Illness perceptions in mothers with postpartum depression - Midwifery

Midwifery 29 (2013) 779–786

Contents lists available at SciVerse ScienceDirect

Midwifery journal homepage: www.elsevier.com/midw

Illness perceptions in mothers with postpartum depression Tineke Baines, Clin Psy D (Clinical Psychologist)a,c,d, Anja Wittkowski, Clin Psy D (Lecturer, Clinical Psychologist)a,b,e,n, Angelika Wieck, Dr Med, FRCPsych (Consultant, Senior Lecturer)a,b a

University of Manchester, Manchester M13 9PL, UK Manchester Mental Health and Social Care NHS Trust, Laureate House, Wythenshawe Hospital, Southmoor Road, Manchester M23 9LT, UK c Child and Adolescent Mental Health Service (CAMHS), Fieldhead House, Bradford, UK d Learning Disabilities CAMHS Team, Acorn Centre, Barnsley, UK e Mother and Baby Unit, Wythenshawe Hospital, Southmoor Road, Manchester M23 9LT, UK b

a r t i c l e i n f o

abstract

Article history: Received 23 February 2012 Received in revised form 27 June 2012 Accepted 29 June 2012

Objective: to examine perceptions of mothers experiencing postpartum depression utilising the revised Illness Perception Questionnaire (IPQ-R), to explore relationships between illness perceptions, depression severity and perceptions of maternal bonding, and to assess the psychometric properties within this population. Design: longitudinal correlational design. Setting: North West of England, UK. Participants: 43 mothers, who screened positive for postpartum depression (mean age 29.36 years) with babies whose mean age was 4 months. Methods: participants were recruited through health services. Participants completed the IPQ-R and measures of depression severity and maternal bonding. Illness perceptions and depression severity were assessed at 2 time points, 4 weeks apart. Findings: mothers endorsed IPQ-R subscales of cyclical timeline, consequences, emotional representations, treatment and personal control. IPQ-R subscale scores and depression severity correlated significantly at Time 1. Initial IPQ-R subscale scores, however, were not associated with and accounted for little variation in depression severity at Time 2. IPQ-R identity and consequence subscales positively correlated with perceived bonding difficulties. Key conclusions and clinical implications: the IPQ-R was shown to be a reliable measure of illness perceptions in mothers experiencing postpartum depression. The maternal illness perceptions endorsed in this study have implications for clinical practice. Interventions aimed at developing a more coherent understanding of depression may enhance beliefs of personal control over symptoms, reduce the number of perceived symptoms and associated emotional distress. Educating mothers regarding the benefits of interventions may be important in increasing the number of mothers accessing support for postpartum depression. & 2012 Elsevier Ltd. Open access under CC BY license.

Keywords: Illness perceptions Postnatal depression Questionnaires

Introduction One of the most widely used theoretical models to explore how individuals perceive their physical health problems is the Self-Regulation Model (SRM) developed by Howard Leventhal. The SRM proposes that a causal relationship exists between illness beliefs and health outcomes, which is mediated by coping behaviours (Leventhal et al., 1984; Hampson et al., 1990, 1995). The SRM also proposes that illness perceptions are represented as five main dimensions (Leventhal and Diefenbach, 1991) (1) n Corresponding author at: Division of Clinical Psychology, School of Psychological Sciences, University of Manchester, 2nd Floor Zochonis Building, Brunswick Street, Manchester M13 9PL, UK. E-mail address: [email protected] (A. Wittkowski).

0266-6138 & 2012 Elsevier Ltd. Open access under CC BY license. http://dx.doi.org/10.1016/j.midw.2012.06.020

causal—beliefs about the cause(s) of the illness, (2) identity—beliefs concerning the illness’ label and symptoms, (3) timeline—perceptions about the time course of an illness, characterised along the acute–chronic dimension where individuals may perceive their illness as chronic or acute, or cyclical in nature (where the condition appears under a particular set of circumstances, such as after stressful life events), (4) cure–control—beliefs about how the condition is treated and effectiveness of available treatment and (5) consequences—perceived effect(s) of the illness on an individual’s life. The Illness Perception Questionnaire (IPQ) (Weinman et al., 1996) was developed as a measure of the SRM dimensions. It was later revised (IPQ-R) (Moss-Morris et al., 2002) to include an assessment of emotional representation (one’s emotional response to the illness) and illness coherence (the sense of having a

