Postnatal quality of life assessment: validation of the Mother-Generated Index
*Dr A. Symon, School of Nursing and Midwifery, University of Dundee, Dundee DD1 9SY, UK.
Quality of life is multifactorial, but may not be adequately assessed using existing validated health measures. The Mother-Generated Index, a subjective tool to measure postnatal quality of life, was assessed in a study of 102 women in Tayside, Scotland. They specified the most important areas of their lives having had a baby, and scored these. The Index showed good correlation with established measures of physical and psychological wellbeing, and accurately tracked attitudes towards the baby and the partner. Face, criterion and construct validity were all demonstrated. As both content and scoring are subjectively determined, the Index encourages a holistic assessment.
The incidence of postnatal morbidity has been comprehensively described in recent years1. The focus on obvious morbidity such as anaemia, infections and haemorrhage has been widened to include other areas such as sexual function, backache, painful perineum and constipation, and screening for postnatal depression is well established2. However, there is currently no validated tool which measures the mother's overall quality of life.
Quality of life is complex and all-embracing, and may be affected by many factors, including physical, mental, emotional, social, sexual and spiritual parameters. Calman3 defines it as ‘the extent to which hopes and ambitions are matched by experience’, with the aim of medical care being to ‘narrow the gap between a patient's hopes and aspirations and what actually happens’. Many aspects may be beyond the clinician's control, and there is an understandable tendency to concentrate on those aspects which are most amenable to some form of treatment or intervention, with a corresponding focus on health-related quality of life measurement4. However, this approach may not reflect the ‘social reality of the situation … Measures of health status lack focus, and may obscure factors of prognostic importance to an individual patient’5. While mothers may attend a health centre in the first year after childbirth for a variety of reasons, including family planning and cervical cytology, routine screening is not automatic for many postnatal women beyond the standard six-week examination, and morbidity may go undetected. Mothers may be left with specific health problems, or broader quality of life issues that may in turn affect their general health. Due to these concerns, we sought to validate a postnatal quality of life assessment tool.
A pilot study established the Mother-Generated Index (MGI) as a workable tool6. Like other quality of life tools, it makes its assessment from the point of view of the recipient of care. This avoids a reliance on predefined lists of variables, which may not accord with an individual's values or sense of importance. Predefined lists may assess psychological wellbeing or the prevalence of physical morbidity, but may miss the relevance to that particular mother of issues such as family relationships or the need to return to work. An index whose content and scoring are generated by the person in question confers on him/her a degree of autonomy which we believe may help practitioners to see that person more holistically.
Following the pilot study, minor modifications were made to the layout of the MGI form to make it easier on the eye. These included using a visual analogue scale for the Step 2 scores (see below). Because of these changes, further assessment of the instrument was required in order to determine its validity in both the puerperium (at six to eight weeks) and later in the first year following childbirth (we have chosen six to eight months). Validity can be assessed in terms of face validity (a subjective judgement about whether the instrument appears to reflect the relevant areas [motherhood in our case]); criterion validity (how well the instrument correlates with established assessment tools); and construct validity (the instrument's conformity to theoretical perspectives).
Ethical approval was sought and obtained from the local Trust Research Ethics Committee. Mothers under the age of 16, and those whose baby had died or who was seriously ill were excluded. A week prior to the scheduled visit, the woman was sent an explanatory letter which outlined the study, and informed her of her right not to be included without any risk to the care offered to herself or her baby. The local Trust consent form for research was completed and signed by the mother immediately prior to completing the forms.
The MGI is a single-sheet three-step questionnaire, and was completed by the mother along with the validators with the health visitor present. In the MGI, the mother specifies a maximum of eight areas of her life which have been affected by having had a baby, and identifies these as either positive, negative or neither (Step 1). She then gives each area a score out of 10 according to how she has felt about this issue over the previous month (Step 2). The average of these scores gives an Index, which is a broad assessment of quality of life. In Step 3, she allocates 20 ‘spending points’ between the areas, with more points given to the areas she deems most important. The distribution of points informs the health care practitioner of the woman's perspective on these issues.
Validation was sought from concurrent use of the Edinburgh Postnatal Depression Scale2, Glazener et al.'s1 Postnatal Morbidity Index (PNMI) and Kumar et al.'s7 postnatal version of the Maternal Adjustment and Maternal Attitudes (MAMA) scale. The PNMI is comprised of three indices measuring maternal physical morbidity [PNMI (M)], baby physical morbidity [PNMI (B)] and an assessment of maternal reaction to the baby (‘Baby descriptor’). The postnatal MAMA scale is comprised of five subscales: ‘Body image’, ‘Somatic symptoms’, ‘Marital relationship’, ‘Attitudes to sex’ and ‘Attitudes to baby’.
