On the Calculation of the Prevalence of Transsexualism

1
On the Calculation of the Prevalence of
Transsexualism

• Prof. Femke Olyslager1, PhD, and Prof. em. Lynn
  Conway2
• 1 Department of Information Technology,
• Ghent University, Ghent, Belgium
• 2 Department of Electrical Engineering and
  Computer Science,
• University of Michigan, Ann Arbor, Michigan,
  USA

2
Introduction

• Introductory remark
• Only male-to-female transsexualism is discussed
  here!
• The key-question
• What are the chances that my child is
  transsexual?
• The prevalence of transsexualism.
• Observation
• Literature(1) P 1 in
  11,900
• In discrepancy with simple sanity-checks!(2)
• Actions
• Analyse the old reports on prevalence
• Develop a mathematical framework
• Derive commensurable prevalence numbers
• (1) A. Bakker et al., The prevalence of
  transsexualism in the Netherlands, Acta Psych.
  Scand., 1993.
• (2) L. Conway, How frequently does
  transsexualism occur?, LynnConway.com, 2001.

3
Deriving prevalence

• How was prevalence derived?
• Problem A
• E.g. the Netherlands(1,2)
• Prevalence increases over time, as more patients
  treated.
• Problem B
• Who is counted?
• Those satisfying TS definition, those
  on hormone therapy, those having SRS, ... ?
• Leads to incommensurable numbers.
• (1) A. Bakker et al., The prevalence of
  transsexualism in the Netherlands, Acta Psych.
  Scand., 1993.
• (2) P.L.E. Eklund et al., Prevalence of
  transsexualism in the Netherlands, Brit. J.
  Psych., 1988.

P
4
Deriving prevalence

• Types of prevalence
• P(TS) the prevalence of transsexualism
• P(SH) the prevalence of transsexual people who
  have sought help
• P(HT) the prevalence of those on hormone
  therapy
• P(ST) the prevalence of those who have
  socially transitioned
• P(SRS) the prevalence of those who have
  undergone SRS
• Inequalities (in general)
• P(TS) gt P(SH) gt P(HT) gt P(ST) gt P(SRS)
• or
• P(TS) gt P(SH) gt P(ST) gt P(HT) gt P(SRS)
• Allows e.g. to derive P(HT) from P(SRS) if
  P(SRS)/P(HT) is known
• Provides lower bounds for P(TS)

5
Deriving prevalence

• Suppose transsexuality were visible at birth
• Suppose transsexuality is always identified at
  some point in life
• Assumes constant demographics
• Assumes regime situation for identified
  individuals
• Correction for non-constant demographics
• L average age of identification
• Note This is a different from recurring
  conditions such as the flu, where

6
Inherent vs. active prevalence

• Inherent prevalence
• P(TSI) 1 in 4,000 population 4,000,000
• 1,000 identified transsexuals? No!
• Age of identification L 35 years life
  expectancy E 70 years
• 500 identified transsexuals!
• Active prevalence
• Relation between active and inherent
  prevalence
• L average age of identification
• E life expectancy

7
An example

• Prevalence as given in Bakker et. al.(1)
• Inherent prevalence
• 34 new patients annually (1,2)
• Average age to start HT L 32 years
• Male births in 1990 32 1958 (3) 120,000
• Active prevalence
• If life expectancy E 75 years, then
• (1) A. Bakker et al., The prevalence of
  transsexualism in the Netherlands, Acta Psych.
  Scand., 1993.
• (2) P.L.E. Eklund et al., Prevalence of
  transsexualism in the Netherlands, Brit. J.
  Psych., 1988.
• (3) Central bureau for statistics in the
  Netherlands, http//www.cbs.nl.

8
Other methods

• Other methods to estimate the prevalence of
  transsexualism
• triangulations and sanity-checks
• Example 1 Fellows of the IEEE(1)
• 5,500 Fellows (mostly males)
• 3 Fellows had SRS
• Estimate
• If P(SRSA) 1 in 11,900 then only 1.6 chance to
  have 3 cases
• (1) IEEE Institute for Electrical and
  Electronics Engineers.

9
Other methods

• Example 2 SRS surgeries on US residents(1)
• Over 1,500 SRS on US residents
• over 1,000 SRS in US over 500 SRS elsewhere
• 2,000,000 US male births per year
• Since P(TSI) gt P(SHI) gt P(HTI) gt P(STI) gt
  P(SRSI)
• P(TSI) gt 1 in 500
• (1) L. Conway, How frequently does
  transsexualism occur?, LynnConway.com, 2001.

10
Other methods

• Example 3 estimates based on Thai Kathoey
• Counting Kathoey (1) 6 per 1,000
• P(STA) 1 in 167
• Demographic profile of Kathoey (2) 27.7 of
  Kathoey had SRS
• P(SRSA) 0.277 x P(STA) 1 in 600
• Average age of SRS L 24.1 years Life
  expectancy E 73 years
• P(SRSI) x P(SRSA) 1 in 400
• Demographic profile of Kathoey (2) 48.3 of
  Kathoey who had no SRS want SRS
• P(TSI) 1 in 200
• (1) S. Winter, Counting Kathoey, Transgender
  Asia papers, 2002.

11
Conclusions

• Past reports failed to answer the key question
• What are the chances that my child is
  transsexual?
• Analysis of past reports
• P(SRSI) 1 in 4,500 to 1 in 2,000
  P(TSI) 1 in 2,000 to 1 in 1,000
• Estimates from more recent reports
• P(TSI) 1 in 500
• Transsexuality is not that uncommon!
• Important for transgender health care and
  recognition by society!
• Further reading (draft hardcopies available from
  authors)
• F. Olyslager and L. Conway, On the Calculation
  of the Prevalence of Transsexualism, submitted
  to the International Journal of Transgenderism.