780

T. Baines et al. / Midwifery 29 (2013) 779–786

comprehensive understanding of the illness). In addition, the timeline scale was separated into acute–chronic and cyclical and the control scale into personal control and treatment control. The IPQ and IPQ-R have been used extensively in physical health and have been shown to be valid and reliable measures of illness beliefs (Moss-Morris et al., 2002). Within the last decade research has applied the SRM to explore illness perceptions in mental health. For example, the IPQ has been used with depressed participants: firstly in primary care patients (Brown et al., 2001), then in an all-female sample with a history of depression approximately half of whom were still actively depressed (Fortune et al., 2004) and more recently in a sample of women experiencing antenatal depression who were followed up over time to assess treatment use (O’Mahen et al., 2009). Across these three studies, individuals perceived their depression as having many symptoms and negative consequences, being cyclical in nature and amenable to control/cure. Commonly attributed causes of depression included; stress, own behaviour, hereditary, relationship and physical health difficulties. Women experiencing depression within the perinatal period endorsed medical illness causes, attributing their depression to their own state of mind and pregnancy-related changes, such as hormones, lack of sleep and difficulties in adjusting to being pregnant (O’Mahen et al., 2009). In terms of correlations between illness dimensions, Fortune et al. (2004) noted that women who endorsed more depressive symptoms perceived their depression as chronic, with many negative consequences. Women who perceived their depression as chronic were also less likely to believe depression was controllable and viewed it as having more personal consequences. Overall, the IPQ was found to be a reliable measure of illness perceptions for depression and discriminated between women actively depressed and those who were not. However, during interviews with participants, a number of the causal subscale items were deemed non-applicable to depression (Fortune et al., 2004). In terms of outcomes, perceptions of depression were significantly associated with current and past treatment-seeking behaviour, medication adherence and coping strategies. Beliefs regarding the cause, seriousness and response to medication had the greatest influence on treatment use and medication adherence (Brown et al., 2001). In addition, perceived illness chronicity influenced treatment-seeking (O’Mahen et al., 2009). However, as the original version of the IPQ was used in these three studies, illness representations and coherence were not explored. As cross-sectional designs were used, the direction of relationships between illness perceptions, illness severity, coping and treatment-seeking behaviour could not be determined (Brown et al., 2001; Fortune et al., 2004). Despite some exploration of depression-related perceptions, there has been little research examining illness beliefs of mothers experiencing depression after the birth of a child and the impact of these beliefs on bonding. Postpartum depression (PPD) affects 10–15% of mothers (O’Hare and Swain, 1996). There is evidence suggesting the nature of symptoms differs little between depression within the postnatal period and depression experienced at any other time within the life cycle (Whiffen and Gotlib, 1993). However, studies highlight that mothers experiencing PPD report a greater frequency of aggressive thoughts towards their baby than depressed mothers with non-postpartum onset (Wisner et al., 1999). Crucially, PPD has been associated with negative outcomes for children whose mothers experience this depression, including poorer cognitive and emotional development, poor attachment and behavioural problems (Murray et al., 1996). Even mild depressive symptoms can have a significant impact on maternal bonding (Moehler et al., 2006). PPD can also be the first episode in

a life-long pattern of recurrent depression (Boyce and Stubbs, 1994) and is associated with an increased usage of health services (Dennis, 2004) alongside significant cost expenditure for health services (Civic and Holt, 2000). Therefore, effective treatment and adequate support for mothers with depression after childbirth are paramount. Clinical guidelines (National Institute of Clinical Excellence, 2007) recommend that, depending on severity, women experiencing depression during the postnatal period should be offered access to guided self-help and other forms of effective talking therapies. However, some mothers with PPD are reluctant to seek professional help for fear they will be admitted to a psychiatric unit, get ‘locked up’ or have their baby taken away from them (Hall, 2006). In a qualitative study conducted by McIntosh (1993), less than half of the depressed mothers interviewed (18 out of 38) sought assistance: Health-care professionals were regarded either as threatening or as an inappropriate solution to the problem. To enable appropriate opportunities for mothers to disclose their feelings and seek support from services, it is essential that health-care professionals have an understanding of how mothers perceive their depression during this time. Identifying illness perceptions in mothers with PPD may help to refine treatment and determine whether aspects of illness perception are predictive of depression severity. Using the IPQ-R, this preliminary study examined illness perceptions of mothers experiencing PPD. The aims of the study were to examine (1) perceptions of mothers experiencing postpartum depression utilising the revised Illness Perception Questionnaire (IPQ-R), (2) to explore relationships between illness perceptions, depression severity and perceptions of maternal bonding, and (3) to assess the psychometric properties within this population.

Methods Design A longitudinal correlational questionnaire design was used to explore illness perceptions and the psychometric properties of the IPQ-R.

