medicaid_utils.adapted_algorithms.py_nyu_billings package¶

Submodules¶

medicaid_utils.adapted_algorithms.py_nyu_billings.billings_ed module¶

Python implementation of The New York University (NYU) Emergency Department (ED) visit algorithm. This algorithm is the most widely used tool for retrospectively assessing the probability that ED visits are urgent, preventable, or optimally treated in an ED, using administrative data (Billings, Parikh, and Mijanovich 2000b; Feldman 2010).

class medicaid_utils.adapted_algorithms.py_nyu_billings.billings_ed.BillingsED[source]¶

Bases: object

This class packages functions to perform NYU/ Billings algorithm based classification of ED visits

cat = 'injury'¶

data_folder = '/home/runner/work/medicaid-utils/medicaid-utils/medicaid_utils/adapted_algorithms/py_nyu_billings/data'¶

dct_category_dx_codes = {'acs': ['033', '390', '391', '037', '045', '345', '481', '483', '485', '486', '462', '463', '465', '011', '012', '013', '014', '015', '016', '017', '018', '491', '492', '494', '496', '493', '681', '682', '683', '686', '590', '260', '261', '262', '521', '522', '523', '525', '528', '382', '428', '413', '614', '7803', '4721', '2501', '2502', '2503', '2508', '2509', '2500', '2512', '5589', '5990', '5999', '2765', '2680', '2681', '4660', '4822', '4823', '4829', '5184', '4010', '4019', '4111', '4118', '3200', '2801', '2808', '2809', '7834', '7070', '7071', '7078', '7079', '40201', '40211', '40291', '40210', '40290', '40200'], 'alcohol': ['291', '303', '3050', '3575', '4255', '5353', '5710', '5712', '5713', '7903', 'V704', 'V113', 'V791', '76071'], 'drug': ['292', '304', '3052', '3053', '3054', '3055', '3056', '3057', '3058', '3059', '3576', '6483', '6555', '7795', '76072', '76073', '76075'], 'injury': ['8', '9', 'E8', 'E90', 'E91', 'E92', 'E93', 'E94', 'E96', 'E97', 'E98'], 'psych': ['290', '293', '294', '295', '296', '297', '298', '299', '300', '301', '302', '306', '307', '308', '309', '310', '311', '312', '313', '314', '315', '316', '317', '318', '319', 'E95', '6484', 'V110', 'V111', 'V112', 'V114', 'V115', 'V116', 'V117', 'V118', 'V119', 'V710', 'V790']}¶

