key: cord-0734450-m71rwh67 authors: Yu, X. title: Did elderly people living in small towns or rural areas suffer heavier disease burden during the COVID-19 epidemic? date: 2020-05-06 journal: nan DOI: 10.1101/2020.05.01.20087791 sha: cada8208da7b9b4a68f7980d8beacaba732b72ab doc_id: 734450 cord_uid: m71rwh67 Background: Health disparities were often overlooked during the emerging epidemic. Objectives: This study examined geographic differences in the rates of health care use and deaths among elderly patients. Methods: Based on individual patient records, multivariate Poisson and logistic models were used to calculate adjusted incidences of COVID-19 and probabilities of emergency department (ED) visits, hospitalizations and deaths. Results: Of 8,203 elderly patients, 11% died. Elderly people living in small metropolitan areas were half as likely to be diagnosed with COVID-19. Elderly female patients living in small metropolitan areas had much lower rates of ED visits (23% vs. 34%; Odds Ratio (OR): 0.58; 95%confidence interval (CI): 0.41-0.81; p=0.002) and hospitalizations (22% vs. 31%; OR: 0.62; 95%CI: 0.44 - 0.87; p=0.006) than those living in large metropolitan areas. Furthermore, those living in non-metropolitan areas were more likely to be hospitalized than those living in large metropolitan areas (44% vs. 33%; OR: 1.46; 95%CI: 1.07-1.99; p=0.016), especially among elderly men (51% vs. 35%; OR:1.86; 95%CI: 1.18-2.93; p=0.008). Finally, there was a significant linear trend in hospitalization rates among elderly male patients (p for trend = 0.01). Conclusions: Profound health disparities exist in the time of emerging epidemic. Since December 2019, the novel Severe Acute Respiratory Syndrome associated coronavirus (SARS CoV2) (Zhu et al., 2020) has infected over 3 million people and claimed more than 216,000 lives worldwide (https://coronavirus.jhu.edu/, accessed on April 28, 2020). Unlike the 2003 SARS virus that had limited transmissibility before symptom onset (Peiris, Yuen, Osterhaus, & Stohr, 2003) , the novel SARS CoV2 can be transmitted from pre-symptomatic and asymptomatic patients (Bai et al., 2020; Li, Li, He, & Cao, 2020) and cause sudden symptom exacerbation among mildly symptomatic patients, often leading to cytokine storm and acute respiratory distress syndrome (ARDS) (Guan et al., 2020; C. Huang et al., 2020) . The unprecedent scale of epidemic has forced many countries to adopt aggressive mitigating measures such as social distancing, closing schools and business, and prohibiting large gatherings (Anderson, Heesterbeek, Klinkenberg, & Hollingsworth, 2020; Ferguson et al., 2020; Pan et al., 2020) . Consequently, the epidemic in the US has slowed down significantly and many states have reached a turning point by April 15, 2020, as demonstrated in our previous study (Yu, 2020) . Elderly people were disproportionally affected by the epidemic, as about 80% deaths occurred among people aged 65 or above (https://www.cdc.gov/nchs/nvss/vsrr/covid19/index.htm). Due to their physiologically weak immunity and high prevalence of comorbidities in which two thirds of elderly people had two or more chronic conditions (Chavan, Kedia, & Yu, 2017) , elderly people might be more likely to have severe disease if infected by the virus. However, timely diagnosis and treatment might be impeded by myriads of health care access barriers such as lack of transportation, difficulties in communicating with health care providers, and complexity of health care system (Fitzpatrick, Powe, Cooper, Ives, & Robbins, 2004 ; Hill, Perez-Stable, All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020 May 6, . . https://doi.org/10.1101 May 6, /2020 doi: medRxiv preprint Anderson, & Bernard, 2015) . Many elderly patients had prolonged stay in the intensive care unit and one fourth of them eventually died (Richardson et al., 2020) . Unfortunately, health disparities might be overlooked during the emerging epidemic (Hill et al., 2015) . Awakened by this, many states started reporting the numbers of cases, hospitalizations and deaths by age and racial/ethnicity groups, depicting a disproportional disease burden among certain racial groups and some vulnerable elderly populations (e.g., https://floridahealthcovid19.gov/). On the other hand, despite wide availability of maps representing the epidemic process, health disparities by geographic areas have not been rigorously examined. Existing reports and maps often focused on the description of the epidemic(e.g., https://coronavirus.jhu.edu/map.html and similar websites driven by a GIS system), but none have carefully explored the disparities underlying the reported case counts with appropriate epidemiological methods. In this study, we aimed to uncover health disparities during the epidemic between urban and rural elderly people. Using individual patient records obtained from the Florida Department of Health, we employed multivariate Poisson and logistic regressions to calculate adjusted incidence of COVID-19 and probabilities (rates) of COVID-19 related emergency department (ED) visits, hospitalizations, and deaths among elderly people residing in Florida. We hypothesized that those living in small metropolitan or non-metropolitan areas might have lower rates of visiting an