In the article by Mu et al. (2021), the authors have re‐analysed the published article and found some errors that are corrected in this version.
The values under Result section are corrected to “d = 0.589”; N = 6.
In Figure 1, the content in text boxes 5 and 7 is corrected as follows: “1 additional record identified from reference lists”; and “6 studies included in meta‐analysis” respectively. The revised figure is as follows.
FIGURE 1.
Flow chart of selected articles
In the section 6.1 Study characteristics, the following sentences were wrong:
The seven studies included provided information on 2,277 women with PPD and were conducted across five countries, including Australia (1/7), Sweden (1/7), the United States (1/7), Canada (1/7) and the United Kingdom (3/7). Due to loss during follow‐up, only a total of 856 patients were included in the final analysis. The sample sizes were between 42–910 patients each.
The mean age of the participants in the studies ranged from 30.55 (4.99)–35.6 (11.9) years and 29.81 (6.09)–34.3 (11.3) years for the intervention and control groups respectively. Six of the studies had a follow‐up period of 10–17 weeks after intervention.
It has been corrected as follows:
The six studies included provided information on 1,980 women with PPD and were conducted across five countries, including Australia (1/6), Sweden (1/6), the United States (1/6), Canada (1/6) and the United Kingdom (2/6). Due to loss during follow‐up, only a total of 646 patients were included in the final analysis. The sample sizes were between 42–910 patients each.
The mean age of the participants in the studies ranged from 30.55 (4.99)–32.3 (4.7) years and 29.81 (6.09)–32.2 (5.7) years for the intervention and control groups respectively. Five of the studies had a follow‐up period of 10–17 weeks after intervention
The values under 6.2 Heterogeneity test results section are corrected to “(I2 = 53.4%, p < .001)”
The values under 6.3 Overall effects and subgroup analysis section are corrected to “effect size of 0.589 (95% CI: 0.442–0.785)”
In addition, the last paragraph under same section has been corrected from “The RR was highest in studies followed up for over 12 weeks after intervention (0.727, 95% CI: 0.637–0.831). (Table 2). In the subgroup analysis of intervention types, we found that five studies conducted on CBT interventions showed higher heterogeneity (I2 = 67.0%, p = .013) (Table 2).” to “The RR was highest in studies followed up for over 12 weeks after intervention (0.683, 95% CI: 0.570–0.819). (Table 2). In the subgroup analysis of intervention types, we found that four studies conducted on CBT interventions showed higher heterogeneity (I2 = 71.0%, p = .030) (Table 2).
TABLE 2.
Subgroup analysis of internet‐based interventions for prevalence of postpartum depression according to fellow‐up, type of intervention, number of sessions and professional support
| Stratification group | N | Sample size | RR a (95% CI) | Z‐statistics | p | Heterogeneity I 2 (%) | Chi‐square, χ 2 (df) |
|---|---|---|---|---|---|---|---|
| Comparison | |||||||
| Total | 6 | 646 | 0.589 (0.442–0.785) | 3.61 | <.001 | 53.4 | 10.73(5) |
| Fellow‐up | |||||||
| >12W | 3 | 525 | 0.683 (0.570–0.819) | 4.62 | <.001 | 46.2 | 3.72(2) |
| <12W | 2 | 80 | 0.488 (0.302–0.788) | 2.93 | .003 | 18.3 | 1.22(1) |
| 12W | 1 | 41 | 0.257 (0.105–0.628) | 2.98 | .003 | / | / |
| Type of intervention | |||||||
| Cognitive behavioural therapy | 4 | 232 | 0.526 (0.294–0.941) | 2.17 | .030 | 71.0 | 10.35(3) |
| Behavioral activation | 2 | 414 | 0.626 (0.507–0.772) | 4.37 | <.001 | 0.0 | 0.52(1) |
| Number of sessions | |||||||
| <8 sessions | 2 | 80 | 0.418 (0.246–0.708) | 3.24 | .001 | 63.2 | 2.72(1) |
| ≥8 sessions | 4 | 566 | 0.638 (0.475–0.857) | 2.98 | .003 | 53.6 | 6.46(3) |
| Professional support | |||||||
| Exclusive therapist support | 5 | 607 | 0.629 (0.471–0.839) | 3.15 | .002 | 51.8 | 8.31(4) |
| Without therapist support | 1 | 39 | 0.367 (0.179–0.755) | 2.73 | .006 | / | / |
ES: Effect Size and represents the prevalence of depressive mood.
In Figure 2, “Kessler D et al 2009 (2009)” under study ID & “Events study and Events gratral” have been removed. Also, RR (95% CI) for overall value is corrected to “0.59 (0.44, 0.79)”. The revised figure is as follows:
FIGURE 2.
Forest plot for prevalence of PPD outcomes of intervention versus control groups
The value for women under 7. Discussion is corrected to “a total of 1980 women”
In Table 1, first row has been removed and the revised table is as follows.
TABLE 1.
