The impact of supportive counselling on women's psychological wellbeing after miscarriage – a randomised controlled trial

Authors

  • GWS Kong,

    Corresponding author
    1. Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong, China
    • Correspondence: Dr GWS Kong, Department of Obstetrics & Gynecology, The Chinese University of Hong Kong, Prince of Wales Hospital, New Territories, Hong Kong, China. Email: gracekong@cuhk.edu.hk

    Search for more papers by this author
  • TKH Chung,

    1. Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong, China
    Search for more papers by this author
  • IH Lok

    1. Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong, China
    Search for more papers by this author

Abstract

Objective

To assess the effectiveness of supportive counselling after miscarriage.

Design

Randomised controlled trial.

Setting

University hospital.

Sample

Two hundred and eighty women with miscarriage.

Method

Women were randomised to receive supportive counselling from a nurse (at diagnosis and 2 weeks later) or routine care. Psychological wellbeing was measured with the General Health Questionnaire (GHQ-12) and Beck Depression Inventory (BDI).

Main outcome measures

Primary outcome measured the proportion of women suffering psychological distress (GHQ-12 score ≥4) at 3 months after miscarriage. Secondary outcomes were GHQ-12 and BDI scores at 6 weeks, 3 and 6 months.

Results

There was no difference in the proportion of women suffering psychological distress at 3 months after miscarriage (17.1% in counselling group versus 24.4% in control group; 95% CI –0.034 to 0.177; P = 0.19). However, for the subgroup of women (n = 152) with high baseline GHQ-12 scores, the median GHQ-12 score in the counselling group was significantly lower than the control group at 6 weeks (median score 3 versus 4.5 in counselling and control groups; P = 0.04) and 3 months (median score 1 versus 2.5 in counselling and control groups; P = 0.03). Similarly, for women with high baseline BDI scores (BDI > 12), the proportion for women continuing to score high was significantly lower in the counselling group 6 weeks after miscarriage (33.3 versus 61.1% in counselling group and control group; P = 0.03).

Conclusions

Although the results of current study do not justify routine counselling of all women following miscarriage, a supportive counselling programme for selected women with high levels of psychological distress is promising and merits further investigation.

Introduction

Although emotional distress following miscarriage can be significant and enduring, women are not routinely provided with professional psychological support or follow-up care. Previous studies have suggested that psychological distress following miscarriage is often unrecognised by medical professionals attending the women.[1, 2] A report found that only one-quarter of women received follow-up care after miscarriage, despite such care being desired by more than 90% of these patients.[3] Surveys of patient satisfaction following miscarriage indicate that a high percentage of the patients are angry about and dissatisfied with the emotional support that they had receive.[4-6]

Anecdotal studies indicate that when follow-up care is offered, it may have a positive effect on emotional adjustment after pregnancy loss.[7, 8] Only six randomised studies (including a pilot randomised controlled trial involving 19 patients) have assessed psychological intervention in miscarriage.[9-15] The sample sizes were small and discordant results were reported. The most updated Cochrane database of systemic reviews on psychological wellbeing post-miscarriage, reported that the lack of good quality adequately-powered studies means that the effectiveness of psychological support such as counselling cannot be determined.[16] To date, there is a substantial lack of intervention studies in this area and the effectiveness of psychological intervention for women suffering pregnancy loss has not been evaluated properly. The purpose of this randomised controlled trial was to assess the effectiveness of a supportive counselling programme compared with the ‘standard’ care for psychological wellbeing following miscarriage.

Methods

Study design and participants

This randomised controlled study was performed in a university-affiliated, tertiary referral hospital in Hong Kong caring for over 6000 deliveries a year. Over a 1-year recruitment period (February 2004 to January 2005), women who had been admitted with a diagnosis of miscarriage were approached to participate in this study. Women were excluded if (i) they were unwilling to participate, (ii) they had a current psychiatric disorder requiring active treatment, (iii) pregnancies were diagnosed as having congenital fetal abnormalities, (iv) the women were not Chinese, (v) they were visitors to Hong Kong (e.g. tourists for whom arranging follow up would be difficult). This study was approved by the institutional ethics committee.

