OBJECTIVES: Network meta-analysis is often performed with aggregate level data (AD).  A challenge with meta-regression models using AD is that the association between a patient level covariate and relative treatment effects of the compared interventions at the study level may not reflect the individual level effect-modification. In this paper, non-linear network meta-analysis models for combining individual patient data (IPD) and AD are presented to reduce bias and uncertainty of treatment effects in the presence of heterogeneity due to patient characteristics. METHODS: The first method uses the same model form for IPD and AD. With the second method, the model for AD is obtained by integrating an underlying IPD model over the joint within-study distribution of covariates. With a simple simulation study the two modeling approaches are compared.  RESULTS: Having IPD for a subset of studies improves estimation of treatment effects with network meta-analysis in the presence of patient level heterogeneity and inconsistency. Of the two proposed non-linear models for combining IPD and AD, the second approach seems less affected by bias. Additional studies, however, are needed to assess the value of both methods. CONCLUSIONS: Overall, for network meta-analysis it is recommended to use IPD when available, rather than treating all studies as AD.