dct_recode_non_startswith = {'0049': '0059', '0085': '0059', '0088': '0059', '0340': '462', '0380': '0381', '0389': '0381', '0542': '5280', '0579': '05700', '0740': '5280', '0792': '075', '0799': '075', '07999': '075', '1104': '1100', '1108': '1100', '1109': '1100', '1110': '1100', '1319': '13101', '25001': '25000', '25060': '25080', '25061': '25080', '25081': '25080', '25082': '25080', '25083': '25080', '2510': '2512', '2740': '2749', '2761': '2765', '2768': '2767', '28261': '28260', '28262': '28260', '2910': '29181', '29532': '29530', '29590': '29530', '2989': '29530', '30001': '30000', '3009': '30000', '30300': '30301', '30400': '30390', '30420': '30561', '30500': '30390', '30560': '30390', '3058': '30390', '30590': '30390', '3061': '30000', '30781': '7840', '3080': '3089', '3090': '311', '3469 ': '7840', '34690': '7840', '3559': '3549', '37200': '0779', '37203': '0779', '37214': '0779', '37230': '0779', '38100': '3814', '38101': '3829', '38200': '38110', '40291': '4019', '40391': '4019', '40493': '4019', '41001': '41091', '41011': '41091', '41021': '41091', '41041': '41091', '41071': '41091', '41090': '4111', '41400': '4149', '41401': '4149', '4260': '42613', '42732': '42731', '4279': '42789', '43311': '4359', '43401': '4359', '43411': '4359', '43491': '4359', '436': '4359', '4549': '4540', '4553': '4550', '4554': '4552', '4556': '4550', '4558': '4552', '460': '4660', '4610': '4619', '4620': '462', '463': '462', '4658': '4660', '4659': '4660', '4660': '4660', '4661': '46619', '46611': '46619', '4730': '4739', '4770': '4781', '4779': '4739', '4809': '485', '481': '485', '4821': '485', '4824': '485', '48283': '485', '48289': '485', '486': '485', '490': '4660', '5210': '5220', '5225': '5220', '5233': '5231', '5234': '5231', '5239': '5231', '5282': '5280', '53081': '5301', '53100': '5339', '53140': '5339', '53240': '5339', '53390': '5339', '53550': '53500', '5400': '541', '5401': '541', '5409': '541', '56081': '5609', '57400': '57420', '57410': '57420', '5750': '57420', '5770': '5772', '5771': '5772', '5889': '586', '5921': '5920', '5929': '5920', '5941': '5920', '5950': '5990', '5959': '5990', '59780': '0980', '6089': '6084', '6142': '6149', '6143': '6149', '6159': '6149', '61610': '6160', '6168': '6160', '6269': '6268', '6331': '6339', '6338': '6339', '63491': '63490', '63492': '63490', '64003': '64000', '64093': '64000', '64183': '64303', '64193': '64000', '64313': '64303', '64390': '64303', '64413': '64410', '6809': '6806', '68102': '68100', '68110': '68100', '68111': '68100', '6850': '6851', '6926': '6918', '6929': '6918', '7079': '7071', '70890': '7089', '71590': '71946', '71596': '71946', '71690': '71946', '71696': '71695', '71697': '71946', '71940': '71946', '71941': '71946', '71943': '71946', '71949': '71947', '7210': '7235', '7213': '7243', '72190': '7248', '72210': '7243', '7231': '7235', '7241': '7243', '7242': '7248', '7245': '7248', '72709': '72705', '73399': '7329', '7803': '78039', '78031': '78039', '78052': '78050', '7821': '6918', '78601': '30000', '78700': '7870', '78701': '7870', '78702': '7870', '78703': '7870', '7880': '5920', '78840': '7884', '78841': '7884', '78842': '7884', '78843': '7884', '7887': '0980', '7907': '0381', '8024': '8028', '80700': '80701', '80702': '80700', '80703': '80701', '81201': '81200', '81209': '81200', '81220': '81200', '81240': '81200', '81241': '81200', '81305': '81200', '81342': '81341', '81343': '81383', '81344': '81341', '81381': '81341', '81400': '81341', '81600': '81500', '81601': '81500', '81602': '81500', '82101': '82100', '82380': '82382', '82381': '82382', '8240': '8242', '8244': '8242', '8246': '8242', '8248': '8242', '8250': '82520', '82525': '82520', '8260': '82520', '83101': '83100', '83104': '83100', '83209': '83200', '8408': '8409', '8419': '8409', '84210': '84200', '84213': '84209', '8441': '8448', '8449': '8439', '84509': '84500', '84510': '84500', '8469': '8460', '8471': '8472', '8479': '8460', '8709': '8708', '87200': '87201', '87261': '38420', '8728': '87201', '8731': '8730', '87340': '87320', '87341': '87320', '87342': '87320', '87343': '87320', '87344': '87320', '87353': '87320', '87360': '87320', '87364': '87320', '8738': '8744', '8748': '8744', '8760': '8750', '8770': '8750', '88003': '88000', '88100': '88000', '88110': '88120', '8820': '88102', '8821': '8822', '8831': '8830', '8832': '8830', '8840': '88000', '8910': '8900', '8921': '8920', '8930': '8920', '9104': '9105', '9110': '9100', '9114': '9105', '9130': '9100', '9134': '9135', '9140': '9141', '9150': '9100', '9160': '9100', '9164': '9175', '9165': '9175', '9170': '9100', '9180': '9181', '9189': '9181', '9190': '9100', '9195': '9194', '9196': '9156', '9212': '920', '9213': '920', '9219': '920', '9222': '9221', '9223': '9221', '92232': '92231', '92310': '92300', '92311': '92300', '92320': '92300', '9233': '92300', '9238': '92300', '9239': '92300', '92401': '92400', '92411': '92400', '92420': '92400', '92421': '92400', '9243': '92400', '9245': '92400', '9330': '933', '9331': '933', '9351': '933', '938': '936', '9404': '9409', '94321': '94120', '94421': '94120', '94500': '94120', '94522': '94120', '94524': '94420', '94526': '94120', '9462': '94120', '9490': '94120', '9592': '9593', '9630': '9654', '9690': '9654', '9694': '9691', '9700': '9691', '99810': '9981', '99811': '9981', '99812': '9981', '99813': '9981', '9985': '99883', '99859': '99883', '99889': '99883', 'V589': 'V5889', 'V709': 'V708 ', 'v709': 'V708 '}¶

