COVID-19 needs no passport: the interrelationship of the COVID-19 pandemic along the U.S.-Mexico border

We analyzed standardized COVID mortality rates along the U.S. Mexico border. We found that along the U.S.-Mexico border region there were higher COVID-19 mortality rates than non-border areas. Despite differing policies on either side of the border, proximal cities experienced linked waves of COVID-19 in terms of peaks and troughs. To inform the disparity between border and non-border areas, we plotted standardized mortality ratio (SMR) against human development index (HDI), a socioeconomic proxy. In US border states, there was a negative correlation between SMR and HDI resulting in a higher SMR in the border region compared to the rest of the counties. However, in Mexican border states, there was no association between SMR and HDI. In the U.S. border states, there was a negative correlation between SMR and vaccination percentage opposite to national trends. Additionally, in both the U.S. and Mexico, border regions were more vaccinated than non-border regions. Our findings overall highlight the importance of targeted regional interventions that consider both sides of the border in unison, and counter the notion that nations can isolate themselves through strict border policies.

Waves of COVID-19 were seen in border cities usually beginning on the Mexican side and surging on the U.S. side shortly after. Reasons for this trend are unclear but may lie in the unidirectional policies that existed surrounding border travel. Starting in March 2020, the U.S. government enacted restrictions on inbound traffic limited to “essential travel”. Reasons deemed as “essential” were citizen or lawful resident return, medical, educational, work, emergency or public health, cross-border trade, government or diplomatic, or military [32]. Conversely, the Mexican government did not impose any inbound travel restrictions [33]. This resulted in U.S. citizens and permanent residents being able to cross the border freely, while Mexican citizens were not able to cross into the U.S. Furthermore, the rates at which state and municipality restrictions were implemented varied between the U.S. and Mexico [34]. In March, California was one of the first states in the U.S. to enact statewide stay-at-home orders, while physical distancing orders in bordering Tijuana and Mexicali were not in effect until weeks later [35]. As initial cases in Mexico were likely imported from the U.S., the heavy interaction between individuals on the Mexican side early in the epidemic lead to Tijuana and Mexicali being early epidemic centers for Mexico like what New York was in the U.S. [35, 36] Later in April, as case counts remained low in California and as restrictions began to lift, cases were likely imported from Mexicali to bordering Imperial County and southern San Diego leading to pressure on hospital centers in those areas. A similar phenomenon occurred on the eastern border leading to high mortality rates in Laredo and Brownsville (Fig. 3).

Reasons for disparities in COVID indicators at the border areas compared to non-border areas are likely sociostructural. The U.S. border region is predominantly populated by Latino individuals and much has been reported on the U.S. side of the increased risk of COVID-19 in Latino and Black communities [37]. Yet, the virus does not have special receptors to target based on race or ethnicity, and it has yet to be proven if there is a particular genetic susceptibility in these racial/ethnic groups that increases these populations’ risk. Hence, the increased risk of disease acquisition is most likely due to sociostructural factors that lead to situations of numerous and prolonged interactions with other individuals that allow for disease transmission. For example, high density housing leads to rapid spread of the virus [38]. Additionally, many individuals due to their documentation status did not qualify for certain economic benefits to mitigate the economic impact of the epidemic. This forced these vulnerable individuals to continue working and be at high-risk for SARS-CoV-2 acquisition. U.S. border counties are amongst the poorest in the country, this tied with high rates of undocumented individuals in these areas, lead to a large segment of the population who lack basic primary and preventative medical care [39]. For mortality outcomes, pre-existing conditions like diabetes or kidney disease play a significant role in Latino communities and has been consistently associated with a higher risk of mortality from COVID [40].

On the Mexican side additional challenges were present. Testing was severely limited to the point where Mexico was last in testing per capita out of all OECD nations early in the pandemic [41]. This was evident in our data, where mortality rates in Mexico were similar to those in the U.S.; but, case rates in Mexico were on an order of magnitude lower than those in the U.S. This points to a severe lack of testing on the Mexican side, and only limited to symptomatic and sicker individuals. Health infrastructure is severely lacking along the border [42]. Lacking access on the Mexican side of the border is possibly leading many individuals to cross the border to access hospital centers in the U.S. also likely inflating that sides case mortality rates [43]. However, border municipalities have a higher development index then the rest of Mexico. This is partly due to the high number of high-density maquiladoras (factories) that manufacture products, including health products for the U.S. market. These maquiladaroas were pressured by the U.S. to continue operating. In turn, these maquiladoras were early centers for COVID-19 clusters as workers cited not having adequate personal protective equipment and an inability to physically distance within their work environments [44]. Workers’ rights groups have striked as result of these conditions and now that vaccines are available, specific efforts are being made to vaccinate workers in maquiladoras [45].

