Global determinants of cervical cancer outcomes: a health systems analysis

Abstract

Objective

Significant global health disparities persist in cervical cancer, with over 85% of cases and deaths occurring in low- and middle-income countries (LMICs). In many settings, access to screening, vaccination and treatment is limited. Despite cervical cancer being largely preventable through human papillomavirus (HPV) vaccination and early detection, many women around the world face inadequate healthcare infrastructure, lack of awareness, cultural stigma and gender barriers to seeking care. Therefore, we evaluated global health system metrics that may inform efforts to improve equity in access to cervical cancer care globally.

Methods and analysis

We derived estimates of age-standardised mortality-to-incidence ratios (MIRs) for female patients with cervical cancer of all ages from the Global Cancer Observatory (GLOBOCAN 2022 database), with complete data available for 121 of 185 countries in the database. We selected the following health system metrics: health spending as a per cent of gross domestic product (GDP), physicians/1000 population, nurses and midwives/1000 population, surgical workforce/1000 population, GDP per capita, Universal Health Coverage Service Coverage Index (UHC index), availability of pathology services, Human Development Index (HDI), Gender Inequality Index (GII) (a combined metric of health, empowerment and economic agency), radiotherapy centres/1000 population, out-of-pocket expenditure as percentage of current health expenditure, availability of cervical cancer screening, and HPV coverage.

We evaluated the association between MIR and each metric using univariable linear regressions. To account for multiple comparisons, a Bonferroni correction using a p<0.0045 was applied. Metrics meeting this threshold were included in multivariable models. Variation inflation factor (VIF) allowed exclusion of variables with significant multicollinearity. R² defined goodness of fit.

Subgroup analysis was done by country income level using 2025 World Bank classifications. Separate regression models were then done for high-income and low-middle-income countries.

Results

On univariable analysis, 12 of the 13 metrics were significantly associated with MIR of cancer (p<0.001 for all). HPV coverage was not significantly associated with MIR (β=–0.0007, p=0.290). After including metrics that were significant on univariable analysis, HDI demonstrated significant collinearity (VIF=19). Therefore, after correcting for multicollinearity, the final multivariable model with 11 variables had R² of 0.81.

On multivariable analysis, the following variables were independently associated with lower (improved) MIR for cancer: (1) nurses/midwives per 1000 population (β=–0.0072, p=0.029) and (2) UHC index (β=–0.0022, p=0.026). In addition, greater gender inequality was associated with greater (worse) MIR (β=0.30, p=0.003). Further subgroup analysis by income classification showed that GII remained significantly associated with lower MIRs among HICs (β=–0.526, p=0.033), and only UHC index was significantly associated among LMICs (β=–0.0023, p=0.047)

Conclusion

This comprehensive and global analysis of health system metrics suggests promoting progress towards UHC and strengthening the nursing/midwifery workforce may be independently associated with improved cervical cancer MIR. Furthermore, greater gender inequality was associated with worse MIR. These findings may inform efforts to improve global cervical cancer care and underscore the importance of reducing gender inequality to improve global cervical cancer outcomes.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Cervical cancer disproportionately affects women in low- and middle-income countries, reflecting persistent gaps in prevention, early detection and access to treatment.

• Prior analyses have linked outcomes to socioeconomic factors such as the Human Development Index, but few studies have examined how specific, comparable health-system indicators relate to cervical cancer mortality-to-incidence ratios across countries.

• The independent contribution of gender inequality to cervical cancer outcomes has been suggested but not rigorously assessed alongside health system performance metrics in global analyses.

WHAT THIS STUDY ADDS

• Across 121 countries, higher densities of nurses and midwives and greater Universal Health Coverage (UHC) service coverage were independently associated with lower cervical cancer mortality-to-incidence ratios.

• Gender inequality was independently associated with worse cervical cancer outcomes, even after adjusting for economic and health system factors.

• Determinants differed by economic context: gender inequality was a dominant predictor of outcomes in high-income countries, while UHC coverage remained a significant predictor in low- and middle-income countries.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Strengthening UHC and expanding the nursing and midwifery workforce may be high-impact strategies for reducing cervical cancer mortality, especially in resource-limited settings.

