GOTM: a Goal-oriented Framework for Capturing Uncertainty of Medical Treatments

GOTM: a Goal-oriented Framework for
Capturing Uncertainty of Medical Treatments

Davoud Mougouei, David M. W. Powers School of Computer Science, Engineering, and Mathematics Flinders University
Adelaide, Australia
{davoud.mougouei,david.powers}@flinders.edu.au
Abstract

It has been widely recognized that uncertainty is an inevitable aspect of diagnosis and treatment of medical disorders. Such uncertainties hence, need to be considered in computerized medical models. The existing medical modeling techniques however, have mainly focused on capturing uncertainty associated with diagnosis of medical disorders while ignoring uncertainty of treatments. To tackle this issue, we have proposed using a fuzzy-based modeling and description technique for capturing uncertainties in treatment plans. We have further contributed a formal framework which allows for goal-oriented modeling and analysis of medical treatments.

I introduction

It has been widely recognized in the literature that associations between symptoms and medical disorders are uncertain [1, 2, 3, 4]. In other words, it is not known for sure that a set of medical symptoms would indicate a certain medical disorder [3]. In a similar spirit, associations between elements of a treatment plan (namely treatment goals and medical interventions) [5] are also uncertain [6, 7, 8, 9]. In other words, employing medical interventions designed for satisfaction of a certain treatment goal may not fully satisfy that goal [10, 11, 12, 13]. Hence, efficient medical models must capture uncertainties in treatment plans (guidelines).

Fuzzy logic, probability theory, and possibility theory have long been used in medical expert systems for capturing uncertainty of medical diagnosis [14, 15, 4]. The existing techniques for treatment modeling [16] however, have mainly ignored considering uncertainties in medical plans. Ying et. al [8] underlined uncertainty of HIV (Human Immunodeficiency Virus) treatment plans by proposing a fuzzy discrete event system for capturing uncertainty in finding an optimal regiment for treatment of HIV infection. Although the work underlines captures uncertainty associated with selection of medical interventions [17] it fails to consider uncertainties of medical interventions themselves.

In this paper, we have proposed using fuzzy logic for capturing uncertainty in treatment modeling of medical disorders. We have further contributed a goal-oriented framework referred to as GOTM (Goal-Oriented Treatment Modeling) which allows for goal-oriented modeling and description of medical treatments. Goal-oriented approach toward treatment modeling in GOTM provides structured arguments for choosing medical interventions [16] that satisfy certain treatment goals [16]. Fuzzy algebraic structure of treatment models in GOTM allows for capturing uncertainty of treatment goals and medical interventions in medical plans.

The remainder of this paper is organized as follows. Section II presents GOTM framework and its application on a hypothetical treatment plan, and Section III concludes our work with a summary of major contributions and future work.

Ii The Gotm Framework

Fig. 1: The GOTM Framework

Based on our earlier works in software requirement engineering [18, 19] we have contributed GOTM in order to consider uncertainty in development of medical treatment plans (guidelines). As depicted in Figure 1, GOTM consists of two main components: 1) the treatment modeling process, and 2) the treatment description process.

The process starts with development of treatment model [20]. Treatment model then, will be formally described to allow for automated analysis of treatment plans.

Ii-a Modeling Treatment Guidelines

As explained earlier, treatment plans (guidelines) are technical documents which capture treatment goals and medical interventions required for satisfaction of those goals [16]. Various techniques [21, 22, 23, 24] could be used for development of treatment plans. Modeling and description processes in GOTM nevertheless, are independent of the approach used for development of treatment plans.

Fig. 2: a Sample Treatment Model

The process of treatment modeling in GOTM starts with identification of high-level treatment goals from existing treatment plans. Then high-level treatment goals are extracted and refined into lower-level treatment goals and/or medical interventions. The refinement process continues until all of the treatment goals are refined to achievable medical interventions [18, 25]. The treatment model of a sample medical guideline is depicted in Figure 2. Junction points in Figure 2 represent logical AND-operator and their absence denotes OR-operator among treatment goals/medical interventions.

