Skip to main content

Table 5 Average marginal effect of expected use and cost & covariate[−adjusted baseline trend significance tests

From: Impacts of innovation in dental care delivery and payment in Medicaid managed care for children and adolescents

Total Observations = 1,236,967

Expected Use: ZINB

Expected Cost: Two-part Model ($)

Dependent Variables: Dental Service Types

Baseline Trend Diff.(1)

Average Marginal Effect

Baseline Trend Diff.(5)

Average Marginal Effect

All (2)

R_Full (3)

R_Reduce (4)

 

All (6)

R_Full (7)

R_Reduce (8)

Any Dental Service (counting all following categories)

−0.023(− 0.056, 0.010)

0.027(− 0.007, 0.061)

0.027(− 0.006, 0.060)

0.024 (− 0.010, 0.058)

−0.573 (−1.309, 0.162)

0.645** (0.012,1.278)

0.636** (0.012,1.259)

0.585* (− 0.067,1.237)

Any Preventive Service (CDT D1000-D1999)

− 0.007 (− 0.017, 0.003)

0.007 (− 0.006, 0.021)

0.007 (− 0.006, 0.020)

0.006 (− 0.007, 0.020)

−0.174 (− 0.402, 0.055)

0.166 (− 0.082, 0.414)

0.163 (− 0.083, 0.409)

0.141 (− 0.110, 0.393)

Any Diagnostic Service (CDT 0001–0999)

− 0.009 (− 0.021, 0.004)

0.009** (0.000, 0.018)

0.009** (0.000, 0.017)

0.008* (− 0.001, 0.017)

− 0.134 (− 0.296, 0.028)

0.142 (− 0.030, 0.313)

0.140 (− 0.030, 0.310)

0.125 (− 0.051, 0.300)

Fluoride Varnish (CDT D1206)

− 0.001 (− 0.006, 0.003)

0.002 (− 0.001, 0.004)

0.004 (− 0.003, 0.011)

0.001 (− 0.003, 0.004)

−0.017 (− 0.077, 0.043)

0.007 (− 0.029, 0.042)

0.006 (− 0.029, 0.042)

0.003 (− 0.031, 0.036)

Topical Fluoride, except Varnish (CDT D1208)

0.000 (− 0.001, 0.001)

0.001** (0.000, 0.002)

0.004 (− 0.015, 0.023)

0.000 (− 0.002, 0.002)

−0.005 (− 0.019, 0.009)

0.011 (− 0.005, 0.027)

0.011 (− 0.005, 0.026)

0.010 (− 0.007, 0.027)

Sealants (CDT D1351)

−0.001 (− 0.004, 0.002)

0.003** (0.000, 0.006)

0.003** (0.000, 0.006)

0.000 (− 0.004, 0.004)

−0.027 (− 0.090, 0.036)

0.064* (− 0.007, 0.135)

0.064* (− 0.006, 0.133)

0.062* (− 0.010, 0.135)

Caries Arrest, by Silver Diamine Fluoride (CDT D1354)

−0.000 (− 0.005, 0.004)

0.002 (− 0.005, 0.009)

0.002 (− 0.005, 0.009)

0.044*** (0.041, 0.048)

−0.011 (− 0.076, 0.054)

0.026 (− 0.058, 0.109)

0.026 (− 0.058, 0.110)

0.024 (− 0.061, 0.109)

Any Restorative Service (D2000–2999)

−0.003 (− 0.008, 0.002)

0.004 (− 0.002, 0.009)

0.003 (− 0.002, 0.009)

0.004 (− 0.002, 0.009)

−0.198 (− 0.489, 0.094)

0.205 (− 0.066, 0.476)

0.203 (− 0.062, 0.467)

0.216 (− 0.060, 0.492)

Any Extraction Services (D7140, D7210, D7250)

−0.000 (− 0.001, 0.001)

0.000 (− 0.001, 0.001)

0.000 (− 0.001, 0.001)

0.000 (− 0.001, 0.001)

−0.033 (− 0.159, 0.093)

0.049 (− 0.073, 0.172)

0.048 (− 0.076, 0.172)

0.037 (− 0.083, 0.157)

  1. 1. 95% confidence interval (CI) in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1
  2. 2. All values and their 95% confidence intervals shown here are covariate-adjusted differences between control and PREDICT groups. For baseline period, they are model-adjusted differences between baseline values. For DiD, they are model-adjusted average marginal differences between PREDICT and control group, considering their pre- and post- values
  3. 3. Model Setups (all models used “cluster” for county, robust standard errors generated):
  4. a. Baseline differences: utilization/cost = function (secular time [1–8 representing 8 baseline quarters], dummy[control vs. PREDICT], interaction[time * control/PREDICT], seasonality dummies, all usual covariates)
  5. b. DiD – “All”: utilization/cost = function (dummy[pre-period vs. post-period],], dummy[control vs. PREDICT], interaction [dummy[pre-period vs. post-period * control/PREDICT], all usual covariates)
  6. c. DiD – R_Full”: same as “All” model above, but with addition of “Seasonality” dummy variable
  7. d. DiD – R_Reduce: function (dummy[pre-period vs. post-period], dummy[control vs. PREDICT], interaction[pre/post* control/PREDICT], seasonality dummies), but no other covariates included in this model
  8. 4. Setting of the Difference-in-Differences Average Marginal Effects:
  9. a. Use: [Use (PREDICT, post-intervention period 2016–2017) – Use (PREDICT, pre-intervention period 2014–2015)] – [Use (CONTROL, post-intervention period 2016–2017) – Use (CONTROL, pre-intervention period 2014–2015)],
  10. b. Cost: same structure as “Use”, replace all “use” items by their corresponding “cost” items
  11. 5. AME on Expected Use illustrate the change in expected use as an impact from the PREDICT program
  12. 6. AME on Expected Cost illustrate the change in the expected cost as an impact from the PREDICT program
  13. 7. The ZINB models on “Use of Fluoride Varnish”, “Use of Topical Fluoride”, and “Use of Caries Arrest” did not achieve convergence, so the coefficient on their effects are not statistically reliable
  14. 8. Significance levels were calculated using the original un-rounded numbers, so the calculation of the face value (4-digit point rounded) in the table may not show significance levels as indicated due to rounding omissions
Back to article page

BMC Health Services Research

ISSN: 1472-6963

Contact us