Miscarriage was defined as a pregnancy loss occurring spontaneously before 24 weeks of gestation and was diagnosed based upon both ultrasound and clinical findings. All women were managed in the hospital according to a previously described protocol,[17] which has been the standard treatment for over 10 years in the center. All subjects were recruited shortly after receiving treatment (medical evacuation with Misoprosol or surgical evacuation of product of conception or expectant management if indicated) for miscarriage and before they were discharged home. The study protocol, randomisation and the follow-up plan were explained to the women by one of the two research nurses. The socio-demographics and clinical characteristics of women with the diagnosis of miscarriage were collected by means of a semi-structured interview that included age, education level, employment status, parity, and current and past obstetric history. These characteristics were assessed as potential risk factors for psychological distress. The participants were asked to complete three psychometric questionnaires at baseline, including the 12-item General Health Questionnaire (GHQ-12), Beck Depression Inventory (BDI) and Dyadic Adjustment Scale (DAS).

Instruments

The GHQ-12 is a measure of current mental health. It is a self-administered questionnaire for identifying individuals with general psychological distress.[18] The scale asks whether the respondent has experienced a particular symptom or behaviour recently. The GHQ is well validated with a high sensitivity and specificity. A validated 12-item Chinese version was used in this study.[19] The BDI is a self-reported inventory with 21 multiple-choice questions.[20] It is one of the most widely used instruments for measuring the severity of depression. A validated Chinese version was used for this study.[21] Both GHQ-12 and BDI used both continuous and categorical variables. The categorical approaches served to define cases with psychological distress, with the dichotomised cutoff scores of GHQ-12 and BDI defined in previous miscarriage studies.[22, 23] A cutoff score ≥4 in GHQ-12 is used to define women with definitive psychological distress,[22] whereas a BDI score ≥12 is used to define women with probable depressive disorder.[23] The DAS scale is a 32-item, self-reported measure of the marital relationship adjustment.[24] The total score is obtained by adding all four subscales ranges from zero to 151, with higher scores indicating more positive dyadic adjustment. A score of 100 or less on the DAS suggests possible marital discord.[24] The validated Chinese version was used.[25]

Randomisation and blinding

Women who agreed to participate in the study gave written consent and were then randomised by a research nurse. Randomisation was carried out according to the computer-generated random number using block sizes between two and ten to ensure approximately equal group sizes. The results of group allocation from randomisation were sealed in a set of opaque and sequentially numbered envelopes, and released by a research nurse after randomisation. Randomisation was into one of the following two groups:

Group 1 (Counselling group)

Women in this group received supportive counselling from a single nurse counsellor after completion of baseline questionnaires in the hospital before discharge. They were then followed up by the same nurse counsellor 2 weeks later by telephone for a further counselling session.

Group 2 (Control group)

Women in this group were managed according to our routine clinical practice and attended by the clinical staff as usual. No specific counselling or follow-up care was arranged.

Supportive counselling

As there is no standardised counselling method for miscarriage, we developed a programme that was easy-to-use by a nurse counsellor, problem-focused, likely to be acceptable to the women, and likely to be replicable. The basic elements used in the counselling programme included listening, explaining and giving information on miscarriage, guidance and advice on further pregnancy, and encouraging hope.

The individual counselling programme comprised only two sessions, for brevity and potential acceptability to the women. The first session usually lasted about 60 minutes and was conducted while the woman was still in hospital. The second session usually lasted about 30 minutes, was conducted 2 weeks after miscarriage and aimed at reinforcing the previous information given; allowing the women to discuss their feelings, worries and physical concerns; discovering possible underlying stress factors. We adopted telephone counselling format for the second session mainly for the convenience of the women. The detail of the counselling intervention is shown in Figure S1. The nurse counsellor had received prior training in supportive counselling techniques, grief counselling, interpersonal psychotherapy, and identification of common psychiatric disorders.

GHQ-12 and BDI were also sent to women 6 weeks, 3 and 6 months after miscarriage including self-addressed freepost envelopes. Those who failed to complete or return the questionnaires were contacted by the research nurse via telephone or by post (when telephone contact failed) to ensure maximal compliance. The research nurse who was assigned to the questionnaire data collection was blinded to the randomisation allocation.

