BackgroundPerinatal depression is minor or major depression episodes since pregnancy to 12 months after childbirth,and can be recognized by clinical diagnosis and psychological assessment.Perinatal depression brings severe and lasting harms to both expected mothers and their offspring.Relieving perinatal depression meets the guideline of Healtht China Action(20192030)on depression prevention and control.Mindfulness was described as an intentionally and non-judgmentally way to pay attention to the present moment.Previous researches reported promising effects of mindfulness-based interventions on perinatal depression,and potentials on anxiety,stress and fear of childbirth during perinatal period.Psychological treatment using internet-based interventions were mainly delivered via the internet with but not limited to smartphone-based applications.Internet-based interventions are convenient,anonymous and easily accessed,and therefore a suitable approach for our community.Internet-based mindfulness interventions(IMBIs)provide an opportunity for perinatal depression treatment.However,the effects of internetbased mindfulness interventions on perinatal depression keep unknown.Review studies and empirical studies are two different approaches in evidence-based medicine.Systematic review and meta-analysis present reliable results by combining and analyzing data from different studies conducted on similar research topics.Randomized controlled trial(RCT)is regarded as the standard evidence of clinical research which is accomplished by randomly allocation and binding.This study therefore adapts both systematic review,meta-analysis and RCT to explore the scientific question "Do IMBIs improve perinatal depression?" The combination of the two approaches is able to provide higher level of evidence.The "Personalized causal pathway hypothesis" addressed ppersonalized intervention for depression treatment and suggested to recognize the moderator that is able to change the intervention mechanism.As far as mindfulness intervention,to cultivate mindfulness is considered the basic mechanism and can be used to explore the moderators under the context.Therefore,this study aims to explore the moderators that could influence the IMBIs improving perinatal depression by influence the mindfulness level using moderated mediation models.Study 1 Effects of internet-based mindfulness interventions on perinatal depression:a systematic review and meta-analysisObjectives:(1)This study firstly aims to assess the status of IMBIs during perinatal period by systematic review.(2)This study secondly aims to evaluate the effects of IMBIs on perinatal depression by meta-analysis so that answer the scientific question "Do IMBIs improve perinatal depression?"(3)This study thirdly aims to evaluate the effects of IMBIs on secondary outcomes by meta-analysis.Methods:Subjects and free words were both used to search in PubMed,PsycInfo,Embase,Scopus,CINAHL,Cochrane Central Register of Controlled Trial,CNKI and Wangfang Data from inception to November 2020.RCT,quasi-experimental studies and pre-post design studies were all included.Two independent researchers searched the databases and extracted and synthesized data.Descriptive analyses were used to in systematic review,and random effect model and standard mean difference(SMD)were used in meta-analysis.Results:(1)Study characteristics:A total of 10 publictaions were included in this study.Eight studies were published sicne 2018 and six studies were RCT.The quality appraisal resulted that 7 studies were high/critical risk studies.No low risk study was found.(2)Characteristics of IMBIs:Complex mindfulness interventions with other intervention components were the main intervention formats(6/10).Self-help intervention without guidance was the main guide format(7/10).Website(6/10)and smartphone-based applications(3/10)were main internet mediums.Pregnancy stage was the main implementation stage(6/10).The most frequent intervention period is 8 weeks and people usually practiced mindfulness daily.Eight studies reported complete rate and half of them had a completion rate range from 55%to 57%.The median of complete rate was 56%.(3)Follow-up status:Nine studies reported the response status at immediate postintervention and the dropout rate was 8.2%~61%,with more than half exceeded 30%%(5/9).Threes studies evaluated the long-term follow-up status and the dropout rate was 36.4%~91%.