The emergence of common health conditions across the life course: evidence from the Born in Bradford family cohort.

Poverty and deprivation severely impact women's health in the UK, reducing life expectancy and diminishing the number of years lived in good health, particularly within disadvantaged communities 1 . These health disparities are often exacerbated for women from Black, Asian, and other minority ethnic backgrounds, who frequently contend with intersecting inequalities tied to both socioeconomic status and ethnicity 2 . Factors such as poor living conditions, chronic stress, limited access to healthcare and adverse environmental exposures accumulate across the lifespan 3 , contributing to long-term physical and mental health problems, including those arising during the perinatal period, while hindering opportunities for healthy living. Addressing both poverty and systemic ethnic inequalities is therefore essential for improving women's overall and reproductive health. Bradford, a city in West Yorkshire, England, faces significant challenges related to these issues. Characterised by a young population, high levels of deprivation (ranking 13 th most deprived local authority out of 317 in England, with almost a fifth of children living in low income families) 4 , and notable ethnic diversity, with 57% identifying as White British and 26% of Pakistani heritage 5 . Concerns regarding high rates of illness in Bradford led to the establishment of the Born in Bradford (BiB) cohort study. Between 2007 and 2010, the BiB study recruited 12,453 women during 13,776 pregnancies (resulting in 13,858 births) to investigate how socioeconomic, environmental, behavioural, and genetic factors influence health and well-being 6 . As these mothers now enter middle age, the BiB cohort presents a unique opportunity to examine the accumulation of health conditions following childbirth. This study therefore aimed to (1) describe the rates of common non-communicable health conditions among mothers in the BiB cohort; and (2) determine if the incidence of these conditions differs between the two largest ethnic groups in Bradford.

This study utilised data from the BiB cohort. Between 2007 and 2010, BiB enrolled 12,453 pregnant women during 13,776 pregnancies who were booked for delivery at the Bradford Royal Infirmary, resulting in 13,858 births 6 . At recruitment, participants completed a baseline questionnaire, providing demographic information, and consented to the linkage of their electronic health records (EHR). Linkage to routine health data has been successfully achieved for over 98% of the cohort participants. For this study, primary care data containing information on health conditions were extracted from SystmOne, the clinical EHR system used universally across all GP practices in Bradford. This study also leverages a collaboration between BiB and MuM-PreDiCT, a UK-wide research project focused on pregnant women with two or more long-term health conditions 7 . We adopted a list of 79 long-term health conditions identified by MuM-PreDiCT as important factors in multimorbidity, originally selected for their prevalence, potential to impact on pregnancy outcomes, patient importance, and data availability. For the present study, while our exploratory analysis assessed the incidence of all 79 health conditions within the BiB cohort, we report trends for 21 conditions (19 physical and 2 mental health) with an incidence of ≥2% per 000 person-years in at least one of the major ethnic groups in Bradford. Many of these 21 health conditions aggregate several diagnoses; see Table 1 for details and definitions. Information on ethnicity was obtained from research questionnaires and routine data sources, all of which were self-reported by the participant and categorised using the same classification as the 2001 UK census 8 . White European ethnicity included those who identified as White British, Irish, Gypsy/Irish traveller, and White Other; South Asian ethnicity was assigned to participants of Pakistani, Indian, and Bangladeshi heritage. Ethics approval for the Born in Bradford study was granted by the National Health Service Health Research Authority Yorkshire and the Humber (Bradford Leeds) Research Ethics Committee (reference: 07/H1302/112, date of approval 01/04/2008). Informed consent for data collection and linkage to routine healthcare records was provided by participants at recruitment to the BiB cohort study. Further information on our privacy policy can be found here: https://borninbradford.nhs.uk/privacy-policy/ . To ensure sufficient data for identifying existing health conditions, we excluded participants with fewer than two years of medical records prior to recruitment. Baseline prevalence for each condition was defined as having either a GP-recorded diagnostic code or a relevant prescription during the two years before the recruitment date (“time 0”). Prevalence was calculated separately for each ethnic group and is reported as percentages. Incidence rates per 1,000 person years (IR/1000) were calculated from the date of recruitment, with follow-up ending at the earliest of: last GP-recorded event (up to 31 st December 2021), withdrawal from the study, relocation out of the area, or death. Individuals with prevalent cases were excluded from the incidence calculations. For the main categories of conditions, we used age-adjusted Cox proportional hazards models to estimate hazard ratios (HR), comparing the risk of developing each health condition between South Asian and White European participants (the reference group). An HR of 1 indicates no difference in risk between groups; values above 1 suggests a higher risk among South Asians, while values below 1 suggest a lower risk. HR were not calculated for individual conditions within broader categories due to small sample sizes. Analyses were conducted using Stata/SE version 17 software ( https://www.stata.com/ ).

