Residential and healthcare mobility during pregnancy among women living with HIV in the UK, 2009 – 2019

Introduction: The extent to which individuals living with HIV experience residential and healthcare mobility during pregnancy in the UK is unknown. We aimed to determine a minimum estimate of residential and healthcare mobility during pregnancy in people living with HIV in the UK in 2009 – 2019 to explore patterns of and factors associated with mobility and to assess whether mobility was associated with specific HIV outcomes. Methods: We analyzed data from the Integrated Screening Outcomes Surveillance Service to assess pregnancies with HIV in the UK and included livebirths and stillbirths with estimated delivery in 2009 – 2019. Residential mobility was defined as changing residential postcode between notification and delivery, and healthcare mobility was defined as changing NHS Trust or Strategic Health Authority (SHA) in that same timeframe. We used logistic regression to determine factors associated with residential and healthcare mobility and with detectable delivery viral load. Results: Among 10 305 pregnancies, 19.6% experienced residential mobility, 8.1% changed NHS Trust, and 4.5% changed SHA during pregnancy. Mobility was more likely to be experienced by younger women, migrants, and those with new antenatal diagnosis; residential but not healthcare mobility declined over time. In a fully adjusted model, mobility was not associated with having a detectable viral load at


INTRODUCTION
Around 900 pregnancies occur in people living with HIV in the UK annually, but little is known about the extent to which this population experiences residential and/or healthcare mobility (i.e., changing home address or healthcare provider) during pregnancy.However, around 82% of pregnant women with HIV are migrants, mostly from sub-Saharan Africa but with an increasing proportion from Eastern Europe, including some arriving in the UK while pregnant and who may experience additional challenges in accessing timely healthcare [1][2][3].
Residential mobility, which could result in changing healthcare providers, is common among young adults and may occur more frequently in pregnancy or after giving birth [4].Data on mobility in pregnancy and its impact on health outcomes in general are limited, although around 10%-15% of pregnant women in the UK change residence [4,5].Relocation can be a stressful life event, particularly if moving country, with potential acute disruption to healthcare access, continuity of care, health behaviours, and social support networks alongside possible longer-term impacts [6][7][8].Mobility in pregnancy may thus be associated with adverse parental and child outcomes, particularly in the presence of chronic conditions, such as diabetes or HIV, where optimum management requires regular monitoring and/or sustained treatment [6][7][8].
No research has been conducted to date in the UK to investigate residential or healthcare mobility during pregnancy in the context of HIV.Our objectives were to determine a minimum estimate of the proportion of people living with HIV in the UK experiencing residential and healthcare mobility during pregnancy in 2009-2019 to explore patterns of and factors associated with mobility and to assess whether mobility was associated with key outcomes, including delivery with detectable HIV RNA viral load (VL) and loss-to-care (LTC) among infants exposed to HIV.

MATERIALS AND METHODS
The Integrated Screening Outcomes Surveillance Service (ISOSS) is commissioned by NHS England as part of the NHS Infectious Diseases in Pregnancy Screening Programme to conduct comprehensive population-based active surveillance of all pregnancies with HIV diagnosed in the UK since 1989 (formerly the National Study of HIV in Pregnancy and Childhood) [1].Since 2020, ISOSS is an England-only service.Reports of all pregnancies in people living with HIV diagnosed by delivery are submitted to ISOSS by maternity and paediatric NHS respondents at each unit nationally via a secure online portal, including data on sociodemographics, treatment, pregnancy management, delivery, and outcomes.Data are collected under legal permissions granted under Regulation 3 of the Health Service (Control of Patient Information) Regulations 2002 [2].
Data on gender have been routinely collected since April 2023, with no reports of pregnancies to non-binary or transgender people to date.Therefore, the data presented in this paper only refer to 'women' in the context of pregnancy.
Analyses were restricted to pregnancies ending in livebirth or stillbirth with an estimated date of delivery between 1 January 2009 and 31 December 2019; all had known residential partial postcode and hospital at notification and delivery.Residential mobility was defined as changing residential postcode between ISOSS notification and delivery and/or arrival of non-UK-born woman during pregnancy.Two healthcare mobility groups were defined using the same interval: changing NHS Trust and changing Strategic Health Authority (SHA); changing SHA was restricted to England deliveries (n = 9759), since SHAs are within NHS England only.Pregnancies where women arrived in the UK while pregnant were counted as 'mobile' cases for healthcare mobility.'Delivery VL' results were those measured from 36 weeks' gestation to 7 days postpartum.Children were classified as LTC if their HIV status was indeterminate and they were older than 3 years or if reported as LTC by respondents.
We used logistic regression models to identify factors associated with mobility; initial models adjusted for delivery year, and a multivariable model adjusted for maternal region of birth, age group, parity, and HIV diagnosis timing.Further logistic regression models were used to explore whether mobility was associated with detectable delivery VL (>50 copies/mL); adjusted models included delivery year, maternal birth region, maternal age group, timing of antiretroviral treatment (ART) start, and mobility (model 1: residential, model 2: NHS Trust, model 3: SHA).In a sensitivity analysis, residential mobility was included in models 2 and 3, and healthcare mobility was included in model 1.Data were analysed using STATA (StataCorp, v17.0,College Station, TX, USA).