• The independent effect of gender inequality highlights the need for gender-responsive cancer policies that address social, economic and structural barriers to prevention and care.

• Findings can guide nations in prioritising system-level interventions and inform future research evaluating the implementation and cost-effectiveness of health system reforms aimed at reducing global inequities in cervical cancer outcomes.

Introduction

Cervical cancer continues to contribute significantly to the global burden of cancer.¹ Cervical cancer is the fourth leading cause of cancer morbidity and mortality worldwide, with an estimated 660 000 cases diagnosed and 350 000 women dying from the disease in 2022.1,3 Cervical cancer disproportionately affects women in low- and middle-income countries (LMICs), where more than 75% of the world’s population reside and where over 85% of cases and deaths occur.⁴ In 2020, mortality rates reached as high as 55.7 per 100 000 women in Eswatini, compared with just 1.0 in Switzerland—indicative of the high socioeconomic disparities that persist in LMICs which affect cervical cancer survivorship.⁴ This distribution closely reflects disparities in access to effective primary prevention in the form of vaccination and secondary preventive measures through screening.⁴ At diagnosis, women in many LMICs often present with advanced stages of the disease, commonly attributable to inadequate healthcare infrastructure, the lack of awareness and significant cultural and gender-associated barriers that influence access to timely diagnosis and care.⁵

Women in less resourced settings with higher cervical cancer rates are disproportionately affected by gender inequality, faring worse in terms of education levels, employment opportunities, workforce participation and reproductive health access than their counterparts in developed nations.⁵ With an estimated 11 million women from LMICs being diagnosed with cervical cancer in the next decade, it is essential for health systems around the world to strengthen cancer systems. A greater understanding of healthcare system factors associated with cervical cancer outcomes may inform cancer planning.

Therefore, we used the Global Cancer Observatory (GLOBOCAN) 2022 estimates of cervical cancer burden at the country level and evaluated health system factors that may be associated with cervical cancer outcomes. GLOBOCAN 2022 is a project of the International Agency for Research on Cancer (IARC) which provides contemporary estimates of cancer burden in 185 countries using high-quality registry data and statistical modelling where direct data are unavailable.² We made use of countries’ cancer systems and national cervical cancer estimates to identify factors that may be most strongly associated with cervical cancer outcomes. We hope to inform efforts to improve equity in access to cervical cancer care globally.

Methods

The Standards for Reporting Qualitative Research guidelines were used to guide the synthesis and reporting of this study.⁶

Data sources

Estimates of age-standardised mortality-to-incidence ratios (MIRs) were derived for 185 countries from GLOBOCAN 2022 for female patients with cervical cancer of all ages. GLOBOCAN derives incidence and mortality estimates using a tiered approach that prioritises high-quality population-based cancer registry data where available, with systematic evaluation of registry completeness and validity conducted by IARC. For countries lacking robust registries, GLOBOCAN applies standardised modelling methods—drawing on MIRs, neighbouring-country data and established regional patterns—to generate comparable estimates while maintaining internal consistency across settings.² MIR was used as a proxy for the survival experience.⁷

Health system metrics were collected from several databases, including the WHO, the World Bank, the DIrectory of RAdiotherapy Centres (DIRAC) and the United Nations Development Programme (UNDP). The health system metrics included in this study were selected a priori based on their documented relevance in prior global cervical cancer and health systems research. To minimise selection bias, we restricted inclusion to indicators with standardised, publicly available country-level data from major international sources, ensuring consistent definitions and measurement across settings.

Gross domestic product (GDP) per capita adjusted for purchasing power parity, health spending as a per cent of GDP, physician workforce per 1000 population, nurses and midwives per 1000 population and surgical workforce per 1000 population were taken from the World Bank. Out-of-pocket expenditure as a percentage of current health expenditure, Universal Health Coverage (UHC) index, availability of pathology services defined binarily, availability of cervical cancer screening defined binarily and human papillomavirus (HPV) coverage (last dose for all females) were taken from the WHO. Human Development Index (HDI) and Gender Inequality Index (GII) were taken from the UNDP. The UHC index serves as a composite indicator that measures progress towards Sustainable Development Goal indicator 3.8.1 (coverage of essential health services). The GII is a composite measure reflecting inequality in opportunity between men and women in the three dimensions: reproductive health (maternal mortality and adolescent birth rates), empowerment (measured by education levels and parliamentary positions) and labour market participation.⁸ Additionally, using DIRAC data and World Bank 2023 population data, we derived radiotherapy (RT) centres per 1000 population. Because each indicator was obtained from a single database, there were no conflicting data.