As one example, the OR-operator between and in treatment model of Figure 2 indicates that in order to satisfy the treatment goal , either of medical interventions or need to be employed. Also the AND-relation between treatment goals and specifies that both and are required for satisfaction of treatment goal .

Ii-B Description of Treatment Guidelines

As discussed earlier, associations between treatment goals/ medical interventions are uncertain [7, 8, 9, 12, 10]. In other words, treatment goals/medical interventions may partially contribute [25, 26] to satisfaction of their higher-level treatment goals in a treatment model. In order to care for this partiality, we have employed a fuzzy-based description process [27, 28] for describing associations between treatment goals/medical interventions in a treatment model.

The description process employed was first introduced by Mougouei et. al in  [19] for capturing uncertainty in modeling and description of security requirements [18]. The process was later successfully adopted for capturing partiality in visibility requirement engineering of commercial websites [29]. The main components of our description process are explained in the following sections.

Ii-B1 Construction of Fuzzy Grammar of Treatment

A fuzzy grammar [30] of treatment will be constructed to describe treatment model of a medical disorder. Such grammar is identified by a quintuple of where denotes the top-level treatment goal, is the set of treatment goals, is the set of treatment requirements and specifies the set of Fuzzy Derivation Rules (FDRs).

For treatment model of Figure 2, , and = “top-level treatment goal”. The membership function specifies the membership value of FDRs as given by (1). in (1) denotes the degree to which a treatment goal (requirement) contributes to satisfaction of a treatment goal .

(1)

FDRs can be identified based on relations among treatment goals/requirements in treatment models of medical disorders. For instance, the derivation rule in Table I explains that and in treatment model of Figure 2 contribute to satisfaction of top-level treatment goal . In addition, the fuzzy membership value of specifies that “ AND ” can almost completely satisfy the treatment goal . The membership values of FDRs will be specified by their corresponding fuzzy membership functions. For instance, .

 

Rule Membership Value Rule Membership Value

 

TABLE I: Derivation Rules for TM of Figure 2

Ii-B2 Calculation of Impact

Let to be members of , then specifies a derivation chain for . For a pair of in which denotes a treatment requirement and denotes a treatment goal on a derivation chain of , impact is given by the membership function where specifies the strongest contribution of to satisfaction of goal .

The value of is calculated through taking supremum over membership values of all derivation chains which generate . The membership value of a derivation chain is calculated by taking minimum over membership values of all FDRs on that derivation chain. Equation (2) demonstrates the calculation of impact. Fuzzy operators and denote fuzzy OR (maximum) and fuzzy AND (minimum) operators respectively [31].

(2)

As one example for calculation of impact, is calculated for the requirement in treatment model of Figure 2. For to be generated, there are two derivation chains in treatment model of Figure 2: , and . And so, we have: .

The impact values for treatment requirements of Figure 2 are listed in Table II. For each treatment requirement , different impact values are calculated with respect to different treatment goals on the derivation chains of . For instance, , and .

Goal )
0.9 0.7 0.6 0.85 0.8 0.7 0.7 0.65 0.6 0.6
0.9 0.7 0 0 0 0.7 0.7 0.65 0.6 0.6
0.85 0.7 0.6 0.85 0.8 0.7 0.7 0.65 0.6 0.6
0.9 0.7 0 0 0 0.7 0.7 0.65 0.6 0.6
0 0 0.6 0 0 0 0 0 0 0
0 0 0 0.9 0.8 0 0 0 0 0
0 0 0 0 0 0 0 0 0.6 0.6
0 0 0 0 0 0.85 0 0 0 0
0 0 0 0 0 0 0.75 0.65 0 0
0 0 0 0 0 0 0 0 0.6 0.6
0 0 0 0 0 0 0 0.85 0 0
0 0 0 0 0 0 0 0 0.8 0.85
TABLE II: Calculated Impact Values for Treatment Requirements of Figure 2

Iii Conclusion and Future Work

This paper has proposed caring for uncertainty in treatment modeling of medical disorders. To this end, we have proposed using a fuzzy-based technique for modeling and description of treatment models. The algebraic structure of the proposed modeling and description technique allows for capturing uncertainties of associations between goals/requirements of treatment guidelines. We have further contributed a goal-oriented framework which allows for goal-oriented modeling and description of treatment guidelines. Goal-oriented approach toward modeling treatment guidelines in GOTM provides structured arguments to choose from medical requirements those satisfying particular treatment goals.

