Catastrophic and impoverishing health expenditures in fragmented public health systems: lessons from Mexico, 2000–2022

Abstract

Background

Universal Health Coverage (UHC) requires that all individuals access comprehensive health services without facing financial hardship. Fragmented health systems, characterised by multiple uncoordinated financing schemes, can weaken financial risk protection and disproportionately expose vulnerable populations to out-of-pocket expenditures (OOP). Evidence on how fragmentation affects financial protection in Latin America is limited. This study examines the long-term trends in catastrophic, impoverishing, and excessive health expenditures in Mexico’s fragmented public health system.

Methods

We analysed 470,104 households from the Mexican National Household Income and Expenditure Survey (ENIGH, 2000–2022), corresponding to an estimated 341.3 million household-wave observations nationally. Financial hardship was measured using standard indicators: catastrophic health expenditure (CHE), impoverishing health expenditure (IHE), and excessive health expenditure (EHE). Descriptive results used a more granular insurance breakdown, whereas the main analyses grouped households into four main public insurance categories: uninsured, Seguro Popular/INSABI, social security, and mixed public coverage. Heckprobit models estimated adjusted probabilities of financial hardship, accounting for selection among households with positive health expenditure. Analyses incorporated the complex survey design and expansion factors, and sensitivity analyses used alternative thresholds for CHE and EHE.

Findings

Financial protection improved for most households between 2000 and 2014, particularly among Seguro Popular affiliates. After 2020, following the replacement of this public insurance scheme with INSABI, protection deteriorated, with CHE and EHE reaching their highest levels among mixed insurance households by 2022. Social security households consistently experienced the lowest financial burden, while uninsured and INSABI-affiliated households faced substantially higher risk. Increased use of private healthcare, especially for medicines, contributed to rising OOP. Sensitivity analyses confirmed these trends.

Conclusion

Fragmentation in Mexico’s public health system has intensified inequalities in financial protection. The coexistence of multiple uncoordinated insurance schemes has failed to protect households from economic hardship, particularly those in the informal sector. Achieving meaningful UHC will require integrated financing reforms that standardise benefits, reduce administrative complexity, and address persistent gaps in service provision.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13561-026-00717-z.

Background

Achieving Universal Health Coverage (UHC) requires that all individuals receive the health services they need without suffering financial hardship [1]. Health insurance plays an essential role in safeguarding households from financial risks [2]. However, fragmentation can undermine the protective potential of insurance by creating disparities in coverage, benefits, and access [3].

Fragmented health systems, often characterised by multiple unintegrated financing schemes, disjointed service provision, and misaligned governance structures [4–6], compromise health system efficiency, exacerbate inequities in both access and quality of care [7–9], and weaken financial risk protection [2, 10–12]. Evidence from low- and middle-income countries (LMICs) indicates that fragmentation is associated with a heightened risk of out-of-pocket expenditures (OOP) and exposure to catastrophic health spending (CHE) [6, 13, 14], with these adverse effects unequally distributed across populations. In highly fragmented systems, different population segments are often covered by distinct insurance schemes with varying benefit packages. While some plans provide comprehensive coverage, others offer only limited benefits, leaving individuals enrolled in less protective schemes more vulnerable to incurring CHE when seeking healthcare [15–18]. Comparative analyses in Latin America reveal that CHE prevalence ranges from 1% to 25% across countries [19], disproportionately affecting rural populations, the poor, older adults, and those without insurance [19].

The detrimental effects of fragmentation are evident in numerous LMICs. In Tanzania, the coexistence of multiple insurance funds produces notable disparities in healthcare utilisation [20]. In India, the proliferation of insurance arrangements has failed to mitigate high OOP payments [21], while in South Africa, multiple concurrent health insurance schemes are associated with substantial inequities in access to care [22]. These patterns underscore the urgent need for systemic reforms that consolidate and coordinate financing mechanisms, strengthen risk pooling, and harmonise benefit packages [23]. Despite this, evidence remains limited on how fragmented health systems perform in terms of financial protection in LMICs, particularly in Latin America.

Mexico, a large LMIC in Latin America with a fragmented health system [16, 24], offers a compelling case for examining the interplay between health system fragmentation and financial protection. Its health system encompasses a large public sector covering approximately 90% of the population, alongside a smaller but growing private sector. The public sector itself comprises two distinct systems: contributory social security agencies, which cater to the salaried population, and a non-contributory subsystem that serves individuals without social security—principally informal workers, the self-employed, and the unemployed [24]. This dual structure has fostered stark disparities in access, quality, and financial protection. While social security beneficiaries benefit from relatively good comprehensive coverage, the informal sector, which constitutes more than half of the population, has historically faced higher OOP expenditures and greater risk of CHE [17, 25].

Health system fragmentation in Mexico manifests across three interrelated dimensions. In financing, contributory social security schemes such as IMSS and ISSSTE are supported through payroll-based contributions, providing broader benefits and higher per-capita spending compared to non-contributory programmes—formerly Seguro Popular and, since 2020, INSABI—which are funded through general taxation [17, 24, 26]. In service delivery, distinct provider networks cater to each population group, differing in infrastructure, referral pathways, and continuity of care [16, 17, 24]. Governance is characterised by parallel administrative structures and decentralised management across institutions and states, which hampers coordination and undermines efficient resource allocation [16, 24]. Together, these structural divisions account for the persistent inequities in access to healthcare and financial protection observed in Mexico.

Fragmentation also has direct and tangible implications for household financial vulnerability. Although public health expenditure increased during the 2000s—particularly following the introduction of Seguro Popular—the continuation of segmented financing mechanisms, uneven service quality, and underfunded non-contributory schemes limited the health system’s capacity to prevent substantial OOP spending [16, 17, 24]. In Mexico, OOP expenditures largely comprise payments for medicines, outpatient consultations, diagnostics, and hospital services, reflecting both supply shortages and the increasing reliance on private providers to meet healthcare needs [25, 27].

Recent policy shifts and health events, including the dismantling of Seguro Popular in 2019 and the transition to INSABI together with the onset of the COVID-19 pandemic, further eroded financial protection [18]. Against this backdrop, this study analyses the evolution of catastrophic and impoverishing health expenditures between 2000 and 2022. We conceptualise financial protection as the capacity of the health system to safeguard households from the financial hardship associated with healthcare utilisation [18, 19, 28]. This concept is operationalised through internationally standardised and validated indicators [2, 19, 28–30]: CHE, defined as OOP payments exceeding a specified share (10%, 25%, 30%, or 40%) of household capacity to pay; and impoverishing health expenditure (IHE), which occurs when healthcare spending drives households below the poverty line. By applying these complementary metrics, we offer a precise and contextually grounded assessment of financial protection within Mexico’s fragmented health system. Specifically, we: (i) examine trends in CHE and impoverishment across populations; (ii) assess disparities in financial protection associated with Mexico’s fragmented insurance architecture; and (iii) draw policy lessons for other LMICs seeking to strengthen financial risk protection amid fragmentation.

Methods

Study design and data

We conducted a cross-sectional quantitative study using data from the 2000–2022 waves of Mexico’s National Household Income and Expenditures Survey (ENIGH, in Spanish) [31]. This bi-annual survey, established in 1992, gathers information on household (unit of observation) income and expenditures, as well as the economic activities of household members. The survey is managed by the National Institute for Statistics and Geography (INEGI in Spanish), the primary statistical institute in Mexico. Like other household income and expenditure surveys worldwide, ENIGH is recognised as the “gold standard” for measuring health-related spending in Mexico. Its probabilistic, stratified, two-stage, and clustered sampling design enables the production of national and state-representative estimates, with distinctions between urban and rural areas. INEGI-trained personnel conducted face-to-face interviews with around 500,000 households. After excluding 2.2% of observations due to incomplete information or implausible values in key study variables, the final analytical sample comprised 470,104 Mexican households, representing 341.3 million households across the study period.

Outcomes

We analysed changes in the CHE, IHE, and EHE. First, CHE was defined as a binary variable with a value of one when the share of health expenditure (HE) ≥ 30% of a household’s capacity to pay (CTP) [27]. Following COICOP 2018 guidelines [32], HE are classified into four categories: medicines and other health products (drugs, medical supplies, assistive devices, and therapeutic appliances); outpatient care (preventive care, outpatient dental services, and other outpatient services excluding preventive and dental care); inpatient care services (hospitalisation, surgeries, and overnight treatments); and other services (diagnostic imaging, laboratory tests, emergency patient transportation, and rescue services). CTP was defined as total expenditure (TE) minus subsistence expenditure (SE), representing the poverty line. SE was estimated as the average food expenditure of households that allocate between 45% and 55% of their total expenditure to food. This value was then adjusted using a consumption equivalence scale factor of 0.56, based on data from 59 countries [2]. For households where food expenditure (FE) exceeds SE (i.e., FE > SE), CTP was calculated as TE − FE. For households where FE is equal to or greater than SE (i.e., FE ≤ SE), CTP was calculated as TE − SE.

Second, IHE is defined as expenditure on health care that results in a household i falling below the prevailing poverty line or deepening its impoverishment if it is already poor [33]. In our study, IHE was defined as a binary variable equal to 1 when a household’s total expenditure before health payments was at or above subsistence expenditure (TE ≥ SE) and fell below subsistence expenditure after health spending (TE − HE < SE), and 0 otherwise. Finally, excessive spending on health care EHE [34], includes all households experiencing CHE or IHE, capturing a broader measure of financial hardship due to health spending.

