Economía y salud
BOLETÍN INFORMATIVO - Año 2020. Septiembre nº 95
www.aes.es
En este número Opinión
Portada

Opinión
Temas
Tesis
Noticias
Buzón de sugerencias

Brothers in arms



Pedro Pita Barros
BPI | Fundação “la Caixa” Professor of Health Economics
Nova School of Business and Economics
Email: ppbarros@novasbe.pt
Twitter: @pitabarros

 

Portugal and Spain, neighbours in the Iberian Peninsula, are, so far, different in how they were affected by COVID-19. At least, they have been perceived and portrayed that way in the public eye and media.

Rigorous cross-country comparisons are less obvious than usually conceded, as we have a mix of elements to account for: the dynamics of the pandemic, the policy responses by the public authorities, the behavioural changes by the population and how well the health system reacts to the new demands, determined by the interactions of disease, policies and behaviours.

It is relevant to be clear about the objectives pursued by governments’ actions, as performance should be measured against them. The main objective of many public authorities in reaction to the COVID-19 can be summarised by the famous expression “to flatten the curve”. The congestion of health systems leads to failure in treatments and to extra deaths that can, and should, be avoided.

With this objective, the “right” metric for comparison is additional deaths created by congestion of the health system. The lower this value, the better is the performance. An indirect indicator, with noise but easier to assess in real time, is how overwhelmed, how congested is the health system at any point in time.

Comparing Portugal and Spain at this lens, Portugal´s National Health Service (NHS) did not experience extreme congestion due to COVID-19 patients. In Spain, the initial stages of COVID-19 led to dramatic situations in some areas (Madrid stands out). So, if we accept the difference in congestion levels, a natural question is what motivated it.

The two major factors, in my view, are luck and quick decision-making.

The luck element for Portugal was that Italy and Spain were hit first, and that this was widely reported in the news. Portuguese authorities and citizens were able to learn quickly from this elapsed time. With the TV news reporting “horror” stories from Italian and Spanish hospitals, public opinion pressure in Portugal mounted and led to early decisions on lockdown and movement restrictions by authorities, and early compliance with containment measures by citizens. Indeed, they anticipated by several days the official start of lockdown measures, with some activities closing and grocery shopping taking place prior to the state of emergency declaration.

Portugal’s advantage was early adoption of measures, relative to the start of COVID-19 cases, not just on how well they were received by the population. Publicly available data from Google Community Mobility Reports (Figure 1) show similar drops in magnitude in the relative mobility of people in Portugal and Spain (albeit slightly deeper in the latter), but it occurred several days earlier in Portugal relative to the start of the lockdown. Thus, I advance the conjecture that early decisions by public authorities and by the population were a demand-side effect that favoured Portugal.

 

Figure 1 Changes in mobility in Portugal and Spain by type of activity/location

Source: author’s calculations from Google Community Mobility Reports.

 

I turn now to supply-side effects, where luck in being a laggard and quick learning provided for a better situation in Portugal. The Portuguese Medical Association created a group, in February, to follow the international evolution of COVID-19. Some members later provided advice to the Portuguese Government. This resulted in the public announcement, during the early days of the pandemic, of a strategy represented by 80-15-5: 80% of COVID-19 patients would have mild enough symptoms to be treated at home, 15% would need admission to hospital care, and 5% would require use of intensive care.

This approach supported a general acceptance that admission to the hospital was not essential for many patients, reducing considerably the workload, and avoiding congested situations in NHS hospitals. The quest for protective equipment to be used by health professionals was also pursued early on.

Following COVID-19 patients at home relied on a strong and widely disseminated network of NHS’ primary care units. This network provided, outside the media spotlights, the continuous work to avoid unnecessary hospitalisations. Although there is absence of regular information on the results of this strategy, government’s documents published when the measures of further easing the lockdown and downgrading the state of emergency were announced (May 15) reported that 97.2% of confirmed COVID-19 patients were being treated at home, 2.3% were treated at the hospital and 0.5% at ICU (Portugal went into state of emergency on March 18, ending it on May 2 (Saturday), and started first and second phases of de-confinement on May 4 and 18, respectively).

Over time, Spanish hospitals adjusted to the demands of COVID-19. Current operation is already far from the crisis reports of the initial days.

The difference of Portugal combined luck and early decisions.

Up to this point, my performance metric was how the health system, and in particular hospitals, coped with the COVID-19 pressure. However, the majority of international comparisons use either total number of cases, or number of new daily cases, or deaths, or daily deaths, either in absolute terms or normalised by population size. Several websites provide illustrative graphics of countries’ positions in these metrics. Although tempting to use these data, there is more to it than just compiling and reporting the numbers.