Measures The Illness Perceptions Questionnaire revised (IPQ-R) The IPQ-R, modified for PPD, was used to measure illness perceptions across the SRM dimensions: identity, cause, timeline (acute–chronic and cyclical), illness coherence, consequences, cure–control (personal and treatment control) and emotional representations. As the authors recommend modifications, 2 subscales were revised. The identity subscale was revised to include the Diagnostic and Statistical Manual of Mental Disorders Fourth Edition (American Psychiatric Association (APA), 1994) criteria for major depressive disorder, namely: ‘depressed mood’, ‘loss of interest and/or pleasure in activities’, ‘weight loss/gain (without dieting)’, ‘feelings of restlessness’, ‘feelings of worthlessness’, ‘difficulty concentrating’, ‘fatigue/loss of energy’, ‘sleep difficulties’ ‘recurrent thoughts about death’ and ‘thoughts of suicide or self-harm’. The remaining items of the identity subscale were ‘fluctuations in mood’, ‘over preoccupation with baby’s well-being and complete disinterest in baby’. Participants were asked whether they had experienced each symptom since their depression (yes/no) and whether they believed it was related to their depression and/or to being a mother with a young child. Yes responses on symptoms attributed to depression were given a


score of 1 and produced the identity subscale score. The identity scale comprised of 14 items. Six of the 18 items of the original causal subscale, deemed to be unlikely causes of depression after childbirth, were removed (e.g., ‘a germ or virus’), while ‘ageing’ was changed to ‘age’. Nine items were added including ‘limited social support’, ‘my pregnancy was unplanned’, ‘there were complications during the pregnancy’, ‘history of depression—postnatal or otherwise’, ‘history of other mental health problems’, ‘hormonal changes’, ‘difficult labour’, ‘post birth complications with self or baby’ and ‘unrealistic expectations of motherhood’. This subscale consisted of 21 items. Whilst the timeline subscales (acute–chronic and cyclical), illness coherence, cure–control (personal and treatment control) and consequences subscales remained unchanged, 1 item was removed from the emotional representation subscale (‘I get depressed when I think about my illness’), because the participants were depressed. Participants were asked to rate their level of agreement with each item on a 5-point Likert scale. Some items were reverse scored (1, 4, 8, 15, 17–19, 23–27, 35). Participants also had the option of rating items as ‘not relevant to my depression’ or ‘too difficult or cannot answer’; items rated as such were not allocated a score.

The Edinburgh Postnatal Depression Scale (EPDS) The EPDS (Cox et al., 1987) is an effective screening tool and a reliable and validated measure of depression after childbirth (Cox and Holden, 2003). It has 10 items scored on a 4-point Likert scale, with a maximum score of 30. A score of 12 or above can identify women with major depression. However, a cut-off score of 10 significantly reduces the risk of failing to identify mothers within community samples who may display milder depression (Cox et al., 1987). Therefore, a minimum score of 10 was used as the study inclusion criterion.

The Patient Health Questionnaire 9 (PHQ-9) The PHQ-9 (Kroenke et al., 2001) is a reliable and validated measure of depression severity. Participants are asked to score themselves on 9 statements based on the DSM-IV (1994) criteria for depression. Statements are scored on a Likert scale ranging from ‘0’ (not at all) to ‘3’ (nearly every day), producing a maximum score of 27 (1–4¼ minimal depression, 5–9 ¼mild depression, 10–4 ¼moderate depression, 15–19¼moderately severe depression and 420 ¼severe depression). The PHQ-9 has been used to assess depression severity of mothers in the postpartum period (Pawar et al., 2011).

The Postpartum Bonding Questionnaire (PBQ) The 25-item PBQ (Brockington et al., 2001) is a self-report measure assessing maternal perceptions of bonding with their baby. After service user feedback, 5 items, which were believed to potentially cause offense, were removed from the measure. This decision was supported by other studies reporting difficulties with these items (Brockington et al., 2001; Reck et al., 2006; Wittkowski et al., 2007, 2010). Consequently, only total sores were calculated and analysed. The maximum score was 100. A higher score indicates problematic bonding. A demographic questionnaire was also used to obtain social and demographic information, such as participants’ ages, infant’s age, marital and occupational status, number of children, current and past psychological difficulties and treatment use.

781

Participants and recruitment All participants were mothers who had a child under the age of 1 year and who screened positive for PPD on the EPDS. Participants were excluded if their baby had a developmental delay suggestive of a learning disability. Participants were recruited via health visitors, perinatal outpatient and inpatient services, Children and Parent Services, Children’s Centres across Greater Manchester, United Kingdom, and MIND centres nationwide. Participants were initially approached regarding the research by professionals/staff who they were familiar with. Online adverts were placed on relevant maternal, mental health and PPD websites.

Procedure This study had full National Health Service (NHS) and University ethical and research approval. Participants were given questionnaire packs, including a participant information sheet, a consent form, contact details for the main researcher, copies of all the questionnaires, and a stamped addressed envelope. They could complete measures online via a secure server or return paper copies. Participants had the option of being assisted by the researcher or their health-care professional to offer support in case completion of the questionnaires led to any potential emotional distress for participants. In order to assess test–retest reliability and the predictive validity of the IPQ-R, participants were asked to complete the IPQ-R and PHQ-9 again after 4 weeks (Field, 2009).