It is important to note that we augmented the classification of subsistence expenditure with expenditure on food outside the home due to its importance to Mexican households. Following the literature [35], we also included both the monetary and non-monetary components of household spending, with the latter being derived from self-consumption or gifts received from other households (at market prices) and institutional contributions, such as government subsidies or private organisations’ transfers. Expenditure figures are those incurred during the quarter before the survey and were expressed in constant international dollars of 2018, adjusted for international purchasing power parity dollar (PPP) for household final consumption expenditure (Int-US$).

Covariates

We used several household and contextual characteristics [13, 18, 36] to control for potential confounders in the relationship between financial protection and health insurance status.

Health insurance

We stratified households according to the insurance status of one or more household members into seven mutually exclusive groups. These groups were designed to reflect key institutional distinctions in Mexico’s health system [24, 37]. Specifically, the categories included: (i) fully uninsured households; (ii) households covered by formal employee social security institutions (IMSS, ISSSTE, SEDENA, Marina, PEMEX); (iii) households covered exclusively by programmes for informal workers and the uninsured (SP/INSABI); (iv) households with mixed public coverage, i.e., members covered by both social security and SP/INSABI; (v) households covered only by private insurance (refers to voluntarily purchased health insurance obtained outside the statutory public system, either as a substitute for or a complement to public coverage); (vi) households with both social security and private insurance; and (vii) households with SP/INSABI and private insurance. This classification aligns with common distinctions used in health financing studies internationally, which often differentiate populations by formal versus informal sector coverage and by public versus private combinations. Similar approaches have been applied in fragmented systems [38–41].

Household head

Age (years), sex, schooling level (none, elementary, secondary, high school and college), employment during the last month (unemployment, informally employed or formally employed) and marital status (married/free union, divorced/separated/widowed and single).

Household characteristics

Composition (unipersonal, nuclear, extended or composite), number of adult equivalents [42], proportion of family members aged 0–5 and ≥ 65 years, women of reproductive age (10–54 years), the demographic dependence rate, and a factorial asset and housing material standardised index as a measure of socioeconomic status (SES) [43]. The higher values indicate greater assets, better housing conditions, and participation in any government conditional/non-conditional transfers program.

Area of residence

Rurality/urbanity (urban ≥ 2,500 inhab), and a factorial social-deprivation index-based data collected by the 2020 Population Census on municipal access to basic public services, housing conditions and salary [44], where the higher values indicate a greater social municipal development. We additionally incorporated measures of health resource density—including outpatient facilities, consulting rooms, hospital beds, physicians, dentists, physicians in training, and nurses—per 1,000 inhabitants without social security, alongside the proportion of the population lacking social security. These indicators were obtained from the Health Equipment, Human Resources and Infrastructure Information Subsystem (SINERHIAS), administered by the Ministry of Health [45], as well as from population projections of individuals without social security for 1990–2012 produced by the Ministry of Health, and from the 2015 and 2020 population censuses and intercensal surveys. The indicators were completed using linear interpolation and extrapolation to generate a continuous data series for the study period [46]. Finally, the 32 Mexican states were grouped into seven socio-economic regions, with region one representing the lowest and region seven the highest level of development [47].

Analysis

Main analyses were conducted among households grouped into four main categories capturing public insurance coverage: (i) uninsured; (ii) SP/INSABI affiliates; (iii) social security affiliates; and (iv) mixed public coverage (SP/INSABI and social security). This grouping allowed us to focus on over 95% (see Fig. 1) of households while preserving the ability to identify differences in financial protection and access across the major institutional arrangements. Within this framework, we first described the sample characteristics over the entire study period, including household HE, HE per adult equivalent, total expenditure, food expenditure, capacity to pay, and the proportion of households with positive HE. For these households, we reported mean HE and its share relative to CTP. We also examined trends in the probability of incurring healthcare expenditure by health insurance category and the distribution of healthcare expenditure across medicines and other health products, outpatient care, inpatient services, and other healthcare services. Descriptive statistics were calculated as weighted means or percentages with 95% confidence intervals (CIs), stratified by the four public insurance categories.

Fig. 1.

Health insurance coverage among Mexican households, 2000–2022. Note: SP refers to Seguro Popular and INSABI to the Institute of Health for Wellbeing (Instituto de Salud para el Bienestar). INSABI replaced SP on 1 January 2020 to ensure free provision of health services, medicines, and related supplies for individuals without social security coverage across all levels of care. Mixed public health insurance refers to households affiliated with SP/INSABI and social security institutions. All estimates account for survey design and expansion factors, based on data from the 2000–2022 waves of the Mexican National Household Income and Expenditure Survey (ENIGH) [31]

To analyse variation in the risk of financial hardship due to HE, we estimated the CHE, IHE, and EHE health expenditures by insurance status and survey year. For each outcome, we fitted survey-weighted two-stage probit (Heckprobit) models [48] to account for potential selection bias from restricting the analysis to households with positive HE [49]. In the first stage (selection equation), the probability of any HE was modelled as a function of insurance status, survey year, household head characteristics (sex, age, education, marital status), and household sociodemographic characteristics, including the demographic dependency ratio, percentage of children < 5 years, older adults ≥ 65 years and women of reproductive age, number of adult equivalents, participation in social programmes, municipal deprivation index, and socioeconomic region fixed effects. We also included the density of health resources (outpatient facilities, consulting rooms, hospital beds, physicians, dentists, physicians-in-training, and nurses) per 1,000 inhabitants without social security, and the percentage of the population lacking social security. Interaction terms between insurance status, participation in social programmes, and survey year were included to capture differential temporal trends.

In the second stage (outcome equation), we modelled the probability of CHE, IHE, or EHE conditional on positive HE, controlling for household head characteristics (sex, age, education, marital and employment status), household composition (presence of children < 5 years, older adults ≥ 65 years, and household type), socioeconomic status, urban/rural residence, and state fixed effects. Predicted probabilities were obtained using the margins command in Stata, specifying the interaction between insurance status and year. Analyses were conducted at the mean values of covariates (using the atmeans option), so reported figures represent the model-based of CHE, IHE, and EHE for a household with average characteristics. Estimates are presented as percentages with 95% CIs.

This approach allows for comparisons of changes over time between groups in the likelihood of financial hardship due to health spending across insurance groups, while accounting for differences in observed household characteristics. Our focus is therefore on relative trajectories between populations covered by different insurance schemes and the uninsured. The estimated rho and its high statistical significance (p < 0.001). A high and significant rho (p < 0.001) indicates that unobserved factors influencing whether a household is included in the analysis are also related to unobserved factors determining whether it experiences catastrophic health expenditure. This confirms that selection bias exists and justifies using the two-stage Heckman model, which explicitly corrects for this bias. To assess robustness, we conducted extensive sensitivity analyses. These included varying the HE/CTP thresholds used to define CHE and EHE (≥ 10%, ≥ 25%, and ≥ 40%), as well as estimating alternative additive model specifications with sequential inclusion and exclusion of covariate blocks—particularly those related to health resource density. Across all specifications, the main results remained stable in magnitude, direction, and statistical significance, supporting the robustness of our findings to alternative modelling assumptions and proxy measurement choices. All analyses were performed using Stata v18MP [50].

Results

Between 2000 and 2022, the analytical sample comprised 470,104 households, representing approximately 341.3 million households at the national level. As shown in Fig. 1, following the implementation of SP, there was a continuous expansion in the coverage of households enrolled in SP/INSABI, reaching its peak between 2012 (32.8%) and 2018 (31.1%). However, after the replacement of SP by INSABI in 2020, coverage declined notably. The share of households without public insurance consistently decreased until 2016, reaching its lowest point over the 22 years (8.3%). However, after 2020, this trend reversed, with the uninsured population rising sharply to 26.1% by 2022. Meanwhile, the proportion of households with social security coverage remained relatively stable, comprising approximately 40% of the sample across the entire period. In contrast, households with mixed public insurance coverage exhibited a gradual but sustained increase, particularly during the last decade, going from 5.3% in 2008 to 22.2% in 2016.

Household characteristics differed systematically across insurance groups (Table 1). In 2000–2022, 80.2% (95%CI: 79.8‒80.7) of household heads in SP/INSABI households were informally employed, compared with 54.2% (95%CI: 53.5‒55.0) among mixed insurance households and 35.4% (95%CI: 34.9‒35.9) in those with social security. Educational attainment also differed markedly: 47.2% (95%CI: 46.6‒47.8) of SP/INSABI household heads and 37.0% (95%CI: 36.3‒37.8) of mixed insurance household heads had completed elementary education only, compared with 26.0% (95%CI: 25.4‒26.5) among heads of social security households. Furthermore, 20.8% (95% CI: 20.1–21.4) of SP/INSABI households, 6.6% (95% CI: 6.2–6.9) of mixed insurance households, and 2.1% (95% CI: 1.9–2.2) of social security households were below the poverty line. Demographic dependency ratios were lowest among mixed insurance households (63.5%, 95% CI: 62.6–64.4), compared with SP/INSABI households (82.6%, 95% CI: 81.8–83.4) and social security households (58.8%, 95% CI: 58.2–59.4). Socioeconomic status was highest among social security households (37.0%, 95% CI: 36.9–37.1), followed by mixed insurance households (35.0%, 95% CI: 34.9–35.1) and SP/INSABI households (28.8%, 95% CI: 28.6–29.0).

