Risk adjustment tools for analyzing patient electronic discharge records

Abstract

A system and method are disclosed for examining and effectively managing resource allocation in a health care organization or facility (e.g., a hospital, a hospice, or a nursing home). The disclosed technology relies upon an analysis of the electronic discharge records of a health care organization in a manner that allows extraction of only those records that were generated for patients having a specified condition (e.g., septic shock, coronary artery disease, auto-immune disease, etc.) or fall into a particular class based upon resource usage (e.g., length of hospital stay, type of surgery, or quantity and type of pharmaceuticals taken). Note that discharge records often fail to explicitly specify the condition of interest. To accomplish selective extraction, the content of the discharge records is matched against one or more "key explanatory variables" such as a "selection vector" which is collection of patient codes that implicitly specify the condition of interest.

Claims

What is claimed is:

1. In a computer system, a method of classifying patients according to an underlying condition, the method comprising:

providing a selection vector of patient condition codes which specify a selected medical condition without explicitly reciting that selected medical condition;

providing a laboratory cost driver which specifies a laboratory test which may be performed during the course of a patient's treatment;

analyzing a plurality of electronic discharge records of one or more health care organizations to determine which of those electronic discharge records recite patient data which match the (i) patient condition codes of the selection vector and (ii) the laboratory test of the laboratory cost driver; and

selecting those electronic discharge records matching at least one of the selection vector and the laboratory cost driver, wherein the selected electronic discharge records represent a class of patients having the selected medical condition, wherein the selection vector has a sensitivity of at least about 80 percent for the selected medical condition and a specificity of at least about 70 percent for the selected medical condition.

2. The method of claim 1, wherein the selected class of patients has an expected deviation from a normal amount of reimbursement.

3. The method of claim 1, wherein the patient condition codes are provided as at least one of ICCS codes, DRG codes, CPT-4 codes, ICD-8 codes, HICS codes, ICD-9 codes, and ICD-10 codes.

4. The method of claim 1, wherein the selection vector has an associated cost parameter which may be used to determine an expected cost for treating a patient having an electronic discharge record matching the selection vector.

5. The method of claim 1, wherein the selected medical condition is septic shock.

6. The method of claim 1, wherein the laboratory cost driver includes an associated cost parameter which may be used to determine an expected cost for treating a patient having an electronic discharge record matching the laboratory cost driver.

7. The method of claim 6, wherein the laboratory cost driver is chosen such that its effect on the cost of treatment is at least about three times the cost of the associated laboratory test.

8. The method of claim 1, further comprising calculating an amount of resources expected to be consumed based upon an expression including as variables a plurality of laboratory cost drivers.

9. The method of claim 8, wherein the variables of the expression further comprise the selection vector.

10. The method of claim 8, wherein the expression is a linear expression.

11. The method of claim 8, wherein the expression is used in risk adjustment to classify patients based upon the amount of resources expected to be consumed.

12. The method of claim 9, wherein the expression is used to confirm that medical treatment guidelines are being followed by a health care organization.

13. A method implemented on one or more computer systems for classifying patients according to risk, the method comprising:

determining whether patient information provided as one or more electronic records matches a selection vector specifying a collection of patient conditions;

determining whether said patient has had one or more tests from a prescribed group of tests performed, and

classifying said patient in a specified risk group based upon whether that patient's electronic discharge records match said selection vector and whether that patient's electronic discharge records for tests performed indicates that the patient has had specific ones of said defined tests performed, wherein the selection vector has a sensitivity of at least about 80 percent for a specified medical condition and a specificity of at least about 70 percent for the specified medical condition.

14. The method of claim 13, wherein the specified risk group correlates with an expected cost for treating said patient, which expected cost is derived from a statistical analysis of a group of patients.

15. The method of claim 13, wherein said selection vector contains patient condition codes indicative of septic shock.

16. The method of claim 15, wherein the patient condition codes include at least (a) one or more codes specifying one or more organ systems which have failed together with (b) one or more codes specifying an infection.

17. The method of claim 16, wherein the patient condition codes set forth in the selection vector are provided in the format of at least one of ICCS codes, DRG codes, CPT-4 codes, ICD-8 codes, ICD-9 codes, and ICD-10 codes.

18. The method of claim 13, wherein the defined tests include tests indicative of sepsis.

19. The method of claim 18, wherein the defined tests include at least one test that is selected from the group consisting of bacterial tests, acids/base tests, coagulation tests, and fluid sample tests.