Power calculation and statistical analyses Based on data provided by Fortune et al. (2004), a sample size of n¼30 provided 80% power to detect an underlying correlation of .5 or more as statistically significant (the significance and power levels were set at .05 and .80, respectively; Field, 2009). A minimum of 20 participants were required to have 80% power of detecting significant differences between mothers who experienced past psychological difficulties and mother who had not. The distribution of the data was assessed using Shapiro–Wilk tests of normality. Correlational analyses were used to assess the IPQ-R’s psychometric properties; Spearman correlation coefficients were calculated to examine the predictive validity, intersubscale correlations, associations between IPQ-R subscale scores and test–retest reliability. Wilcoxon tests were also used to assess differences between IPQ-R scores between Time 1 and 2. Cronbach’s alpha was used to assess the internal consistency and inter-item correlations for IPQ-R subscales. Univariate and multiple regressional analyses (beta) were used to explore whether IPQ-R subscale scores were predictive of depression severity.

Findings Sample characteristics Forty-three women participated in this study. The mean age of mothers was 29.36 (SD 5.79, range¼ 18–40). Their infants had a mean age of 4 months (SD 2.05). Just under half (46.5%, n ¼20) of participants had their baby via normal vaginal birth, 30% (n¼ 13) via caesarean section and 2% (n ¼1) via forceps. The majority of participants (65%, n ¼28) identified their ethnicity as White British. Full participant social and demographic information is displayed in Table 1.

782


Table 1 Social and demographic participant information. Social and demographic variables

n

%

Marital status Single Married Divorced Other Number of children One Two Three Four Ethnicity White British White Irish Mixed White and Asian Black or Black British African Asian or Asian British Pakistani Educational attainment G.C.S.E. level or equivalent A-level or equivalent Degree level or equivalent Employment status Full time Part time Housewife Student Unemployed Recruitment source Online or poster advert Health visitor Parenting course GP Psychologist Psychiatrist Midwife Health during pregnancy Good Okay Poor Difficult Complications Support available from friends Always Most of the time Some of the time Rarely Never Support available from family Always Most of the time Some of the time Rarely Never Support available from husband/partner Always Most of the time Some of the time Rarely Never

10

23

21 1 2

49 2 5

19 11 3 1 28

44 26 7 2 65

3 1 1 1 10

7 2 2 2 23

15 9

35 22

11 13 6 3 1

26 30 14 7 2

13 13 4 1 1 1 1

30 30 9 2 2 2 2

15 11 2 3 1

35 26 5 7 2

10 11 12 3 5

23 26 30 7 12

13 8 6 2 4

30 19 14 5 9

18 6 4 3 3

42 14 9 7 7

positive correlation between EPDS and initial PHQ-9 total scores (rs ¼.69; po.01). Just over half of the mothers reported receiving a diagnosis of PPD. Most mothers (72%, n ¼31) identified themselves as currently experiencing psychological problems: depression (35%, n¼15), depression and anxiety (35%, n ¼15), and depression and post-traumatic stress disorder (2%, n ¼1). Of those reporting current difficulties, 87% (n ¼27) were receiving treatment (including talking therapy and medication). Just under half of the sample reported experiencing past psychological problems (46.5%, n¼20); depression (27.9%, n ¼12) and anxiety (14%, n ¼6). Sixty per cent (n ¼12) of participants with past problems reported having received treatment in the past.

Illness perceptions As many of the IPQ-R subscales at Time 1 were not normally distributed, parametric analyses were not appropriate. Descriptive statistics for the IPQ-R subscales are shown in Table 2. Mothers endorsed just over half the symptoms within the identity subscale, perceiving their depression to have a high number of consequences, that it was amenable to personal control and treatment, was cyclical in nature and caused a high emotional response. Commonly experienced symptoms included depressed mood (endorsed by 95% of mothers), difficulties concentrating (87.5%), loss of interest/pleasure in activities (82.5%), fatigue/loss of energy (80%) and sleep difficulties (77.5%). Many mothers attributed sleep difficulties (90%), fatigue/loss of energy (80%) and over-preoccupation with baby’s well-being (65%) to being a mother with a young child. As there were an insufficient number of participants to undertake a factor analysis of the causal subscale, the percentages of participants in agreement (either agreed or strongly agreed) were calculated for each item. The most commonly endorsed causes of depression after childbirth were stress or worry (endorsed by 85% of the sample), hormonal changes (75%), own emotional state (75%), family problems (67.5%), mental attitude (62%) and own behaviour (55%). The least endorsed causes were age (12.5%) and accident or injury (10%).

Bonding The PBQ scores were not normally distributed. The PBQ median was 13.5 (range ¼0–73), which was higher compared to previous studies of mothers with PPD (mean ¼10.46, SD¼7.43) (Moehler et al., 2006) and with no depression (median¼7, range¼0–40) (Wittkowski et al., 2007). The range of PBQ scores was indicative of considerable variation in maternal responses.

