nonrandom (version 1.42)

ps.estimate: Estimation of propensity score based treatment effects

Description

Estimation of propensity score based treatment effects

Usage

ps.estimate(object,resp,treat = NULL, stratum.index = NULL,
  match.index = NULL, adj = NULL, weights = "rr", family = "gaussian",
  regr = NULL, ...)

Arguments

object

an object of class 'stratified.pscore', 'stratified.data.frame', 'matched.pscore', 'matched.data.frame', 'matched.data.frames' or a data frame.

resp

an integer or a string indicating the outcome variable in the data and in the matched data if ps.match() is previously used.

treat

an integer or a string indicating the treatment variable in the data and in the matched data if ps.match() is previously used. If the class of the input object is 'stratified.pscore' or 'matched.pscore', no specification is ne

stratum.index

an integer or a string indicating the vector containing the stratum indices. No specification is needed if ps.makestrata() is previously used.

match.index

an integer or a string indicating the vector containing the matching indices. No specification is needed if ps.match() is previously used.

adj

a formula or a vector of integers or strings indicating covariates. The default is NULL, i.e. no additional adjustment for covariates is done in stratified or matched data. If adj is a formula, it must be formulated as 'outcome~tr

weights

a string indicating how to weight the stratum-specific estimates to obtain an overall estimate in case of stratified data. The default is 'rr' indicating stratum-specific weights proportional to the stratum-specific number of observations. In

family

a description of the error distribution and link function to be used in the model (see glm).

regr

a formula or a vector of integers or strings indicating covariates in data. The default is NULL, i.e. no regression model is fitted. If regr is a formula, it must be formulated as 'outcome~treatment+covariates' according to

...

further arguments passed to or from other methods.

Value

• ps.estimate returns an object of the same class as the input object. The number and the manner of values depends on the scale of resp:
• data
  a data frame containing the input data.
• data.matched
  a data frame limiting 'data' only to matched observations. It is only available if ps.match() is previously used.
• name.resp
  a string indicating the name of the outcome variable.
• resp
  a numeric vector indicating the outcome variable labeled by 'name.resp'.
• name.stratum.index
  a string indicating the name of the selected stratum indices. It is only available if ps.makestrata() is previously used.
• stratum.index
  a numeric vector containing the selected stratum indices labeled by 'name.stratum.index'. It is only available if ps.makestrata() is previously used.
• intervals
  a vector of characters indicating the interval used for stratification. It is only available if ps.makestrata() is previously used.
• stratified.by
  a string indicating the name of the selected stratification variable. It is only available if ps.makestrata() is previously used.
• formula.pscore
  a formula describing formally the propensity score model fitted at last.
• model.pscore
  an object of class glm containing information about the propensity score model fitted at last.
• name.pscore
  a string indicating the name of propensity score estimated at last.
• pscore
  a numeric vector containing the estimated propensity score labeled by 'name.pscore.
• name.treat
  a string indicating the name of the treatment variable.
• treat
  a numeric vector containing the treatment index labeled by 'name.treat'.
• matched.by
  a string indicating the name of the selected matching variable. It is only available if ps.match() is previously used.
• name.match.index
  a string indicating the name of the selected matching indices. It is only available if ps.match() is previously used.
• match.index
  a numeric vector containing the selected matching indices labeled by 'name.match.index' whereas '0' indicates 'no matching partner found'. It is only available if ps.match() is previously used.
• match.parameters
  a list of matching parameters including caliper, ratio, who.treated, givenTmatchingC and bestmatch.first. It is only available if ps.match() is previously used.
• lr.estimation
  a list containing information about the regression model based treatment effect estimates. It correspond to the argument regr.

  [object Object],[object Object],[object Object],[object Object],[object Object],[object Object]

• ps.estimation
  a list containing information about the estimated propensity score based treatment effects. The list entries depend on the scale of outcome. [object Object],[object Object],[object Object],[object Object],[object Object]

Details

Propensity score methods are used to estimate treatment effects in observational data. The treatment effects are estimated without adjustment and can to be interpreted as marginal effects. However, it is additionally possible to adjust for residual imbalances in strata or in the matched data (using adj) and also to apply traditional regression (using regr).

The usage of ps.estimate() depends slightly on the class of the input object. If either ps.makestrata() or ps.match() are previously used, treat, match.index and stratum.index are not needed, contrary to the case where the input object is a data frame. If both match.index and stratum.index are specified, stratum.index will be ignored. If one specifies adj and regr as formulas, they must be identical in form of 'outcome~treatment+covariates' and 'outcome' and 'treatment' must agree with resp and treat.

See Also

glm, formula, lmer

Examples

## STU1
data(stu1)
stu1.ps <- pscore(data    = stu1, 
                  formula = therapie~tgr+age)
stu1.match <- ps.match(object          = stu1.ps,
                       ratio           = 2,
                       caliper         = 0.5,
                       givenTmatchingC = FALSE,
                       matched.by      = "pscore",
                       setseed         = 38902)
stu1.est <-
  ps.estimate(object  = stu1.match,
              resp    = "pst",
              adj     = "tmass",
              regr    = pst~therapie+tgr+age)

## PRIDE
data(pride)
pride.ps <- pscore(data        = pride,
                   formula     = PCR_RSV~SEX+RSVINF+REGION+
                                 AGE+ELTATOP+EINZ+EXT,
                   name.pscore = "ps")
pride.strata <- ps.makestrata(object = pride.ps,
                              breaks = quantile(pride.ps$pscore,  
                                                seq(0,1,0.2)),
                              stratified.ps = "ps")
pride.est <-
  ps.estimate(object  = pride.strata,
              resp    = "SEVERE",
              family  = "binomial",
              adj     = "AGE",
              regr    = SEVERE ~PCR_RSV+SEX+ETHNO+FRUEHG+HERZ+
                        ELTATOP+REGION+AGE+TOBACCO+VOLLSTIL+EXT+
                        EINZ+KRANKSUM,
              weights = "rr")