2.1. Pregnant women with complex social factors: a model for service provision
This guideline was commissioned in acknowledgement of unaddressed problems in pregnant women with complex social factors; that is, women whose social situation may impact adversely on the outcomes of pregnancy for them and their baby. A number of key reports have highlighted some of the issues faced by these women.
Saving Mothers’ Lives (2007)3 drew attention to the fact that socially excluded women are at higher risk of death during or after pregnancy than other women. The vulnerable women with socially complex lives who died were far less likely to seek antenatal care early in pregnancy or to stay in regular contact with maternity services. Overall, 17% of the women who died from direct or indirect causes booked for maternity care after 22 weeks of gestational age or had missed over four routine antenatal visits compared to 2% of the general population. Compared to women who booked prior to 20 weeks, women who booked late or missed more than four routine appointments were more likely to be: black African or Caribbean; experiencing domestic abuse; substance misusers; known to social services or child protection services; or unemployed.
Maternity Matters (2007)4 included a section focussing on ‘The Equality Impact Assessment’. This reported that women in all vulnerable groups are likely to recognise their pregnancy later, to first see a health professional later and to book later for antenatal care. Maternity Matters4 also highlighted the fact that commissioners need to understand what barriers in their current services may prevent these women from seeking care early, or maintaining contact with their maternity services, and to overcome these by providing more flexible services at times and places which meet their needs.
Perinatal Mortality (2009)5 demonstrated that women from non-white ethnic groups and women in the most deprived population quintile had stillbirth and neonatal death rates that were twice those of white women and those resident in the least deprived areas. It is also recognised that maternal stress in pregnancy has a detrimental effect on subsequent childhood development.
The following general principles were established in the ‘Changing Childbirth’ report6 and the NICE guideline ‘Antenatal care: routine care for the healthy pregnant woman’ (2008)1 (clinical guideline 62):
Women should be the focus of maternity care, with an emphasis on providing choice, easy access and continuity of care.
Care during pregnancy should enable a woman to make informed decisions, based on her needs, having discussed matters fully with the healthcare professionals involved. It is essential that women, their partners and their families be treated with kindness, respect and dignity. This includes respect for their views, beliefs and values. Good communication between healthcare professionals and women is vital. Any information given should be evidence-based and supported by appropriate, written information. All information should also be accessible to women with additional needs, for example those with sensory or learning disabilities, and to women who do not speak or read English.
The NICE clinical guideline 62 ‘Antenatal care’ (2008)1 recommends that, ideally, booking should occur by 10 weeks of gestation in order to facilitate screening for haemoglobinopathies and other conditions. In addition to allowing screening, this will also enable good history taking, the provision of advice as early as possible and the organisation of a booking scan at 12–14 weeks.
This guidance is especially pertinent for women with complex social problems.
If a woman has additional health problems complicating pregnancy (such as hypertension or diabetes) in addition to social problems, the relevant NICE guideline should also be consulted to effectively direct clinical care.
In planning this guideline it was intended that four exemplar populations would be used to represent women with complex social factors that might impact on their health during pregnancy as well as pregnancy outcomes. These groups were chosen from groups highlighted in Saving Mothers’ Lives3 as having poorer pregnancy outcomes than the general population and in consultation with the guideline’s stakeholders. The four exemplar populations are:
women who are substance misusers (including drugs and/or alcohol)
recent migrants, refugees, asylum seekers, and women with little or no English
young women aged under 20
It is recognised that this is not a comprehensive list and that there is a range of other complex social factors which may impact on pregnancy (such as learning disabilities, families where one or both parents are unemployed, imprisonment and poverty). It is also recognised that vulnerable women often have multiple needs and may experience a number of complex social factors at the same time. Although the majority of the recommendations in the guideline have been written for the specific population groups listed above, there have been some general principles for care identified which can be applied to all vulnerable women with complex social factors (see chapter 3). It is hoped that healthcare professionals consider the full range of complex social factors which may impact on pregnancy and apply these recommendations accordingly. While it can be helpful to consider general findings for a particular vulnerable population, when designing a service for that group it is important to consider each woman’s needs individually and to recognise that she may have needs that cross those population boundaries and require antenatal input from a range of health and social care professionals, as well as third sector agencies.