Outcome measures

The effect of intervention (supportive counselling) was assessed in the following two ways. The primary outcome was to assess the proportion of women suffering psychological distress (defined as GHQ-12 score ≥4) at 3 months after miscarriage in both intervention and control groups. Secondary outcomes were GHQ-12 and BDI scores in both groups at 6 weeks, 3 and 6 months after miscarriage. The proportion of women with a probable depressive disorder, i.e. a BDI score ≥12, was also assessed. Women's GHQ-12 and BDI scores in both groups were assessed at 6 weeks and 6 months after miscarriage to observe any short-term or more enduring effects of the counselling programme. Also assessed was the interaction between time since miscarriage on the psychometric outcomes in each group. A subgroup analysis was also conducted on women who had high initial scores on the GHQ-12 and BD1 measures, and also of women with and without a prior history of psychiatric illness.

Statistical analysis

Student's t-test and the Mann–Whitney U-test were used to compare continuous variables. The Chi-squared or Fisher exact test was used to compare categorical data where appropriate. The trend of psychological well-being over time was assessed using repeated-measures analysis of variance (rANOVA), with time-since-miscarriage entered as the within-subject variable. Bonferroni correction was used for adjustment for multiple comparisons. Both complete case analysis and intention-to-treat analyses were performed in this study.[26] Complete case analysis involved only those women who returned the questionnaires. Intention-to-treat analysis involved all women who returned the questionnaires and who failed to do, using the last observation carried forward (LOCF) method, whereby the last available measurement for each individual at the time point prior to withdrawal from the study was retained in the analysis, as described in the study performed in a similar area.[11] Statistical analysis was performed using the Statistical Package for Social Sciences version 17.0; a P-value of <0.05 was defined as statistical significance.

Sample size calculation

A previous longitudinal study[17] suggested that 25% of women scored ≥4 on the GHQ-12 at 3 months after miscarriage (denoting psychological distress) and that this was reduced to 10% of women by 1 year. The sample size calculation of the current randomised controlled trial was based on the clinical assumption that the intervention (i.e. counselling) would shorten the psychological recovery and reduce the proportion of women with psychological distress (GHQ-12 ≥ 4) at 3 months from 25% down to 10%, a proportion equivalent to that at 1 year after miscarriage. One hundred women per arm (200 women in total) are required to give a power (β) of 80% with a type I error (α) of 0.05. With an assumed a loss to follow-up rate of 25%, it was calculated that 134 women would be needed to be recruited to each arm.

Results

Characteristics of women who refused and those who agreed to participate

A total of 368 women fulfilling the recruitment criteria were invited to join the study (Figure 1). Of these, 88 (23.9%) declined but all agreed to have their background socio-demographic and clinical data used for comparison. The women who refused to participate were older, had been married for longer, fewer had received tertiary or higher education, more were multiparous and more had reported the current pregnancy being an unplanned one. There were no statistically significant differences in gestational age at diagnosis, history of previous pregnancy losses or proportion receiving medical versus conservative management of miscarriage, between women who refused and those who agreed to participate (Table S1).

Figure 1.

Recruitment and data analysis flow chart.

Characteristics of participants

Of the 280 cases recruited, 12 refused to continue the study after being randomised and refused to complete the baseline psychometric scales, and were therefore excluded for further analysis (counselling group = 8, controlled group = 4) (Figure 1). There were no differences in the socio-demographic or clinical data between these 12 refusals and the remaining 268 participants. There were no differences between those who were randomised to the counselling or control group in terms of their socio-demographic and clinical characteristics, their baseline DAS scores or psychometric scores of GHQ-12 and BDI (Table S2).

All 132 women randomised to the counselling group received the first psychological counselling sessions. Ten women were unable to be contacted 2 weeks after miscarriage and thus the second session of counselling was not conducted. However, all of them had returned the psychometric questionnaires and were therefore included in the data analysis. Figure 1 illustrates the subject progress through the different phases of the randomised trial with the number of women who failed to return the questionnaires (Loss) at various stages stated. Psychometric outcome measures are presented in Table S2.

