Preconception indicators and associations with health outcomes reported in UK routine primary care data: a systematic review
Senior Research Fellow School of Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, UK MRC Lifecourse Epidemiology Centre, University of Southampton, Southampton, UK NIHR Southampton Biomedical Research Centre, University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK
GP and NIHR In Practice Fellow Primary Care Research Centre, Faculty of Medicine, University of Southampton, Southampton, UK
PhD candidate Centre for Public Health, School of Medicine, Dentistry and Biomedical Sciences, Queen’s University Belfast, Belfast, UK
Clinical Associate Professor Institute for Women’s Health, University College London, London, UK
MRC Clinical Research Fellow Warwick Medical School, University of Warwick, Coventry, UK
Lecturer in Nutrition and Behaviour Change Centre for Public Health, School of Medicine, Dentistry and Biomedical Sciences, Queen’s University Belfast, Belfast, UK
GP NHS Surrey Heartlands, UK
Professor of Public Health School of Primary Care, Population Sciences and Medical Education, Faculty of Medicine, University of Southampton, Southampton, UK NIHR Southampton Biomedical Research Centre, University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK NIHR Applied Research Collaboration Wessex, Southampton, UK
Professor of Sexual and Reproductive Health Institute for Women’s Health, University College London, London, UK
Professor of Epidemiology and Human Development School of Human Development and Health, Faculty of Medicine, University of Southampton, Southampton, UK MRC Lifecourse Epidemiology Centre, University of Southampton, Southampton, UK NIHR Southampton Biomedical Research Centre, University of Southampton and University Hospital Southampton NHS Foundation Trust, Southampton, UK
*Correspondence to Danielle Schoenaker University of Southampton MRC Lifecourse Epidemiology Centre Tremona Road, Southampton SO16 6YD, UK D.Schoenaker@soton.ac.uk https://orcid.org/0000-0002-7652-990XAbstract
Background
Routine primary care data may be a valuable resource for preconception health research and informing provision of preconception care.
Aim
To review how primary care data could provide information on the prevalence of preconception indicators and examine associations with maternal and offspring health outcomes.
Design and Setting
Systematic review of observational studies using UK routine primary care data.
Method
Literature searches were conducted in five databases (March 2023) to identify observational studies that used national primary care data from individuals aged 15-49 years. Preconception indicators were defined as medical, behavioural and social factors that may impact future pregnancies. Health outcomes included those that may occur during and after pregnancy. Screening, data extraction and quality assessment were conducted by two reviewers.
Results
From 5,259 records screened, 42 articles were included. The prevalence of 30 preconception indicators was described for female patients, ranging from 0.01% for sickle cell disease to >20% for each of advanced maternal age, previous caesarean section (among those with a recorded pregnancy), overweight, obesity, smoking, depression and anxiety (irrespective of pregnancy). Few studies reported indicators for male patients (n=3) or associations with outcomes (n=5). Most studies had low risk of bias, but missing data may limit generalisability.
Conclusion
Findings demonstrate that routinely collected UK primary care data can be used to identify patients’ preconception care needs. Linking primary care data with health outcomes collected in other datasets is underutilised but could help quantify how optimising preconception health and care can reduce adverse outcomes for mothers and children.
How this fits in
- Provision of preconception care is not currently embedded into routine clinical practice but may be informed by routinely collected primary care data.
- This systematic review demonstrates that UK primary care data can provide information on the prevalence of a range of medical, behavioural and social factors among female patients of reproductive age, while limited research has examined male preconception health or associations with maternal and offspring health outcomes.
- Routinely recorded electronic patient record data can be used by primary healthcare professionals to search for preconception risk factors and thereby support individualised preconception care, while aggregate data can be used by public health agencies to promote population-level preconception health.
- Further data quality improvements and linkage of routine health datasets are needed to support the provision of preconception care and future research on its benefits for maternal and offspring health outcomes.