ED, being hospitalized and higher mortality rates than those living in large or medium metropolitan areas. Data sources: All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. We merged all these data sources by gender, age, and county. As of April 25,2020, there were 32,138 confirmed cases in the file. We excluded 30 patients who did not have age information, and additional 1 patient without county information, resulting in 32,107 COVID-19 cases of all age groups and 8,203 elderly cases included in the analysis. Patient's age was grouped into <25, 25-49, 50-64, 65-74, and 75+. However, except for Table 1 which gave an overview of COVID-19 epidemic in Florida, the main analyses were restricted to patients aged 65 or older, as the focus of this study was about health disparities among elderly people. Metropolitan status of each county was classified as large metropolitan areas and their suburbs (1 million or more people), medium metropolitan areas (250,000 -1 million), small metropolitan All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. . https://doi.org/10. 1101 /2020 areas (50,000 -250,000), and non-metro areas (counties with 50,000 or less people, including rural areas). The first COVID-19 case was recorded on March 2, 2020, and the whole epidemic was divided into early stage (before April 1, 2020) and late stage (after April 1, 2020). During the late stage, many control measures were strictly enforced, including stay-at-home rule issued by the Florida state government on April 3, 2020. In addition to describing numbers of cases, hospitalizations and deaths, we calculated the adjusted incidence rates (per 1,000 persons) for COVID-19 cases with age and sex specific population in each county as the proper denominator. Poisson regressions were used to calculate incidences based on summary counts, adjusted for county, age, sex, metropolitan status and period. Incidence rates were predictive margins from the model. Similarly, logistic regressions were used to calculate probabilities (rates) of ED visits, hospitalizations, and deaths using the individual line list file. The probabilities were obtained from predictive margins as well. Furthermore, we mapped the adjusted hospitalization rate per 100 cases for each county based on a Poisson model. SAS 9.4 and Stata 16.1 were used in the analysis. A p value less than 0.05 was considered statistically significant. No multiple comparisons were adjusted. Table 1 presented an overview of COVID-19 epidemic in Florida as of April 25, 2020. Of 32,107 confirmed cases, 25.8% cases had visited ED, 16.2% were hospitalized, and 3.5% died. Although cases aged 65 or older accounted for only 25% of total cases, they were twice to three All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. . https://doi.org/10.1101/2020.05.01.20087791 doi: medRxiv preprint times more likely to be hospitalized, and five to ten times more likely to die than younger people. Overall, they accounted for 54% of hospitalizations and 82% deaths. Among those 8,203 elderly patients, one third of them visited an ED or were hospitalized. About 11.1% of them died. Elderly men were slightly more likely to visit an ED, be hospitalized and die. About 7% elderly patients lived in small or non-metropolitan areas. They had similar unadjusted rates of ED visits, hospitalizations and deaths compared with those living in large or medium metropolitan areas. Those who were diagnosed before April 1 were more likely to visit an ED, be hospitalized, or die than those who were diagnosed after April 1, indicating more mildly symptomatic patients were detected in late period. The adjusted incidences of COVID-19 cases were presented in Table 2 , taking account of age and sex distributions in each county. Those living in small metropolitan areas had lower incidences than those living in large metropolitan areas. For example, for male patients aged 65-74 living in small metropolitan areas, the adjusted incidence was about half of that of large metropolitan areas (0.97 vs. 2.08 per 1000 persons; Rate Ratio (RR): 0.50; 95% confidence interval (95%CI): 0.29-0.85; p=0.01). Similar reduction was observed among elderly female patients (0.81 vs. 1.52 per 1000 persons; RR: 0.54; 95%CI: 0.33-0.86; p=0.01). In addition, elderly female patients aged 75 or above and living in non-metropolitan areas had much higher incidence than those living in large metropolitan areas (3.45 vs. 1.79 per 1,000 persons; RR: 1.97; 95%CI: 1.01-3.84; p=0.05). Table 3 presented predicted probabilities (rates) of ED visits, hospitalizations, and deaths for each age and sex group and by metropolitan status. There was no difference in rates of ED visits, hospitalizations and deaths between large and medium metropolitan areas. However, there were striking differences between males and females and between those living in small or non-metro All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. Furthermore, elderly male patients living in small metro also had higher hospitalization rates, leading to significant linear trends in hospitalization rates for both age groups among elderly male patients (p for trend = 0.01 for all males, and p=0.01 for age 65-74, and p=0.02 for age 75+). Thirdly, due to small number of deaths in each group, the