Suicide Rates: Percentages and Rates, Age and Gender
The Centers for Disease Control and Prevention (CDC) recently released a report on suicide rates, finding that suicides in the United States had increased by 24 percent between 1999 and 2014. Like most people, I learned of this report after reading upsetting headlines about this increase. My local newspaper, the Los Angeles Times reported that "US Suicides Have Soared Since 1999."
As sociologists, we learn to look at the original data to get the real story beyond the headlines. What do the data tell us? Is it the same story as being told in news reports?
Percentages vs. Rates
That 24 percent increase really got my attention. When I looked at the original CDC report, I saw that the rate had changed from 10.5 suicides per 100,000 in 1999 to 13.0 in 2014. Any increase that represents a loss of life is important, and worth asking questions to learn more about how to address and prevent if possible. But note that a rate increase of 2.5 per hundred thousand is less dramatic sounding that 24 percent, yet they both represent the same change.
What's the difference between percentages and rates? We are more familiar with percentages, and they are easier to understand. Percentages, like the one in this report, are calculated by taking the change in rates from time one to time two (2.5) and dividing by the original rate (10.5) to get .24 which is then multiplied by 100, thus the 24 percent change.
Rates consider the likelihood of the event—in this case suicide—taking place within the population. To calculate the rate, you would divide the number of suicides in 1999 by the number of the people in the population. According to the CDC, in 1999 there were 29,199 suicides within a population of approximately 279 million. This comes to about .00010465, essentially the odds that any one person in the population will commit suicide. Because this is such a small number, we multiply by 100,000, giving us that rate of 10.5 per hundred thousand people in 1999.
Rates also control for population changes. In 2014, the CDC recorded 42,773 suicides, but the population also grew from 279 million in 1999 to about 319 million in 2014. The 2014 number of suicides represents many more deaths, but the rate allows us to accurately compare multiple time periods.
Rates are very instructive, but don't have the same emotional appeal that percentages do. And while it is not incorrect, the use of percentages evokes a different emotional reaction than reporting the less dramatic rate change.
Graphing Long-Term Trends
Seeing long-term trends graphically also helps us visually comprehend the meaning of the data. If you take a look at the graph below, you see that the increase during this time period has been gradual.
The shape of the slope of any line on a graph is shaped by the scale. For instance, the Y-axis above is separated in increments of five; if it were in increments of 10 the line would be flatter, and if the increments were smaller the slope would be sharper.
Gender and Age
One thing should stand out on the graph above: male suicide rates have been and remain much higher than for females. Using our sociological imaginations, we might examine this issue more deeply and conduct research to understand why this is the case. What sort of pressure might boys and men experience? Do boys and men have access to, and seek help from, mental health resources at the same rate as girls and women do? These are some questions that public health officials might consider in trying to prevent future suicides.
And yet a U.S. News & World Report headline, "Suicide Rate Triples Among Girls," creates the impression that suicide has become a bigger problem for young girls. Take a look at the graph below:
SOURCE: NCHS, National Vital Statistics System, Mortality.
1Significantly higher than rates for all other age groups (p < 0.05).
NOTES: For all age groups, the difference in rates between 1999 and 2014 is significant (p < 0.05). Suicides are identified with codes U03, X60–X84, and Y87.0 from the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision. Access data table for Figure 2.
Yes, the rate for ten- to fourteen-year-olds did triple from .5 to 1.5 per 100,000 (or 1 suicide for every million girls in 1999 to 3 in a million in 2014), but they are still the least likely group to commit suicide. Among women, 45- to 64-year-olds are the most likely to commit suicide, and that rate has grown (by 63 percent, if we want to use a dramatic figure for emphasis). As the note above explains, these changes are statistically significant, meaning that they are unlikely the result of chance.
Why might middle-age be the time when women's suicide risk peaks? What factors might have exacerbated these factors in recent years? We can't know for sure without further research, but we might ask what unique kinds of pressure women in this age group might face. Is it that women in this age group often face the pressure of caring for children—including financial support for young adult children— and aging parents (that's why middle-aged people are often called the "sandwich generation")?
Despite this rise, it is important to note that male suicide rates within each age group are higher than those of women, as detailed below. As with women aged 45 to 64, men in this age group also experienced a sharp rise in suicide (43 percent). Note that men in this group are about three times more likely than women to commit suicide. Whatever issues might lead people to take their life at this age might be exacerbated for men, and might point to the need for more research and outreach programs to support some of the challenges people face at this time in life.
1Significantly lower than rates for all other age groups (p < 0.05).
2Significantly higher than rates for all other age groups except 75 and over (p < 0.05).
3Significantly higher than rates for all other age groups (p < 0.05).
NOTES: For all age groups, the difference in rates between 1999 and 2014 is significant (p < 0.05). Suicides are identified with codes U03, X60–X84, and Y87.0 from the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision. Access data table for Figure 3.
One thing should stand out in the graph above: in both 1999 and 2014, the group of males most likely to commit suicide is the 75 and over age category, while suicide rates for women get lower as they pass age 65. Why might this be? What sort of mental health resources might we create for older men in particular?
If there is any good news in the new data on suicide, it is that suicide rates among both men and women age 75 and over have fallen. As I wrote about in a previous post , as our population ages, the emotional well being of the elderly—and working to prevent suicides among the elderly—will take on greater importance.