20. The method of claim 19, wherein the defined tests include at least one of the following tests: aerobic bacterial culture, anaerobic bacterial culture, and complete blood count.

Description

COPYRIGHT NOTICE

A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the xerographic reproduction by anyone of the patent document or the patent disclosure in exactly the form it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.

BACKGROUND OF THE INVENTION

Generally, the present invention relates to computerized systems and methods for analyzing hospital electronic discharge data to accurately predict the resources required to treat a patient. Further, the present invention relates to computerized systems and methods applied to such electronic discharge data to allow statistical patient classification based upon certain key explanatory variables of a patient's condition.

In the age of managed health care and government medical care programs (e.g., Medicare in the US), it has become increasingly important for hospitals or other patient care institutions to accurately monitor their costs and justify their treatment procedures. Hospitals contract with HMO's (health maintenance organizations) and other managed care providers to provide services for the patients enrolled by the HMO. HMOs contract with employers to enroll employees in their health care plans.

If two hospitals are vying for a contract from the same HMO, the hospital having a lower cost per patient should normally win such contract--all other features being equal. However, the cost differential may not necessarily be due to inefficiencies in the higher cost hospital. It may simply be that the higher cost hospital is treating, on average, sicker patients or patients requiring more care. If the managed health care provider can be made to understand this, it may actually decide to contract with the higher cost hospital. Similarly, when a managed health care provider is contracting with an employer, the health care organization may wish to charge some employers more than others to enroll their employees. This can be justified if the employer has a class of employees that are, on average, likely to be sicker or require more care than the employees of some other employer. Obviously, the health care provider must convince the employer that it is charging a higher premium for that employer's employees because they are actually more of a high cost risk. Unfortunately, no precise mechanism exists for this purpose--although a measure known as a "DRG" has been applied.

In the mid-1980's, the United States began, through it's Medicare program, to reimburse hospitals and other health care institutions a fixed dollar amount based upon a "diagnosis related group" (DRG) determined from certain medical conditions recorded by the hospital or organization attending to the patient. This practice has spread beyond the government and throughout the industry to HMOs, PPOs, etc. Each DRG classification is comprised of one or more codes (e.g., "ICD-9 codes" currently used in the US as specified in the International Classification of Disease, 9th Revision, Clinical Modification) each of which represents a specific medical condition. For example, ICD-9 code 42.0 represents a patient that is HIV positive, ICD-9 code 789.59 represents a patient which has septic shock, ICD-9 code 410.9 represents a patient who had a myocardial infarction. All told, there are several thousand such ICD-9 codes.

Obviously, a hospital must properly identify the ICD-9 codes for each of its patients. Hospitals now have internal staff groups (medical records librarians) whose sole function is to identify proper ICD-9 codes and input them into patient records. To do this, such hospital financial staff groups take the clinical records provided by the physicians and nurses attending to a patient and attempt to determine which ICD-9 codes fit with the clinical records available to them. The hospital financial group then inputs a collection of ICD-9 codes in a standard electronic form (a "UB-92" form in the US). These and similar forms will often be referred to herein generically as hospital or health care "electronic discharge records."

The codes are entered in a specified order (e.g., the first ICD-9 code represents the principal condition for which the patient was treated, and subsequent ICD-9 codes may represent the conditions present at the hospital admission and other conditions that develop during the patient's stay). In the US, the codes are entered pursuant to a protocol specified by The Health Care Financing Agency ("HCFA"). The electronically formatted ICD-9 codes (together with other demographic information) for each patient are then used to classify patients into particular DRGs. A specific protocol governs DRG classification; for example, the principal diagnostic code, the presence of comorbidity conditions, the patient's age, the use of surgical procedures, and death or survival together dictate the DRG classification.

Unfortunately, there is considerable variance in costs for treating all those patients falling within a specific DRG. This is because DRGs do not do a particularly good job of classifying patients with sufficient specificity. For example, a DRG associated with a particular condition may provide reimbursement at a rate of $100,000 per patient. However, within the class of patients meeting this DRG, some may only cost the hospital $50,000 while others may cost the hospital $500,000 or more. Obviously, if one hospital has more than its share of patients falling within the DRG category for the particular condition and yet costing well in excess of the approximately $25,000 average charge, the hospital would like to be able to explain its additional costs.

