The study employed a nonconcurrent multiple-baseline A-B design across participants, in which A is the baseline phase and B is the intervention phase.24 Different lengths of baseline were used to ensure that scores on the outcome measures were stable and that any improvements in scores on outcome measures could be attributed to the intervention rather than to time, expectancy of receiving treatment, regression to the mean, or multiple testing. Only those effects that can be clearly observed are interpreted as providing some protection against type I error, but we supplemented visual analysis with formal clinical significance testing of the outcome measures. At entry to the study, participants were assigned to a no-intervention baseline that ranged from 3 to 6 weeks.
The following 8 survey instruments were used in this study: 1) Edinburgh Postnatal Depression Scale, 2) Structured Clinical Interview, 3) Beck Depression Inventory−second edition, 4) Hospital Anxiety and Depression Scale, 5) Postpartum Bonding Questionnaire, 6) Metacognitions Questionnaire−30, 7) Positive Beliefs about Rumination Scale, and 8) Negative Beliefs about Rumination Scale.
The Edinburgh Postnatal Depression Scale (EPDS) is a 10-item self-report measure designed to identify women at risk of developing PPD25 and is the most widely used screening measure in community samples.26 Scores on the EPDS range from 0 to 30, where scores of 10 to 12 indicate possible PPD and scores greater than 13 suggest probable PPD. A score of 10 on the EPDS was used as a cutoff for inclusion into the current study, as this is the score most commonly used in similar studies and is indicative of possible PPD. Participants completed this measure weekly during baseline and the intervention and at follow-up.
The Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV), Axis 1 Disorders (SCID-I), was used to verify the presence of major depression in participants and to ensure that this was the primary diagnosis.27
The Beck Depression Inventory (BDI-II) is a 21-item self-report measure that assesses the presence and severity of depressive symptoms, attitudes of depression, and somatic-affective and cognitive symptoms occurring in the previous 2 weeks.28 Although some items of the BDI-II may apply to all new mothers (eg, changes in sleep), it is still the best assessment of depression severity and has good specificity, sensitivity, and positive predictive value for moderate to severe depression with women in the postpartum period.29 Total scores range from 0 to 63. A score from 0 to 13 suggests an absence of or minimal depression, 14 to 19 indicates mild depression, 20 to 28 signifies moderate depression, and a score greater than 29 suggests severe depression. The BDI-II has been used as an outcome measure in similar studies.8,14,30,31 Participants completed the BDI-II pre- and posttreatment and at follow-up assessments.
The 14-item self-report Hospital Anxiety and Depression Scale32 (HADS), designed to assess the severity of both depression and anxiety, has been widely used in both community and clinical samples.33,34 The HADS has been proposed as an appropriate tool to assess postpartum anxiety as well as depression, which has been lacking in previous research.35 Scores range from 0 to 21 for both subscales. A score of 0 to 7 indicates a lack of depressive/anxious symptoms. A score between 8 and 10 suggests a “borderline” case, 11 to 15 suggests moderate symptoms, and a score greater than 16 indicates the presence of severe symptoms. This measure was completed weekly.
The 25-item Postpartum Bonding Questionnire36 (PBQ) is a widely used measure of a mother's feelings or attitudes toward her infant.37,38 Although it has 4 subscales (ie, impaired bonding, rejection and anger, anxiety about care, and risk of abuse), because of concerns over its factor structure,37,38 only total scores were analyzed. These range from 0 to 125, with higher scores indicating a potentially impaired relationship between mother and infant. This measure was completed pre- and posttreatment and at follow-up assessments.