probabilities of dying were similar across all sex and age groups, except for elderly male patients aged 65-74 and living in small metropolitan areas. They had a lower death rate than those living in large metropolitan areas. However, this significance might be incidental. In addition, after small metropolitan areas were combined with non-metro areas, the patterns were similar (appendix Table 1 ). Compared with those living in large or medium metropolitan areas, elderly male patients living in small or non-metro areas were more likely to visit an ED and be hospitalized, while elderly female patients living in these areas were less likely to visit an ED and be hospitalized. All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. . The adjusted hospitalization rates were mapped by counties (Figure 1 ). In addition to Miami-Dade County (right southeast corner), there were pockets of small metropolitan or non-metro counties that had hospitalization rates of over 60% (pink color), and many other small metropolitan counties that had 40-60% hospitalization rates (brown color). This was the first study that documented significant health disparities during the COVID-19 epidemic between large urban areas and small towns or rural areas. Elderly people living in small metropolitan areas had lower incidence of COVID-19 than those living in large or medium metropolitan areas. However, the patterns of ED visits and hospitalizations were opposite between elderly male patients and female patients living in small metropolitan areas. For example, among elderly patients living in small metropolitan areas, female patients had lower rates of ED visits and hospitalizations, while male patients had higher rates of ED visits and hospitalizations compared with those living in large metropolitan areas. There was a significant linear trend in hospitalization rates across county metropolitan status among elderly male patients. On the other hand, the death rates were similar across all regions. Our findings confirmed deficiency in providing health care to elderly people living outside of large or medium metropolitan areas (Casper et al., 2016; Singh et al., 2019) . Elderly people living in small towns or rural areas were known for lacking adequate health care (Odoi, Nagle, Roberson, & Kintziger, 2019) . In the time of emerging epidemic such as COVID-19, such problems may be aggravated when health care resources were under pressure of a run. However, the lower rates of ED visits and hospitalizations among elderly female patients, while higher rates of ED visits and hospitalizations among elderly male patients living in small metropolitan areas, required careful explanations. It was unclear whether this was due to differences in disease All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. . 1 0 severity, symptom tolerance, health care seeking behavior, or availability and accessibility to health care. Women were known to have lower tolerance of pain(Ruau, Liu, Clark, Angst, & Butte, 2012), but this might not be applicable to infectious diseases. In fact, that gender differences only occurred among those living in small metropolitan or non-metro areas suggested this might be more likely due to system level factors rather than individual behaviors. Furthermore, the reasons for a lower incidence of COVID-19 among small metropolitan areas were complicated. The dispersed residence in small towns and rural areas may deter the virus transmission, leading to a relatively lower incidence of COVID-19. On the other hand, there might not be enough detection kits available in these areas, resulting in an artificially lower incidence compared with large metropolitan areas. In addition, health care facilities in the US were mostly concentrated in large cities. Many small town or rural hospitals were not equipped to manage infectious patients. Patients with mild symptoms might be triaged to self-care at home, without being diagnosed and lab confirmed. For elderly people, this was not ideal, as the respiratory symptoms may exacerbate suddenly. Many of these severe cases were likely transferred to hospitals in larger cities, often enduring all kinds of troubles during the process. Although our knowledge of the 2019 SARS-CoV2 was growing rapidly, the treatment outcome was still unsatisfactory. Treating acute respiratory distress syndrome (ARDS) was still a major challenge, often leading to a mortality rate of 50%(C. Richardson et al., 2020) . Many elderly patients, especially those with underlying conditions such as cardiovascular diseases and diabetes, often had more severe diseases than those who were young and healthy. Therefore, a coordinated public health system, together with timely virus detection, case isolation, symptom monitoring and active contact tracing, were more important to curb the epidemic. All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. . https://doi.org/10. 