Table 1.

Household characteristics in assessing catastrophic and impoverishing health expenditures by insurance coverage, Mexico, 2000–2022

Sociodemographic and contextual characteristics	Estimated mean or percentage (95% CI), weighted household sample
	Uninsured	SP/INSABI	Social Security	Mixed Public	Overall
	(n = 90,316,375)	(n = 68,368,745)	(n = 144,383,321)	(n = 38,263,287)	(n = 341,331,728)
Head of Household
Women, %	25.8 (25.2‒26.4)	25.8 (25.3‒26.3)	26.1 (25.7‒26.6)	28.8 (28.2‒29.5)	26.3 (26.0‒26.6)
Age (in yrs.), mean	48.8 (48.5‒49.0)	48.2 (48.0‒48.4)	49.3 (49.1‒49.5)	50.7 (50.5‒50.9)	49.1 (49.0‒49.2)
Schooling, %
None	14.4 (13.6‒15.1)	14.2 (13.7‒14.6)	3.9 (3.7‒4.1)	6.8 (6.4‒7.1)	9.0 (8.8‒9.3)
Elementary	44.4 (43.6‒45.2)	47.2 (46.6‒47.8)	26.0 (25.4‒26.5)	37.0 (36.3‒37.8)	36.3 (35.9‒36.7)
Secondary	20.2 (19.5‒20.8)	25.5 (25.0‒26.1)	23.7 (23.3‒24.1)	28.8 (28.1‒29.4)	23.7 (23.4‒24.0)
High school	9.8 (9.4‒10.1)	8.4 (8.1‒8.7)	17.5 (17.1‒17.8)	13.9 (13.4‒14.4)	13.2 (13.0‒13.4)
College	11.3 (10.6‒12.0)	4.7 (4.4‒4.9)	29.0 (28.4‒29.6)	13.6 (13.1‒14.1)	17.7 (17.3‒18.1)
Unemployment, %	21.7 (21.0‒22.4)	18.5 (18.1‒18.9)	23.8 (23.3‒24.3)	24.4 (23.8‒25.0)	22.3 (22.0‒22.5)
Informally employed in the last month, %	75.9 (75.2‒76.6)	80.2 (79.8‒80.7)	35.4 (34.9‒35.9)	54.2 (53.5‒55.0)	57.2 (56.7‒57.6)
Formally employed in the last month, %	2.4 (2.2‒2.6)	1.3 (1.2‒1.4)	40.8 (40.2‒41.5)	21.4 (20.7‒22.0)	20.6 (20.1‒21.0)
Marital status, %
Married/free union	66.9 (66.2‒67.6)	73.6 (73.1‒74.1)	69.8 (69.3‒70.2)	72.2 (71.5‒72.8)	70.1 (69.7‒70.4)
Divorced/separated/widowed	23.9 (23.4‒24.5)	21.0 (20.5‒21.4)	22.3 (21.9‒22.7)	22.8 (22.2‒23.4)	22.5 (22.3‒22.8)
Single	9.1 (8.8‒9.5)	5.4 (5.1‒5.7)	7.9 (7.7‒8.2)	5.0 (4.7‒5.3)	7.4 (7.2‒7.6)
Household
Composition, %
Unipersonal	15.3 (14.8‒15.8)	8.8 (8.5‒9.1)	9.7 (9.4‒10.1)	1.8 (1.6‒2.0)	10.1 (9.9‒10.3)
Nuclear	64.9 (64.2‒65.5)	66.9 (66.4‒67.4)	66.3 (65.9‒66.7)	52.5 (51.8‒53.2)	64.5 (64.2‒64.8)
Extended	18.9 (18.3‒19.4)	23.5 (23.0‒24.0)	23.0 (22.6‒23.4)	44.4 (43.7‒45.1)	24.4 (24.2‒24.6)
Composite	1.0 (0.9‒1.1)	0.7 (0.6‒0.8)	1.0 (0.9‒1.1)	1.3 (1.1‒1.4)	1.0 (0.9‒1.0)
AE, mean	2.4 (2.4‒2.4)	2.6 (2.6‒2.6)	2.5 (2.5‒2.5)	3.1 (3.0‒3.1)	2.6 (2.6‒2.6)
Family members aged 0–5, %	8.2 (8.0‒8.5)	10.3 (10.1‒10.4)	6.9 (6.8‒7.1)	8.9 (8.7‒9.1)	8.2 (8.1‒8.3)
Family members aged ≥ 65, %	13.7 (13.3‒14.2)	12.4 (12.1‒12.7)	11.7 (11.4‒12.1)	9.3 (9.0‒9.6)	12.1 (11.9‒12.3)
Women of reproductive age (10–54 years), %	30.3 (30.0‒30.6)	31.5 (31.3‒31.7)	33.2 (32.9‒33.5)	35.3 (35.0‒35.6)	32.3 (32.2‒32.5)
Demographic dependence rate, %	75.5 (74.4‒76.5)	82.6 (81.8‒83.4)	58.8 (58.2‒59.4)	63.5 (62.6‒64.4)	68.5 (68.0‒69.0)
SES index, %	29.3 (29.0‒29.7)	28.8 (28.6‒29.0)	37.0 (36.9‒37.1)	35.0 (34.9‒35.1)	33.1 (33.0‒33.2)
TE per AE	2.1 (1.9‒2.2)	1.1 (1.1‒1.1)	2.5 (2.4‒2.5)	1.4 (1.4‒1.4)	2.0 (1.9‒2.0)
FE per AE	0.6 (0.6‒0.7)	0.5 (0.5‒0.5)	0.7 (0.7‒0.7)	0.5 (0.5‒0.5)	0.6 (0.6‒0.6)
CTP per AE	1.6 (1.4‒1.7)	0.7 (0.7‒0.7)	2.0 (1.9‒2.0)	1.0 (0.9‒1.0)	1.5 (1.4‒1.5)
Below the poverty line	14.4 (13.3‒15.4)	20.8 (20.1‒21.4)	2.1 (1.9‒2.2)	6.6 (6.2‒6.9)	9.6 (9.2‒10.0)
Beneficiary of any social program, %	22.4 (21.1‒23.8)	49.4 (48.6‒50.2)	14.1 (13.7‒14.6)	36.0 (35.2‒36.7)	25.8 (25.1‒26.6)
Area of residence, %
Urban	70.0 (67.7‒72.2)	55.4 (54.1‒56.7)	91.3 (90.8‒91.9)	79.7 (78.9‒80.6)	77.2 (76.2‒78.1)
Social deprivation index, %	29.6 (28.9‒30.4)	27.5 (27.2‒27.8)	20.8 (20.4‒21.1)	21.5 (21.3‒21.7)	24.5 (24.2‒24.9)
Density of outpatient health facilities, mean	0.2 (0.2‒0.2)	0.3 (0.3‒0.3)	0.2 (0.2‒0.2)	0.2 (0.2‒0.2)	0.2 (0.2‒0.2)
Density of consulting rooms, mean	0.6 (0.5‒0.6)	0.6 (0.6‒0.6)	0.6 (0.6‒0.6)	0.6 (0.6‒0.6)	0.6 (0.6‒0.6)
Density of inpatient hospital beds, mean	2.0 (1.9‒2.2)	2.4 (2.2‒2.5)	1.5 (1.4‒1.6)	1.7 (1.7‒1.8)	1.8 (1.8‒1.9)
Density of physicians and dentists, mean	1.1 (1.0‒1.2)	1.1 (1.1‒1.1)	1.3 (1.3‒1.4)	1.3 (1.3‒1.3)	1.2 (1.2‒1.3)
Density of physicians-in-training, mean	0.4 (0.4‒0.4)	0.4 (0.3‒0.4)	0.5 (0.5‒0.5)	0.4 (0.4‒0.5)	0.4 (0.4‒0.5)
Density of nurses, mean	1.9 (1.7‒2.0)	2.0 (1.9‒2.0)	2.6 (2.5‒2.7)	2.6 (2.5‒2.6)	2.3 (2.2‒2.4)
Population without Social Security, %	59.7 (58.5‒60.9)	69.7 (69.2‒70.2)	46.4 (45.9‒46.9)	54.8 (54.3‒55.2)	55.5 (54.9‒56.1)
Socioeconomic Region
Lowest	14.1 (12.2‒16.1)	17.7 (16.7‒18.7)	5.0 (4.3‒5.6)	7.6 (7.1‒8.2)	10.2 (9.4‒11.1)
2	22.1 (19.8‒24.5)	27.1 (25.9‒28.4)	14.1 (12.9‒15.3)	18.9 (18.0‒19.8)	19.4 (18.2‒20.5)
3	13.0 (11.3‒14.8)	14.9 (14.1‒15.8)	9.8 (8.9‒10.8)	12.7 (12.0‒13.3)	12.0 (11.1‒12.9)
4	22.6 (19.7‒25.6)	20.9 (19.9‒22.0)	25.3 (23.5‒27.0)	25.5 (24.5‒26.5)	23.7 (22.2‒25.3)
5	9.3 (7.6‒11.0)	8.3 (7.7‒8.8)	15.8 (14.3‒17.4)	14.3 (13.6‒14.9)	12.4 (11.3‒13.6)
6	11.1 (9.1‒13.1)	7.5 (6.8‒8.1)	18.9 (17.3‒20.5)	14.0 (13.2‒14.7)	14.0 (12.8‒15.2)
Highest	7.6 (5.1‒10.0)	3.6 (3.3‒3.8)	11.2 (9.4‒12.9)	7.1 (6.6‒7.5)	8.2 (6.8‒9.6)