Major characteristics of primary studies
| Author (year) | Country | Depressive condition (recruitment) | Intervention (I) | Control (C) | Age (mean (SD)) | Sample size (initial enrolment, N) | Sample size (follow‐up, N) | Outcomes (measures) | Follow‐up (in weeks) |
|---|---|---|---|---|---|---|---|---|---|
| O'Mahen HA et al. (2013) | United Kingdom | EPDS>12 | internet Behavioral Activation(iBA) | TAU |
I:32.3(4.7) C:32.2(5.7) |
I:462 C:468 |
I:181 C:162 |
EPDS | 15 |
| O'Mahen HA et al. (2014) | United Kingdom | DSM‐IV | Netmums Helping With Depression (Netmums HWD) | TAU | N/A |
I:41 C:42 |
I:37 C:34 |
EPDS | 17 |
| Barrera AZ et al. (2015) | United States | EPDS≥10 | Mood management Internet intervention | TAU |
I:30.55(4.99) C:29.81(6.09) |
I:435 C:417 |
I:57 C:54 |
CES‐D | During pregnancy and up to six months postpartum |
| Pugh NE et al. (2016) | Canada | EPDS≥10 | Therapist‐Assisted Internet‐delivered Cognitive Behavior Therapy | Waitlist | N/A |
I:25 C:25 |
I:21 C:20 |
EPDS | 10 |
| Milgrom J et al. (2016) | Australia | DSM‐IV | MumMoodBooster | TAU | N/A |
I:21 C:22 |
I:19 C:22 |
BDI‐II, DSM‐IV | 12 |
| Forsella. E et al. (2017) | Sweden | Major depressive disorder | Internet delivered cognitive behavior therapy | TAU |
I:31.2(3.7) C:30.8(5.3) |
I:22 C:20 |
I:21 C:18 |
MADRS‐S | 10 |
BDI‐II, Beck Depression Inventory‐II; CES‐D, Epidemiological Studies—Depression; DSM‐IV, The Diagnostic and Statistical Manual of Mental Disorders‐IV; EPDS, Edinburgh Postnatal Depression Scale; MADRS‐S, Montgomery‐Åsberg Depression Rating Scale—Self report version.
In Table 2, overall values have been corrected and the revised figure as follows.
In the section 7.3 Efficacy of Internet‐based interventions on the treatment of postpartum depression, the following sentences were wrong.
Internet‐based interventions provided a significant improvement in the prevalence of PPD with effect size of 0.642 in this study, which was supported by effect size of 0.63 obtained in the therapist‐supported iCBT on depressive symptoms (Lau et al., 2017) and effect size of 0.46 obtained for the use of computer‐based or web‐based interventions to alleviate perinatal depressive symptoms (Ashford et al., 2016). Among the studies included in this meta‐analysis, there were six studies (6/7) on network interventions that could efficacy decrease PPD and only one study (1/7) failed to demonstrate a significant reduction in PPD (mood management Internet intervention) (Barrera et al., 2015).
It has been corrected as follows:
Internet‐based interventions provided a significant improvement in the prevalence of PPD with effect size of 0.589 in this study, which was supported by effect size of 0.63 obtained in the therapist‐supported iCBT on depressive symptoms (Lau et al., 2017) and effect size of 0.46 obtained for the use of computer‐based or web‐based interventions to alleviate perinatal depressive symptoms (Ashford et al., 2016). Among the studies included in this meta‐analysis, there were five studies (5/6) on network interventions that could efficacy decrease PPD and only one study (1/6) failed to demonstrate a significant reduction in PPD (mood management Internet intervention) (Barrera et al., 2015).
In the section 7.4 The effect of Internet‐based interventions weakened as the follow‐up duration increased, the following sentences were wrong.
In this meta‐analysis, four studies (4/7) had a follow‐up period of more than 12 weeks and three of these studies (Kessler et al., 2009; O'Mahen et al., 2013, 2014) suggested that the Internet‐based interventions were effective.
It has been corrected as follows:
In this meta‐analysis, three studies (3/6) had a follow‐up period of more than 12 weeks and two of these studies (O'Mahen et al., 2013, 2014) suggested that the Internet‐based interventions were effective.
In the section 7.5 Internet‐based interventions for postpartum depression are mostly based on cognitive behavioural therapy (CBT), the following sentences were wrong.
In this review, five of the studies included (5/7) were based on CBT and the remaining two studies were based on behavioural activation. At the same time, the five studies based on CBT resulted in a significantly improved level of PPD with a medium effect size of 0.628 in this meta‐analysis, which was supported by effect size of 0.63 obtained as a result of therapist‐supported iCBT on depressive symptoms (Lau et al., 2017).
It has been corrected as follows:
In this review, four of the studies included (4/6) were based on CBT and the remaining two studies were based on behavioural activation. At the same time, the five studies based on CBT resulted in a significantly improved level of PPD with a medium effect size of 0.526 in this meta‐analysis, which was supported by effect size of 0.63 obtained as a result of therapist‐supported iCBT on depressive symptoms (Lau et al., 2017).
The authors apologize for these errors.