1101 /2020 This study had some limitations. First of all, not all patient's information was publicly released due to privacy concerns. There were no explicit dates of symptom onset, clinic or ED visits, hospitalizations, and deaths in the data. There was no other information such as race and ethnicity, income and education level in the file, hindering our ability to full examine the roots of disparities (Hill et al., 2015) . However, this was not unique to Florida. Many other states released aggregated data only. To some extent, we had more than enough data that were useful to paint a broad picture, but no good data to help understand the drivers of epidemic process and examine health disparities behind the case counts. Secondly, although elderly patients might be more likely to have symptoms if infected by the virus, we would nevertheless miss many asymptomatic or mildly symptomatic patients who would not seek care or not be detected. The detection kits, especially at the early stage of epidemic and in small metropolitan and rural areas, were not readily available to health providers. We did not know whether the proportion of asymptomatic patients differed between large metropolitan areas and small metropolitan or nometro areas. Thus, we might underestimate the case incidences and overestimate the rates of ED visits and hospitalizations among those living in small metropolitan or non-metro areas. The true health disparities might be worse than our observed disparities. Thirdly, the COVID-19 epidemic was still evolving. Although our previous research indicated that the end of epidemic was near in the US (Yu, 2020), there would still be a lot of new cases to come every day, and the patterns of hospitalizations and deaths by different age, sex and regions would likely be more evident at the end of epidemic. Finally, given that much was still unknown regarding the treatments and consequences of COVID-19, elderly people might be impacted more profoundly by the epidemic and health disparities due to the diseases might be long lasting. All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020 . . https://doi.org/10.1101 /2020 There were some unique strengths in our study. To our knowledge, this was the first study using individual patient information to examine urban-rural disparities in the current COVID-19 epidemic. Health disparity issue was often neglected in the time of emerging epidemic, which was the reasons for urgent calls to tabulate cases and deaths by age, gender and ethnicities. Our research pointed to another dimension that should be incorporated in epidemic reports. Furthermore, unlike common descriptive reports that focused on numbers of cases, hospitalizations and deaths, we employed analytical methods to uncover hidden health disparities that were not evident in the aggregated tables. For example, comparing crude rates in Table 1 and adjusted rates in Table 2 and 3, only after careful adjustments did health disparities emerge. Therefore, our study called for more good data, more transparent reporting, and more appropriate analyses. In summary, although elderly people living in small metropolitan or non-metro areas had a lower incidence of COVID-19, elderly male patients living in these areas were more likely to have ED visits and hospitalizations, while elderly female patients living in these areas were less likely to have ED visits and hospitalizations compared with those living in large metropolitan areas. Profound health disparities exist in the time of emerging epidemic, and health care providers, especially those serving vulnerable population, should be vigilant about undiagnosed patients. All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. . This study was deemed exempt from ethical review, as it used publicly available data and no human subjects were directly involved in the study. No inform consent was necessary. The authors declared no conflict of interest with respect to the research, authorship, and/or publication of this article. The authors received no external financial support for the research, authorship, and/or publication of this article. All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. All rights reserved. No reuse allowed without permission. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint (which this version posted May 6, 2020. . How will country-based mitigation measures influence the course of the COVID-19 epidemic? Presumed Asymptomatic Carrier Transmission of COVID-19 Changes in the Geographic Patterns of Heart Disease Mortality in the United States Physical and Functional Limitations in US Older Cancer Survivors, J of Palliative Care & Medicine Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand Barriers to health care access among the elderly and who perceives them Clinical Characteristics of Coronavirus Disease 2019 in China The National Institute on Aging Health Disparities Research Framework Clinical features of patients infected with 2019 novel coronavirus in Wuhan Rapid asymptomatic transmission of COVID-19 during the incubation period demonstrating strong infectivity in a cluster of youngsters aged 16-23 years outside Wuhan and characteristics of young patients with COVID-19: a prospective contact-tracing study Asymptomatic and Presymptomatic Infectors: Hidden Sources of COVID-19 Disease Geographic disparities and temporal changes in risk of death from myocardial infarction in Florida Association of Public Health Interventions With the Epidemiology of the COVID-19 Outbreak in Wuhan The severe acute respiratory syndrome Presenting Characteristics, Comorbidities, and Outcomes Among Male