Skip to main content

Table 3 Risk of study participation and outcome measurement bias - the QUIPS tool

From: The use of non-prescribed antibiotics; prevalence estimates in low-and-middle-income countries. A systematic review and meta-analysis

1 Abdulraheem et al., 2016 [32, 48] Low: Despite the study being conducted in a Health facility the sample size and the response rate are very high with 1150 and 95,5% respectfully. Low: Used a descriptive and comparative statistical data analysis was processed with the SPSS. Simple and multiple logistic regression models were used to evaluate associations between participant characteristics and reported usage of antibiotics.
2 Aditya, S. et al. 2013 [34] High: although high participation rate of 94.8 and 97.4%, clear comparison of responders vs. non-responders. The sample was not representative of the study population. We also question the self-administered questionnaire High: Measurement of outcome is valid, reliable and similar for all subjects. However, data collected were self-reported which may introduce some bias in the behaviors of the respondents studied.
3 Albawani, S.et al., 2016 [41] Low: A convenience sampling method was used to distribute 363 questionnaires were to consumers attending 10 selected community pharmacies. The pharmacies were carefully selected to represent different areas in Sana’a City. Low: Clear operationalization of outcome measure, the bivariate analysis was used to identify the risk factors associated with the use of antibiotics during self-medication Access.
4 Bilal, M. et al.,2016 [30] Low: Although the sample was selected from hospital patients, the questionnaire went through a pilot phase in which 30 people who conformed to the inclusion criteria. Thesampling was from an appropriate sampling frame. Low: Measurement of outcome is similar for all subjects. The questionnaire was then given to the volunteers to fill out. It contained 5 sections (A, B, C, D and E). Participants were given the option of answering the questionnaire themselves or having the researcher fill it based on verbal responses.was validated
5 Israel, E. et al., 2015 [42, 49] Low: The sample size was calculated according to a formula described by Badger et al. The questionnaires were randomly distributed to 526 civil servants based on the various ministries, departments, and units. The response rate was 89.5%. Low: the outcome of measurement was conducted in a similar way, additionally the results were analysed by one-way ANOVA, using SPSS, all data were expressed as mean and difference between groups were considered significant at P = .05.
6 Ramay, B. et al., 2015 [48] Low: Good participation rate, good participation groups. The Sample size was calculated for each pharmacy using Epidat 4.0 based on a population of 350 patients arriving to the pharmacy weekly, assuming that 50% of the population self-medicates, a precision of 5% and a 95% confidence level. Low: Data was collected by a questionnaire that was designed based on instruments used in previous studies. The questionnaire was validated by interviewing 20 customers with .
7 Ramay, B. et al., 2016 [31] Low: Only customers arriving to pharmacies to purchase antibiotics without a prescription were invited to participate in the study, and the inclusion criteria as well as the age of the participants was considered. There is a reasonable, response rate of 60%. Low: Measure of outcome similar for all subjects, criteria are clearly stated.
8 Sah et al., 2016 [50] High: Although the response rate was 99% the selection of the nursing students was not clear, so who to approach was still at the discretion of the interviewer and was therefore likely to be biased. High: Measure of outcome was similar for all subjects. Although comprised of two components. The questionnaire was too short considering the objectives of the study.
9 Senadheera, et al, 2017 [46] High: Sample size was determined by the formula to estimate a population proportion and the prevalence of SMA was assumed to be 50%. Moderate: Prevalence of SMA was calculated by dividing the number who used an antibiotic without a valid prescription (SMA) in the last 3 months by the total respondent. The estimated prevalence to SMA was calculated within 5% with 95% confidence.
This might not a reflect of the answers of the participants.
10 Shah, S. et al., 2014 Moderate: The sample was representative it was calculate based on previous studies that had reported the SMA prevalence’s. However, the study included all male and female students enrolled in undergraduate or postgraduate programs in universities of Karachi who understood English. This might incur a certain level of bias. Moderate: Clear use of a study specific questionnaire that used closed questions.
11 Shiavong, A. et al., 2017 [51] Moderate: Using ordinary sample size calculation the authors (estimated prevalence 50%, precision 5%, confidence interval 95%), additionally 10 villages were randomly selected for the household survey. Low: A pretested structured interview form was performed, and the final tool contained both closed and open-ended questions was considered.