COVID-19 in Italy: Considerations on official data

Perspective

COVID-19

in

Italy:

Considerations

on

of

ficial

data

Gino

Sartor

,

Marco

Del

Riccio

*,

Irene

Dal

Poz

,

Paolo

Bonanni

,

Guglielmo

Bonaccorsi

a_{DepartmentofHealthSciences,UniversityofFlorence,Florence,Italy} b_{InstituteofAdvancedStudies,UniversityofWarwick,Coventry,UnitedKingdom}

ARTICLE INFO

Articlehistory: Received16April2020

Receivedinrevisedform15June2020 Accepted17June2020 Keywords: COVID-19 SARS-CoV-2 Italy ABSTRACT

COVID-19representsamajorpublichealthissueinItaly;estimatingthesizeoftheoutbreakcoulddirect publichealthpoliciesandinformusoftheextentofthereorganizationneededinthehealthcaresystem, theefficacyofquarantinemeasures,andeventuallyontheachievementofherdimmunity.Tochartthe realextentofCOVID-19infectioninItalyofficialdataneedtobeinterpreted,consideringvariousaspects such as the "suspected-case" definition that changed during recentmonths, the management of asymptomatic and untested symptomatic cases, the system for reporting deaths, and short-term fluctuations.Alltheseaspectsshouldbeconsideredwhenreflectingonthemeaningoftheofficial COVID-19figuresinItaly.Regionalizationofthehealthcaresystemandfragmentationofdatarepresentreal challengesinthemanagementoftheCOVID-19outbreakinItaly.Theauthors’opinionisthattransparent andaccuratereportingcouldguidepolicy-makingandhelpreorganizehealthservices.

©2020TheAuthor(s).PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases. ThisisanopenaccessarticleundertheCCBY-NC-NDlicense( http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction

COVID-19representsasignificantpublichealthissue.Mar11, 2020,theWorldHealthOrganization(WHO)officiallydeclaredthe COVID-19 outbreak a pandemic (World Health Organization (WHO),2020).Thisviruscaneasilyspread(Wangetal.,2020;Li etal.,2020)andcanleadtoasymptomaticcases(Tianetal.,2020; Mizumotoetal.,2020),mildsyndromes(Guanetal.,2020;Wuand McGoogan, 2020) as well as severe manifestations, requiring hospitalizationandIntensiveCareUnits(ICU)(WuandMcGoogan, 2020).Italyisamongtheworld'sworst-hitcountries(Ministryof Health,Italy,2020)andthefirstamongWesterncountriestoorder anationallockdown.

Accurateestimatesofthesizeoftheoutbreakcouldinfluence publichealthpolicies.TheycouldinformCovid-19researchonthe necessary reorganization of a healthcare system and on the efficacyofquarantinemeasures.Finally,accurateestimatescould improveourunderstandingofhowclosewearetotheacquisition ofherdimmunity.

National bulletins provide us withofficial data onanalyzed swabs,confirmedcumulativecases,homeisolationcases, hospi-talized cases, ICU cases, and deaths (Ministry of Health, Italy, 2020).However, Russel et al. estimate that Italy is one of the countries with the highest percentage of unreported cases of

COVID-19(Russelet al.,2020).Theregionalization ofdataflow representsanotherchallenge.Dataiscommunicateddaily–inan aggregatefashion- byeveryregiontothenational authorities; however,thereisnostandardizedcollectionofdata.Inparallel, public health services upload individual data to a surveillance system’s online platform. While the first data flow is region-dependent and has the advantage of being updated daily, the second is more standardized but notifications refer to cases diagnosedonpreviousdays.Inthispaper,wewillrefertothefirst dataflow,whichhasextensivelybeenusedbyscientistsandmedia toanalyzethepandemic'sevolution.Regardingthisdataflow,the lack of uniformity among Italian regional data raised the unconvincinghypothesisofdifferentlyvirulentSARS-CoV-2types. Thus,totracktherealextentofCOVID-19infectioninItaly,itneeds tobeinterpretedaccordingtothefollowingpoints.

