EFFECTS OF COST SHARING ON PHYSICIAN UTILIZATION N一 套豪§馨豪 字 oE82 g8=22 乐的 百 三85v9x 52 ∧ 道自 Copyright a 2001 John Wiley Sons, Ltd Health Econ.10:457-471(2001)EFFECTS OF COST SHARING ON PHYSICIAN UTILIZATION 467 Copyright © 2001 John Wiley & Sons, Ltd. Health Econ. 10: 457–471 (2001) Table 1. Fixed effects (WLSDV) estimation results for the levels model (1) t-statistic ln(price) Trend Robust Robust t-statistic Adjusted White Adjusted Durbin – S.E.a testb R2 S.E.a Watsonc GP home visits per person −16.3 −0.011 0.002 −5.5 0.9896 1.570 1.488 −0.390 0.024 ln(rate of act pop) 0.001 3.0 0.9997 0.196 1.300 0.003 −3.0 ln(rate of WOPI) −0.082 0.027 0.001 4.0 ln(rate of low-income WOPI) 0.9998 1.043 1.717 0.044 0.071 0.6 0.004 GP office visits per person ln(rate of act pop) 0.02 0.001 20.0 0.9997 0.325 1.241 −0.156 0.014 −11.1 0.001 32.0 0.9994 1.056 1.040 0.032 −2.3 0.028 −0.063 ln(rate of WOPI) 0.001 17.0 ln(rate of low-income WOPI) 0.9994 0.100 1.212 −0.048 0.031 −1.5 0.017 Specialist visits per person 0.001 23.0 0.9975 1.032 1.273 ln(rate of act pop) −0.104 0.018 −5.8 0.023 0.001 33.0 0.9975 2.497 1.454 0.033 −2.0 ln(rate of WOPI) −0.055 0.027 0.002 15.5 ln(rate of low-income WOPI) 0.9966 2.515 1.245 0.092 0.152 0.6 0.031 a Robust standard error estimates using Arellano’s method [17]. b White test statistic 1.96 indicates stat significant heteroscedasticity at 95% level. c All adjusted Durbin–Watson statistics indicate positive autocorrelation, see the table in [16].