Divide and conquer: novel patient-segmentation approach leads to new opportunities: anonymized patient-level data can be used to segment patients according to clinical factors, therapy usage, and demographics, allowing pharmaceutical companies to reach the right physicians and patients with the right messages.

Patient segmentation puts a new twist on the age-old concept of "divide and conquer." Instead of a strategy for maintaining power over vast empires, it describes the way pharmaceutical companies can benefit by segmenting the patient population and concentrating their resources on the group(s) that will benefit most from their products.

The concept of patient segmentation is not new. For years, marketers have segmented their patient populations with limited insight into real-world, evidence-based clinical information. They relied mostly on primary market research (which was sometimes limited by small sample sizes and inaccuracies related to physician and patient recall), in combination with dispensed prescription data, to understand how product utilization varied by patient segment.

The variables used in segmentation were typically restricted to age and gender, which are limited in their ability to provide insights into a patient's clinical profile. For example, these factors do not address the comorbidities that are common among patients in a particular segment.

Today, anonymized patient-level data (APLD) combines patient demographics with actual evidence-based treatment patterns for millions of de-identified patients, allowing companies to take their market assessments and patient segmentations to a more sophisticated and comprehensive level. This information can identify distinct patient segments based on their disease attributes, treatment specifics, and/or key demographics, including the presence of comorbidities, which can lead to the identification of new opportunities and advantages in the market.

In the following example, a major player in the proton-pump inhibitor (PPI) class used an APLD-driven patient-segmentation approach to improve its performance in the gastroesophageal reflux disease (GERD) market.

The Means to Understand

The company's brand team had observed that all products in the class were experiencing a high volume of patient churn, including brand switches (i.e., patients switching to and from their brand) and restarts (i.e., patients restarting their brand after quitting). They hypothesized that utilization varied significantly based on patients' clinical profiles. If this theory were borne out, the company intended to identify and concentrate its efforts on segments representing the highest current and potential value to its brand. The manufacturer used de-identified patient-level data to segment the market and analyze actual utilization patterns.

The segmentation revealed four distinct patient segments characterized by different demographics, clinical factors, and therapy usage (Figure). Patients in segment B, which represented 40% of the market, tended to be older, male, and suffered from the fewest number of comorbidities. Their clinical characteristics indicated a good potential fit with the company's product. Most (68%) were treated with PPIs, and registered the longest duration of therapy with the greatest rate of compliance. Therefore, this segment presented the greatest potential value to the brand.

The Insight to Differentiate

Further analysis revealed only a small amount of share variability occurred for each brand in the PPI class across each segment (Table). For example, if a brand had a 6% share of segment A, it likely had a similar share of segments B, C, and D. This suggested a potential lack of product differentiation between the available brands in the minds of physicians and the patients they treat. Until this point, share in this market had largely been driven by share of promotional voice.

To combat this, the company capitalized on its newfound knowledge of patients in segment B by developing tailored messaging for physicians. Rather than simply recommending that physicians use the brand for patients who suffer from GERD, the brand team designed physician messaging that described the specific clinical and demographic profile of the patients who would benefit most from the therapy (i.e., older patients, males, patients diagnosed with diseases of the esophagus), and was aligned with the product's labeling.

The upfront investment in a more complete assessment of customer behavior paid big dividends in downstream marketing efficiencies. The salesforce was subsequently able to detail segments of physicians who were treating patients in segment B, and deliver more tailored messages. Marketing and sales teams were able to measure the success of the company's more informed strategy by monitoring changes in product uptake among key segments of physicians and patients.

By using de-identified patient-level data to identify and size the patient segments that would benefit most from the brand, the company was better positioned to differentiate its brand in a relatively undifferentiated market, providing potential advantages over competitors who were communicating a more general message. Furthermore, the company was able to disseminate their message more efficiently by reaching out to specific segments of physicians and patients, rather than trying to blanket the market. Enabling the brand team to take a more granular look at the market gave the company a significant marketing edge.

The Power to Gain Share

De-identified patient-level data allow companies to better assess their markets and seize new opportunities. Marketers can operate with a deeper understanding of the patients who their products are meant to treat, enabling companies to segment, position, and communicate with them more effectively. Furthermore, it can be translated directly into sales strategies and tactics, ensuring a comprehensive and effective approach to the market.

Mr. Carroll is Director, Product Management, and Ms. Tirrell is Product Marketing Manager, at PharMetrics, a unit of IMS. For more information about IMS and PharMetrics, please call (617) 393-8484 or E-mail apld@us.imshealth.com. TABLE: PERCENTAGE OF PATIENTS BY SEGMENT AND PROTON-PUMP INHIBITOR BRAND Brand Segment A Segment B Segment C Segment D Brand A 4.2% 4.7% 5.4% 4.7% Brand B 4.5% 0.4% 1.9% 0.7% Brand C 28.7% 23.7% 25.6% 25.1% Brand D 54.0% 65.8% 57.1% 61.0% Brand E 8.6% 5.4% 10.1% 8.5% Total 100.0% 100.0% 100.0% 100.0% Figure. Patient segmentation of gastroesophageal reflux disease (GERD) market. PPI = Proton-pump inhibitor; Rx = prescription; GI = gastrointestinal; [H.sub.2] = histamine receptor. Segment A--Untreated "At Risk" 28.5% * Youngest (mean age = 45 yr, with 17% < 18 yr) * Untreated with GI Rxs (89%) * Fewest comorbidities * Diagnoses:

Diverticula of the intestine

GI hemorrhage

GI systems symptoms

Functional digestive disorder Segment B--Primary GERD 40.0% * Older (mean age = 55 yr, with 70% 35-64 yr) * Greater number of males * Largest number treated with PPIs (68%) * Highest utilization of PPIs (days supplied) * Longest therapy duration and greatest rate of compliance * Fewest other related Rxs * Diagnoses:

Diseases of the esophagus

GERD Segment C--Secondary GI Rx 25.4% * Older (mean age = 52 yr, with 58% 35-64 yr) * Largest number treated with [H.sub.2] blockers (34%) * Shortest therapy duration and lowest rate of compliance * Diagnoses:

Gastritis & duodenitis

GERD

H. pylori Segment D-Treated "High Risk" 6.1% * Oldest (mean age = 55 yr, with 29% [greater than or

equal to] 65 yr) * Mostly treated with PPIs (65%) * Highest number of comorbidities, medical procedures,

other related Rxs, hospital stays, and office visits * Diagnoses:

Diseases of the esophagus

Dyspepsia/stomach disorders

Gastritis & duodenitis

GERD & ulcers Note: Table made from pie chart.