Article Text

PDF

Development of a multidimensional labour satisfaction questionnaire: dimensions, validity, and internal reliability
  1. L F P Smith, research partner1
  1. 1East Somerset Research Consortium, Westlake Surgery, High Street, West Coker, Somerset BA22 9AH, UK
  1. Dr L F P SmithESReC{at}globalnet.co.uk

Abstract

Background—No published quantitative instrument exists to measure maternal satisfaction with the quality of different models of labour care in the UK.

Methods—A quantitative psychometric multidimensional maternal satisfaction questionnaire, the Women's Views of Birth Labour Satisfaction Questionnaire (WOMBLSQ), was developed using principal components analysis with varimax rotation of successive versions. Internal reliability and content and construct validity were assessed.

Results—Of 300 women sent the first version (WOMBLSQ1), 120 (40%) replied; of 300 sent WOMBLSQ2, 188 (62.7%) replied; of 500 women sent WOMBLSQ3, 319 (63.8%) replied; and of 2400 women sent WOMBLSQ4, 1683 (70.1%) replied. The latter two versions consisted of 10 dimensions in addition to general satisfaction. These were (Cronbach's alpha): professional support in labour (0.91), expectations of labour (0.90), home assessment in early labour (0.90), holding the baby (0.87), support from husband/partner (0.83), pain relief in labour (0.83), pain relief immediately after labour (0.65), knowing labour carers (0.82), labour environment (0.80), and control in labour (0.62). There were moderate correlations (range 0.16–0.73) between individual dimensions and the general satisfaction scale (0.75). Scores on individual dimensions were significantly related to a range of clinical and demographic variables.

Conclusion—This multidimensional labour satisfaction instrument has good validity and internal reliability. It could be used to assess care in labour across different models of maternity care, or as a prelude to in depth exploration of specific areas of concern. Its external reliability and transferability to care outside the South West region needs further evaluation, particularly in terms of ethnicity and social class.

  • Women's Views of Birth Labour Satisfaction Questionnaire (WOMBLSQ)
  • labour
  • questionnaire

Statistics from Altmetric.com

Key messages

  • A quantitative multidimensional maternal labour satisfaction questionnaire has been developed which comprises 10 dimensions in addition to general satisfaction.

  • It possesses good internal reliability and validity, but its generalisability and external reliability need further testing.

  • It could be used to compare models or systems of labour care as a stand alone instrument, or as a screening test prior to in depth qualitative work.

What this paper adds to the subject

To date there has been no published multidimensional satisfaction instrument to assess labour care. This instrument could be used in addition to standard measures of labour care, such as perinatal mortality and maternal morbidity, to give a more rounded overall picture of the care which women receive in labour in the UK.

Maternal and perinatal mortality are the traditional national and international measures of the quality of labour care. Fortunately, maternal deaths have become extremely rare and perinatal mortality and morbidity such as cerebral palsy are insensitive indicators of various components of pregnancy care.1 For a full picture of the quality of labour care one needs a patient centred measure such as women's satisfaction.

Measurement of patient satisfaction is not easy.25 To compare between women, institutions, or aspects of labour care, a quantitative measure is needed.6 Medical and nursing psychometric satisfaction measures exist,711 but only one has been published for (antenatal) maternity care.12 When designing a satisfaction questionnaire (instrument) one must consider a range of potential dimensions361317 including continuity of care, availability of carers, access, interpersonal skills, and technical competence.2781822 “Home made” satisfaction questionnaires tend to overestimate satisfaction, as do those which ask questions about satisfaction in general terms.1323

Care in childbirth continues to evolve—for example, following publication of the “changing childbirth” report24—and there are still many organisational models for labour.2526 Because these models emphasise various aspects of care—such as greater continuity, greater accessibility, and greater shared care—or provide care at different sites and/or from different professional groups, there is a need to have a valid reliable multidimensional questionnaire to assess the quality of labour care from a woman's perspective. Such assessment could allow future models of care to be more focused on women's specific needs, not just on the needs of the professionals or the “maternity care system”. This paper describes the development of such a questionnaire, the Women's Views of Birth Labour Satisfaction Questionnaire (WOMBLSQ).

