Design and sample

The present study is a retrospective secondary data analysis or a city-wide Covid-19 registry within Mexico City. Data was gathered from the Sistema Nacional de Vigilancia Epidemiológica (Sinave) of Mexico, designed to track Covid-19 possible cases in the city and made available by the Health Ministry of Mexico City on its open data portal. This database includes characteristics of patients attended for the possible presence of Covid-19 and the institution that provided medical care and was last consulted on January 22, 2021. At the moment of enquiry, the database had the information of 1 491 245 suspected cases for Covid-19 defined by the presence of clinical symptoms described elsewhere.²¹

We included those individuals with a positive RT-PCR or rapid antigen test for SARS-CoV-2 between the months of March and December 2020. Those individuals with missing data on the result of RT-PCR o rapid antigen test for SARS-CoV-2 were excluded. The Unique Key of Health Establishments (CLUES) catalogue was used to obtain the level of care (first, second or third) of health institutions in Mexico City. This catalogue is made available to the public by the Dirección General de Información en Salud (DGIS) Information of the national ministry of health, and its purpose is to identify all health establishments in the country and their main characteristics.²²

Dependent variable: Case Fatality Rate

The study’s dependent variable was defined as the death or survival status among patients with a positive Covid-19 test. In order to construct it, a dichotomous variable was generated based on the case evolution indicated in the Sinave database. A value of “1” was assigned when the case evolution result was death and “0” for any other result.

Institutional variables

The institutional variable of interest for this study corresponds to the institution where patients were treated; generating a categorical variable considering five options: private sector, IMSS, ISSSTE, SS, and other health institutions (Red Cross, IMSS -Oportunidades, Petróleos Mexicanos [Pemex], Secretaría de la Defensa Nacional [Sedena], Secretaría de Marina [Semar] and university hospitals). This variable and the municipality where the establishment is located were obtained from the Sinave database. The level of care of the establishment (primary, secondary or tertiary), and the number of clinics and beds, were obtained from the CLUES catalogue; these variables were matched with the Sinave database using the name of the establishment present in both datasets. Finally, two variables regarding the number of patients per medical office and the number of patients per bed were created by dividing the number of SARS-CoV-2 positive patients by the number of medical offices and beds. This was done in order to have variables considering the patient saturation of medical establishments.

Cofounding variables

Study population

Our analysis included 349 890 persons with a positive Covid-19 lab test during the period March-December 2020. From this total, 43 974 were hospitalized, and 305 916 were ambulatory patients. The mean age of patients attended for the possible presence of Covid-19 during March-December 2020 was 42.5 (±16.9) years old. Those who died were older than those who survive either among the hospitalized (63.2 ± 13.9 vs. 61.3 ± 14.5 years old) or ambulatory (53.3 ± 16.4 vs. 40.3 ± 15.8). Among all, 50.9% were female, and 59.7% were non-health-related workers (table I).

The attention of Covid-19 within healthcare facilities in Mexico City

Concerning the place of attention, 75.5% of the cases were attended on first-level clinics, 13.3% in second-level hospitals and 11.3% on third-level hospitals, with significant differences between healthcare subsystems (p<0.001) going from 6.7% of ministry of health patients attended on third-level hospitals to 15.7% of patients from IMSS, 16.2% of private institutions patients, 65.1% of patients from ISSSTE and 82.5% of patients of other institutions attended at this level. There is a significant difference in the level of attention between hospitalized and ambulatory patients (p<0.001). Considering hospitalized patients, 4.7% were attended in first-level clinics, 55.1% in second-level hospitals and 40.2% in third-level hospitals. Meanwhile, 85.6% of ambulatory patients were attended in first-level clinics, 7.2% in second-level hospitals and 7.2% in third-level hospitals. Most cases (76.5%) were attended in facilities of the ministry of health, followed by IMSS (17.7%), ISSSTE (1.8%) and private facilities (1.7%). In contrast, higher proportion of the deaths occurred when the attention was given in the IMSS (59.1% and 65.5% for the hospitalized and the ambulatory, respectively), followed by the ministry of health (23.6 and 24.1% for the hospitalized and the ambulatory, repectively), the ISSTE (10 and 3.9% for the hospitalized and the ambulatory, respectively), and the private sector (1.5 and 0.9% for the hospitalized and the ambulatory, respectively). The municipality in Mexico City presenting the higher proportion of cases (11.2%) was Cuauhtémoc, and the one who presented the least (2.3%) was Milpa Alta. The municipality with higher proportion of deaths was Azcapotzalco (16.9%; 2 986 hospitalized and 239 ambulatories), and the one with the least was Xochimilco (0.04%; 0 of the hospitalized and 7 in the ambulatory) (table I and table II). For all the variables, we observed significant differences in the distribution of variables under study for those who died for Covid-19 and those who survived the disease, with the exception of obesity for hospitalized cases.

