diff --git a/jwst/badpix_selfcal/badpix_selfcal.py b/jwst/badpix_selfcal/badpix_selfcal.py
index 2c2b19a857c..1b6b421019c 100644
--- a/jwst/badpix_selfcal/badpix_selfcal.py
+++ b/jwst/badpix_selfcal/badpix_selfcal.py
@@ -3,6 +3,7 @@
 import jwst.datamodels as dm
 from jwst.outlier_detection.outlier_detection_ifu import medfilt
 from stdatamodels.jwst.datamodels.dqflags import pixel
+import warnings
 
 
 def badpix_selfcal(minimg: np.ndarray,
@@ -52,7 +53,9 @@ def badpix_selfcal(minimg: np.ndarray,
     elif dispaxis == 1:
         kern_size = (1, kernel_size)
 
-    smoothed = medfilt(minimg, kern_size)
+    with warnings.catch_warnings():
+        warnings.filterwarnings("ignore", category=RuntimeWarning, message="All-NaN")
+        smoothed = medfilt(minimg, kern_size)
     minimg_hpf = minimg - smoothed
 
     # Flag outliers using percentile cutoff
@@ -82,7 +85,7 @@ def apply_flags(input_model: dm.IFUImageModel, flagged_indices: np.ndarray) -> d
         Flagged data model
     """
 
-    input_model.dq[flagged_indices] += pixel["DO_NOT_USE"] + pixel["OTHER_BAD_PIXEL"]
+    input_model.dq[flagged_indices] |= pixel["DO_NOT_USE"] + pixel["OTHER_BAD_PIXEL"]
 
     input_model.data[flagged_indices] = np.nan
     input_model.err[flagged_indices] = np.nan
diff --git a/jwst/badpix_selfcal/badpix_selfcal_step.py b/jwst/badpix_selfcal/badpix_selfcal_step.py
index bfb6a43aaae..f886b3d4efb 100644
--- a/jwst/badpix_selfcal/badpix_selfcal_step.py
+++ b/jwst/badpix_selfcal/badpix_selfcal_step.py
@@ -1,5 +1,6 @@
 
 
+import warnings
 from ..stpipe import Step
 from . import badpix_selfcal
 import numpy as np
@@ -64,7 +65,7 @@ def process(self, input, selfcal_list=None, bkg_list=None):
         are included in the MIN frame from which outliers are detected.
         If selfcal_list and/or bkg_list are specified manually, they are appended to any
         selfcal or background exposures found in the input association file.
-        If selfcal_list and bkg_list are both set to None and input is 
+        If selfcal_list and bkg_list are both set to None and input is
         a single science exposure, the step will be skipped with a warning unless
         the force_single parameter is set True.
         In that case, the input exposure will be used as the sole background exposure,
@@ -96,7 +97,9 @@ def process(self, input, selfcal_list=None, bkg_list=None):
         selfcal_3d = []
         for i, selfcal_model in enumerate(selfcal_list):
             selfcal_3d.append(selfcal_model.data)
-        minimg = np.nanmin(np.asarray(selfcal_3d), axis=0)
+        with warnings.catch_warnings():
+            warnings.filterwarnings("ignore", category=RuntimeWarning, message="All-NaN")
+            minimg = np.nanmin(np.asarray(selfcal_3d), axis=0)
         bad_indices = badpix_selfcal.badpix_selfcal(minimg, self.flagfrac_lower, self.flagfrac_upper, self.kernel_size, dispaxis)
 
         # apply the flags to the science data
@@ -155,7 +158,7 @@ def _parse_inputs(input, selfcal_list, bkg_list):
 
             if len(sci_models) > 1:
                 raise ValueError("Input data contains multiple science exposures. "
-                                    "This is not supported in calwebb_spec2 steps.")
+                                 "This is not supported in calwebb_spec2 steps.")
             input_sci = sci_models[0]
 
         elif isinstance(input_data, dm.IFUImageModel) or isinstance(input_data, dm.ImageModel):