The study, led by Muhammad Khairul Alam and Rumana Rois, draws on the Bangladesh Demographic and Health Survey of 2022, a nationally representative household survey that collects detailed information on fertility, contraceptive use, and household characteristics. Rather than treating unmet need as a single undifferentiated category, the researchers separated it into its two distinct components: the need for spacing births, which applies to women who want to wait at least two years before their next pregnancy, and the need for limiting family size, which applies to women who want no more children. This distinction matters because the two groups face different barriers, require different counseling approaches, and respond to different kinds of contraceptive services.

The analytical strategy unfolded in two stages. First, the team built a conventional multivariable multinomial logistic regression model, the workhorse method for estimating how demographic and socioeconomic factors shift the odds of belonging to one category rather than another. This model produced adjusted odds ratios, the familiar risk profiles that program planners can act on directly. Second, the researchers trained two ensemble machine learning classifiers, Random Forest and XGBoost, on the same data to see whether these flexible algorithms could capture non-linear relationships and complex interactions between variables that a regression equation would smooth over or miss entirely.

Working with survey data of this kind presents a well-known technical challenge: severe class imbalance. Most women in the survey either use contraception or have no need for it, so the categories representing unmet need contain relatively few observations. Machine learning classifiers trained on imbalanced data tend to simply predict the majority class, achieving deceptively high accuracy while failing to identify the minority cases that matter most. To counter this, the researchers applied a synthetic hybrid resampling technique based on SMOTE, the Synthetic Minority Over-sampling Technique, which generates synthetic examples of the underrepresented classes to balance the training data and give the algorithms a fair chance to learn the rarer patterns.

Feature selection posed another methodological question: with dozens of potential predictors, which ones genuinely carry signal? The team used the Boruta algorithm, a wrapper method built on Random Forests that compares the importance of real variables against shuffled, randomized versions of themselves. Only features that consistently outperform their random shadows are retained, providing a conservative filter against noise. To interpret what the trained models had actually learned, the researchers turned to Shapley Additive Explanations, or SHAP, a technique borrowed from cooperative game theory that assigns each feature a contribution value for every individual prediction, making it possible to visualize both global patterns and local interactions across the entire dataset.

The bivariate analysis revealed significant associations between unmet need and several key characteristics: maternal age, parity, household wealth, urban or rural residence, and the woman’s decision-making power within the household, all at the conventional threshold of statistical significance. The multivariable regression then sharpened these findings into striking effect sizes. Women with four or more children faced dramatically elevated odds of unmet need, with an adjusted odds ratio of 20.60 for spacing and 2.24 for limiting. In other words, high-parity mothers who still wanted to space a future birth were more than twenty times as likely to be without contraception as their lower-parity counterparts, a signal that current family size alone does not capture the full complexity of reproductive intentions.

Wealth emerged as a protective factor, particularly for limiting needs. Women in the richest household quintile had an adjusted odds ratio of 0.48 for unmet need for limiting, meaning their odds were roughly half those of the poorest women. This gradient points to the role of economic access: wealthier households can more easily reach clinics, pay for methods, absorb the indirect costs of seeking care, and negotiate the logistics of obtaining supplies. The SHAP analysis added exploratory depth, suggesting non-linear interactions between residence and factors such as parity and education, hinting that the barriers facing rural women may operate differently depending on how many children they have and how much schooling they received.

Yet the study delivers an honest verdict on the machine learning component. The predictive accuracy of the Random Forest and XGBoost models was only moderate, with macro-averaged areas under the receiver operating characteristic curve comparable to those achieved by traditional regression. The sophisticated algorithms did not outperform the simpler statistical model in ways that would justify relying on them for population-level inference. The authors therefore conclude that conclusions about the population should be drawn primarily from the multinomial logistic regression, while the SHAP-based insights into interactions remain exploratory rather than definitive. This is a valuable caution for a field where machine learning is sometimes presented as a wholesale replacement for classical epidemiological methods.

The cross-sectional design of the survey imposes further limits. A single snapshot in time cannot establish causation, and unmet need for spacing is often a fluid, transient state shaped by relationships, recent births, and shifting intentions. A woman classified as having an unmet need today may adopt a method next month or decide she wants another child soon. The researchers argue that this fluidity means programs should not treat the unmet need gap as a fixed deficit to be closed by a single intervention, but as a dynamic condition that requires continuous, responsive contact with women across their reproductive lives.

The practical implications are concrete. The authors recommend moving away from broad media campaigns toward localized, face-to-face counseling that addresses the specific circumstances of each woman. They call for ensuring a stable supply of spacing methods in rural areas, where stockouts and distance can quietly push women into the unmet need category. Perhaps most strikingly, they emphasize involving husbands and mothers-in-law in family planning conversations, reflecting evidence that reproductive decisions in Bangladesh are frequently negotiated within the household rather than made by women alone. Turning contraception from a private, sometimes contested choice into a shared household conversation, the study suggests, may be the key to closing a gap that decades of national programs have not yet eliminated.