SP refers to Seguro Popular and INSABI to the Institute of Health for Wellbeing (Instituto de Salud para el Bienestar). INSABI replaced SP on 1 January 2020 to ensure free provision of health services, medicines, and related supplies for individuals without social security coverage across all levels of care. Mixed public health insurance refers to households affiliated with SP/INSABI and social security institutions. AE: Adult equivalents. Women of reproductive age: Defined as those aged 10–54 years. TE: Total quarterly household expenditure per adult equivalent (in thousands of international dollars, adjusted for purchasing power parity [PPP] for household final consumption expenditure), including spending on food and beverages (FE), transport and communications, housing and utilities, personal care, education, health (HE), and other items. FE: Total quarterly food expenditure per adult equivalent (PPP-adjusted international dollars). CTP: Total quarterly capacity to pay per adult equivalent (PPP-adjusted international dollars). SES: Socioeconomic status index, constructed from household assets and dwelling conditions using factor loadings from 2000. Poverty line: Subsistence expenditure, defined as the average food expenditure of households allocating 45%–55% of total spending to food, adjusted by a consumption equivalence scale (0.56) and estimated using data from 59 countries [2]. Beneficiary of a social programme: Participation in any government conditional or unconditional programme, notably the discontinued Prospera programme (formerly Progresa or Oportunidades). Social deprivation index: Factorial index constructed from the 2000, 2010, and 2020 Population Censuses and the 2005 and 2015 Intercensal Surveys, with linear predictions for 2002–2008 and 2012–2018 and standardisation based on 2000 factor loadings. Density of health resources: Outpatient health facilities, consulting rooms, inpatient hospital beds, physicians and dentists, physicians-in-training, and nurses per 1,000 inhabitants without social security. All estimates account for survey design and expansion factors, based on data from the 2000–2022 waves of the Mexican National Household Income and Expenditure Survey (ENIGH) [26]

Across the study period, 68.4% (95%CI: 67.6‒69.1) of households with mixed insurance reported any health expenditure. This level was similar to that of uninsured households (69.1%, 95%CI: 68.2‒70.0) and to the level observed among households with social security coverage (67.9%, 95%CI: 67.3‒68.6), and slightly higher than that of households covered by SP/INSABI (65.8%, 95%CI: 65.1‒66.5) (Table 2). The average quarterly health expenditure over the whole period was highest among households with social security coverage (Int-US$314.2), followed by uninsured households (Int-US$271.7), households with mixed insurance (Int-US$245.8), and SP/INSABI households (Int-US$175.9). Among households with positive HE, it represented approximately 4.9% (95%CI: 4.9‒5.0) of total household expenditure across the sample, with slight increases among SP/INSABI and uninsured households. When considering capacity to pay, households affiliated with SP/INSABI allocated 8.2% (95%CI: 8.1‒8.4) to health expenditure, households with mixed insurance 7.3% (95%CI: 7.1‒7.5), and households with social security 6.5% (95%CI: 6.4‒6.6).

Table 2.

Healthcare expenditure by insurance coverage, Mexico, 2000–2022

Health expenditure indicator	Estimated mean or percentage (95% CI), weighted household sample
	Uninsured	SP/INSABI	Social Security	Mixed Public	Overall
	(n = 90,316,375)	(n = 68,368,745)	(n = 144,383,321)	(n = 38,263,287)	(n = 341,331,728)
Reported HE > 0, %	69.1 (68.2‒70.0)	65.8 (65.1‒66.5)	67.9 (67.3‒68.6)	68.4 (67.6‒69.1)	67.9 (67.4‒68.4)
If HE > 0
HE, mean	271.7 (251.0‒292.4)	175.9 (165.1‒186.6)	314.2 (301.3‒327.0)	245.8 (224.4‒267.2)	268.2 (258.2‒278.1)
HE per AE, mean	131.7 (121.0‒142.5)	77.4 (72.0‒82.9)	137.8 (132.3‒143.4)	87.5 (77.7‒97.4)	118.8 (114.1‒123.4)
HE as a share of TE, %	5.2 (5.1‒5.4)	5.0 (4.9‒5.1)	4.7 (4.6‒4.8)	4.8 (4.7‒5.0)	4.9 (4.9‒5.0)
HE as a share of FE, %	23.8 (21.9‒25.7)	18.4 (17.4‒19.5)	23.2 (22.0‒24.4)	21.0 (19.1‒22.9)	22.2 (21.3‒23.0)
HE as a share of CTP, %	7.9 (7.7‒8.1)	8.2 (8.1‒8.4)	6.5 (6.4‒6.6)	7.3 (7.1‒7.5)	7.3 (7.2‒7.4)

SP refers to Seguro Popular, and INSABI to the Institute of Health for Wellbeing (Instituto de Salud para el Bienestar), which replaced SP on 1 January 2020 to provide free health services, medicines, and related supplies for individuals without social security coverage across all levels of care. Mixed public health insurance refers to households affiliated with SP/INSABI and social security institutions. AE denotes adult equivalents, and women of reproductive age are defined as those aged 10–54. TE is total quarterly household expenditure per adult equivalent (in thousands of international dollars, PPP-adjusted for household final consumption expenditure), including spending on food and beverages (FE), transport and communications, housing and utilities, personal care, education, health (HE), and other items. Following COICOP 2018 guidelines [32], health expenditures are classified into four categories: Medicines and other health products (drugs, medical supplies, assistive devices, and therapeutic appliances); Outpatient care (preventive care, outpatient dental services, and other outpatient services excluding preventive and dental care); Inpatient care services (hospitalisation, surgeries, and overnight treatments); and Other services (diagnostic imaging, laboratory tests, emergency patient transportation, and rescue services). FE is total quarterly food expenditure per adult equivalent (PPP-adjusted), and CTP is total quarterly capacity to pay per adult equivalent (PPP-adjusted). All estimates account for survey design and expansion factors, based on data from the 2000–2022 waves of the Mexican National Household Income and Expenditure Survey (ENIGH) [26]

The survey-weighted probability of incurring any positive HE varied across insurance groups and over time (Fig. 2). After 2020, this probability increased in all groups: uninsured households (67.4%, 95% CI: 66.1–68.7 vs. 55.0%, 95% CI: 53.0–57.1 in 2018), SP/INSABI households (73.7%, 95% CI: 72.7–74.7 vs. 63.8%, 95% CI: 62.7–64.9 in 2018), social security households (74.1%, 95% CI: 73.3–75.0 vs. 61.4%, 95% CI: 60.4–62.4 in 2018), and mixed insurance households (78.2%, 95% CI: 77.1–79.3 vs. 66.9%, 95% CI: 65.6–68.2 in 2018). By 2022, the probability of HE was 75.9% (95% CI: 74.5–77.3) among mixed insurance households, 68.7% (95% CI: 67.9–69.5) among social security households, 67.8% (95% CI: 66.8–68.8) among uninsured households, and 73.3% (95% CI: 72.0–74.5) among SP/INSABI households.

Fig. 2.

Trends in probability of household healthcare expenditure by health insurance coverage, Mexico, 2000–2022. Note: SP refers to Seguro Popular, and INSABI to the Institute of Health for Wellbeing (Instituto de Salud para el Bienestar), which replaced SP on 1 January 2020 to provide free health services, medicines, and related supplies for individuals without social security coverage across all levels of care. Mixed public health insurance refers to households affiliated with SP/INSABI and social security institutions. AE denotes adult equivalents, and women of reproductive age are defined as those aged 10–54. Following COICOP 2018 guidelines [32], health expenditures are classified into four categories: Medicines and other health products (drugs, medical supplies, assistive devices, and therapeutic appliances); Outpatient care (preventive care, outpatient dental services, and other outpatient services excluding preventive and dental care); Inpatient care services (hospitalisation, surgeries, and overnight treatments); and Other services (diagnostic imaging, laboratory tests, emergency patient transportation, and rescue services). All estimates account for survey design and expansion factors, based on data from the 2000–2022 waves of the Mexican National Household Income and Expenditure Survey (ENIGH) [31]

The composition of HE also differed by insurance group (Fig. 3). In all groups, medicines and other health products accounted for the largest share of spending. Over the entire study period, this share was highest among SP/INSABI households (64.5%), followed by mixed insurance households (63.6%), uninsured households (63.3%), and social security households (61.9%). Outpatient and inpatient services represented a larger proportion of spending among social security households (30.8%) and, to a lesser extent, SP/INSABI households (28.2%).

Fig. 3.