Take new daily confirmed cases of COVID-19. To compare this number across countries, we need some sort of benchmark or conceptual framework to interpret it. We can start with the simplest conceptual background for an epidemic, the basic Susceptible-Infectious-Recovered (SIR) epidemiological model, which became very popular in the past three months. A basic characteristic of it is that dynamics of contagion is based on infected (and contagious) persons meeting susceptible (to infection) people. At the early stages of the pandemic, the overall size of the population is basically irrelevant to the growth of the infected group. Thus, looking at the absolute numbers of each country at early stages will reveal if a country is doing better than another one, or not, in terms of progression of the disease.

Normalising by total population size, at such early stages, for the same dynamic process, will make smaller countries look worse than it is the case. Of course, after some time, the population still free of infection becomes smaller more rapidly in the smaller country. So, after some point, normalising by the total population will become necessary if we want to compare how, in the end, each country was affected. The lower proportion of a population infected, the better. The problem for the analyst is to decide when to move from the absolute number to normalised values. The same argument can, obviously, be applied to comparisons based on deaths. The use of deaths as an indicator is often pointed out as having the advantage of less noise in measurement than confirmed cases (a variable that may also depend on the testing capacity and the testing policy of each country). It is also a measure that performs better if the health system is better at treating people. Several sources publish variants of measures comparing across countries. I use here the free Financial Times graphics (at 20 May 2020), with normalisation for the time point of the pandemic (see Figure 2).

 

Figure 2 Number of deaths and cases in Portugal and Spain

 

From these pictures, the situation in the two countries is not that different at start, and in per capita terms, the number of cases even looked more problematic in Portugal, despite its NHS hospitals not facing major congestion levels, unlike what was reported in some Spanish hospitals (where the problems caused by COVID-19 patients were further compounded by infection of health care professionals).

However, looking at deaths, we have a quite distinct evolution in the two countries. Naturally, as the pandemic surge becomes more under control, the same absolute reduction in the total number of new cases will look smaller in larger countries, and direct inspection of figures becomes more difficult to inform our analysis. The natural log scale hides from human eye perception the large difference that may exist between the two countries, though the more important aspect is that Portugal is not an outlier in the lower end of the spectrum, while Spain is on the high side of the distribution.

Comparisons based on deaths are not problem free. They are affected by the way each country computes the death toll (patients dying with COVID-19 versus patients dying from COVID-19). And if our interest is in lost lives, then we should probably also compute the deaths occurring due to people not going to health facilities given their fear of contagion. Currently, there is no direct account for this that is widely available and comparing countries.

Overall, comparing countries in their response and how badly they were hit by COVID-19 requires a more detailed and complex background framework than just piling up numbers.

My main view, at this stage and with the available information, is that Portugal and Spain started very differently the early stages of the pandemic. A mix of luck (of being hit first) and of quick decisions by authorities and population saved Portugal from severe congestion at hospitals. Spain had a delayed response, forcing a longer and deeper lockdown, and had a higher cost in deaths (the economic cost is too early to assess).


Deja tu opinión:

Nombre:
E-mail: (No visible)
Titular:
Opinión:
¿Cuál es el número de en medio?
3 7 (9) 1 6


AES-Secretaría
TecnoCampus Mataró-Maresme
Edifici TCM2 | P2. O3.

Av. Ernest Lluch, 32
08302 Mataró (Barcelona)

Tel. 93 755 23 82
secretaria@aes.es
 

Editores del boletín: Cristina Hernández-Quevedo (c.hernandez-quevedo@lse.ac.uk) y Jorge Mestre-Ferrándiz (jormesfer13@gmail.com).

Editora de redacción: Cristina Hernández-Quevedo (c.hernandez-quevedo@lse.ac.uk).


El Comité de Redacción del Boletín / Blog de Economía y Salud está compuesto por: Ruth Puig Peiró (en representación de la Junta Directiva de AES), Helena Hernández y Myriam Soto Ruiz de Gordoa (en representación de AESEC), Javier Mar y Miguel Ángel Negrín (en representación de EEconAES), Patricia Cubí-Mollá y Borja García-Lorenzo (en representación de EvaluAES), Carmen Pérez Romero y Elisa Gómez Inhiesto (en representación de GestionAES), Ariadna García Prado (Universidad de Pamplona) y Luz María Peña Longobardo (Universidad Castilla-La Mancha).


Han colaborado en este número: Enrique Castellón, Laura Coll-Planas, Francisco Escribano Sotos, Beatriz González López-Valcárcel, Ildefonso Hernández, Álvaro Hidalgo, Pere Ibern, Guillem Lopez i Casasnovas, Roberto Martínez Lacoba, Ricard Meneu, Patricia Moreno Mencía, Marta Ortega-Ortega, Vicente Ortún, Isabel Pardo García, Salvador Peiró, Sara Pinillos Franco, Pedro Pita Barros, María Jesús Pueyo Sánchez, Pere A. Taberner, Néboa Zozaya.