Nov. 18, 2004 — My reaction to the article on smoking rates was instant suspicion. (See "V.I. Has Lowest Smoking Rate in the Nation").
How could the V.I. have the best non-smoking rate in the nation when on so many other health and socioeconomic variables it comes out at or near the bottom?
Using the same Centers for Disease Control data set as the one that dealt with smoking I found that the V.I. is 53rd of 54 jurisdictions when it comes to the percent of pregnant mothers having pre-natal care in the first three months of pregnancy. The national average is 83.7 percent and in the territory it is 64.0 percent. Similarly, the V.I. is 54th on the matter of the percentage of 65+ people getting flu shots in the previous 12 months; the U.S. average is 66.4 percent and in the islands it is 32.2 percent.
But when it comes to smoking the CDC report shows the V.I. with a 10 percent rate, compared to 22.1 percent among adults nationwide.
While I am neither a public health specialist nor a trained statistician I have been conducting or analyzing surveys on many subjects over the last 35 years. For example, in 1969 I conducted the first survey of illegal aliens ever funded by the U.S. government, and subsequently conducted surveys of various groups of immigrants, refugees and farm workers. I have worked extensively with sampling methodologies and survey-instrument design: I know that you want to capture a good random sample of those you are studying, and you want to ask questions in a totally neutral manner. Because of this I think I can spot an unlikely set of responses.
The CDC data is based on a sample survey of adults in each of the states and territories.
After checking several other public health indicators, as noted above, I did what the professionals do in the business, I looked at the "internals" of the smoking survey; in other words, when you break out the surveyed population into segments, what do you find?
What I found was remarkable. These are the ethnic breakouts of adult smoking prevalence in the U.S. as a whole and in the Virgin Islands:
(Ethnic Group, U.S. Smoking Rate, V.I. Rate)
Whites, 23.3%, 31.7%
Blacks, 21.7%, 4.0%
Hispanics, 18.5%, 11.0%
While in every state of the Union there is some difference between black and white smoking rates, with many states showing higher white rates than black ones, and some showing the reverse, these rates are all relatively close together, as the national averages are in the little table above.
But not in the Virgin Islands, where whites are shown as being eight times more likely than blacks to smoke. Does one see this on the streets in the islands? I doubt it.
The range of black adults smoking in the 50 states is from a high of 36.2 percent in tobacco-rich, if low-income, Kentucky to 15.6 percent in Massachusetts. But it is only 4.0 percent in the Virgin Islands, which is hard to believe.
The other ethnic distributions in the V.I. are unusual compared to those of the mainland, but not outlandish. The range of white smokers is from 32.2 percent in Kentucky (again) to 12.3 percent in Mormon-dominated Utah, with the V.I. whites being second to those in Kentucky, at 31.7 percent. Nationally Hispanic usage rates ranges from to 47.5 percent (Kentucky) to 13.0 percent (Hawaii) with V.I.s rate being only 11.0 percent,
But something is either very, very different about the incidence of smoking among V.I. blacks, or, more likely, something is wrong with the survey.
Meanwhile, as I scrolled through reams of CDC statistics on public health, a disturbing element showed itself.
As I sought public health comparisons of the V.I. to the 50 states, as shown above for pre-natal care and flu shots, I found, again and again, that while data were collected for all 50 states, and sometimes for the 50 states and Puerto Rico, and in one instance for the 50 states, Puerto Rico and Guam there was nothing for the V.I.
In an admittedly casual sampling of the CDC's many data sets, I found no V.I. information on the following significant public health subjects: infant deaths, child immunization, mental health expenditures, diabetes prevalence, heart disease deaths, rate of motor vehicle deaths and Pap smear rates but rates were always available for the mainland, and in four instances for the mainland and Puerto Rico.
Is the CDC funding these studies in the 50 studies and in Puerto Rico, but not in the V.I.?
Does the CDC fund infant mortality rate studies on the Mainland, in Puerto Rico and in Guam, but does it deliberately ignore the V.I.? Or, more likely, is some island variable at work here?
Perhaps this is a subject for another article.
Editor's note: David S. North, a semi-retired former federal employee living in the Washington, D.C., area, reports and writes frequently for the Source on government affairs and economics.
We welcome and encourage readers to keep the dialogue going by responding to Source commentary. Letters should be e-mailed with name and place of residence to firstname.lastname@example.org.