Secondary Logo

Journal Logo

Covid Pandemic

A survey on the frequency of COVID-19-like symptoms on students and staff of the University of Milan

Negri, Evaa; La Vecchia, Carlob

Author Information
European Journal of Cancer Prevention: May 2021 - Volume 30 - Issue 3 - p 282-284
doi: 10.1097/CEJ.0000000000000609
  • Free


Italy, and in particular Lombardy (10 million inhabitants), continue to be heavily affected by the coronavirus disease 2019 (COVID-19) pandemic. According to official data, on 11 May, there were over 219 000 cases and 30 739 deaths in Italy, and over 81 800 cases and 15 054 deaths in Lombardy (Protezione Civile, 2020). As noted, this may have substantial impact on cancer prevention, control and mortality (Mauri et al., 2020; Omarini et al., 2020; Kluge et al., 2020; Lyman and Kuderer, 2020; Negri et al., 2020; Wise, 2020).

The official figures are; however, substantially underestimated: the recorded cases, in fact, essentially include hospital admissions, plus a limited number of positive subjects from PCR swabs performed using a nonsystematic approach, and variable over time.


To estimate the number of possible contagions from COVID-19, a short survey was made available through a web link to all students and staff on the Milan University ‘la Statale’ website in the period 14–30 April 2020. It was anonymous at the source and included a COVID-19-related series of questions (fever, headache, cold, cough, anosmia, gastrointestinal complaints, and separately fever over 38.5°C) in the previous three weeks. It also included similar information on the cohabitants. The overall period covered by the investigation, therefore, runs from 24 March to 30 April. The survey was approved by the Milan University Ethical Committee on 2 April 2020.

The survey included 14 383 participants. Of these, 9 were excluded due to missing data, leaving a dataset of 14 374 subjects aged 18–77, median age 23.

The chi-square test was used to assess the statistical significance of differences in the frequency of symptoms across strata of selected covariates. Logistic regression models were used to estimate prevalence odds ratios (OR) and the corresponding 95% confidence intervals (CI) adjusted for the mutual confounding of various covariates.


Table 1 gives the prevalence of COVID-19-like symptoms. In the period under consideration, from March 24 to 30 April 3138 subjects (21.8%) reported COVID-19-like symptoms, and 219 (1.5%) fever above 38.5°C; 217 subjects performed at least one swab. Of these, 46 were positive (21.3% of those performed, 0.3% of the total of respondents).

Table 1 - Prevalence (%) of COVID-19-like symptoms in the last three weeks in a survey made available to students and personnel on the University of Milan’s website. Refers to the period 24 March–30 April 2020
No Yes % Yes (95% CI)
COVID-19-like symptoms 11 236 3138 21.8 (21.2–22.5)
Fever >38.5°C 14 155 219 1.5 (1.3–1.7)
Did you undergo a throat swab? 14 157 217 1.5 (1.3–1.7)
Did you have a positive swab?
 On the total sample 14 328 46 0.3 (0.2–0.4)
 On the total swabbed 171 46 21.2 (16.0–26.6)

The frequency of subjects reporting COVID-19-like symptoms was similar in women and men, slightly higher in the young (largely students), smokers and overweight subjects (Table 2).

Table 2 - Prevalence of COVID-19-like symptoms in the last three weeks in strata of selected covariates, corresponding odds ratio (OR) and 95% confidence intervals (CI) University of Milan, Italy, 24 March to 30 April 2020
Strata Prevalence of COVID-19-like symptoms
No Yes % Yes OR (95% CI)a
 Male 3717 1049 22.0 1b
 Female 7519 2089 21.7 1.00 (0.92–1.09)
P  = 0.72
Age (years)
 <30 8864 2576 22.5 1.27 (1.15–1.41)
 30+ 2372 562 19.2 1b
P  < 0.0001
 No 9253 2541 21.5 1b
 Yes 1983 597 23.1 1.08 (0.98–1.20)
P  = 0.08
BMI (kg/m2)c
 <25 (normol weight) 9480 2576 21.4 1b
 25+ (overweight) 1745 559 24.3 1.24 (1.12–1.39)
P = 0.002
aEstimates from a logistic regression model including all the variables in table.
bReference category.
cData were missing for 11 subjects.

Table 3 provides the multivariate ORs of reporting a fever over 38.5°C in relation to sex, age, smoking and BMI. The only significant result was the lower risk in women (OR = 0.65, 95% CI, 0.49–0.85) than in men.

Table 3 - Prevalence of fever >38.5°C in the last three weeks in strata of selected covariates, corresponding odds ratio (OR) and 95% confidence intervals (IC)
Strato Prevalence of fever >38.5°C
No Si % SI OR (95% CI)a
 Male 4672 94 2.0 1b
 Female 9483 125 1.3 0.65 (049–0.85)
P = 0.002
Age (years)
 <30 11 262 178 1.6 1.17 (0.82–1.66)
 30+ 2893 41 1.4 1b
P = 0.59
 No 11 611 183 1.6 1b
 Yes 2544 36 1.4 0.87 (0.61–1.25)
P = 0.56
BMI (kg/m2)c
 <25 (normal weight) 11877 182 1.5 1b
 25+ (overweight) 2267 37 1.6 1.01 (0.70–1.46)
P = 0.73
aEstimates from a logistic regression model including all the variables in table.
bReference category.
cData were missing for 11 subjects.

