Introduction and Objectives: Cluster analysis is a statistical method for categorizing groups of individuals with similar properties or characteristics. This method is useful for examining the relationship between complex urological problems and other health conditions or lifestyle factors. The method has previously been shown to differentiate men with mixed patterns of voiding and storage symptoms in a large Canadian sample . Assessing symptom co-occurrence can increase understanding of the disease etiology and has patient management implications. We used this method to classify men with storage, voiding, and post-voiding symptoms in the Boston Area Community Health (BACH) Survey .
Methods: The BACH Survey is a racially and ethnically diverse sample (n = 2301 men) of community-dwelling residents of Boston, Massachusetts aged 30 to 79. Fourteen urologic symptoms measured by participant self-report were used in this analysis. Cluster analyses were conducted using hierarchical and nonhierarchical (k-means) methods. Within clusters, demographic characteristics and risk factors for urologic symptoms, severity (measured by frequency of occurrence), and the interference of symptoms with daily activities were also assessed.
Results: Seventy percent (70%) of men in the BACH sample reported 1 or more urologic symptom. Among these symptomatic men, 5 statistically distinct symptom clusters were identified. Approximately half of the symptomatic men were designated as Cluster 1, which included individuals with a low prevalence and frequency of urologic symptoms and a correspondingly low level of interference with usual activities. Intermediate levels of symptom frequency and prevalence were noted in Clusters 2 through 4, which were characterized by mixed patterns of voiding, storage, and post-voiding symptoms. Cluster 5 consisted predominantly of older men (mean age: 58.9), with a high prevalence and frequency of urologic symptoms (mean number of symptoms = 9.9 + 2.1) from all 3 categories, and with markedly elevated levels of psychosocial distress and comorbid cardiovascular disease (P < 0.0001). These men also had higher rates of kidney and bladder infections and prior urologic surgery. Men with increased waist circumference and more sedentary lifestyles were over-represented in the more symptomatic clusters.
Conclusions: Five clusters of urologic symptoms were identified among symptomatic men in the BACH Survey. Clusters showed consistent co-occurrence of storage, voiding, and post-micturition symptoms. Increasing frequency and severity of urologic symptoms was associated with psychosocial and cardiovascular comorbidities. These results are consistent with previous findings by Norman et al.  and provide an empirically based classification for examining etiologic mechanisms, risk factors, and comorbidities associated with common urologic symptoms in aging men. Clinical implications include the need for a more comprehensive approach to assess and treat men with common urological symptoms.
KEYWORDS: Lower urinary tract symptoms (LUTS), cluster analysis, heart disease, diabetes, hypertension
- Norman, R. W., et al. (1994). Br J Urol 74: 542-50.
- McKinlay, J. B. and C. L. Link. (2007). "Measuring the Urologic Iceberg: design and implementation of the Boston Area Community Health (BACH) Survey." Eur Urol 52: 389-396.
Supported by: Grant, D. K. 56842, from the National Institute of Diabetes, Digestive and Kidney Diseases (NIDDK). Additional funding support by Pfizer, Inc.