Subject of Research: Unmet need for spacing and limiting family planning in Bangladesh, analyzed with regression and machine learning on national survey data

Article Title: Factors associated with the unmet need for spacing and limiting family planning: Evidence from a national survey

Article References: Alam, M. K., & Rois, R. (2026). Factors associated with the unmet need for spacing and limiting family planning: Evidence from a national survey. PLOS Global Public Health, 6(10), e0007412. https://doi.org/10.1371/journal.pgph.0007412

Image Credits: AI Generated

DOI: 10.1371/journal.pgph.0007412

Keywords: family planning, unmet need, Bangladesh, Demographic and Health Survey, contraception, machine learning, Random Forest, XGBoost, SHAP, multinomial logistic regression, SMOTE, reproductive health

The research, conducted by Zlatko Nikoloski and Elias Mossialos of the London School of Economics and Political Science together with Aletto Koetlisi of UNICEF Lesotho, delivers the first comprehensive assessment of universal health coverage for the country. Universal health coverage, the cornerstone target of the United Nations Sustainable Development Goals in health, rests on two pillars that must be evaluated together: people must be able to obtain the services they need, and using those services must not push them into financial hardship. Most national and global monitoring frameworks lean heavily on the second pillar, tracking catastrophic health expenditure, typically defined as household health payments exceeding 10 or 25 percent of total household consumption or income. On that metric, Lesotho looks remarkably good.

Using data from the 2023-24 Demographic and Health Survey and the 2017-18 Household Budget Survey, the researchers calculated that just 0.9 percent of Lesotho’s households face catastrophic health spending at the 10 percent threshold, and a vanishingly small 0.2 percent at the more stringent 25 percent threshold. By comparison, many lower-middle-income countries report catastrophic expenditure rates several times higher, and global monitoring reports routinely flag out-of-pocket payments as a leading driver of poverty worldwide. If Lesotho’s numbers were taken at face value, the country would seem to have largely solved the problem of medical impoverishment. The study’s central contribution is to show why that inference would be wrong.

The same surveys revealed that 28.1 percent of households reported at least one member forgoing health care that they believed they needed. That figure, the authors argue, is the missing half of the story. The technical insight is subtle but consequential: catastrophic expenditure can only be incurred by households that actually use health services and pay for them. When anticipated costs, distance, or prior experience of unaffordable bills deter households from seeking care in the first place, no expenditure is recorded, and the headline indicator improves even as the underlying health system fails. In econometric terms, the observed sample of health spenders is not a random sample of the population; it is a selected sample from which the most cost-sensitive households have removed themselves.

To confront this selection problem directly, the team employed a selection model, a statistical framework designed for exactly this kind of censored data. The model jointly estimates the decision to seek care and the level of spending conditional on seeking care, allowing the two processes to be correlated. The results were stark: poorer households in Lesotho are significantly less likely to seek care when they need it. In other words, the households most vulnerable to financial catastrophe are precisely the ones absent from the expenditure data, because they forgo treatment rather than risk the bill. The low catastrophic expenditure rate is, on this reading, partly an artifact of healthcare avoidance rather than evidence of effective financial protection.

The study also constructed a composite universal health coverage index for Lesotho, which scored 74.1, a level broadly comparable to other lower-middle-income countries. That headline figure, however, conceals sharp inequities. Applying concentration index analysis, a technique borrowed from health economics that measures whether service use is distributed evenly across the socioeconomic gradient, the researchers found pro-rich inequity in maternal and child health services. Wealthier households use these services disproportionately, and the analysis traced the drivers to household socioeconomic status and maternal education. Children born to mothers with more schooling and families with more resources are systematically more likely to receive the interventions that keep mothers and infants alive.

These findings arrive at a moment when the global health community is actively debating how to measure progress toward universal health coverage. The World Health Organization and the World Bank track two headline indicators: coverage of essential services and the incidence of catastrophic health spending. Both are useful, but the Lesotho study exposes a structural blind spot. A country can post excellent numbers on both indicators while a large share of its population silently goes without care. Unmet need, captured through survey questions about forgone treatment, is not currently a headline metric in most UHC monitoring frameworks, and the authors argue that it must be. Without it, the indicators reward systems that deter use, penalizing none and protecting no one.

The implications extend well beyond Lesotho’s borders. Many low- and middle-income countries rely heavily on out-of-pocket payments at the point of care, and in settings where formal insurance is thin and household budgets are precarious, the decision to seek care is an economic calculation as much as a medical one. If the price of a clinic visit, transport, or medicines is expected to exceed what a family can bear, the rational response is to stay home. The study’s selection model makes the mechanism visible in the data: the absence of spending among the poor is not evidence of low need or low cost, but of rationing by price at the household level. Monitoring frameworks that ignore this will systematically overstate progress in exactly the countries where progress is hardest to achieve.

