Improving the Performance of Risk Adjustment Systems: Constrained Regressions, Reinsurance, and Variable Selection

Risk Adjustment in Health Insurance Markets: Do Not Overlook the "Real" Healthy

OBJECTIVES

The goals of this paper are (1) to identify groups of healthy people and (2) to quantify the extent to which the Dutch risk adjustment (RA) model overpays insurers for these groups.

BACKGROUND

There have been strong signals that insurers in the Dutch regulated health insurance market engage in actions to attract healthy people. A potential explanation for this behavior is that the Dutch RA model overpays insurers for healthy people.

METHODS

We identify healthy groups using 3 types of ex-ante information (ie, information available before the start of the health insurance contract): administrative data on prior spending for specific health care services (N = 17 m), diagnoses from electronic patient records (N = 1.3 m), and health survey data (N = 457 k). In a second step, we calculate the under/overpayment for these groups under the Dutch RA model (version: 2021).

RESULTS

We distinguish eight groups of healthy people using various "identifiers." Although the Dutch RA model substantially reduces the predictable profits that insurers face for these groups, significant profits remain. The mean per person overpayment ranges from 38 euros (people with hospital spending below the third quartile in each of 3 prior years) to 167 euros (those without any medical condition according to their electronic patient record).

CONCLUSIONS

The Dutch RA model does not eliminate the profitability of healthy groups. The identifiers used for flagging these groups, however, seem inappropriate for serving as risk adjuster variables. An alternative way of exploiting these identifiers and eliminating the profitability of healthy groups is to estimate RA models with "constrained regression."

Identifying undercompensated groups defined by multiple attributes in risk adjustment

Objective

To identify undercompensated groups in plan payment risk adjustment that are defined by multiple attributes with a systematic new approach, improving on the arbitrary and inconsistent nature of existing evaluations.

Methods

Extending the concept of variable importance for single attributes, we construct a measure of ‘group importance’ in the random forests algorithm to identify groups with multiple attributes that are undercompensated by current risk adjustment formulas. Using 2016–2018 IBM MarketScan and 2015–2018 Medicare claims and enrolment data, we evaluate two risk adjustment scenarios: the risk adjustment formula used in the individual health insurance Marketplaces and the risk adjustment formula used in Medicare.

Results

A number of previously unidentified groups with multiple chronic conditions are undercompensated in the Marketplaces risk adjustment formula, while groups without chronic conditions tend to be overcompensated in the Marketplaces. The magnitude of undercompensation when defining groups with multiple attributes is many times larger than with single attributes. No complex groups were found to be consistently undercompensated or overcompensated in the Medicare risk adjustment formula.

Conclusions

Our method is effective at identifying complex undercompensated groups in health plan payment risk adjustment where undercompensation creates incentives for insurers to discriminate against these groups. This work provides policy-makers with new information on potential targets of discrimination in the healthcare system and a path towards more equitable health coverage.