The reliability of the IPQ-R for PPD Postpartum depression (PPD) While the EPDS total scores were not normally distributed, PHQ9 scores were. The EPDS median was 17 (range¼10–27) confirming that participants were experiencing probable depression, with 39 participants scoring 12 or above, 2 scoring 11 and 4 scoring 10. The mean PHQ-9 score of 12.66 (SD 6.58) at Time 1 and 11.32 (SD 6.15) at Time 2 indicated that most participants experienced moderately severe depression. A paired samples t-test showed a significant reduction between PHQ-9 scores for participants who completed the measure at both Time 1 and 2 (n¼19), t(18)¼ 2.51; po.05. As expected, a Spearman’s correlation coefficient revealed a significant

The reliability of the modified IPQ-R subscales was examined in terms of internal consistency, test–retest reliability, inter-item correlations and inter-subscale correlations.

Internal consistency Cronbach’s alpha scores were calculated to assess the IPQ-R’s internal consistency (see Table 2). Alpha values should ideally range between .70 and .90 (Streiner and Norman, 1995). All subscale values were within this desired range, with the exception of the cyclical timeline subscale (rs ¼.44).


783

Table 2 IPQ-R median scores, Cronbach’s alpha, inter-item correlations and test–retest reliabilities. IPQ-R subscales

Identity (14 items) Timeline acute–chronic (6 items) Consequences (6 items) Personal control (6 items) Treatment control (5 items) Illness coherence (5 items) Timeline cyclical (4 items) Emotional representations (5 items) n

Maximum possible score for each subscale

Median (range) (n ¼40)

Alpha (a)

Mean inter-item correlation

Test–retest (Spearmen’s rho) (n ¼20)

Test–retest Wilcoxon z scores

p-Values

14 30 30 30 25

9 (0–14) 15 (4–29) 21 (6–26) 22 (11–30) 18 (8–25)

.80 90 .86 .87 .70

.35 .59 .49 .54 .37

rs ¼ .49* rs ¼.69** rs ¼.50* rs ¼.57** rs ¼.65**

.72 .80 .35 .28 .14

.47 .42 .73 .78 .89

25 20 25

15.5 (4–25) 14 (8–20) 19 (5–25)

.96 .44 .90

.82 .33 .63

rs ¼.75** rs ¼.40* rs ¼.67**

2.21 .36 .99

.03* .72 .32

p o .05. p o.01.

nn

Test–retest reliability The stability of the IPQ-R subscales was assessed over a 4-week period. Spearman’s correlations, calculated to assess degrees of association between subscale scores for participants who completed the IPQ-R both at Time 1 and 2 (n¼20), showed significant positive correlations, ranging from r¼.40 to .75. Wilcoxon tests assessing differences between IPQ-R scores over 4 weeks, indicated that illness coherence beliefs did show a slight increase over this period (z¼ 2.21; po.05) (see Table 2). No significant differences were found between the scores at Time 1 and 2 for the other IPQ-R subscales (z values ranging from .141 to .99; p4.05). These results suggest IPQ-R subscales were reliable over a 4-week period. Inter-item correlations Cronbach’s alpha was used to assess inter-item correlations for items within each IPQ-R subscale. Briggs and Cheek (1986) recommend that correlations ranged between .20 and .40. Mean inter-item correlations ranged from .35 to .82 (see Table 2). As the acute–chronic timeline, consequences, personal control, illness coherence and emotional representations subscales values were above the recommended range, some items within these subscales may be synonymous or redundant. Inter-subscale correlations Using Spearman’s correlations, significant inter-correlations were found between the IPQ-R subscales (Table 3). Mothers, who perceived their depression as having many negative consequences and symptoms, viewed their depression as more chronic, were less likely to have a coherent understanding of their depression and had a strong emotional response. In contrast, mothers with a coherent understanding of their depression were more likely to perceive that their own actions and treatment were able to control symptoms. Chronicity was linked to a poorer coherent understanding, a stronger emotional response and more depressive symptoms. Unsurprisingly, a greater number of depressive symptoms were associated with a stronger emotional response. Predictive validity of the IPQ-R and associations with depression severity Spearman’s correlations were calculated between initial IPQ-R subscale and total PHQ-9 follow-up scores to assess whether illness perceptions were predictive of depression severity 4 weeks later (see Table 3). However, no significant correlations were identified (correlations ranged from rs ¼ .003 to .31, p4.05). Calculation of R2s values indicated that IPQ-R subscales accounted