For Lesotho itself, the authors outline a policy agenda grounded in their findings. Affordability barriers must be addressed directly, whether through reduced user fees, better protection at the point of care, or targeted support for the poorest households. Maternal education emerges as a powerful lever, with the equity analysis pointing to schooling as a determinant of whether women and children receive essential services. The study also points to expanding community health workers, who can carry basic services into remote mountain communities where facility-based care is hours away, and to increasing investment in primary health care targeted at rural and vulnerable populations. Lesotho’s geography, with dispersed settlements across difficult terrain, makes outreach models particularly consequential for whether nominal entitlements translate into actual access.

The broader lesson is a caution against reading health financing statistics at face value. Numbers that look like success can encode failure, and the direction of the error is not random: it flatters systems that exclude the poor. The authors’ proposal is straightforward in principle, if demanding in practice, that financial protection indicators be complemented everywhere by measures of unmet need, so that a household that never visits a clinic because it cannot afford to is counted as a system failure rather than a statistical blessing. As countries race toward the 2030 deadline for universal health coverage, the Lesotho study offers a reminder that the goal is not merely to make health care cheap on paper, but to make sure people actually receive it, and that the instruments used to measure progress must be sharp enough to tell the difference.

Subject of Research: Financial risk protection and unmet healthcare need in universal health coverage measurement in Lesotho

Article Title: Low catastrophic expenditure, high unmet need in Lesotho: Rethinking financial protection indicators for universal health coverage

Article References: Nikoloski, Z., Koetlisi, A., & Mossialos, E. (2026). Low catastrophic expenditure, high unmet need in Lesotho: Rethinking financial protection indicators for universal health coverage. International Journal for Equity in Health. https://doi.org/10.1186/s12939-026-03026-3

Image Credits: AI Generated

DOI: 10.1186/s12939-026-03026-3

Keywords: universal health coverage, catastrophic health expenditure, Lesotho, unmet need, health equity, concentration index, selection model, out-of-pocket payments, maternal and child health, primary health care, health financing, low- and middle-income countries

The study, led by Long Chen and colleagues at the Arthritis Clinic and Research Center of Peking University People’s Hospital, pooled data from six harmonized longitudinal cohorts: the China Health and Retirement Longitudinal Study (CHARLS), the English Longitudinal Study of Ageing (ELSA), the Health and Retirement Study (HRS) in the United States, the Korean Longitudinal Study of Aging (KLoSA), the Mexican Health and Aging Study (MHAS), and the Survey of Health, Ageing and Retirement in Europe (SHARE). These surveys collectively follow hundreds of thousands of older adults using comparable questionnaires, a harmonization effort coordinated through the Gateway to Global Aging Data platform and supported by the US National Institute on Aging. Because the same walking-difficulty questions appear across all six cohorts, the researchers could ask whether the signal holds across radically different health systems, economies, and cultures.

The mortality findings are the study’s statistical backbone. Using cohort-specific Cox proportional hazards models combined through a DerSimonian-Laird random-effects meta-analysis, the team found that older adults reporting walking difficulty faced roughly double the mortality risk of those without it. The pooled hazard ratio was 2.21, with a 95 percent confidence interval of 2.06 to 2.36, based on five cohorts that provided the necessary survival data. In plain terms, an older adult who struggles to walk was more than twice as likely to die during follow-up as an otherwise comparable peer who walks without difficulty. That is an effect size comparable to major established risk factors, captured here with nothing more than self-reported mobility status.

The sheer diversity of the cohorts, however, meant the association was not perfectly uniform. The researchers reported substantial between-cohort heterogeneity, quantified by an I² statistic of 73.8 percent, meaning that nearly three-quarters of the variability across studies reflected real differences rather than chance. To probe this, they ran a sensitivity analysis using restricted maximum likelihood estimation with Hartung-Knapp inference, a more conservative statistical framework designed for small numbers of pooled studies. That analysis produced a hazard ratio of 2.18 with a wider confidence interval of 1.87 to 2.53, and a 95 percent prediction interval spanning 1.66 to 2.87. The prediction interval matters: it tells us that if the study were repeated in a new country, the true hazard ratio would very likely fall between about 1.7 and 2.9. The direction and magnitude of the association are robust; the exact strength varies by setting, likely reflecting differences in cause-of-death patterns, healthcare access, and baseline functional status across nations.

