Authors response to article retraction

Dear Editor

Regarding the retraction decision of the article entitled ‘Benefits of vaginal misoprostol prior to IUD insertion in women with previous caesarean delivery: a randomised controlled trial, The European Journal of Contraception & Reproductive Health Care, 23:1, 32–37, DOI: 10.1080/13625187.2018.1428297.’

The authors need to clarify some data to the authors

The investigating team had asked for the raw data, ethical committee approval, informed consent and the protocol submitted. These files were checked and no concerns were raised about them.

The editor decided to retract the paper based on the following issues.

1. They were concerned about submission of the article within 1 month of the study completion which is usually ca be done through professional cooperation between the authors.

2. The editors claimed that visual examination of numerical baseline variables age, BMI and previous CS shows (a) that the distribution in the treatment group substantially differs from the control group and (b) that variable values in the treatment group appear to have been fabricated in blocks. In response to this comment the authors statistically assessed the visual examination of the baseline numerical variables shows no substantial difference between both groups or any data fabrication (Figure 1). This was confirmed by the non-significant p-values when these variables were compared, as shown in Table 1 in the retracted published article (Figure 2). On the contrary, the visual examination of the numerical outcomes shows a substantial difference between both groups (Figure 3). This also was confirmed by the significant p-values when these variables were compared, as shown in table 2 in our published article (Figure 4).

3. The editor described that the null hypothesis of no autocorrelation is strongly rejected for several variables in the treatment group (age, number of CS, difficulty inserting), as is the number of CS in the control group and they claimed that these findings are confirmed by the Wald-Wolfowitz runs test of randomness. Regarding the testing for autocorrelation, it is not suitable for this research model. It is the correlation of a signal with a delayed copy of itself as a delay function. Sometimes, it is known as serial correlation in the discrete-time case [1]. Therefore, if we repeat this test with the same variables but in a different order, we will get a different result every time we change the order of the variables!!!

4. Finally, when analysing the data using the Wald–Wolfowitz runs test of randomness, we found no significant difference between both groups for all numerical variables, as shown in Figure 5. These non-significant p-values indicate that both groups were taken from the same population sample size, i.e., the data have arisen from a genuine randomised controlled trial [2].

Figure 1. Visual examination of the baseline numerical variables.

Figure 2. Table 1 in our published article.

Figure 3. Visual examination of the numerical outcomes.

Figure 4. Table 2 in our published article.

Figure 5. Wald-Wolfowitz runs test of randomness.

Disclosure statement

No potential conflict of interest was reported by the author.