有鑑於全民健保實施以來，健保欠費或健保中斷者人數攀升，健保局於2003年實施為期一年之「健保特赦條款」，本研究運用健保局行政資料分析該方案參與者之特性，檢視該方案是否真正照顧到弱勢族群。研究結果顯示205,754人或107,272戶之參與者，以15~44歲、男性、失業者為居多數，通常被排除在現有社會救助體系之外。此外，透過5%之樣本戶所得資料分析，結果指出參與者樣本之經濟劣勢程度與現有低收入戶相近，屬近貧階級。然而，參與者之持續欠費與無力負擔延緩積欠的保費，顯示該方案雖保障一時之就醫可及性，卻未解決貧窮邊緣戶長期無法繳納健保費的困境。此外，本研究也發現有超過一成的方案受益家戶實無經濟困難之虞。因此，本研究建議一方面有必要放寬全民健保保費補助之經濟門檻，另一方面應對於以人口、家庭、社會特性為目標人口的保費補助設有經濟門檻之配套，甚或全面改以經濟門檻為單一補助資格標準。

In respect to an increasing number of enrollees who dropped out the National Health Insurance (NHI) program due to premium arrears, the Bureau of National Health Insurance enacted a 1-year premium relief program for the economically underprivileged people in 2003. Eligible beneficiaries could apply for either premium waiver or installment according to their economic status. This study aims to analyze the characteristics of program participants using the NHI administration data. The results show that 205,754 individuals or 107,272 households were mainly aged between 15-44, male, unemployed, those who were conventionally excluded from formal social assistance systems. In addition, by analyzing the tax information of sample (5% of households), the results indicate that a majority of the program participants were marginally above the poverty line, and hence, they were not qualified as low-income to be exempted from the NHI premium. Moreover, although this 1-year program had protected this near-poor population's temporary access to care from 2003 to 2004, the outstanding arrears and new premium payments will continue to accumulate after the termination of this relief program. Clearly, they won’t be able to afford such a large financial burden. Therefore, a more generous and long-term premium subsidy policy, instead of a short-term relief program, is critical for the near poor. Besides, the results show that over 10% of sampled households were quite affluent in contrast with the household income of the general population. Therefore, the study suggests that the premium subsidy criteria should be either exclusively income-tested or income-tested in complement to a categorical arrangement for specific population groups.