The problem is exacerbated because the medical records librarians responsible for entering the ICD-9 codes and other information frequently fail to enter those codes that directly specify patient conditions. For example, ICD-9 code 038.9 specifies that a patient has sepsis. However, analysis of a large sample of UB-92 records has shown that only 27 percent of all sepsis patients actually have had code 038.9 recorded. Obviously, it is therefore not possible to identify all patients having sepsis (or most any other serious condition for that matter) by simply analyzing a hospital's electronic discharge records (e.g., the UB-92 records). This problem arises in part because of the large number of codes and the associated complexity of the ICD-9 code system. It also arises in part because the poor linking between a hospital's clinical production system (discharge summaries, doctors and nurses notes, etc.) and its financial billing system (ICD-9 coded records).

To address these and other issues, some "risk adjustment tools" have been developed. In some cases, these attempt to explain the statistical variance in cost of treating the patients falling under a particular DRG (or other generic patient classification). Early risk adjustment tools were simple models used by insurance companies for health insurance underwriting and pricing individual premiums. More sophisticated modern tools were developed by deploying powerful statistical software to analyze patient databases and identify patterns in patient records. These tools were then provided as software to gauge risk within certain patient populations.

All risk adjustment tools model some combination of demographic and/or health status data. See "Risk Adjustment" by C. Lee and D. Rogal, produced for The Robert Wood Johnson Foundation, March 1997. Demographic models generally include some or all of the following variables: age, sex, family status, geographic location, and welfare status. Measures of health status can include survey data of health, diagnoses, and data reflecting prior utilization. Health status models may include demographic variables as predictors of health.

Some more recent risk adjustment tools employ ICD-9 codes to classify patients. See the above-referenced "Risk Adjustment" paper, pages 14-15. Unfortunately, the best of the existing tools typically can explain no more than about 45 percent of the variance in patient cost within a given DRG. This leaves over 50 percent of the variance unexplained. In addition, known risk adjustment tools were developed using fairly narrow databases, and therefore predict best for a specific population group or health care setting.

In view of the above, medical records analysis technology could be improved to provide a method for ascribing the variance within a generic class of patients (such as all patients grouped into the DRG for particular patient condition category).

SUMMARY OF THE INVENTION

This invention provides technology for identifying effective resource allocation in a health care organization or facility (e.g., a hospital, hospice, or nursing home). The technology is based upon a statistical analysis of health care data in a manner that allows extraction of "key explanatory variables" such as "selection vectors" that select the electronic discharge records of only those patients that have a specified condition (e.g., septic shock, coronary artery disease, auto-immune diseases, red blood cell disorders, leukemias, chirrosis, chest pain, adult diabetes, etc.) or fall into a particular class based upon resource usage (e.g., length of hospital stay, type of surgery, or quantity and type of pharmaceuticals taken).

Various types of "key explanatory variables" may be employed to analyze the electronic "discharge records" of health care organizations. The key explanatory variables correspond to one or more pieces of information, contained in electronic discharge records such as UB-92 records described above, which record some specific event or condition associated with the patient. Common to all key explanatory variables is a disproportionately large impact on resource usage or a strong indication of a resource intensive medical condition. Further all key explanatory variables are premised upon the identification of a particular condition that may not be immediately apparent from inspection of any given electronic discharge record or other electronic hospital record. For example, very often the medical condition sepsis or septic shock is not specifically identified in the electronic discharge records of a given patient. Certain key explanatory variables are identified based upon a statistically significant correlation between sepsis (preferably high resource usage cases of sepsis) and specific information contained in patient electronic discharge records.

One example of a key explanatory variable is a "selection vector" listing one or more "patient condition codes" (e.g., ICD-9 codes) which together implicitly specify a condition such as septic shock. An example of a selection vector for the medical condition septic shock may specify patient condition codes for (1) failure of a major system such as the renal system or the cardiovascular system and (2) infection of an organ or organ system outside the system which has failed.

Another key explanatory variable is a "laboratory cost driver" corresponding to a laboratory test which when performed on the patient indicates a very large (disproportionate) effect on resource usage. Information as to whether or not such a laboratory test was conducted (and how many such tests were conducted) typically resides in certain electronic records for patients of a health care organization. Note that the UB-92 discharge record format described above typically does not specify laboratory test information in detail, so other types records may have to be evaluated to identify laboratory cost drivers.

In one example, the use of an anaerobic bacterial culture test (sometimes indicating a suspicion of sepsis) costs less than approximately $100, but is shown by statistical analysis to be associated with an additional patient cost of about $1,000. Further, each additional anaerobic bacterial culture test adds about another $1,000 to the cost of treating the patient.

