Limits...
Assessing Women's Preferences and Preference Modeling for Breast Reconstruction Decision-Making.

Sun CS, Cantor SB, Reece GP, Crosby MA, Fingeret MC, Markey MK - Plast Reconstr Surg Glob Open (2014)

Bottom Line: None of the preference models performed significantly worse than the best performing risk averse multiplicative model.We hypothesize an average theoretical agreement of 94.6% for this model if participant error is included.There was a statistically significant positive correlation with more unequal distribution of weight given to the seven attributes.

View Article: PubMed Central - PubMed

Affiliation: Department of Biomedical Engineering, The University of Texas at Austin, Austin, TX ; Department of Plastic Surgery, The University of Texas MD Anderson Cancer Center, Houston, TX.

ABSTRACT

Background: Women considering breast reconstruction must make challenging trade-offs amongst issues that often conflict. It may be useful to quantify possible outcomes using a single summary measure to aid a breast cancer patient in choosing a form of breast reconstruction.

Methods: In this study, we used multiattribute utility theory to combine multiple objectives to yield a summary value using nine different preference models. We elicited the preferences of 36 women, aged 32 or older with no history of breast cancer, for the patient-reported outcome measures of breast satisfaction, psychosocial well-being, chest well-being, abdominal well-being, and sexual wellbeing as measured by the BREAST-Q in addition to time lost to reconstruction and out-of-pocket cost. Participants ranked hypothetical breast reconstruction outcomes. We examined each multiattribute utility preference model and assessed how often each model agreed with participants' rankings.

Results: The median amount of time required to assess preferences was 34 minutes. Agreement among the nine preference models with the participants ranged from 75.9% to 78.9%. None of the preference models performed significantly worse than the best performing risk averse multiplicative model. We hypothesize an average theoretical agreement of 94.6% for this model if participant error is included. There was a statistically significant positive correlation with more unequal distribution of weight given to the seven attributes.

Conclusions: We recommend the risk averse multiplicative model for modeling the preferences of patients considering different forms of breast reconstruction because it agreed most often with the participants in this study.

No MeSH data available.


Related in: MedlinePlus

We attempted to evaluate consistency between the participant (patient model) and multiattribute utility theory (MAUT model). Both models begin with the same initial conditions (true participant subconscious preferences). However, the patient model has a feedback mechanism that we cannot model (eg, we have no direct access to a patient’s subconscious or conscious preference). Each block is a potential source of error. We only have control over “MAUT” and “consistency evaluation” and partial control over “preference elicitation.” Note that reductions in consistency may be due to changes in preferences in the patient model due to feedback. Misranking outcomes is a major source of error in “consistency questions” on both the patient and MAUT side. It is difficult to discern if a participant indeed makes a mistake.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4120963&req=5

Figure 6: We attempted to evaluate consistency between the participant (patient model) and multiattribute utility theory (MAUT model). Both models begin with the same initial conditions (true participant subconscious preferences). However, the patient model has a feedback mechanism that we cannot model (eg, we have no direct access to a patient’s subconscious or conscious preference). Each block is a potential source of error. We only have control over “MAUT” and “consistency evaluation” and partial control over “preference elicitation.” Note that reductions in consistency may be due to changes in preferences in the patient model due to feedback. Misranking outcomes is a major source of error in “consistency questions” on both the patient and MAUT side. It is difficult to discern if a participant indeed makes a mistake.

Mentions: If we assume that participants incorrectly ordered outcomes with error resolutions >0.10, then consistency improves another 14.8% on average (15.7% for the risk-averse multiplicative model). Therefore, the theoretical consistency may be as high as 94.6%. There are a number of sources that contribute to error in consistency (Fig. 6). We are only able to control for error originating from the preference models (hence this study) and, to a limited extent, preference elicitation.


Assessing Women's Preferences and Preference Modeling for Breast Reconstruction Decision-Making.

Sun CS, Cantor SB, Reece GP, Crosby MA, Fingeret MC, Markey MK - Plast Reconstr Surg Glob Open (2014)

We attempted to evaluate consistency between the participant (patient model) and multiattribute utility theory (MAUT model). Both models begin with the same initial conditions (true participant subconscious preferences). However, the patient model has a feedback mechanism that we cannot model (eg, we have no direct access to a patient’s subconscious or conscious preference). Each block is a potential source of error. We only have control over “MAUT” and “consistency evaluation” and partial control over “preference elicitation.” Note that reductions in consistency may be due to changes in preferences in the patient model due to feedback. Misranking outcomes is a major source of error in “consistency questions” on both the patient and MAUT side. It is difficult to discern if a participant indeed makes a mistake.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4120963&req=5

Figure 6: We attempted to evaluate consistency between the participant (patient model) and multiattribute utility theory (MAUT model). Both models begin with the same initial conditions (true participant subconscious preferences). However, the patient model has a feedback mechanism that we cannot model (eg, we have no direct access to a patient’s subconscious or conscious preference). Each block is a potential source of error. We only have control over “MAUT” and “consistency evaluation” and partial control over “preference elicitation.” Note that reductions in consistency may be due to changes in preferences in the patient model due to feedback. Misranking outcomes is a major source of error in “consistency questions” on both the patient and MAUT side. It is difficult to discern if a participant indeed makes a mistake.
Mentions: If we assume that participants incorrectly ordered outcomes with error resolutions >0.10, then consistency improves another 14.8% on average (15.7% for the risk-averse multiplicative model). Therefore, the theoretical consistency may be as high as 94.6%. There are a number of sources that contribute to error in consistency (Fig. 6). We are only able to control for error originating from the preference models (hence this study) and, to a limited extent, preference elicitation.

Bottom Line: None of the preference models performed significantly worse than the best performing risk averse multiplicative model.We hypothesize an average theoretical agreement of 94.6% for this model if participant error is included.There was a statistically significant positive correlation with more unequal distribution of weight given to the seven attributes.

View Article: PubMed Central - PubMed

Affiliation: Department of Biomedical Engineering, The University of Texas at Austin, Austin, TX ; Department of Plastic Surgery, The University of Texas MD Anderson Cancer Center, Houston, TX.

ABSTRACT

Background: Women considering breast reconstruction must make challenging trade-offs amongst issues that often conflict. It may be useful to quantify possible outcomes using a single summary measure to aid a breast cancer patient in choosing a form of breast reconstruction.

Methods: In this study, we used multiattribute utility theory to combine multiple objectives to yield a summary value using nine different preference models. We elicited the preferences of 36 women, aged 32 or older with no history of breast cancer, for the patient-reported outcome measures of breast satisfaction, psychosocial well-being, chest well-being, abdominal well-being, and sexual wellbeing as measured by the BREAST-Q in addition to time lost to reconstruction and out-of-pocket cost. Participants ranked hypothetical breast reconstruction outcomes. We examined each multiattribute utility preference model and assessed how often each model agreed with participants' rankings.

Results: The median amount of time required to assess preferences was 34 minutes. Agreement among the nine preference models with the participants ranged from 75.9% to 78.9%. None of the preference models performed significantly worse than the best performing risk averse multiplicative model. We hypothesize an average theoretical agreement of 94.6% for this model if participant error is included. There was a statistically significant positive correlation with more unequal distribution of weight given to the seven attributes.

Conclusions: We recommend the risk averse multiplicative model for modeling the preferences of patients considering different forms of breast reconstruction because it agreed most often with the participants in this study.

No MeSH data available.


Related in: MedlinePlus