Patterns and predictors of domestic violence and abuse enquiry in South East London maternity settings: Cross-sectional analysis of routine electronic health record data collected between 2019 and 2023

A cross-sectional analysis of over 7,000 maternity appointments in South East London reveals that despite national guidelines recommending universal domestic violence and abuse (DVA) screening, enquiry rates were inconsistent and significantly influenced by demographic and socioeconomic factors, suggesting midwives selectively prioritize screening based on perceived risk rather than adhering to routine protocols for all patients.

The Gist

Imagine a maternity ward as a busy train station where every pregnant person is supposed to get a "safety check" before their journey begins. In the UK, the rulebook (called NICE guidelines) says that midwives should ask every single person a specific, sensitive question: "Are you currently experiencing domestic violence or abuse?" This is meant to be a universal safety net, like checking the brakes on every car before it leaves the garage.

However, a new study from South East London looked at the actual logs from 2019 to 2023 to see if this rule was being followed for everyone, or if some people were getting skipped.

Here is what they found, broken down simply:

1. The "Safety Net" Had Holes

Out of nearly 8,000 first appointments, midwives asked the question about 79% of the time. That sounds like a lot, but it means one in five pregnant people were never asked, even though the rules say they should be.

It's like a teacher who is supposed to check every student's homework but only checks the homework of students who look like they might need help, or only checks when the classroom is quiet.

2. The Pandemic Was a "Storm"

The study covered the years of the COVID-19 pandemic. During the lockdowns, the rate of asking these questions dropped significantly.

• Why? Many appointments moved to video calls or phone calls. Midwives worried that if a partner or family member was in the room (or on the other end of the line), it wasn't safe to ask such a personal question.
• The Result: The "safety check" became less frequent during the storm, though it seemed to bounce back a bit by early 2023.

3. Who Got Asked? (The "Risk" Filter)

The most interesting part of the study is who got asked and who didn't. The midwives didn't ask everyone randomly; they seemed to be subconsciously (or consciously) filtering people based on who they thought was "at risk."

People who were MORE likely to be asked:

• Black women and those born in Sub-Saharan Africa: They were asked more often than White women or those born in the UK.
• Women living in poorer neighborhoods: If the area was more deprived, the question was asked more often.
• Single women: Women who were not married or living with a partner were asked more often.
• Women living alone (but with support): Even if they had help, living alone made them a target for the question.
• Women with unplanned pregnancies: If the pregnancy wasn't planned, they were asked more.

People who were LESS likely to be asked:

• Women needing an interpreter: If a woman didn't speak English and needed a translator, she was less likely to be asked. This is like a mechanic refusing to check the engine because they can't talk to the driver.
• Women born in North America: They were asked less often.
• Teenagers: Younger women (under 18) were less likely to be asked.
• Women whose partners had mental health issues: Surprisingly, these women were asked less often.

4. The Big Takeaway: "Guessing" vs. "Checking"

The authors suggest that midwives aren't following the "ask everyone" rule. Instead, they are acting like security guards at a club who decide who to check based on how they look or where they are from.

If a midwife thinks, "This woman looks like she might be in danger," they ask. If they think, "She looks safe," or "It's too complicated to ask right now" (like with a language barrier), they don't.

The paper argues that this is a problem because:

1. Bias: It relies on the midwife's gut feeling, which can be wrong or biased.
2. Missed Opportunities: The women who are actually at the highest risk (like those with language barriers or those in abusive relationships who are trying to hide it) are the ones getting skipped.

Summary

The study found that while the rule is to ask everyone about domestic violence, in practice, midwives are picking and choosing based on a mix of practical problems (like language) and their own assumptions about who is at risk. This means the safety net is catching some people but letting others slip through the cracks.

Technical Summary

Technical Summary: Patterns and Predictors of Domestic Violence and Abuse Enquiry in South East London Maternity Settings

Problem Statement
Domestic violence and abuse (DVA) is a significant public health issue that can escalate during the perinatal period, increasing risks of adverse pregnancy outcomes and maternal/neonatal mortality. While the UK National Institute for Health and Care Excellence (NICE) recommends routine, universal enquiry about DVA during all maternity booking appointments, evidence suggests this is not consistently practiced. Barriers include limited time, lack of private space, and potential conscious or unconscious bias regarding which patients are perceived as "at risk." Furthermore, the impact of the COVID-19 pandemic on DVA enquiry rates remains unclear, as healthcare delivery shifted toward virtual consultations and restricted partner attendance. This study addresses the gap in understanding the actual rates of DVA enquiry and the specific demographic, socioeconomic, and clinical predictors that influence whether a midwife asks a pregnant individual about current DVA experiences.

