Materials and methods
Randomly selected samples of men 18 to 40 years of age were surveyed in each
country. This survey was a part of a larger omnibus survey that is routinely conducted in
the UK and Italy every 2 to 4 weeks. The demographic characteristics of the sample in
each country were comparable to that of the general population in that country. A
personal interviewer identified participants, and the respondent then completed a
questionnaire in his primary language without any assistance from the interviewer.
The questionnaire was initially developed in the United States in American English to
evaluate the impact of hair loss on men's health-related quality of life. The instrument
was translated from American-English into French for use in France, German for use in
Germany, Italian for use in Italy and British-English for use in the UK using the standard
forward-backward method [9-11]. Prior to data collection, each translated version of the
questionnaire was pilot-tested for comprehension  with 10 men in each country.
Men rated their hair loss using a categorical scale with seven responses: "a full head of
hair", "only a little hair loss", "some hair loss", "moderate hair loss", "a good bit of hair
loss", "a lot of hair loss", and "I am bald". In addition, men circled the classification that
they felt best matched their hair loss pattern, using the Norwood/Hamilton scale [1, 2].
Men were also asked questions that were specific to hair loss, including previously
validated measures  of satisfaction with their hair appearance (on top of their head,
frontal hairline, and overall), degree of bother due to hair loss, extent of concern about
aging, and perceived noticeability to others . A "hair loss distress" domain score was
calculated as the sum of responses to questions regarding bother and concern about
looking older due to hair loss, whereas a domain score for thinning/shedding included
three questions pertaining to shedding while grooming . Demographic data,
respondent's medical history and data about hairstyle, hair color, hair length and family
history of disease were also collected. Other questions, identified as most relevant in a
previous survey , were included on an exploratory basis to investigate the relationship
between hair loss and worry about appearance and relationships, helplessness, self-
confidence and self-consciousness, preoccupation, impact on social life and job
opportunities, and other pertinent areas specifically related to hair loss. The response
scales of these questions were modified to better conform to the rest of the present
survey. Additionally, data were collected about men's awareness of hair loss products
and remedies, such as surgery, drug therapy, wigs, and hair styling.
Respondents were categorized into one of the four groups based on their self-reported
degree of hair loss: "a full head of hair", "a little hair loss", "some hair loss" or "moderate
hair loss", and finally, "a good bit of hair loss", "a lot of hair loss" or "bald".
Differences among men with various degrees of hair loss were assessed using general
linear models (GLM) . Covariates in the model included marital status (never
married, married/cohabitating/widowed/divorced), age (18-25, 26-30, 31-40), education
level (¾ high school, > high school), and interactions of hair loss and age. Analyses
were performed separately for each country. Bonferroni adjustments for multiple testing
were made for all statistical tests . Spearman rank correlations were used where
appropriate to investigate relationships between individual variables.