We are extending the work by implementing a tool based on the GOTM and studying a real world medical plan.

References

  • [1] F. Griffiths, E. Green, and M. Tsouroufli, “The nature of medical evidence and its inherent uncertainty for the clinical consultation: qualitative study,” bmj, vol. 330, no. 7490, p. 511, 2005.
  • [2] F. A. Sonnenberg and J. R. Beck, “Markov models in medical decision making a practical guide,” Medical decision making, vol. 13, no. 4, pp. 322–338, 1993.
  • [3] K.-P. Adlassnig et al., “A fuzzy logical model of computer-assisted medical diagnosis,” Methods of Information in Medicine, vol. 19, no. 3, pp. 141–148, 1980.
  • [4] K.-P. Adlassnig, G. Kolarz, W. Scheithauer, H. Effenberger, and G. Grabner, “Cadiag: Approaches to computer-assisted medical diagnosis,” Computers in biology and medicine, vol. 15, no. 5, pp. 315–335, 1985.
  • [5] I. I. Rosen, R. G. Lane, S. M. Morrill, and J. A. Belli, “Treatment plan optimization using linear programming,” Medical Physics, vol. 18, no. 2, pp. 141–152, 1991.
  • [6] M. C. Politi, P. K. Han, and N. F. Col, “Communicating the uncertainty of harms and benefits of medical interventions,” Medical Decision Making, vol. 27, no. 5, pp. 681–695, 2007.
  • [7] H. Pan, Y. Li, Y. Cao, and Z. Ma, “Model checking fuzzy computation tree logic,” Fuzzy Sets and Systems, vol. 262, pp. 60–77, 2015.
  • [8] H. Ying, F. Lin, R. D. MacArthur, J. Cohn, D. C. Barth-Jones, H. Ye, L. R. Crane et al., “A fuzzy discrete event system approach to determining optimal hiv/aids treatment regimens,” Information Technology in Biomedicine, IEEE Transactions on, vol. 10, no. 4, pp. 663–676, 2006.
  • [9] S. Y. Belal, A. F. G. Taktak, A. Nevill, and A. Spencer, “An intelligent ventilation and oxygenation management system in neonatal intensive care using fuzzy trend template fitting,” Physiol. Meas, vol. 26, pp. 555–570, 2005.
  • [10] M. C. Politi, P. K. Han, and N. F. Col, “Communicating the uncertainty of harms and benefits of medical interventions,” Medical Decision Making, vol. 27, no. 5, pp. 681–695, 2007.
  • [11] M. H. Mishel, “Uncertainty in illness,” Image: The Journal of Nursing Scholarship, vol. 20, no. 4, pp. 225–232, 1988.
  • [12] B. B. Germino, M. H. Mishel, M. Belyea, L. Harris, A. Ware, and J. Mohler, “Uncertainty in prostate cancer. ethnic and family patterns.” Cancer Practice, vol. 6, no. 2, pp. 107–113, 1998.
  • [13] M. van Herk, P. Remeijer, and J. V. Lebesque, “Inclusion of geometric uncertainties in treatment plan evaluation,” International Journal of Radiation Oncology* Biology* Physics, vol. 52, no. 5, pp. 1407–1422, 2002.
  • [14] D. K. Owens and H. C. Sox, “Medical decision-making: probabilistic medical reasoning,” in Medical Informatics.   Springer, 2001, pp. 76–131.
  • [15] A. Ciabattoni, D. P. Muiño, T. Vetterlein, and M. El-Zekey, “Formal approaches to rule-based systems in medicine: the case of cadiag-2,” International Journal of Approximate Reasoning, vol. 54, no. 1, pp. 132–148, 2013.
  • [16] A. Ten Teije, S. Miksch, and P. Lucas, Computer-based medical guidelines and protocols: a primer and current trends.   Ios Press, 2008, vol. 139.
  • [17] D. Meltzer, “Addressing uncertainty in medical cost–effectiveness analysis: implications of expected utility maximization for methods to perform sensitivity analysis and the use of cost–effectiveness analysis to set priorities for medical research,” Journal of health economics, vol. 20, no. 1, pp. 109–129, 2001.