Distribution of healthcare expenditure by component and insurance coverage, Mexico, 2000–2022. Note: SP refers to Seguro Popular, and INSABI to the Institute of Health for Wellbeing (Instituto de Salud para el Bienestar), which replaced SP on 1 January 2020 to provide free health services, medicines, and related supplies for individuals without social security coverage across all levels of care. Mixed public health insurance refers to households affiliated with SP/INSABI and social security institutions. AE denotes adult equivalents, and women of reproductive age are defined as those aged 10–54. Following COICOP 2018 guidelines [32], health expenditures are classified into four categories: Medicines and other health products (drugs, medical supplies, assistive devices, and therapeutic appliances); Outpatient care (preventive care, outpatient dental services, and other outpatient services excluding preventive and dental care); Inpatient care services (hospitalisation, surgeries, and overnight treatments); and Other services (diagnostic imaging, laboratory tests, emergency patient transportation, and rescue services). All estimates account for survey design and expansion factors, based on data from the 2000–2022 waves of the Mexican National Household Income and Expenditure Survey (ENIGH) [31]

The financial protection indicators exhibited distinct trends across insurance groups (Table 3). From 2000 to 2014, CHE declined across all groups, particularly among SP/INSABI households (from 4.8%, 95% CI: 4.2–5.5 in 2006 to 1.7%, 95% CI: 1.5–2.9 in 2014) and mixed insurance households (from 5.4%, 95% CI: 4.7–6.1 in 2006 to 2.0%, 95% CI: 1.7–2.3 in 2014). However, after 2014, this trend reversed, with CHE rising to 3.3% (95% CI: 3.0–3.5) and 3.0% (95% CI: 2.7–3.2) in 2020 and 2022, respectively, among SP/INSABI households, and 3.6% (95% CI: 3.3–4.0) and 3.3% (95% CI: 3.0–3.7) among mixed insurance households. The IHE remained below 1% across all groups between 2006 and 2014 but showed an upward trend post-2014 (Table 3). By 2022, IHE reached 1.2% (95% CI: 0.9–1.5) in SP/INSABI households and 1.0% (95% CI: 0.7–1.2) in mixed insurance households. EHE, combining both CHE and IHE, followed a U-shaped trend, decreasing from 3.9% in 2000 to 1.9% in 2014 and then rising back to 3.2% in 2022 among uninsured and social security households. In SP/INSABI households, EHE decreased from 5.8% (95% CI: 5.1–6.4) in 2006 to 2.4% (95% CI: 2.1–2.7) in 2014, before increasing to 3.9% (95% CI: 3.6–4.3) in 2022. Similarly, in mixed insurance households, EHE declined from 5.9% (95% CI: 5.2–6.7) in 2006 to 2.5% (95% CI: 2.1–2.8) in 2014 and then rose to 4.1% (95% CI: 3.7–4.4) in 2022.

Table 3.

Incidence of catastrophic, impoverishing, and excessive health expenditure prevalence by health insurance status, Mexico, 2000–2022

	2000	2002	2004	2006	2008	2010	2012	2014	2016	2018	2020	2022
CHE, threshold 30% (rho = 0.705, p < 0.001)
Uninsured	3.2 (2.6‒3.9)	2.8 (2.4‒3.3)	2.7 (2.4‒3.1)	4.1 (3.6‒4.6)	2.4 (2.1‒2.7)	1.5 (1.2‒1.7)	1.8 (1.5‒2.2)	1.4 (1.2‒1.7)	2.2 (2.0‒2.5)	2.4 (2.1‒2.6)	2.7 (2.5‒3.0)	2.5 (2.3‒2.7)
SP/INSABI				4.8 (4.2‒5.5)	2.9 (2.6‒3.2)	1.8 (1.5‒2.1)	2.2 (1.8‒2.6)	1.7 (1.5‒2.0)	2.6 (2.4‒2.9)	2.8 (2.6‒3.1)	3.3 (3.0‒3.5)	3.0 (2.7‒3.2)
Social Security	3.5 (2.8‒4.1)	3.0 (2.6‒3.5)	2.9 (2.5‒3.4)	4.4 (3.9‒4.9)	2.6 (2.3‒2.9)	1.6 (1.4‒1.8)	2.0 (1.6‒2.3)	1.6 (1.3‒1.8)	2.4 (2.2‒2.6)	2.5 (2.3‒2.8)	2.9 (2.7‒3.2)	2.7 (2.5‒2.9)
Mixed Public				5.4 (4.7‒6.1)	3.3 (2.9‒3.7)	2.0 (1.7‒2.3)	2.5 (2.0‒2.9)	2.0 (1.7‒2.3)	3.0 (2.7‒3.2)	3.2 (2.9‒3.5)	3.6 (3.3‒4.0)	3.3 (3.0‒3.7)
IHE (rho = 0.205, p = 0.118)
Uninsured	0.5 (0.3‒0.8)	0.4 (0.3‒0.6)	0.4 (0.3‒0.6)	0.6 (0.5‒0.8)	0.5 (0.3‒0.6)	0.4 (0.2‒0.5)	0.4 (0.2‒0.6)	0.5 (0.3‒0.7)	0.8 (0.5‒1.0)	0.8 (0.5‒1.0)	0.9 (0.7‒1.1)	0.9 (0.7‒1.1)
SP/INSABI				0.8 (0.6‒1.1)	0.7 (0.5‒0.9)	0.5 (0.3‒0.7)	0.5 (0.3‒0.8)	0.7 (0.4‒0.9)	1.0 (0.8‒1.3)	1.1 (0.8‒1.4)	1.2 (0.9‒1.4)	1.2 (0.9‒1.5)
Social Security	0.4 (0.3‒0.6)	0.3 (0.2‒0.4)	0.3 (0.2‒0.4)	0.5 (0.3‒0.6)	0.4 (0.2‒0.5)	0.3 (0.2‒0.4)	0.3 (0.2‒0.5)	0.4 (0.2‒0.5)	0.6 (0.4‒0.8)	0.6 (0.4‒0.8)	0.7 (0.5‒0.9)	0.7 (0.5‒0.9)
Mixed Public				0.7 (0.5‒0.9)	0.5 (0.4‒0.7)	0.4 (0.3‒0.6)	0.4 (0.3‒0.6)	0.5 (0.4‒0.7)	0.8 (0.6‒1.0)	0.9 (0.6‒1.1)	1.0 (0.8‒1.2)	1.0 (0.7‒1.2)
EHE, threshold 30% (rho = 0.543, p < 0.05)
Uninsured	3.8 (3.1‒4.4)	3.2 (2.7‒3.6)	3.1 (2.7‒3.5)	4.7 (4.1‒5.2)	2.8 (2.4‒3.1)	1.8 (1.5‒2.1)	2.2 (1.7‒2.6)	1.9 (1.6‒2.2)	2.7 (2.4‒3.0)	2.9 (2.5‒3.2)	3.4 (3.1‒3.7)	3.1 (2.9‒3.4)
SP/INSABI				5.8 (5.1‒6.4)	3.5 (3.1‒3.9)	2.3 (1.9‒2.6)	2.7 (2.3‒3.2)	2.4 (2.1‒2.7)	3.4 (3.1‒3.7)	3.6 (3.3‒3.9)	4.2 (3.9‒4.5)	3.9 (3.6‒4.3)
Social Security	3.9 (3.3‒4.5)	3.3 (2.8‒3.8)	3.2 (2.8‒3.6)	4.8 (4.2‒5.3)	2.8 (2.5‒3.2)	1.8 (1.6‒2.1)	2.2 (1.8‒2.6)	1.9 (1.7‒2.2)	2.8 (2.5‒3.0)	2.9 (2.6‒3.2)	3.5 (3.2‒3.7)	3.2 (3.0‒3.5)
Mixed Public				5.9 (5.2‒6.7)	3.6 (3.2‒4.0)	2.4 (2.0‒2.7)	2.8 (2.3‒3.3)	2.5 (2.1‒2.8)	3.5 (3.2‒3.8)	3.7 (3.4‒4.1)	4.3 (4.0‒4.7)	4.1 (3.7‒4.4)

The percentage of households that incur catastrophic (CHE), impoverishing (IHE), and excessive (EHE) health expenditures, with 95% confidence intervals in parentheses, is reported. Seguro Popular (SP, People’s Insurance) was implemented in 2003 and later replaced by the Instituto de Salud para el Bienestar (INSABI, Institute of Health for Wellbeing) on 1 January 2020. INSABI was mandated to guarantee free access to health services, medicines, and medical supplies for individuals without social security coverage across all levels of care. Mixed public health insurance refers to households affiliated with SP/INSABI and social security schemes. Following COICOP 2018 guidelines [32], health expenditures were classified into four categories: (i) Medicines and other health products (drugs, medical supplies, assistive devices, and therapeutic appliances); (ii) Outpatient care (preventive care, outpatient dental services, and other outpatient services excluding preventive and dental care); (iii) Inpatient care services (hospitalisation, surgeries, and overnight treatments); and (iv) Other health services (diagnostic imaging, laboratory tests, emergency transportation, and rescue services). Probabilities of CHE, IHE, and EHE were estimated using a two-stage maximum-likelihood probit (Heckprobit) model [48]. In the first stage (selection equation), the probability of any HE was modelled as a function of insurance status, survey year, household head characteristics (sex, age, education, marital status), and household sociodemographic characteristics, including the demographic dependency ratio, percentage of children < 5 years, older adults ≥ 65 years and women of reproductive age, number of adult equivalents, participation in social programmes, municipal deprivation index, and socioeconomic region fixed effects. We also included the density of health resources (outpatient facilities, consulting rooms, hospital beds, physicians, dentists, physicians-in-training, and nurses) per 1,000 inhabitants without social security, and the percentage of the population lacking social security. Interaction terms between insurance status, participation in social programmes, and survey year were included to capture differential temporal trends. In the second stage (outcome equation), we modelled the probability of CHE, IHE, or EHE conditional on positive HE, controlling for household head characteristics (sex, age, education, marital and employment status), household composition (presence of children < 5 years, older adults ≥ 65 years, and household type), socioeconomic status, urban/rural residence, and state fixed effects. Predicted probabilities were obtained using the margins command in Stata, specifying the interaction between insurance status and year. Analyses were conducted at the mean values of covariates (using the atmeans option). The estimated rho and its high statistical significance (p < 0.001) indicate that the probability of selection and the probability of the outcome are strongly correlated in their errors, confirming that selection bias is relevant and that the Heckprobit model is appropriate. All estimates incorporate survey design features and expansion factors. Data are drawn from the Mexican National Household Income and Expenditure Survey (ENIGH), 2000–2022 [26]

Table 4.

