San Francisco Transgender Benefit: Estimating Cost and Utilization for the “Worst Case” (1997-2001)
In 2001, the City and County of San Francisco ("San Francisco") became the first major U.S. employer to publicly remove discriminatory transgender access exclusions in its health insurance plans for employees, retirees and their dependents in order to explicitly cover medically necessary treatment for transgender transition. See our complete set of resources on San Francisco's Transgender Benefit.
In 2001 as today, very little cost and utilization data exists to help insurance experts anticipate the impact of providing inclusive insurance benefits. As a result, San Francisco's "transgender benefit" was launched in 2001 for the self-funded, Preferred Provider Organization (PPO) plan, known as the "City Plan," and was designed based on rough, nonscientific data reviews dating back to 1997 using "worst case" assumptions to counter fears of unknown demand and high costs.
We now know that these fears were unfounded, that initial cost projections were vastly overestimated — on a magnitude of at least 35 times over the actual costs. Further, the actual costs and utilization in San Francisco demonstrate that inclusive benefits are actuarially insignificant for large health plans:
Breaking Down San Francisco's Projections
Leading up to San Francisco's transgender benefit in the City Plan in 2001, actuaries attempted to estimate total costs for transitioning individuals with conservative estimates using unscientific, rough estimates of actual utilization data from the Canadian province of British Columbia, where services were covered under the province's public health insurance plan. The actuarial estimates recognized some understanding of the medical needs of transitioning people, but often fell short.
Several interrelated "unknowns" needed to be addressed:
- Number of claimants — how many people would need these services? How many would file claims in any given year?
- Cost of claims — how much would such services for any given individual amount to in their lifetime? How much would total costs for annual claims amount to?
Number of Claimants
San Francisco's actuaries used rough data from British Columbia to estimate the minimum number of claimants San Francisco could expect, but these conservative estimates were vastly overstated.
Describing how actuarial firm Towers Perrin built concerns about a "magnet effect" and "worst case" utilization into their official cost predictions, actuary Michael Kramer explained it was his firm's fiduciary duty to make estimates on the high side, "so the health care system isn't caught off guard." "The variable that sort of dwarfs all other variables is the unknown demand... It's possible that people wanting sex-change [sic] services could be drawn to San Francisco and seek a city job to pay for them."
To structure coverage conservatively, San Francisco's projections utilized rough data indicating that 50 people had accessed services, but also assumed this derived from a smaller-than-actual population of 1 million people, resulting in a roughly-estimated rate of 0.05 claimants per thousand adults.
Reexamining British Columbia "Data"
While there are significant questions regarding the data utilized for San Francisco's actuarial studies, we know that the reference population of 1 million represented a subset of adults. We also know that the actuaries assumed this was an annual rate, although no information is available to indicate whether the British Columbia estimates were over a period of multiple years or a single year.
Census data shows that, in 2001, British Columbia had an adult population of over 2.4 million and a total population of 3.9 million. The reference population of 1 million adults was probably those aged 20-39. Thus, a more accurate estimate would have been that 0.021 claimants per thousand adults or 0.013 claimants per thousand people had accessed services.
|Age Group||Number of Claimants||Population||Claimants per thousand|
|Adults (aged 20-39)||50||1 million||0.05 (used by SF)|
|All Adults||50||2.4 million||0.021 (more accurate)|
|Total population||50||3.9 million||0.013|
In 2001, San Francisco had somewhere between 25,000 and 37,000 employees (the total numbers of insured individuals including family members and retirees are variously reported). There were reportedly 12-17 known transgender employees, and actuaries assumed that for every known transgender employee there were 1-2 unknown transgender employees.
Utilizing the rough British Columbia rate of 0.05 cases per thousand individuals and making the assumption that San Francisco would likely have a higher incidence of transsexualism than British Columbia, actuaries originally anticipated a worst-case of 35 people might use the San Francisco transgender benefits program each year. With an assumed employee base of just over 37,000 and a total enrollee base of 80,000, this translates into a worst-case, predicted rate of
- 0.946 claimants per thousand employees per year (45 times the rate of 0.021 claimants per thousand adults in British Columbia), or
- 0.438 claimants per thousand enrollees per year (20 times the rate of 0.021 claimants per thousand adults in British Columbia).
Compare our calculated rate for all adults in British Columbia (0.021 per thousand) to the overestimates used to structure San Francisco's coverage in 2001. The worst case utilization of 0.946 claimants per thousand employees was 45 times the rate for British Columbia. More conservatively, the overestimated rate of 0.438 claimants per thousand enrollees was still greater than 20 times that of British Columbia.