(4)Evaluation on perinatal depression:Three studies mainly aimed to evaluate the intervention effects of IMBIs on perinatal depression.Edinburgh Postnatal Depression Scale(EPDS)was the most widely used scale to screen perinatal depression(7/10).(5)Meta-analysis on perinatal depression:At the immediate post-intervention timepoint,IMBIs significantly reduced depression symptoms than control group(SMD=-0.47,95%CI:0.82~-0.12)with a high heterogeneity.Subgroup analysis showed that both complex interventions and interventions with only mindfulness component,both self-help interventions and guided interventions reduced perinatal depression symptoms.IMBIs delivered through website didn’t reduce depression symptoms while non-website interventions reduce depression symptoms.Interventions with 8-week period reduced depression symptoms and interventions with less than 8-week period didn’t reduce depression symptoms.Samples with mental health problems at baseline reported reduced depression symptoms while those healthy women didn’t.Interventions conducted during postnatal period reduced depression symptoms,but studies conducted during pregnancy priod showed high heterogeneity.IMBIs didn’t reduce depression symptoms than control group(SMD=-0.13,95%CI:-0.44~0.18)in the follow-up and the heterogeneity was low.(6)Meta-analysis on perinatal anxiety:At the immediate post-intervention timepoint,IMBIs didn’t reduce anxiety symptoms than control group(SMD=-0.47,95%CI:-0.82~-0.12)with a high heterogeneity and unstable results in sensitivity analysis.One study showed no between-group difference on anxiety symptoms in the follow-up.(7)Meta-analysis on perinatal stress:At the immediate post-intervention timepoint,IMBIs didn’t reduce stress than control group(SMD=-0.01,95%CI:-0.66~0.63)with a high heterogeneity.IMBIs also didn’t reduce stress than control group(SMD=0.19,95%CI:0.76~1.13)and the heterogeneity was low in the follow-up.(8)Meta-analysis on mindfulness:At the immediate post-intervention timepoint,IMBIs improved mindfulness than control group(SMD=0.61,95%CI:0.22-1.00)with a high heterogeneity.IMBIs also improved mindfulness ability than control group(SMD=0.58,95%CI:0.06~1.09)and the heterogeneity was low in the follow-up.Conclusions:(1)Studies on IMBIs are in the initial stage.Quality of included studies is poor.Few studies mainly focused on perinatal depression.More high-quality empirical studies are needed in the future.(2)This study answered the scientific question "Do IMBIs improve perinatal depression?"from a review perspective.IMBIs reduced perinatal depression symptoms at immediate postintervention,but the long-term effect of internet-based mindfulness intervention on perinatal depression need further exploration.Different mindfulness format(complex intervention/only mindfulness technique)and different guidance format(self-help intervention/guided intervention)didn’t affect the intervention effects.Non-wesit-based format and eight-week are appropriate for IMBIs and women with mental health problems benefit from IMBIs.Intervention performed in postnatal period reduced depression symptoms at immediate postintervention and effects of prenatal interventions need to be studied in the future.(3)IMBIs also improved mindfulness.IMBIs didn’t reduce perinatal anxiety symptoms and stress,but the results for anxiety showed high heterogeneity which need to be further studied.Study 2 Effects of a smartphone-based mindfulness intervention on perinatal depression:a randomized controlled trialObjectives:(1)This study firstly aims to assess the effects of a smartphone-based mindfulness intervention on perinatal depression so that answer the scientific question "Do IMBIs improve perinatal depression?"