Sensitivity analyses of the incidence of catastrophic, impoverishing, and excessive health expenditures by health insurance status, Mexico, 2000–2022

	2000	2002	2004	2006	2008	2010	2012	2014	2016	2018	2020	2022
CHE, threshold 10% (rho = 0.766, p < 0.001)
Uninsured	17.5 (16.2‒18.9)	15.7 (14.6‒16.8)	14.8 (13.8‒15.9)	18.6 (17.6‒19.5)	12.1 (11.4‒12.9)	7.4 (6.7‒8.0)	9.8 (8.8‒10.8)	7.4 (6.8‒8.0)	10.3 (9.6‒10.9)	10.1 (9.4‒10.7)	12.5 (11.9‒13.1)	11.1 (10.6‒11.6)
SP/INSABI				21.6 (20.4‒22.8)	14.5 (13.6‒15.4)	9.0 (8.3‒9.7)	11.8 (10.7‒12.8)	9.0 (8.4‒9.7)	12.4 (11.7‒13.0)	12.1 (11.5‒12.7)	14.9 (14.3‒15.5)	13.3 (12.7‒13.9)
Social Security	18.3 (16.9‒19.7)	16.4 (15.3‒17.5)	15.5 (14.4‒16.7)	19.4 (18.3‒20.4)	12.7 (12.0‒13.5)	7.8 (7.2‒8.4)	10.3 (9.3‒11.3)	7.8 (7.2‒8.4)	10.8 (10.3‒11.4)	10.6 (10.0‒11.2)	13.1 (12.6‒13.6)	11.6 (11.2‒12.1)
Mixed Public				21.9 (20.6‒23.2)	14.7 (13.7‒15.6)	9.2 (8.4‒9.9)	12.0 (10.9‒13.1)	9.2 (8.5‒9.9)	12.5 (11.9‒13.2)	12.3 (11.7‒13.0)	15.1 (14.4‒15.8)	13.5 (12.8‒14.1)
CHE, threshold 25% (rho = 0.818, p < 0.001)
Uninsured	4.6 (3.9‒5.3)	4.1 (3.5‒4.6)	3.9 (3.5‒4.4)	5.5 (4.9‒6.1)	3.6 (3.2‒3.9)	2.1 (1.8‒2.4)	2.8 (2.4‒3.3)	2.1 (1.9‒2.4)	3.1 (2.8‒3.4)	3.2 (2.9‒3.5)	3.8 (3.5‒4.1)	3.4 (3.1‒3.7)
SP/INSABI				6.3 (5.6‒7.0)	4.1 (3.7‒4.5)	2.5 (2.2‒2.8)	3.3 (2.8‒3.8)	2.5 (2.2‒2.8)	3.6 (3.3‒3.8)	3.7 (3.4‒3.9)	4.4 (4.1‒4.7)	3.9 (3.6‒4.2)
Social Security	4.7 (4.0‒5.4)	4.2 (3.6‒4.7)	4.0 (3.5‒4.6)	5.6 (5.0‒6.3)	3.7 (3.3‒4.0)	2.2 (1.9‒2.4)	2.9 (2.5‒3.4)	2.2 (2.0‒2.5)	3.2 (2.9‒3.4)	3.3 (3.0‒3.5)	3.9 (3.7‒4.2)	3.5 (3.3‒3.7)
Mixed Public				6.7 (5.9‒7.5)	4.4 (3.9‒4.9)	2.7 (2.3‒3.0)	3.6 (3.0‒4.1)	2.7 (2.4‒3.0)	3.8 (3.5‒4.1)	3.9 (3.6‒4.3)	4.7 (4.3‒5.1)	4.2 (3.8‒4.6)
CHE, threshold 40% (rho = 0.603, p < 0.01)
Uninsured	1.5 (1.1‒1.9)	1.4 (1.1‒1.6)	1.6 (1.3‒1.9)	2.1 (1.8‒2.4)	1.2 (1.0‒1.4)	0.7 (0.6‒0.9)	1.0 (0.7‒1.2)	0.9 (0.7‒1.0)	1.3 (1.1‒1.5)	1.4 (1.2‒1.6)	1.6 (1.4‒1.8)	1.4 (1.2‒1.6)
SP/INSABI				2.5 (2.1‒2.9)	1.5 (1.3‒1.7)	0.9 (0.8‒1.1)	1.2 (0.9‒1.5)	1.1 (0.9‒1.2)	1.6 (1.4‒1.8)	1.7 (1.5‒1.9)	1.9 (1.7‒2.1)	1.7 (1.5‒1.9)
Social Security	1.8 (1.4‒2.2)	1.7 (1.3‒2.0)	2.0 (1.6‒2.4)	2.5 (2.1‒2.8)	1.5 (1.3‒1.7)	0.9 (0.7‒1.1)	1.2 (0.9‒1.4)	1.0 (0.9‒1.2)	1.6 (1.4‒1.8)	1.7 (1.5‒1.9)	1.9 (1.7‒2.1)	1.7 (1.5‒1.9)
Mixed Public				3.0 (2.5‒3.5)	1.8 (1.5‒2.1)	1.1 (0.9‒1.3)	1.4 (1.1‒1.8)	1.3 (1.1‒1.5)	2.0 (1.7‒2.2)	2.1 (1.8‒2.3)	2.3 (2.1‒2.6)	2.1 (1.8‒2.3)
EHE, threshold 10% (rho = 0.724, p < 0.001)
Uninsured	17.8 (16.4‒19.2)	15.9 (14.8‒17.0)	15.0 (13.9‒16.1)	18.8 (17.8‒19.8)	12.4 (11.5‒13.2)	7.6 (6.9‒8.3)	10.0 (8.9‒11.1)	7.7 (7.0‒8.3)	10.6 (9.8‒11.3)	10.4 (9.6‒11.2)	12.9 (12.2‒13.6)	11.4 (10.8‒12.0)
SP/INSABI				22.0 (20.8‒23.2)	14.8 (13.8‒15.8)	9.4 (8.5‒10.2)	12.2 (11.0‒13.4)	9.4 (8.7‒10.2)	12.8 (12.1‒13.6)	12.6 (11.8‒13.4)	15.5 (14.8‒16.2)	13.8 (13.1‒14.5)
Social Security	18.5 (17.1‒19.9)	16.6 (15.5‒17.7)	15.6 (14.5‒16.8)	19.5 (18.5‒20.6)	12.9 (12.1‒13.8)	8.0 (7.3‒8.7)	10.5 (9.4‒11.6)	8.1 (7.4‒8.7)	11.1 (10.4‒11.8)	10.9 (10.2‒11.6)	13.5 (12.9‒14.1)	12.0 (11.4‒12.5)
Mixed Public				22.1 (20.8‒23.5)	14.9 (13.9‒16.0)	9.4 (8.6‒10.3)	12.2 (11.0‒13.5)	9.5 (8.7‒10.3)	12.9 (12.1‒13.7)	12.7 (11.9‒13.5)	15.6 (14.8‒16.3)	13.9 (13.1‒14.6)
EHE, threshold 25% (rho = 0.712, p < 0.01)
Uninsured	4.9 (4.2‒5.7)	4.3 (3.8‒4.8)	4.2 (3.8‒4.7)	6.0 (5.4‒6.6)	3.8 (3.4‒4.2)	2.4 (2.0‒2.7)	3.2 (2.7‒3.7)	2.5 (2.2‒2.8)	3.5 (3.1‒3.8)	3.6 (3.2‒3.9)	4.4 (4.0‒4.7)	3.9 (3.6‒4.3)
SP/INSABI				7.1 (6.4‒7.9)	4.6 (4.2‒5.1)	2.9 (2.6‒3.3)	3.8 (3.3‒4.4)	3.1 (2.7‒3.4)	4.2 (3.9‒4.5)	4.3 (4.0‒4.7)	5.3 (4.9‒5.6)	4.8 (4.4‒5.1)
Social Security	5.0 (4.3‒5.7)	4.3 (3.8‒4.9)	4.3 (3.8‒4.8)	6.0 (5.4‒6.7)	3.9 (3.5‒4.2)	2.4 (2.1‒2.7)	3.2 (2.7‒3.7)	2.5 (2.2‒2.8)	3.5 (3.2‒3.8)	3.6 (3.3‒3.9)	4.4 (4.1‒4.7)	4.0 (3.7‒4.2)
Mixed Public				7.2 (6.4‒8.0)	4.7 (4.2‒5.2)	3.0 (2.6‒3.4)	3.9 (3.3‒4.5)	3.1 (2.7‒3.5)	4.3 (3.9‒4.6)	4.4 (4.0‒4.8)	5.3 (4.9‒5.7)	4.8 (4.4‒5.2)
EHE, threshold 40% (rho = 0.436, p < 0.01)
Uninsured	2.1 (1.6‒2.6)	1.9 (1.5‒2.2)	2.1 (1.7‒2.4)	2.7 (2.3‒3.1)	1.7 (1.4‒2.0)	1.1 (0.9‒1.4)	1.4 (1.0‒1.7)	1.4 (1.0‒1.7)	2.0 (1.6‒2.3)	2.1 (1.6‒2.5)	2.3 (2.0‒2.7)	2.2 (1.9‒2.5)
SP/INSABI				3.5 (3.0‒4.0)	2.2 (1.9‒2.6)	1.5 (1.2‒1.9)	1.8 (1.4‒2.2)	1.8 (1.5‒2.2)	2.6 (2.2‒3.0)	2.7 (2.2‒3.1)	3.1 (2.7‒3.4)	2.8 (2.5‒3.2)
Social Security	2.3 (1.8‒2.8)	2.0 (1.6‒2.4)	2.2 (1.8‒2.6)	2.9 (2.5‒3.4)	1.8 (1.5‒2.1)	1.2 (1.0‒1.5)	1.5 (1.1‒1.9)	1.5 (1.2‒1.8)	2.1 (1.8‒2.5)	2.2 (1.8‒2.6)	2.5 (2.2‒2.8)	2.3 (2.0‒2.6)
Mixed Public				3.6 (3.1‒4.2)	2.3 (1.9‒2.7)	1.6 (1.2‒1.9)	1.9 (1.4‒2.3)	1.9 (1.5‒2.2)	2.6 (2.2‒3.0)	2.8 (2.3‒3.2)	3.1 (2.8‒3.5)	2.9 (2.5‒3.3)

