Hospital quality measures: are process indicators associated with hospital standardized mortality ratios in French acute care hospitals?

TitreHospital quality measures: are process indicators associated with hospital standardized mortality ratios in French acute care hospitals?
Publication TypeJournal Article
Year of Publication2017
AuthorsNgantcha, M, Le Pogam, M-A, Calmus, S, Grenier, C, Evrard, I, Lamarche-Vadel, A, Rey, G
JournalBMC health services research
Date Published08/2017
Mots-clésHospital Process Indicators, Hospital Standardized Mortality Ratio, Quality of care, SIMEX method

BACKGROUND: Results of associations between process and mortality indicators, both used for the external assessment of hospital care quality or public reporting, differ strongly across studies. However, most of those studies were conducted in North America or United Kingdom. Providing new evidence based on French data could fuel the international debate on quality of care indicators and help inform French policy-makers. The objective of our study was to explore whether optimal care delivery in French hospitals as assessed by their Hospital Process Indicators (HPIs) is associated with low Hospital Standardized Mortality Ratios (HSMRs).

METHODS: The French National Authority for Health (HAS) routinely collects for each hospital located in France, a set of mandatory HPIs. Five HPIs were selected among the process indicators collected by the HAS in 2009. They were measured using random samples of 60 to 80 medical records from inpatients admitted between January 1st, 2009 and December 31, 2009 in respect with some selection criteria. HSMRs were estimated at 30, 60 and 90 days post-admission (dpa) using administrative health data extracted from the national health insurance information system (SNIIR-AM) which covers 77% of the French population. Associations between HPIs and HSMRs were assessed by Poisson regression models corrected for measurement errors with a simulation-extrapolation (SIMEX) method.

RESULTS: Most associations studied were not statistically significant. Only two process indicators were found associated with HSMRs. Completeness and quality of anesthetic records was negatively associated with 30 dpa HSMR (0.72 [0.52-0.99]). Early detection of nutritional disorders was negatively associated with all HSMRs: 30 dpa HSMR (0.71 [0.54-0.95]), 60 dpa HSMR (0.51 [0.39-0.67]) and 90 dpa HSMR (0.52 [0.40-0.68]).

CONCLUSION: In absence of gold standard of quality of care measurement, the limited number of associations suggested to drive in-depth improvements in order to better determine associations between process and mortality indicators. A smart utilization of both process and outcomes indicators is mandatory to capture aspects of the hospital quality of care complexity.

Alternate URL



Alternate JournalBMC Health Serv Res
Citation Key / SERVAL ID8220
Peer reviewRefereed
PubMed ID28830422
PubMed Central IDPMC5568353

Institut universitaire de médecine sociale et préventive
Route de la Corniche 10, 1010 Lausanne - Switzerland
+41 21 314 72 72 |

Go to top