation of the item pool, the determination of the instrument format, the assessment of the items by the expert panel, item testing, item evaluation, and scale optimization (DeVellis, 2017). Content Validity Ratio (CVR) and Content Validity Index (CVI) were used to evaluate content validity, while exploratory factor analysis (EFA) with the principal component analysis (PCA) method was used to evaluate construct validity. A reliability test was carried out using .

Setting and samples
Instrument testing was conducted in Bandung City, West Java, Indonesia. A total of 20 women were included in the pre-test, and the instrument was then tested on 200 women via a Google form. The sample size was determined based on the assumption that a sample size of 200 is sufficient for most cases with a factor analysis of no more than 40 items (DeVellis, 2017), and must exceed the number of items, from 2 to 20 respondents per item with a minimum number of 100 respondents (Anthoine et al., 2014). The inclusion criteria were being married, had never been diagnosed with cervical cancer and were willing to participate in the study.

Measurement and data collection
The content validity ratio (CVR) and content validity index (CVI) assessment form was used by experts to evaluate the items. The first step was developing the instrument by determining the construct through a literature review. The search for articles was carried out on the PubMed, CINAHL, Google Scholar, and Garuda databases, with inclusion criteria including the publication year of 2010-2019, using Indonesian or English, and articles were published in the Science and Technology Index (SINTA) and Scopus indexed journals. The keywords used in Indonesian were "risk factors AND cervical cancer" and in English were "risk factors AND cervical cancer AND Indonesia." A total of 29 articles met the inclusion criteria and were used in the formulation of the item pool.
Family history of disease, including family history of cancer and family history of cervical cancer (Arfailasufandi et al., 2016;Rahmah et al., 2017;Yuviska & Amirus, 2015); and lifestyle, including smoking habits, history of cigarette smoke exposure (Natphopsuk et al., 2012;Nindrea, 2017;Dewi et al., 2013;Putri et al., 2019), no daily fruit and vegetable consumptions (Rasjidi, 2009;Syatriani, 2011) were taken. Personal hygiene included no changing sanitary napkins every 4-6 hours a day during menstruation, no drying out the vagina after micturition, no washing the genital area after sexual intercourse, and using panty liners and public toilets frequently (Dianti & Isfandiari, 2015;Fitrisia et al., 2019;Dewi, 2017). Sexual behaviors which were pooled included the number of sexual partners (Umami et al., 2017;Yuliani et al., 2019), bleeding after sexual intercourse, and intercourse during menstruation (Damayanti, 2013;Nindrea, 2017). The use of contraceptives, like types of contraception and duration of using the contraceptive (Dewi et al., 2015;Fitrisia et al., 2019;Jasa, 2016;Partiwi et al., 2015); and cervical cancer prevention measures, such as the history of HPV vaccine and IVA or pap smear screening were derived (Dewi et al., 2015;Kusumawati et al., 2016;Putri et al., 2019). Factors related to husbands, consisting of uncircumcised husbands (Syatriani, 2011), number of husbands' marriages (Rimmi et al., 2016;Soans et al., 2012 The instrument format consisted of risk factor items and answer choices. The answer choices included two choices, with a score of 1 (risky) and 0 (no risk). The risk answer column was red, and the non-risk answer column was green. Respondents gave a checkmark () according to their own risk factors. The instrument category was divided into two categories; risky and not risky. The determination of instrument categories was in accordance with the theory and literature review about cervical cancer risk factors. If there was only one factor contained in the instrument item in women, it was categorized as at risk of cervical cancer. This was adjusted to the purpose of developing instruments and in accordance with several instruments that have been previously developed using dichotomous answer choices.
Three experts consisting of psychometricians, nurses, and doctors were involved in the assessment of instrument items. All experts had had a minimum education level of master degree with a working period of >5 years, so they had qualified experiences. The assessment of items by experts used the content validity ratio (CVR) and content validity index (CVI) assessment forms. The content validity assessment was carried out in two stages with the calculation of CVR and CVI, as well as a comparison of the results of CVR and CVI (Bolarinwa, 2016). CVR was assessed using a Likert scale, divided into "1=essential", "2=important but not essential" and "3=not essential". The CVI assessment used a four-point Likert scale, including "1=not relevant", "2=content cannot be reviewed without revision", "3=relevant with minor revisions", and "4=very relevant". The I-CVI value should be 1.00 if the expert panels are five or fewer (Hendryadi, 2017). In the first stage, 9 items were not essential with the CVR of -0.33. As a result, 29 items had a positive value of 0.33-1, meaning that half the panelists considered that the items were essential/important. The CVI calculation showed that 27 items were not valid with an I-CVI value of <1 and an S-CVI value of 0.71 (71%). After comparing the results of the CVR and CVI calculations, 14 items were removed and 32 items were included in the second stage of testing. In the second stage, 6 items were removed because the CVI value did not meet the requirements of the three experts, leaving 26 items meeting the content validity, with a CVR value of 0.33-1 and I-CVI from each expert of 1.00 and S-CVI of 1.00.
The instrument with 26 valid items based on the expert judgment was tested on 20 respondents to determine the respondents' understanding of the items in July 2020. After that, it was continued with item testing on 200 respondents in August 2020. Google forms were used to test items by respondents. Next, a construct validity test was conducted on 26 items using principal component analysis (PCA). Five items were removed because the Measure of Sampling Adequacy (MSA) value was <0.5. Twenty-one items had a Kaiser Meyer Olkin Measure of Sampling Adequacy (KMO MSA) value of 0.805 (≥0.6) and an MSA value of each item of >0.5, indicating the adequacy of the sample used for factor analysis and had a strong significance with a value of Bartlett's test of sphericity of p=0.000 (p<0.05).