From the selection vectors, laboratory cost drivers, and other key explanatory variables associated with a particular condition, an expression can be developed for predicting the resource usage required for various patients. By applying this expression to the electronic discharge records (and other appropriate records if necessary) for a group of patients, the patients can be classified according to whether they have a particular condition and how much resource usage should be expected with their treatment.

The statistical information obtained in accordance with this invention may be employed for various purposes. For example, it may be employed in a risk adjustment computer software tool to accurately classify patients based upon conditions which require differential resource usage. Preferably, the classification has greater granularity than the DRG classification used for reimbursement purposes. That is, it subdivides a given DRG into multiple classifications by risk of death and cost. This allows a comparison of outcomes and costs across organizations for purposes of quality and cost management. Note that it is possible that a given subdivision may straddle more than one DRG.

The tools of this invention may also be employed to help define specific guidelines for treatment of patients. Patients falling into specific risk classifications, when subject to the analysis of this invention, may have certain conditions that could not otherwise be previously identified from the electronic discharge records. Guidelines to make more efficient use of available resources could then be specifically developed for the condition at issue.

Still further, the invention may be employed to determine whether new or existing guidelines are being followed. Because the invention accurately predicts resource usage for a certain set of conditions (and for following guidelines for those conditions), it may be employed to determine whether a health care organization is using more resources than expected. If so, it is likely that the staff is not following the guidelines in a large proportion of the cases.

These and other features and advantages of the invention will be described in more detail below with reference to the associated drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A is a block diagram of a generic computer system useful for implementing the present invention.

FIG. 1B is a depiction of a hospital electronic discharge record that may be processed by computer systems and methods of this invention.

FIG. 2A is a process flow diagram depicting the steps employed to generate a selection vector in accordance with one preferred embodiment of this invention.

FIG. 2B is a process flow diagram depicting the steps employed to identify a laboratory cost driver in accordance with one preferred embodiment of this invention.

FIG. 3A is a block diagram illustrating how septic shock selection vectors were generated in one example.

FIG. 3B is a graph ranking four hospitals in their costs of treating similarly severe septic shock patients.

FIG. 3C is a graph ranking the four hospitals of FIG. 3B in terms of their treatment outcomes (death rates specifically) in treating similarly severe septic shock patients.

FIG. 4 is a block diagram illustrating how laboratory cost drivers were generated in one example.

FIGS. 5A and 5B together are a list of terms for a linear expression incorporating a selection vector, various laboratory cost drivers, and other key explanatory variables to predict the total cost of a hospital admission.

FIG. 6A is a block diagram illustrating the inputs and outputs of a software program for analyzing health care organizations in accordance with a preferred embodiment of this invention.

FIG. 6B is a process flow diagram of a method for analyzing health care organizations in accordance with a preferred embodiment of this invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

The present invention provides methods and systems for analyzing electronic discharge records of patients in order to classify the patients according to a specified underlying condition that is not necessarily reflected in data elements in the discharge records. Further, the invention predicts or models resource usage by specific patients having a specified underlying condition. In the following description, various specific details are set forth in order to fully illustrate preferred embodiments for accomplishing the classification and modeling of this invention. For example, certain specific applications of the invention (e.g., identifying patients who had septic shock and modeling the cost of their treatment) will be described. It will be apparent, however, that the invention may be practiced without limitation to the specific details and applications presented herein.

DEFINITIONS

Some of the terms used herein are not commonly used in the art. Other terms are used in the art but not necessarily consistently. Therefore, the following definitions are provided as an aid to understanding the description that follows. The invention as set forth in the claims should not strictly be limited by these definitions.

The term "health care organization" refers to any organization which treats or manages patients. Examples of health care organizations include hospitals, health maintenance or other capitated payment organizations, preferred provider organizations, hospices, nursing homes, home care businesses, etc. For purposes of this invention, the health care organization or some associated entity will generate electronic records of the patients for financial management or billing purposes.

The term "patient resources" refers to resources used to treat a patient while that patient is under the care of a health care organization. In the context of a hospital, patient resources may include a cost component of a patient's treatment, the type of room a patient occupies (e.g., room size, attendant facilities, etc.), the length of time the patient stays in the hospital, the type of surgery, if any, required, the types and quantities of pharmaceuticals taken, the types and numbers of tests employed, etc. If the health care organization is a capitated payment organization, patient resources may also include such items as the number and type of nurse visits, special equipment, etc. The term "electronic discharge records" or "patient electronic profile" refers to those records maintained by a hospital or other health care organization for the purpose of accounting, reimbursement, comparison to other health care organizations, or a similar purpose. Such records are readable by computer systems, and particularly those systems deploying software that reads and analyzes such records. Often, such records contain multiple fields, each designated for entry of one or more ICD-9 codes or other patient condition codes. An example of one such record is a UB-92 record illustrated in FIG. 1B. Another example of such record is an ICCS code record maintained by Commission of Professional and Hospital Associations ("CPHA") which records the particular laboratory tests that were used to treat a patient. Of course, the present invention may utilize other forms and formats which recite patient condition codes, laboratory related cost drivers, or other information having a strong statistical correlation to resource usage and/or outcomes.

