diff --git a/claims_hosp/delphi_claims_hosp/__init__.py b/claims_hosp/delphi_claims_hosp/__init__.py
index 8a1e78f6c..00f9ec6c9 100644
--- a/claims_hosp/delphi_claims_hosp/__init__.py
+++ b/claims_hosp/delphi_claims_hosp/__init__.py
@@ -12,7 +12,6 @@
 from . import indicator
 from . import load_data
 from . import run
-from . import smooth
 from . import update_indicator
 from . import weekday
 
diff --git a/claims_hosp/delphi_claims_hosp/indicator.py b/claims_hosp/delphi_claims_hosp/indicator.py
index 4ad3ef7df..69a40cf8e 100644
--- a/claims_hosp/delphi_claims_hosp/indicator.py
+++ b/claims_hosp/delphi_claims_hosp/indicator.py
@@ -12,15 +12,19 @@
 # third party
 import numpy as np
 import pandas as pd
+from delphi_utils import Smoother
 
 # first party
 from .config import Config
-from .smooth import left_gauss_linear
 
 
 class ClaimsHospIndicator:
     """Class to fit a hospitalizations indicator using CLI counts from claims-based data."""
 
+    smoother = Smoother("savgol",
+                        poly_fit_degree=1,
+                        gaussian_bandwidth=Config.SMOOTHER_BANDWIDTH)
+
     @staticmethod
     def gauss_smooth(num, den):
         """Smooth using the left_gauss_linear.
@@ -33,8 +37,8 @@ def gauss_smooth(num, den):
             tuple: (array of smoothed num, array of smoothed den)
 
         """
-        num_smooth = left_gauss_linear(num)
-        den_smooth = left_gauss_linear(den)
+        num_smooth = ClaimsHospIndicator.smoother.smooth(num)
+        den_smooth = ClaimsHospIndicator.smoother.smooth(den)
         den_smooth = np.clip(den_smooth, 0, None)
         num_smooth = np.clip(num_smooth, 0, den_smooth)
         return num_smooth, den_smooth
diff --git a/claims_hosp/delphi_claims_hosp/smooth.py b/claims_hosp/delphi_claims_hosp/smooth.py
deleted file mode 100644
index 56b132fa2..000000000
--- a/claims_hosp/delphi_claims_hosp/smooth.py
+++ /dev/null
@@ -1,40 +0,0 @@
-"""
-This file contains a left gauss filter used to smooth a 1-d signal.
-
-Code is courtesy of Addison Hu (minor adjustments by Maria).
-
-Author: Maria Jahja
-Created: 2020-04-16
-Modified: 2020-09-27
-    - partially concede few naming changes for pylint
-
-"""
-import numpy as np
-
-from .config import Config
-
-
-def left_gauss_linear(arr, bandwidth=Config.SMOOTHER_BANDWIDTH):
-    """
-    Smooth the y-values using a local linear left Gaussian filter.
-
-    Args:
-        arr: one dimensional signal to smooth.
-        bandwidth: smoothing bandwidth (in terms of variance)
-
-    Returns: a smoothed 1D signal.
-
-    """
-    n_rows = len(arr)
-    out_arr = np.zeros_like(arr)
-    X = np.vstack([np.ones(n_rows), np.arange(n_rows)]).T
-    for idx in range(n_rows):
-        weights = np.exp(-((np.arange(idx + 1) - idx) ** 2) / bandwidth)
-        XwX = np.dot(X[: (idx + 1), :].T * weights, X[: (idx + 1), :])
-        Xwy = np.dot(X[: (idx + 1), :].T * weights, arr[: (idx + 1)].reshape(-1, 1))
-        try:
-            beta = np.linalg.solve(XwX, Xwy)
-            out_arr[idx] = np.dot(X[: (idx + 1), :], beta)[-1]
-        except np.linalg.LinAlgError:
-            out_arr[idx] = np.nan
-    return out_arr
diff --git a/claims_hosp/tests/test_smooth.py b/claims_hosp/tests/test_smooth.py
deleted file mode 100644
index 76b71b057..000000000
--- a/claims_hosp/tests/test_smooth.py
+++ /dev/null
@@ -1,15 +0,0 @@
-# third party
-import numpy as np
-import pytest
-
-# first party
-from delphi_claims_hosp.smooth import left_gauss_linear
-
-
-class TestLeftGaussSmoother:
-    def test_gauss_linear(self):
-        signal = np.ones(10)
-        assert np.allclose(left_gauss_linear(signal)[1:], signal[1:])
-
-        signal = np.arange(1, 10) + np.random.normal(0, 1, 9)
-        assert np.allclose(left_gauss_linear(signal, 0.1)[1:], signal[1:])