The short-form Metacognitions Questionnaire39 (MCQ-30) was used to measure metacognitive beliefs, judgments, and thought-monitoring tendencies, which are assumed to maintain psychological disorders. Participants must rate their beliefs about particular thoughts on a 4-point scale ranging from “do not agree” to “agree very much.” Total scores range from 30 to 120; higher scores indicate a potentially unhelpful thinking style. The MCQ-30 has good internal consistency (Cronbach's alpha ranging from 0.72 to 0.93) and good test-retest reliability (r= 0.75 total score; r= 0.79 positive beliefs; r= 0.59 uncontrollability/danger; r= 0.69 confidence; r= 0.74 need for control; r= 0.87 cognitive self-consciousness).39
The 9-item Positive Beliefs about Rumination Scale40 (PBRS) and the 13-item Negative Beliefs about Rumination Scale41 (NBRS) assess positive and negative metacognitive beliefs about rumination, respectively. On a 4-point scale, participants rate how much they agree with a list of statements; a high score on both scales indicates a greater belief in rumination as a useful coping strategy. The PBRS has high internal consistency (Cronbach's alpha = 0.89) and good test-retest reliability (r= 0.85) and reasonable convergent and discriminant validity, including significant correlations with measures of rumination and depression.40 Total scores on the PBRS range from 9 to 36, with higher scores suggesting a stronger belief rating. The NBRS has good internal consistency, test-retest reliability, and convergent and concurrent validity.41 Cronbach's alphas are 0.80 and 0.83 for NBRS1 and NBRS2, respectively. Total scores on the NBRS range from 13 to 52, with higher scores suggesting a stronger belief rating. These measures were completed pre- and posttreatment and at follow-up assessments.
After gaining full ethical approval from the Stockport National Health Service and the University of Manchester's research ethics committees, participants were recruited via a primary care mental health team, general practitioners (GPs), and health visitors in greater Manchester, United Kingdom. Participants who were referred to the study had already been given a diagnosis of PPD by the referrer. Based on a risk protocol, if during the assessment or baseline period, there were any concerns regarding risk or significant lowering of mood, the researcher used a behavioral activation strategy in order to improve mood. If this was not successful or if there were significant concerns regarding risk, then the participant would be removed from the study, and routine procedures regarding risk would be employed.
To ensure that informed consent was given, participants were given a Participant Information Sheet, which outlined the nature of the study, and were asked to take a week to think about it and discuss with family and/or friends before signing a consent form. All treatment sessions were conducted in the participant's own home by a third-year trainee clinical psychologist who was trained and supervised by the developer of the MCT treatment.18 This treatment was conducted following a manualized protocol18 and consisted of 8 to 12 weekly sessions that lasted for one hour. Because of similarities between symptoms of PPD and depression and similarities between the cognitive processes of worry and rumination, the treatment protocol was based on the MCT model for depression. The protocol is outlined in Table 1.
Table 1. Typical Metacognitive Therapy Treatment Protocol
| Session || Aim || Techniques || Homework |
|1||Normalize symptoms, socialize to MCT model, and elicit formulation.||Teach attention training technique (ATT).||Listen to ATT recording.|
|2||Challenge negative beliefs about rumination starting with beliefs about uncontrollability of rumination.||Introduce detached mindfulness and rumination postponement. Conduct experiments in session.||Listen to ATT recording and apply detached mindfulness and rumination postponement.|
|3||Challenge negative beliefs about rumination, explore activity levels and avoidant coping.||Conduct experiments in session.||Listen to ATT recording, apply detached mindfulness and worry postponement, increase activity levels.|
|4||Check uncontrollability belief and challenge positive beliefs about rumination.||Socratic dialogue and guided discovery.||As above and conduct agreed-on experiments.|
|5||Continue challenging positive beliefs about rumination and explore unhelpful coping strategies.||Guided discovery and Socratic questioning. Identify advantages and disadvantages of rumination. Threat monitoring intervention.||As above, conduct experiments in which threat monitoring is banned and then increased.|
|6||Explore negative beliefs about depression and/or anxiety.||Guided discovery and Socratic questioning.||As above.|
|7||Generalize techniques. Start developing a relapse prevention strategy.||Old plan versus new plan. How did the participant respond to intrusive thoughts in the past, and how do they respond now?||As above.|
|8||Complete relapse prevention plan and identify potential future triggers.||As above.||As above.|
Home visits were used to facilitate attendance at sessions (ie, no need to arrange child care) and because it is envisaged that this treatment may be delivered in the future by a range of health care professionals who support women with PPD (such as health visitors).