Data analysis
The content validity test used the content validity ratio (CVR) and the content validity index (CVI). Construct validity with principal component analysis (PCA) was conducted to explain the proportion of limited variance (shared variance) through varimax rotation analysis (DeVellis, 2017). The most recommended approach is the Kaiser and Cattell method based on the eigenvalue of 1.0 and should not be less than 1.0 because it reflects unstable factors (DeVellis, 2017). Correlation matrix analysis was used to analyze the pattern of co-variation and correlation between items through Barlett's test of sphericity with a significance value of p<0.05 and Kaiser-Meyer-Olkin (KMO) with a value of 0.6. In addition, to perform further factor analysis, an antiimage correlation item value of >0.5 was required. In the advanced factor analysis, the loading factor value of 0.40 was used as significance in defining the factor. Internal consistency reliability test was calculated using Kuder-Richardson 20 (KR-20) >0.70 (ri>0.70).

Ethical considerations
This study had received ethical approval from the Research Ethics Committee of Universitas Padjadjaran with a reference number of 466/UN6.KEP/EC/2020. Permission from the National Unity and Politics of the Bandung City Health Office was obtained before conducting the research. Informed consent was included in the questionnaire via Google form. The respondents who agreed to participate continued filling out the questionnaire. Table 1 shows that after passing the item assessment process and content validity testing, the self-risk assessment instrument consisted of 26 items with a CVR value of 0.33-1, the I-CVI of each expert of 1.00, and S-CVI of 1.00. Husband's income 0.33 1 4

Content validity
Wife's income 0.33 1 5 Have experienced excessive white vaginal discharge 0.33 1 6 Have experienced foul-smelling and itchy vaginal discharge 1 1 7 Have experienced bleeding outside the menstrual period 1 1 8 Have experienced bleeding after menopause 1 1 9 Age at first sexual intercourse 1 1 10 Number of delivery 1 1 11 Have a family history of cancer 1 1 12 Have a family history of cervical cancer 1 1 13 Have a smoking habit 1 1 14 Have a history of exposure to cigarette smoke in the environment where you live or work