Suspectedcasedefinition

Atthebeginningoftheepidemic,thedefinitionof“suspected case”tobetestedforSARS-CoV-2wasbasedonsymptoms(fever, cough, and flu-like symptoms) and a history of traveling or residence in China or Italian “red areas." This testing policy explains why the regions of Lombardy and Veneto initially collected more samples. The definition was later revised as a personwithacuterespiratory infectionassociated witheithera historyoftravelorresidenceinanareareportinglocal transmis-sion of COVID-19, close contact witha confirmed or probable COVID-19 case, or requiring hospitalization (Italian Ministryof *Correspondingauthor.

E-mailaddress:marco.delriccio@unifi.it(M.DelRiccio). https://doi.org/10.1016/j.ijid.2020.06.060

1201-9712/©2020TheAuthor(s).PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.ThisisanopenaccessarticleundertheCCBY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

InternationalJournalofInfectiousDiseases98(2020)188–190

ContentslistsavailableatScienceDirect

International

Journal

of

Infectious

Diseases

Health,2020).Nonetheless,Veneto decidedtoextendtestingto asymptomaticpeoplewithclosecontactwithconfirmedcasesor asymptomaticpeopleinredareas.

Theresultisthattheproportionofthepopulationtestedvaries among regions. The percentage of samples over the total populationranges from0.92% in Campania to 5.46% in Veneto (dataupdatedApr21)(MinistryofHealth,Italy,2020).

Takingtheproportionofthetestedpopulationinto consider-ation helps usinterpret official data.For example, Veneto (the regionthattestedthemost)andPiedmonthaverecordedsimilar numbersofconfirmedcasesuntilmid-April.However,thenumber ofdeathsand thenumber ofhospitalizedandICUpatientswas much higher inPiedmont. This suggeststhat a vastnumber of infections went undetected in Piedmont. Eventhough most of these undetected cases probably presented mild or moderate symptoms, some critical patients might have died without a confirmeddiagnosis.

Short-termfluctuations

Short-termfluctuationsmightcompromisethereliabilityofdaily data.Thesefluctuationsareduetolaboratorydelays–e.g.,results fromdelaysduetolaboratorysaturation–orcalendareffects.The numberofswabstendstodroponweekends,andthuswenoticea weeklypattern inthetimeseriesofthenumber ofswabs andpositive cases(Figure1).Techniques,suchasmovingaverages,shouldbe consideredinordertoaccountforthisvariability.

Asymptomaticanduntestedsymptomaticcases

The number of asymptomatic and untested symptomatic subjects is essential to track the size of the outbreak. In the literature, there remains a high level of uncertainty in the

estimatesofthepercentageofasymptomaticcarriers(Centrefor Evidence-Based Medicine (CEBM), 2020). We hypothesize that untestedsymptomaticsubjectsdevelopmostlymildormoderate syndromes.Tabata etal.,forexample,foundthat43 outof104 patients were mildly symptomatic on the “Diamond Princess” (Tabataetal.,2020).

Thereasonsforunderdiagnosisofasymptomaticand mildly/ moderatelysymptomaticpatientsmaybe:

a)politicalandhealthcarechoices;

b)geographicproximityto“redareas”(Giordanoetal.,2020); c)strengthandflexibilityofpublichealthservicesandprimary careservices;

d)availabilityofswabsand/orlaboratoriestoanalyzesamples (Rubinoetal.,2020).

SaturationofICU

ItisworthremarkingthatthenumberofICUpatientsisgivenas dailyprevalencedata,whilethetotalnumberofpositivecasesand deathsiscumulative.Weshouldconsiderthatthetotalnumberof ICUcasesmayreflectthesaturationofICUcapacityandnotthe totalneedforICUbeds(Grassellietal.,2020).Moreover,someof thedeadandhealedsubjectsmusthavepreviouslybeentakencare of inICU. Zhouetal. foundthatout of54 deathsin a hospital setting,42hadbeentreatedinICU,i.e.,77.7%)while12werenever admittedtoICU-22.3%,probablybecauseofoldage,comorbidities orlackof ICUbeds(Zhouet al.,2020).However,deathsdo not alwaysoccurinhospitalwardsorICUs.Particularlyinthoseareas where deaths and recoveries now significantly outnumber prevalentICUcases,thepercentageofsubjectswhowouldneed intensive treatment over the total of infected cases is under-estimated.