  • [18] D. Mougouei, M. Moghtadaei, and S. Moradmand, “A goal-based modeling approach to develop security requirements of fault tolerant security-critical systems,” in Computer and Communication Engineering (ICCCE), 2012 International Conference on.   IEEE, 2012, pp. 200–205.
  • [19] D. Mougouei and W. Nurhayati, “A fuzzy-based technique for describing security requirements of intrusion tolerant systems,” International Journal of Software Engineering and its Applications, vol. 7, no. 2, pp. 99–112, 2013.
  • [20] S. M. Weiss, C. A. Kulikowski, S. Amarel, and A. Safir, “A model-based method for computer-aided medical decision-making,” Artificial intelligence, vol. 11, no. 1-2, pp. 145–172, 1978.
  • [21] J. Fox, N. Johns, and A. Rahmanzadeh, “Disseminating medical knowledge: the proforma approach,” Artificial intelligence in medicine, vol. 14, no. 1, pp. 157–182, 1998.
  • [22] Y. Shahar, S. Miksch, and P. Johnson, “The asgaard project: a task-specific framework for the application and critiquing of time-oriented clinical guidelines,” Artificial intelligence in medicine, vol. 14, no. 1, pp. 29–51, 1998.
  • [23] L. Ohno-Machado, J. H. Gennari, S. N. Murphy, N. L. Jain, S. W. Tu, D. E. Oliver, E. Pattison-Gordon, R. A. Greenes, E. H. Shortliffe, and G. O. Barnett, “The guideline interchange format,” Journal of the American Medical Informatics Association, vol. 5, no. 4, pp. 357–372, 1998.
  • [24] P. D. Clayton, T. A. Pryor, O. B. Wigertz, and G. Hripcsak, “Issues and structures for sharing medical knowledge among decision-making systems: The 1989 arden homestead retreat,” in Proceedings of the Annual Symposium on Computer Application in Medical Care.   American Medical Informatics Association, 1989, p. 116.
  • [25] D. Mougouei, W. N. W. A. Rahman, and M. M. Almasi, “Measuring security of web services in requirement engineering phase,” International Journal of Cyber-Security and Digital Forensics (IJCSDF), vol. 1, no. 2, pp. 89–98, 2012.
  • [26] D. Mougouei, W. N. W. A. Rahman, and M. Moein Almasi, “Evaluating fault tolerance in security requirements of web services,” in Cyber Security, Cyber Warfare and Digital Forensic (CyberSec), 2012 International Conference on.   IEEE, 2012, pp. 111–116.
  • [27] D. Mougouei, W. Nurhayati, and M. Eshraghi Eavri, “Fuzzy-based intrusion tolerance for web-services.” 2013, pp. 83–88.
  • [28] ——, “Fuzzy-based intrusion tolerance for web-services.” International Journal of Advances in Computer Science and its Applications, vol. 4, no. 1, pp. 83–88, 2014.
  • [29] D. Mougouei and M. K. Yeung, “Visibility requirements engineering for commercial websites.” International Journal of Software Engineering & Its Applications, vol. 8, no. 8, pp. 11–18, 2014.
  • [30] J. N. Mordeson and D. S. Malik, Fuzzy automata and languages: theory and applications.   CRC Press, 2002.
  • [31] D. Mougouei and W. N. W. A. Rahman, “Fuzzy description of security requirements for intrusion tolerant web-services,” in The Second International Conference on Cyber Security, Cyber Peacefare and Digital Forensic (CyberSec2013).   The Society of Digital Information and Wireless Communication, 2013, pp. 141–147.
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
""
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
   
Add comment
Cancel
Loading ...
193803
This is a comment super asjknd jkasnjk adsnkj
Upvote
Downvote
""
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters
Submit
Cancel

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test
Test description