The first session involved eliciting a formulation and socializing the participant to the model and aimed to show participants how their existing coping strategies (ie, ruminating/worrying) maintained their symptoms of PPD. Participants were taught attention training (ATT) to increase awareness of their attentional focus and reduce the amount of attention given to internal thoughts and symptoms. Attention training is based on principles of cognitive psychology, with participants practicing attention focusing, switching, and divided attention. They can apply this technique to shift the focus of attention away from negative thoughts (ie, worrying/ruminating).
The first goal of treatment was to challenge negative beliefs about the uncontrollability of rumination. Participants were taught detached mindfulness42 and worry/rumination postponement to reduce levels of rumination and worry. Detached mindfulness is a technique that allows a person to become detached from her or his thoughts and to observe them rather than engaging with them (by analyzing or suppressing them). When participants noticed a worrying thought, they were asked to acknowledge their worry, apply detached mindfulness to it, and postpone their worry until a specified time. Participants were allowed to worry as much as they wanted to within this period of 15 minutes if they needed to.
Maladaptive coping strategies, such as avoidance of activities, were addressed. Experiments were set up between sessions so that participants could challenge their metacognitive beliefs; for example, participants were asked to engage in activities to test the belief that they were unable to concentrate or perform effectively. When negative beliefs had decreased to zero levels of conviction, positive beliefs about rumination and worry were addressed through guided discovery and Socratic questioning.
Negative beliefs about depression and anxiety were addressed because participants may monitor themselves for signs that they are developing symptoms of PPD again. Participants were encouraged to compare their new way of managing worries with their old way. They were also invited to identify potential future triggers and consider how they would manage these using their new plan.
Participants who received more than 8 sessions of MCT were slower to challenge negative beliefs about rumination (eg, “I have no control”) and generalize techniques. Extra sessions focused on conducting experiments in session and verbal reattribution aimed at challenging these beliefs.
Inclusion and Exclusion Criteria
Participants were at least 18 years of age and scored 10 or more on the Edinburgh Postnatal Depression Scale. Postpartum depression was required to be the participant's primary difficulty, and they were stable on pharmacologic treatment for one month (if applicable).
Participants were excluded from the study if they were actively psychotic or used alcohol or other substances. Because of limited resources, recruitment of participants was restricted to those who did not require the use of an interpreter.
Clinical and Reliable Change Calculations
Mean baseline EPDS scores were above the cutoff for PPD, with 3 participants presenting with severe depression, as measured by the BDI-II. The cutoff point for the EPDS was calculated using Jacobson et al's formula.44 This calculation indicated that participants must score below 9 on the EPDS posttreatment to ensure that they have moved from a clinical population to a “normal” population. This figure was lower than the cutoff score of 10 for possible PPD.25
Similar calculations were conducted for the anxiety and depression subscales of the HADS. The anxiety subscale required participants to report scores of less than 11 posttreatment, which was the same as the cutoff point suggested by the authors.32 The depression subscale required participants to score less than 8 posttreatment to be considered to have made a clinically significant change. This score was lower than the recommended published cutoff score of 11 for this subscale.
In terms of the reliable change index (RCI), a 4-point change on the EPDS was required to ensure that reliable change had occurred. The BDI-II required a change of 11 points to ensure that a statistically significant change had occurred. An RCI for the HADS indicated that a change of 4 points was needed on both scales to ensure that a reliable change had occurred. Therefore, to demonstrate both reliable and significant clinical change in scores posttreatment, scores had to decrease by the number of points as suggested by the RCI calculations and had to be below c (ie, EPDS <9, HADS-Anxiety <11, HADS-Depression <8).