Psychological outcome measures at 6 weeks

Seventy-three women (27.2%) failed to return the psychometric questionnaires 6 weeks after miscarriage (counselling group = 36, control group = 37), leaving 96 women in the counselling group and 99 women in the control group for data analysis. Women in the attrition group had higher baseline median (interquartile range) GHQ-12 scores than those who returned the questionnaires: 6 (3–8) versus 4 (2–7), P = 0.03. There were no differences in other socio-demographic variables, clinical characteristics or baseline BDI or DAS scores. No differences were found in either the median scores or proportions of women with a high score (GHQ-12 ≥ 4 or BDI ≥ 12) between the counselling and control groups at 6 weeks after miscarriage in either the complete case analysis or the intention-to-treat analysis (Table 1).

Table 1. Post-intervention psychometric measures of patients randomised to counselling and control group
 Complete case analysisIntention to treat
Counselling groupControl groupCounselling groupControlgroup
  1. BDI, Beck Depression Inventory; GHQ-12, 12-item General Health Questionnaire; IQR, interquartile range; = number; SD, standard deviation.

  2. Data are presented in mean (SD), number of median (IQR) where appropriate.

  3. Mann–Whitney U-tests were used for comparing median values.

  4. Chi-squared test were used for comparing proportion values.

  5. All P-values in this table were non-significant.

GHQ-12
Post-intervention 6 weeksn = 96n = 99n = 132n = 136
Median (IQR)2 (0–4)2 (0–6)3 (0–6)3 (0–7)
Score ≥4, n (%)32 (33.3)37 (37.4)56 (33.3)60 (44.1)
Post-intervention 3 monthsn = 105n = 115n = 132n = 136
Median (IQR)0 (0–2.5)1 (0–3)1 (0–3)1 (0–4.75)
Score ≥4, n (%)18 (17.1)28 (24.4)32 (24.2)42 (30.9)
Post-intervention 6 monthsn = 104n = 110n = 132n = 136
Median (IQR)0 (0–2)1 (0–1)0 (0–3)1 (0–3)
Score ≥4, n (%)16 (15.3)15 (13.6)30 (22.7)27 (19.9)
BDI
Post-intervention 6 weeksn = 96n = 99n = 132n = 136
Median (IQR)3 (1–10)6 (2–12)4 (2–12)7 (2–13)
Score ≥4, n (%)18 (18.7)26 (26.3)33 (25)41 (30.1)
Post-intervention 3 monthsn = 105n = 115n = 132n = 136
Median (IQR)2 (0–6)3 (0–8)3 (0–7)4 (1–10)
Score ≥4, n (%)14 (13.3)17 (14.9)24 (18.2)27 (19.9)
Post-intervention 6 monthsn = 104n = 110n = 132n = 136
Median (IQR)1 (1–5)1 (0–5)2 (0–7)2 (0–8.75)
Score ≥4, n (%)8 (7.7)12 (10.9)20 (15.2)26 (16.9)

Psychological outcome measures at 3 months

Forty-eight women (17.9%) failed to return the psychometric questionnaires at 3 months (counselling group = 27, control group = 37), leaving 105 women in the counselling group and 115 women in the control group for data analysis. There were no statistical differences in socio-demographic variables, clinical characteristics or baseline psychometric characteristics between those lost to follow up and those who returned the questionnaires. One quarter of women scored above the high score cut-off (≥4) in GHQ-12 in the control group. Although the percentage of women scoring ≥4 in GHQ-12 was 30% less in the counselling group than in the control group (17.1 versus 24.4%, P = 0.19), the difference did not reach a statistical significance. No differences were found in the BDI scores between the counselling and the control groups (Table 1).

Psychological outcome measures at 6 months

Fifty-four women (20.1%) failed to returned the psychometric questionnaires 6 months after miscarriage (counselling group = 28, control group = 26), leaving 104 women in the counselling group and 110 women in the control group for data analysis. When compared with those who returned the questionnaires, women lost to follow-up had a significantly higher baseline median (IQR) GHQ-12 [6 (3–9) versus 4 (2–7), P = 0.002] and BDI [12 (4–19.5) versus 8 (4–13), P = 0.03] scores. There were no differences in their age or gestation at diagnosis. No differences were found in the psychometric scores between the counselling group and the control group (Table 1).

