Health Plan Payment in Switzerland

How the COVID-19 pandemic can distort risk adjustment of health plan payment

“…Where payments to plans are fixed ex ante—an approach assumed in our hypothetical example and common practice in the Netherlands—under-/overpayment can occur both at the level of individual insurers and at the level of the entire system. Where payments are designed to sum to zero (ex post)—such as in the Swiss basic health insurance [ 3 ] and in the US Marketplaces [ 4 ]—under-/overpayment might occur at the insurer level, but not at the market level. When it comes to the effects of distortions on under-/overpayment it is also important to emphasize that the analysis in this paper is limited to one year (or one contract period).…”

“…The aforementioned changes in utilization and spending are likely to affect risk adjuster flags in health plan payment systems. For example, a reduction of regular hospital treatments is likely to result in fewer disease flags in morbidity classifications based on diagnoses from hospital settings (such as the Diagnoses-based Cost Groups used in the Netherlands [ 2 ], the Hierarchical Morbidity Groups used in Germany [ 11 ], the Hierarchical Condition Categories used in the United States [ 4 , 5 ] and the indicator based on prior hospitalization used in Switzerland [ 3 ]). Below, we describe how such effects complicate the calibration of payment weights and the calculation of payments to insurers.…”

“…Such risk adjusters are commonly referred to as ‘prospective’ risk adjusters, as opposed to ‘concurrent’ risk adjusters that are based on information from the payment year itself. Prospective risk adjusters are used in Germany (where Hierarchical Morbidity Groups are based on diagnoses from the prior year [ 11 ]) and Switzerland (where Pharmaceutical Cost Groups and an indicator for yes/no hospitalization are based on information from the prior year [ 3 ]). The Netherlands goes even further using information from up to five consecutive prior years as a basis for spending-based risk adjusters [ 2 ].…”

“…Moreover, in systems where payments to plans are ‘fixed’ ex ante—such as in the Netherlands [ 2 ]—under-/overpayment might also occur at the market level. In systems where payments are designed to sum to zero (ex post)—such as in Switzerland [ 3 ] and in the US Marketplaces [ 4 ]—under-/overpayment might occur at the insurer level, but not at the market level.…”

“…For example, the Netherlands uses nationwide spending data from year t -3 (in combination with morbidity flags from t -4 as well as other risk adjusters) to calibrate weights for payment year t [ 2 ]. Switzerland uses somewhat more recent spending data, from t -1 in combination with morbidity flags from t -2 as well as some other risk adjusters [ 3 ]. Systems in the United States, such as the Marketplaces (Affordable Care Act) and Medicare Advantage rely on calibration data from both a prior period and a subpopulation [ 4 , 5 ].…”