df_eddxs = prindx nonemerg emergpc emedpa emednpa 0 0030 0.000000 1.000000 0.0 0.000000 1 0059 0.371429 0.457143 0.0 0.171429 2 0090 1.000000 0.000000 0.0 0.000000 3 01190 0.000000 0.000000 1.0 0.000000 4 03400 0.000000 1.000000 0.0 0.000000 .. ... ... ... ... ... 654 V718 1.000000 0.000000 0.0 0.000000 655 V719 1.000000 0.000000 0.0 0.000000 656 V72 1.000000 0.000000 0.0 0.000000 657 V741 1.000000 0.000000 0.0 0.000000 658 V802 1.000000 0.000000 0.0 0.000000 [659 rows x 5 columns]¶

filename = 'billings_ed.py'¶

classmethod get_nyu_ed_categories(dx_code: str) → Tuple[int, int, int, int, int, int, float, float, float, float][source]¶

Returns probabilities for each of the NYU ED categories, based on the input diagnosis code

Parameters:

dx_code (str) – Diagnosis code

Returns:

unclassified (int) – The code did not meet any NYU/Billings category
injury (int) – Injury
drug (int) – Drug Related (excluding alcohol)
psych (int) – Mental Health Related
alcohol (int) – Alcohol Related
peds_acs_ed (int) – Pediatric ambuilatory care sensitive ED visit
ne (float) – Non-Emergent probability
epct (float) – Emergent, Primary Care Treatable probability
edcnpa (float) – Emergent, ED Care Needed, Preventable/Avoidable
edcnnpa (float) – Emergent, ED Care Needed, Not Preventable/Avoidable

Examples

>>> BillingsED.get_nyu_ed_categories('4659')

classmethod get_nyu_ed_proba_for_dx_code(dx_code: str) → dict[source]¶

Merge with EDDXs gets the probbolities attached with the code

ed_category	description
nonemerg	Non-Emergent
emergpc	Emergent, Primary Care Treatable
emedpa	Emergent, ED Care Needed, Preventable/Avoidable
emednpa	Emergent, ED Care Needed, Not Preventable/Avoidable

Parameters:: dx_code (str) – Diagnosis code
Returns:: Dictionary with probability keys (‘nonemerg’, ‘emergpc’, ‘emedpa’, ‘emednpa’) and float values, or empty dict if the code is not found.
Return type:: dict

Examples

>>> BillingsED.get_nyu_ed_proba_for_dx_code('4659')

classmethod get_special_categories(dx_code: str) → dict[source]¶

Categorizes the diagnosis code into below categories

category_code	description
injury	Injury
psych	Mental Health Related
alcohol	Alcohol Related
drug	Drug Related (excluding alcohol)
acs