Subgroup analysis on women with high baseline GHQ-12 scores (≥4)

With the complete case analysis, the median scores of GHQ-12 for women with high baseline scores (n = 152) were significantly lower for women in the counselling group than those in the control group at both 6 weeks (three in counselling group versus 4.5 in control group; P = 0.04) and at 3 months (one in counselling group versus 2.5 in control group; P = 0.03) after miscarriage. Particularly for women with high baseline GHQ-12 scores who did not have a history of psychiatric illness (n = 145), the effect of counselling was particularly significant (three in counselling group versus five in control group; P = 0.01 at 6 weeks; and one in counselling group versus 2 in control group; P = 0.009 at 3 months) (Table 2). The proportion of women continuing to score ≥4 in GHQ-12 at 6 weeks and 3 months was also lower in the counselling group (Table 2), although this difference did not reach statistical significance. Intention-to-treat analysis also yielded similar results with significant differences found in the median scores of GHQ-12 between groups at 6 weeks and 3 months (Table 2). Six months after miscarriage, there were no differences in either the median scores or the proportion of high scorers in GHQ-12 between the counselling or control group. No differences were found between groups for the initial ‘low’ scorers (i.e. GHQ-12 < 4) at any time point.

Table 2. Subgroup analysis on GHQ-12 scores of high baseline GHQ-12 scorers
 Complete case analysisIntention to treat
Counselling groupControl groupP valueCounselling groupControl groupP value
  1. GHQ-12, 12-item General Health Questionnaire; IQR, interquartile range; = number; SD, standard deviation.

  2. Data are presented in mean (SD) or number of median (IQR) where appropriate.

  3. Mann–Whitney U-tests were used for comparing median values.

  4. Chi-squared test were used for comparing proportion values.

  5. All P-values in this table were non-significant.

  6. P < 0.05 indicates significant differences between counselling group and control group. P = NS, not statistically significant.

Overall initial high GHQ-12 scorers
Post-intervention 6 weeksn = 55n = 50 n = 79n = 73 
Median (IQR)3 (0–6)4.5 (2–7)0.044 (1–7)4 (1–7)0.02
Scores ≥4, n (%)25 (45.5)30 (60)NS49 (62.0)53 (72.6)NS
Post-intervention 3 monthsn = 61n = 62 n = 79n = 73 
Median (IQR)1 (0–3.5)2.5 (1–5)0.032 (0–5)1 (0–5)NS
Scores ≥4, n (%)15 (24.5)24 (38.7)NS29 (36.7)35 (47.3)NS
Post-intervention 6 monthsn = 58n = 57 n = 79n = 73 
Median (IQR)0 (0–3)0 (0–2.5)NS1 (0–5)1 (0–5)NS
Scores ≥4, n (%)13 (22.4)13 (22.8)NS27 (34.2)24 (32.4)NS
Initial high GHQ-12 scores without history of psychiatric illness
Post-intervention 6 weeksn = 51n = 49 n = 74n = 71 
Median (IQR)3 (0–5)5 (2–7)0.014 (1–4)4 (1–7)0.01
Scores ≥4, n (%)22 (43.1)30 (61.2)NS45 (60.8)52 (73.3)NS
Post-intervention 3 monthsn = 56n = 61 n = 74n = 71 
Median (IQR)1 (0–3)2 (1–5)0.0091 (0–5)1 (0–5)0.03
Scores ≥4, n (%)12 (21.4)23 (37.7)NS26 (35.1)33 (46.5)NS
Post-intervention 6 monthsn = 53n = 55 n = 74n = 71 
Median (IQR)0 (0–2.5)0 (0–2)NS1 (0–5)1 (0–5)NS
Scores ≥4, n (%)11 (20.8)12 (21.8)NS25 (33.8)23 (32.4)NS

Subgroup analysis on women with high baseline BDI scores (>12)

With the complete case analysis, for the initial high scorers (BDI ≥ 12) 6 weeks after miscarriage, the proportion of women continuing to score high in the counselling group was half of that in the control group and the differences were statistically significant (P = 0.03 and P = 0.01 for women without a history of psychiatric illness) (Table 3). No differences in the BDI scores were found 3 and 6 months after miscarriage. No differences were found between groups for the initial ‘low’ scorers (i.e. BDI < 12) at any time points.

Table 3. Subgroup analysis on BDI scores of high baseline BDI scorers
 Complete case analysisIntention to treat
Counselling groupControl groupP-valueCounselling groupControl groupP-value
  1. BDI, Beck Depression Inventory; IQR, interquartile range; = number; SD, standard deviation.

