In California, no reporting requirement exists for self-insured employers that would have provided the data necessary to conduct this study. For this reason, data on self-insured claims were excluded from Peterson et al. (1998). For the study described in this report, RAND undertook an unprecedented effort to collect essential data from self-insured employers in California. This work was accomplished with assistance from the California Self-Insured Plans (the self-insured employers regulator), the California Self-Insurers Association (the employers lobbying group), and the Commission on Health and Safety and Workers Compensation.
The data collected from self-insured employers were then connected to wage data from the State of California to create a one-of-a-kind database of proprietary employer data linked to state administrative data. In this chapter, we describe the self-insured data collection effort and administrative data link. (More information on the self-insured data can be found in Appendix A.) This chapter also briefly describes the data on claims at insured firms in California, which are described in greater detail in Peterson et al. (1998) and Reville (1999).
In response to a request from CHSWC asking for estimated earnings losses at self- insured employers in California, RAND contacted a sample of 150 private, self-insured firms (out of a total of 466) and 150 public self-insured firms (out of a total of 432) and requested data on all indemnity claims from 1991 through 1996. The sample was based on the number of claims at the employer; therefore, the result was a representative sample of claims data from self-insured employers (rather than just a sample of self-insured firms), and the sample was stratified by employer size to increase the probability of selection for small self-insured employers.1
Specifically, we requested data on benefit amounts paid and incurred, injury dates, and individual identifiers to facilitate linking to earnings data maintained by the State of California EDD. Because we were concerned that requesting numerous data elements would lead to a lower response rate, we requested additional data that could be provided optionally, such as disability ratings, settlement method, and litigation indicators.____________________