diff --git a/src/systems/abstractsystem.jl b/src/systems/abstractsystem.jl
index 51a495e4ff..efa586fc87 100644
--- a/src/systems/abstractsystem.jl
+++ b/src/systems/abstractsystem.jl
@@ -235,7 +235,7 @@ function SymbolicIndexingInterface.parameter_index(sys::AbstractSystem, sym::Sym
     if has_index_cache(sys) && (ic = get_index_cache(sys)) !== nothing
         idx = parameter_index(ic, sym)
         if idx === nothing ||
-           idx.portion isa SciMLStructures.Discrete && idx.idx[2] == idx.idx[3] == 0
+           idx.portion isa SciMLStructures.Discrete && idx.idx[1] == idx.idx[2] == 0
             return nothing
         else
             return idx
diff --git a/test/symbolic_indexing_interface.jl b/test/symbolic_indexing_interface.jl
index 89b6907f4b..400ba1fa6c 100644
--- a/test/symbolic_indexing_interface.jl
+++ b/test/symbolic_indexing_interface.jl
@@ -2,6 +2,7 @@ using ModelingToolkit, SymbolicIndexingInterface, SciMLBase
 using ModelingToolkit: t_nounits as t, D_nounits as D, ParameterIndex,
                        SymbolicContinuousCallback
 using SciMLStructures: Tunable
+using OrdinaryDiffEq
 
 @testset "System" begin
     @parameters a b
@@ -236,3 +237,14 @@ end
     @test timeseries_parameter_index(sys, p[1, 1]) ===
           ParameterTimeseriesIndex(1, (1, 1, 1, 1))
 end
+
+@testset "Indexing with symbols work for discrete parameters" begin
+    t = ModelingToolkit.t_nounits; D = ModelingToolkit.D_nounits
+    @variables x(t) = 1.
+    @parameters p(t) = 1.
+    ev = ModelingToolkit.SymbolicDiscreteCallback((t == 1) => [p ~ Pre(p)*2], discrete_parameters = p)
+    @mtkcompile sys = System([D(x) ~ p], t; discrete_events = ev)
+    prob = ODEProblem(sys, [], (0., 2.))
+    sol = solve(prob, Tsit5())
+    @test sol[p] == sol[:p] == sol.ps[p] == sol.ps[:p]
+end