  2. Data are presented in mean (SD) or number of median (IQR) where appropriate.

  3. Mann–Whitney U-tests were used for comparing median values.

  4. Chi-squared test were used for comparing proportion values.

  5. All P-values in this table were non-significant.

  6. P < 0.05 indicates significant differences between counselling group and control group. P = NS, not statistically significant.

Overall initial high BDI scorers
Post-intervention 6 weeksn = 30n = 36 n = 45n = 51 
Median (IQR)9.5 (2.75–23)13 (7.25–21)NS13 (5–24)15 (10–21)NS
Scores ≥12, n (%)10 (33.3)22 (61.1)0.0325 (55.5)37 (72.5)NS
Post-intervention 3 monthsn = 33n = 43 n = 45n = 51 
Median (IQR)4 (1–13)8 (3–15)NS9 (1–18)10 (4–18)NS
Scores ≥12, n (%)9 (27.3)14 (32.6)NS19 (42.2)22 (43.1)NS
Post-intervention 6 monthsn = 30n = 39 n = 45n = 51 
Median (IQR)1.5 (0–11.25)3 (0–14)NS7 (0–16.5)8 (0–21)NS
Scores ≥12, n (%)7 (23.3)11 (21.6)NS18 (40.0)20 (39.2)NS
Initial high BDI scores without history of psychiatric illness
Post-intervention 6 weeksn = 28n = 35 n = 42n = 49 
Median (IQR)9 (2.25–19.0)13 (7–12)NS12 (4.75–23)15 (9.5–21)NS
Scores ≥12, n (%)8 (28.6)21 (60)0.0122 (52.4)35 (71.4)NS
Post-intervention 3 monthsn = 30n = 42 n = 42n = 49 
Median (IQR)3 (0.75–11.25)7.5 (2.75–15)NS8.5 (1–16.25)8 (4–17.5%)NS
Scores ≥12, n (%)7 (23.3)13 (31)NS17 (40.5)20 (40.1)NS
Post-intervention 6 monthsn = 27n = 37 n = 42n = 49 
Median (IQR)1 (0–8)3 (0–13.5)NS4.5 (0–16)8 (1–19.5)NS
Scores ≥12, n (%)5 (18.5)10 (27)NS16 (38.1)19 (38.8)NS

Interaction between time and counselling

Time since miscarriage alone had a significant effect on the reduction of both the GHQ-12 and BDI scores from baseline through to 6 months after miscarriage (< 0.001) (Table S3). Although the mean psychometric scores appeared to be lower in general across time in the counselling group, counselling did not exert an additional effect to time in terms of reduction of the psychometric scores (Figure 2).

Figure 2.

Repeated measures of GHQ-12 and BDI scores across time between counselling and control group.

Discussion

Main findings

A session of supportive counselling with a trained nurse counsellor, delivered immediately and at 2 weeks after diagnosis for miscarriage, did not show a statistically significant effect in reducing psychological distress of women after miscarriage. It also failed to show any additional effect to time on the reduction of the psychometric scores along the longitudinal course. Nevertheless, a 30% reduction in the proportion of women with high GHQ-12 scores (indicative of definitive psychological distress) was evident by 3 months post miscarriage in the counselling compared with the standard care group, suggesting a potential clinical beneficial effect, albeit not a statistically significant one. However, among the subset of women who had high baseline scores on the GHQ-12 and BDI questionnaires, there was a statistically significant difference observed between counselling and standard care groups, in terms of lower scores and reduced proportions of women scoring highly at 6 weeks in the counselling group. This suggests that a ‘selective’ counselling programme aimed at women with high baseline levels of psychological distress might be beneficial for improving emotional wellbeing in this group in the first weeks after miscarriage.

Strengths and limitations

The current programme was developed for its brevity in terms of the number of sessions, session length and simplicity in application. Modification in the counselling programme may help to improve its therapeutic effect. More thorough and comprehensive counselling sessions may be necessary for better development of rapport between woman and counsellor and to personalise carings suggested by Swanson[10] who indicated that several hour-long counselling sessions may be required for each woman. However, the intensiveness and comprehensiveness of the counselling sessions should be balanced with woman's level of accceptance.