Article notes
Competing Interest Statement
KMG has received reimbursement for speaking at conferences sponsored by companies selling nutritional products, and is part of an academic consortium that has received research funding from Abbott Nutrition, Nestec, BenevolentAI Bio Ltd. and Danone, outside the submitted work. No competing interests declared for other authors.
Funding Statement
DS is supported by the National Institute for Health and Care Research (NIHR) through an NIHR Advanced Fellowship (NIHR302955) and the NIHR Southampton Biomedical Research Centre (NIHR203319). MM is supported by the UK Medical Research Council (MR/W01498X/1). KMG is supported by the UK Medical Research Council (MC_UU_12011/4), the NIHR (NIHR Senior Investigator (NF-SI-0515-10042) and NIHR Southampton Biomedical Research Centre (NIHR203319)) and Alzheimers Research UK (ARUK-PG2022A-008).
Introduction
Preconception care is the provision of biomedical, behavioural and social interventions to people of reproductive age (15-49 years) before conception may occur with the aim of improving short-and longer-term parental and child health outcomes.1 Primary care teams have a key role in providing preconception care as identified by patients and healthcare professionals.2, 3 Preconception care delivered in primary care improves knowledge and preconception health behaviours in female patients, but there is currently less evidence about male patients or the impact on pregnancy and longer-term health outcomes.4, 5 In line with the National Institute for Health and Care Excellence (NICE) Clinical Knowledge Summary on preconception advice and management, primary care teams are encouraged to consider discussions about preconception health when appropriate, and to assess, manage and potentially optimise a range of physical and mental health conditions, health behaviours, and social needs prior to potential pregnancy.6 However, routine provision of preconception care is not currently widespread in UK clinical practice.7
To build the case for implementation of strategies and guidelines that optimise the population’s preconception health, the UK Preconception Partnership proposed an annual report card to describe and monitor preconception health.8 Our scoping review to inform national surveillance identified 65 preconception indicators (medical, behavioural and social risk factors that may impact potential future pregnancies among individuals of reproductive age) that are recorded in existing UK routine health data.9 A first report card was produced based on 23 indicators recorded in the national Maternity Service Data Set (MSDS), demonstrating that nine in 10 women in England enter pregnancy with at least one potentially modifiable risk factor for adverse pregnancy and birth outcomes.10, 11 Similarly, an analysis of primary care data from the Royal College of General Practitioners Research and Surveillance Centre found that 91% of women of reproductive age have a behavioural or medical risk factor for adverse pregnancy outcomes.12 These studies have to date focussed on preconception health of women (not men), and have not examined trends and trajectories in medical, behavioural and social indicators during the years leading up to pregnancy. Doing so would improve our ability to identify the population’s preconception care needs throughout their reproductive years. Routinely collected primary care data is potentially a unique resource to describe and monitor preconception health, and to examine the impact of (changes in) preconception indicators on improving outcomes such as gestational diabetes and preterm birth.
To inform future research and surveillance, and develop policy and clinical practice recommendations, we aimed to systematically review the literature to explore how UK routine primary care data could provide information on the prevalence of preconception indicators and examine associations with maternal and offspring health outcomes.
Methods
Search strategy and selection criteria
The protocol for this review was registered with PROSPERO,13 and the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA 2020) guideline used to ensure transparent reporting.14 A search strategy was developed, and searches conducted on 27 March 2023 (from inception date) in five databases: MEDLINE (Ovid), EMBASE (Ovid), Scopus, CINAHL, Web of Science (Supplementary Table 1). Supplementary searches using ‘preconception’ and ‘prepregnancy’ terms were conducted using databases from the British Journal of General Practice, and UK primary care datasets.13 Reference lists of included articles were screened for additional studies.
Articles were selected if they included findings from an observational study among individuals of reproductive age (15-49 years), used national patient-level routine primary care data collected in England, Wales, Scotland and/or Northern Ireland, and reported on the prevalence of at least one preconception indicator identified from our previous scoping review (Table 1).9 Articles not including new/original peer-reviewed results, and conference abstracts, were excluded.
