Market Segmentation of Adults with Chronic Pain

Cluster analyses of a nationally representative sample of U.S. adults with chronic pain in one or more joints (e.g. knee, hip, shoulder) revealed that:

  •  adults who suffer from chronic joint pain are not a homogenous population
  •  alternative cluster modeling algorithms produce different segmentation solutions
  •  best solution is the one that produces intra-cluster homogeneity and inter-cluster differences, as well as discriminant dimensions that align well with targeted messaging
  •  chronic pain segments varied in terms of pain sites, comorbidities, lifestyle, emotional well-being, and likelihood to seek medical treatment

Pain perception is subjective. Each person's experience of pain is shaped by emotional and cognitive processing of the physical sensation. Thus it is arguably impossible to objectively quantify the level of one person’s pain with another’s, because pain perception varies greatly across individuals. Thus direct self-reports of pain in survey data remain an essential means of measuring population pain sites & severity.

The current analysis aims to create distinct clusters of adults suffering from chronic pain, specifically chronic pain the joints of the body including pain in the hip, knee, shoulder, etc.  A segmentation of chronic joint pain sufferers was needed to inform patient messaging and product development in this diverse market. Focus was placed on psychological differentiators of chronic pain frequency and severity experienced by U.S. adults, who may or may not have received an official diagnosis.

Heterogeneity in the Chronic Pain Population.jpg

Around 75 million U.S. adults experience joint pain lasting 3 months or more, defined as “pain, aching, or stiffness in or around a joint” - including pain in the hip, knee, shoulder, elbow, wrist, fingers, ankle, and toes. Some of these adults take the initiative to see a doctor for their chronic pain; some do not. Some are obese; some are not. Some may receive an official diagnosis of arthritis, gout, lupus, or fibromyalgia; some do not. Some report being limited in daily activities due to joint pain; some do not. Some of these adults experience pain in other sites such as back and neck; some do not. Some of these adults suffer from other concurrent chronic conditions; some do not. Clearly, the population of adults who suffer from chronic joint pain is not homogenous.

A target population that is not homogenous presents an opportunity for segmentation. Segmenting a broad population into subgroups can help to inform patient messaging and product development in a diverse market. The core objective of segmentation is to extract coherent clusters whereby similarities along attributes of interest are maximized within cluster, and differences are maximized between clusters. Careful thought on input dimensions is critical because multiple segmentation solutions could be derived from the same data. Input dimensions should be directly relevant to the primary research question, and should yield maximum differentiation among respondents. To this end, cluster modeling was performed to extract coherent clusters in terms of the key differentiators of chronic joint pain frequency and severity, including the following input dimensions available from the NHIS 2015 survey data:

Physical Factors

  • Diagnosed chronic conditions 
  • Pain sites: joints as well as other sites
  • Exercise frequency
  • BMI
  • Smoking 
  • Alcohol consumption

Psychological Factors

  • Pain perceptions
  • Feelings of depression, anxiety, worthlessness, pessimism
  • Perceived social support vs. isolation
  • Sleep disruptions
  • Financial stress
  • Work stress

Different clustering algorithms applied on the same data will produce different segmentation solutions. In this study, we attempted 4 alternative clustering algorithms, namely:

  • Agglomerative Hierarchical Clustering
  • K-means Clustering 
  • Partitioning Around Medoids (PAM) k-medoids Clustering 
  • Latent Class Clustering
method summary.jpg

Details on the process underlying each technique is provided in the summary table above. As is often the case, more than one viable segmentation solution can be developed with the same data. The alternative segmentation solutions were evaluated using a set of common metrics that permit comparisons across methods in terms of validity and stability. A good cluster solution should yield high homogeneity within each cluster and high heterogeneity between clusters. Various indices of intra-cluster homogeneity, inter-cluster separation, and cluster stability suggested that a 5-cluster solution worked best for this target sample, and the PAM and k-means clustering solutions appear to perform the best. 

Key Discriminants in Alternative Segment Classifications.jpg

The final step in evaluating the alternative solutions was to study the key discriminants that were most important in segment classification. As shown above, the top discriminants in the k-means solution were disproportionately concentrated in various pain sites in hip, neck, lower back, and whether the person had received a diagnosis of arthritis, gout, lupus, or fibromyalgia; whereas the discriminants in the PAM solution was more evenly distributed across pain sites as well as other pertinent dimensions. Because we wanted the segments to touch on more even spread of discriminants that are relevant and useful in differentiating between different types of chronic pain patients, we will select the PAM solution to round up this analysis. 