Mortality is only half the story. The second pillar of the study examined whether walking difficulty also captures what health services researchers call measured need—the diagnosable, quantifiable burden of illness that care ought to respond to. Here the evidence was emphatic. Older adults with walking difficulty had more than three times the odds of multimorbidity, the co-occurrence of multiple chronic conditions, with an odds ratio of 3.05 and a remarkably tight confidence interval of 2.98 to 3.13 across the pooled data. In absolute terms, walking difficulty came with a 25.1 percentage-point higher prevalence of multimorbidity. The authors also linked mobility limitation to limitations in activities of daily living and instrumental activities of daily living, the everyday tasks such as dressing, bathing, shopping, and managing medications that determine whether an older person can live independently. In short, the person who says walking is hard is very often the person carrying the heaviest load of disease and disability.

This is precisely what makes walking difficulty attractive as a policy instrument. Clinical risk scores require laboratory tests, diagnosis codes, and trained assessors. Socioeconomic gradients require detailed income and education data. But a mobility question costs nothing, is understood across cultures, and is already embedded in the world’s major aging surveys. If health systems want to identify the highest-need older adults for outreach, case management, or preventive care, the study suggests that a functional question can triage risk almost as well as instruments that are far more expensive to administer.

The study’s most consequential findings, however, concern what happens after high-need older adults enter the healthcare system. Using cohorts with linked healthcare utilization data—CHARLS, HRS, SHARE, and MHAS—the researchers applied the relative index of inequality (RII) and the slope index of inequality (SII), standard epidemiological tools for quantifying social gradients across the entire education distribution rather than just comparing extremes. After statistically adjusting for measured need, they asked a pointed question: among people who are equally sick, does care reach them equally?

The answer, in three of the four cohorts, was no. For any-care contact—whether an older adult had any contact with a physician or health service—higher educational position was associated with greater need-adjusted healthcare response in CHARLS, HRS, and SHARE. The MHAS estimate pointed in the same direction but was statistically imprecise. Because the population studied was restricted to older adults with walking difficulty—people already identified as high need—the gradients cannot be explained away by the educated simply being healthier. The sicker-poorer pattern had already been held constant; what remained was a systematic tendency for care to track social position.

To make the abstract indices tangible, the researchers translated the education-related gaps into standardized scenarios expressed per 1,000 high-need older adults. In HRS and SHARE, the scenario gaps amounted to roughly 15.6 to 15.7 additional care contacts per 1,000 high-need adults favoring the more educated, a modest but measurable tilt. In Mexico’s MHAS, the estimated gap was 70.0 additional contacts per 1,000, with a confidence interval of 27.8 to 101.0. In China’s CHARLS, the scenario gap ballooned to 296.7 additional contacts per 1,000 high-need adults, with a confidence interval of 74.9 to 456.8—an enormous differential in a country where hundreds of millions of people are entering old age. The authors are careful to note that these standardized scenarios are policy benchmarks rather than causal estimates of what would happen under any specific intervention; they quantify the size of the inequity, not the effect of fixing it.

The evidence on cost-related unmet need—older adults skipping care because they could not afford it—was clearest in SHARE, the European survey, where the linked data allowed the cleanest measurement. This finding connects the study to a broader policy conversation: in systems with high out-of-pocket costs, financial barriers compound the mobility barriers that already make it harder for disabled older adults to reach care. The study’s education gradient likely operates through multiple channels—health literacy, the ability to navigate complex systems, persistence in seeking appointments, and the financial means to act on symptoms—rather than through discrimination alone.

What emerges is a two-part message with global relevance. First, walking difficulty is a pragmatic, validated marker of the highest-need older adults, doubling mortality risk and flagging a multimorbidity burden three times higher, and it can be ascertained anywhere a questionnaire can reach. Second, within this high-need population, healthcare response still follows the gradient of education, with the magnitude of the inequity varying dramatically—from modest gaps in Europe and the United States to a nearly 300-contact difference per 1,000 high-need adults in China. The magnitude and certainty of the gradient differed across cohorts, and the authors stress that heterogeneity itself is informative: health systems are not equally equitable, and the tools to measure that now exist at near-zero cost. As populations age worldwide, monitoring whether care is distributed according to need rather than social position may depend on nothing more sophisticated than asking older adults a question they can answer in a single sentence—and then holding systems accountable for the answer.

Subject of Research: Walking difficulty as a marker of mortality risk, health need, and education-related inequality in healthcare response among older adults across six national aging cohorts.

Article Title: Walking difficulty as a marker of high health need and education-related healthcare response in older adults: a cross-national multicohort study

Article References: Walking difficulty as a marker of high health need and education-related healthcare response in older adults: a cross-national multicohort study. (n.d.). https://doi.org/10.1186/s12916-026-05226-8

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05226-8

Keywords: walking difficulty, older adults, mortality, health equity, multimorbidity, education, healthcare utilization, aging cohorts, unmet need, functional limitation, cross-national study, BMC Medicine