Influence of healthcare subsystem on Covid-19 CFR

The multivariable logistic regression model for CFR in hospitalized patients showed that the probability of dying from Covid-19 for a patient treated at a facility within the IMSS (OR 5.11, 95%CI 4.16,6.27), ISSSTE (OR 2.10, 95%CI 1.69,2.62), SS (OR 1.94, 95%CI 1.57,2.40) or other institutions (OR 1.7, 95%CI 1.33,2.18, table III) is higher than those in private institutions, even after adjustment for age, sex, comorbidities and symptom severity. As for ambulatory patients, the probability of dying was higher for patients treated within IMSS facilities (OR 36.57, 95%CI 16.62,80.46), ISSSTE (OR 9.19, 95%CI 3.92,21.55), SS (OR 5.29, 95%CI 2.37,11.81), and other institutions (OR 9.56, 95%CI 4.11,22.23) than for those attending in private institutions (table IV). The probability was higher for men; for those dedicated to the household, unemployed or retired subjects; for those requiring intubation; those with diagnosis of clinical pneumonia or a higher number of respiratory symptoms; those with diabetes or hypertension as preexisting conditions, and were attended in facilities with a higher number of patients per bed. As expected, the probability of dying increased with age. The pre-existence of obesity or asthma, being attended in a third-level care centre, and the number of patients per medical office were not statistically significant.

Changes across time and level of care in Covid-19 CFR

CFR associated with Covid-19 within hospitalized patients decreased after the initial peak in March-April, when it was higher than any of the months from May to October; however, increased CFR was observed in November compared to March and April after adjusting for relevant covariates. Treatment within a third-level healthcare hospital was associated with an increased risk for CFR associated with Covid-19 compared to second-level care (OR 1.11, 95%CI 1.03,1.19, table III). In contrast, for ambulatory patients, CFR was consistently lower in subsequent months. Finally, when considering the level of care within the model for ambulatory patients, odds for CFR associated with Covid-19 was higher for the third and second levels of care (OR 4.60, 95%CI 3.80,5.56; and OR 3.36, 95%CI 2.87,3.92, respectively) compared to ambulatory cases attending first-level clinics, and after controlling by relevant covariates (table IV).

Supplementary material

Health system in Mexico is divided in three subsystems: employment based social insurances (IMSS, ISSSTE), public assistance services for uninsured (SS), and a private sector with service providers and insurers.8

Among the employment-based social insurances, IMSS is the oldest and largest institution in the country, insuring 68 158 455 beneficiaries in 2018.⁹ It was created in 1943, providing insurance to formal workers of private companies, retired formal workers, students at public universities, domestic workers, and the families of those affiliated, as well as some families that voluntarily pay for the insurance.³⁵ Therefore, nearly 50% of Mexicans are affiliated with it. IMSS financing comes from a tripartite scheme with the workers, their bosses, and the federal government contribute.

ISSSTE was created in 1959 and provides insurance to workers of the executive, legislative and judicial power branches of the federal government and the workers of the federal autonomous dependencies such as the office of the Attorney General, the Constitutional Autonomous Organizations, and the Jurisdictional Autonomous Organs.¹⁰ However, it does not provide insurance to the military forces (Sedena and Semar) and the national oil company Pemex; each has its insurance institutions. The institutions that ensure State workers are financed through a double scheme, with contributions from the worker and the State.³⁶

It was until 1983 when the government declared the right of health protection for all in the Federal Constitution.¹¹ Since then, there have been several efforts to achieve universal health access coverage for all Mexicans. In response to these needs, in 2004, the Seguro Popular (Popular Insurance) was created. Any Mexican could affiliate to this institution that was financed by public taxes and by client fees that were calculated according to the insured’s socioeconomic status. This insurance would only cover some of the most common and severe health ailments; if the health service needed was not in CAUSES, the patient had to pay out-of-pocket. By 2019, 89% of the population had health coverage.¹² The Seguro Popular Services were provided in health facilities run by the Federal Ministry of Health or by local government’s ministries or institutions.³⁷

November 29 of 2019, the Federal Government reformed the National Health Act to substitute the Seguro Popular for the Instituto Nacional de Salud para el Bienestar (Insabi).¹³ Insabi aims to achieve universal health coverage, universal coverage of services, significant reductions in out-of-pocket spending, eradicating corruption from health services, and improving people’s well-being.¹⁵ In March 2020, a reengineering plan for the Health Ministry and health services was proposed son the Insabi could be implemented effectively to achieve its goals. This plan centres the health services on critical and integrated care in primary health care.¹⁴ Along with these changes, the Federal Government announced the centralization of purchases of medicine and health care supplies and the regularization of health workers as federal employees (in contrast to local government employees in Seguro Popular).¹⁵ The Organization’s Governing Statutes for the Insabi were not approved until November 2020.¹⁶ Therefore, the SARS-CoV-2 pandemic arrived during one of the most ambitious and essential health transitions in recent Mexican history.