Method

FACE AND CONTENT RELIABILITY

There were three development versions of the WOMBLSQ proper, resulting in a fourth or final version (WOMBLSQ4). Questions were selected from five sources to ensure content and face validity. These were fieldwork, a published survey manual,27 the North American PSQ-III satisfaction questionnaire,28 specially written questionnaires from an initial literature review, and further new questions were added after version 2 was tested (see below). The original fieldwork (reported elsewhere)2930 enabled us to discard or to rewrite questions which were frequently skewed or not answered. WOMBLSQ2 contained open questions which were analysed to ensure that no important areas were being excluded from the questionnaire. If this was so, the missing content areas were added to WOMBLSQ3.

Many questions were deliberately very positively or negatively worded to enhance the respondents' ability to express minimal dissatisfaction.13 All “questions” were statements which required respondents to mark a seven point Likert scale from “totally disagree” to “totally agree” to enhance the sensitivity of subsequent dimensions.61316

The questionnaire was developed over 4 years in several NHS trusts and general practices in the old South West region of England. WOMBLSQ1 was tested in three trusts, WOMBLSQ2 and 3 in nine trusts, and WOMBLSQ4 in three different trusts. Midwives were asked to give questionnaires to consecutive women postnatally within 10 days of birth. This included women who had delivered at home, in community units, or large hospitals. Except for WOMBLSQ4 which was posted 6–12 weeks postnatally to consecutive women on a list of births, no reminders were used. WOMBLSQ3 and 4 had additional sections on demography and pregnancy details to judge the generalisability of results and to allow testing of construct validity.

INTERNAL RELIABILITY

Questionnaires were analysed using the SPSS-PC statistical package. Repeated stepwise principal components analysis (PCA) with varimax rotation was used to produce factors (dimensions),3132 each of which consisted of one or more questions which were then re-read as a group intuitively to label the dimensions. The aim was to evolve groups of questions which each addressed a specific dimension. This procedure has been used in a similar situation to produce an antenatal satisfaction questionnaire12 and other patient satisfaction questionnaires.78

The remaining questions were then checked to ensure content validity. New questions were added where either important topic areas were missing following the statistical analysis (reduced content validity) or where the internal reliability of a dimension was low and the next version evolved.

SCALE GENERATION

Scale scores were generated to allow easily comprehensible comparisons between individual dimensions. This is important if the questionnaire is to be used to assess the relative strengths and weaknesses of various aspects of labour care. To produce a scale score for each of the dimensions identified, constituent scale questions were added (with negatively worded questions being reversed). This total was then transformed so that the minimum possible score was always 0 (total dissatisfaction with that dimension) and the maximum possible score 100% (total satisfaction with that dimension).

CONSTRUCT VALIDITY

Construct validity should be assessed by examining the compatibility of dimensions with primary research evidence about how different groups of women should score. In addition, individual dimensions were tested against the transformed general satisfaction dimension. There should be moderate correlation between dimensions which are related to, but distinct from, “satisfaction” as a global concept.89

The transformed individual dimensions were tested against (1) depression (on the Edinburgh Postnatal Depression scale of 0–20 with a higher score indicating more severe depression) and age (years) (Pearson correlation coefficients calculated); (2) duration of breast feeding (recorded on an ordinal scale of: never attempted, tried only a few times, breast fed but now discontinued, still breast feeding but mixed; solely breast feeding), educational level (primary or secondary school, sixth form/college, professional or technical qualification, diploma, degree course) and social class (unskilled manual, partly skilled manual, skilled manual, skilled non-manual, professional/managerial) (Spearman correlation coefficients calculated); and (3) place of delivery (home, small community unit, large hospital), parity (multiparous or primiparous), induction of labour (yes/no), mode of delivery (vaginal delivery, forceps/Ventouse delivery, caesarean section), and labour complications (yes/no) (one way analysis of variance followed by Student-Newmann-Keuls (SNK) post hoc tests if appropriate). Because of the large number of comparisons being analysed the p value was set at 1%.

