Why SQLape is better than DRG-like solution?
Patients are assigned to a single category, which is supposed to summarize all its characteristics. This inevitably leads to categories that are difficult to interpret medically.
In the race for the explanatory power of the classification system, other systems include predictive variables (complications, ventilation times, etc.) which can be harmful for patients.
Current classification systems have too many outliers, diverting significant hospital resources and leaving uncertainty about whether additional costs stem from complications or pricier medical practices.
Classification depends enormously on the choice of the principal diagnosis. Patients who have had several surgeries are underpaid. Moreover it leads to bias in medical coding.
Our algorithm identifies several significant issues for each case. The classification of diseases is based on the organs affected and the pathology (infection, cancer, etc). Procedures are based on the organs operated on, the operative gesture and access. This typology is significantly more meaningful for doctors.
The tool compares observed and expected costs according to the patient profile, distinguishing the contribution of extreme cases, iatrogenic complications and medical practices. This is very important for setting the cost control strategy.
Diagnoses likely to be complications or immediate causes of death and the given care (ventilation for instance) are not involved in the classification and thus they are not skewing the results.
The hierarchy between the problems (primary or secondary) is established by the grouper and does not depend on the coding.
Current patient classification systems often have around 1000 categories. Instinctively it seems that the more the better but it's not. There are only 6000 hospital stays in an average US hospital per year. Too many categories means less statistical accuracy and interpretability.
The number of categories is limited to around 200 diagnosis and 200 procedures. It allows for better statistical accuracy and fewer extreme cases.
Traditional Patient Classification Tools