Rural-urban differences in the mental health of perinatal women: a UK-based cross-sectional study

Mental illness, most commonly depression and anxiety, affects up to 20% of women in the UK during pregnancy and in the first year after childbirth [1]. Understanding the reasons for this problem requires consideration of the social, economic and geographical environments in which mothers live, which can greatly differ within a given country. International data, including from high-income, middle-income and low-income countries, suggest that perinatal mental illness is more prevalent in rural women compared to their urban counterparts and to the general population [2, 3]. Other negative outcomes have been reported to be more frequent in women from rural areas including later initiation of prenatal care [4], later and less use of contraceptives [5], as well as higher chances of a small-for-gestational-age birth and low birth-weight [6], all of which may add to stress levels and increase the risk of mental illness. Possible mediators for these rural-urban differences include difficulties in accessing healthcare services, regional variations in cultural practices, sociodemographic and lifestyle factors [3, 6, 7], as well as varying levels of hazardous environmental, occupational and transportation conditions [8]. Social support and socioeconomic status are particularly important because they are related to many of the above factors [3, 9, 10].

In the UK context, rural-urban status has been found to be associated with behavioural and emotional problems in children, although this was explained by the quality of their schools [11]. Also relevant is the finding that in the UK increased neighbourhood deprivation has been found to predict poor maternal mental health outcomes [12], particularly amongst older mothers [13], and that regional differences in the rate of referrals to perinatal mental health services can vary up to 20% [14]. However, to the best of our knowledge no study has looked specifically at the association between rural-urban status and maternal mental health outcomes in the country, and much of the existing literature is based on exclusively rural or urban samples which hinders comparisons between the two groups.

The aim of the present study was to assess the association between rurality (i.e., living in a rural area) and the risk for two common mental illnesses (depression and anxiety) during the antenatal and postnatal periods, i.e. perinatal period, in women from a county in Northern England (UK).

Participant flow is shown in Fig. 1. A total of 295 participants were entered in the main analysis, all of whom had provided valid outcome data on the three measures and a valid postcode. Excluded participants did not differ significantly from others in age and perinatal stage; whether they differed in terms of IMD or urban/rural status cannot be determined due to missing postcodes. Approximately half of women heard about the study during a health visitor’s home visits (n = 137, 46.9%), others in antenatal or postnatal appointments at the hospital (n = 96, 32.9%), social media (n = 55, 18.8%), parenting groups (n = 3, 1.0%) or through other channels (n = 1, 0.3%).

The characteristics of study participants are presented in Table 1, by urban (n = 165, 55.9%) and rural (n = 130, 44.1%) groups, and overall (N = 295). There were no significant differences between urban and rural groups except for the perinatal stage; those from urban areas, compared to those from rural areas, were more likely to be currently pregnant (60.0% vs 47.7%, p = .035). Of those women who were pregnant, the median number of gestational weeks was 28 (IQR 15); of the postnatal women, the median number of weeks of the baby was 7 (IQR 12).

The median IMD decile was 7 for the whole sample (IQR 3), corresponding to being in the 40% least deprived geographical area (technically known as a Lower Layer Super Output Area, or LSOA) in the country. In the urban group the number of women in each of the three IMD decile groups was roughly similar, but in the rural group a considerably larger proportion of women were in the second tertile, although differences were non-significant between the two groups. The characteristics of the participants are broadly representative of those in the region where the study took place with regard to average IMD score, ethnicity, female educational qualifications and employment rate [34] .

Table 2 provides data on access to services and social support. No significant differences existed between urban and rural groups with respect to any of these variables. Over a fifth (21.4%) had looked or asked for support or information about feelings of anxiety and/or depression during pregnancy or since the birth of their child. Amongst these women, partners were the most common reported source of support/information.

Of a possible range of 4 to 28, perceived social support was highest with regards to a significant other (median 28, IQR 3), followed by support from family (median 26, IQR 4) and friends (median 25, IQR 6). Taking the sample overall, the median value for perceived social support was 78 (IQR 12), of a possible range of 12–84.

All models, across the three measures, showed a higher prevalence of mental illness in rural women (Table 3). In the unadjusted analysis (model 1), the prevalence of testing positive for depression and/or anxiety was 11.5% higher for rural compared to urban mothers, which was statistically significant (urban 28.5% vs rural 40.0%, p = .038). In all other models, differences though indicative were not statistically significant and confidence intervals were generally wide. For three of the four outcomes, adjusting for IMD and perinatal stage (model 2) was associated with a decrease in the odds ratio, which demonstrates that these variables explained some of the association as expected. For three of the four outcomes, the addition of social support (model 3) either led to increases in the odds ratio or to no change. The only exception was anxiety, where the odds ratio decreased slightly (from 1.24 to 1.16) in model 3, which indicates that social support had some explanatory value in the hypothesised direction.