Selection process
Search results were collated in EndNote and duplicates removed, before uploading to Covidence software. Titles and abstracts, followed by full text articles, were screened independently by two reviewers for inclusion. Disagreements or uncertainties were resolved through discussion.
Data extraction and synthesis
A standardised data extraction form was developed and piloted. Data were extracted by one reviewer and checked by a second reviewer. Disagreements were resolved between the two reviewers. All extracted data on study characteristics (grouped by primary care database), prevalence of preconception indicators, and measures of association between preconception indicators and outcomes (grouped by preconception indicator), were presented in tables. Meta-analysis was not conducted due to heterogeneity in preconception indicator definitions and inclusion and exclusion criteria of study populations.
Risk of bias assessment
Risk of bias was assessed for study findings on the prevalence of preconception indicators using the 10-item scale developed by Hoy et al rating internal and external validity.15 The Newcastle-Ottawa Scale (NOS) was used to rate risk of bias of study findings on associations between preconception indicators and health outcomes based on seven items related to selection, comparability, and exposure/outcome.16 Risk of bias was assessed by one reviewer, and checked by a second reviewer. Disagreements were resolved between the two reviewers. Studies were classified as low, moderate or high risk of bias (findings on prevalence),15 and good, fair or poor data quality (findings on associations)16 (scoring guides in Supplementary Tables 2, 3A-3B).
Results
From 9,401 identified records, 4,142 duplicates were removed and after title and abstract screening (n=5,259), 117 full-text articles were evaluated for eligibility (Figure 1). 42 articles were included, reporting findings from 11 primary care databases such as the Clinical Practice Research Datalink (CPRD) and The Health Improvement Network (THIN).
Most articles reported findings from primary care databases that included patients from three (n=1) or all four UK nations (n=30), or from England (n=6), Scotland (n=3) or Northern Ireland only (n=2) (Table 2, Supplementary Table 4). In 11 studies, a primary care dataset was linked with at least one other dataset, such as Hospital Episodes Statistics (HES), Office for National Statistics (ONS) mortality register, community prescribing data, or the Avon Longitudinal Study of Parents and Children. All studies included data on female patients; three studies also reported preconception indicators for male patients.
Prevalence of preconception indicators
Articles reported findings on 30 preconception indicators across seven of the 12 domains identified in our scoping review.9 Most studies included people of reproductive age irrespective of past/future pregnancy (n=26), while other studies included women with a pregnancy or birth recorded during the study period (n=15) or women with a recorded pregnancy and their partners (n=1) (Supplementary Table 5).
To obtain population-level estimates of preconception indicators, prevalence data were extracted only if reported (or could be calculated) for the overall study population of females or males of reproductive age (i.e. not if reported only in sub-populations such as patients with a specific condition or characteristic) (Table 3). Data on overall prevalence were available for 21 of the 42 studies, with the other 21 studies reporting prevalence estimates only in sub-populations. Additional preconception indicators reported in sub-populations included housing, domestic abuse, routine GP check-up in the past year, paternal age, previous pregnancy loss, history of assisted reproduction, alcohol consumption, substance misuse, cervical screening, and cardiovascular disease (Supplementary Table 4).
