Understanding the High Prevalence of Low-Prevalence Chronic Disease Combinations: Databases and Methods for Research. 2. Introduction


Individuals with multiple chronic conditions (MCC) represent a growing percentage of the population. Some chronic diseases commence at birth, while others occur later in life, and they may be caused by genetic, behavioral, environmental, or infectious factors. Chronic diseases may become acute at times, may impair functioning or may be asymptomatic. While estimates vary by data source and methodology, those from the following studies are illustrative: in 2006, 28% of the population had MCC and by 2010 this increased to about 32% (RWJF, 2000 & Abt Associates, 2013). In 2010, 14% of Medicare beneficiaries with 6+ chronic conditions accounted for 46% of total Medicare spending (CMS, 2012). As individuals age, they are more likely to acquire MCC, however the rate of comorbidities is also increasing in the under 65 years-of-age population. As a high need population, the MCC cohort represents a large percentage of healthcare service utilization and cost. For example, persons with disabilities (the vast majority of whom have multiple chronic conditions) make up only 15% of the United States Medicaid population, but account for 43% of nearly the $350 billion per year in expenditures nationwide (Kaiser Family Foundation, 2009 & CMS, 2011).

Research on multiple chronic conditions has been scant in recent decades but is growing as the affected population increases. Understandably, current MCC research has focused primarily on studying the impact of high-prevalence diseases (i.e., hypertension, hyperlipidemia, diabetes, arthritis, etc.) in terms of patient outcomes, care utilization and cost. However, an understudied group comprises patients with less prevalent combinations of MCC. How this group may change over time as individuals acquire new chronic conditions, or certain conditions change in intensity, has only recently been examined. Overall, there are many unique constellations of MCC; for example, a recent study of approximately 32 million Medicare beneficiaries found over 2,000,000 unique disease combinations (Sorace et al. 2011). The distribution of constellations of diseases results in a curve with a very “long tail” of complex patients. Exhibit 1 depicts the beginning of Medicare’s long tail distribution. Sources and methods for studying the long tail are the primary focus for this white paper.

Exhibit 1: The Beginning of Medicare’s Long Tail: Prevalence of Top 250 Disease Combinations

Exhibit 1: The Beginning of Medicare’s Long Tail: Prevalence of Top 250 Disease Combinations

Constellations categorized as “rare” can result from combinations of common chronic conditions and/or less common or rare diseases. In other words, there are multiple pathways to becoming less prevalent (See Exhibit 2), and combining less prevalent combinations may account for as much as 79% of Medicare expenditures and 32% of beneficiaries (Sorace et al. 2011). Unique constellations are especially complex when multiple organ systems are involved and the combination of diseases, or treatments interact with one another. Developing treatment strategies for these complex patients is extremely difficult.

Exhibit 2: Multiple Chronic Condition Combination Types

MCC Types Example
Rare or less prevalent Condition in Combination with a Rare Condition Multiple sclerosis and schizophrenia
Rare or less prevalent Condition in Combination with a Moderately Common Condition Multiple sclerosis and lung cancer
Rare Condition or less prevalent in Combination with Common Chronic Conditions Multiple myeloma, hypertension and depression
Combinations of Moderately Common Chronic Conditions with Common Chronic Conditions Breast cancer, COPD, and arthritis
Unique Combinations of Common Chronic Conditions Hypertension, hyperlipidemia, chronic back pain, and depression

Acknowledging the “long tail” is important in interpreting the results of many types of healthcare studies. If the long tail is not accounted for, the following can potentially occur:

  • Quality measures may show skewed calculations due to inaccurately classified individuals. For example, a person with type 2 diabetes and Alzheimer’s disease may not be a good candidate for tight glycemic control.
  • Healthcare costs are inaccurately calculated. The patient with heart disease and MS may have all of their healthcare utilization and cost attributed to their heart disease when it is really a combination of the two or the majority of the cost is due to MS-related service utilization.
  • Randomized controlled trials (RCTs) may be designed inappropriately, causing the results to not be generalizable to non-experimental settings. For example:
    • The patients enrolled in the trial do not represent the comorbidities present in the actual patient population.
    • Complex patients may have higher attrition compared to other patients (e.g., MCC patients fall out of a study arm).
    • Investigators do not necessarily randomize for complexity or check to see if randomization has been successful for patients who may have MCC.
    • Even if complex patients are involved in RCTs, patients with different patterns of complexity will likely be encountered in the future, which may limit the generalizability and long term implications of results.
  • Disease management guidelines for a specific chronic disease may not work when combined together with other chronic conditions, and, in some cases, may contradict other guidelines (Boyd et al. 2005.)

View full report


"rpt_LowPrevMCCData.pdf" (pdf, 1.37Mb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®