Skip to main content

Table 1 Estimates of homoscedastic probit model and heteroskedastic probit model, HIECS, 2015

From: Predictors of catastrophic out-of-pocket health expenditure in rural Egypt: application of the heteroskedastic probit model

Variables

Probit model

Heteroskedastic probit model

 

Coef.

Marginal effect

Coef.

"Mean model"

Marginal effect

Intercept

− 2.407*** (0.504)

 

− 2.398* (1.239)

 

Male-headed household

0.363*** (0.087)

0.045*** (0.009)

0.816*** (0.204)

0.061** (0.053)

Age of household head

− 0.012 (0.0127)

− 0.001 (0.002)

− 0.010 (0.026)

− 0.001 (0.001)

Squared age of household head

0.000 (0.000)

0.000 (0.000)

− 0.010 (0.026)

0.001 (0.001)

Illiterate household head

0.126* (0.071)

0.018* (0.010)

0.239 (0.159)

0.049** (0.071)

Educational degree of household head (Dummy variables)

  Having a primary or lower secondary degree

  Having a secondary or post-secondary degree

  Having a university degree

  Having a postgraduate degree

  None (omitted category)

− 0.196 (0.091)

− 0.151* (0.091)

− 0.469** (0.154)

− 0.028 (0.474)

− 0.015 (0.011)

− 0.019* (0.010)

− 0.048*** (0.011)

− 0.003 (0.063)

− 0.151 (0.212)

− 0.352* (0.199)

− 1.273** (0.182)

− 1.311 (2.176)

− 0.013 (0.023)

− 0.053* (0.170)

− 0.094*** (0.021)

− 0.083* (0.063)

Employed household head

− 0.255*** (0.474)

− 0.038** (0.012)

− 0.528** (0.162)

− 0.072** (0.069)

Geographical area: rural lower Egypt

0.722* (0.367)

0.126* (0.065)

1.868* (1.003)

0.321** (0.091)

Household size

− 0.0206 (0.027)

− 0.002 (0.003)

0.158** (0.060)

0.032** (0.021)

Proportion of females

0.437*** (0.132)

0.059*** (0.018)

0.592** (0.283)

0.082** (0.031)

Proportion of children in household

0.186 (0.169)

0.025 (0.023)

0.297 (0.397)

0.064 (0.317)

Proportion of elderly in household

0.271* (0.135)

0.037* (0.018)

0.200** (0.260)

0.029** (0.012)

Presence of chronic diseases

0.324*** (0.064)

0.039*** (0.007)

0.769*** (0.152)

0.097*** (0.032)

Presence of disability

0.164 (0.012)

0.025 (0.021)

0.118 (0.261)

0.101 (0.073)

Proportion of sick members in household

0.218* (0.19)

0.031* (0.017)

0.506* (0.269)

0.063** (0.072)

Presence of insurance coverage

− 0.213*** (0.064)

− 0.027*** (0.007)

− 0.447** (0.159)

− 0.042*** (0.029)

Household total expenditure a

− 0.003 (0.002)

− 0.009 (0.007)

− 0.167*** (0.012)

− 0.071*** (0.47)

Proportion of wage earners in household

− 0.219 (0.178)

− 0.027 (0.022)

− 4.052** (0.495)

− 0.271*** (0.149)

Proportion of educated members in household

− 0.551*** (0.164)

− 0.066*** (0.019)

− 1.373** (0.473)

− 0.332*** (0.017)

Proportion of children covered by Health Insurance

− 0.355 (0.266)

− 0.039 (0.029)

− 1.595 (1.014)

− 0.071 (0.017)

Proportion of elderly covered by Health Insurance

0.238 (0.912)

0.035 (0.721)

0.934 (1.127)

0.011 (0.23)

Type of health insurance coverage (Dummy

variables)

  Covered by HIO

− 0.028 (0.174)

− 0.004 (0.023)

− 0.229 (0.423)

− 0.061 (0.093)

  Covered by private insurance

0.391 (0.433)

0.068 0.093)

0.617 (1.02)

0.252 (0.093)

  Covered by employer-provided private insurance

− 0.279 (0.259)

− 0.031 (0.024)

− 0.594 (0.656)

− 0.081 (0.074)

  Covered by occupational syndicates

0.391 (0.284)

0.068 (0.061)

1.285* (0.684)

0.168* (0.051)

  Other (omitted category)

    

Health provider (Dummy variables)

  HIO hospitals

4.793** (0.169)

0.027 (0.028)

0.161* (0.400)

0.012* (0.021)

  Public hospitals

4.275*** (0.101)

0.085 (0.017)

0.153* (0.224)

0.031* (0.021)

  Private hospitals

0.531*** (0.064)

0.066*** (0.009)

1.097* (0.146)

0.389** (0.071)

  Pharmacy

− 0.081 (0.124)

− 0.026 (0.012)

− 0.418 (0.245)

− 0.076 (0.032)

  Other(omitted category)

    

 Using outpatient health services

0.131 (0.101)

0.019 (0.017)

0.049* (0.242)

0.091* (0.021)

 Het-test χ2

------

 

52.24***

 

 LogL

− 1566.968

 

− 1541

 

 Likelihood ratio test

229.13***

   

 McFadden’s pseudo R2

0.069

   

 AIC

3175.94

 

3125.69

 

 BIC

3317.12

 

3273.603

 

 Number of fisher scoring iterations

6

 

10

 

N

6143

 

6143

 
  1. ***p<0.001, **p<0.01, *p<0.05, ( ): robust standard errors against misspecification.
  2. aCoefficient and standard error of household total expenditure in latent scale model is{0.003*** (0.002)}.
  3. LogL is the maximum value of the likelihood function.
  4. Likelihood ratio test measures the significance of the overall model.
  5. Pseudo R2shows the amount of variance explained by explanatory variables.
  6. Wald χ2 test the significance of the full model; if all slopes equal 0.
  7. Het-test χ2 is the likelihood ratio test of the null hypothesis of homoscedasticity; it tests the model with heteroskedasticity against the full model with out.
  8. AIC and BIC are information metrics that penalize the inclusion of additional variables and used to select the appropriate model.
  9. Number of fisher scoring iterations indicates how quickly the iterative weighted least squares (IWLS) algorithm terminates.