BFD-4733: Prior Auth

apps/bfd-model-idr/.gitignore

@@ -21,6 +21,7 @@ out/*
 !out/ExplanationOfBenefit-Carrier.json
 !out/ExplanationOfBenefit-DME.json
 !out/ExplanationOfBenefit-Pharmacy.json
+!out/ExplanationOfBenefit-PriorAuth.json
 sushi/fsh-generated/*
 validator_cli.jar
 ReferenceTables/source-to-target-mappings

apps/bfd-model-idr/README.md

@@ -78,6 +78,20 @@ curl -X GET "http://localhost:8080/matchboxv3/actuator/health"
 
 As new versions of IGs are released, they may have multiple nested dependencies. This takes up a significant amount of memory if loaded directly into Matchbox. To eliminate heap errors while still being able to accurately validate profiles and terminology, we download the FHIR Packages locally, untar them, and upload relevant resources directly to Matchbox. The list of resources + packages are in matchbox_profiles.txt. To add a new IG reference, follow the syntax in that file. Packages are only uploaded using docker compose up (by calling setup_matchbox.py), so restart the composition if adding more dependencies. 
 
+### Generating Sample JSON from Synthetic CSVs
+
+To generate or update sample data files from the generated synthetic CSVs (located in `out/`), you can use the sample generation scripts. This only exists for prior auth data for now - it will be expanded to other use cases gradually.
+
+#### Prior Authorization Sample Generator
+
+To generate a prior authorization JSON sample from the synthetic CSVs based on tracking number.
+
+```sh
+python generate_prior_auth_sample.py --utn=<utn-here>
+```
+
+This will search for the specified UTN in `out/SYNTHETIC_PRAUC.csv`, collect the segments, and get it into the format that we map using FML. 
+
 ### Create FHIR files with synthetic data
 
 Requires Matchbox to be active.

apps/bfd-model-idr/augment_sample_resources.py

@@ -1,4 +1,5 @@
 import json
+import os
 import sys
 import yaml
 from dataclasses import asdict, dataclass, field
@@ -15,6 +16,31 @@
 cond_sk_info_file = "sample-data/CLM_RLT_COND_SGNTR_MBR_POC.csv"
 cond_sk_df = pd.read_csv(cond_sk_info_file, dtype={"CLM_RLT_COND_SGNTR_SK": str})
 
+synth_df = None
+synth_prvdr_file = "out/SYNTHETIC_PRVDR_HSTRY.csv"
+if os.path.exists(synth_prvdr_file):
+    try:
+        synth_df = pd.read_csv(synth_prvdr_file, dtype={"PRVDR_SK": str})
+    except Exception:
+        pass
+
+
+def lookup_provider_history(npi_num: str | int | None) -> dict | None:
+    if not npi_num:
+        return None
+    npi_str = str(npi_num).strip()
+    if not npi_str or npi_str == "None":
+        return None
+    matching_rows = df[df["PRVDR_SK"] == npi_str]
+    if not matching_rows.empty:
+        return json.loads(matching_rows.iloc[0].to_json())
+    if synth_df is not None:
+        matching_rows = synth_df[synth_df["PRVDR_SK"] == npi_str]
+        if not matching_rows.empty:
+            return json.loads(matching_rows.iloc[0].to_json())
+    return None
+
+
 cur_sample = sys.argv[1]
 cur_sample_data = {}
 with Path(cur_sample).open("r") as file:
@@ -42,8 +68,11 @@
 
 def load_profile_map():
     profile_map = {}
-    paths = [Path("dictionary-support-files/ExplanationOfBenefit.yaml"), 
-             Path("dictionary-support-files/ExplanationOfBenefit-Pharmacy.yaml")]
+    paths = [
+        Path("dictionary-support-files/ExplanationOfBenefit.yaml"),
+        Path("dictionary-support-files/ExplanationOfBenefit-Pharmacy.yaml"),
+        Path("dictionary-support-files/ExplanationOfBenefit-PriorAuth.yaml"),
+    ]
 
     for p in filter(Path.exists, paths):
         with p.open("r") as f:
@@ -257,6 +286,13 @@ class Provider:
 ]
 provider_list = []
 