Five segments of adults suffering from chronic joint pain have the following key attributes:

  1. "Diagnosed & Secure" segment have all received a clinical diagnosis of arthritis, gout, lupus, or fibromyalgia to which they could attribute their chronic joint pain; their treatment is working well as only 18% of them report limited in their daily activities due to chronic joint pain, despite the fact that this is the oldest segment (average age 63). They also report the lowest average scores on depression, anxiety, sleep problems, and financial stress; and the highest scores on trust in neighbors.
  2. "Self-reliant" segment are the least likely (only 1 in 5) to have ever seen a doctor for their chronic joint pain; and they are also least likely to report any limitations in daily activities due to chronic joint pain. They have a high incidence of concurrent back pain (71%) but otherwise relatively low on other comorbidities. Those who are employed report the highest level of work stress, resulting from harassment, lack of support from supervisor, or fear of losing their job. They report relatively high rates of physical exercise, and also relatively high rates of smoking and alcohol consumption.
  3. "Pervasive Pain" segment report the highest number of chronic joint pain sites - average of 8 sites compared to 2-3 sites in other segments. They also report the highest incidence of pain in most sites including shoulder, knee, hip, neck, lower back, and headache / severe migraine; and the highest scores on depression, anxiety, sleep problems, and financial stress; and the lowest scores on trust in neighbors. This segment has the highest incidence of heart conditions.
  4. "Disruptive Pain" segment have the highest number of people who report limitations in their daily activities due to chronic joint pain (87%) even though their average number of joint pain sites is only 3, mostly in the knees and lower back. They have next to highest scores on depression, anxiety, sleep problems, and financial stress; and moderate trust in neighbors. This segment also has the highest incidence of hypertension and high cholesterol. 
  5. "Active & Fit" segment talk to their doctors about their chronic joint pain, but less than 1% have received a clinical diagnosis of arthritis, gout, lupus, or fibromyalgia. They report the highest rates of physical exercise, including vigorous workouts and strength training. Knees are where they tend to experience chronic pain. It is likely that the chronic joint pain for some members in this segment is related to sports injuries.
Pain Segment Profiles by Prevalence of Comorbidities.jpg
Pain Segment Profiles by Physical Limitations x Age.jpg

Finally, we provide a few examples of how the differences across these 5 segments have implications for chronic pain patient messaging: 

"Diagnosed & Secure” segment messaging could focus on Adherence. These patients appear to be pretty well treated for their pain; however, they are also the segment clinicians would be most concerned about dropping off.  Since all of them have serious diagnoses of arthritis, gout, lupus, or fibromyalgia, they are likely to be on the most aggressive treatment regimens. And a growing body of evidence indicates that when patients start to feel better, some of them begin to slip off their regimens because they are feeling well. Therefore, the key message for this segment is adherence to therapy, as well as solutions to maximize patients' compliance with their prescribed treatment regimens.   

"Self-reliant" segment messaging could focus on Improved Access and Innovative Alternatives. This segment is clearly the most resistant to seeking medical treatment. Hence, one solution for these people could be more education about alternative treatment options – perhaps if they don’t want to see a medical doctor, they could go to a physical therapist to talk about their pain instead. Or chiropractic doctors or acupuncture doctors, or even meditation practice; all of which now have accumulating evidence on benefits for pain management. Access would also be helpful as they could be more inclined to see a healthcare provider who comes to their workplace and can talk to them there, without them having to make a separate trip to a doctor’s office. 

"Pervasive Pain" and "Disruptive Pain" segments messaging could focus on Counseling and Integrated/Holistic Care. Both of these segments have the highest unmet need, because they are still suffering greatly, either due to a high number of pain sites, or due to high degree of limitations in daily activities. These patients are the most viable targets for chronic pain medications, but more importantly, they are in need of supportive care including counseling, and social or community activities such as creative arts interventions. In addition, it is critical that these patients have consistent primary care that takes an integrated approach to treat their multiple chronic conditions. The solution for them should be a holistic one, since it is unlikely the chronic pain will go away with only medications. Deeper levels of treatment are needed to get to the real reason behind their pain.

"Active & Fit" segment messaging could focus on Physical Therapy and Injury Prevention. This segment is open to seeking medical treatment, and have NOT been diagnosed with any serious chronic condition that could account for their joint pain. Instead, the high rates of vigorous exercise suggest that their chronic joint pain is more likely due to sports injuries or otherwise overuse/soreness issues that often creep up on active people. This segment is therefore appropriate targets for shorter term medications, as well as training solutions available in physical therapy clinics, gyms, and other service providers that could teach them how to care for themselves and better prevent injury.

Source: Chang, LinChiat, Tucker Hurtado, and Susan Weber. 2017. Market Segmentation of U.S. Adults Suffering from Chronic Pain. Paper to be presented at the annual conference of the European Survey Research Association (ESRA) to be held in Lisbon, Portugal.

Data: All estimates are derived from weighted statistics from a nationally representative sample of U.S. adults who responded to the 2015 National Health Interview Survey (NHIS) conducted by the U.S. National Center for Health Statistics (NCHS), especially the 10,000 respondents who reported living with chronic pain in 1+ joints.