Patients classification System

tools for
Cost management

tools for
Quality management

tools for

4  Why to choose US

4  implementation


4  contact

4  Logic

4  diagnoses

4  procedures

4  inpatients

4  outpatients

4  insurance

4  readmissions

4  reoperations

4  complications

4  deaths

4  length of stay

4  day surgery

4  unjustified stays


If the observed length of stay is higher than expected, it means that it might be reduced. But the analyst should remind that the expected length of stay correspond to the set of hospitals with high performances (see expected values below).

Similarly to the costs, the interpretation depend on the context. If the length ratio is higher among stays followed by a potentially avoidable readmission, this gives some evidence that the physicians have guessed that those patients were at more at risk of readmission. On the contrary – if the length ratio is lower – it is of concern, because it could be associated with premature discharges. A high length ratio for patients placed in nursing homes indicate that it is probably difficult to find a free nursing bed in the region, questioning the organization of the health system in the region.

The separated analysis of unjustified stays and candidates for one day surgery is necessary to compare hospitals. Otherwise, shorter stays might possibly be explained by a higher proportion of inappropriate inpatients’ admissions.

Observed length of stay (LS1)                                                                                                 

Expected length of stay

-   average (LS0)
- minimal (LS0min)

- maximal (LS0max)


Expected lengths of stay are based on average lengths per diagnosis and procedure category among Swiss hospitals with high performances in 2010 (benchmark):
- potentially avoidable readmissions rate ratio (RAR) < 1.2
- iatrogenic complications rate ratio (RCP) < 1.2
- premature deaths rate ratio (RPD) < 1.2
- inpatients costs ratio (RCO) < 1.0

Expected values are adjusted for uncompleted stays (transfer, death), complexity (number of SQLape® diagnoses and procedures categories) to avoid perverse financial effects.
Complications are excluded from adjustment. 95% confidence intervals are provided: LS0min et LS0max.

Length ratio (RLS): observed rate (LS1) / expected rate (LS0).

Statistical significance

A   significantly lower than expected rate (CO1 < CO0min)       
B   not significantly different from CO1 (CO0min < CO1 < CO0max)
C   significantly higher than expected rate (CO1 > CO0max).

© SQLape s.à.r.l. 2014. Last update: 20.05.2014