The sample comprised 9,784 mothers, of whom 52% were of South Asian ethnicity; see Figure 1 for further details of the sample selection, and Table 2 for participant characteristics. The prevalence and IR/1000 for each condition are presented in Table 3 . The most prevalent conditions among White European and South Asian women were common mental health disorders (29.3% vs 15.7% respectively), eczema (25.3% vs 26.4%), and asthma (22.5% vs 9.1%). The highest IR/1000 was observed for common mental health disorders (55.9 vs 35.4 in White European and South Asian women respectively). High IR/1000 was observed in South Asian women for eczema, allergic rhinoconjunctivitis, and migraine. For the individual health conditions grouped within broader categories, the IR/1000 were statistically significantly higher in White European women compared to South Asian women for fibromyalgia, endometriosis, genital prolapse, and all conditions classified under common and other mental health disorders apart from dissociative disorders. Conversely, South Asian women had higher rates of hypothyroidism, seborrhoeic dermatitis, lichen planus, systemic lupus erythematosus, back pain, osteoarthritis, polycystic ovarian syndrome, uterine fibroids, infertility, urinary incontinence, and all conditions included in other chronic headaches. a Prescribed in the absence a diagnosis of depression or anxiety. b Please note that the IR/1000 for a broader health condition may not equal the total of the conditions included within them due to some individuals having more than one of the individual conditions, or because of rounding . Cumulative hazards plots for the health condition examined are presented in Figure 2a to Figure 2e and show the different patterns of the conditions over time by ethnic group. Some had very similar trajectories over the whole study period (rheumatological and orthopaedic conditions, cancer, irritable bowel syndrome and asthma); some showed ethnic divergence from the birth of the BiB child (eczema, allergic rhinoconjunctivitis); some conditions diverged in the early years (other skin, thyroid and gynaecological disorders, migraine/other headaches, diabetes); whilst others diverged between 7 – 10 years after recruitment (gallstones, hypertension, peripheral neuropathy, pelvis floor dysfunction and neuromuscular disorders). The increasing divergence in the rate of common mental health disorders in WE compared to SA women was apparent shortly after recruitment, whereas other mental health disorders had a similar trajectory in both ethnic groups until 2 – 3 years after recruitment before increasing more rapidly in WE women. The HRs for each health condition are shown in Figure 3 . Cancer (HR 0.53 [95%CI 0.40, 0.70]), other mental health disorders (HR 0.56 [95%CI 0.44, 0.70]), neuromuscular disorders (HR 0.63 [95%CI 0.49, 0.82]) and common mental health disorders (HR 0.69 [95%CI 0.64, 0.74]) were more common in White European mothers. South Asian mothers were at statistically significantly higher risk of developing 14 out of the 21 health conditions, with diabetes (HR 3.92 [95%CI 3.13, 4.91]) and chronic liver diseases (HR 2.98 [95%CI 2.29, 3.88]) have the largest hazard ratios.