The guideline focuses on women who are recent migrants (defined here as women who have come to the country of residence within the past year), asylum seekers, refugees and women with little or no English (or language spoken in country of residence) rather than migrant women generally, as these sub-groups are highlighted within Saving Mothers’ Lives3 as being particularly at risk of poor birth outcomes. These poor outcomes are thought to be related specifically to difficulty accessing services due to language barriers and a lack of knowledge and understanding of how the health and social care system works. Since improving access is a major focus of this guideline, this was felt to be a more appropriate targeting of the migrant population group than considering all migrant women as a whole.
The guideline describes how access to care can be improved, how contact with antenatal carers can be maintained, what additional consultations and supports are required and what additional information should be provided for pregnant women with complex social factors.
Specific issues considered include:
consideration of the most appropriate healthcare setting for maternity care provision
practice models for overcoming barriers and facilitating access, including access to appropriate interpreting services and all necessary care
ways of communicating information to women so that they can make appropriate choices
optimisation of resources.
2.7. Guideline development methodology
This guidance was commissioned by NICE and developed in accordance with the guideline development process outlined in NICE’s ‘The guidelines manual 2007’,26 with post development phases carried out as per NICE’s ‘The guidelines manual 2009’.27
summarises the key stages of the process and which version of the guidelines manual was followed at each stage.
Stages in the NICE guideline development process and versions of ‘The guidelines manual’ followed at each stage.
In accordance with NICE’s Equality Scheme, ethnic and cultural considerations and factors relating to disabilities have been considered by the GDG throughout the development process and specifically addressed in individual recommendations where relevant. This includes consideration of target populations which include women with little or no English, asylum seekers, refugees and recent migrants, substance misusing women and young women aged under 20. Further information is available from: www.nice.org.uk/aboutnice/howwework/NICEEqualityScheme.jsp.
Forming clinical questions and search strategies
Five clinical questions were developed based on the scope of the guideline. The questions focussed on access to care, barriers to care, maintaining contact with care, additional consultations, support and information needed over and above that set out in the NICE clinical guideline 62: ‘Antenatal care’ (2008)1. These questions were asked for each of the guideline populations which are:
women who misuse substances
women who are recent migrants, refugees, asylum seekers, or who speak little or no English,
young women aged under 20
The main purpose of the guideline is to provide guidance on how services can be organised in order to improve women’s access to and contact with antenatal care. In order to determine the clinical and economic effectiveness of care provision it is necessary to review evidence that shows which service interventions lead to improved pregnancy outcomes: this requires findings from comparative studies reporting ‘hard’ outcomes, such as birthweight and gestation at birth. While acknowledging that this approach would reduce the number of potential studies for inclusion, it was felt to be very important that this distinction was made in order that recommendations could be made based on evidence of effectiveness. Where this evidence was found to be lacking research recommendations have been made.
Due to the complex nature of the interventions of interest, and lack of certainty among the GDG over what terms would be appropriate to describe some of these, it was decided that searches would be carried out for a particular population rather than by guideline question, and so four broad searches were run covering all five questions for each population. This approach was similar to that adopted by Tina Lavender, Soo Downe, Kenny Finnlayson and Denis Walsh, for the report ‘Access to antenatal care: A systematic review’ (unpublished report; February 2007)28 and reflects the uncertainty inherent in the review questions which focus on antenatal care provision generally rather than specific interventions within antenatal care. Searching in this way increases the sensitivity of the search at the expense of specificity.
Four search strategies were developed to capture studies examining antenatal service provision for each of the guideline’s target populations. For each population, searches were run in Medline (1950 onwards), Embase (1980 onwards), Cumulative Index to Nursing and Allied Health Literature (CINAHL; 1982 onwards) and three Cochrane databases (Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews and the Database of Abstracts of Reviews of Effects) as well as PsycInfo. In addition, for three out of the four populations the ASSIA, Sociological Abstracts and Social Services Abstracts databases were searched. These databases were not searched for the population of women experiencing domestic abuse (or for the re-run searches) as the subscription to the databases was discontinued prior to this being carried out. However, the decision to stop the subscription was taken only after the contribution made by the social science databases had been investigated. It was found that the three social science databases contributed less than 5% of the total number of hits obtained across all three populations searched (young women aged under 20, 1.7%; recent migrants, 5.9%; substance misusers, 4.9%).
