Status Plus




Exploratory cluster analysis of US males 40-70 years old focused on erectile dysfunction and associated risk factors, using National Health and Wellness Survey data

Goldstein, I1; Goren, A2; Liebert, R2; Tang, WY3; Hassan, TA3

1: Alvarado Hospital, United States; 2: Kantar Health, United States of America; 3: Pfizer Inc, United States of America

Objectives: Studies often focus on one patient profile element, risk factor, and/or condition in evaluating association with health outcomes. The real world can be more complex; patients have comorbidities, multiple risk factors, and differing profiles. This study aims (a) to identify natural clusters of male characteristics from a general population sample and (b) to see erectile dysfunction (ED) dynamics in these natural profiles.

Material and Methods: National Health and Wellness Survey (NHWS) 2016 patient-reported data on males age 40-70 (n=15,652) were analysed. Hierarchical agglomerative clustering identified clusters where categorical predictors included demographics, health characteristics/behaviors, ED risk (hypertension, dyslipidemia, diabetes, and depression), and provider visits in the past 6 months. Multinomial logistic regression assessed the independent utility of variables in predicting cluster membership compared with the “healthier” control cluster.

Results: Four natural clusters were found for men in this US sample 40-70 years of age (table below)

Cluster #1

Cluster #2

Cluster #3

Cluster #4

High ED (25.2%)

Lower ED (12.4%)

High ED (27.9%)

High ED (20.4%)

Elderly: M=63.8 years

Youngest: M=50.9 years

Older: M=56.5 years

Younger: M=53.5 years

Retired: 95.2%

Employed: 94.8%

Employed: 93.4%

Not Employed: 98.8%

Highly comorbid

Low comorbid

Highly comorbid

Highest comorbid rates

Low Smoking: 12.6%

Mod. Smoking: 15.1%

Low Smoking: 10.6%

High Smoking: 26.0%

Mod. Alcohol: 66.0%

High Alcohol: 73.8%

High Alcohol: 71.0%

Low Alcohol: 52.0%

Overweight+: 77.8%

Overweight+: 68.1%

Overweight+: 86.7%

Overweight+: 70.4%

High Exercise: 31.7%

High Exercise: 34.8%

Mod. Exercise: 22.8%

Low Exercise: 14.6%

In regression (Cluster #2 as ref), age, hypertension, dyslipidaemia, diabetes and employment remain predictors across membership to clusters. ED remained a predictor for all except the oldest cluster (#1).

Conclusions: The four natural clusters show unique prevalent profiles among US adult males 40-70. Sustained rates of ED found within these natural clusters, even despite differences in previously identified protective factors such as younger age, drives attention to multifactorial considerations in evaluation of ED risk.


Work supported by industry: yes, by Pfizer (industry initiated, executed and funded study). The presenter or any of the authors act as a consultant, employee (part time or full time) or shareholder of an industry.

Go Back