(2)This study secondly aims to assess the effects of a smartphone-based mindfulness intervention on secondary outcomes(anxiety,stress and fear of childbirth).Methods:This is a two-group parallel randomized controlled trial.Adult women at risk of perinatal depression[EPDS≥10 or 9-item Patients Health Questionnaire(PHQ-9)≥5)]were recruited at a tertiary comprehensive hospital of Shandong Province from June 2018 to January 2020.They were randomly allocated to smartphone-based mindfulness intervention or control group.The sample size should reach 168 with 84 women in each group.Mindfulness-based intervention was developed on the basis of mindfulness-based cognitive therapy.It contains eight modules for eight weeks with mindfulness course,formal practice and informal practice in each module.The formal practice lasts 15 to 25 minutes and requires 6 days one week.Data were collected at baseline(T1,pre-intervention),4 weeks after randomization(T2,inter-intervention),8 weeks after randomization(T3,immediate post-intervention),18 weeks after randomization(T4,late pregnancy and before childbirth),and 6 weeks after childbirth(T5).The researcher who collected data was blind for the randomization.The primary outcome of the current study was perinatal depression assessed by EPDS.Secondary outcomes included anxiety symptoms assessed by Generalized Anxiety Disorder(GAD-7),stress assessed by Perceived Stress Scale(PSS)and fear of childbirth assessed by Wijma Delivery Expectancy/Experience Questionnaire(WDEQ).The analysis followed intention-to-treat rules and Generalized Estimating Equation(GEE)was used to evaluate the intervention effects.Two tailed a was set at 0.05.Results:(1)Recruitment:A total of 1140 pregnant women were recruited and a total of 168 were randomly allocated with each group containing 84 women.The dropout rate at immediate postintervention(T3)was 22.0%(37/168),and the overall dropout rate through the intervention was 34.5%(58/168).Univariate analysis was conducted to compare the basic characteristics between study samples and dropout sample,and found that participants with an advanced gestational age at baseline tended to drop out more frequently during follow up(103.33±15.414 vs 96.43±13.836,P=0.004).(2)Baseline characteristics:The mean age of the sample was 29.91±4.015 years old.The mean gestational week was 14.12 weeks and 64.9%of the sample was primipara.Univariate analysis was conducted to compare the basic characteristics between smartphone-based mindfulness intervention group and control group,and results showed that more participants of intervention group planned pregnancy than controlled group(67.5%vs 86.6%,P=0.004).There’s no significant between-group difference on age,gestational days,weight gain,BMI before pregnancy,BMI now,race,education,employment status,family income,parity,history of abortion,history of induced abortion,history of embryo damage,intended pregnancy,pregnancy reaction,history of chronic disease,EPDS,PHQ-9,GAD-7,PSS and WDEQ(all P>0.05).This suggested that the randomization was successful.(3)Intervention compete status:Forty-four participants completed at least four weeks’intervention and met the criteria of the current study for "completed intervention".The complete rate was 52.4%(44/84).(4)Between-group comparison on scores of perinatal depression and secondary outcomes:Independent sample t test showed that participants in intervention group reported significantly lower scores on perinatal depression(6.49±4.497 vs 9.09±6.241,P=0.012)and nearly medium effect size(d=0.47,95%CI:0.10-0.84)at T3,lower scores on perinatal anxiety(4.32±2.773 vs 5.93±4.079,P=0.032)and nearly medium effect size(d=0.46,95%CI:0.04~0.87)at T4,and lower scores on fear of childbirth(37.46±21.055 vs 46.25±19.312,P=0.027)and nearly medium effect size(d=0.44,95%CI:0.05~0.82)at T2,than control group.(5)Between-group comparison on positive perinatal depression symptoms:Chi-square test showed no difference on positive depression symptoms between intervention group and control group at baseline(P<0.05).But intervention group reported significantly lower positive depression symptoms than control group at T3 and T4(T3:23.8%vs 44.4%,P=0.018;T4:21.6%vs 44.2%,P=0.019).(6)Between-group comparison on perinatal depression remission:More than 4 points of EPDS decreased compared to baseline was considered as depression remission,which can be marked as reliable change index(RCI)≥ 4.Chi-square test showed that intervention group had significantly larger proportion of participants who were considered depression remission than control group at T3(38.1%vs 16.7%,P=0.010).(7)GEE analysis on perinatal depression:Main effect of time and interaction effect of time×group showed statistical significance(Waldχ2=14.095,P=0.007;Waldχ2=15.484,P=0.004).Between-group comparison at each timepoint illustrated that intervention group had significantly lower perinatal depression scores on T3 and the estimated mean difference reached 2.64 points(95%CI:0.72~4.56,P=0.007).