Statistical methods

All statistical analyses were performed using Stata V.16.1. We evaluated the association between MIR and each metric using univariable linear regressions. To account for multiple comparisons across the 11 tested health system variables, we applied a Bonferroni correction, dividing the α-level of 0.05 by 11, yielding an adjusted significance threshold of p<0.0045 for inclusion in multivariable models. Variation inflation factor (VIF) analysis allowed exclusion of variables with significant multicollinearity by removing variables with VIF >10. For the final multivariable model, α=0.05 defined statistical significance, given the ecological design of the study.⁹ R² evaluated goodness of fit.¹⁰

No imputation was done for missing values. Countries were included in the analysis only if complete data were available for all variables included in each model. The list of included countries is presented in online supplemental table 1. To assess potential bias, we compared the distribution of HICs and LMICs between included and excluded countries. A χ² test revealed no significant difference in exclusion rates between high income countries (HICs) and LMICs (χ²=2.47, p=0.116), suggesting no systemic bias in data completeness by income level (online supplemental table 2).

Subgroup analysis was done to explore associations by economic context. Countries were stratified by income level using the 2025 World Bank income classification system.¹¹ Separate multivariable linear regression models were then generated for HICs (n=44) and LMICs (n=77). HDI was excluded due to multicollinearity (VIF >10). Analysis was conducted in a similar fashion to the main analysis.

Results

Overall, 185 countries’ data were included based on the availability of cancer incidence and mortality data in GLOBOCAN. Data availability ranged from 98 (52.9%, HPV coverage) to 185 (100%, GDP per capita, RT centres per 1000 population). On univariable analysis, only 12 of the 13 metrics were significantly associated with MIR of cervical cancer with p<0.001 for all except HPV coverage (β=–0.0007, p=0.290). For instance, on linear modelling, the availability of a national cervical cancer screening programme was associated with a –0.19-point reduction in the MIR for cervical cancer. A 0.10-point reduction in GII (scaled from 0 to 1) was associated with a 0.065 increase in cervical cancer MIR. The univariable results are summarised in table 1.

Table 1. Univariable analyses associating health system variables with cervical cancer MIR for female patients.

Health system measure	Beta (β)	CI lower	CI higher	P value	Countries with data
Health spending as % of GDP	−0.0221954	−0.0291775	−0.0152132	<0.001	171
Physicians per 1000 population	−0.0646292	−0.0736497	−0.0556087	<0.001	167
Nurses/midwives per 1000 population	−0.0284479	−0.0319653	−0.0249305	<0.001	166
Surgical workforce per 1000 population	−0.002774	−0.0032795	−0.0022684	<0.001	144
UHC index	−0.0076282	−0.0084961	−0.0067603	<0.001	175
Pathology services (yes vs no)	−0.1564238	−0.2126245	−0.1002231	<0.001	174
HDI	−0.8330788	−0.917276	−0.7488815	<0.001	177
Radiotherapy centres per 1000 population	−88.50648	−102.3259	−74.68706	<0.001	185
GDP per capita	−4.76E−06	−5.55E−06	−3.97E−06	<0.001	185
GII	0.6509115	0.5790224	0.7228006	<0.001	165
Out-of-pocket expenditure as a percentage of current health expenditure (%)	0.0030399	0.0017961	0.0042836	<0.001	171
Cervical cancer screening programme (yes vs no)	−0.187293	−0.2302777	−0.1443082	<0.001	173
HPV vaccination coverage	−0.0006568	−0.0018824	0.0005689	0.290	98

Health system indicators were obtained from the World Bank: health spending as a per cent of GDP (2021), physicians per 1000 population (2016–2021), nurses and midwives per 1000 population (2016–2021), surgical workforce per 1000 population (2013–2018), out-of-pocket expenditure as a percentage of current health expenditure and GDP per capita (international dollars). The UHC index, 2021 was obtained from WHO; it is scored from 0 to 100, with higher values indicating greater service coverage. Pathology service availability, cervical cancer screening and HPV vaccination coverage were taken from the WHO Cancer Country Profiles 2020. The HDI, 2024 and GII, 2022 (lower scores denote greater gender equality) were obtained from UNDP. Radiotherapy centre counts were derived from DIRAC (International Atomic Energy Agency), and population denominators from the 2023 World Bank data set were used to calculate radiotherapy centres per 1000 population.