+# we only use CLM_SRVC_PRVDR_GNRC_ID_NUM for part D events (we filter for PRVDR_SRVC_NPI_)
+if cur_sample_data.get("CLM_TYPE_CD") not in (1, 2, 3, 4):
+    billing_column = "PRVDR_BLG_PRVDR_NPI_NUM"
+else:
+    billing_column = "CLM_SRVC_PRVDR_GNRC_ID_NUM"
+
+
 # There may be an opportunity to consolidate even the duplicate NPIs into a
 # single careTeam reference, but we should wait to get feedback on this
 # The reason being: it's possible to lose context on rendering vs ordering
@@ -270,16 +306,18 @@ def create_billing_and_service_provider(billing_col_name):
     if qualifier in ("01", None):
         # Only pull NPI data if it's an NPI
         npi_num = cur_sample_data.get(billing_col_name)
-        prvdr_hstry_for_npi = json.loads(df[df["PRVDR_SK"] == str(npi_num)].iloc[0].to_json())
-        provider_object.NPI_TYPE = "2" if prvdr_hstry_for_npi.get("PRVDR_LGL_NAME") else "1"
-        provider_object.PRVDR_SK = npi_num
-        provider_object.PRVDR_LAST_OR_LGL_NAME = (
-            prvdr_hstry_for_npi["PRVDR_LGL_NAME"]
-            if provider_object.NPI_TYPE == "2"
-            else prvdr_hstry_for_npi["PRVDR_LAST_NAME"]
-        )
-        if prvdr_hstry_for_npi.get("PRVDR_1ST_NAME"):
-            provider_object.PRVDR_1ST_NAME = prvdr_hstry_for_npi.get("PRVDR_1ST_NAME")
+        prvdr_hstry_for_npi = lookup_provider_history(npi_num)
+        if prvdr_hstry_for_npi:
+            provider_object.NPI_TYPE = "2" if prvdr_hstry_for_npi.get("PRVDR_LGL_NAME") else "1"
+            provider_object.PRVDR_SK = npi_num
+            provider_object.PRVDR_LAST_OR_LGL_NAME = (
+                prvdr_hstry_for_npi["PRVDR_LGL_NAME"]
+                if provider_object.NPI_TYPE == "2"
+                else prvdr_hstry_for_npi["PRVDR_LAST_NAME"]
+            )
+            if prvdr_hstry_for_npi.get("PRVDR_1ST_NAME"):
+                provider_object.PRVDR_1ST_NAME = prvdr_hstry_for_npi.get("PRVDR_1ST_NAME")
+
         if cur_sample_data.get("CLM_BLG_PRVDR_OSCAR_NUM"):
             provider_object.PRVDR_OSCAR_NUM = cur_sample_data.get("CLM_BLG_PRVDR_OSCAR_NUM")
         if cur_sample_data.get("CLM_BLG_PRVDR_TAX_NUM"):
@@ -308,17 +346,20 @@ def create_careteam_provider(careteam_column):
     if qualifier in ("01", None):
         # Only pull NPI data if it's an NPI
         npi_num = cur_sample_data.get(careteam_column)
-        prvdr_hstry_for_npi = json.loads(df[df["PRVDR_SK"] == str(npi_num)].iloc[0].to_json())
-        provider_object.NPI_TYPE = "2" if prvdr_hstry_for_npi.get("PRVDR_LGL_NAME") else "1"
-        provider_object.PRVDR_SK = npi_num
-        # set a default name using PRVDR_HSTRY if not available.
-        provider_object.PRVDR_CARETEAM_NAME = (
-            prvdr_hstry_for_npi["PRVDR_LGL_NAME"]
-            if provider_object.NPI_TYPE == "2"
-            else prvdr_hstry_for_npi["PRVDR_LAST_NAME"]
-            + ", "
-            + prvdr_hstry_for_npi["PRVDR_1ST_NAME"]
-        )
+        prvdr_hstry_for_npi = lookup_provider_history(npi_num)
+        if prvdr_hstry_for_npi:
+            provider_object.NPI_TYPE = "2" if prvdr_hstry_for_npi.get("PRVDR_LGL_NAME") else "1"
+            provider_object.PRVDR_SK = npi_num
+            # set a default name using PRVDR_HSTRY if not available.
+            provider_object.PRVDR_CARETEAM_NAME = (
+                prvdr_hstry_for_npi["PRVDR_LGL_NAME"]
+                if provider_object.NPI_TYPE == "2"
+                else f"{prvdr_hstry_for_npi.get('PRVDR_LAST_NAME') or ''}, {prvdr_hstry_for_npi.get('PRVDR_1ST_NAME') or ''}".strip(", ")
+            )
+        else:
+            provider_object.NPI_TYPE = "1"
+            provider_object.PRVDR_SK = npi_num
+            provider_object.PRVDR_CARETEAM_NAME = "Practitioner"
 