The study was approved by the relevant local research ethics committees.

Results

DEVELOPMENT

The details of the four versions of WOMBLSQ are given in table 1. Analysis of open questions in WOMBLSQ2 suggested four missing content areas: relevant questions were added to WOMBLSQ3. These were support from one's husband or partner, being in control or not, being left alone, and immediate postnatal care; only two of these were retained after analysis of version 4.

RESPONDENTS

WOMBLSQ3 was sent out to 500 women in nine trusts of whom 319 (63.8%) replied. WOMBLSQ4 was sent to 2400 women in three trusts of whom 1683 (70.1%) replied. Respondents to both were comparable with respect to social class, educational level, and employment in early pregnancy (table 2). The clinical details of the respondents to both of these postings are as follows.

WOMBLSQ3

Of 376 respondents (mean (SD) age 29.1 (5.82) years, range 15–45), 346 (92.5%) were living with their husband or partner, 365 (98.1%) were of white ethnic origin, 189 (51.6%; 10 not known) were primiparous, 298 (79.7%; two not known) had a vaginal delivery, 32 (8.6%) had an instrumental delivery, 44 (11.8%) had caesarean section, 77 (20.7%; four not known) had their labour induced, 123 (33.1%) had labour complications, 251 (69.1%; 13 not known) planned to deliver in a large hospital, 90 (24.8%) in a community unit, and 22 (6.1%) at home; mean (SD) gestation at birth was 39.8 (1.55) weeks (range 33–43).

WOMBLSQ4

Of 1683 respondents, 1420 (86.4%) were screen negative on the Edinburgh Postnatal Depression scale with 40 women not answering the question; 317 (19%; 14 not known) never attempted breast feeding, 146 (8.7%) tried for a short period but quickly gave up, 534 (32.0%) breast fed for some weeks, 407 (24.4%) were still breast feeding in addition to other milk, 265 (15.9%) were still exclusively breast feeding when surveyed.

DIMENSIONS AND INTERNAL RELIABILITY

PCA of the WOMBLSQ3 respondents suggested 11 dimensions comprising 32 questions, including general satisfaction (two questions). These were (table 1): professional support, women's expectations, knowing carers, pain in labour, labour environment, home assessment in early labour, pain immediately after labour, holding the baby, being in control, and support from husband/partner. PCA of the WOMBLSQ4 respondents confirmed these dimensions. Each of the 30 questions highly loaded onto only one dimension (see appendix).

The overall reliability of the total scale excluding general satisfaction was good with Cronbach's alpha of 0.860 and 0.892, respectively, and values for individual dimensions of 0.55–0.90 and 0.62–0.91, respectively, for WOMBLSQ3 and WOMBLSQ4.

CONSTRUCT VALIDITY

The individual dimensions of WOMBLSQ4 were tested against the general satisfaction scale and found to be moderately correlated (table 3). Intercorrelations between dimensions were mostly acceptable.

Planned place of delivery was significantly associated with satisfaction with professional support, expectations of labour, pain in labour, knowing carer, and control subscales (table 4). Those women whose labour was induced were less likely to be satisfied with their expectations of labour subscale (F=16.38, df=1, 360; p<0.0001; 49.3% v 63.2% for those not induced). Primiparous women were less likely to be satisfied with their expectations of labour (F=33.8, df=1, 353; p<0.0001; 52.7% v 68.5%) and with knowing their carers in labour (F=6.81, df=1, 360; p<0.01; 38.5% v 46.9%). Type of delivery was also related to satisfaction with expectations of labour (F=95.96, df=2, 360; p<0.0001) and with pain after delivery (F=7.15, df=2, 368; p<0.001). Women having a vaginal delivery were more satisfied with their expectations of labour (68.0%) than those having instrumental delivery (32.3%) and those having a caesarean section (24.1%). Women having a vaginal delivery (48.9%) were also more satisfied with their pain relief immediately after delivery than those having a caesarean section (40.4%), but not than those having an instrumental delivery (44.1%). Maternal age and gestation at delivery were unrelated to any of the satisfaction subscales.

