The French Court of Auditors is not a scientific body and publishes no articles in economics journals. Too bad. Their statement about a positive correlation, in France, between physician density and healthcare expenses, is not evidence-based. It is also an example of an organization that goes beyond its only accounting skills. Its tweet is distressing from the point of view of scientific credibility and I will demonstrate that.
1 / The statement concerning correlation is scientifically nonsense. Let us look at the kind of proof! It is only maps...
ANY scientist knows that correlations are not evidence of causality. Rereading or reading this excellent synthesis will go down some maybe even at the CC! https://fr.wikipedia.org/wiki/Cum_hoc_ergo_propter_hoc
2 / In this case, an inverse causality can also explain the correlation. In a population with higher incomes the density of doctors increases. In a more solvent population (it can be by free treatment-CMU -Complémentaire -etc) the density increases. It is amazing that it is not even mentioned.
3 / This assertion is "limited" to "liberal" physicians. We know that there is no real private sector in France. The liberal physicians are paid civil servants with tariffs and the public physicians are salaried civil servants. If one wants to rely on serious arguments of the literature and not on a fantasized vision of the world of health (this vision which consists in saying that there would be a sector of "liberal" health professionals in France) one must compare with other countries. But all studies in the EU are comparisons between a real private sector (what is the sector of practitioners out of convention id est less than 1000 participants in France all in outpatient medicine) and the contracted sector id est where the payer is the state. And these comparisons are not in favour of correlations as evoked by the CC. For example this work (https://www.uni-salzburg.at/fileadmin/oracle_file_imports/1461242.PDF).
http://www.oecd.org/els/health-systems/35987490.pdf
On this table we see that many counter-examples undermine the correlation as causal evidence: Greece, the US, are at the antipodes of the expense and density ...
On this correlation of several countries, the hypothesis of a correlation and therefore of causality is unlikely given the low coefficient of correlation. Moreover, it is clear that if there is a correlation, it is for the poorly developed countries that is to say either with a low medical density or with low expenses (Mexico).
http://www.oecd.org/els/health-systems/35987490.pdf
4 / The CC ignores what highly specialized agencies have studied in detail about the care and non-health needs of populations. The OECD is extremely precise and factual in its work, which is very regular in the area of health care expenditure and population parameters. For example, and this has certainly not escaped the members of the CC, health expenditures in developed countries are not correlated to the densities of doctors ...
6 / Can we expect a decline in medical density to reduce healthcare costs? Probably not or a very weak one, this review of the literature demonstrates this.
"In the absence of more accurate information, it should be considered cautiously."
https://www.sciencedirect.com/science/article/pii/S0168851008002789
1 / The statement concerning correlation is scientifically nonsense. Let us look at the kind of proof! It is only maps...
 |
| https://www.ccomptes.fr/fr/documents/40737 |
ANY scientist knows that correlations are not evidence of causality. Rereading or reading this excellent synthesis will go down some maybe even at the CC! https://fr.wikipedia.org/wiki/Cum_hoc_ergo_propter_hoc
2 / In this case, an inverse causality can also explain the correlation. In a population with higher incomes the density of doctors increases. In a more solvent population (it can be by free treatment-CMU -Complémentaire -etc) the density increases. It is amazing that it is not even mentioned.
3 / This assertion is "limited" to "liberal" physicians. We know that there is no real private sector in France. The liberal physicians are paid civil servants with tariffs and the public physicians are salaried civil servants. If one wants to rely on serious arguments of the literature and not on a fantasized vision of the world of health (this vision which consists in saying that there would be a sector of "liberal" health professionals in France) one must compare with other countries. But all studies in the EU are comparisons between a real private sector (what is the sector of practitioners out of convention id est less than 1000 participants in France all in outpatient medicine) and the contracted sector id est where the payer is the state. And these comparisons are not in favour of correlations as evoked by the CC. For example this work (https://www.uni-salzburg.at/fileadmin/oracle_file_imports/1461242.PDF).
http://www.oecd.org/els/health-systems/35987490.pdf
On this table we see that many counter-examples undermine the correlation as causal evidence: Greece, the US, are at the antipodes of the expense and density ...
