Sign In
CPSI Share               



Vignette: Tallahassee Memorial

The CEO of Tallahassee Memorial, Duncan Moore learned that the organizations’ HSMR was 30% higher than the national average. Moore declared a new quality aim: to improve HSMR by 30% in the first stage, benchmark levels of performance in the second stage, and thereafter reach for perfection – no needless deaths. Tallahassee Memorial’s first task was to understand their patterns of death. They did this by pulling data from patient charts and inserting it into a 2x2 mortality table, shown below.

  • Box 1:  Patients admitted for comfort care only and to the Intensive Care Unit (ICU) on admission but who were fully expected to die.
  • Box 2:  Patients admitted for comfort care only and to a regular floor as the first stop in the hospital.
  • Box 3:  Critically ill patients not admitted for comfort care and who were admitted to an ICU as their first stop in the hospital.
  • Box 4:  Patients who were admitted to hospital, not for comfort care, and whose condition deteriorated and later died.


2x2 Mortality Table
  Admitted to ICU Not admitted to ICU
Admitted for Comfort Care only Box 1
(Range 0-14%)
Box 2
(Range 0-40%)
Not Admitted for Comfort Care Box 3
(Range 16-64%
Box 4
(Range 18-64%)


Box 1 and 2 do not contain needless deaths but provide data on how to improve palliative and comfort care services. Box 3 contains needless deaths such as patients that develop complications from critical care such as Ventilator Acquired Pneumonia or Central Line Associated Bloodstream Infections. While these complications from critical care have been thought of as inevitable, a number of hospitals have completely eliminated them, and they are needless errors. Needless deaths are also common in Box 4 and tend to relate to patients whose conditions have deteriorated, and these are important indicators of the cultural and organizational characteristics that determine high reliability and resilience. For example these patterns could be due to failure to respond to nursing concerns regarding a deteriorating patient, or the failure of multiple physicians to develop a coordinated care plan. The work around measuring HSMR at Tallahassee Memorial demonstrated that using big dot measures, such as whether the patient is alive or dead after the hospital experience, relate to system level attributes like communication, staffing levels, end of life care planning and ICU organization.

Adapted from Reinertsen and Schellekens (2005), 10 Powerful Ideas for Improving Patient Care, pgs. 18-21.