The percentage of households that incur catastrophic (CHE), impoverishing (IHE), and excessive (EHE) health expenditures, with 95% confidence intervals in parentheses, is reported. Seguro Popular (SP, People’s Insurance) was implemented in 2003 and later replaced by the Instituto de Salud para el Bienestar (INSABI, Institute of Health for Wellbeing) on 1 January 2020. INSABI was mandated to guarantee free access to health services, medicines, and medical supplies for individuals without social security coverage across all levels of care. Mixed public health insurance refers to households affiliated with SP/INSABI and social security schemes. Following COICOP 2018 guidelines [32], health expenditures were classified into four categories: (i) Medicines and other health products (drugs, medical supplies, assistive devices, and therapeutic appliances); (ii) Outpatient care (preventive care, outpatient dental services, and other outpatient services excluding preventive and dental care); (iii) Inpatient care services (hospitalisation, surgeries, and overnight treatments); and (iv) Other health services (diagnostic imaging, laboratory tests, emergency transportation, and rescue services). Probabilities of CHE, IHE, and EHE were estimated using a two-stage maximum-likelihood probit (Heckprobit) model [48]. In the first stage (selection equation), the probability of any HE was modelled as a function of insurance status, survey year, household head characteristics (sex, age, education, marital status), and household sociodemographic characteristics, including the demographic dependency ratio, percentage of children < 5 years, older adults ≥ 65 years and women of reproductive age, number of adult equivalents, participation in social programmes, municipal deprivation index, and socioeconomic region fixed effects. We also included the density of health resources (outpatient facilities, consulting rooms, hospital beds, physicians, dentists, physicians-in-training, and nurses) per 1,000 inhabitants without social security, and the percentage of the population lacking social security. Interaction terms between insurance status, participation in social programmes, and survey year were included to capture differential temporal trends. In the second stage (outcome equation), we modelled the probability of CHE, IHE, or EHE conditional on positive HE, controlling for household head characteristics (sex, age, education, marital and employment status), household composition (presence of children < 5 years, older adults ≥ 65 years, and household type), socioeconomic status, urban/rural residence, and state fixed effects. Predicted probabilities were obtained using the margins command in Stata, specifying the interaction between insurance status and year. Analyses were conducted at the mean values of covariates (using the atmeans option). The estimated rho and its high statistical significance (p < 0.001) indicate that the probability of selection and the probability of the outcome are strongly correlated in their errors, confirming that selection bias is relevant and that the Heckprobit model is appropriate. All estimates incorporate survey design features and expansion factors. Data are drawn from the Mexican National Household Income and Expenditure Survey (ENIGH), 2000–2022 [26]

Results were consistent across CHE and EHE thresholds, with mixed-insurance and SP/INSABI households facing greater hardship than those with social security.

Discussion

This population-based study analysed different financial protection indicators within the context of Mexico’s structurally fragmented health system. Between 2000 and 2022, financial protection in Mexico followed a U-shaped trend—improving steadily until 2014 and deteriorating thereafter. Households with social security coverage consistently experienced the lowest impoverishing health expenditures, while those affiliated with SP/INSABI or holding mixed public insurance faced greater financial hardship. These findings are consistent with prior research highlighting the stronger protective effects of contributory social health insurance schemes relative to non-contributory or fragmented arrangements [13, 18, 19, 25]. These disparities reflect persistent inequities in financial protection within a fragmented system where insurance schemes differ in financing, benefits, and service readiness [13, 18, 51–53].

Importantly, the transition from SP to INSABI appears to have exacerbated these inequities. Rather than consolidating coverage and strengthening effective access, the reform introduced policy ambiguity and institutional discontinuities that weakened financial protection, particularly among households without social security and those in more vulnerable socioeconomic positions [18, 54, 55]. Recent evidence indicates that INSABI’s implementation was accompanied by confusion regarding entitlements, limited operational capacity, and disruptions in service provision, including access to medicines [17, 25]. These institutional shortcomings likely increased households’ reliance on private providers and OOP, contributing to the post-2019 deterioration in financial protection observed in our results [18, 56].

Similar patterns of financial vulnerability have been documented in other Latin American countries with fragmented health systems. For instance, in Colombia and Peru, multiple parallel insurance schemes have led to persistent inequalities in financial protection, with informal and mixed-insurance households facing the greatest risk of CHE [19]. By contrast, countries that have advanced towards more integrated financing arrangements and harmonised benefit packages—such as Chile and Costa Rica—report lower levels of CHE and more equitable access to care [19]. These regional comparisons suggest that Mexico’s experience is emblematic of broader challenges faced by LMICs attempting to achieve UHC under conditions of institutional fragmentation [17, 25, 56].

The coexistence of multiple, weakly coordinated insurance schemes in Mexico has not mitigated financial vulnerability; rather, it appears to have amplified it. Households with mixed public insurance—those with overlapping affiliations to social security and SP/INSABI—often face equal or greater financial risk than households covered exclusively by SP/INSABI. This finding underscores that overlapping coverage does not guarantee effective protection when benefit packages are inconsistent, administrative responsibilities are fragmented, and referral pathways between subsystems are poorly coordinated [17, 25, 55, 56]. In such contexts, institutional misalignment undermines continuity of care and increases out-of-pocket spending despite formal insurance coverage [25, 56]. These results reinforce the argument that progress towards UHC requires not only expanded enrolment, but also harmonisation of benefit packages, improved coordination across schemes, and strengthened financing and capacity of public providers [57–60].

Despite substantial expansions in public healthcare coverage during the study period, we observe a sustained increase in the proportion of households seeking private healthcare, regardless of insurance status. This trend was particularly pronounced among households with mixed public insurance, which consistently reported greater reliance on private providers than households covered solely by SP/INSABI (Appendix 1). One plausible explanation is dissatisfaction with the quality, accessibility, or responsiveness of public healthcare services. Although ENIGH does not capture all drivers of healthcare-seeking behaviour, prior research suggests that institutional barriers between public schemes, fragmented benefit packages, and perceptions of lower quality or longer waiting times in public facilities may prompt households to seek private care and incur OOP [13, 17]. Evidence from Mexico indicates that long waiting times, staff shortages, and frequent medicine stockouts in public facilities push insured households towards private consultations [25, 61]. This “bypassing” behaviour—seeking private care despite coverage—may therefore represent an important mechanism linking insurance fragmentation to higher CHE and IHE among mixed-insurance households [18, 62, 63].

While the cross-sectional design of ENIGH does not allow for causal testing of these mechanisms, our findings motivate several testable hypotheses for future research. These include whether fragmented benefit packages, referral barriers between public schemes, or perceived quality differentials systematically drive private healthcare utilisation, and whether private expenditures disproportionately contribute to CHE among vulnerable subgroups. Household surveys could be enhanced by collecting more detailed data on healthcare utilisation, including the type of facility used (public vs. private), service category (outpatient, medicines, diagnostics, inpatient care), reasons for choosing private services, household health status (e.g., chronic conditions, disability) and linked OOP [64]. Such data would allow a more precise assessment of the causal pathways linking insurance fragmentation, private care utilisation, and financial risk, informing targeted policy interventions to improve coverage integration, reduce CHE and improve financial protection.