Deathsreporting

TheItalianNationalInstituteofStatistics(ISTAT)adoptedthe WHOICD-10codes‘U07.1COVID-19,virusidentified’and‘U07.2 COVID-19,virusnotidentified’tocodedeathsfromCOVID-19as confirmed ornot bylaboratory testing.Nonetheless, weexpect underreportingandmisclassification.

In Nembro (province of Bergamo), one of the mostaffected municipalities,thenumberof expecteddeathsinthefirstthree monthsof2020undernormalconditionswas35.158deathswere instead registered by municipal offices. The number of deaths officiallyattributedtoCOVID-19is31.Itisreasonabletoexpectthat thedeathsinexcessoccurredathomeorinnursingcarehomes (CorrieredellaSera(CdS),2020).

However,we cannotexclude somecasesof misclassification due to ascertainment bias, i.e., coding deaths from severe respiratorysyndromewith‘U07.2COVID-19,virusnotidentified’.

Figure1.NumberofdailyswabsanddailynewcasesinPiedmontandVeneto.

Table1

Regionalpopulation,numberofsamples,percentageofsamplesoverthepopulation,thecumulativenumberofreportedcases,percentageofcasesoverpopulation, percentageofcollectedsamplesovercases,hospitalizationratio,andhospitalvs.homecareratio.Hospitalizationratioisthepercentageofhospitalizedcasesovertotalcases (thisvalueisthemeanoftheratioscalculateddailysincethebeginningofthepandemic).Hospitalvs.HomeCareratioisthepercentageofhospitalizedcasesovercases managedathome(thisvaluerepresentsthemeanoftheratioscalculateddailysincethebeginningofthepandemic)(Anon,2020).

OFFICIALDATA Apr21

Population Samples Samples*100 /population

Cumulative Cases

Cases*100 /population

Samples/Cases Hospitalization ratio(mean)

HospitalvsHome Careratio(mean) Emilia-Romagna 4,459,477 134,878 3.02 13,244 0.30 10.18 0.34 0.48 Lazio 5,879,082 100,031 1.07 4,402 0.07 22.72 0.45 0.76 Lombardy 10,060,574 277,197 2.76 33,978 0.34 8.16 0.46 0.80 Marche 1,525,271 44,332 2.91 3,218 0.21 13.78 0.34 0.46 Piedmont 4,356,406 105,434 2.42 14,811 0.34 7.12 0.39 0.63 Veneto 4,905,854 268,069 5.46 10,077 0.21 26.60 0.21 0.23 Italy 60,359,546 1,450,150 2.40 78,671 0.18 18.43 0.35 0.46

Casefatalityratio(CFR)

Accordingtoofficialstatistics, 13.40%ofconfirmedcasesdiedby Apr21,varyingbetween4.43%inUmbriaand18.52%inLombardy. Mizumotoetal.suggestthatelevateddeathriskestimatescanbe the consequence of a breakdown of the healthcare system (MizumotoandChowell,2020).ThisisconsistentwiththeItalian figuresindicatingthatNorthernregionsofLombardyandEmilia Romagnahaveboththehighestcumulativeincidencesandhighest CFRs.However,atimelagoccursbetweeninfectionsanddeaths.If thecurveofdeathsfollowsthecurveofcases,thedeathcurvein first-hitregions(LombardyandEmiliaRomagna)mightprecede thoseofotherregions.

TheItalianofficialCFRishigherthanthoseofSouthKoreaand Germany.BothSouthKoreaandGermanyanalyzedalotofsamples and,asaresult,haveawidercoverageofdiagnosis,increasingthe denominatoroftheCFR (Rodriguez-Moralesetal.,2020).Verity etal.foundCFRvalues rangingfrom2.7% to3.6%(Verityetal., 2020).AsnotedbyOnderetal.,theproportionofpositivecases overtotalsamplespositivelycorrelateswiththeCFR(Onderetal., 2020).