Methodology
The study employed an observational cohort design using cross-sectional data extracted from routine electronic health records (EHR) via the BadgerNet system at St Thomas' Hospital in South East London. The dataset covered the period from January 1, 2019, to March 31, 2023, encompassing pre-pandemic, pandemic, and post-pandemic periods.

• Cohort: The analysis included 7,932 booking appointments for 7,007 unique women.
• Outcome Variable: DVA enquiry was dichotomized into "enquiry made" (responses of yes, no, or declined) versus "no enquiry made" (recorded as "unable to ask").
• Predictors: Variables included demographics (age, ethnicity, country of birth), socioeconomic factors (deprivation index, housing, employment), social factors (marital status, support status), and reproductive/health indicators (mental health, smoking, alcohol use, pregnancy history).
• Statistical Analysis:
  • Missing data were handled using multiple imputation by chained equations (MICE) with 20 iterations.
  • Linear mixed modeling (binomial generalized linear mixed models) was used to test associations between predictors and DVA enquiry, adjusting for booking date and midwifery team as a random effect.
  • Temporal trends were analyzed using cubic spline modeling to account for non-linear fluctuations over time.
  • Feature selection via elastic-net penalized regression was performed to identify the most relevant variables for a final multivariable model, addressing multicollinearity among predictors.

Key Contributions

• First European Investigation: This is the first study in Europe to utilize routine EHR data to comprehensively examine patterns and predictors of DVA enquiry in maternity settings.
• Objective Data Source: By using recorded clinical data rather than self-reports from clinicians, the study avoids reliance on perceived barriers or self-reported adherence to guidelines, providing a more objective view of practice patterns.
• Pandemic Impact Analysis: The study specifically isolates the effect of the COVID-19 pandemic on enquiry rates, a period where clinical workflows were significantly disrupted.

Results

• Overall Enquiry Rate: DVA enquiry occurred in 79.4% of booking appointments. However, the rate was not uniform, fluctuating significantly over time and dropping to as low as 40–60% during certain weeks.
• Temporal Trends: There was a gradual decline in enquiry rates over the study period. The pandemic period was associated with a significant reduction in enquiry compared to the pre-pandemic period (OR=0.87), with a further decline in the post-pandemic period (OR=0.70). Rates appeared to recover to pre-pandemic levels by late 2022/early 2023.
• Demographic Predictors:
  • Increased Enquiry: Black-identifying women (OR=1.28) and those born in Sub-Saharan Africa (OR=1.37) or Southern Europe (OR=1.34) were more likely to be asked than White or UK-born women.
  • Decreased Enquiry: Women born in North America were less likely to be asked (OR=0.59). Adolescent pregnancies (<18 years) were less likely to be asked (OR=0.50).
• Socioeconomic and Social Predictors:
  • Increased Enquiry: Women living in more deprived neighborhoods, those in rented accommodation or with no fixed abode, and single women were more likely to be asked.
  • Decreased Enquiry: Women requiring an interpreter were less likely to be asked in multivariable models (aOR=0.73).
• Multivariable Findings: After adjustment, women born in Sub-Saharan Africa or Southern Europe, and those living alone but with additional support, remained more likely to be asked. Conversely, women born in North America and those requiring an interpreter remained less likely to be asked. Ethnicity as a broad category lost significance in the multivariable model, suggesting specific regional origins were the driving factors.

Significance and Claims
The paper claims that despite national guidelines recommending universal enquiry, a substantial proportion of pregnant individuals are not asked about DVA. The findings suggest that midwives may be consciously or unconsciously prioritizing enquiry for women they perceive to be at greater risk (e.g., those from specific ethnic backgrounds, living in deprivation, or single), while potentially overlooking others (e.g., those with language barriers or from North America).

The authors conclude that the observed patterns reflect both practical barriers (such as language and interpreter availability) and biases regarding risk perception. The study highlights a critical need for system improvements and training to ensure that DVA enquiry is truly universal and not contingent on a patient's demographic profile or the midwife's assessment of risk. The authors modestly note that while the study identifies these disparities, further qualitative research is required to understand the specific clinical decision-making processes and barriers faced by staff.