Improvements of risk factor analysis

[jreps]

The risk factor analysis will have the following modifications:

1. Addition of min standardized mean difference input - this will mean only risk factors with a standardized mean difference >= min standardized mean difference input will be output. This should significantly reduce the number of rows returned by risk factor analysis.

2. Risk factor analysis will run in three modes (Efficient, CohortIncidence and PatientLevelPrediction). The mode is a global input. The efficient mode will compare target vs cases this can be run very efficiently. The CohortIncidence and PatientLevelPrediction will compare non-case vs case but have different logic for who is considered a non-case based on exclusions done in the CohortIncidence and PatientLevelPrediction R packages. This will enable users to priorities speed (Efficient mode) or ensure case counts match other analyses (CohortIncidence and PatientLevelPrediction).