On Some Counterintuitive Implications of Using a Likelihood Ratio Rule to Classify Patients as Functional or Dysfunctional