Conclusions

Regionalizationofthehealthcaresystemandfragmentationof datarepresent challenges in themanagement of the COVID-19 outbreakinItaly.Thelackofastrongcentralizedresponsetothe emergency resulted in different regional policies, especially in termsoftestingstrategies.

Table 1 shows regional differences in surveillance and containment strategies and measures. In particular, it emerges howsomeregionstestedmorethanothers(e.g.,Venetocollected 26.60samplesforeach case while Piedmontonly 7.12for each case)andhowdifferentwasthemanagementofthediseaseamong differentareas(e.g.,Venetohospitalized,onaverage,21%of COVID-19 patients, while Lombardy hospitalized, on average, 46% of COVID-19patients).Itisofutterimportancetodevelopasystem forinforminginfectiousdiseasesbasedonaconstantlyupdated anduniqueplatformatanationallevel,possiblyfedbylaboratory data and linked to clinical records and other administrative databases. A transparent and accurate reporting could guide policy-makingand help reorganize health services ( Rodriguez-Moralesetal.,2020).

Funding

Theauthor(s)receivednospecificfundingforthiswork. EthicalApproval

Notapplicable. Conflictofinterests

The authors declare that they have no known competing financial interests or personal relationships that could have appearedtoinfluencetheworkreportedinthispaper.

References

Anon.AnalisideimodelliorganizzatividirispostaalCOVID-19:FocussuLombardia, Veneto,EmiliaRomagna,Piemonte,LazioeMarche.[Lastaccessed:May2020].

Available from:. 2020. https://altems.unicatt.it/altems-ALTEMS-COVID19_Is-tantReport4-report.pdf.

Centre for Evidence-BasedMedicine (CEBM). COVID-19:What Proportion are Asymptomatic?. [AccessedApr,2020].Availablefrom:.Oxford:CEBM;2020. https://www.cebm.net/covid-19/covid-19-what-proportion-are-asymptomat-ic/.

CorrieredellaSera(CdS).TheRealDeathTollforCovid-19isatLeast4Timesthe OfficialNumbers.Rome:CdS.[Lastaccessed:Apr2020].Availablefrom:.2020. https://www.corriere.it/politica/20_marzo_26/the-real-death-toll-for-covid- 19-is-at-least-4-times-the-official-numbers-b5af0edc-6eeb-11ea-925b-a0c3cdbe1130.shtml.

GiordanoG,BlanchiniF,BrunoR,etal.ModellingtheCOVID-19epidemicand implementationofpopulation-wideinterventionsinItaly[publishedonline aheadofprint,2020Apr22].NatMed2020;1–6,doi:http://dx.doi.org/10.1038/ s41591-020-0883-7.

GrasselliG,PesentiA,CecconiM.CriticalcareutilizationfortheCOVID-19outbreak in Lombardy, Italy: early experience and forecast during an emergency response. JAMA [Internet] 2020;(March 13), doi:http://dx.doi.org/10.1001/ jama.2020.4031Availablefrom:.

GuanW-J,NiZ-Y,HuY,LiangW-H,OuC-Q,HeJ-X,etal.Clinicalcharacteristicsof coronavirusdisease2019inChina.NEnglJMed[Internet].2020;(Feb28) 10.1056/NEJMoa2002032.Availablefrom:https://pubmed.ncbi.nlm.niZ.-Y. ItalianMinistryofHealth.CircularLettern. 7922,Mar9Availablefrom:.2020.

https://www.fnopi.it/wp-content/uploads/2020/03/Circolare_9_marzo_2020. pdf.

LiQ,GuanX,WuP,WangX,ZhouL,TongY,etal.Earlytransmissiondynamics inWuhan,China,ofnovelcoronavirus–infectedpneumonia.NEnglJMed 2020;1–9.