DIRAC, DIrectory of RAdiotherapy Centres; GDP, gross domestic product; GII, Gender Inequality Index; HDI, Human Development Index; HPV, human papillomavirus; MIR, mortality-to-incidence ratio; UHC index, Universal Health Coverage Service Coverage Index; UNDP, United Nations Development Programme.

Only 12 of the 13 metrics demonstrated significant univariable association and were included in the initial multivariable model, which demonstrated strong explanatory power (n=121, R²=0.8160, online supplemental table 3). However, on VIF analysis, VIF=19.87 for HDI, suggesting multicollinearity. Therefore, the multivariable model was rerun without HDI. After correcting for multicollinearity, the final multivariable model with 11 variables had R² of 0.8105, maintaining strong explanatory power. The multivariable results are summarised in table 2.

Table 2. Multivariable analyses associating health system variables with cervical cancer MIR after removing variables with significant multicollinearity. This model had R² of 0.8105. N=121.

Health system measure	Beta (β)	CI lower	CI higher	P value	VIF
Health spending as % of GDP	0.0011516	−0.0044199	0.0067231	0.683	1.70
Physicians per 1000 population	−0.0128213	−0.0299071	0.0042644	0.14	5.83
Nurses/midwives per 1000 population	−0.0071663	−0.0135736	−0.0007589	0.029	5.24
Surgical workforce per 1000 population	0.0002765	−0.0004292	0.0009823	0.439	4.39
UHC index	−0.0021872	−0.0041054	−0.000269	0.026	5.37
Pathology services (yes vs no)	−0.0003551	−0.042359	0.0416488	0.987	1.51
Radiotherapy centres per 1000 population	−6.78249	−22.44294	8.877957	0.393	2.29
GDP per capita	−5.31E−07	−1.58E−06	5.18E−07	0.318	3.66
GII	0.2987919	0.1072679	0.4903158	0.003	7.66
Out-of-pocket expenditure as a percentage of current health expenditure (%)	−0.0003087	−0.0012288	0.0006114	0.507	1.36
Cervical cancer screening programme (yes vs no)	−0.0080604	−0.0473759	0.0312551	0.685	1.59

Health system indicators were obtained from the World Bank: health spending as a per cent of GDP (2021), physicians per 1000 population (2016–2021), nurses and midwives per 1000 population (2016–2021), surgical workforce per 1000 population (2013–2018), out-of-pocket expenditure as a percentage of current health expenditure and GDP per capita (international dollars). The UHC index, 2021 was obtained from WHO; it is scored from 0 to 100, with higher values indicating greater service coverage. Pathology service availability and cervical cancer screening were taken from the WHO Cancer Country Profiles 2020. The GII, 2022 (lower scores denote greater gender equality) were obtained from UNDP. Radiotherapy centre counts were derived from DIRAC (International Atomic Energy Agency), and population denominators from the 2023 World Bank data set were used to calculate radiotherapy centres per 1000 population.

DIRAC, DIrectory of RAdiotherapy Centres; GDP, gross domestic product; GII, Gender Inequality Index; MIR, mortality-to-incidence ratio; UHC index, Universal Health Coverage Service Coverage Index; UNDP, United Nations Development Programme; VIF, variation inflation factor.

On multivariable analysis, the following variables were independently and significantly associated with lower (improved) MIR for cancer: (1) nurses/midwives per 1000 population (β=–0.0072, p=0.029) (figure 1a) and (2) UHC index (β=–0.0022, p=0.026) (figure 1b). In addition, greater gender inequality was associated with greater (worse) MIR for cervical cancer (β=0.30, p=0.003) (figure 1c).

Figure 1. Cervical cancer MIR plotted against nurses/midwives per 1000 population (a), UHC Service Coverage Index (b) and Gender Inequality Index (c). MIR, mortality-to-incidence ratio; UHC, Universal Health Coverage.

