3.18 Mortality Rate

(Download as PDF)

Description of Indicator

Relationship to Quality Medical care should be effective
Type of Indicator Hospital outcome, Hospital level
Proposed Data Source Trauma Registry, Administrative Data
Definition Number of patients admitted to hospital* with an injury diagnosis who die‡ per 100 patients†
Numerator All patients age 18 years and older admitted to hospital with an injury diagnosis who die‡
Denominator All patients age 18 years and older admitted to hospital with an injury diagnosis
Benchmark Not specified at presentβ
Risk Adjustmentα Age, sex, pre-existing conditions and a validated ISS [e.g. abbreviated ISS (AIS) or International Classification of Diseases–based ISS (ICISS)]

* Indicator is restricted to patients admitted to hospital. Patients who die prior to hospital arrival or who die in the ED prior to being admitted to hospital are excluded.
‡ We propose calculating two risk adjusted mortality rates:
1) One for deaths during initial hospital stay = in hospital mortality.
2) One for deaths during the first 12 months following injury = mortality rate 12 months following injury.
Although data for in hospital mortality is easier to obtain than data for 12 month mortality, both measures provide potentially important and slightly different evaluations of patient care outcome.

† How to calculate Risk-Adjusted Mortality Rate:
Risk-adjusted Mortality = [Observed Mortality Rate/Risk-adjusted Expected Mortality Rate (x100)] x Overall Mortality Rate in the standard population.
Note: Standard population refers to a population of institutions under evaluation (e.g. institutions contributing data to a national trauma registry or centrally collected administrative data bank).
Alternatively risk adjusted mortality can be calculated directly from parameter estimates from a multivariable risk adjusted model examining data from individual institutions or from multiple institutions88.
ß Mean mortality across all centers excluding the center under evaluation is one possible benchmark that can be considered88.

 

Summary
This indicator is intended to monitor risk-adjusted mortality in hospital and 12 months from injury and allow comparisons across institutions.

 

Panel Review
Panelists agreed that this is a key measure as it addresses a widely accepted outcome. The panel discussed the challenges related to data collection, analysis, risk adjustment and interpretation. They highlighted that multiple approaches for risk adjustment for this indicator exist, each with their respective strengths and limitations. These include using a Standardized Mortality Ratio (SMR) or direct estimation from parameters derived from multivariable regression models. If SMR is used then particular attention needs to be paid to the external standard applied to ensure that it accurately reflects a standard population. It was suggested that risk adjustment may be most effective and best accepted if applied using a validated and published instrument.

 

Trauma Center Review
The trauma centers suggested that the definition of this indicator could be refined so that it does not capture patients with what are expected to be fatal injuries. Centers emphasized the importance of the indicator, but highlighted challenges of implementation including data linkages (e.g. trauma registries and death records) to provide 12 month follow up.


Review of Literature & Evidence
Face Validity: No studies identified.
Construct Validity: Most studies show no correlation with individual measures of process but some correlation with composite measures of process47. Two studies showed poor agreement between risk adjustment using TRISS and ASCOT89,90.
Reliability: No studies identified.
Risk Adjustmenta: Several variables have been evaluated in studies of risk adjustment of patient mortality90,91.
Utilization: Measures of hospital mortality are used by a large number of trauma centers: USA 52% (105/200), Canada 71% (25/35), Australasia 8% (1/12).
Nakahara et al.92 developed a simplified method to predict survival probability with the objective of risk adjustment. The model included only three predictors: age, anatomical injury severity described in the Injury Surveillance Guidelines, and a physiological status parameter90.
The simplified model may allow for efficiencies in data collection by enabling the use of injury surveillance data for both injury prevention and risk adjustment in quality evaluation88.

 

Source
Risk adjusted mortality has been proposed as a quality indicator by professional trauma societies, patient advocacy groups and researchers2,89,90.

 

References
2. American College of Surgeons Committee on Trauma. Resources for Optimal Care of the Injured Patient 2006. Chicago: American College of Surgeons; 2006.
47. Cryer HG, Hiatt JR, Fleming AW, Gruen JP, Sterling J. Continuous use of standard process audit filters has limited value in an established trauma system. J Trauma. 1996;41(3):389-394; discussion 394-385.
88. Moore L, Hanley JA, Turgeon AF, Lavoie A, Eric B. A new method for evaluating trauma centre outcome performance: TRAM-adjusted mortality estimates. Ann Surg. 2010;251(5):952-958.
89. Glance LG, Osler TM, Dick A, Mukamel D. The relation between trauma center outcome and volume in the National Trauma Databank. J Trauma. 2004;56(3):682-690.
90. Glance LG, Osler TM, Dick AW. Evaluating trauma center quality: does the choice of the severity-adjustment model make a difference? J Trauma. 2005;58(6):1265-1271.
91. Osler T, Rutledge R, Deis J, Bedrick E. ICSS: an international classification of disease-9 based injury severity score. J Trauma. 1996;41(3):380-386, discussion 386-388.
92. Nakahara S, Ichikawa M, Kimura A. Simplified alternative to the TRISS method for resource-constrained settings. World J Surg. 2011;35(3):512-519.