Context: The costs might be direct (e.g. hospital or ambulatory costs) or indirect (e.g. years lost because of premature death or disability).

Most patients have several diseases and allocating cost only to the main disease – as it is done by DRGs – would overestimate the costs of these diseases and underestimate those of the other diseases. This might falsify public health priorities and render cost reduction attempts ineffective. What is more, determining which morbid condition is the main one is sometimes arbitrary.

Method: Costs are allocated to multiple diseases (disease j for a patient i) by an iterative proportional repartition of total cost per patient Yi.

Example:

Initial data – iteration 1

Patients iCostsCox arthrosisPneumonia
A10,00010,000
B5,0005,000
C16,0008,0008,000
Cost weight ()9,0006,500
(58%)(42%)

Iteration 2

Patients iCostsCoxarthrosisPneumonia
A10,00010,000
B5,0005,000
C16,0009,2906,710
Cost weight ()9,6455,855
(62%)(38%)

Iteration 3

Step 20CostsCoxarthrosisPneumonia
110,00010,000
25,0005,000
316,00010,6675,333
Cost weight ()10,3335,167
(67%)(33%)

Scientific validation: Rousson V, Rossel JB, Eggli Y. Estimating Health Cost Repartition Among Diseases in the Presence of Multimorbidity. Health Serv Res Manag Epidemiol. 2019;6.

Strength of the indicator: Allocating health care costs per diseases enables a better comparison between services and between illnesses. Moreover, it is the sole way to analyze efficiency of care, since the effect should be analyzed separately for each disease: one might heal from a specific disease but worse for another one.

Another advantage of this methods it that it proportionally allocates the cost of interactions among the multiple diseases.

Limitation: The cost of possible missing diseases is allocated to identified ones.