         provider_object.careTeamType = careteam_header_columns.get(careteam_column)
         if qualifier:
@@ -351,17 +392,18 @@ def create_rendering_line_provider(npi_num):
     provider_object = Provider(
         PRVDR_SK=npi_num, careTeamType="rendering", PRVDR_CARETEAM_NAME="N/A"
     )
-    prvdr_hstry_for_npi = json.loads(df[df["PRVDR_SK"] == str(npi_num)].iloc[0].to_json())
-    provider_object.NPI_TYPE = "2" if prvdr_hstry_for_npi.get("PRVDR_LGL_NAME") else "1"
-    provider_object.PRVDR_CARETEAM_NAME = (
-        prvdr_hstry_for_npi["PRVDR_LGL_NAME"]
-        if provider_object.NPI_TYPE == "2"
-        else prvdr_hstry_for_npi["PRVDR_LAST_NAME"] + ", " + prvdr_hstry_for_npi["PRVDR_1ST_NAME"]
-    )
+    prvdr_hstry_for_npi = lookup_provider_history(npi_num)
+    if prvdr_hstry_for_npi:
+        provider_object.NPI_TYPE = "2" if prvdr_hstry_for_npi.get("PRVDR_LGL_NAME") else "1"
+        provider_object.PRVDR_CARETEAM_NAME = (
+            prvdr_hstry_for_npi["PRVDR_LGL_NAME"]
+            if provider_object.NPI_TYPE == "2"
+            else f"{prvdr_hstry_for_npi.get('PRVDR_LAST_NAME') or ''}, {prvdr_hstry_for_npi.get('PRVDR_1ST_NAME') or ''}".strip(", ")
+        )
     return provider_object
 
 # now we go through the line items!
-for line_item in cur_sample_data["lineItemComponents"]:
+for line_item in cur_sample_data.get("lineItemComponents", []):
     # we only care about PRVDR_RNDRNG_PRVDR_NPI_NUM
     cur_rendering_providers = [
         x.PRVDR_SK for x in provider_list if getattr(x, "careTeamType", None) == "rendering"

apps/bfd-model-idr/claims_generator.py

@@ -13,6 +13,7 @@
 from claims_adj import AdjudicatedGeneratorUtil
 from claims_other import OtherGeneratorUtil
 from claims_pac import PacGeneratorUtil
+from claims_priorauth import PriorAuthGeneratorUtil
 from claims_static import INSTITUTIONAL_CLAIM_TYPES, PHARMACY_CLM_TYPE_CDS, PROFESSIONAL_CLAIM_TYPES
 from claims_util import four_part_key, match_line_num
 from generator_util import (
@@ -35,6 +36,7 @@
     CLM_VAL,
     CNTRCT_PBP_NUM,
     PRVDR_HSTRY,
+    PRAUC,
     GeneratorUtil,
     RowAdapter,
     adapters_to_dicts,
@@ -566,6 +568,89 @@ class _ClaimsFile(StrEnum):
             f.META_LST_UPDT_SK,
         ],
     )
+    PRAUC = (
+        PRAUC,
+        [
+            f.MBI_NUM,
+            f.CAN,
+            f.EQUAT_BIC,
+            f.UTN,
+            f.SEGMENT_COUNT,
+            f.CURRENT_SEGMENT,
+            f.UTN_VALID_ST_DT,
+            f.UTN_VALID_EN_DT,
+            f.PA_IND,
+            f.CLM_TYPE,
+            f.HCPCS_OR_CPT_OR_HIPPS,
+            f.ICD_PROC_IND,
+            f.ICD_PROC_CODE,
+            f.MAC_ID,
+            f.PA_FILLER,
+            f.ICN_DCN,
+            f.PA_DT_ADDED,
+            f.PA_DT_UPDATED,
+            f.SERVICE_CNTS,
+            f.PA_DECISION,
+            f.PA_REQ_SUB_DT,
+            f.PA_REQ_REC_DT,
+            f.PA_DECISION_DT,
+            f.PA_DECISION_EXP_DT,
+            f.ORDER_REFER_NPI,
+            f.RENDER_NPI,
+            f.OPERATE_NPI,
+            f.SVC_RENDER_ST,
+            f.PRICE_MOD1,
+            f.PRICE_MOD2,
+            f.PLACE_OF_SERV,
+            f.ICD_DIAG_IND_1,
+            f.ICD_DIAG_CODE_1,
+            f.ICD_DIAG_IND_2,
+            f.ICD_DIAG_CODE_2,
+            f.ICD_DIAG_IND_3,
+            f.ICD_DIAG_CODE_3,
+            f.ICD_DIAG_IND_4,
+            f.ICD_DIAG_CODE_4,
+            f.ICD_DIAG_IND_5,
+            f.ICD_DIAG_CODE_5,
+            f.NPI,
+            f.NAME,
+            f.CMS_CERT,
+            f.REV_CODE_1,
+            f.REV_CODE_2,
+            f.REV_CODE_3,
+            f.REV_CODE_4,
+            f.REV_CODE_5,
+            f.REV_CODE_6,
+            f.REV_CODE_7,
+            f.REV_CODE_8,
+            f.REV_CODE_9,
+            f.REV_CODE_10,
+            f.REV_CODE_11,
+            f.REV_CODE_12,
+            f.REV_CODE_13,
+            f.REV_CODE_14,
+            f.REV_CODE_15,
+            f.REV_CODE_16,
+            f.REV_CODE_17,
+            f.REV_CODE_18,
+            f.REV_CODE_19,
+            f.REV_CODE_20,
+            f.COND_CODE_1,
+            f.COND_CODE_2,
+            f.COND_CODE_3,
+            f.COND_CODE_4,
+            f.OCCUR_CODE_1,
+            f.OCCUR_CODE_2,
+            f.OCCUR_CODE_3,
+            f.OCCUR_CODE_4,
+            f.TOB,
+            f.MR_COUNT_IND,
+            f.MR_COUNT_ST_DT,
+            f.MR_COUNT_END_DT,
+            f.ATT_PHY_NPI,
+            f.RRB_EXCL_IND,
+        ],
+    )
 