Parameters:: dx_code (str) – Diagnosis code
Returns:: Dictionary with keys ‘acs’, ‘psych’, ‘drug’, ‘alcohol’, ‘injury’ and integer values (0 or 1).
Return type:: dict

Examples

>>> BillingsED.get_special_categories('E8600')
{'acs': 0, 'psych': 0, 'drug': 0, 'alcohol': 1, 'injury': 0}

classmethod is_peds_acsed(dx_code: str) → int[source]¶

Checks if the diagnosis code meets pediatric ACS ED visit criteria

Parameters:: dx_code (str) – Diagnosis code
Returns:: 1 if the code meets pediatric ACS ED criteria, 0 otherwise.
Return type:: int

Examples

>>> BillingsED.is_peds_acsed('493')
1
>>> BillingsED.is_peds_acsed('999')
0

logger_name = 'medicaid_utils.adapted_algorithms.py_nyu_billings.billings_ed'¶

lst_special_ed_categories = ['acs', 'psych', 'drug', 'alcohol', 'injury']¶

package_folder = '/home/runner/work/medicaid-utils/medicaid-utils/medicaid_utils/adapted_algorithms/py_nyu_billings'¶

pdf_nonstartswith_recode = origin target 10 78700 7870 10 78701 7870 10 78702 7870 10 78703 7870 11 78840 7884 .. ... ... 181 9694 9691 182 9985 99883 182 99859 99883 182 99889 99883 183 V589 V5889 [322 rows x 2 columns]¶

pdf_recode = origin target ... deleted starts_with 0 [042, 043, 044] 042 ... 0 1 1 [0781] 0781 ... 0 1 2 [2501] 2501 ... 0 1 3 [2780] 2780 ... 0 1 4 [4151] 4151 ... 0 1 .. ... ... ... ... ... 179 [9592] 9593 ... 0 0 180 [9630, 9690] 9654 ... 0 0 181 [9694] 9691 ... 0 0 182 [9985, 99859, 99889] 99883 ... 0 0 183 [V589] V5889 ... 0 0 [182 rows x 6 columns]¶

pdf_startswith_recode = origin target 0 042 042 0 043 042 0 044 042 1 0781 0781 2 2501 2501 3 2780 2780 4 4151 4151 5 493 493 6 5301 5301 7 5751 5751 8 6860 6860 9 7807 7807 12 7890 7890 13 850 850 14 8540 854 14 8541 854 15 9898 9898 16 9955 9955 17 9970 9970 19 v72 V72 19 V72 V72 62 682 682 62 684 682 62 6869 682¶

classmethod recode_diag_code(dx_code: str) → str[source]¶

Recodes diagnosis code

Parameters:: dx_code (str) – Diagnosis code
Returns:: Recoded diagnosis code.
Return type:: str

Examples

>>> BillingsED.recode_diag_code('4659')
'4659'

medicaid_utils.adapted_algorithms.py_nyu_billings.billings_ed.get_nyu_ed_proba(df: DataFrame, date_col: str, index_col: str, cms_format: str = 'MAX') → DataFrame[source]¶

This functions returns probabilities for dxgroup codes in the input dataframe being urgent, preventable, or optimally treated ED visits. This function recodes the passed dx codes to a set of dx codes used in (Billings, Parikh, and Mijanovich 2000b; Feldman 2010) study, adds columns (‘injury’, ‘psych’, ‘alcohol’ and ‘drug’), indicating the type of ED service.

Parameters:

df (dask.DataFrame) – dataframe with dx codes
date_col (str) – Date column name
index_col (str) – Index column name
cms_format ({'MAX', 'TAF'}) – CMS file format

Return type:

pd.DataFrame

Examples

>>> # Requires a dask DataFrame with ED claims and diagnosis columns
>>> pdf = get_nyu_ed_proba(df, 'srvc_bgn_date', 'MSIS_ID')