Table 4 reports the prevalence of COVID-19-like symptoms in cohabitants. There was a strong association between symptoms in the respondent and in cohabitants: 64% of subjects with symptoms reported at least one cohabitant with symptoms, compared to 14% of asymptomatic subjects. The corresponding OR was 11.4, with a 95% CI, 10.4–12.6.

Table 4 - Prevalence of COVID-19-like symptoms in cohabitants according to the presence of symptoms in the responder, corresponding odds ratio (OR) and 95% confidence intervals (CI)
COVID-19-like symptoms in at least one cohabitant
No Yes % Yes OR (95% CI)a
Symptoms in the responderb
 No 8809 1413 13.8 1c
 Yes 1023 1839 64.3 11.4 (10.4–12.6)
P < 0.0001
aEstimate from a logistic regression model including terms for responder’s age, sex, smoking, BMI and number of cohabitants.
bSubjects with none or more than 10 cohabitants were excluded, plus other 11 subjects for which the variable symptoms in cohabitants were missing.
cReference category.


The young age of the majority of the population – our students – suggests that a relevant proportion of these symptoms can reasonably be attributed to COVID-19, and not to chronic conditions that can be associated with these symptoms mainly in the elderly. In addition, both seasonal flu and most other viral and respiratory tract diseases are infrequent in late March–April.

Nevertheless, part of the reported symptoms may be related to other nonspecific (viral) conditions. It is possible, in any case, that a part of the symptoms – and the majority of fever episodes above 38.5°C – is due to COVID-19. Assuming that only half of the reported symptoms are attributable to COVID-19, our investigation confirms that around 10% of our academic community would have been affected by COVID-19 between late March and end April.

It is known that women have a lower frequency of moderate and severe COVID-19 since genes of ACE2 receptors are on the X chromosome (Ciaglia et al., 2020; Shang et al., 2020; Sharma et al., 2020), and women tend to have a more favourable lifestyle profile than men, and hence less frequent comorbidity. The lower risk of high fever in women, an indication of severity, indirectly confirms that a relevant proportion of reported symptoms are attributable to COVID-19.

The strong intra-nucleus of cohabitation contagiousness is an indication that the symptoms were caused by a new pathogen – most likely SARS-CoV-2 – against which there was no immunity in the population at the beginning of the epidemic.

Other individuals may have suffered from similar symptoms before March 24, although the February–early March data may be confused by the end of the seasonal flu epidemic. To these estimates, we should add the completely asymptomatic subjects infected with COVID-19, and those with symptoms so mild as not to be reported – even if in recent months all of us have been so impressed by the COVID-19 pandemic that most of us tended to pay attention to minor symptoms, too.

Although not representative of the general population, these data suggest that – even ignoring asymptomatic cases – COVID-19 affected a substantial proportion of our academic community in March and April 2020. This again suggests that the number of persons affected by COVID-19 was much greater in northern Italy than the number of recorded cases.


The key credit of this survey goes to all the students and staff of ‘La Statale’ who participated. The Teaching & Learning Innovation and Multimedia Technology Centre, University of Milan, in particular Dr. Patrizia Angelillo, and the Vice Rector for research strategies and policy, Prof. Maria Pia Abbracchio provided support for the survey.

The study was funded by Internal University Funds.

Conflicts of interest

There are no conflicts of interest.


Ciaglia E, Vecchione C, Puca AA (2020). COVID-19 infection and circulating ACE2 levels: protective role in women and children. Front Pediatr 8:206.
Kluge HHP, Wickramasinghe K, Rippin HL, Mendes R, Peters DH, Kontsevaya A, et al. (2020). Prevention and control of non-communicable diseases in the COVID-19 response. Lancet 395:10238.
Lyman GH, Kuderer NM (2020). Personalized cancer supportive care in COVID-19 era. Ann Oncol 31:835–837.
Mauri D, Kamposioras K, Tolia M, Alongi F, Tzachanis D, Petricevic B, et al. (2020). Summary of international recommendations in 23 languages for patients with cancer during the COVID-19 pandemic. Lancet Oncol 21:759–760.
Negri E, Scarpino V, La Vecchia C (2020). Prevalence of COVID-19-like symptoms in Italy and Lombardy, March-April 2020, and their implications on cancer prevention, diagnosis and management. Eur J Cancer Prev doi: 10.1097/CEJ.0000000000000604.
Omarini C, Maur M, Luppi G, Narni F, Luppi M, Dominici M, et al. (2020). Cancer treatment during COVID-19 pandemic: do not postpone, do it! Eur J Cancer 133:29–32.
Protezione Civile (2020). COVID-19 Italia - Monitoraggio situazione. Accessed 13 May 2020.
Shang J, Ye G, Shi K, Wan Y, Luo C, Aihara H, et al. (2020). Structural basis of receptor recognition by SARS-CoV-2. Nature 581:221–224.
Sharma G, Volgman AS, Michos ED (2020). Sex differences in mortality from COVID-19 pandemic: are men vulnerable and women protected? JACC Case Rep. doi: 10.1016/j.jaccas.2020.04.027.
Wise J (2020). Covid-19: Cancer mortality could rise at least 20% because of pandemic, study finds. BMJ 369:m1735.

COVID-19; neoplasms; prevalence; symptoms; survey

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.