Medicines constitute the largest share of out-of-pocket healthcare expenditure among SP/INSABI and mixed-insurance households (64.5% and 63.6%, respectively). Although our analysis documents this expenditure pattern, it does not aim to establish a causal relationship between medicine spending and CHE, as medicine expenditure is a component of total OOP by construction. Nonetheless, the prominence of medicines as a driver of financial burden warrants careful institutional interpretation. Emerging evidence suggests that high medicine-related OOP reflects a combination of limited effective coverage for chronic disease treatments, supply-side failures in public facilities, and disruptions in procurement and distribution mechanisms following recent reforms [17, 25]. Differences in medicine reimbursement policies and formulary enforcement between social security institutions and SP/INSABI may further exacerbate these patterns [55]. Future research could examine whether high medicine spending is driven by: (i) limited coverage or reimbursement policies for chronic disease medications in SP/INSABI; (ii) supply-side limitations in public facilities that force households to purchase medicines privately; or (iii) behavioural factors, such as perceptions of quality, adherence to prescribed treatments, or preferences for brand-name drugs. Stratified analyses by household health status, particularly the presence of chronic conditions, would help determine whether medicine expenditure is indeed a principal contributor to CHE among high-risk subgroups. Furthermore, future waves of ENIGH could be enhanced by collecting granular data on medicine use, including: (i) type and chronicity of medications consumed, (ii) whether medicines are obtained free-of-charge or purchased OOP, (iii) insurance scheme-specific reimbursement limits or formulary restrictions, and (iv) stockouts or access issues in public facilities. Collecting these variables would enable researchers to link medicine expenditure to both coverage policies and financial protection outcomes, thereby identifying high-risk subcategories and informing targeted policy interventions—such as expanding reimbursement lists or strengthening public medicine supply chains. By incorporating these improvements, future analyses could provide a more precise understanding of the contribution of medicine expenditure to CHE, helping to guide policies aimed at reducing financial risk, particularly among vulnerable households [17, 25].

The sharp rebound in the proportion of uninsured households—from 8.3% in 2016 to 26.1% in 2022—raises additional concerns. Although the cross-sectional nature of ENIGH prevents reconstruction of individual insurance trajectories, recent evidence suggests that this reversal may reflect coverage losses among populations previously insured under SP, potentially due to administrative barriers and implementation challenges associated with INSABI [65]. Future research could explore the characteristics of newly uninsured populations, such as whether they were previously covered by SP and lost coverage due to complex INSABI application procedures, or whether they are predominantly low-income households with heightened financial vulnerability. Stratifying uninsured households by prior coverage status (“long-term uninsured” versus “newly uninsured”) could reveal differences in CHE and EHE risk across subgroups, thus capturing internal heterogeneity among the uninsured and the true impact of reversed insurance coverage. To enable this, future ENIGH waves could incorporate retrospective questions on prior coverage, reasons for losing or changing insurance, and household income and expenditure shocks linked to health events. These enhancements would allow researchers to more accurately assess financial risk among newly uninsured households and inform targeted policy interventions to protect vulnerable groups.

These findings have clear policy implications. In the short term, the most urgent priority is to ensure equitable access to essential medicines through pooled purchasing and reliable public supply chains, as previous studies have documented that medicine costs remain the leading cause of CHE and IHE [25]. Expanding availability and affordability of medicines would have an immediate impact on household financial burden. In the medium term, reforms should focus on integrating health financing—potentially through a single-purchaser model or harmonised benefit packages across schemes—to reduce inefficiencies and inequalities created by fragmented risk pools [17, 55, 65]. These structural reforms would yield longer-term gains by improving equity and financial sustainability. Finally, targeted efforts to improve service quality and accessibility in public facilities are essential to reduce reliance on private care and ensure that coverage translates into effective financial protection for all households.

This study has several limitations. Although ENIGH provides high-quality, nationally representative data on household expenditure, self-reported measures may be subject to recall and reporting bias. Estimates also depend on standard—but debated—definitions of subsistence expenditure and capacity to pay; however, extensive sensitivity analyses confirmed that the main trends and relative differences across insurance groups are robust to alternative specifications.

Measures of health resource availability are derived from interpolated administrative data and should therefore be interpreted as proxies, particularly given the absence of comparable information for social security institutions [17]. Importantly, sensitivity analyses that sequentially excluded or added health resource density variables yielded consistent results, suggesting that the main conclusions are not driven by measurement error in these proxies. Nonetheless, these indicators may not fully capture subnational variation in actual access or service readiness, especially for populations covered by social security [17]. In addition, the classification of household insurance status cannot distinguish whether multiple coverages apply to the same individual or to different household members, nor does it capture differences in service quality or effective utilisation across schemes [65]. Our analysis focuses on catastrophic, impoverishing, and excessive health expenditure and does not account for coping strategies such as borrowing, asset depletion, or forgone care, which may lead to an underestimation of the true burden of health-related financial risk.

Despite controlling for a wide range of household- and contextual-level characteristics, unobserved factors—such as health status, severity of illness, or care-seeking preferences—may still confound the estimated associations, as ENIGH does not include direct measures of morbidity or disability. Incorporating population health information at the household level would strengthen future analyses of the relationship between health needs and financial hardship.

Importantly, our study design does not allow us to establish a causal relationship between health system fragmentation and financial protection outcomes. Nevertheless, our findings show that persistent and systematic inequalities in financial protection endure within a fragmented system, even when overall trends in HE evolve similarly across groups. The stable ranking of financial risk by insurance status over more than two decades indicates that fragmentation shapes how common system pressures are translated into unequal financial consequences, thereby reinforcing long-standing inequities in protection against health-related financial hardship.

Finally, the 2020 and 2022 measurements of overall and household OOP were partially influenced by the COVID-19 pandemic. Although our models incorporate state- and time-fixed effects, extensive household-level covariates, and Heckprobit specifications to account for selection and potential confounding, it remains difficult to fully disentangle the effects of COVID-19 from those of the policy transition to INSABI [18]. Nonetheless, the timing of events suggests that the observed increases in CHE, IHE, and EHE are more plausibly linked to health policy changes than to the pandemic, in particular those occurring after 2022: reductions in the Ministry of Health budget from 2015 [24, 25, 37, 65], the dismantling of SP, and the cancellation of conditional cash transfer programmes were already affecting households prior to 2020 [65, 66], whereas the pandemic began to exert measurable effects only by mid-2020. ENIGH 2020, conducted between 21 August and 28 November [31, 67], captured expenditures in the preceding quarter, largely reflecting the cumulative effects of policy changes with only early signals of the pandemic, while ENIGH 2022 captured spending during the post-pandemic period, according to the timeline defined by Mexico’s health authorities (COVID-19 period—from April 2020 to March 2022—and post-COVID period—from April 2022 to October 2023) based on trends in infections, hospitalizations, and mortality, and their implications for healthcare service utilisation [68]. Future research could integrate complementary data sources or exploit variation in pandemic intensity and policy rollout across states to more precisely isolate the independent effects of health system reforms and pandemic shocks on financial protection outcomes.

Due to households’ considerable financial burden from OOP on medicines, Mexico should implement policies to expand coverage and improve the affordability of essential medications. This could include negotiating better prices through pooled purchasing and ensuring a consistent supply of medicines across all public insurance programs. By lowering the financial barriers associated with medication costs, these policies would directly impact one of the leading causes of CHE and IHE. Additionally, improving the responsiveness, quality, and accessibility of public healthcare facilities would reduce the incentive for vulnerable households to seek private care. This would lower OOP and alleviate financial hardship, especially among vulnerable populations. The Mexican health system’s recent reforms have reduced financial protection, increasing CHE and EHE. Strengthening public health financing and addressing systemic fragmentation are critical to achieving UHC. Future research should explore policy mechanisms to enhance financial protection and reduce healthcare access disparities.

Abbreviations

UHC

Universal Health Coverage

ENIGH

National Household Income and Expenditure Survey

CHE

Catastrophic health expenditure

IHE

Impoverishing health expenditure

EHE

Excessive health expenditure

OOP

Out-of-pocket expenditures

LMICs

Low- and middle-income countries

SP

Seguro Popular

INSABI

Instituto de Salud para el Bienestar

INEGI

National Institute for Statistics and Geography

CTP

Capacity to pay

COICOP

Classification of individual consumption by purpose

TE

Total expenditure

SE

Subsistence expenditure

FE

Food expenditure

HE

Health expenditure

PPP

Purchasing power parity

SES

Socioeconomic status

SINERHIAS

Health Equipment, Human Resources and Infrastructure Information Subsystem

CIs

95% confidence interval

Authors’ contributions

DCG jointly conceived the idea for the paper. ESM led the formal analysis and performed data curation. ESM supported the conceptual framework. DCG, ESM and OGD collaboratively drafted the initial manuscript. OGD, TH, CPA, AMB and LF provided critical feedback across multiple revisions. ESM serves as a guarantor of the work, having full access to all study data and taking responsibility for the data&apos;s integrity and the analysis&apos;s accuracy. All authors reviewed, revised, and approved the final manuscript.

Funding

This study was funded by the NIHR GHPSR researcher-led grant NIHR150067 using UK aid from the UK Government to support global health research. The views expressed in this publication are those of the author(s) and not necessarily those of the NIHR or the UK government. Diego Cerecero-Garcia is funded by the President’s Scholarship (Imperial College London).

Data availability

Data analyzed were obtained from the public repository hosted by the National Institute of Statistics and Geography of Mexico (INEGI for its acronym in Spanish), available at https://www.inegi.org.mx/programas/enigh/.

Declarations

Ethics approval and consent to participate

This study involved no human participants, and was approved by the Research, Biosafety and Ethics Committees of the National Institute of Public Health in Mexico (ID: 2358/1826/S21-2022).

Consent for publication

Not applicable.

Competing interests

OGD served as Director General for Performance Evaluation at the Ministry of Health of Mexico during 2000–2006, the initial implementation period of the Seguro Popular. The remaining authors declare no competing interests.