The first stage of this research was a literature review to develop a blueprint for the questionnaire. The questionnaire items were derived from the Guidelines for Antibiotic Use (Ministry of Health of the Republic of Indonesia, 2021), the Antimicrobial Resistance Control Program in Hospitals (Ministry of Health of the Republic of Indonesia, 2015), Antimicrobial stewardship interventions: A practical guide (World Health Organization, 2021), The Core Elements of Hospital Antimicrobial Stewardship Programs (Centers for Disease Control and Prevention, 2019), the WHO AWaRe Antibiotic Book (World Health Organization, 2023), and previous studies (Fabre et al., 2024)(Setiawan et al., 2019).

The blueprint consisted of several aspects. The statement items were taken from various books and previous research related to antimicrobial stewardship. The knowledge questionnaire contained 20 statements using the Guttman scale, which was used to measure respondents' level of knowledge regarding antimicrobial stewardship. The attitude questionnaire contained 20 statements using the Likert scale, which was used to measure respondents' attitudes toward antimicrobial stewardship. The practice questionnaire contained 20 statements using the Likert scale, which was used to measure respondents' practice toward antimicrobial stewardship.

Second, the questionnaire was consulted with experts (expert judgment). A total of 8 experts consisted of pharmacist practitioner in hospitals and academic pharmacists with the criteria for a minimum of a Master's degree and a pharmacist profession. The results of the 8 expert judgments were calculated using the Content Validity Index (CVI) method. The questionnaire instrument was provided to expert judgment, followed by an assessment using the Content Validity Index (CVI). The CVI evaluation included four aspects: relevance, clarity, simplicity, and ambiguity. Experts will assess the relevance of the items from 1-4 and then calculate the CVI value.

Subsequently, the relevance ratings were recoded as 1 (relevance scale of 3 or 4) and 0 (relevance scale of 1 or 2). The CVI was calculated by dividing the number of experts in agreement by the total number of experts. An acceptable statement item value is a CVI value > 0.83 (Yusoff, 2019). Based of expert judgment consultation, several questionnaire item were revised.

Third, after all statement items were revised, a pilot test on 30 pharmacist respondents fulfill the inclusion criteria. The inclusion criteria for this study were pharmacists working in hospitals and willing to participate in the study. The exclusion criteria for this study were pharmacists working outside of pharmaceutical services. After that, the results of the pilot test on 30 respondents were subjected to reliability testing by entering data into the Excel program and processing the results with SPPS software version 25.

The blueprint instrument, each aspect contains more than one statement item. This ensures that if an item is not used after validity and reliability testing, there are still other items available for that aspect. The instrument blueprint is presented in Table 1.

Questionnaire validity testing aims to measure the extent to which a questionnaire accurately measures the intended construct. The questionnaire instrument is given to expert validators and then assessed using the Content Validity Index (CVI). The CVI encompasses four assessment aspects: relevance, clarity, simplicity, and ambiguity. The relevance rating is then calculated by recoding it as 1 (relevance scale 3 or 4) or 0 (relevance scale 1 or 2) (Yusoff, 2019). The result of CVI value is presented in Table 2.

According to Table 2, the knowledge variable questionnaire has a CVI value of 0.75–1.0. One item, Q3, had a CVI value <0.83, so it was excluded. Therefore, the CVI test for the knowledge variable reduced the number of items from 20 to 19 for reliability testing.

The attitude variable questionnaire has a CVI value of 0.75–1.0. Two items, Q19 and Q20, had a CVI value <0.83, so they were excluded. Therefore, the CVI test for the attitude variable reduces the number of statements from 20 to 18 for reliability testing.

The practice variable questionnaire has a CVI value of 0.75–1.0. Four items, Q3, Q4, Q14, and Q19, had a CVI value <0.83, therefore, these four items were excluded. Therefore, the CVI test for the practice variable reduced the 20 statement items to 16 statement items for reliability testing.

The reliability test was conducted on 30 pharmacist respondents who practice in hospitals around the Solo area. Reliability was measured by examining the correlation between the score of one item and the total score of all other items in the instrument. The corrected item-total correlation value was >0.25. Table 3 shows the results of the corrected item-total correlation value for the knowledge variable.

