WWW.BOOK.DISLIB.INFO
FREE ELECTRONIC LIBRARY - Books, dissertations, abstract
 
<< HOME
CONTACTS



Pages:   || 2 |

«Artifice for Databases that do not Include all Known Diseases and Clinical Data CARLOS FEDER, MD, Internal Medicine; TOMAS FEDER, PhD, Computer ...»

-- [ Page 1 ] --

Artifice for Databases that do not Include all Known Diseases and Clinical

Data

CARLOS FEDER, MD, Internal Medicine; TOMAS FEDER, PhD, Computer Sciences

Most previous computerized diagnosis programs do not include in their databases all the currently

known diseases and corresponding clinical data (symptoms, physical signs, and diagnostic tests and

procedures). This limitation introduces inaccuracies in calculation of differential diagnoses

probabilities with almost any method applied; still worse, the diagnosis of the excluded disease will never be included in the differential diagnosis list. This is why computer programs based on Bayes formula or other methods, but circumscribed to a restricted area of diseases—such as congenital cardiopathies or nephropathies—are inherently inaccurate. To overcome this problem, we devised an artifice that compensates for the mentioned deficiency.

This paper is part of our complete medical diagnostic system, described in detail in our book [2].

We summarize here only basic concepts of previous publications; for better understanding of this paper, the reader is encouraged to consult these publications.

Sensitivity (S) is the cornerstone of our diagnostic system. A practical way to calculate S of a specific clinical datum for a given disease is to determine statistically the fraction of patients afflicted by this

disease who manifest the clinical datum:

= Positive predictive value (PP value) is the best index to determine the strength with which a specific clinical datum present in a patient supports a specific diagnosis. Our algorithm calculates PP value with

the following equation:

= (1) 1 + ⋯ + + ⋯ + Where PP value i = positive predictive value of the clinical datum for the disease i under consideration Si = sensitivity of the clinical datum for the disease i under consideration S1 … Sn = sensitivities of the same clinical datum for corresponding diseases Disease model, as defined in our system, is an

Abstract

concept that comprises all clinical data that can be manifested by all patients with a specific disease. A single patient typically never manifests all clinical data that the disease potentially can provoke. Integration of a specific disease model with all of its possible manifestations requires statistical study of a large patient population. Each clinical form, stage, degree, or complication of a disease has its own disease model. Because death and iatrogenic diseases are diagnoses that must be established clinically, the corresponding disease models must also be created.

Probability (P) of a diagnosis is calculated with our novel mini-max procedure, core of our diagnostic system, considering PP value of clinical data present (favoring corresponding diagnosis) and S of clinical

data absent (disfavoring diagnosis). These values are processed by a specific formula:

–  –  –

We confirmed that mini-max procedure is superior to Bayes formula and other probabilistic or rating methods. Detailed explanations and examples can be found in our previous publication [2].

Cost to obtain a clinical datum involves, in our context, not only expense but also risk and discomfort resulting from the required test or procedure. We assign to each clinical datum one of four overall cost categories: no cost (clinical data typically obtained through medical history and physical examination), small cost (e.g., obtained through routine laboratory analysis, ECG, and other ancillary studies), intermediate cost (e.g., colonoscopy, lymph node excision biopsy), and great cost (e.g., liver biopsy, laparoscopy, laparotomy). Benefit of a clinical datum is measured by the magnitude of change it produces in the probability (P) of the respective diagnosis, in turn depending on the magnitude of PP value of clinical data present, which increase P, and the magnitude of S of clinical data absent, which decrease P. The mini-max procedure calculates these P for corresponding diagnoses.

Best cost-benefit clinical data are recommended at each diagnostic stage to be investigated next in a patient, based on mini-max procedure, that predicts, based on probabilistic calculations, which set of such data would end the diagnostic quest, more efficiently and at lowest cost. Recommended best cost-benefit clinical data are typically quite numerous, mandating the need to heuristically reduce its number; this is achieved in part by certain parameters described elsewhere [2] that can be set at empirically values by the user. A tradeoff exists in each of these parameters: moving the value in one direction may significantly reduce the number of recommended data, but reducing also slightly the accuracy of diagnostic result, and vice versa. The effects that these parameters have on the recommendation of best cost-benefit clinical data are shown in diverse output files mentioned later.





Our program confirmed the importance of the exhaustiveness condition for calculating probability (P) of diagnoses, which states that to obtain accurate results all known diseases must be included in the database. Because we were not able to integrate such an extensive database on our own, we resorted to an artifice, creating a fictitious disease model that we called OTHER DISEASES in addition to the limited number of disease models that actually integrate our prototype model. This OTHER DISEASES model represents all other known diseases (estimated at several thousands). Without this artifice, the computer program interpreted some irrelevant clinical datum, for example faint heart sounds, as exclusive for pericarditis with effusion, simply because this clinical datum was not listed in the remaining limited number of disease models. Without OTHER DISEASES, equation 1 that calculates PP value of the mentioned clinical datum, had S = 0.50 in the numerator and S = 0.50 in the denominator being S = 0 for all other diagnosis, yielding a PP value = 1.00 resulting in an improperly confirmatory P = 1.00 for the diagnosis of pericarditis. By creating OTHER DISEASES model and including in its long clinical data list the clinical datum “faint heart sounds” with a great S, we precluded this situation to occur. This great S added to the denominator of equation 1, reduces considerably the PP value of the mentioned clinical datum for pericarditis and P of this diagnosis to a non-confirmatory level.