The COVID-19 pandemic had important economic implications in Central America that lead to waves of asylum seekers migrating to the U.S.-Mexico border [15]. Prior to the pandemic, the US instituted a Remain at Home Policy where asylum seekers remained in a third country (in this case Mexico) while their application was being reviewed. This led to refugees remaining along the U.S.-Mexico border in dense migrant camps [15]. Migration across the border has been a point of contention in the political realm of the U.S. with concerns that cases were being imported. While individuals crossing the border were likely not responsible for COVID-19 surges throughout the U.S. [46], the conditions asylum seekers experience are cause for concern and led to difficulties in protecting these individuals from outbreaks in congregate settings [15, 47]. There have been efforts to vaccinate individuals migrating to the U.S. [48], but policies still restrict asylum seekers from entering despite recent opening of the U.S.-Mexico border to vaccinated individuals [49].

Wide sweeping vaccination which cuts across all populations in a society is an essential component to handling any infectious disease. Vaccination for COVID-19 in the first year of availability was variable geographically throughout the U.S. In our data we observed an inverse correlation between vaccination rates and HDI at the border which went against national trends. Possible explanations for these observations lie in states’ and counties’ efforts to target vulnerable socioeconomic groups for increased vaccination resources. Arizona created a subset 1 S which listed groups such as minorities, tribal communities, and uninsured individuals as priorities during all phases of vaccination [50]. California enacted a plan using a Healthy Places Index to divide zip codes throughout the state in four quartiles. Allocation was then conducted 70% by age and 30% by sector with a double allotment for the lowest quartile [50]. New Mexico included a caveat which, based on the CDC’s Social Vulnerability Index (SVI), reallocated up to 25% of supply to higher SVI areas or regions with > 10,000 cases per 100,000 [50]. Texas’ plan included identification of vulnerable groups such as minorities, those experiencing homelessness, and those with disabilities who could contribute to the vaccine allocation decision making process [50]. However, after initial vaccine rollout, local Texas jurisdictions argued that vaccination clinics were predominantly located in wealthier and whiter areas [51‐53]. Dallas attempted to modify plans and target Black and Latino neighborhoods; however Texas health officials stated this practice was unacceptable and threatened to cut off all supply [52]. Houston and Austin also enacted policies which identified vulnerable areas based on socioeconomic status, zip code, and COVID-19 burden [54, 55]. With more supply than demand, Texas health officials voted to no longer allocate vaccinations starting on May 10, 2021, and instead created a system in which providers could order shipments [56]. As of writing, many of the racial gaps in vaccination have narrowed or closed [57]. Of note, it is to the authors’ knowledge that many Mexican nationals or residents crossed into the U.S. to get vaccinations early during the vaccination rollout [58]. Potential reasons included inability to access vaccines in Mexico or to access vaccines that were not available in Mexico (e.g., Pfizer or Moderna) at the time. The magnitude of this phenomena is unknown and may overestimate vaccination coverage, especially in the border regions.

On the Mexican side of the border, vaccination began with a rapid start but since slowed. The country was first in Latin America to begin vaccination and doses administered peaked in July. However, poor supply chain and mismanagement of planning seem to have plagued further coverage [59]. Based on our analysis we observed that states along the U.S. border had higher ratios of vaccine doses. This information correlates with the broad push to vaccinate border residents and reopen border traffic and commerce. The U.S. government offered a multitude of doses to support the northern region of Mexico [60]. Additionally, starting in November, vaccines were offered to migrants as a part of the Remain in Mexico program [61].

Limitations of our study lie in the data sourcing and how representative it is of COVID-19 morbidity and mortality. We hoped to avoid testing limitations by focusing specifically on mortality. However, due to healthcare access and public health policies individuals may not be counted or missed [62]. Additionally, the methods through which counties, municipalities, states, and countries count deaths due to COVID-19 vary and can be incredibly complex [63]. When investigating reasons for disparities in COVID-19 outcomes, our metrics for socioeconomic conditions were not perfect comparisons due to differing education systems in the U.S. and Mexico and many other variables likely play a role into COVID-19 related mortality. For example, forced migration is an important factor in healthcare disparities and outcomes that was not assessed. Lastly, Mexican vaccination data more granular than national statistics are severely limited which impacts our ability to investigate its contribution to COVID-19’s spread in the region.