On subgroup multivariable analysis, among HICs (n=44), GII remained significantly associated with lower MIR for cancer (β=–0.526, p=0.033), while others were not significantly associated (online supplemental table 4). Among LMICs (n=77), only the UHC index was significantly associated with lower MIR (β=–0.0023, p=0.047), while all other indicators showed no significant association (online supplemental table 5).

While our multivariable model demonstrates statistically significant associations between health system factors and cancer MIR, we note that the beta coefficients for these associations are small in obsolete terms. For example, an increase of five nurses per 1000 population is associated with only a 0.036 unit decrease in MIR. Similarly, the UHC index—ranging from 0 to 100—had a beta of −0.002, suggesting modest changes per unit. These findings should be interpreted with caution. However, given the population-scale impact of cancer outcomes, even incremental improvements in MIR may represent substantial public health gains when aggregated globally.

Discussion

Cervical cancer remains a stark reflection of global inequalities in healthcare access, as most cases and deaths occur in LMICs.¹² Our study aimed to identify key health system factors that are associated with improved outcomes, as measured by MIRs in order to inform cancer care systems strengthening. Our study revealed that all 12 determinants of health analysed—reflecting healthcare access and delivery, economy and financial protection and gender-related inequalities significantly influence cervical cancer outcomes on univariate analysis. Additionally, the multivariate model highlighted three independent predictors: nursing/midwife capacity, progress towards universal healthcare, and gender inequality.

These findings offer insights into potential points for health system prioritisation. The finding that nursing/midwifery workforce density was inversely and independently associated with MIR on multivariate analysis reinforces their role in the prevention and early delivery of care for cervical cancer.¹³ Nurses and midwives have vital roles in the provision of primary care for cervical cancer not only because they perform screening procedures but also because of their roles in patient and community education and the coordination of care for patients in underserved and rural areas in LMICs.¹⁴ This suggests that strengthening the healthcare workforce by focusing on training, retaining and supporting nurses and midwives must be prioritised in national cervical cancer care strategies, especially in areas where physicians are scarce.13,15

We have also found that the UHC index—a composite measure of 14 tracer measures that reflects coverage on essential services on four domains: reproductive, maternal and newborn and child health, infectious diseases, non-communicable diseases and service capacity and access¹⁶—was independently associated with reduced MIRs. Though prior research has shown that a greater UHC index is associated with reduced mortality rates for patients with cervical cancer,¹⁷ often in the context of access to cervical cancer screening availability,¹⁸ our work examined a multitude of other metrics covering workforce density, economic factors, healthcare delivery and gender inequality. This underscores the importance of progression towards UHC which plays a critical role in the provision of cervical cancer care by: facilitating preventive strategies with HPV vaccination, early screening and treatment of precancerous lesions of the cervix, as well as access to treatment modalities for both early and more advanced forms of the disease.¹⁷

Our finding that HPV vaccination coverage was not significantly associated with cervical cancer MIR on univariable analysis (p=0.290) should be interpreted with caution, given the substantial data gaps in global HPV vaccination reporting. HPV coverage had the highest degree of missingness among all variables, limiting the robustness of this estimate and precluding its inclusion in the multivariable model without major reductions in sample size. The lack of statistical significance, therefore, reflects limitations in data availability rather than the true biological impact of HPV vaccination on cervical cancer outcomes. This aligns with extensive evidence demonstrating that HPV vaccination is a fundamental driver of primary prevention.^{4 5} Accordingly, our findings do not refute the established role of HPV vaccination in reducing cervical cancer burden but instead underscore the urgent need for more complete, standardised global reporting systems for HPV vaccination coverage—particularly in LMICs.^{18 19}

Additionally, while our results support the importance of UHC expansion in improving cervical cancer outcomes, especially in resource-limited settings, we emphasise that implementation remains uneven across countries.20,24 In South Africa and Tanzania, for example, recognition of the link between HIV and cervical cancer has led to targeted screening programmes for women living with HIV and the integration of cervical cancer services into existing UHC-linked HIV care platforms.¹² Recent implementation research among Aboriginal and Torres Strait Islander women in remote Western Australia further illustrates how context-specific, community-aligned strategies—such as self-collection paired with point-of-care HPV testing and same-day specialist assessment—can substantially improve screening participation among historically underserved populations, reinforcing the need for culturally grounded approaches to achieve equitable cervical cancer prevention.²⁵ These targeted, context-specific approaches illustrate how UHC infrastructure can be leveraged to expand equitable access to prevention and screening.¹²