(8)GEE analysis on anxiety:Main effect of time and interaction effect of timexgroup showed statistical significance(Waldχ2=17.373,P=0.002;Waldχ2=11.908,P=0.018).Between-group comparison at each timepoint showed no significant difference between intervention group and control group(all P>0.05).(9)GEE analysis on stress:Main effect of time showed statistical significance(Waldχ2=11.155,P=0.025).Between-group comparison at each timepoint showed no significant difference between intervention group and control group(all P>0.05).(10)GEE analysis on fear of childbirth:Main effect of time showed statistical significance(Waldχ2=11.088,P=0.011).Between-group comparison at each timepoint showed no significant difference between intervention group and control group(all P>0.05).Conclusions:(1)Dropout rate of the smartphone-based mindfulness intervention was 22.0%~34.5%,and the completion rate was 52.4%.This study used RCT design,obeyed the evaluator-blinding rule,followed up to 32 weeks,performed advanced statistical analyses and utilized intentionto-treat analyses,suggesting a high research quality.(2)This study answered the scientific question "Do IMBIs improve perinatal depression?"from an emperical perspective.Smartphone-based mindfulness intervention reduced perinatal depression symptoms and positive rate of perinatal depression,and improved perinatal depression remission compared to control group.The short-term intervention effect showed medium effect size.The findings in the current study are same with findings in review studies,supporting the effects of IMBIs on perinatal depression.The conclusion provides advanced evidence for the application of IMBIs in perinatal period.(3)The smartphone-based mindfulness intervention also reduced perinatal anxiety symptoms but failed to improve stress and fear of childbirth.Study 3 Targeting population of internet-based mindfulness intervention on perinatal depression:moderated mediation analysesObjectives:(1)This study aims to clarify the mediation effect of mindfulness between mindfulness intervention and depression symptoms.(2)This study aims to explore the moderators for IMBIs relieving perinatal depression symptoms through enhancing mindfulness using moderated mediation analysis on the basis of perinatal women’s charcateirstics.Methods:The "Personalized casual pathway hypothesis" was the theory model and moderated mediation analysis was the statistical model.We hypothesized that intervention group(IMBIs/control group)influenced participants’ mindfulness change during the intervention,and then influenced their depression symptom change,which can be represented as a mediation path "group→mindfulness change→depression symptom change".Individuals’ characteristics can be moderators influencing the first half of the above mediation path "group→mindfulness change".This study was conducted on the basis of Study 2.Only mindfulness before and after intervention were evaluated and the participants were in pregnancy period at that time.Mindfulness was assessed by Five Facet Mindfulness Questionnaire(FFMQ)at T1 and T3.Group was set as independent variable,individuals’ characteristics at T1 was the moderators,mindfulness changes from T1 to T3 was the mediator,and depression symptom change from T1 to T3 was dependent variable.Individuals characteristics contained socio-demographics(age,BMI before pregnancy,education,job and family income)and pregnancy-related characteristics(parity,intended pregnancy,pregnancy reaction,history of abortion,history of embryo damage and history of chronic disease),which were included as binary variables.SPSS 23.0 PROCESS was used in the current study to perform moderated mediation analysis by bootstrap(n=5000).Model 4 was firstly used to test mediation model and model 7 was then used to test the moderated mediation analysis.The mediation model and moderated mediation model were supported when zero was not included in the 95%CI of indirect effect or INDEX.Results:(1)Between-group comparison on mindfulness:Indenpent sample t test showed no difference between IMBIs group and control group on mindfulness at T1(79.38±7.613 vs 79.12±7.652,P=0.840).IMBIs group showed significantly higher mindfulness scores than control group at T3(81.35±9.908 vs 77.63±9.430,P=0.044).Repeated measuring Chi-square showed significant time×group effect(F=5.840,P=0.017,partial η2=0.055),suggesting IMBIs improved mindfulness significantly.