The prevalence of preconception indicators reported across studies and primary care databases varied widely, possibly due to differences in preconception indicator definitions, year of data collection (Table 3), and study populations (Supplementary Table 5). The prevalence of preconception indicators defined in line with our scoping review (i.e. excluding individual methods of contraception, and prescribed folic acid supplements),9 ranged from 0.01% for sickle cell disease to >20% for each of advanced maternal age, previous caesarean section (among those with a recorded pregnancy), overweight, obesity, smoking and diagnosis of depression and anxiety among female patients (irrespective of pregnancy). Only three studies reported preconception indicators for male patients, showing for example that the prevalence of depression among fathers (9.2%) was lower compared with mothers (22.2%),17 and the proportion of patients prescribed valproate was comparable among female (0.31%) and male patients (0.37%) in 2004, but much lower among females (0.16%) than males (0.36%) in 2018.18
Associations of preconception indicators with maternal and offspring outcomes
Five studies reported associations of preconception indicators (contraception prescription [n=1], sexually transmitted disease [n=1] and polycystic ovary syndrome [PCOS] [n=3]) with pregnancy and birth outcomes (Table 4). Outcome data were obtained from primary care data and/or linked HES data. Where two studies reported on comparable indicators and outcomes, consistent findings were shown for associations of PCOS with preterm delivery (<37 weeks gestational age) (positive association), high birthweight (>4kg) (no association) and low birthweight (<2.5kg) (inconclusive findings).19, 20
Risk of bias and data quality
Risk of bias for findings on the prevalence of preconception indicators was generally low (n=18/21 studies), however, none of the studies received a minimal score (no bias) (Supplementary Table 2). Potential biases were introduced based on representativeness and sampling frame (e.g. excluding women with no pregnancy reported or no linked data available), and indicator definition and measurement (e.g. reporting individual methods of contraception rather than population prescribed contraception, or reliance on medication prescription rather than dispensing data). Moreover, details of non-response (e.g. impact of missing data) were not reported in approximately half the studies. Data quality for studies examining associations of preconception indicators with health outcomes was rated as good for four of the five studies (Supplementary Tables 3A-3B).
Discussion
Summary
This systematic review found that UK routine primary care data can provide valuable information on patients’ medical, behavioural and social risk factors before (a potential) pregnancy. Based on 42 included studies among people of reproductive age or women with a pregnancy recorded during the study period, the prevalence of 30 preconception indicators was reported. Findings showed that >20% of female patients of reproductive age would benefit from support with smoking cessation, and management of weight, depression and anxiety. This would optimise their own health, and improve their chance of a successful pregnancy and healthy baby if that is something they want. Limited research has used primary care data to examine preconception indicators among male patients, or associations of preconception indicators with pregnancy outcomes and longer-term maternal and offspring health outcomes.
Strengths and limitations
This is the first systematic review to demonstrate how national routine primary care databases can be used to describe the population’s preconception health, to inform clinical practice and future research directions. Comprehensive, prospectively registered review methods were used. Our search was limited to UK primary care data and findings may not be generalisable to other countries. Preconception indicators were selected based on our previous scoping review;9 so potentially relevant indicators not included in this review or not reported in the included studies would have been missed. Moreover, some preconception indicators (such as dietary intake and physical activity) are not routinely recorded in general practice.
Comparison with existing literature
Findings from our review complement our previous preconception report card based on the MSDS,10 showing that national routine health data are a valuable resource to describe and monitor women’s preconception health. Half of the preconception indicators identified in this review were also reported in the MSDS, with comparable prevalence estimates for most indicators (e.g. teenage pregnancy, previous caesarean delivery, overweight, obesity), while other indicators may be underreported in primary care (e.g. over the counter folic acid supplementation) or in the MSDS (e.g. mental health conditions).10 Published primary care data reported an additional 15 indicators not included in the MSDS (e.g. fertility problems, contraception, relevant medical conditions, teratogenic medication use). Linkage of these (and other) national routine health datasets would enhance the quality of preconception report cards and surveillance (Box 1). Based on linkage of primary care and HES datasets, findings from our review (n=2 studies19, 20) confirm the previously reported association of PCOS with increased risk of preterm delivery.21
Findings from our review are also in line with previous research reporting primary care data quality issues.22–24 Studies included in our review documented substantial missing data (20-60%) for ethnicity and BMI category, likely varying across sub-populations. Coding quality is related to financial incentives such as the Quality and Outcomes Framework (QOF), which may improve accurate recording of selected indicators but also distort prevalence estimates over time.22 The prevalence of some preconception indicators may be underestimated as not all conditions are solely diagnosed and coded in general practice (e.g. sexually transmitted disease),25 or medications and supplements prescribed (e.g. contraception, folic acid supplements).26 Another commonly reported limitation is the representation of selected general practices in research databases,22, 23 often limited to practices that use one of four main software platforms to manage electronic patient records (EPRs) and further determined by voluntary ‘opt ins’.22, 23 As a result, primary care databases may underrepresent specific regions and bias national prevalence estimates of preconception indicators and associations with health outcomes.