Hedging Against a Magnet Affect
To limit the possibility that transgender people might deliberately seek employment or coverage under the plan (the feared "magnet effect" causing "adverse selection") actuaries also proposed limiting the availability of the benefit to only those employees, retirees or dependents who had been members of the plan for more than one year. This was later revised - see San Francisco Transgender Benefit: Actual Claims Experience and Plan Changes (2001-2006).
Costs of Services / Lifetime Maximum Dollar Cap
In 2001, the cost of services in British Columbia ranged from $4,000 to $75,000, depending on what was clinically appropriate and necessary for the person.
Actuaries also presumed that claims for the known transsexual employees would be low because some were already "in various stages of treatment and [would] not need to undergo every procedure were these benefits available." While a few people might have costs in the higher range, most total costs could be expected to be much lower, and other people might utilize no services at all.
Estimating that costs might run as high as $37,000 for male-to-female procedures and $77,000 for female-to-male procedures, actuaries suggested a lifetime cap on costs per person of $50,000.This was also later revised - see San Francisco Transgender Benefit: Actual Claims Experience and Plan Changes (2001-2006).
To offset the total projected costs, the actuaries further suggested that all City Plan participants be charged an additional $1.70 per month. Because employee and retirees could select a plan that did not include the transgender benefit, the actual amount that would be raised by this measure was uncertain. The actuaries must have assumed that all those requiring these services would shift to the City Plan, but fertility and other benefits had also been added at the same time and the impact on enrollment was still speculative.
If the worst-case scenario occurred, and 35 individuals utilized the maximum $50,000 benefit each year, the City Plan would require $1.75 million to cover services each year. If all 80,000 employees, retirees and dependents paid $1.70 monthly surcharge, San Francisco would collect over $1.63 million annually, just shy of the "worst case," but enough to cover more than 32 individuals at the maximum benefit.
Shortcomings of the 2001 Estimates
Clearly, not every person requiring surgical interventions would require the most expensive interventions, and of those who did few would be likely to reach the maximum dollar amount in one year. The 1997 report by the San Francisco Human Rights Commission articulates these arguments. Given there were at most 17 known transgender employees at the time, that not all would require surgery, and that cost estimates were based on the maximum dollar value of the range of possible services, these estimates were overinflated. But by how much? Only with time and transparent reporting of utilization would anyone know for sure. Nevertheless, the number and magnitude of claims were in reality far less than originally projected by actuaries - see San Francisco Transgender Benefit: Actual Claims Experience and Plan Changes (2001-2006).
 Gordon, Rachel. "S.F. Set To Add Sex Change Benefits City would be first to include option." San Francisco Chronicle. Feb. 16, 2001. Available at: http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2001/02/16/MN202072.DTL
"In most cases for all other surgeries for city workers, the full cost is covered under the health insurance plan. The discrepancy isn't lost to transgender activists and their supporters. They're also concerned that a city worker will have to be on the payroll for at least a year before the sex- change benefits can be used, whereas all other medical procedures are available immediately. The restrictions were added to limit the potential financial burden on the the Health Service System fund, which provides medical insurance for 37,000 workers employed by the city, the San Francisco Unified School District, the community college district and 17,000 retirees."
 The actual numbers cited here are as reported by Human Rights Commission staff by email communication Dec. 22, 2009. News reports and insiders indicate that Towers Perrin performed the initial research, however it is unclear whether a written report was made. We do not know whether the number "50" represented new cases each year, or whether it included all transition services or only surgical cases. Census for British Columbia in this period indicates 2.4 million adults, and we do not know what smaller age range of adults was used to reduce the total to 1 million. However, the 2001 census suggests that the age brackets between 20 and 39 totaled just over 1 million people.
 Towers Perrin actuaries apparently assumed there were 1 million adults in British Columbia, which was a gross underestimation. Total population 3,907,738 in 2001. http://www.bcstats.gov.bc.ca/data/cen01/c2001hl.asp. Of these 61.4% or 2,399,352 were between the ages of 20 and 65. http://www.bcstats.gov.bc.ca/data/cen01/facts/cff0101.PDF
 Insurance Coverage for Transsexual Employees of the City and County of San Francisco. (October 1997). "Potential costs will be ameliorated by the fact that: (1) some City and County employees for whom these treatments are applicable are known to be in various stages of treatment and do not need to undergo every procedure were these benefits available. Also, (2) in some instances, coexistent medical conditions such as HIV disease, hemophilia, liver disease, extreme hypertension, and certain blood disorders, will preclude individuals from undergoing certain aspects of the sex reassignment treatment." (p.5)
Earlier in the document: "In considering these costs, it should be noted that some City and County employees for whom these treatments are applicable are known to be in various stages of treatment and do not need to undergo every procedure were these benefits available. To date, we know of only 12 transsexual individuals who are employed in the City and County of San Francisco, among the 27,000 employees." Note that news articles in 2001 cite the figure of 17 known employees, and an employee base of 37,000 employees.