Connect with us

Insight

Automating inequality: When AI undervalues women’s care needs

Published

on

By Morgan Rose, chief science officer at Ema

Artificial intelligence is supposed to make care smarter, faster, and fairer, but what happens when it quietly learns to see women as less in need?

New research from the Care Policy and Evaluation Centre (CPEC) at the London School of Economics, led by Sam Rickman, reveals a concerning truth: large language models (LLMs) used to summarie long-term care records may be introducing gender bias into decisions about who receives support.

The Study

Researchers analysed real case notes from 617 older adults receiving social care in England. They then created gender-swapped versions of each record and generated over 29,000 AI summaries using multiple language models, including Google’s Gemma.’

The goal was simple: would AI treat men’s and women’s needs the same way?

It didn’t.

The Results

  • Google’s Gemma model consistently downplayed women’s physical and mental health issues compared to men’s.
  • Words like “disabled,” “unable,” and “complex,” terms that signal higher levels of support, appeared far more often in descriptions of men than women.
  • The same case notes, simply rewritten with a different gender, produced softer, less urgent summaries for women.

In other words, when the algorithm rewrote her story, her needs shrank.

The Cost of Softer Language

Language isn’t neutral. In healthcare, it’s the difference between monitor and act.

Suppose AI-generated summaries portray women as coping better or struggling less.

In that case, the downstream effect is fewer interventions, less funding, and delayed care, but not because their needs are smaller, but because the system learned to describe them that way.

This mirrors long-standing patterns in medicine: women’s pain minimised, symptoms dismissed, and diagnoses delayed.

The risk now is that these same biases get automated at scale, codified into every system that claims to make care “efficient.”

Why This Matters for Femtech

Femtech founders, clinicians, and AI builders have a responsibility to notice what’s hiding in the data.

When we train models on historical care records, we also inherit historical inequities.

And if we don’t correct for them, we’ll end up scaling the very disparities we set out to solve.

At Ema, we build for women’s health with this reality in mind:

  • Language is clinical data. Every word shapes care pathways.
  • Bias is not neutralised by scale. It’s magnified by it.
  • Ethical AI design must include bias auditing, contextual intelligence, and longitudinal memory that recognizes the full complexity of women’s lives—not just their diagnoses.

The Path Forward

Fixing this isn’t about scrapping AI.

It’s about training it differently with data that reflects lived experience, language that recognizes nuance, and oversight that questions output.

Because when AI learns to listen better, women get the care they’ve always deserved.

Source:

Rickman, S. et al., AI tools risk downplaying women’s health needs in social care, Care Policy and Evaluation Centre (CPEC), London School of Economics, 2025.

Insight

Working from home linked to higher fertility, research finds

Published

on

Working from home is linked to 0.32 more children per woman when both partners do it at least once a week, research across 38 countries suggests.

The study found that among working adults aged 20 to 45, estimated lifetime fertility, meaning children already born or fathered plus plans for future children, rises when one or both partners work remotely.

In the US, the increase was even higher at 0.45 children per woman.

On average, women whose partners did not work from home had 2.26 children.

When the woman worked from home at least one day a week, this rose to 2.48. When both partners did so, it increased to 2.58.

If the man worked from home at least one day a week, the increase was more limited at 2.36 children.

The research, by Steven J. Davis and colleagues and published as a working paper by the National Bureau of Economic Research, points to three possible explanations.

Remote working may make it easier to balance childcare with paid work, leading some couples to have more children.

Families with children may also be more likely to look for remote roles. Or the growing availability of those roles may lift fertility by opening up more parent-friendly jobs.

“All three stories align with the idea that WFH jobs make it easier for parents to combine child rearing and employment,” the report suggests.

The pattern held both after the pandemic, between 2023 and 2025, and before it, between 2017 and 2019.

The implications for national fertility rates vary mainly because working-from-home rates differ widely between countries.

Among workers aged 20 to 45, the share working from home at least one day a week ranges from 21 per cent in Japan to 60 per cent in Vietnam. The UK ranks third globally and leads Europe at 54 per cent.

The report estimates that, if “interpreted causally”, remote working accounts for 8.1 per cent of US fertility, equal to about 291,000 births a year as of 2024.

The researchers note that while this may sound modest, it is larger than the effect of government spending on early childhood care and education in the US.

“Bringing WFH rates to the levels that currently prevail in the United States, United Kingdom, and Canada has the potential to materially boost fertility in many other countries,” the report suggests.

However, the research cautions against broad policy approaches, saying the desire for remote work varies widely between individuals, and that it is not practical in every job or organisation.

“Thus, policy interventions that push for a one-size-fits-all approach to working arrangements are likely to yield unhappier workers and lower productivity,” it warns.

A UK Parliament report has also found that remote and hybrid work can boost employment, with parents, carers and people with disabilities likely to benefit most from more flexible working options.

Continue Reading

Insight

Radiotherapy may cut lymphoedema risk

Published

on

Radiotherapy to the armpit instead of surgery may cut lymphoedema risk in some breast cancer patients, early trial results suggest.

Lymphoedema is swelling of the arm or armpit that can happen after surgery to remove lymph nodes.

New findings suggest axillary radiotherapy may be as effective at killing any remaining cancer cells while being less likely to trigger this complication.