The term "key explanatory variable" is used herein to refer to any indicator which, when applied to a health care organization's electronic discharge records, will select those records associated with patients having a particular condition, a specified level of resource usage and/or a specified medical outcome. Examples include selection vectors of ICD-9 codes, laboratory cost drivers corresponding to tests performed on the patient, the geographic region within the country where the patient was treated, the number of beds or rooms in the health care organization, whether or not the patient died, whether or not surgery was required, whether or not the patient stayed in the intensive care unit, etc.

A "patient condition code" refers to a code employed to specify a patient condition for entry in an electronic discharge record. The code may take the form of a text string or a numerical string or other appropriate format. Patient condition codes are typically specified by a government agency or other large organization. Often, such codes are provided for the purpose of calculating or determining a reimbursement level to a health care organization for a particular patient. Examples of patient condition codes include DRG codes, ICCS codes, ICD-8 codes, ICD-9 codes, ICD-10 codes, CPT-4 codes, etc. The ICD-9 codes are presented in "ICD-9.CM," Fifth Edition, International Classification of Diseases, 9th Revision, Clinical Modifications, McGraw-Hill, Inc., New York, N.Y. (1996) and CPT-4 codes are presented in "AMA Physicians' Current Procedural Terminology CPT 97," CPT Intellectual Property Series, Chicago, Ill., 1996, both of which are incorporated herein by reference for all purposes.

COMPUTER SYSTEMS FOR IMPLEMENTING THE INVENTION

Embodiments of the present invention as described herein employ various process steps involving data stored in or transferred through computer systems. The manipulations performed in implementing this invention are often referred to in terms such as identifying, selecting, or comparing. Any such terms describing the operation of this invention are machine operations. Useful machines for performing the operations of embodiments of the present invention include general or special purpose digital computers or other similar devices. In all cases, there is a distinction between the method of operations in operating a computer and the method of computation itself. Embodiments of the present invention relate to method steps for operating a computer in processing electrical or other physical signals to generate other desired physical signals.

Embodiments of the present invention also relate to an apparatus for performing these operations. This apparatus may be specially constructed for the required purposes, or it may be a general purpose computer selectively activated or reconfigured by a computer program and/or data structure stored in the computer. The processes presented herein are not inherently related to any particular computer or other apparatus. In particular, various general purpose machines may be used with programs written in accordance with the teachings herein, or it may be more convenient to construct a more specialized apparatus to perform the required method steps. The required structure for a variety of these machines will appear from the description given below.

In addition, embodiments of the present invention further relate to computer readable media or computer program products that include program instructions and/or data (including data structures) for performing various computer-implemented operations. The media and program instructions may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well known and available to those having skill in the computer software arts. Examples of computer-readable media include, but are not limited to, magnetic media such as hard disks, floppy disks, and magnetic tape; optical media such as CD-ROM disks; magneto-optical media such as floptical disks; and hardware devices that are specially configured to store and perform program instructions, such as read-only memory devices (ROM) and random access memory (RAM). Examples of program instructions include both machine code, such as produced by a compiler, and files containing higher level code that may be executed by the computer using an interpreter.

FIG. 1A illustrates a typical computer system in accordance with an embodiment of the present invention. The computer system 100 includes any number of processors 102 (also referred to as central processing units, or CPUs) that are coupled to storage devices including primary storage 106 (typically a random access memory, or RAM), primary storage 104 (typically a read only memory, or ROM). As is well known in the art, primary storage 104 acts to transfer data and instructions uni-directionally to the CPU and primary storage 106 is used typically to transfer data and instructions in a bi-directional manner Both of these primary storage devices may include any suitable computer-readable media such as those described above. A mass storage device 108 is also coupled bi-directionally to CPU 102 and provides additional data storage capacity and may include any of the computer-readable media described above. Mass storage device 108 may be used to store programs, data and the like and is typically a secondary storage medium such as a hard disk that is slower than primary storage. It will be appreciated that the information retained within the mass storage device 108, may, in appropriate cases, be incorporated in standard fashion as part of primary storage 106 as virtual memory. A specific mass storage device such as a CD-ROM 114 may also pass data uni-directionally to the CPU.

