Background:
Dose-response meta-analysis (DRMA) is widely employed to establish the potential dose-response relationship between continuous exposures and disease outcomes. However, no method is readily available for exploring the relationship between a discrete exposure and a binary or continuous outcome.

Methods:
We proposed a piecewise linear (PL) DRMA model, based on both a one-stage and two-stage approach, as a solution to this issue. We used sleep (continuous exposure) and parity (discrete exposure) data as examples to illustrate how to apply a PL model in DRMA using a Stata code. We also empirically compared the slopes of the PL model with a simple linear as well as a restricted cubic spline (RCS) model.

Results:
Both a one-stage and two-stage PL DRMA model fitted well in our two examples, with similar results. Both examples showed an obvious 'piecewise effect'. In our example, the PL model showed a better fitting effect and practical, reliable results compared to a simple linear model, while results were similar for the RCS model.

Conclusions:
In conclusion, piecewise linear function is a simple and valid method for DRMA and can be used for discrete exposures. It may represent a superior model to the linear model in DRMA and an alternative to a non-linear model.

Patient or healthcare consumer involvement:
None.