Characteristics of respondents
Based on Table 2, the characteristics of respondents on item testing with the minimum age of the respondent was 19 years, and the maximum age was 66 years. Most of the respondents had a higher education level (70%) and worked as private employees (44%).   Table 3 shows that the factor analysis carried out on 21 items in the self-risk assessment instrument through rotated component matrix analysis resulted in six factors. First factors (items 13, 16, 18, 19, 20, 25), second factors (item 5, 6, 15), third factors (items 3, 4, 9, 23), fourth factors (items 14, 17, 26), fifth factors (items 7, 11, 12) and a factor of six (items 10, 21) had a loading factor value of each item of ≥0.4.   Table 4 shows that the KR-20 value was 0.807 (ri>0.70), so that the self-risk assessment instrument (SiNara) was reliable.  Table 5 describes the items of the self-risk assessment instrument adjusted to the results of the factor analysis. Based on the results of the content and construct validity test, there was a change in the number of items from 38 items in the item pool to 21 items. Have a smoking habit 2

Optimization of scale length
Have a habit of washing out the genitalia area after sexual intercourse 3 Have a number of sexual partners 4 Have experienced bleeding after sexual intercourse 5 Have had sexual intercourse during menstruation 6 Husband has a history of venereal diseases such as gonorrhea, syphilis or HIV disease 7 Have experienced excessive white vaginal discharge 8 Have experienced foul-smelling and itchy vaginal discharge 9 Have a habit of changing pads during menstruation 10 Husband's income 11 Wife's income 12 Age at first sexual intercourse 13 Have had an IVA or pap smear 14 Have a history of exposure to cigarette smoke in the environment where you live or work 15 Have a habit of washing the vagina with soap or cleaning fluids 16 Husband has a smoking habit 17 Have experienced bleeding outside the menstrual period 18 Have a family member who has been diagnosed with cancer 19 Have a female family member who has had cervical cancer 20 Number of delivery 21 Have experienced complaints during used contraceptive