CPU 102 is also coupled to an interface 110 that includes one or more input/output devices such as such as video monitors, track balls, mice, keyboards, microphones, touch-sensitive displays, transducer card readers, magnetic or paper tape readers, tablets, styluses, voice or handwriting recognizers, or other well-known input devices such as, of course, other computers. Finally, CPU 102 optionally may be coupled to a computer or telecommunications network using a network connection as shown generally at 112. With such a network connection, it is contemplated that the CPU might receive information from the network, or might output information to the network in the course of performing the above-described method steps. The above-described devices and materials will be familiar to those of skill in the computer hardware and software arts.

The hardware elements described above may implement the instructions of multiple software modules for performing the operations of this invention. For example, instructions for matching selection vectors to patient electronic profiles (in electronic discharge records for example) may be stored on mass storage device 108 or 114 and executed on CPU 108 in conjunction with primary memory 106.

A computer system such as system 100 typically analyzes electronic discharge records or other appropriate patient electronic profiles when operated in accordance with this invention. Such records may be stored on a machine readable medium such as devices 108 or 114 and analyzed by CPU 108 in conjunction with primary memory 106. FIG. 1B depicts some important features of a UB-92 (HCFA-1450) form 150 which is a standard electronic discharge record. UB-92 forms are conventionally required for Medicare reimbursement. For each patient treated by a health care organization, a medical records librarian reviews the patient's clinical records (as provided by an attending health care professional such as a doctor) and fills out a UB-92 form. The organization then submits the completed form to Medicare or other entity for reimbursement. As mentioned, the reimbursement level depends upon classification of the treatment into an appropriate DRG based upon analysis of the UB-92 form.

Form 150 includes various fields for medical records librarians to enter pertinent information chronicling a patient's treatment. Within form 150, a section 152 contains fields for various pieces of generic patient information such as name, address, birthdate, sex. It also includes less generic information which is specific to the patient's treatment such as a patient control number, a federal tax number, and a date on which the condition occurred.

Also within electronic medical record 150, a section 154 provides a listing of various expenses actually incurred by the hospital in treating the patient. As shown in this example, there are entries for pharmacy products, medical/surgical supplies, various tests and laboratory supplies, physical therapy, and operating room usage. Each of these items includes a cost to the hospital. These costs are summed and provided in a "total" line in the UB-92 form. Note that Medicare (or other payer using the UB-92 form) does not reimburse the hospital based upon these costs.

A section 156 includes the various patient condition codes that the medical records librarians enter when analyzing clinical records for the patient's treatment. In general, UB-92 forms such as form 150 include fields for a minimum of five diagnosis codes (e.g., septic shock, cancer, etc.) and fields for a minimum of three codes for procedures (e.g., surgery). Some patient treatments are limited and therefore do not require all eight patient condition code sections to fully describe the treatment. Within section 156, a principal diagnostic code field 158 specifies the diagnosis which largely determines the reimbursement. And "other diagnostic codes" section 160 lists such auxiliary diagnoses as comorbidities. While not shown in FIG. 1B, one example might provide a principal diagnostic code of 038.40 (the ICD-9 code for gram negative septicemia) and secondary (other) codes including disseminated intervascular coagulation (ICD-9 286.6) and acute venticular failure (ICD-9 391.8).

Below the diagnostic codes sections, multiple procedure code sections are provided. A "principal procedure" code 162 is provided for the main procedure performed while the patient was being treated. An "other procedure" code section 154 is provided for auxiliary procedures.

SELECTION VECTORS AND LABORATORY COST DRIVERS AS KEY EXPLANATORY VARIABLES

As indicated above, one key explanatory variable is a selection vector specifying one or more encoded patient conditions which can be matched against electronic discharge records from a health care organization. The selection vector should be carefully chosen and preferably validated by, for example, a statistical analysis such as that set forth below. The resulting selection vector should extract only those records of patients having a specified medical condition for which the vector was designed. The specified medical condition selected by the selection vector may be septic shock or any of a number of other conditions. A partial listing of such conditions and associated DRG codes follows.