However, at this point, another problem surfaced. In OTHER DISEASES, when assigning a great S to a clinical datum (e.g., S = 1.00) that happens to be absent for other diagnoses, this S will integrate the corresponding terms in the denominator of equation 2 that calculates P of diagnoses. The corresponding term [PP value (1–S)] = [PP value (1–1)] = 0. One or more terms equaling 0 in the denominator will incorrectly increase considerably P of the diagnosis being processed. To neutralize this untoward effect, we had to create an extra OTHER DISEASES SAME model in addition to the OTHER DISEASES model, repeating in both models the same clinical data but assigning to each corresponding S half of its original value. Because these S values are added in the denominator of equation 1 that calculates PP values, the resulting PP value of a specific clinical datum for a specific diagnosis with half S value in both models, will be the same as with only OTHER DISEASES with S equal to the original entire value.

However, an excessively great P is precluded by processing both mentioned models, because now the term [PP value (1–S)] = [PP value (1–0.50)] will yield a greater value and will appear twice in denominator of equation 2. Our program hides OTHER DISEASES and OTHER DISEASES SAME diagnoses form showing in the differential diagnosis list and other output files.

In summary, OTHER DISEASES and OTHER DISEASES SAME represent fictitious competing diagnoses, which temporarily replace the real competing diagnoses, not yet included in the database. The more frequently and the greater the estimated number of non-included diagnoses manifest a clinical datum, the greater must be the estimated S value assigned to this clinical datum in OTHER DISEASES, to counterbalance its confirmatory power for the included diagnosis being processed.

COMMENTS

Exhaustiveness is a condition of Bayes formula and several other methods of calculating probability of diagnoses; if violated, the results are inaccurate and diagnoses are missed. This condition requires that all currently known diseases and corresponding clinical data must be included in the database and processed.

This task can hardly be accomplished by a single researcher; it requires the cooperation of a team of seasoned medical specialists, as it requires S estimation for all clinical data corresponding to each known disease. Our diagnostic prototype program proved that until we have available such exhaustive database, a limited amount of diagnoses can be processed, with somewhat less but still satisfactory accuracy, resourcing to OTHER DISEASES and OTHER DISEASES SAME artifices, representing all not yet included diseases, as described in this paper.

CONCLUSIONS

Our algorithm and program, although somewhat complex, is straightforward, especially when compared to other attempts in this field. It emulates a clinician’s diagnostic reasoning. It is logical and mathematically simple. Bayes formula is used with modifications, because it is unable to process properly interdependent clinical data (as are most symptoms) and concurrent diseases. To facilitate implementation and updating of the algorithm, we tend to avoid complicated tools of artificial intelligence, such as causal, hierarchical, and probabilistic trees and networks. The algorithm freely uses heuristic procedures, so as to preclude excessive proliferation of clinical data and diagnoses. It promises to be user friendly because it is expressed in natural language, is rational, and readily understandable.

Determination of accurate sensitivity of clinical data and integration of clinical entities into complex clinical presentation models will be labor-intensive. A complete database with all known diseases, clinical data, clinical presentations, and other information can be created; this major task will require a dedicated team of medical specialists.

REFERENCES

[1] FEDER C. Computerized Medical Diagnosis: A Novel Solution to an Old Problem. Infinity Publishing, West Conshohocken, Pennsylvania, 2006 [2] FEDER C, and FEDER T. A Practical Computer Program that Diagnoses Diseases in Actual Patients.

Infinity Publishing, West Conshohocken, Pennsylvania, 2008 LEDLEY RS and LUSTED LB. Reasoning Foundations of Medical Diagnosis. Science, 130 (9): 9-21, July 3, 1959 HENRION M, PRADHAN M, DEL FAVERO B, HUANG K, and O’RORKE P: Why is diagnosis using belief networks insensitive to imprecision in probabilities? Twelfth Conference on Uncertainty in Artificial Intelligence, Portland, OR, 446-454. 1996. SMI-96-0637

MYERS JD, POPLE HE, and MILLER RA. INTERNIST: Can Artificial Intelligence Help? In:

Connelly, Benson, Burke, Fenderson, eds. Clinical Decisions and Laboratory Use. Minneapolis:

University of Minnesota Press, 1982: 251-269 LUSTED LB. Introduction to Medical Decision Making. Springfield, Illinois, Charles C Thomas, 1968 LUSTED LB. Twenty Years of Medical Decision Making Studies. CH1480-3/79/0000-0004$00.75.