In an exploratory analysis, we compared women in the two most urban categories (n = 18) to those in the two most rural categories (n = 23) (Table 4). Only the descriptive data of this analysis are presented due to the small sample size. Both groups were comparable in most variables except perinatal stage, consistent with the full sample analysis. Although women in the most rural group experienced less socioeconomic deprivation, they were also more likely to look for support/information about depression/anxiety, and to report lower levels of social support especially from friends.

To our knowledge, this is the first quantitative study that compared mental health outcomes between perinatal women from rural and urban areas in the UK. Overall, our findings suggest that women in rural areas are at higher risk of perinatal depression and anxiety than their urban counterparts, with the odds ratio varying between 1.16 and 1.72 across outcomes. After accounting for potential confounding factors, though indicative, findings of a positive association between rurality and risk of perinatal mental illness did not reach statistical significance.

Rural and urban groups were comparable in all variables except perinatal stage; women in the urban group were more likely to be pregnant, whereas those in the rural group were more likely to have given birth already. This is likely to be due to the fact that health visitors who assisted us with recruitment happened to be more involved in rural areas and in postnatal visits at the time of the study. Even after adjusting for perinatal stage, our data revealed a trend, though non-statistically significant, for higher propensity for depression and anxiety in rural women. This suggests that other factors may exist that account for those differences. Some of those possible factors may not have been captured in our survey and include, as previously noted, greater expenditure with transport, poorer transport infrastructure and longer travel times [35], fewer options for childcare provision (other than a children’s centre) [36] or to attend antenatal and postnatal groups. Additional reasons may be a lack of stimulating activities and distance to amenities such as shops, church/temple or college [37], or to the gym, cinema or coffee shops, which can facilitate social interaction [38]. As measured in our study, (perceived) social support appeared to contribute little by way of explanation in the main analysis. However, our exploratory analysis suggested that rural-urban differences in social support exist, namely that women in very rural areas lack support from friends and, to lesser extent, from family, though this sub-analysis was based on a small sample.

The percentage of women at risk of anxiety in this study (11.5 to 13.9%) is similar to that reported in a nationally-representative study conducted in England [39]. However, the risk of depression appears to be unusually high in our study. Based on the Whooley test, 28.5% were positive screens, whereas on the EPDS, that number was 22.4%. Estimated rates of maternal mental illness range widely, depending on the assessment method, the timing of the assessment, and population characteristics. A literature review has reported the prevalence of maternal depression to be as high as 20% antenatally and between 12 to 16% postnatally [40], whereas the National Health Service (NHS) in England estimates both depression and anxiety rates to be 15 to 20% [1]. It is important to reiterate that none of the tools used in this study, including the EPDS (due to the choice of cut-off), aimed to assess rates of depression and anxiety as such, but rather, identify how many women would be at higher risk. For example, it is well-known that the Whooley is highly sensitive and should only be pre-diagnostic [22]. Likewise, the number of positive EPDS screens (22.4%) was higher than the previous figures reported above, and this may be due to various reasons:

This study had several other limitations. The sample was relatively small and fell below our target sample of 348 which limited the power of our analysis. It consisted mainly of White-British women from an area of above average levels of affluence. As such, findings may not be generalisable to other parts of the UK or other countries. Although we assessed social support in various ways, we may still have missed some important aspects of this construct. The cross-sectional design of this study makes it impossible to make conclusions on causality. No information is available on the women who declined to take part. However, as discussed earlier, our sample compared well with the profile of the region where the study took place in all relevant sociodemographic variables [34], which suggests that no significant selection bias occurred.

The focus of this study was to compare rural and urban areas rather than estimating prevalence rates in each group. In addition, it is worth noting that both in the main and in the exploratory analysis more urban women lived in areas of higher socioeconomic deprivation (i.e., first tertile), although national data have shown that such geographical differences are not always clear-cut [35]. It is likely that financial and professional development support need to be part of efforts to increase the (mental) health of perinatal women in less affluent neighbourhoods, in addition to other forms of social support.

Future studies should aim to recruit a larger number of participants, ideally with matched rural and urban participants, to further explore whether those in rural areas are at greater risk of mental illness. This analysis will benefit from including participants in both ends of the spectrum, i.e., the most rural and most urban. Further quantitative and qualitative research is needed to investigate the contextual factors and potential mechanisms of such differences, e.g., role of friends, partners and extended family, or socioeconomic deprivation, to inform early intervention for mothers at risk of anxiety and depression.

This study suggested that perinatal women from rural areas are at higher risk of depression and anxiety than their urban peers. Although this was not statistically significant when key confounding variables were considered, the direction of this association was consistent across analyses. More work is needed to replicate these findings in a larger sample in order to better understand the interplay between rurality, social deprivation and accessibility to support. This is necessary to identify and understand any inequalities that may exist between rural and urban areas in terms of antenatal and postnatal services, so that the right support can be provided where it is most needed, both to mothers and children at such a critical period of their lives.