Implications for research and clinical practice
Our findings demonstrate that many preconception indicators are routinely recorded in EPRs, allowing primary healthcare professionals to search for risk factors and provide individualised preconception care. A digital risk screening template has been developed in the Ardens Clinical Decision Support System based on the NICE Clinical Knowledge Summary,6 to support primary healthcare professionals to improve their preconception care practice, screening, coding and recording of indicators. Further work is required to co-develop practical guidance and resources to support integration of preconception care into every day clinical practice (Box 1).
Our findings identify the need to use standardised definitions when reporting preconception indicators (Box 1). Due to heterogeneity in definitions, the prevalence of preconception indicators across UK nations, and changes over time, could not be directly compared across studies. However, Lee and colleagues applied standardised definitions to CRPD (UK) and SAIL data (Wales), showing comparable prevalence estimates for some indicators (e.g. obesity, depression), but higher (e.g. smoking, underweight, anxiety, asthma) or lower (e.g. advanced maternal age) prevalence for other indicators, when comparing pregnant women in Wales with those in the UK overall.27 Moreover, standardised reporting within the same database showed, for example, increases over time in the prevalence of type 2 diabetes (1995-2017),29 alongside decreases in poor diabetes control (2004-2017).29, 30
Lastly, the limited reporting of male preconception indicators, and associations of preconception health with pregnancy, maternal and offspring health outcomes, calls for further research. Many of the preconception indicators reported for female patients are also relevant to male patients (e.g. smoking, obesity), with increasing evidence suggesting better paternal preconception health is associated with reduced risks of infertility and adverse pregnancy and offspring health and developmental outcomes.31–33 To enable further research, improvements are needed in the way that families (i.e. biological parents and their children) can be identified and data linked.17, 34 Primary care data also provide a unique opportunity to examine trajectories of preconception health during reproductive years irrespective of pregnancy, and to quantify the extent to which these reduce adverse pregnancy and offspring health outcomes. Future research would be enhanced by linkage of primary care and other routine health datasets beyond the identified existing linkages (e.g. MSDS and Community Services Data Set) to determine the short-and longer-term benefits of preconception care (Box 1).
Conclusion
Routinely collected primary care data in the UK provide a valuable resource for research and surveillance, and can guide to provision of preconception care. Improvements in coding and reporting, and linkage of general practice systems and other national routine health datasets, would inform evidence-based provision of preconception care in primary care.
Supporting information
Data Availability
All data produced in the present work are contained in the manuscript.
Funding
DS is supported by the National Institute for Health and Care Research (NIHR) through an NIHR Advanced Fellowship (NIHR302955) and the NIHR Southampton Biomedical Research Centre (NIHR203319). MM is supported by the UK Medical Research Council (MR/W01498X/1). KMG is supported by the UK Medical Research Council (MC_UU_12011/4), the NIHR (NIHR Senior Investigator (NF-SI-0515-10042) and NIHR Southampton Biomedical Research Centre (NIHR203319)) and Alzheimer’s Research UK (ARUK-PG2022A-008). For the purpose of Open Access, the author has applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising from this submission.
Competing interests
KMG has received reimbursement for speaking at conferences sponsored by companies selling nutritional products, and is part of an academic consortium that has received research funding from Abbott Nutrition, Nestec, BenevolentAI Bio Ltd. and Danone, outside the submitted work. No competing interests declared for other authors.