The results come from the pilot phase of a phase III randomised international clinical trial looking at whether axillary radiotherapy has a lower risk of lymphoedema than axillary lymph node dissection in breast cancer patients who have had chemotherapy or hormone therapy before surgery, and whose cancer has spread to only one or two lymph nodes.

The trial will also assess overall survival and disease-free survival.

The researchers stressed that these are preliminary results from two years of follow-up in the pilot study, and that clinicians should wait for results from the ongoing phase III trial before considering changes to clinical practice.

Amparo Garcia-Tejedor, from the Functional Breast Unit at Bellvitge University Hospital in Spain and the Institut Català d’Oncologia, is leading the trial.

She said studies had already shown that axillary radiotherapy was a good alternative to axillary lymph node dissection in patients whose first line of treatment was surgery.

She said: “In situations where patients have received chemotherapy or hormone therapy before surgery, it is expected that results could be similar. However, robust prospective data are not yet fully established or published.

“Many patients treated with neoadjuvant therapy experience a significant reduction in axillary disease burden and ultimately present with only one or two lymph nodes that are positive for cancer metastases, which often correspond to the sentinel lymph node, while the remaining axillary nodes are negative.

“This observation strongly suggested that further axillary surgery might be unnecessary in a substantial proportion of patients and that a strategy of de-escalation should be explored.”

From June 2021 to April 2023, the ADARNAT trial recruited 272 breast cancer patients whose disease might have spread to one or more lymph nodes.

The patients had received neoadjuvant therapy and, at the time of surgery, had metastatic cancer in one or two sentinel lymph nodes, the lymph nodes where cancer typically spreads first.

Patients were randomised to receive either axillary radiotherapy or axillary lymph node dissection, and patients in both groups also received radiotherapy to areas of the breast and chest. Results were available for 46 patients in the radiotherapy group and 56 in the surgery group, with a median follow-up of two years.

No cancer recurred in the axillary area in the radiotherapy group, compared with one recurrence in the surgery group, or 1.8 per cent.

Cancer spread to other parts of the body in 4.4 per cent of radiotherapy patients and 5.5 per cent of surgery patients, and there were two deaths in the surgery group, or 4.3 per cent.

Lymphoedema was more common after surgery, at 26.7 per cent, than after radiotherapy, at 18.9 per cent, although the researchers said this difference was not statistically significant. Disease-free and overall survival rates were similar after two years.

Garcia-Tejedor said: “These results indicate that ART instead of ALND is feasible and has good cancer outcomes at two years.

“While some specialists have already begun to substitute axillary lymph node dissection with axillary radiotherapy without waiting for definitive results, the only way to determine with certainty whether this strategy is truly safe and effective is through participation in a well-designed clinical trial such as the one we are now conducting.

“This is particularly important given that the study population includes patients with residual axillary disease and, therefore, a potentially worse prognosis.

“In this context, treatment decisions should not be made without robust evidence.

“Our trial is designed to provide the necessary data to definitively answer this question and to ensure that any future change in standard practice is safe in terms of cancer outcomes and is also beneficial for patients.”

Maria Laplana-Torres, a radiation oncologist at the Hospital Clínic de Barcelona, presented pilot-phase results showing that although axillary radiotherapy was linked to more skin damage from radiation, this was usually temporary and easily treated.

Acute skin damage of grade 2 or above occurred in 27.8 per cent of radiotherapy patients compared with 13.3 per cent after surgery.

It mainly involved skin redness, pigment changes or, in some cases, skin peeling. There were no significant differences in later skin damage between the two groups.

She said: “Some patients experienced mild, temporary difficulty raising the arm above the shoulder or lifting it to the side. These limitations were usually short-lived and did not affect everyday activities.

“We found that treating the axilla with radiotherapy instead of extensive surgery can avoid a more aggressive operation without compromising treatment safety in patients with sentinel lymph node involvement.

“One and two years after treatment, there were no meaningful differences in arm mobility or quality of life between the two groups, although there was a more favourable trend in the ART patients.

“These results show that axillary radiotherapy may be a safe and less invasive option for some women treated with chemotherapy or hormone therapy before surgery.’

“This kind of research is essential to continue improving patient outcomes and to define safer, equally effective therapeutic approaches.”

More than 500 patients have now joined the main phase III trial. The researchers estimate that about three more years will be needed to complete recruitment, followed by five years of follow-up to fully assess cancer outcomes.

Continue Reading

Insight

Report makes the case for an incentive change in health data

Published

on

In a new report, “The Case for Incentive Change in Healthcare Data,” WHIS Lead Producer Poppy Howard-Wall explores why healthcare’s biggest data challenge may not be technical but economic.

Integrating learnings from Poppy’s conversations with senior leaders at the ViVE Summit, the report highlights how fragmented data and misaligned incentives continue to limit the industry’s ability to deliver truly longitudinal care.

Howard-Wall writes: “For the women’s health industry, where many conditions have historically been under-researched and longitudinal datasets remain incomplete, the consequences of fragmented data infrastructure are even more pronounced.

“Artificial intelligence promises to accelerate discovery, improve diagnosis and enable more proactive care. But its potential is inseparable from the data ecosystems that support it.

“In the absence of strong economic incentives for deeper integration, the question becomes how the industry is beginning to navigate this constraint and what signals are emerging about the future of healthcare data and AI in women’s health.”

Read the report here.

Continue Reading

Trending

Copyright © 2025 Aspect Health Media Ltd. All Rights Reserved.