Searches to identify economic studies were undertaken using the above databases and the NHS Economic Evaluation Database (NHS EED). None of the searches was limited by study type, date or language of publication (although publications in languages other than English were not reviewed). There was no attempt to search grey literature (conferences, abstracts, theses and unpublished trials) and hand searching of journals not indexed on the databases was not undertaken.
Towards the end of the guideline development process, the searches were updated and reexecuted to include evidence published and indexed in the databases by 2 September 2009. Full details of the systematic searches, including the sources searched and the search strategies for each review question, are presented in appendix G.
Criteria for deciding inclusion/exclusion of studies
Studies from all countries and all dates were considered for inclusion. Studies were considered for inclusion if they involved the specific target population as defined in the PICO tables (population, intervention, comparison, outcome tables; see appendix H). Studies involving exclusively small indigenous groups not common in the UK, such as Australian Aborigines, Native Americans and Inuit, were excluded. Studies originally considered for inclusion could be later excluded if the GDG felt the target population was so different from a UK population that it would make generalisation of findings impossible. Where this was the case, the reason for exclusion was detailed in the excluded studies tables.
The target population for women who misuse substances, including alcohol, was defined as regular use of substances and/or alcohol, for example on a weekly basis, to an extent where physical dependence and/or harm to their health or that of their unborn baby is a risk. The level of substance misuse defined within each paper considered for inclusion was discussed with the GDG where it was unclear whether the population was one that would be defined as ‘substance misusers’ and decisions to include or exclude made on the basis of this discussion and with the GDG’s agreement. The population of recent migrants, refugees, asylum seekers and women with little or no English was identified by explicit mention of these terms, or by GDG consensus that the included population was likely to represent one or more of these groups.
Recent migrants were defined as women who had come to the resident country within the past year. However, few papers used this definition and it was often necessary to decide inclusion/exclusion based on level of understanding of English (or mother tongue of the resident country), degree of acculturalisation or understanding of health care services where this was reported. Studies including women from black and ethnic minority groups were excluded if it appeared the sample was primarily made up of women who were not recent migrants and spoke English, or if there was no information reported that would allow a judgement to be made regarding these criteria. Where the study sample was made up of a combination of women falling across categories, half or more of the sample had to comprise women who were recent migrants, non-English speaking, asylum seekers or refugees for the study to be included.
For intervention questions (Q1a, Q2, Q3 and Q4) only comparative studies were included. Types of study considered included randomised controlled trials (RCTs), non-randomised controlled studies, observational cohort studies, before and after studies and case–control studies. Comparisons included cohort studies with groups that received different service interventions drawn from the same hospital, comparisons between services in different hospitals/clinics and comparisons between outcomes obtained for the antenatal service under investigation and population figures for those outcomes, for example for that town, city, region or US state. Comparison groups would usually be receiving standard or ‘usual’ care, or any alternative model of care as described by the authors. Comparative studies with sample sizes of five or less in any of the study groups were excluded.
For questions addressing barriers to access (Q1b), qualitative and descriptive studies were also considered for inclusion. Individual or very small case reports (n=1 or 2) were excluded. Interventions aimed at improving affordability of health care in countries where women would be expected to pay for antenatal care (such as the USA) were excluded on the advice of the GDG as it was felt these interventions were not applicable to a UK NHS setting.
For the substance misuse population, programmes aimed at helping women manage their substance misuse were excluded from the scope as these do not form part of antenatal care (that is, this care is not delivered by midwives or obstetricians). However, programmes which combined substance misuse management with aspects of antenatal care, for example advice and information regarding healthy eating for pregnancy, preparation for parenthood, assessing progress of pregnancy and so on, were included as these focus on services that fall within the remit of antenatal maternity care.
Outcomes for clinical questions were decided by the GDG prior to reviewing being conducted and are presented in the PICO tables in appendix H. Studies which did not report outcomes of interest were excluded.
