Method and Apparatus for Estimating Patient Populations

• US 20110184761A1
• Filed: 01/25/2011
• Published: 07/28/2011
• Est. Priority Date: 01/25/2010
• Status: Abandoned Application

• Current Assignee: Cerner Innovation, Inc. (Oracle Corporation)
• Original Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Inventors: Farooq, Faisal, Rahmanian, Farbod, Krishnapuram, Balaji, Yu, Shipeng
• Application Number:
  US13/013,314
• Publication Number:
  US 20110184761A1
• Time in Patent Office: Days
• US Class Current: 705/3
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

The methods and apparatuses of the present invention provide for a continuous abstraction of randomly sampled patient data and shortened data processing cycle times when an accurate sample population size is unknown at the beginning of the sampling process. The present invention estimates an initial medical patient population size for the purpose of randomly sampling that population. The estimated population size is calculated based on historical patient population data and is corrected at the end of the sample time period. Under-sampling is remediated at the end of the sample time period, during which continuous sampling of the patient data is carried out to provide interim and immediately available sampled patient data. Criteria for medical patient population sizing and sampling are provided by health care organizations responsible for administrating health care quality improvement standards.

Claims

1. A method of predictive sampling of a medical patient population comprising:

estimating an initial patient population (M) for a sample time period, said initial patient population based on historical sample populations for said sample time period;

calculating an initial sample population (m) for said sample time period, said initial sample population based on said estimated initial patient population and minimum sample size tables;

sampling said initial patient population (M) randomly until m population members are sampled;

recalculating a minimal sample population (n) for said sample time period, said minimal sample population (n) based on an actual patient population (N) for said time period and said minimum sample size tables; and

correcting for an under-sampling of said actual patient population (N).

2. The method of claim 1 wherein said step of sampling further comprises:

randomly determining a sample staring point between a first patient within said initial patient population and a k^th patient within said initial patient population, wherein k=M/m; and

sampling said initial sample patient population every k^th element until m population members are sampled.

3. The method of claim 1 wherein said step of correcting further comprises:

determining that said under-sampling results from said actual patient population (N) being greater than said initial patient population (M) for said sample time period; and

resampling a reconstituted patient population (N′

) upon said determination, said reconstituted patient population being said actual patient population (N) without said previously sampled population members.

4. The method of claim 3 wherein said step of resampling comprises:

recalculating an additional sample population (m′

), said additional sample population based on a difference between said minimal sample population (n) and said initial sample population (m); and

resampling said reconstituted patient population (N′

) randomly for said additional sample population (m′

).

5. The method of claim 1 further comprising storing a plurality of patient records in a patient database, said plurality of patient records including historical medical data related to said medical patient population, said historical sample populations being derived from said historical medical data.

6. The method of claim 5 further comprising submitting said patient records from said initial sample population to a health care quality standards provider.

7. The method of claim 1 wherein said medical patient population is sampled for one of:

quality of care analysis, inclusion in a clinical trial, or inclusion in a meaningful use initiative.

8. The method of claim 7 wherein said medical patient population is sampled for quality of care analysis and said minimal sample size tables pertain to a core measure, said quality of care analysis being conducted by a health care quality standards provider.

9. The method of claim 8 wherein said core measure is one of:

heart failure, acute myocardial infarction, pneumonia, surgical care improvement, stroke, or venous thromboembolism.

10. The method of claim 8 wherein said health care quality standards provider is the Center for Medicare and Medicaid Services.

11. The method of claim 1 wherein said step of sampling occurs continuously and in real-time as the patient population is admitted to a health care provider, said actual patient population being unknown until the end of said sample time period.

12. The method of claim 1 wherein said minimum sample size data is a percentage of said patient populations.

13. The method of claim 1 wherein said minimum sample size data is a fixed number of patients.

14. A computer-based predictive sampling system, said sampling system coupled to a sample population database having historical population data for medical patients, said sampling subsystem also coupled to a standard-based sampling requirements database having minimum sample size tables, said sampling system comprising:

an estimation subsystem for estimating an initial patient population (M) and collecting an actual patient population (N), said estimation system also estimating an initial sample population (m) based on said minimum sample size tables, said initial patient population (M) being based on said historical population data, said initial patient population (M) and said actual patient population (N) being determined before and after said sample time period respectively;

a verification subsystem for calculating a minimal sample population (n) for said sample time period based on said actual patient population (N) and said minimum sample size tables, said verification subsystem further calculating a reconstituted patient population (N′

) upon a determination of undersampling during said sample time period; and

a sampling subsystem for sampling randomly said initial patient population (M) and said reconstituted patient population (N′

), said random sampling performed on said initial patient population (M) for m population members, said reconstituted patient population being determined by said verification subsystem to be said actual patient population (N) without said previously sampled population members, said random sampling additionally performed on said reconstituted patient population (N′

) based on a difference between said minimal sample population (n) and said “

m”

sampled population members.

15. The computer-based system of claim 14 wherein said predictive sampling system is part of a computerized medical quality measures system, said minimum sample size tables are provided by a health care quality standards provider, and patient data related to said sampled patients is transmitted to said health care quality standards provider for evaluation.

16. A method of sampling a patient population for quality of care analysis, said quality of care analysis being conducted by a health care quality standards provider, the method comprising:

storing a plurality of patient records in a patient database, said plurality of patient records including historical data related to a target patient population, said target patient population having a common core measure;

estimating an initial patient population (M) for a sample time period, said initial patient population being based on said historical data;

calculating an initial sample patient population (m) for said sample time period;

said initial sample patient population based on said estimated initial patient population and minimum sample size data provided by said health care quality standards provider;

sampling said initial patient population (M) randomly until m patient population members are sampled;

submitting after each sampling said patient records for said initial sample patient population to said health care quality standards provider;

determining an actual patient population (N) for said time period;

recalculating a minimal sample patient population (n) for said sample time period based on said actual patient population (N) and said minimum sample size data provided by said health care quality standards provider;

recalculating an additional sample patient population (m′

), said additional sample patient population based on a difference between said minimal sample patient population (n) and said initial sample patient population (m);

resampling randomly a reconstituted patient population (N′

) until m′

patient population members are sampled, said reconstituted patient population being said actual patient population (N) without the members previously sampled; and

submitting said patient records for said additional sample patient population to said health care quality standards provider.

17. The method of claim 16 wherein said first sampling step includes:

randomly determining a sample staring point between a first patient within said initial patient population and a k^th patient within said initial patient population, wherein k=M/m; and

sampling said initial sample patient population every k^th element until m patients are sampled.

18. A method of predictive sampling of a medical patient population comprising:

real time and continuous sampling of the medical patient population when the final medical patient population size is unknown.