    DRG           Admissions Type
    76 & 77       Other Resp Surg Complications or Comorbidities
    204           Disorders of Pancreas
    277 & 278     Cellulitus
    316           Renal Failure
    144 & 145     Other Circulatory Complications or Comorbidities
    15            Transient Ischemic Attack and Precerebral Occlusion
    130 & 131     Peripheral Vascular Accident Complications or
                  Comorbidities
    294 & 295     Diabetes Adult
    395           Red Cell Disorders
    24 & 25       Seizures, Headaches, Complications or Comorbidities
    403 & 404     Lymphomas, Leukemias
    188 & 189     Other Digestive Complications or Comorbidities
    202           Chirrosis, Hepatitis
    127           Heart failure with Shock
    475           Respiratory Involvement with Ventilator
    416           Septicemia Adult
    121, 122 & 123 Circulatory, Acute Myocardial Infarction
    79 & 80       Respiratory Infection
    174 & 175     Gastro-intestinal tract Hemorrhage
    488, 489 & 490 HIV
    296 & 297     Nutri-Metabolic
    182 & 183     Esophagitis Complications or Comorbidities
    140           Angina
    320 & 321     Kidney Urinary Track Infection Complications or
                  Comorbidities
    138 & 139     Cardiac Arrhythmia Complications or Comorbidities
    143           Chest Pain

                   ICD-9 Codes for Septic Shock
    Shock Codes
    Septic shock                        785.59
    Pulmonary shock                     518.5
    Renal Failure
    Acute                               586.x
    Post OP                             997.5
    Post trauma                         958.5
    Post labor                          669.3
    Toxemia                             593.9
    Post abortion      634.3 through 639.3 (only codes ending on .3)
    Pulmonary Failure
    ARDS                                518.82
    Resp. failure                       518.18
    On Respirator                       v46.1
    Heart Failure
    Fail. w/congestion                  428.0
    Left ventricle fail.                  428.1
    Acute sudden fail.                  428.9
    Cardio-resp. fail.                  799.1
    Circulatory fail.                   799.8
    Septic myocarditis                  422.92
    Toxic myocarditis                   422.93
    Acute hf & renal fail.                  404.93
    Acute vent. fail.                   391.8
    Coagulation Failure
    D.I.C.                              286.6
    Liver Failure
    Hepatic fail.                       572.8
    CNS Failure
    Cerebrovascular                     437.8
    collapse
    Infections
    Septicemia                          038.xx
    Bacteremia                          790.7
    Enterobacteremia                    038.49
    Infect. due to device                  996.62
    Bacterial Conditions
    Specific bacterias               001.x-005.x
                                     008.x-009.x
                                     020.x-041.x
                                     097.x-098.x
                                     100.x-104.x
    Note that ".x" represents any possible number for this digit.

            Gram Stain                              $730
            Transfusion Cross-Match                 $482
            Chemistry Panel 12                      $261
            Unspecified Lab Test                    $535
            Urinary Microscopy                     -$564
            Unspecified Bacteria Test               $189
            Serum Magnesium                         $463
            Bacterial Sensitivity                   $218
            Random Glucose                          $92
            A.P.P.T.*                               $516
            MIC**                                   $426
            Blood Gases                             $603
            Leukocyte Differential                  $375
            Chemistry Panel 7                       $658
            Specimen Collection                     $102
            Aerobic or Anaerobic Bacterial Culture        $995
            Aerobic Culture                         $498
            Complete Blood Count (CBC)             $1,002
    *Activated Partial Protrombintine
    **Minimize Inhibitory Concentration

                                Vector 1         Low Risk
                               Septic Shock        Sepsis
            Intercept             $5615           $2681
            ICU                   2977             2950
            Surgery               3595             2754
            Died                  -4301           -1963
            Amylase                690             216
            Blood gases            576             846
            CBC                    998             445
            Chem. 20               716             619
            Creatin. Kinase       -1425             72
            Coagulation            -45             230
            Creatinine            -596             -107
            Culture                -17             511
            Glucose                246              91
            Gram Stain            2071             -498
            Electrolytes           765             880
            Magnesium              262             661
            Specimen col.          137              99
            Urine                 -1664            -190
            Cross Match            40              848
            R square                0.73            0.52
            Adj. R. square          0.72            0.52
            Death Rate             50.9%            7.7%