for little variation within follow-up PHQ-9 score ranks (between 0% and 10%). This suggests that illness perceptions were not predictive of depression severity within this timeframe. However, significant correlations were found between IPQ-R subscale and PHQ-9 total scores at Time 1. Mothers, who endorsed more depressive symptoms, believed their depression was chronic and perceived a stronger emotional response, also rated their depression as more severe, while mothers with perceived personal control had lower depression severity scores. Univariate regression analyses revealed that the identity (beta¼.54, p o.01), personal control (beta ¼ .40; p o.05) and emotional representations (beta¼.51, p o.05) subscales at Time 1 were significant predictive of depression severity at Time 1. The acute/chronic timeline (beta¼.30, p4 .05) subscales and whether or not mothers had previously experienced psychological difficulties (beta¼ .11, p 4.05) were not significant predictors of depression severity. However, within stepwise multiple regression analyses only the IPQ-R identity subscale was shown to be a significant predictor. This may be explained by the fact that the identity, personal control and emotional representations IPQ-R subscales were all significantly correlated with one another. Relationship between illness perceptions and bonding Spearman’s correlations were calculated to explore the relationship between illness perceptions and maternal perceptions of bonding (Table 3). Significant positive correlations were noted for PBQ scores and IPQ-R identity and consequence subscales, suggesting mothers who endorsed more symptoms and more negative depression-related consequences appeared to have more difficulties in bonding with their baby. Illness perceptions of mothers with and without past psychological difficulties A series of independent samples t-tests failed to reveal significant differences in illness perceptions between mothers who experienced past psychological difficulties (n ¼20) and those who had not (n ¼13) (t-values ranged from t(30)¼.14 to 1.02; p4.05) (see Table 4). This suggests that mothers who were experiencing psychological difficulties for the first time had similar illness beliefs to those who experienced difficulties previously. Discussion The IPQ-R was shown to be a reliable measure for assessing illness perceptions in mothers experiencing depression after

784


Table 3 Spearman’s inter-correlations between IPQ-R subscales and PBQ total scores at Time 1 and PHQ-9 total scores at Time 1 and 2. Scales

Identity

Timeline acute– chronic

Consequences

Personal control

Treatment control

Timeline cyclical

Illness coherence

Emotional representations

– – – – – – – .41** .13 (.02) .45*

.27* – – – – – – .30* .22 (.05)

.42** .52** – – – – – .18 .04 (.00)

.27* .18 .14 – – – – –.38* .003 (.00)

.07 .25 .17 .26* – – – .20 .26 (.08)

.09 .17 .02 .003 –.04 – – .22 .16 (.03)

.21* .34* .28* .58** .27* .04 – .14 .31 (.1)

.50** .51** .53** .36* .03 .13 .45** .44* .13 (.02)

.27

.31*

.24

.19

.1

.17

.28

Identity Timeline acute–chronic Consequences Personal Control Treatment control Timeline cyclical Illness coherence PHQ-9 total score time 1 PHQ-9 total score time 2 (R2s ) PBQ total score n

p o .05. p o.01.

nn

Table 4 Independent samples t-tests between IPQ-R subscale scores for mothers with and without past psychological problems. IPQ-R subscales

Identity Timeline acute–chronic Consequences Personal control Treatment control Timeline cyclical Illness coherence Emotional representations

Past psychological difficulties (n ¼ 20)

No past psychological difficulties (n ¼ 13)

M

SD

M

SD

9.1 15.55 20.20 21.15 16.75 14.15 15.5 17.80

3.5 7.58 5.13 4.75 2.95 2.80 6.07 4.73

8.62 14.69 19.1 20.92 18 14.31 13.54 17.08

2.9 6.58 5.12 4.54 4.10 2.18 5.81 5.42

childbirth. All subscales apart from the cyclical timeline scale showed acceptable levels of internal consistency, although analyses indicated some items on multiple IPQ-R subscales may be redundant. All subscales were shown to be reliable over a 4-week period. Whilst mothers in this sample endorsed a moderate number of symptoms and viewed their depression as having negative consequences with an emotional impact, they also perceived depression to be time-limited and that symptoms could be controlled by their own actions or treatment. These observations are largely consistent with the findings of the previous studies using the IPQ in depression (Fortune et al., 2004; O’Mahen et al., 2009). Mothers appropriately identified depressed mood, difficulties concentrating, loss of interest/pleasure in activities, fatigue/loss of energy and sleep difficulties as symptoms related to their depression. This sample attributed the cause of their depression to stress or worry, hormonal changes, own emotional state, family problems, mental attitude and own behaviour, which is consistent with previous research (O’Mahen et al., 2009). Surprisingly, there were no significant differences between illness perceptions of mothers experiencing psychological difficulties for the first time and mothers with a past history. There were also associations between illness perceptions. For example, mothers who perceived their depression as having many negative consequences were less likely to believe they had a coherent understanding of their depression; they viewed their depression as chronic, with many symptoms and responded more emotionally. In contrast, mothers who had a more coherent understanding of their depression were more likely to believe their own actions and treatment could control their symptoms. A greater degree of perceived depressive symptoms was, perhaps unsurprisingly, related to a greater emotional response. Fortune et al. (2004) also found positive correlations between their IPQ