Discussion
The purpose of this study was to develop the Indonesian self-risk assessment instrument for cervical cancer. This study showed that the Indonesian self-risk assessment for cervical cancer (SiNara) instrument had been developed, and the instrument had good validity and reliability score. Content validity test showed that there were 26 items with a CVR value of 0.33-1 and I-CVI as well as S-CVI of 1 (100%), meaning that all experts agreed that these items were valid content. The construct validity test was carried out on 21 items, because there were five items, including items 1, 2, 8, 22, and 24 omitted due to MSA values of <0.5. The varimax rotation resulted in six factors, namely factor one consisting of six items, factor two consisting of three items, factor three consisting of four items, factor four consisting of three items, factor five consisting of three items, and factor six consisting of two items. The internal consistency reliability testing was conducted Nurse Media Journal of Nursing, 11(2), 2021, 226 Copyright © 2021, NMJN, e-ISSN 2406-8799, p-ISSN 2087-7811 only once and then analyzed (Gray et al., 2013). The six factors had each item loading factor value of ≥0.4. The Kuder-Richardson 20 (KR-20) value showed that the instrument had good reliability on internal consistency, with the KR-20 value of >0.70 (Fraenkel et al., 2011).
The instrument, "SiNara", consists of six factors with 21 valid and reliable items. Compared to the previous self-risk assessment instrument, there were three new items produced in this study, including having a habit of washing out the vagina with soap or cleaning liquid (Dianti & Isfandiari, 2015;Fitrisia et al., 2019;Syatriani, 2011), having a history of cigarette smoke exposure in the environment where they live or work (Arfailasufandi et al., 2016;Dewi, 2017;Putri et al., 2019), and experiencing complaints during the use of contraceptive agents (Damayanti, 2013;Darmayanti et al., 2015;Nindrea, 2017;Partiwi et al., 2015). Various risk factors in the instrument can evidently increase the risks of cervical cancer, either directly or indirectly. Smoking is one of the factors that can increase the risk of cervical cancer in women in Indonesia (Nindrea, 2017). Also, exposure to cigarette smoke can increase the risk of cervical pre-cancerous lesions (Arfailasufandi et al., 2016;Dewi, 2017;Putri et al., 2019). A study showed that women with partners who smoked for 20 years with a total of 20 packs per year had an increased risk of cervical cancer (Natphopsuk et al., 2012). It can be concluded that not only active smokers who have a risk but women who were in a smoking environment can also increase the risk of being infected with HPV.
Habits of sexual activity and personal hygiene could increase the risk of cervical cancer. Some studies showed that having more than one sexual partner increases the risk of cervical cancer (Jasa, 2016;Umami et al., 2017;Yuliani et al., 2019). First sexual activity at an average age of 16 or <20 years can also increase the risk of cervical cancer (Arfailasufandi et al., 2016;Darmayanti et al., 2015;Fitrisia et al., 2019;Nindrea, 2017). In addition, personal hygiene factors such as the use of antiseptics in the genital area, rarely changing pads during menstruation, and no washing out the genitalia area after sexual intercourse can increase the risk of cervical cancer (Dianti & Isfandiari, 2015;Syatriani, 2011). Thakur et al. (2015) pointed out that poor genital hygiene is one of the risks of cervical cancer. Furthermore, several factors related to sexual activity or other disease conditions can also increase the risk of cervical cancer, including bleeding after sexual intercourse, bleeding outside the menstrual period, excessive white vaginal discharge, and foulsmelling and itchy vaginal discharge (Nindrea, 2017;Trifitriana et al., 2017).
Women with low socioeconomic status tend to have cervical cancer risk (Chandrawati, 2016;Umami et al., 2017). Socio-economic status related to nutrition is one of the factors that influence the incidence of cervical cancer (Syatriani, 2011). Most Indonesian women were known to have never done early detection of cervical cancer. This is indirectly influenced by economic status, which can be several barriers, such as inadequate access to health services and a lack of knowledge that cancer can be detected (Lee & Lee, 2017;Nurhasanah & Afiyanti, 2014;Putri et al., 2019). Another factor that could increase the risk of cervical cancer in Indonesia was genetic factors. The previous studies showed that women with a family history of cancer have an increased risk of cervical cancer (Arfailasufandi et al., 2016;Rahmah et al., 2017;Yuviska & Amirus, 2015). In addition, multiparity women were more at risk of developing cervical cancer (Arfailasufandi et al., 2016;Chandrawati, 2016;Damayanti, 2013;Hidayat et al., 2014;Lestariningsih & Martini, 2013;Sundari et al., 2017). The use of contraceptives is also a risk factor for cervical cancer. The results of the study state that the long-term use of hormonal contraceptives increases the risk of cervical cancer (Damayanti, 2013;Darmayanti et al., 2015;Fitrisia et al., 2019;Nindrea, 2017;Partiwi et al., 2015). Furthermore, the use of oral contraceptives for more than 5 years can escalate the risk of cervical cancer (Dewi et al., 2015;Jasa, 2016;Yuliani et al., 2019).

Implication and limitation
This study has some implications for clinical practice and nursing research. First, the Indonesian self-risk assessment instrument for cervical cancer "SiNara" has good validity and reliability to be used independently by women to detect cervical cancer. The existence of this instrument is expected to increase women's awareness to carry out early detection of cervical cancer through the IVA test or pap smear. Second, the instrument consists of three new items, which were not published previously by another researcher, and were specific to the Indonesian setting. Third, the "SiNara" instrument in the Indonesian language has relatively short items and is easy to fill out so that it can be extensively used by Indonesian women and women's health practitioners/ researchers. This study has several limitations. The first was item pools formulation used secondary data through literature reviews, not primary data. The second limitation was no research on cervical cancer risk factors related to culture or customs. Therefore, it is still needed to develop this instrument by examining it from a cultural perspective that may increase the risk of cervical cancer. The third limitation was the sample used, which only consisted of a population that had smartphones. However, online data collection using smartphones was very helpful for research during a pandemic.

Conclusion
The Indonesian self-risk assessment instrument for cervical cancer "SiNara" had been valid and reliable with relatively short items and used the Indonesian language. The "SiNara" instrument is recommended for every Indonesian woman to assess the risk of cervical cancer independently. Moreover, this instrument can be used by further researchers to identify the risk of cervical cancer in women in Indonesia on a large scale. It is recommended for further researchers to use primary data besides literature review to develop more items.