     def __init__(
         self,
@@ -725,6 +810,7 @@ def generate(
         CLM_RLT_COND_SGNTR_MBR: [],
         PRVDR_HSTRY: [],
         CNTRCT_PBP_NUM: [],
+        PRAUC: [],
     }
     load_file_dict(files=files, paths=list(paths))
     gen_utils.cntrct_pbp_num = [row.kv for row in files[CNTRCT_PBP_NUM]]
@@ -1084,10 +1170,14 @@ def generate(
                     clm_lines=clm_line_dcmtns,
                     clm_line_num=clm_line_num,
                 )
-                tracking_num = clm_line.get(f.CLM_LINE_PA_UNIQ_TRKNG_NUM) or (
-                    init_clm_line_dcmtn[f.CLM_LINE_PA_UNIQ_TRKNG_NUM]
-                    if init_clm_line_dcmtn
-                    else None
+                tracking_num = (
+                    clm_line.get(f.CLM_LINE_PA_UNIQ_TRKNG_NUM)
+                    or clm_line.get(f.CLM_LINE_PMD_UNIQ_TRKNG_NUM)
+                    or (
+                        init_clm_line_dcmtn[f.CLM_LINE_PA_UNIQ_TRKNG_NUM]
+                        if init_clm_line_dcmtn
+                        else None
+                    )
                 )
                 if tracking_num:
                     out_tables[CLM_LINE_DCMTN].append(
@@ -1167,6 +1257,19 @@ def generate(
 
     print("Done generating synthetic claims data for provided BENE_SKs")
 
+    # Generate synthetic prior authorization data (SYNTHETIC_PRAUC)
+    print("Generating synthetic prior authorization data...")
+    pa_util = PriorAuthGeneratorUtil()
+    prauc_rows = pa_util.gen_prior_auths(
+        gen_utils=gen_utils,
+        files=files,
+        out_tables=out_tables,
+        generated_type_1_npis=generated_type_1_npis,
+        generated_type_2_npis=generated_type_2_npis,
+    )
+    out_tables[PRAUC].extend(prauc_rows)
+    print(f"Done generating synthetic prior authorization data.")
+
     _save_claims_data({_ClaimsFile(k): v for k, v in out_tables.items() if k in _ClaimsFile})
 
 

apps/bfd-model-idr/claims_other.py

@@ -19,6 +19,7 @@
     CLM_ANSI_SGNTR,
     RowAdapter,
     gen_basic_id,
+    gen_npi_id,
 )
 
 _faker = Faker()
@@ -82,7 +83,7 @@ def gen_provider_history(
         generated_type_1_npis = set()
         generated_type_2_npis = set()
         for idx, provider_history in enumerate(all_provider_historys):
-            prvdr_sk = gen_basic_id(field="PRVDR_SK", length=9)
+            prvdr_sk = gen_npi_id(field="PRVDR_SK")
             # make half of providers type 1 npi and half type 2
             # type 1 npis never have a legal name
             # need to return both the subsets of type 1/2 npis that were used so that