    CONDITIONS    ICD9 CODES INCLUDED CODE EXPLANATION
    Septicemia    003.1              Salmonella septicemia
                  038.xx             Septicemia
                  639.0              Septicemia due to childbirth
                  999.3              Septicemia due to infusion
    Bacterial     031.8              B. Bronchisepticia
    infection     033.1              Bordatella parapertussis
                  036.0              Meningococcus
                  040.xx             Other bacterial diseases
                                     Bacterial infections in conditions
                  041.xx             classified elsewhere and of
                                     unspecified site
                  998.5              Post op infection
                  Procedure code 99.21 Infusion of antibodies
    Site-specified 042.1              HIV causing other specific
    infections,                      infections
    except        320.xx             Bacterial meningitis
    lungs or      322.xx             Unspecified meningitis
    kidneys       440.24             Athrosclerosis with gangrene
                  540.x              Acute appendicitis
                  567.0 and 567.2    Peritonitis, infectious
                  577.0              Acute pancreatitis
                                     Infection or inflammation due to
                  996.6x             implant
                  998.5              Post op. infection
    Kidney        590.xx             Infection of kidney
    infection     595.0              Acute cystitis
                  599.0              Bacteriuria
    Lung          481.xx             Pneumococcal pneumonia
    Infections    482.xx             Other bacterial pneumonia
                  484.3              Pneumonia in whooping cough
                                     Pneumonia due to inhalation of
                  507.0              food or vomitus
                  513.0              Lung abscess
    Fever         780.6              Fever of unknown origin
    Disorders of  276.2              Acidosis
    electrolytes  276.3              Alkalosis
                  276.4              Mixed acid-base disorder
                  276.7              Hyperkalemia
                  276.8              Hypokalemia
                  276.9              Not classified
    Fluid         276.0              Hyperosmolality
    disorders     276.5              Volume depletion
                  276.6              Fluid overload
    Dysrythmias   427.0              Tachycardia
                  427.1              V. Tach.
                  427.3x             Atrial fibrillation
                  427.4x             Ventricular fibrillation
                  427.6x             Premature beats
                  427.8x             Other dysrythmias
                  427.9              Other rhythm disorders
    Shock         785.59             Septic shock
                                     Post operative shock, endotoxic,
                  998.0              hypovolemic, septic
    Toxic shock   040.89             Toxic shock syndrome
                  639.5              Shock due to sepsis in childbirth
    Renal failure 586                Renal failure unspecified
                  584.xx             Acute renal failure
                  996.73             Complic. of renal dialysis
                                     Renal failure, tubular necrosis,
                  997.5              anuria due to complications
                  788.5              Oliguria/anuria
                  403.91             Hypertension with renal failure
                  Procedure code 39.65 Hemo-dialysis
                  Procedure code 39.95 ECMO
    Pulmonary     518.5              Pulmonary insufficiency
    failure                          following trauma and surgery
                  518.81             Respiratory failure
                  518.82             ARDS
                  799.1              Cardio-respiratory failure
                  V44.0              Tracheostomy
                  V461               Respirator
                  Procedure code 31.1 Tracheostomy
                  Procedure code 96.7x Continuous mechanical
                                     ventilation
                  Procedure code 39.66 Percenteral cardio-pulm. bypass
    Hepatic failure 997.4              Hepatic failure
                  570.x              Hepatic failure, coma
                  572.2              Hepatic coma
                  Procedure code 50.92 Hemo-dialysis for hepatic
                                     assistance
    CNS failure   293.xx             Transient organic psychosis
                  434.xx             Cerebral infarct
                  310.xx             Organic brain syndrome
                  997.0              Anoxic brain damage
                                     Mechanical complication of
                  996.2              nervous system device
                  780.02             Transient Alteration of
                                     awareness
    Cardiac failure 227.5              Cardiac arrest
                  428.x              Heart failure
                  799.1              Cardio-respiratory failure
                                     Heart syncope
                  992.1              Heart syncope with arrest due to
                  997.1              complications
                  585.5              Cardiac shock
    Coagulophaty  286.6              DIC
                                     Phlebitis or thromboplebitis due
                  997.2              to complications
                                     Acquired coagulation factor
                  286.7              deficiency
                  286.9              Unspec. coag. defect
                  287.5              Throbocytopenia unspecified

• Inventors
  Simpson, Kit N.;
• Assignee
  iMetrikus, Inc. (Carlsbad, CA)
• Published
  Jul-24-2001
• Current US Classes:
  705/3
• Application #
  145943
• International Classes
  G06F 017/60
• Field of Search
  705/2 705/3 705/4
• Examiner
  Stamber; Eric W.
• Agent
  Beyer Weaver & Thomas LLP
• US Patent References:
  4667292
  5018067
  5724379
  5778345
  5835897
  5845254
  5915241