Descriptive analysis
Among the included 10 305 pregnancies, 84.4% (8695/10305) of mothers were born outside the UK (of whom most [7331/8584] were born in sub-Saharan Africa).Median age at delivery was 33 years (interquartile range [IQR] 29-37) and in two-thirds of pregnancies (4935/7202; 68.5%) the woman had experienced at least one previous pregnancy (ranging from 1 to 7).Most pregnancies were conceived on ART, although 3744 (37.4%) women ART started antenatally (380 in the first trimester, 2955 in the second trimester, and 409 in the third trimester), including 1674 with an antenatal HIV diagnosis.There were 89 stillbirths and 10 216 livebirths.
Residential mobility was identified in 2022 (19.6%) pregnancies, including 225 (3.7%) pregnancies where a non-UK-born mother arrived in the UK during pregnancy.A change in NHS Trust occurred in 8.1% (834/10 305) of pregnancies, including 237 where hospital country (i.e., England, Scotland, Wales, and Northern Ireland) changed between notification and delivery.Among England deliveries, the SHA changed between notification and delivery in 4.5% (443/9759) of pregnancies.Overall, 448 women changed both postcode and NHS Trust; of the women who changed postcode, 22.2% (448/2022) also changed NHS Trust, whereas 46.3% (386/834) of women changing NHS Trust did not have a residential move (Figure S1).

Factors associated with residential and healthcare mobility
Maternal characteristics of mobile and non-mobile women are presented in Table 1.Maternal age (younger), region of birth (non-UK), and lower parity were associated with both residential and healthcare mobility, as was timing of arrival in the UK for migrant women.A higher proportion of mobile than non-mobile women received their HIV diagnosis during the pregnancy, started ART later in pregnancy, and had a detectable delivery VL.
In analyses adjusted for delivery year only, maternal age, region of birth, parity, and timing of HIV diagnosis and ART start were significantly associated with all three mobility types (data not shown).In the fully adjusted models, delivery year was associated with decreased odds of residential mobility and increased odds of healthcare mobility (Table 2).Older women had lower odds of residential mobility, but maternal age was not associated with changing NHS Trust or SHA during pregnancy.Women born outside the UK were more likely to experience any mobility than were UK-born women; strengths of association were greatest for healthcare mobility, particularly change in SHA, with women born in sub-Saharan Africa having 4.6 (95% confidence interval [CI] 2.70-7.86)times the odds of changing SHA than UK-born women.Women who had previously given birth were less likely to have residential and healthcare mobility than were nulliparous women.Antenatal HIV diagnosis was associated with two-fold increased odds for changing Trust and nearly three-fold increased odds for changing SHA. ) were more likely to deliver with a detectable VL than women who did not change; however, this association was not significant for residential mobility (AOR 1.16 [95% CI 0.98-1.37];p = 0.0811).Further investigation of associations between mobility in pregnancy and detectable delivery VL, adjusting for year, maternal age, origin (born in or outside UK), and timing of ART start, showed that no mobility types were associated with this outcome (Table S2).Results from sensitivity analyses including both types of mobility did not change from the simpler model presented (data not shown).