(2)Mediation analysis:Model 4 was conducted to test the simple mediation model"group→mindfulness change→depression symptom change".The indirect effect of mindfulness was-0.151,95%Bootstrap CI:-0.340~-0.007,suggesting that mindfulness change was the mediator between group and depression symptom change and moderated mediation analysis can be conducted.(3)Moderated mediation analyses:1)Age:The INDEX of age was 0.356,95%Bootstrap CI:0.001~0.832,suggesting a significant moderated mediation model.Further analysis showed that group predicted mindfulness change positively in non-advanced-age women(B=0.636,P=0.034)while group didn’t predict mindfulness change in advanced-age women(B=-0.474,P=0.341).There was no direct effect from group to depression symptom change(B=-0.291,P=0.129).The indirect effect from group to depression symptom change differed between age group.For nonadvanced-age women,the indirect effect of mindfulness change was significant(B=-0.204,95%CI:-0.452~-0.017).However,for advanced-age women,the indirect effect of mindfulness was non-significant(B=0.152,95%CI:=0.123~0.518).2)BMI before pregnancy:The INDEX of BMI before pregnancy was 0.264,95%Bootstrap CI:-0.011~0.828,suggesting a non-significant moderated mediation model.3)Education:The INDEX of education was 0.044,95%Bootstrap CI:-0.245~0.380,suggesting a non-significant moderated mediation model.4)Job:The INDEX of job was 0.065,95%Bootstrap CI:-0.200~0.416,suggesting a nonsignificant moderated mediation model.5)Family income:The INDEX of income was-0.149,95%Bootstrap CI:-0.586~0.082,suggesting a non-significant moderated mediation model.6)Parity:The INDEX of parity was 0.356,95%Bootstrap CI:0.020~0.817,suggesting a significant moderated mediation model.Further analysis showed that group predicted mindfulness change positively in primipara women(B=0.798,P=0.001)while group didn’t predict mindfulness change in multipara women(B=-0.311,P=0.419).There was no direct effect from group to depression symptom change(B=-0.291,P=0.129).The indirect effect from group to depression symptom change differed between age group.For primipara e women,the indirect effect of mindfulness change was significant(B=-0.256,95%CI:-0.555~-0.023).However,for multipara women,the indirect effect of mindfulness was non-significant(B=0.100,95%CI:-0.093~0.369).7)Intended pregnancy:The INDEX of intended pregnancy was 0.098,95%Bootstrap CI:-0.386~0.449,suggesting a non-significant moderated mediation model.8)Pregnancy reaction:The INDEX of early pregnancy reaction was 0.010,95%Bootstrap CI:-0.233~0.244,suggesting a non-significant moderated mediation model.9)History of abortion:The INDEX of education was 0.008,95%Bootstrap CI:0.335~0.268,suggesting a non-significant moderated mediation model.10)History of embryo damage:The INDEX of history of embryo damage was-0.006,95%Bootstrap CI:-0.376~0.368,suggesting a non-significant moderated mediation model.11)History of chronic disease:The INDEX of education was 0.060,95%Bootstrap CI:0.441~0.632,suggesting a non-significant moderated mediation model.Conclusions:(1)IMBIs reduced depression symptom by improving mindfulness,which was the mechanism of IMBIs improving prenatal depression.(2)Age was a moderator of IMBIs improving depression symptom through minduflness.Women in non-advanced age were able to enhance mindfulness from IMBIs and then improved depression symptom.However,the result was not found in advanced-age women.This finding suggested that non-advanced-age women were target population for smartphone-based mindfulness intervention on improving depression symptom.(3)Parity was a moderator of IMBIs on improving depression symptom through mindfulness.Primipara women were able to enhance mindfulness from IMBIs and then improved depression symptom.However,the result was not found in multipara women.This finding suggested that primipara women were target population for smartphone-based mindfulness intervention on improving depression symptom.(4)Intervention focused on non-advanced-age women and multipara women,and aimed at improving postnatal depression and the whole perinatal depression need to be developed in the future. |