We hypothesised that we would find at least moderate correlation between the MGI and these validators, and that a low MGI score would be associated with validator scores suggesting a lack of wellbeing.
Six health visitors based in three areas of Tayside, Scotland, administered the MGI and validators by structured interview in the woman's home or health centre at either six to eight weeks or six to eight months postpartum. Data were entered into EPI-Info, and were exported to Microsoft Excel for analysis. Analysis of variance between mothers with low MGI scores and all others was calculated using ANOVA. SPSS was used to measure the statistical significance of the correlations between MGI and validator scores.
It was estimated (based on the previous pilot work) that, in order to assess the MGI effectively, between 50 and 60 mothers would need to be recruited at each stage of the study. A total of 102 mothers (51 at six to eight weeks and 51 at six to eight months) registered in five general practices in Tayside were recruited.
Neither the areas cited by the mothers nor their respective scores appeared to be significantly associated with variables such as age, parity or whether she lived alone or with one or more others. At six to eight months, mothers who were unemployed had significantly lower scores than those who were working (4.8 compared with 6.3; P = 0.012, F = 4.16, calculated by ANOVA). Other than that, socio-economic status (as measured by her postcode or occupation for herself and/or her partner) was not significantly associated with scores.
Data on type of delivery were unfortunately incomplete, although this (and the other variables noted here) were found in the earlier pilot results not to be significant6.
Eighty-five mothers made a positive comment about their baby or partner, or about being in a family, or about feelings of fulfilment. Many aspects of life were cited, reflecting issues such as physical health, emotional wellbeing, family relationships, personal time, social life and thoughts about returning to work; only six mothers cited a physical problem.
The MGI scores were significantly correlated with the validators at both six to eight weeks (Table 1) and six to eight months (Table 2). In addition, and at both stages, mothers with MGI scores of 5 or less had a significantly higher incidence of physical problems, significantly higher EPDS scores and significantly lower MAMA scores. Mothers with MGI scores of 5 or less at six to eight months also had babies with a higher incidence of physical problems, and were less likely to view their babies in a positive light (Tables 1 and 2).
Table 1. Correlation between MGI and validator scores at six to eight weeks‡ and comparison of validator scores† for mothers with MGI scores of 5 or less and above 5.
|Correlation||−0.402**||0.006 N/S||0.121 N/S||−0.418**|| 0.473**|
|MGI score ≤5||6||2.25||3.8||9.4||159.4|
|MGI score >5||4.13||2.92||4.2||6.3||182.3|
|P||0.02207*|| N/S|| N/S||0.0288*|| 0.00033***|
|F||5.59|| || ||5.07||14.92|
Table 2. Correlation between MGI and validator scores at six to eight months‡ and comparison of validator scores† for mothers with MGI scores of 5 or less and above 5.
|MGI score ≤5||4.8||3.3||4.3||12.8||165.5|
|P||0.006411||0.01910||0.02910|| 0.0000312|| 0.007511|
Mothers who made positive comments about their baby had higher ‘Attitudes to baby’ MAMA subscale scores: 39.6 compared with 36.9 (P = 0.0018; F = 10.25). Mothers who made positive comments about their partner had significantly higher ‘Marital relationship’ MAMA subscale scores: 39.9 compared with 35.4 (P = 0.00153; F = 10.62). Those who made a positive comment about their lives (e.g. ‘feel fulfilled’) had higher MAMA total scores (180.8 compared with 172.8; P = 0.041; F = 3.94). Those mothers with EPDS scores of 12 or more had significantly lower MGI scores than those with EPDS scores of below 12 (4.8 compared with 6.4; P = 6.28 E−06; F = 22.7).
Having previously established that the MGI was a workable tool6, the aim of this study was to test its validity. We acknowledge that the study had limitations. With just 102 participants, it was relatively small scale, and we did not assess the degree of support experienced by the mothers, and so we cannot say how this may affect quality of life. The population in which this tool was assessed is not ethnically diverse, and the MGI does require the mother to be able to articulate her feelings, although she can ask the health visitor to write these down for her. We did not carry out test–retest, and cannot say how sensitive the MGI may be to mothers' changing perceptions of quality of life over time. However, we found similar degrees of correlation between the MGI and the validators in the first and second phases of the study, and so believe that it has reasonable reliability. Mothers also cited very similar areas in the two phases.