Our study has also found an independent association of gender inequality and MIR. Gender inequality was assessed using the GII, a composite measure that reflects gender-based disparities between men and women in three dimensions: reproductive health, empowerment and the labour market.⁷ Unlike previous studies that have linked gender inequality to cervical cancer outcomes through its association with HDI,⁵ our study has identified GII as an independent predictor of MIR, even after adjusting for economic factors. This underscores the need for gender-focused policies that cover the dimension of GII including: improving access to reproductive healthcare, increasing awareness and uptake of HPV vaccination and screening modalities through health education programmes for cervical cancer,²⁶ reducing economic insecurity and increasing women empowerment by increasing female labour force participation²⁷ and removing discrimination against women with cancer and improving reproductive care access in healthcare systems for cervical cancer.²⁸ It should be noted, however, that countries with the lowest HDIs also tended to perform poorly with respect to gender inequality indices.⁵ Although this was not tested directly in our study, country health systems must still strive to increase economic empowerment by increasing workforce participation, expanding economic opportunities for women and should have a greater commitment to social equality.⁵ Intersectional social determinants of health should not be overlooked.29,31

Paralleling these findings, results from our univariate analyses underscore the importance of strong health infrastructure and greater health workforce capacity to decrease cervical cancer MIRs. We have found a significant association between workforce density metrics and cervical cancer MIR. Our study has found that the density of physicians and the density of the surgical workforce were inversely associated with MIR on univariate analysis. These associations align with prior evidence that links obstetrical and gynaecological and specialist access to lower mortality rates for advanced stage disease¹³ but also highlight how primary care physicians also play a vital role in the early diagnosis and treatment of cervical cancer since they constitute the majority of the physician workforce.^{13 32} This underscores the need to augment primary care physicians around the globe with targeted training programmes, necessary health resources and referral pathways for cervical cancer care, especially in low-resource settings where they are usually the first point of contact for women with cervical cancer.^{5 13}

The availability of RT centres and pathology services was also associated with improved MIR, solidifying their importance in the provision of care for patients with cervical cancer.³³ These metrics underscore the importance of increasing universal healthcare coverage—particularly on diagnostic and treatment infrastructure, and optimising healthcare service delivery systems, which could help reduce disparities in early detection and treatment, improving cervical cancer MIR.³⁴ Addressing the geographical and financial barriers that hinder access to these facilities, especially in LMICs, should be part of national cancer care plans.³⁴

There was a strong association of HDI, GDP per capita and healthcare spending per capita seen on univariate analysis, reflecting how low socioeconomic conditions correlate with worse cancer outcomes.⁴ The HDI is derived from three dimensions: life expectancy at birth (a proxy for health), expected and mean years of schooling (a reflection of education) and the gross national income (a proxy for the standard of living).³⁵ On multivariable analysis, there was a lack of independent association with these indicators and MIR suggesting that the explanatory power may overlap with other direct health system indicators such as the UHC index and the GII. The implication for this finding may be that while economic development remains relevant in cervical cancer care control, its impact may be largely mediated on health system characteristics such as service coverage and delivery, workforce capacity and gender equity. Thus, our findings suggest that there is a need to prioritise UHC and gender-focused interventions over more GDP-centric development models alone, especially for underserved women.^{28 36}

On subgroup multivariable analysis, our study found that GII was the only metric that was significantly associated with lower cervical cancer MIR among HICs. This suggests that in the setting of robust health infrastructure and resources, gender equity may play a more central role in shaping cancer outcomes.⁵ To contextualise, in the USA, for example, disparities in cervical cancer screening exist among sexual and gender minority individuals, particularly within racial and ethnic minorities, contributing to as much as 60% lower screening rates compared with cisgender heterosexual Caucasian individuals.³⁷ This finding highlights the need for policies and programmes in HICs that address barriers contributing towards gender disparities, ensuring equitable access to cancer care.³⁸ Among LMICs, our study found that the UHC index was the only variable significantly associated with improved cervical cancer MIR. This underscores the critical role that equitable access to cancer care services plays in resource-constrained settings. Although these associations should be interpreted with caution, they support the role of UHC as a foundational strategy to provide essential cervical cancer care services in the primary care setting.³⁹