The large number of ‘hits’ generated using these broad search strategies were then subjected to three rounds of ‘weeding’ (or sifting). This was carried out firstly to exclude opinion papers, letters, editorials, commentaries and so on and studies which did not focus on antenatal care or the target population; secondly to remove papers that did not address the interventions or outcomes of interest. A third round of weeding was carried out by a second reviewer to remove any remaining papers that did not meet the inclusion criteria. These rounds of weeding were necessary due to the very large databases generated by the searches and the need to check previous decision-making in order to minimise the risk of ordering large numbers of irrelevant papers. In addition, a convenience sample of weeded out papers (approximately 10%) was checked by a second reviewer to ensure exclusion criteria had been applied correctly. Hard copies of potentially relevant papers were then obtained and each paper read in its entirety and assessed again to check it met the inclusion criteria as set out in the PICO tables, for relevant methodology and to determine which of the guideline questions it addressed. Any difficulties encountered at any stage during this process were resolved through discussion with a senior methodologist who also read all papers where a decision was sought.
The population of young women aged under 20 was the first to be reviewed. For this population, papers were assigned to the relevant guideline question(s) when they were received as hard copies, prior to assessment against inclusion criteria and quality appraisal. Papers for consideration were then re-read and data extracted. If, at this stage, it became apparent that the paper should be excluded from that question it was assigned as such and then recorded in the excluded studies table for that particular question. At this stage, some papers were moved from inclusion in one question to inclusion in an alternative question, or were included in more than one question depending on the outcomes reported. This tendency to include papers in more than one question reflects the high degree of similarity between the guideline questions and the iterative nature of the reviewing undertaken. Once this process had been undertaken for this population it became apparent that a more efficient system could be adopted. For the other three populations hard copies of papers to be considered were read through and if they did not address any of the guideline questions, had no comparators or did not describe any barriers to care, they were excluded prior to being assigned to a particular question. These were then recorded in an excluded studies table labelled ‘first round exclusions’. Papers that remained for inclusion were then assigned at this stage to a specific guideline question or questions prior to quality appraisal and data extraction.
Many papers were considered for inclusion in more than one question due to the large degree of overlap between question content. This means that an individual paper may have been excluded from one question but included in another depending on reported outcomes. Decisions about inclusion/exclusion of studies were further supported by input from an experienced methodologist based at NICE.
All excluded studies were checked and queries raised where it was not clear whether inclusion/exclusion criteria had been applied correctly. Suggested changes were discussed and agreement reached on where changes needed to be made, and studies added to the reviews as appropriate. The total numbers of ‘hits’ for each population, the number of hard copies assessed for inclusion and final numbers of papers included and excluded are summarised in tables presented in appendix H. Details of excluded studies are presented in appendix F.
Due to the anticipated high degree of heterogeneity between studies and the lack of good quality comparative evidence, meta-analysis of study results was not appropriate. Sub-group analysis was planned for English speaking versus non-English speaking women and older versus younger teenagers where this was not reported in the paper. However, none of the reviewed studies included data in a form that would allow such analyses to be undertaken.
Reviewing and grading the evidence
Evidence relating to effectiveness was reviewed and graded using the hierarchical system presented in . This system reflects the susceptibility to bias inherent in particular study designs.
Levels of evidence for intervention studies.
The type of clinical question dictates the highest level of evidence that may be sought. In assessing the quality of the evidence, each study receives a quality rating coded as ‘++’, ‘+’ or ‘−’. For issues of intervention effectiveness, the highest possible evidence level (EL) is a well-conducted systematic review or meta-analysis of RCTs with a very low risk of bias (EL = 1++) or an individual RCT with low risk of bias (EL = 1+). Studies of poor quality (high risk of bias) are rated as ‘−’. Usually, studies rated as ‘−’ should not be used as a basis for making a recommendation, but they can be used to inform recommendations.
For each clinical question, the highest available level of evidence was sought. However, due to the nature of the interventions under investigation it was anticipated that most of the evidence would be from retrospective observational studies. Where a number of low quality comparative studies were considered, those with small sample sizes (five or less in each group for comparative studies and one or two cases for case reports/series) were excluded along with those with two or more serious flaws (other than small sample size) which would contribute to significant bias.
Summary results and data from each study are presented in the text. More detailed results and data are presented in the evidence tables provided in appendix E. Where possible, dichotomous outcomes are presented as relative risks (RRs) with 95% confidence intervals (CIs), and continuous outcomes are presented as mean differences with 95% CIs or standard deviations (SDs). It should be noted, however, that the findings reported in the included evidence rarely allow this level of analysis.