The current study is limited by the high refusal rate, with almost one quarter of women declining to participate. Women who refused to participate in the study tended to be older, to have lower education levels, be multiparous and have had an unplanned pregnancy. The reason for their refusal was unclear and is only speculative. It has been reported previously that women who are most distressed at the diagnosis of miscarriage, are less likely to participate in an intervention study.[27] We cannot infer this from our current study, as women who refused to participate did not complete the psychological assessments.

A further drawback of our study is that 18–27% of subjects failed to return a questionnaire at various time points. A loss to follow-up rate ranging from 24 to 47% has been reported in other counselling trials involving bereaved individuals after miscarriage or perinatal deaths[10, 11, 28, 29]. As the attrition rate has been taken into account during the sample size calculation in the current study, the statistical power in detecting differences between groups was ensured. However, women who failed to return the questionnaires at 6 weeks and 6 months appeared to have had a higher baseline level of psychological distress after miscarriage. The underlying reasons for failure to return the questionnaires are not known. It is possible that women were not motivated to complete the questionnaires, but we cannot exclude the possibility that women were too distressed to complete them. Nonetheless, the proportion of women failing to complete the study was not significantly different between the counselling and control groups.

Interpretation

The current study utilized a preventive framework with the assumption that it was appropriate to provide routine counselling for all women experiencing miscarriage. This was based on previous research suggesting that follow-up would be desired by more than 90% of women having a miscarriage.[3] Therefore, the counselling programme was initially designed to be delivered in a ‘routine’ manner rather than for a ‘selected’ group of women. Our previous longitudinal study[17] revealed that the psychometric scores reported immediately after miscarriage appeared to predict scores at subsequent follow up. This might suggest that women who are not so distressed initially might be less likely to benefit from the counselling provided them. The lack of beneficial effect in this group of women may have diluted the overall therapeutic effect of the counselling programme. It may thus be more appropriate to pre-select individuals at higher levels of distress to evaluate the effect of psychological intervention.[9] In the subgroup analysis of women who had had baseline scores of distress, psychological distress was found to be lower in the counselling group at 6 weeks and 3 months after miscarriage by either measuring the median scores or the proportions of high scorers in GHQ-12 (6 weeks and 3 months) or BDI (6 weeks), suggesting that counselling may be more effective in this ‘subgroup’ of women.

Lilford et al.[30] have questioned whether ‘routine’ counselling should be offered to all bereaved women with perinatal death or should only be restricted to those with perceived special need, and they failed to demonstrate a beneficial effect of ‘routine’ counselling. However, the results should be interpreted with caution as the study was limited by sample size (22 in treatment group and 35 in the study group) and high attrition rate (51.4% in the treatment group versus 36.3% in the control group). Neugebauer et al.[11] has found interpersonal counselling to be effective in reducing depressive symptoms in a selected group of miscarrying women classified with ‘subsyndromal depression’. Psychological counselling has been found to be beneficial in reducing depressive symptoms in women screening positive for postnatal depression.[28] Future studies aiming to identify a ‘selected’ group of women at risk of developing psychological distress after miscarriage and to study the effect of ‘selective’ counselling are warranted. Given that a previous study[31] also demonstrated that a significant proportion of male partners had emotional distress after miscarriage, the possible impact of psychological counselling on male partners after miscarriage should also be explored in the future.

Conclusions

Although the results of current study do not justify routine counselling of all women following miscarriage, a supportive counselling programme for selected women with high levels of psychological distress may be more promising and merits further investigation.

Disclosure of interests

None.

Contribution to authorship

GWSK contributed to ethics approval, data analysis, manuscript writing and submission, IHL contributed to protocol writing, ethics approval, recruitment, data collection, data analysis and manuscript writing; TKHC contributed to project supervision. GWSK, IHL and TKHC were all responsible for manuscript editing.

Details of ethics approval

The study was granted ethics approval by jointly The Chinese University of Hong Kong – New Territories East Cluster Clinical Research Ethics Committee (CRE-2000.233) in 2003. The trial is registered with Current Controlled Trials (ISRCTN64429365).

Funding

None.

Acknowledgements

The authors would like to thank all the women who participated in the study.

Ancillary