Prediction of Hospital readmission of Diabetic patients & Imporatance of Glucose monitering (HBA1c test) in case of diabetic patients

Deployment - Web app for prediction of early readmission

Business problem statement (GOALS)

Business Problem Understanding

Hospitals engaging in any model are likely to face penalties if their providers cannot improve hospital readmission rates. To combat growing costs, payers across the industry are adding hospital readmission quality measures to their value-based reimbursement programs. In recent years, government agencies and healthcare systems are increasingly focused on 30-day readmission rates as a way to improve quality and also determine the complexity of patient populations. To avoid value-based penalties (disincentives), Hospitals should reduce readmission rates by identifying the Diabetic patients who are having high probability of early re-admission compared to other patients who do not have Diabetes.

Business Objective

The main objective of our work is to come up with the predictive model which helps Hospital Management systems to predict the risk of early readmission of patients who are having Diabetes Mellitus which can further address to escape value based penalty and better patient care.

• The need for glucose level monitoring like HBA1c test to detect in early stages.
• Proper medication and care for diabetic patients
• Better patient engagement, Transitional care & close post-discharge follow-up with the intention of reducing early readmission

Scope of the project

As cost of inpatient care & readmission rates are higher in patients with diabetes Mellitus (DM) compared to other diagnosis. So, it makes perfect sense for the hospital management that, focusing on reducing cost of readmission in case of Diabetic patients to avoid value-based penalties from the government.

Limitations of project

• Our model is not capable of accounting the factors such as socio-economic characteristics of patient,
• With most efforts focused on reducing readmissions, there is a potential to overlook the stress and vulnerability of patients.

Problems faced

• Problem 1) Require domain expertise for Data understanding (highly domain specific i.e Health care)

remedy : consultation with the colleagues who are associated with health care domain

• Problem 2) Encoding data for Multi-variate imputation through chined equations (KNN as a base algorithm with neighbours = 5)

remedy : Nominal features being label encoded, Ordinal features being ordinally encoded & classifying a missing label considering only some features.

• Problem 3) Encountered a feature which is most important but had 53% missing values

remedy : considered them as 'missing category (Not mentioned)' in the analysis

• Problem 4) Data preparation

remedy : tried with label encoding and Dummy encoding, Dummy encoding yielded good results, implemented a pipeline & column transformer class

Data dictionary

Attribute	type	Information
Encounter ID	Numeric	Unique identifier of an encounter
Patient number	Numeric	Unique identifier of a patient
Race	Nominal	Values: Caucasian, Asian, African American, Hispanic, and other
Gender	Nominal	Values: male, female, and unknown/invalid
Age	Nominal	Grouped in 10-year intervals: (0, 10), (10, 20), ..., (90, 100)
Weight	Numeric	Weight in pounds.
Admission type	Nominal	'Integer identifier corresponding to 9 distinct values,
		for example, emergency,urgent, elective, newborn, and not available.
Discharge disposition	Nominal	'Integer identifier corresponding to 29 distinct values,
		for example, discharged to home, expired, and not available.
Admission source	Nominal	'Integer identifier corresponding to 21 distinct values,
		for example, physician referral, emergency room, and transfer from a hospital.
Time in hospital	Numeric	'Integer number of days between admission and discharge.
Payer code	Nominal	'Integer identifier corresponding to 17 distinct values, for example,
		Blue Cross\Blue, Shield, Medicare, and self-pay.
Medical specialty	Nominal	Integer identifier of a specialty of the admitting physician, corresponding to 84 distinct values,
		for example, cardiology, internal medicine, family\general practice, and surgeon.
Number of lab procedures	Numeric	Number of lab tests performed during the encounter.
Number of procedures	Numeric	Number of procedures (other than lab tests) performed during the encounter.
Number of medications	Numeric	Number of distinct generic names administered during the encounter.
Number of outpatient visits	Numeric	'Number of outpatient visits of the patient in the year preceding the encounter.
Number of emergency visits	Numeric	'Number of emergency visits of the patient in the year preceding the encounter.
Number of inpatient visits	Numeric	Number of inpatient visits of the patient in the year preceding the encounter.
Diagnosis 1	Nominal	The primary diagnosis (coded as first three digits of ICD9); 848 distinct values.
Diagnosis 2	Nominal	Secondary diagnosis (coded as first three digits of ICD9); 923 distinct values.
Diagnosis 3	Nominal	Additional secondary diagnosis (coded as first three digits of ICD9); 954 distinct values.
Number of diagnoses	Numeric	Number of diagnoses entered to the system.
Glucose serum test result	Nominal	Indicates the range of the result or if the test was not taken.
		Values: “>200,” “>300,”“normal,” and “none” if not measured.
A1c test result	Nominal	Indicates the range of the result or if the test was not taken.
		Values: “>8” if the result was greater than 8%
		“>7” if the result was greater than 7% but less than 8%
		“normal” if the result was less than 7%, and
		“none” if not measured.
Change of medications	Nominal	'Indicates if there was a change in diabetic medications (either dosage or generic name).
		Values: “change” and “no change”.
23 features for medications	Nominal	'For the generic names: metformin, repaglinide, nateglinide, chlorpropamide,glimepiride, acetohexamide,
		glipizide, glyburide, tolbutamide, pioglitazone, rosiglitazone, acarbose, miglitol,
		troglitazone, tolazamide, examide, sitagliptin, insulin, glyburide-metformin, glipizide-metformin,
		glimepiride-pioglitazone, metformin-rosiglitazone, and metformin-pioglitazone,
		the feature indicates whether the drug was prescribed or there was a change in the dosage.
		Values: “up” if the dosage was increased during the encounter,
		“down” if the dosage was decreased,
		“steady” if the dosage did not change,
		and “no” if the drug was not prescribed.
Diabetes medications	Nominal	'Indicates if there was any diabetic medication prescribed. Values: “yes” and “no”.
Readmitted	Nominal	Days to inpatient readmission.
		Values: “<30” if the patient was readmitted in less than 30 days,
		“>30” if the patient was readmitted in more than 30 days,
		and “No” for no record of readmission.

Missing values

Attribute	Data_type	Total_Entries	Null_values	Percent_Null_values	Unique_values
weight	object	101766	98569	96.86	9
medical_specialty	object	101766	49949	49.08	72
payer_code	object	101766	40256	39.56	17
race	object	101766	2273	2.23	5
Diagnosis1	object	101766	21	0.02	716
Diagnosis2	object	101766	358	0.35	748
Diagnosis3	object	101766	1423	1.40	789

Missing value Imputation

• Race (127 values imputed Using information from patient_nbr attribute i.e Taking information from Some patients who have visited multiple times & not filled certain columns)

• Rest missing values were imputed using Multivariate_imputation_through_chianed_equations (K-Nearest-Neighbour as algorithm)

  • (Race,Diagnosis_1,Diagnosis_2,Diagnosis_3)