MinistryofHealth,Italy.CoronavirusDisease(COVID-19)SituationReportinItaly. [AccessedMar,2020].Availablefrom:.Italy:Rome:MinistryofHealth;2020. http://www.salute.gov.it/portale/nuovocoronavirus/dettaglioContenutiNuovo- Coronavirus.jsp?lingua=italiano&id=5351&area=nuovoCoronavirus&menu=-vuoto.

MizumotoK,ChowellG.Estimatingriskfordeathfrom2019novelcoronavirus disease,China,january-february2020.EmergInfectDis.2020;26(March(6)). Mizumoto K,Kagaya K,ZarebskiA, ChowellG. Estimating theAsymptomatic

Proportion of Coronavirus Disease 2019 (COVID-19) Cases on Board the DiamondPrincessCruiseShip,Yokohama,Japan..p.1–52020;2019(February). OnderG,RezzaG,BrusaferroS.Case-fatalityrateandcharacteristicsofpatients

dyinginrelationtoCOVID-19inItaly.JAMA2020;.

Rodriguez-MoralesAJ,Cardona-OspinaJA,Gutiérrez-OcampoE,Villamizar-PeñaR, Holguin-RiveraY,Escalera-AntezanaJP,etal.Clinical,laboratoryandimaging featuresofCOVID-19:asystematicreviewandmeta-analysis.TravelMedInfect Dis2020;.

RubinoS,KelvinN,Bermejo-MartinJF,KelvinD.AsCOVID-19cases,deathsand fatalityratessurgeinItaly,underlyingcausesrequireinvestigation.JInfectDev Ctries 2020;14(3)265–7, doi:http://dx.doi.org/10.3855/jidc.12734 Published 2020Mar31.

RusselT,HellewellJ,AbbottS,JarvisCI,vanZandvoortK,FlascheS,etal.Usinga Delay-Adjusted Case FatalityRatio to Estimate Under-Reporting.C Repos. Availablefrom:. 2020. https://cmmid.github.io/topics/covid19/severity/glob- al_cfr_estimates.html?fbclid=IwAR0YKO4lcEUvFyi4xzUlbrukES-_Poo-M39iTLgjx4JsGDuyTFPOoRnAS0ZU.

TabataS,ImaiK,KawanoS,IkedaM,KodamaT,MiyoshiK,etal.Theclinical characteristicsofCOVID-19:aretrospectiveanalysisof104patientsfromthe outbreakonboardtheDiamondPrincesscruiseshipinJapan.medRxiv2020;.. Available from: http://medrxiv.org/content/early/2020/04/07/ 2020.03.18.20038125.abstract.

TianS,HuN,LouJ,ChenK,KangX,XiangZ,etal.CharacteristicsofCOVID-19 infectioninBeijing.JInfect[Internet]2020;..Feb 27;S0163-4453(20)30101-30108.Availablefrom:https://pubmed.ncbi.nlm.nih.gov/32112886.

VerityR,OkellLC,DorigattiI,WinskillP,WhittakerC,ImaiN,etal.Estimatesofthe severityofcoronavirusdisease2019:amodel-basedanalysis.LancetInfectDis 2020;.

WangD,HuB,HuC,ZhuF,LiuX,ZhangJ,etal.ClinicalCharacteristicsof138 Hospitalized Patientswith 2019Novel Coronavirus-InfectedPneumonia in Wuhan,China.JAMA2020;1–9.

WorldHealthOrganization(WHO).CoronavirusDisease(COVID-2019)Situation Reports. [AccessedMar,2020].Availablefrom:.Geneva:WHO;2020.https:// www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports.

WuZ,McGooganJM.Characteristicsofandimportantlessonsfromthecoronavirus disease2019(COVID-19)outbreakinChina:summaryofareportof72314cases fromtheChineseCenterforDiseaseControlandPrevention.JAMA[Internet] 2020;(February 24), doi:http://dx.doi.org/10.1001/jama.2020.2648 Available from:.

ZhouF,YuT,DuR,FanG,LiuY,LiuZ,etal.Clinicalcourseandriskfactorsfor mortalityofadultinpatientswithCOVID-19inWuhan,China:aretrospective cohortstudy.Lancet2020;.