Bug: lambda.z.ix still not always correct

[js3110]

Description

Lambda z ix is not always correct.

To reproduce

Run NCA on dummy data, for dose 1 2 and 3.
PT 11106 Dose 2 has lambda.z.ix of 6, when it should be 4,5,6

Expected behaviour

lambda.z.ix should be correct, and match the points in the slope plots, as well as the lambda.z.n.points and the lamda.time.first