t-Values

p-Values

.41 .33 .62 .14 1.02 .17 .92 .41

.68 .74 .54 .89 .32 .87 .36 .69

identity, chronicity, and consequence subscales, with the control/ cure subscale negatively correlating with these three subscales. Depression severity was associated with particular illness perceptions. More depressed mothers seemed to endorse more depressive symptoms, responded to depression with more emotions and believed their depression was chronic. Negative thinking, which is characteristic of depression, may have resulted in mothers having more pessimistic illness perceptions. In contrast, mothers who were less seriously depressed indicated feeling more control over their symptoms. Similar correlations between IPQ dimensions and depression severity have been observed (Fortune et al., 2004). Within multiple regression analyses only the perceived number of symptoms experienced (IPQ-R identity subscale scores) was predictive of depression severity. In terms of bonding, mothers who viewed their depression as having many symptoms and negative consequences also indicated experiencing more difficulties in bonding with their baby. Mothers experiencing a number of depressive symptoms and associated consequences (i.e., relationship or financial difficulties) may well find it difficult to closely attend to their infant and develop a deeper bond. Limitations A number of limitations require consideration. The sample size was relatively small (n ¼43), but comparable to other studies (n ¼41; see Brown et al., 2001). However, the analyses exploring the predictive validity of the IPQ-R and effects of a past history of depression on illness perceptions were underpowered. Caution is also required when interpreting the study’s analyses as the use of multiple comparisons increased the potential risk of statistically significant results being obtained by chance. Studies with larger samples are required to further support the findings of this study.


Whilst there was no formal assessment of diagnosis, the majority of participants did report receiving a diagnosis of PPD and scores on the EPDS and PHQ-9 indicated mothers were experiencing depressive symptoms of moderate severity. As a number of women were accessing support and reported receiving a diagnosis of PPD, the participants sampled may have been much more likely to hold an ‘illness model’ of their mental health difficulties. Mothers not accessing services may hold different perceptions regarding their difficulties and not conceive them within the context of being an ‘illness.’ Illness perceptions and depression severity were reassessed at 4 weeks follow-up. However, a longer follow-up period would allow further exploration of the relationship between illness perceptions and depression severity and changes in illness perceptions over time. At follow-up, there was some decline in PHQ-9 scores, yet mothers were still experiencing moderate depression severity and IPQ-R scores remained relatively stable. Longitudinal intervention studies (where illness beliefs are the target of intervention) may help determine the nature of the relationships between illness perceptions and depression severity and also explore how illness perceptions change over time. Clinical implications Although the small sample size limits drawing strong general conclusions from the study’s results, the findings do suggest the IQP-R with some minor modification may appropriately be used as a measure of illness perceptions in PPD and may help understand how mothers view their illness. The IPQ-R may be a useful assessment tool to help explore the influence of illness beliefs on treatment use. As certain illness perceptions were shown to correlate with depression severity at least in the short term, the measure could be used as a screen for all mothers to see who could potentially be at more risk of experiencing severe depression, based upon how they perceive PPD. These results also highlight a number of implications for clinical practice. Interventions aimed at allowing mothers to develop a more coherent understanding of their depression, for example, through the development of a greater understanding or formulation may improve outcome. Counselling and psychological therapies can help mothers to address perceptions regarding control over their difficulties and develop effective coping strategies, which could lead to improved bonding between mother and baby. The majority of mothers did not perceive their depression as chronic. Some mothers may have viewed their depression as a unique experience comparable to the Baby Blues, whereas those with past psychological difficulties may have viewed their depression as cyclical. Psycho-education on possible depression chronicity may be important in helping mothers to understand the preventative role of treatment (Goodman et al., 2005). As many mothers sampled were accessing support and endorsed a number of treatment control beliefs, education about potential benefits of depression interventions may be valuable in increasing the number of mothers accessing support for PPD. Mothers’ experience of receiving support may have influenced their perceptions regarding treatment, stressing the importance of mothers experiencing interventions as supportive and effective. Finally, the IPQ-R was shown to be a reliable measure of illness perceptions in mothers experiencing PPD, providing valuable insight into their views of depression. Conflict of interest The authors have no conflicts of interest to disclose.

785

Acknowledgements The authors would like to express their sincere gratitude to all mothers who participated in this research study and to staff across all recruitment sites that helped identify participants and distribute questionnaires.