This study describes the prevalence and incidence of 21 common health conditions in mothers participating in the Born in Bradford cohort before and after recruitment and explores differences in the risk of conditions in women of White European and South Asian heritage. Mental health and atopic disorders were the most common health conditions in both ethnic groups. The risk of developing 14 of the 21 health condition was statistically significantly higher in South Asian mothers, with diabetes and chronic liver disease having the greatest difference, at four and three times higher compared to White European women. White European mothers were at higher risk of developing common and other mental health disorders, neuromuscular disorders, and cancer. Mental health disorders . The risk of common and other mental health disorders in South Asian women was 31% and 44% lower respectively than in White European women. Previous research on women participating in BiB found that although minority ethnic women self-report similar levels of psychological distress as White British women, they are less likely to access treatment and therefore have a GP-recorded diagnosis 9 . It is therefore likely that the incidence of common mental health disorders in South Asian women is underestimated using health service data. An alternative explanation may be that stronger social networks evidence in South Asian women may protect against mental ill-health by provision of financial, social or emotional support 10 . Diabetes mellitus . We found that women of South Asian heritage were almost four times more likely to develop diabetes compared to White European women. This is comparable to other recent estimates reported in people aged between 40–69 11 , though the mean age of women by the end of our study was 40 years, suggesting that differential risk starts young and may yet be higher. Chronic liver diseases . The development of chronic liver disease in South Asian-heritage women was three times higher than in White European participants, predominantly due to fatty liver disease. A recent systematic review and meta-analysis found that South Asian men and women have more fat in the liver compared to their White European counterparts despite having similar BMI levels 12 . Although our primary care data did not categorically distinguish between the types of fatty liver disease, the high prevalence of non-alcoholic fatty liver disease (NAFLD) in South Asian-heritage people has been well documented and is associated with obesity, dyslipidaemia, diabetes, and hypertension 13 , the latter two of which are also reported as being higher in South Asian women in our study. Thyroid disorders . Women of South Asian heritage were at almost twice the risk of developing thyroid disorders compared to White European women, with hypothyroidism being the most common. Hypothyroidism can be caused by low iodine levels, and a study of almost 7,000 BiB mothers found that the urinary iodine-to-creatinine ratio was lower in Pakistani compared to White European participants 14 . Another small multi-ethnic study of thyroid function in pregnancy found that women of South Asian heritage had higher levels of serum thyroid stimulating hormone (TSH) compared with ethnic Europeans despite having similar levels of iodine deficiency; a higher prevalence of subclinical hypothyroidism was also observed 15 . Gynaecological disorders . We also found that women of South Asian heritage had higher rates of uterine fibroids, infertility and polycystic ovarian syndrome compared to White European women but lower rates of endometriosis. Systematic reviews have reported that black ethnicity was a strong risk factor for uterine fibroids 16 although black women are less likely to be diagnosed with endometriosis 17 . There is limited literature on the prevalence of these disorders in South Asian women apart from a small US-based study that found South Asian women had a similar prevalence of uterine fibroids as white women 18 . The observed differences in the incidence of most health conditions between ethnic groups can likely be attributed to a combination of social, cultural, genetic, and lifestyle factors 19 – 21 . There is a clear link between socioeconomic deprivation and poor health, and data from England, including our own cohort, indicate that ethnic minorities experience higher deprivation compared to the White population 19 , 21 . Differences in lifestyle behaviours may also contribute to these disparities. While people of South Asian heritage are often less likely to engage in risky behaviours such as smoking and excessive alcohol consumption, studies indicate that they are more likely to be physically inactive, have a poor diet, and experience higher rates of obesity 22 . Genetic and epigenetic factors may also underlie the risk of some health conditions, such as diabetes (including worse diabetes-related outcomes in South Asian compared to White European people), cardiovascular health conditions, and insulin resistance 23 . There is also evidence that people of minority ethnic groups have a higher risk of allergic conditions and autoimmune disorders 24 , with the onset of these condition occurring earlier in non-White compared to White groups 25 . Health inequalities are also evident with regards to health access, with ethnic minority patients reporting poorer experience of healthcare services compared to White British groups 22 . The uptake of population health screening is also lower in ethnic minorities and those living in deprived areas 26 : screening rates for breast, cervical and bowel cancer are lower in South Asian. This lower uptake likely extends to participation in other health improvement programmes, such as those for diabetes management 22 . Strategies focusing on improving health literacy and ensuring cultural sensitivity and appropriateness in the healthcare setting have been recommended in an attempt to address these barriers 22 , 26 , 27 . Using electronic health records, this study provides a comprehensive assessment of the prevalence, incidence, and ethnic differences in risk for 21 non-communicable health conditions in a bi-ethnic sample of 9,784 women enrolled in a longitudinal cohort study. A key strength is the large proportion of South Asian women (52%), which allowed us to identify health inequalities in some conditions that have not previously been widely explored or reported. While to the proportion of South Asian heritage women in the BiB cohort (50%) is higher than their estimated proportion in the general Bradford population (around 22% at the time of enrolment to the BiB study), this reflects the obstetric population attending Bradford Royal Infirmary during the recruitment period, suggesting it is representative of the pregnant women in the area at the time. To accurately determine prevalent conditions, we implemented a minimum requirement of two years of medical records prior to recruitment. This criterion was based on recorded GP events, which might not perfectly align with continuous registration with GP practice and could have introduced some bias in our prevalence estimates. While exploring the impact of area deprivation on incidence would have been valuable, the numbers if event within specific deprivation strata was insufficient for meaningful analysis. Furthermore, while it would have been preferable to present multivariable hazard ratios adjusted for known lifestyle factors such as physical activity and diet, we unfortunately lacked sufficient information on these potential confounders. We acknowledge that these unmeasured lifestyle factors could influence disease risk and incidence rates, potentially leading to an over- or under-estimation of the impact of ethnicity on health disparities. However, it is important to note that this study is primarily descriptive, aiming to highlight the differences in health conditions between these two groups in the BiB cohort and inform future research directions, rather than to investigate the underlying causes of these differences. We were also limited to studying broader ethnic groups of White European and South Asian due to the small numbers within finer ethnic categories, which may have obscured important intra-group differences. Finally, as this study exclusively examines postpartum women in Bradford, it’s generalisability to other UK populations and to men is limited. However, they are likely to represent parous women in similar urban populations with high levels of deprivation and diversity.

In conclusion, our findings reveal the early emergence and distinct ethnic patterns of non-communicable health conditions in women from White European and South Asian heritage. The impacts of these health conditions on health services may be greatest in later adulthood, but the opportunity for prevention is greatest in early life. Tackling upstream modifiable risk factors at an individual level (diet, physical activity, smoking) and at a systems level (environment, housing, education, urban design) remains a crucial but neglected priority for improving long term health outcomes and reducing health service pressures 28 .