Interpretation of the results of survival analyses containing ordered categorical variables

[Pengfei]

I am using survival analysis with several ordered categorical variables set up, and since relevel does not support specifying the reference category for ordered categorical variables:
'''r
Station.flow.type <- factor(Station.flow.type, levels = c(1, 2, 3),
labels = c("Small-scale ", "Medium-scale", "Large-scale "), ordered = TRUE)
Station.type <- factor(Station.type, levels = c(1, 2, 3, 4),
labels = c("one line only", " connecting two lines", " connecting three lines",
"connecting four lines"), ordered = TRUE)
Public.transport.usage <- factor(Public.transport.usage, levels = c(1,2,3),labels = c("Occasionale", "Daily", "Frequent"),
ordered = TRUE)
'''
So I can't interpret the results of the model：
coef exp(coef) se(coef) z Pr(>|z|)
Station.flow.type.L 3.858e-01 1.471e+00 2.582e-03 149.411 < 2e-16 ***
Station.flow.type.Q -1.413e-01 8.683e-01 1.763e-03 -80.129 < 2e-16 ***
Station.type.L -1.638e-01 8.489e-01 1.803e-02 -9.086 < 2e-16 ***
Station.type.Q -9.406e-02 9.102e-01 1.497e-02 -6.282 3.34e-10 ***
Station.type.C -2.443e-01 7.832e-01 1.110e-02 -22.020 < 2e-16 ***
Public.transport.usage.L 2.675e-01 1.307e+00 8.648e-03 30.931 < 2e-16 ***
Public.transport.usage.Q -1.532e-01 8.579e-01 5.231e-03 -29.296 < 2e-16 ***
What all does LQC refer to, I would appreciate any help!

[jrauser]

There is some relevant discussion in this SO post: r - Interpretation of .L, .Q., .C, .4… for logistic regression - Stack Overflow

[nirgrahamuk]

@Pengfei

level does not support specifying the reference category for ordered categorical variables

but you can be explicit and set whatever order you wish.
for example putting middle first then small then large like so

Station.flow.type <- factor(1:3,
  levels = c(1, 2, 3),
  labels = c("Small-scale ", "Medium-scale", "Large-scale "), ordered = TRUE
)

reordered_station_flow_type <- factor(Station.flow.type,
  levels = c(
    "Medium-scale",
    "Small-scale ",
    "Large-scale "
  )
)
Station.flow.type
reordered_station_flow_type

[Pengfei]

Thanks to jrauser and nirgrahamuk for their replies, I benefited a lot and I also found the explanation related to the question: Better Contrasts for Ordinal Variables in R • Aaron Gullickson . Thanks again for your help!