The body of evidence identified for each clinical question was synthesised narratively in clinical evidence statements.
Formal consensus within the guideline
Formal consensus was used for three purposes within the guideline: for deciding the most important barriers to antenatal care (question 1b); for choosing key priorities for implementation; and for choosing key research recommendations.
Anonymous formal consensus voting was conducted among GDG members to identify the most important barriers to care. This was carried out during GDG meetings following presentation of the evidence relating to the barriers question. A form containing all barriers identified from the evidence was distributed to each GDG member, and they were then each asked to place a tick by the five barriers they saw as being most important. When this had been done, all forms were given to a member of the technical team and the number of votes for each barrier was totalled. The number of barriers identified as priorities was four or five, depending upon how the votes were cast. The results from the voting were fed back to the GDG at the same meeting and the barriers with the most votes were then used to form the basis of the recommendations relating to overcoming barriers to care.
Key priorities for implementation were voted for using formal consensus anonymous voting in a similar way. GDG members were asked at a meeting to vote for their top 10 priority recommendations using pencil and paper voting forms. Forms were examined by a technical team member and votes summed. After the first round of voting, eight top priorities were identified. A second round of anonymous voting was then undertaken with all remaining recommendations that had received at least one vote on the first round of voting. The two recommendations with the most votes on the second round of voting were added to the previous eight to give the top 10 key priorities for implementation. Following stakeholder consultation, four of these recommendations were divided into two to aid clarity. Each of the new recommendations made this way were retained as key priorities for implementation, making a total of 14 key priorities in the final guideline.
Key research recommendations were chosen using anonymous voting conducted via e-mail. Again, two rounds of voting were undertaken in a way similar to that described for key recommendations in order to identify five key research recommendations.
Health economics
The purpose of including economic evidence in a clinical guideline is to allow recommendations to be made based on the cost-effectiveness of different forms of care as well as the clinical effectiveness. The aim is to produce guidance that uses scarce health service resources efficiently; that is, providing the best possible care within resource constraints.
The aim of the health economic input to the guideline was to inform the GDG of potential economic issues relating to providing additional specialist services and consultations to improve access and uptake of antenatal care for vulnerable women, and to ensure that recommendations represented a cost-effective use of healthcare resources.
Systematic searches for published economic evidence were undertaken for all the populations included in the guideline but no relevant economic evaluations were identified. One area was identified by the GDG as having significant resources implications and uncertainty surrounding the effectiveness. Therefore, for this guideline an economic evaluation was conducted to support the following area:
A simple economic model was developed in order to present the GDG with the potential consequences of providing various specialist services with differing costs. The service descriptions were based on programmes currently running across the UK. No audit data were available and no good quality analysis work had been carried out to evaluate the efficacy of providing additional services to these vulnerable groups. As there was no good quality evidence on effectiveness of specialist services, the economic model was used to illustrate what level of effectiveness would be required from different services in order for those services to be considered cost effective using the NICE willingness-to-pay threshold. The economic model in its current form does not result in an incremental cost-effectiveness ratio.
The relevance of the evidence provided by this analysis depends on the assumptions included in the model and how they apply to real-world settings. As the analyses are not based on good quality clinical evidence they can only be used to illustrate the problem as we do not know how effective specialist services will be in improving health outcomes. Where new specialist services are set up, auditing and evaluation will provide useful inputs to update this analysis in the future.
Evidence to recommendations
For each guideline question, recommendations for service provision and care were derived using, and linked explicitly to, the evidence that supported them. In the first instance, informal consensus methods were used by the GDG to agree service delivery and clinical effectiveness evidence statements. Evidence summaries derived from qualitative studies describing reported barriers to accessing care were presented in tabular form. Statements summarising the GDG’s interpretation of the evidence and any extrapolation from the evidence used to form recommendations were also prepared to ensure transparency in the decision-making process.
In areas where no substantial good quality evidence was identified, the GDG made consensus statements and used their collective experience and expertise to identify good practice. Health economic modelling was used to support recommendations and this is also explained in the GDG interpretations of evidence. The GDG also identified areas where evidence to answer the guideline questions was lacking and used this information to formulate and prioritise recommendations for future research.
Formal consensus voting was carried out among GDG members to identify the five barriers they considered most important for UK NHS services to address in order to promote access to care. This was carried out independently for each population sub-group and the key barriers identified used to inform the recommendations for that population.