The findings of our study offer strong evidence to support the WHO Cervical Cancer Elimination Initiative Strategy’s emphasis on UHC.³⁸ The strong independent association of the UHC index with MIR for both our main and subgroup analyses highlights the importance of reinforcing and creating universal health programmes in an effort to increase access to cervical cancer diagnostic and treatment modalities particularly among primary care facilities. Our findings of lower MIRs with higher densities of nurses and midwives underscore the call for more investment in healthcare workers as the backbone of comprehensive service delivery for cervical cancer. The independent association of GII and MIR also highlights the call of the Strategy to reduce systemic discrimination in health systems, increase health literacy among women and engage underserved, marginalised women through culturally appropriate services. The other findings in our study also align with the Strategy’s plan of action to achieve 90% treatment and care for cervical cancer by strengthening pathology services, expanding surgical capacity and improving access to RT and chemotherapy.³⁹ Although political commitment, data systems, referral systems and palliative programmes have also been cited to be key components in the Strategy, these were not directly analysed in our study due to limited data availability.

While the multivariable associations were statistically significant, the magnitude of the coefficients should be interpreted with caution. Per-unit changes in variables such as workforce density or UHC index reflect substantial nationwide improvements. Thus, even small beta coefficients may correspond to meaningful population-level reductions in MIR when implemented at scale. Nonetheless, the clinical implications of these findings require further validation in longitudinal or country-specific analyses—involving not just providers and policymakers but also patients and patient advocates⁴⁰—as ecological models cannot directly infer individual-level effects.

Our findings should be interpreted in light of the study’s limitations. While our study identified key health system metrics that are associated with cervical cancer MIR, these health system metrics serve as proxies for more complex healthcare and socioeconomic dynamics. Our model’s R² of 0.81 indicates that the selected variables explain a large proportion of the global variance in cervical cancer MIR; however, only 121 countries had complete data, such that the subgroup analysis by income level may not fully account for selection bias. Although we excluded HDI in the multivariable model due to a high degree of collinearity, the other health system determinants likely remain inter-related. In addition, important contextual factors that influence cervical cancer outcomes—such as cultural norms, political stability and regional health policies—were not captured in the available metrics and may contribute to unmeasured confounding. These findings should therefore not be overinterpreted in the context of individual countries and their unique economic systems; efforts to test and apply these findings must be contextualised to unique population-specific needs.

We note that GLOBOCAN data quality varies across regions, with some countries lacking robust cancer registries and relying on modelled estimations that may potentially impact the accuracy of MIR used in this analysis. Furthermore, MIR reflects population-level mortality and incidence and does not capture individual-level clinical information such as stage at diagnosis, treatment modalities or comorbidities, which are central to understanding survival and quality of care; critically, MIR can be influenced by detection bias, stage migration and data quality issues across countries with varying cancer registration systems. Additionally, we were unable to include metrics on community health worker density per country as WHO and World Bank data demonstrated significant missingness (available only for 65 countries/territories). Furthermore, the cross-sectional nature prevents assessment of temporal relationships between health system factors and cervical cancer outcomes.

Conclusion

In conclusion, this global analysis of health-system metrics highlights the importance of UHC, health workforce expansion and gender equity in reducing cervical cancer MIRs. Beyond primary prevention through screening and vaccination, strengthening nursing and midwifery services and increasing gender equity may be key strategies to improve overall survival. Future efforts on improving cancer care should focus on implementing targeted policies, particularly in LMICs, and evaluate long-term impacts of UHC expansion on cervical cancer outcomes. Furthermore, future works should evaluate the cost-effectiveness of system-level interventions on the expansion of the nursing workforce, increasing UHC coverage and towards improving gender equity to inform resource-prioritising strategies and optimise health policies, especially in resource-constrained settings.

Data availability statement

Data are available upon reasonable request.