Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1998-06-30
2001-04-03
Alam, Hosain T. (Department: 2771)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C705S002000
Reexamination Certificate
active
06212519
ABSTRACT:
FIELD OF THE INVENTION
The invention relates to systems and methods for creating databases of information, and more particularly, for creating databases of statistical data for applications such as diagnosing medical conditions.
BACKGROUND OF THE INVENTION
The process of diagnosing an illness or a disease typically involves recording the demographic information about the patient and documenting the findings that the patient is exhibiting. Together the demographic information and the finding information provide a patient profile. The diagnosing physician can compare the demographic information and the finding information of the patient against a listing of the common and descriptive findings for a particular disease, and can determine the likelihood that such a disease would be present in a patient having the information set forth in the profile.
For this process to work well, the physician must be provided with a database of information that sets forth the descriptive symptoms for a particular disease. Unfortunately, this information is difficult to come by. Frequently, the doctor must rely on personal experience, or search through the literature to identify studies that set forth statistical ranges indicating the likelihood that a particular finding is present in a patient with a particular disease. Often, however, the literature only offers general qualitative discussions as to the frequency or relevancy of a particular finding for a particular disease. Such qualitative expressions include statements that a particular disease is “very frequently” accompanied by a particular finding or that the finding is “almost never absent” in a patient with a particular disease. Similar qualitative expressions are used to describe the likelihood that a patient with a certain demographic profile will have a particular disease.
Although qualitative expressions provide information as to the probability that a particular finding will be seen in a patient suffering from a particular disease, the information is conveyed in a somewhat ambiguous manner and lacks the precision of a numerical statement that typically would be provided by a rigorous statistical analysis. However, patient studies that result in rigorous statistical analysis are few and far between. The scarcity of such publications arises in part from the difficulty in obtaining research support for analyzing a sufficiently large pool of candidates to come away with meaningful statistical data. Consequently, much of the diagnostic information available to a treating physician is in the form of qualitative expressions.
Accordingly, there is a need in the art for a system that can extract reliable and meaningful information from qualitative expressions.
SUMMARY OF THE INVENTION
It is an object of the invention to provide systems and methods that can evaluate qualitative expressions of likelihood to provide a more precise measure of statistical meaning.
The systems and methods of the invention can process a qualitative expression, typically a phrase to generate a numerical value that represents the likelihood of an event as expressed by the given qualitative expression. The qualitative expressions typically are made up of a probability term and possibly a term of degree and one or more terms of negation. Thus, the systems described herein allow numerical information to be captured from qualitative expressions. This, allows further numeric processing and data refining, and facilitates the development of databases of information, such as databases of information about the significance of the existence of a symptom for the diagnosis of a particular disease.
The systems described herein allow the use of qualitative information from separate sources to be combined in a meaningful way. For example, the systems described herein can average numbers that were obtained from the analysis of qualitative expressions from two different articles.
To this end, the systems described herein analyze a qualitative expression by examining each of the terms of that expression. The systems can identify those words which can be understood as probability terms and which can be associated with known numerical values of probability. For example, the qualitative expression may include the probability term “frequently”, which may be known by the system to have, in certain contexts, an associated numerical probability value of about 56%. In one embodiment, the systems include a table of these known probability terms and the associated numerical probabilities. The system then analyzes the other terms in the expression, including terms of degree and terms of negation, and processes how these other terms impact on the numerical probability associated with the probability term. The resulting numerical value is understood to capture in a numerical format the probabilistic information provided by the qualitative expression.
In one embodiment, the invention provides processes for generating a numerical value of likelihood. These processes include the steps of identifying a qualitative expression of likelihood, having at least one term, parsing the qualitative expression as a function of the at least one term to identify a semantic form related to the qualitative expression, identifying for that one term a numerical value of probability associated with that one term, identifying, as a function of the semantic form, a formula for processing the numerical value of probability, and processing the numerical value of probability as a function of the formula to determine a numerical value of likelihood associated with the qualitative expression. In these processes qualitative expressions, such as the expression “most likely absent,” can be identified from the medical literature and employing the processes of the invention, analyzed to generate a numerical value of probability associated with that qualitative expression. The qualitative expressions can include identifying terms that are representative of terms of degree, such as the term “very,” and can identify whether the term is a positive or negative degree term. Similarly, when parsing the qualitative expression, the processes described herein can identify whether the qualitative expression includes terms of negation, such as the term “not.”
In one process, identifying a numerical value of probability includes a step of searching a table for a numerical value of probability associated with the term, and identifying whether the term is a positive or negative probability term. In one process, identifying a formula as a function of a semantic form includes selecting a formula associated with the semantic form. In one practice, the form is selected from a table which includes a plurality of formulas wherein each formula is associated with at least one semantic form. In a further practice, the step of identifying a formula can include a step of generating a formula as a function of combining at least two related formulas. In further practice, the process includes a step of providing a table of weighting factors, wherein each factor is associated with a term of degree. Accordingly, each term of degree can have associated with it a numerical value that indicates quantitatively the effect of the term of degree on the qualitative expression.
The invention described herein can also include systems for generating a database of statistical data of medical information. These systems can include an input device for receiving a text string representative of a qualitative expression of the likelihood of the occurrence of an event, a parser for parsing the qualitative expression into one or more terms, including at least one term representative of a statement of likelihood, and for selecting a semantic form as a function of the parsed qualitative expression associated with the qualitative expression. These systems can also include a selection processor for selecting, for each term in a qualitative expression, a numerical value and a numerical processor for processing the selected numerical value as a function of the formula to provide a nume
Alam Hosain T.
Fleurantin Jean Bolte
Foley Hoag & Eliot LLP
SimulConsult, Inc.
LandOfFree
Systems and methods for quantifying qualitative medical... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Systems and methods for quantifying qualitative medical..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Systems and methods for quantifying qualitative medical... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2553357