Towards the end of the guideline development process, formal consensus methods were used to consider all the guideline recommendations and research recommendations that had been drafted previously. The GDG identified 10 ‘key priorities for implementation’ (key recommendations) and five high-priority research recommendations. The key priorities for implementation were those recommendations likely to have the biggest impact on provision of antenatal care and pregnancy outcomes for at-risk population subgroups in the NHS; they were selected through two rounds of formal voting using pencil and paper during a GDG meeting. The priority research recommendations were selected using two rounds of formal voting carried out electronically via e-mail.
Service survey
A service survey has been undertaken with the help of GDG members to identify examples of current practice within the NHS where services have been designed to deliver care to one of the four target populations for the guideline (young women aged under 20, substance misusers, recent migrants/refugees/asylum seekers/women with little or no English and victims of domestic abuse).
A named person suggested by a GDG member (typically a care provider) was contacted by telephone and/or e-mail and asked if they would be willing to participate in the survey. If they agreed, the questionnaire was administered either electronically via e-mail, via telephone interview, or as a face to face interview. The questionnaire collected data under the following headings:
Access to care (source of referrals; gestation at booking)
Description of the service provided (target population, who provides care, staffing levels, degree of obstetric input, setting, home visiting, additional facilities provided such as childcare, length and frequency of consultations, content of consultations provision of antenatal education/support groups)
Additional consultations (details of appointments over and above ‘routine’ antenatal care as set out in the NICE clinical guideline 62 ‘Antenatal care’ (2008)
1)
Attendance (maintaining contact with services, number of antenatal appointments kept, rate of DNAs (did not attend), attendance at other appointments for example with social services or parole officers, how the service encourages attendance)
Interfaces/links with other services (how the service makes these links, how communication is made with other agencies, problems with communication and how these are overcome)
Training (additional training provided to/identified as being needed by staff providing the specialist service, training provided by specialist staff to other staff members)
Audit data (any audit data including process or clinical outcomes)
Any other information (including identified problems and how these have been overcome).
Each respondent was also asked if they could provide the name and contact details for another service provider involved in antenatal care designed to reach any of the guideline’s target populations. Twenty-three services were suggested by GDG members and other midwives. We attempted to contact all the services and were able to speak to 16 midwives about the services they were involved in. Before the guideline went to consultation (February 2010), one service had been closed. Two midwives contacted were working for the Imperial College NHS Trust and although they worked with different groups of vulnerable women the service provided was the same, so it was decided to include one service description to cover all the groups of women. By using this ‘snowballing’ technique the survey was able to include 14 examples of a range of services using different models. The service descriptions are included as illustrations of what can be provided for enhanced antenatal care for vulnerable women. Some examples are used in the guideline text to illustrate recommendations where the service provided includes components contained in the recommendations.
The service descriptions have not been used to inform the evidence base for recommendations as the survey was carried out as a descriptive exercise rather than in a way that would allow generalisability of findings. In addition, none of the services has yet been evaluated in terms of their impact on key aims, such as improving access to or contact with antenatal care, or effectiveness in improving pregnancy outcomes. Nevertheless, the service descriptions provide examples of innovative services currently provided within the NHS which show how the guideline’s recommendations might be implemented. The need for evaluation of such services is a key message of the guideline.
Full details of survey findings can be found in appendix D.
Stakeholder involvement in the guideline development process
Registered stakeholder organisations were invited to comment on the draft scope of the guideline. Stakeholder comments were each taken into consideration by the guideline’s technical development team and GDG chair in reaching final decisions about its scope. Each comment was responded to individually. Stakeholder organisations also commented on the draft guideline. The GDG considered the comments made and made amendments to the recommendations and text of the full version in light of these comments. Again, individual responses were given for each comment received. The service descriptions reported in the guideline and used to illustrate the recommendations have all been provided by stakeholders, who have also kindly given permission for contact details to be included in the guideline so that other healthcare professionals and commissioners can find out more about a particular service.
Stakeholders were also asked to assist in checking the factual accuracy of the guideline during the pre-publication check. Additional changes were made to the guideline following this consultation in order to address factual errors. A full list of the stakeholders for this guideline can be found in appendix C.
