12
Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US presented by: Xxxxxxx DSCI 5240

Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US

Embed Size (px)

Citation preview

Page 1: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US

Using SAS Predictive Modeling to Investigate the Asthma’s Patient

Future hospitalization Risk.Yehia H. Khalil, University of Louisville, Louisville, KY,US

presented by:

XxxxxxxDSCI 5240

Page 2: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US

Aim

• Develop a predictive model to forecast future Asthma hospitalization

Asthma

• A chronic inflammatory disorder of the airways

• 21 million Americans diagnosed

• Hospitalization rate growing (more than a million cases a year)

• Costs for Asthma: $14 billion

Page 3: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US

Predictive modeling

• Ability to incorporate any type of variable into analysis

• Dynamic; can easily accommodate any information to adjust model

SAS SEMMA methodology

• Sample

• Explore

• Modify

• Model

• Access

Page 4: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US

Source of 2009 Dataset

• Medical Expenditure Panel Survey

• California Health Interview Survey

Survey

• 47,614 adults

• 3,379 adolescents

• 8,945 children

Page 5: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US

Useful Parameters

• Demographics: age, race, marital status

• Health Behaviors: physical activities, fast food, alcohol consumption

• Health Conditions other than Asthma

• Health Insurance

• Poverty Level

• Emergency preparedness module: medication

• Mental or Emotional Condition

Page 6: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US
Page 7: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US
Page 8: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US

Fig. 4 Analysis Diagram

note:

• 40% training

• 30% testing

• 30% validation

Page 9: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US
Page 10: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US
Page 11: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US
Page 12: Using SAS Predictive Modeling to Investigate the Asthma’s Patient Future hospitalization Risk. Yehia H. Khalil, University of Louisville, Louisville, KY,US

Conclusion

• General health conditions, psychological distress and poverty level

affect future hospitalization risk

• Rx coverage and patient disability influence taking medication

regularly and can increase future hospitalization risk

• It is possible to enhance interventions, programs and alternatives to

avoid future hospitalizations