DISCUSSION
About one in five pregnancies from 2009 to 2019 in people living with HIV in the UK experienced residential mobility, and 8% and 4% changed NHS Trust and SHA, respectively.All mobility outcomes were associated with younger age, consistent with results from other studies [4,5,9,10] and being born outside the UK (i.e., migrants).Residential mobility significantly declined over time, albeit with a temporary increase in 2014 and 2015, the reasons for which are uncertain; as our adjusted analysis showed, this decline was independent of maternal age, parity, region of birth, and timing of HIV diagnosis.Meanwhile, healthcare mobility remained stable for much of the study period.Over 85% of pregnancies were delivered in the context of a suppressed VL, reflected in the very low vertical transmission rate of 0.3%.A significantly higher proportion of women with residential or healthcare mobility had detectable delivery VL, although mobility was not associated with this outcome in adjusted analyses, when taking account of timing of ART start and maternal age, both of which were risk factors.These findings suggest that, although mobility does not seem to increase the risk for delivering with a detectable VL after considering other risk factors, healthcare providers could still use residential and healthcare mobility to identify women at potentially higher risk for delivering with detectable VL.In the current era of very low vertical transmission rates, identifying those at higher risk of transmission is a key priority.Investigation of the few cases of vertical transmission now occurring in the UK by an expert clinical review panel has identified unstable housing as a contributing factor in several cases, alongside other social issues [1].
Minimizing the number of HIV-exposed infants LTC before diagnosis is an important aspect of HIV prevention and care [1,11].Although numbers of infants LTC were small, our findings suggest that women who experience mobility while pregnant may need additional support and monitoring postnatally.
Most pregnancies in the UK with HIV diagnosed before conception are to migrants, a group potentially at increased risk for adverse HIV outcomes [1,12].Our findings highlight the importance of enhanced follow-up of pregnant migrants to ensure continued care throughout pregnancy and postnatally, especially as the UK's migrant dispersal policy may contribute to mobility [13], although prevention of vertical HIV transmission is listed as a priority in the Healthcare Needs and Pregnancy Dispersal Policy [14].
Limitations of this study include our inability to examine the impact of gestational age at the time of residential move, which may be important as mobility in the first trimester has been associated with increased risk of adverse birth outcomes across all socioeconomic strata [15,16].Misclassification of mobility 'exposure' may have occurred if a move happened before booking.We were also unable to ascertain reasons for moving (which could be positive or negative) and whether residential postcode and healthcare facility were changed more than once, though women with more mobility could be at higher risk for disrupted HIV care.We were also unable to investigate whether antenatal and/or postnatal appointments were missed more frequently in the mobile group.A further limitation was the high level of missingness for some variables, including delivery VL.
Despite these limitations, this analysis benefits from population-based national surveillance data, representative of all pregnancies with HIV diagnosed in the UK and fills a knowledge gap around residential and healthcare mobility during pregnancy.Future analyses will be able to draw on data on social circumstances now collected within ISOSS.However, specific additional studies (potentially using mixed methods) are needed to gain a better understanding of the impact of mobility on health service use and outcomes in this population.Future research could use a similar framework to examine mobility among pregnant women living with hepatitis B or untreated syphilis, and associated outcomes.

T A B L E 1
Descriptive characteristics of pregnant women living with HIV in the UK by mobility status, 2009-2019.
T A B L E 2 Characteristics associated with residential and healthcare mobility among pregnant women living with HIV in the UK, 2009-2019.Model 1: residential mobility as the outcome, adjusted for year of delivery, maternal region of birth, maternal age group, parity, and timing of HIV diagnosis.NHS trust as the outcome, adjusted for year of delivery, maternal region of birth, maternal age group, parity, and timing of HIV diagnosis.d Model 3: changed Strategic Health Authority as the outcome, adjusted for year of delivery, maternal region of birth, maternal age group, parity, and timing of HIV diagnosis.
a Row percentages among pregnancies (leading to livebirths and stillbirths) with information on the sociodemographic measure and mobility outcome.Pregnancies with missing information are excluded from the denominator.b c Model 2: changed