Respondents to WOMBLSQ4 had a range of postnatal depression scores. These were significantly negatively correlated with the satisfaction scale (p<0.001) and all subscales (p<0.001) except knowing carer (p>0.05) and control (p>0.05; table 5). Duration of breast feeding was positively associated with increasing satisfaction for subscales of professional support, holding her baby, pain in labour, pain after labour, and being in control; it was significantly negatively associated with knowing her carer. Higher educational achievement was negatively correlated with satisfaction on the expectation subscale, knowing carer, and environment subscale; it was significantly positively correlated on the control subscale. Higher social class was significantly positively correlated with the husband/partner support subscale; it was significantly negatively correlated with satisfaction with expectation subscale, knowing carer, and labour environment subscale.

Discussion

This study has produced a new valid reliable instrument with which to assess women's satisfaction with their labour care, and hence strengthen assessment of the quality of labour care. It is short and easily completed by women following delivery or by post. It could be used to evaluate different care models, ideally jointly with neonatal and maternal morbidity following labour to produce a satisfaction score (either a total overall score or one based on its various dimensions), or as a screening instrument to focus efficiently qualitative exploration in areas of low satisfaction from women's perspective. “One off “evaluation is limited as reference ranges have yet to be developed.

The WOMBLSQ has good content validity with questions developed from a review of the literature, interview fieldwork,2930 existing instruments,2728 and women's comments. It also has construct validity: its dimensions relate to a range of clinical and demographic variables which previous work has suggested alter women's perceptions of their labour care: knowing one's carer,14151821 place of delivery,2728 expectations of labour care,33 professional competence,71415 and pain during labour.34 There is no published evidence to support the pain following delivery and the husband support subscales, but these developed out of women's comments.

As one might expect, the instrument can discriminate between women giving birth in different settings (home, community units, large hospitals), having different types of birth, and having a range of problems including induction of labour and complications.182728 As expected, both social class30 and highest level of education13 attainment affected some of the dimensions of satisfaction but these differences are difficult to interpret. Each of the dimensions was moderately correlated with overall satisfaction, but not too strongly; if this was the case, it is likely that such a subscale would be measuring general satisfaction rather than a component of it.89

The WOMBLSQ has good internal reliability when viewed as one scale or when assessed on its individual subscales. Two of the subscales have a Cronbach's alpha of <0.7.35 It may be that these two scales need further development. Overall, the good alpha figures suggest that the dimensions are internally consistent and also separate from “global satisfaction”.

The instrument still has some weaknesses. Further work is needed to assess its test-retest reliability and its generalisability; the respondents were predominately of middle social class, in stable relationships, and of good educational achievement. However, it is sufficiently robust to be used in evaluating service developments as one component of assessing the quality of labour care that women receive.

Appendix Scales derived by principal components analysis (PCA) of WOMBLSQ4. Scales are intuitively named followed by their component questions. Scale means, standard deviations, percentage of variance explained by scale and Cronbach alpha coefficients are given. The final scale about general satisfaction was omitted from PCA with varimax rotation and Kaiser normalisation. Question numbers represent their order on the final questionnaire; negatively worded questions are shown with a minus sign.

Table 1

Parameters of the final and three development versions of the labour satisfaction questionnaires

Table 2

Demographic variables of women who responded to versions 3 and 4 of the labour satisfaction questionnaire

Table 3

Matrix of correlation coefficients for WOMBLSQ4 between all scales

Table 4

Mean satisfaction scores (%) on five subscales categorised by women's place of delivery

Table 5

Testing of construct validity of dimensions against depression, breast feeding, best educational level attained, social class of head of household, place of delivery, parity, whether labour induced, mode of delivery, labour complications and age

References

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.