The MGI appears to be valid. Face validity was satisfactory: mothers completing the form tended to cite aspects of their life that were clearly related to being a mother (this was the one domain referred to on the form). We deliberately did not identify other domains such as tiredness or breastfeeding because we did not want to direct the mothers' thoughts into specific areas of her life. The subjective nature of the MGI is stressed by the mother herself determining what the cited items are. Comments about the baby included ‘Enjoy having a new baby’ and ‘Didn't think I could love as much’. A few negative comments were also made, such as ‘Lack of maternal feelings’, and ‘Slightly jealous of time husband has with daughter’. Positive comments about being in a family included ‘Stronger relationship with husband’ and ‘Feel closer to my mum’. Comments about feeling better in themselves included ‘More purpose to life’, ‘More confidence since having baby’ and ‘Realising how well I can cope’. The wide range of comments and the unique nature of each completed form reflect the view that ‘quality of life, rather than being a description of patients' health status, is a reflection of the way that patients perceive and react to their health status and to other, non-medical, aspects of their life’8.
Although we did not ask the mothers to provide additional comments, feedback from the health visitors indicated that reactions were generally very positive. Comments included ‘Mothers wanted to talk about themselves—the MGI helped them do this’; ‘Once the first statement went down, the rest flowed’; and ‘It enriches the relationship between health visitor and client’. As such, we believe that the MGI's face validity is assured, as it appears to assess the relevant areas of the mother's life; in addition, most mothers found it an acceptable and helpful tool.
Because no existing tool measures postnatal quality of life, there is no ‘gold standard’ against which to test the MGI's criterion validity. As noted earlier, quality of life is a complex and dynamic concept which covers many areas. The statistically significant correlation between the MGI and validator scores suggests that the MGI accurately tracks the physical and psychological health measured by the PNMI and EPDS, and reflects the attitudinal and adjustment notions of the MAMA score. There is a practical limit to the number of other standard tools we could realistically ask these mothers to complete, and so we did not assess social support, which may be influential in determining quality of life. Nevertheless, we believe that reasonable criterion validity was established. The MGI's additional virtue is that it is not restricted to the areas specified by these scales, instead allowing the mother to determine what is considered most (or least) important to her. This does require a certain level of understanding and articulation on the part of the mother, even with health visitor support.
Construct validity can be assessed with reference to how the MGI relates to standard assessments of severity (i.e. can we make accurate inferences about the people in question based on their scores?) There were a number of ways in which this was addressed. Making a positive comment about their baby in the MGI was significantly associated with a higher ‘Attitudes to baby’ subscale MAMA score; similarly, making a positive comment about their partner was significantly associated with higher ‘Marital relationship’ subscale MAMA scores. The mothers whose EPDS scores suggested increased risk of postnatal depression (i.e. a score of 12 or above) had significantly lower MGI scores, suggesting (as might be hypothesised) that women at risk of depression have low quality of life scores.
While mothers with low MGI scores did have a higher score relating to physical and psychological morbidity, there were so few spontaneous comments about physical morbidity, that it was not possible to detect any significant associations between those mothers who did cite a physical problem and either the PNMI (M) or ‘Somatic’ subscale MAMA scores.
Our ‘cutoff point’ in terms of MGI scores (a score of 5 or less) identified mothers whose validator scores were, with two exceptions, significantly different in statistical terms from mothers who scored above 5. This score gives a broad overall quality of life evaluation, although we did not try to verify this by asking the mother to give a separate global rating of her perceived quality of life. In addition to the overall assessment, the distribution of spending points in Step 3 of the tool helps the practitioner to identify which areas of the woman's life are most important to her. What the practitioner does with this information (and therefore how clinically relevant the MGI is) remains to be seen, but we suggest that this practitioner is in a position to assess the importance of the cited measures and the scores the mother assigns to these. We tentatively suggest that a score of 5 or less indicates a woman with low quality of life, and possibly serious issues which require attention. We plan to test the MGI against standard assessments such as the EPDS in a randomised controlled trial in order to assess its usefulness as a screening tool.
The Mother-Generated Index has been tested and has been found to be a valid tool in measuring postnatal women's quality of life. The diversity of mothers' experiences lends weight to the drive to individualise postnatal assessment. A subjective quality of life measurement has the benefit of allowing the woman in question to determine both its content and scoring.
The authors would like to thank Anna MacKay, Alice Smith, Sheila Knight and Janice Gray, health visitors, for their assistance with data collection, and Dr Brian Williams of the Department of Epidemiology at the University of Dundee, for his comments on a late draft of this paper.