Browse Prior Art Database

Patient Targeting Methodology for Healthcare

IP.com Disclosure Number: IPCOM000127331D
Original Publication Date: 2005-Aug-23
Included in the Prior Art Database: 2005-Aug-23
Document File: 2 page(s) / 28K

Publishing Venue

IBM

Abstract

This article describes a process using data mining for patient prospecting in the healthcare industry. The process extends the Promotion Targeting Methodology developed for the retail industry by integrating a healthcare data model with the PTM technology, thereby creating a targeting methodology specific to healthcare. A novel aspect of this process is that it adapts an analytical customer selection methodology developed and validated in the retail industry to the selection of subjects for clinical trials or patients to be targeted for specific treatment protocols in the pharmaceutical and healthcare industries.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

Patient Targeting Methodology for Healthcare

   This article describes a process using data mining for patient prospecting in the healthcare industry. The process extends the Promotion Targeting Methodology (PTM)* developed for the retail industry by integrating a healthcare data model with the PTM technology, to create a targeting methodology specific to healthcare. A novel aspect of this process is that it adapts an analytical customer selection methodology developed and validated for the retail industry to the selection of subjects for clinical trials or patients to be targeted for specific treatment protocols in the pharmaceutical and healthcare industries.

   Healthcare providers (HCP), such as hospitals or HMOs, need to identify patients to target for clinical trials or specialized treatments, e.g., selecting high-risk congestive heart failure (CHF) patients for a supplemental treatment, identifying patients likely to require long-term rehabilitation following scheduled surgery, or targeting currently healthy patients at high risk of developing a serious disease for a wellness program sponsored by an HMO. The HCP may specify a desired number of patients fitting specific criteria to target in a particular program, where the number of patients is limited either by available budget or by competing funding needs. Using specific criteria, including clinical and demographic information on patients, the HCP queries its patient database to identify prospective targets. This process of compiling a list of prospective patients to target is analogous to the process known as "prospecting" for targeted promotions in the retail industry.

   The problem faced by the HCP in prospecting is that, in general, applying the selection criteria to the patient database will result in too few prospects being identified. The HCP is then faced with the problem of relaxing the selection criteria to identify additional prospects in order to attain the desired number of patient prospects. Such a query-based, trial-and-error process of supplementing the initial set of patient prospects may not produce the highest-potential list of prospects. As a result, overall patient benefits may be less than could have been achieved with a better selection technique.

   PTM offers a quantitative way to achieve a set of trial subjects or patients who best fit the selection criteria for the trial or protocol. The approach is based on the data mining technique known as clustering. A clustering model is built using a data mining tool, such as DB2 Intelligent Miner**, which is capable of outputting calculated measurements including assigned cluster ID and confidence of the assigned fit for each record. The confidence statistic (ranging from 0 to 1, where a larger value indicates a better fit of the record to its assigned cluster) represents the degree of certainty that the record has been assigned to the best possible cluster. T...