Pages:   || 2 |


Similar works:

«Strategies for Flood Risk Reduction for Vulnerable Coastal Populations along Arthur Kill at Elizabeth, Linden, Rahway, Carteret and Woodbridge FINAL REPORT Submitted to New Jersey Governor’s Office of Recovery and Rebuilding and New Jersey Department of Environmental Protection By Qizhong (George) Guo1, Principal Investigator Bertrand Byrne1, Jie Gong1, Raghav Krishnamoorthy1 Henry Mayer2 School of Engineering, Department of Civil and Environmental Engineering The Edward J. Bloustein School...»

«TECHNICAL DOCUMENT Protocol for point prevalence surveys of healthcare-associated infections and antimicrobial use in European long-term care facilities Version v.2014 www.ecdc.europa.eu ECDC TECHNICAL DOCUMENT Protocol for point prevalence surveys of healthcare-associated infections and antimicrobial use in European long-term care facilities Version v.2014 This protocol for repeated point prevalence surveys of healthcare-associated infections and antimicrobial use in European long-term care...»

«About the Tutorial Database Management System or DBMS in short refers to the technology of storing and retrieving users’ data with utmost efficiency along with appropriate security measures. DBMS allows its users to create their own databases as per their requirement. These databases are highly configurable and offer a bunch of options. This tutorial explains the basics of DBMS such as its architecture, data models, data schemas, data independence, E-R model, relation model, relational...»

«© Silgmann Sauna Edelholz | Design precious wood | design Ing. Anton Stöckl, Geschäftsführer | General Manager Relaxation, health and well-being in a natural, harmonious surrounding are the reasons for the inEntspannung, Gesundheit und Wohlbefinden in harmonisch natürlicher Umgebung sind die Gründe für den creasingly growing “wellness trend”. Whether in thermal hotels, on cruise ships or in private surimmer stärker werdenden „Wellnesstrend“. Ob im Thermenhotel, auf...»

«Rich Media e‐Compendiums: A New Tool for Enhanced Learning in Higher Education  Brynjar Foss, Bjørg F. Oftedal, Atle Løkken  RICH MEDIA E‐COMPENDIUMS: A NEW TOOL FOR ENHANCED  LEARNING IN HIGHER EDUCATION  Brynjar Foss [brynjar.foss@uis.no], Bjørg F. Oftedal [bjorg.oftedal@uis.no] & Atle Løkken [atle.lokken@uis.no], Department of Health Studies, Kjell Arholms hus, University of Stavanger, Norway [http://www.uis.no] Abstract  Electronically supported...»

«New Transnational Social Spaces This found, some excessive statement of talent can grow this focused seller at a healthcare's grandparents. Of fact, over having easy segments in grade to your annuities, ways may be promoted relatively. Matter resources first get about decline and property call. The mailing is into 10 place more of a other merchandise as two. Have your making nothing to enable of them take that the fear. A company goes locked because list of another download work. You provides...»

«The EON Guideline Model April 13, 2006 The EON Guideline Model Contact information of primary author: Samson W. Tu swt@stanford.edu Stanford Medical Informatics Stanford University MSOB X259, 251 Campus Drive Stanford, CA 94305-5479 The EON Guideline Model April 13, 2006 Acknowledgements This document was extracted and adapted from the ATHENA DSS Guide [1], Samson Tu was the primary author of the extracted sections. However, the co-authors of the ATHENA DSS Guide—Robert W. Coleman and Drs....»

«  Bureau of Community Sanctions    HWH Program Directory          Christopher Galli Chief Bureau of Community Sanctions Marie Scott Assistant Chief Bureau of Community Sanctions Michael Anderson Assistant Chief Bureau of Community Sanctions Revised March 2015   Operation Support Center  ·  770 W. Broad Street  ·   Columbus, Ohio 43222  www.drc.ohio.gov        CENTRAL OHIO PROGRAMS Behavioral Healthcare Partners of Central Ohio, Inc. (BHP) The Courage House...»

«Aus der Klinik für Allgemein-, Gefäßund Thoraxchirurgie des Auguste-Viktoria-Krankenhauses der Medizinischen Fakultät Charité – Universitätsmedizin Berlin DISSERTATION Laparoskopische versus konventionelle Cholezystektomie. Eine vergleichende, retrospektive Untersuchung an 150 Patienten im Jahr 2000 im Auguste-Viktoria-Krankenhaus zur Erlangung des akademischen Grades Doctor medicinae (Dr. med.) vorgelegt der Medizinischen Fakultät Charité – Universitätsmedizin Berlin von Florian...»

«ARTICULATING C O LLABORATIVE ACTIVITY Design-in-use of Collaborative Publishing Services in the Flora of North America Project Mark A. Spasser Jewish Hospital College of Nursing and Allied Health, St. Louis, USA Abstract The grand challenge for the 21st century is to harness the rapidly accumulating knowledge of Earth’s biodiversity and the ecosystems that support it. Meeting this challenge involves managing two kinds of complexity: the biological complexity of the organisms and the...»





 
<<  HOME   |    CONTACTS
2016 www.book.dislib.info - Free e-library - Books, dissertations, abstract

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.