Based on the results of the corrected item-total correlation test, there were 12 statement items with values ​​<0.25, namely Q1, Q2, Q4, Q5, Q6, Q8, Q9, Q13, Q14, Q15, Q16, Q20. These items were removed, and the researcher then conducted a reliability test on items with a corrected item-total correlation value >0.25. Table 4 shows the final results of the reliability test for the knowledge variable.

Based on the final results of the reliability test, 7 items in the knowledge variable statement had a corrected item-total correlation value > 0.25. Therefore, of the 19 questions, 7 items can be used to measure the knowledge variable.

Based on the results (Table 5) of the corrected item-total correlation test, all statement items showed a corrected item-total correlation value of > 0.25, so that 18 statement items can measure attitude variables.

Based on the results of the corrected item-total correlation test for the practice variables, Table 6 shows that three items had values ​​<0.25, namely Q15, Q16, and Q17. These three items were excluded, and the researcher then conducted a reliability test on the items with corrected item-total correlation values ​​>0.25. Table 7 shows the final results of the practice variable reliability test.

Based on the final reliability test results, 13 practice variable statement items had a corrected item-total correlation value > 0.25. Therefore, of the 16 questionnaire items, 13 could be used to measure practice variables.

Based on the results of the reliability test, the Cronbach’s alpha value for the knowledge variable with 7 valid items was 0.867. The Cronbach’s alpha value for the attitude variable with 18 valid items was 0.924, while the Cronbach’s alpha value for the practice variable with 13 valid items was 0.939. Cronbach’s alpha value greater than 0.7 indicates that the items are consistent in measuring the variable (Taherdoost et al., 2016). The Cronbach’s alpha values for the knowledge, attitude, and practice items were all greater than 0.7, the instrument was considered reliable and suitable for use in this study. A measurement is considered to have high validity if the data generated accurately reflects the variables being measured. Content validity assesses the extent to which the elements of a measurement instrument are truly relevant and represent the measurement objectives. This assessment is qualitative and judgmental and is conducted by a panel of experts (Azwar, 2025). A study conducted by (Kurniasari & Ichsan, 2023) which has conducted validity testing using the Content Validity Index (CVI) method on the knowledge, attitudes and practice of pharmacists towards antibiotic resistance in pharmacies resulted in 24 knowledge questionnaire items obtained CVI values ​​between 0.8 and 1.0. A total of 24 attitude questionnaire items obtained CVI values ​​between 0.9 and 1.0 and 24 behavior questionnaire items obtained CVI values ​​between 0.8 and 1.0.

In a previous study, nine experts assessed a draft questionnaire on the development and validation of a questionnaire on knowledge, attitudes, and practices regarding antibiotics for the general public. Based on content validity results, seven items in the knowledge domain, two items in the attitude domain, and seven items in the practice domain were removed due to CVR <0.78 and CVI <80% (Karuniawati et al., 2022).

In this study, the validity and reliability tests used corrected item-total correlation values ​​that showed internal consistency with the aim of ensuring that the measurement instrument provided consistent results when repeated measurements were taken or on the same subject (Azwar, 2025).

The minimum corrected item-total correlation value in this study was 0.25. The item-total correlation coefficient ranges from 0 to 1.00, with a positive or negative sign. The better the discriminatory power, the closer the value is to 1.00. A small coefficient close to 0 or negative sign indicates the item has no discriminatory power. Discriminatory power is the extent to which an item can distinguish between individuals who do or do not have the attribute being measured (Azwar, 2022).

The developed questionnaire may serve as a useful tool for assessing pharmacists’ knowledge, attitudes, and practices related to antimicrobial stewardship. The validated questionnaire can help healthcare institutions identify gaps in pharmacists’ competencies and support the implementation of antimicrobial stewardship programs aimed at reducing antibiotic resistance (Akbar et al., 2021)

This study has several limitations. First, the reliability testing was conducted with a relatively small sample size of pharmacists, which may limit the generalizability of the results. Second, the study did not include construct validity testing, such as exploratory or confirmatory factor analysis, which could provide stronger evidence regarding the structural validity of the instrument. To address the limitations of this study, future research should involve a larger and more diverse sample of pharmacists to improve the generalizability and stability of the reliability estimates. The benefit of validation of the questionnaire is that it ensures that the instrument accurately measures the intended construct, thereby improving the validity of the findings.