We investigated the potential relationship between Universal Periodic Review (UPR) recommendations and health outcomes, using maternal mortality as a key indicator. Specifically, we examined whether supporting UPR recommendations on maternal health from the first three cycles was associated with changes in the maternal mortality ratio (MMR) across countries.

Our preliminary analysis indicates that countries with a higher proportion of accepted recommendations, as well as higher number of UPR recommendations related to maternal health, show a significant correlation with reductions in MMR over time.

Summary of methodology

Identification of health-related recommendations and thematic classification

We first developed a rule-based text classification algorithm to identify health-related themes from all available UPR recommendations (obtained from the Universal Human Rights Index). We used a 2019 report from the WHO to further classify the recommendations into thematic groups. For each thematic group, a dictionary of keywords and term combinations was developed and matched against the recommendation text. A single recommendation could match several thematic groups.

The “Maternal health” thematic category including the following dictionary of keywords: "obstetric," "prenatal," "postnatal," "miscarriage," and "maternal mortality." The algorithm also identified recommendations containing specific combinations of terms linking pregnancy to healthcare access (e.g., "pregnant" appearing in conjunction with "healthcare," "medical care," or "free access"). Recommendations primarily addressing abortion were conditionally classified as maternal health if they explicitly contextualized abortion access within the framework of saving the woman’s life or preserving her physical health. Linguistic false positives were explicitly excluded from the matches.

Statistical analysis

Using a linear mixed-effects model to account for individual country trends, we estimated trends in MMR over time as a function of engagement with UPR recommendations related to maternal health. We specifically tested for a three-way interaction between time, the number of recommendations received, and the proportion of recommendations supported, which allowed us to observe whether engagement with the UPR process was associated with faster rates of MMR reduction.

We visualized these results by comparing the predicted rates of decline for countries with varying levels of recommendation volume (e.g. 5 vs. 15 recommendations) and with varying levels of recommendation acceptance (e.g., 50% vs. 90% support). Pairwise slope comparisons were calculated to assess the statistical differences. All analyses were performed using R.

Summary of Results

We found that a country's support of UPR recommendations on maternal health was associated with a faster decline in MMR. This effect was much stronger when a country received a higher number of the recommendations. When a country received only 5 recommendations (left panel of the figure above), the difference in the rate of MMR reduction between supporting 90% and supporting 50% was minimal but still statistically significant (p = 0.042). However, when a country received 15 recommendations (right panel), supporting 90% of the recommendations was associated with a significantly faster reduction in MMR over time compared to supporting only 50% (p = 0.013).

Conclusion

This preliminary analysis suggests that the UPR process may have impact in contributing towards positive health outcomes as demonstrated by the relationship between the UPR recommendations pertaining to maternal health and reduction of MMR over time.

It is important to note that these results should be interpreted with caution, as this analysis cannot establish causality. Nevertheless, it signifies a potentially important role of engagement with the UPR process and its associated peer review process in enhancing political support and attention for critical health challenges.