References American Psychiatric Association (APA), 1994. Diagnostic and Statistical Manual of Mental Disorder Fourth Edition (DSM-IV). American Psychiatric Association, Arlington. Boyce, P.M., Stubbs, J.M., 1994. The importance of postnatal depression. Medical Journal of Australia 161, 471–472. Briggs, S.R., Cheek, J.M., 1986. The role of factor analysis in the development and evaluation of personality scales. Journal of Personality 54, 106–148. Brockington, I.F., Oates, J., George, S., et al., 2001. A Screening Questionnaire for mother–infant bonding disorders. Archives of Women’s Mental Health 3, 133–140. Brown, C., Dunbar-Jacob, J., Palenche, D.R., Kelleher, et al., 2001. Primary care patients’ personal illness models for depression: a preliminary investigation. Family Practice 18, 314–320. Civic, D., Holt, V.L., 2000. Maternal depressive symptoms and child behaviour problems in a nationally representative normal birthweight sample. Maternal and Child Health Journal 4, 215–221. Cox, J., Holden, J. 2003. Perinatal Mental Health: A Guide to the Edinburgh Postnatal Depression Scale (EPDS). Gaskell, London. Cox, J.L., Holden, J.M., Sagovsky, R., 1987. Detection of postnatal depression: development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry 150, 782–786. Dennis, C.L., 2004. Can we identify mothers at risk for postpartum depression in the immediate postpartum period using the Edinburgh Postnatal Depression Scale? Journal of Affective Disorders 78, 163–169. Field, A., 2009. Discovering Statistics Using SPSS, 3rd edn. Sage, London. Fortune, G., Barrowclough, C., Lobban, F., 2004. Illness representations in depression. British Journal of Clinical Psychology 43, 347–364. Goodman, D.M., Morrissey, S.A., Graham, D., et al., 2005. The application of cognitive behaviour therapy in altering illness representations of Systemic Lupus Erythematosus. Behavioural Change 22, 156–172. Hall, P., 2006. Mothers’ experiences of postnatal depression: an interpretative phenomenological analysis. Community Practice 79, 256–260. Hampson, S.E., Glasgow, R., Toobert, D.J., 1990. Personal models of diabetes and their relations to self-care activities. Health Psychology 9, 632–646. Hampson, S.E., Glasgow, R.E., Foster, L.S., 1995. Personal models of diabetes among older adults: relation to self-management and other variables. The Diabetes Educator 21, 300–307. Kroenke, K., Spitzer, R.L., Williams, J.B., 2001. The PHQ-9: validity of a brief depression severity measure. Journal of General Internal Medicine 16, 606–613. Leventhal, H., Diefenbach, M., 1991. The active side of illness cognition. In: Skelton, JA, Coryle, RT (Eds.), Mental Representation in Health and Illness. SpringerVerlag, New York. Leventhal, H., Nerenz, D.R., Steele, D.F., 1984. Illness representations and coping with health threats. In: Singer, A, Singer, J (Eds.), A Handbook of Psychology and Health. Erlbaum, Hillsdale, NJ. McIntosh, J., 1993. Postpartum depression; women’s help-seeking behaviour and perceptions of cause. Journal of Advanced Nursing 18, 178–184. Moehler, E., Brunner, R., Wiebel, A., et al., 2006. Maternal depressive symptoms in the postnatal period are associated with long-term impairment of mother– child bonding. Archives of Women’s Mental Health 9, 273–278. Moss-Morris, R., Weinman, J., Petrie, K.J., et al., 2002. The revised illness perception questionnaire IPQ-R. Psychological Health 17, 1–16. Murray, L., Fiori-Cowley, A., Hooper, R., et al., 1996. The impact of postnatal depression and associated adversity on early mother infant interactions and later infant outcome. Child Development 67, 12–26. National Institute of Clinical Excellence, 2007. Guidelines for Antenatal and Postnatal Mental Health. CG45. National Institute for Health and Clinical Excellence, London. O’Hare, M.W., Swain, A.M., 1996. Rates and risks of post-partum depression: a meta-analysis. International Review of Psychiatry 8, 37–54. O’Mahen, H.A., Flynn, H.A., Chermack, S., et al., 2009. Illness perceptions associated with perinatal depression treatment use. Archives of Women’s Mental Health 12, 447–450. Pawar, G., Wetzker, C., Gjerdingen, D., 2011. Prevalence of depressive symptoms in the immediate postpartum period. Journal of American Board and Family Medicine 24, 258–261. Reck, C., Klier, C.M., Pabst, K., et al., 2006. The German version of the postpartum bonding instrument: psychometric properties and association with postpartum depression. Archives of Women’s Mental Health 9, 265–271. Streiner, D.L., Norman, G.R., 1995. Health Measurement Scales. A Practical Guide to their Development and Use, 2nd edn. Oxford Medical Publications, New York.

786


Weinman, J., Petrie, K.J., Moss-Morris, R., et al., 1996. The Illness Perception Questionnaire: a new method for assessing illness perceptions. Psychological Health 1, 431–446. Whiffen, V.E., Gotlib, I.H., Comparison of postpartum and nonpostpartum depression, 1993. Clinical presentation, psychiatric history, and psychosocial functioning. Journal of Consultative Clinical Psychology 61, 485–494. Wisner, K.L., Peindl, S., Gigliotti, T., et al., 1999. Obsessions and compulsions in women with postpartum depression. Journal of Clinical Psychiatry 60, 176–180.

Wittkowski, A., Wieck, A., Mann, S., 2007. An evaluation of two bonding questionnaires: a comparison of the Mother-to-infant Bonding Scale with the Postpartum Bonding Questionnaire in a sample of primiparous mothers. Archives of Women’s Mental Health 10, 171–175. Wittkowski, A., Williams, J., Wieck, A., 2010. An examination of the psychometric properties and factor structure of the Post-partum Bonding Questionnaire in a clinical inpatient sample. British Journal of Clinical Psychology 49, 163–172.