On this correlation of several countries, the hypothesis of a correlation and therefore of causality is unlikely given the low coefficient of correlation. Moreover, it is clear that if there is a correlation, it is for the poorly developed countries that is to say either with a low medical density or with low expenses (Mexico).
http://www.oecd.org/els/health-systems/35987490.pdf
4 / The CC ignores what highly specialized agencies have studied in detail about the care and non-health needs of populations. The OECD is extremely precise and factual in its work, which is very regular in the area of health care expenditure and population parameters. For example, and this has certainly not escaped the members of the CC, health expenditures in developed countries are not correlated to the densities of doctors ...
6 / Can we expect a decline in medical density to reduce healthcare costs? Probably not or a very weak one, this review of the literature demonstrates this.
"In the absence of more accurate information, it should be considered cautiously."
https://www.sciencedirect.com/science/article/pii/S0168851008002789
La Cour des Comptes n'est pas un organisme scientifique et ne publie aucun article dans des revues d'économie. C'est dommage. Les affirmations concernant une corrélation entre densité de médecins libéraux et dépenses de soin sont affligeantes du point de vue de la crédibilité scientifique et je vais le démontrer. Mais ces affirmations sont aussi un exemple d'un organisme qui dépasse ses compétences qui ne sont que comptables.
1/ L'assertion concernant la corrélation est scientifiquement un non sens.
TOUT scientifique sait que les corrélations ne sont pas des preuves de causalité. Relire ou lire cet excellente synthèse descillera certains peut être même à la CC! https://fr.wikipedia.org/wiki/Cum_hoc_ergo_propter_hoc
 |
| https://www.ccomptes.fr/fr/documents/40737 |
TOUT scientifique sait que les corrélations ne sont pas des preuves de causalité. Relire ou lire cet excellente synthèse descillera certains peut être même à la CC! https://fr.wikipedia.org/wiki/Cum_hoc_ergo_propter_hoc
2/ Dans ce cas une causalité inverse peut aussi expliquer la corrélation. Dans une population à plus fort revenus la densité de médecins augmente. Dans une population plus solvable (cela peut être par la gratuité des soins-CMU -Complémentaire -etc) la densité augmente. Il est stupéfiant que cela ne soit même pas évoqué.
3/ Cette assertion est "limitée" aux PS libéraux. Or nous savons qu'il n'y a pas de reel secteur privé en France. Les PS libéraux sont des PS fonctionnaires payés à l'acte et les PS publics sont des fonctionnaires salariés. Si l'on veut s'appuyer sur des arguments sérieux de la littérature et non sur une vision fantasmée du monde de la santé (cette vison qui consiste à dire qu'il y aurait un secteur "libéral" ) il faut comparer avec d'autres pays. Mais toutes les études en UE sont des comparaisons entre un vrai secteur privé (ce qu'est le secteur de praticiens hors convention c'est à dire moins de 1000 particiens en France tous en médecine ambulatoire ) et le secteur conventionné c'est à dire où le payeur est l'état. Et ces comparaisons ne sont pas en faveur de corrélations telles qu'évoquées par la CC. Par exemple ce travail (https://www.uni-salzburg.at/fileadmin/oracle_file_imports/1461242.PDF).
 |
| http://www.oecd.org/els/health-systems/35987490.pdf |
Sur cette corrélation de plusieurs pays l'hypothèse d'une corrélation et donc d'une causalité est peu probable vu le faible coefficient de corrélation. De surcroit, on voit clairement que si relation de corrélation il y a, c'est pour les pays peu developpés c'est à dire soit avec une faible densité médicale soit avec des dépenses faibles (Mexique).
 |
| http://www.oecd.org/els/health-systems/35987490.pdf |
4/ La CC ignore ce que des organismes très spécialisés ont étudié en détail au sujet des besoins de soins et non de santé des populations. L'OCDE est extrêmment précis et factuel dans ses travaux qui sont très réguliers dans le domaine des dépenses de soins et des paramètres de des populations. Par exemple et cela n'a certainement pas échappé aux membres de la CC les dépenses de soins dans les pays développés ne sont pas corrélées aux densités de médecins...
6/ Dès lors peut on attendre d'une baisse de la densité médicale une baisse des dépenses de soins? Probablement pas ou alors une baisse très faible, ce travail de revue de la litérature le démontre.
" In the absence of more accurate information, measures encouraging physician supply limitation to contain costs should also be considered cautiously."
https://www.sciencedirect.com/science/article/pii/S0168851008002789
http://users.sussex.ac.uk/~grahamh/RM1web/Eight%20things%20you%20need%20to%20know%20about%20interpreting%20correlations.pdf
Aucun commentaire:
Enregistrer un commentaire