ANALYSIS OF THE WORLD HEALTH SURVEY TOBACCO DATA

1
ANALYSIS OF THE WORLD HEALTH SURVEY TOBACCO DATA
S.M. Halberstadt, K.M. James, M.A. Riedesel,
R.R. Rozeske II, J.L. Winslow, S.J. Wood, J.M.
Legler St. Olaf College, Northfield, MN 55057
INTRODUCTION The objectives of this project
were to obtain standardized smoking rates from
data provided by the World Health Survey (WHS),
launched by the World Health Organization (WHO)
in 2002. The WHS was intended to gather
first-time, comparable data on health risk
factors and demographics from developing
countries. In conjunction with high cholesterol,
high blood pressure, obesity, and alcohol use,
tobacco usage contributes to over a third of
Cardiovascular Disease cases worldwide (Mackay,
Mensah, Mendis, Greenlund, 2004). Using
standardized tobacco-use rates, we are able to
compare rates between countries despite
demographic differences. The information provided
by the WHS has the potential to suggest health
initiatives aimed at reducing tobacco use in
developing countries.

• RESULTS
•  Overall trends are consistent between
  non-standardized smoking rates, and weighted,
  standardized smoking rates
• Eastern Europe (Hungary, Turkey, Georgia, Russia,
  Bosnia, Croatia) and Southeast Asia (Bangladesh,
  Laos, Nepal) exhibit much higher smoking rates
  than did African and Latin American nations.
• Non-smoking rates are greater than occasional
  smoking rates and daily smoking rates in all
  countries.
• The only country in which the occasional smoking
  rate is greater than the daily smoking rate is
  Mexico.
• Male daily smoking rates are greater than female
  daily smoking rates in all countries.
• In some Eastern European countries (e.g. Bosnia
  Herzegovina, the Slovak Republic, Croatia, and
  Hungary) there is a noticeably smaller difference
  between the male and female occasional smoking
  rates compared to other countries.
• In Morocco, Ghana, and Ethiopia the female daily
  smoking rates are nearly zero.
• In Morocco, Ethiopia and Tunisia the female
  occasional smoking rates are nearly zero.

•  
• DESCRIPTION OF WORLD HEALTH SURVEY
• The WHS includes two different surveys an
  individual survey focusing on personal
  demographics and a household survey focusing on
  communal demographics.
• The WHS individual survey also collected data
  concerning major health risk factors for chronic
  diseases. Some of these risk factors include
  high blood pressure, high cholesterol, obesity,
  alcohol consumption, low fruit and vegetable
  intake, physical inactivity, and tobacco use.
• Data from the tobacco section of the survey
  (Section 4000), as well as demographic data
  (Section 1000), were analyzed for 37 of the
  participating countries.
• Tobacco related questions
• Q4000 Do you currently smoke products such as
  cigarettes, cigars, or pipes? and provides three
  possible answers 1. Daily, 2. Yes, but not
  daily, and 5. No, Not at all.
• Q4001 For how many years are you smoking
  daily?
• Q4002-Q4005 On average how many of the
  following products do you smoke each day? and
  ask for the number of daily smoked manufactured
  cigarettes, hand-rolled cigarettes, pipefuls of
  tobacco, and other forms of tobacco not
  mentioned.

• DISCUSSION 
• The same overall trends are found both with and
  without the standardization of smoking rates.
  However, standardization by age and gender is
  still important, because it allows meaningful
  between-country comparisons. Differences between
  rates before and after standardization are
  roughly between 2 and 3 percent, which, in some
  cases, could mean a two-fold increase in daily
  smoking rates.
• Differences in smoking rates between the
  thirty-seven countries participating in the WHS
  may be due to differences in the cultural and/or
  economic characteristics of these countries
  including lifestyle choices and status symbols.
  Regional differences could be due to trade
  agreements, Western influence, and political
  networks. In general, Western relations with
  Eastern Europe and Southeast Asia are far more
  powerful than the links between the Western world
  and Africa, which could be why smoking prevalence
  is much higher in nations with greater proximity
  to the West.
• Differences in daily and occasional smoking rates
  between males and females may be attributed to
  disparities in social acceptability of smoking
  between genders, especially in developing
  countries. Response bias may also be present
  because of such explicit or inexplicit social
  rules.

• METHODS 
• Stata was used throughout the cleaning and
  analyzing process and a do-file was kept to
  document steps and commands for future use.
• Cleaning Process
• Cleaning improved validity of the data by editing
  illogical responses and allowed for complete
  analysis.
• Logical commands were used to eliminate unsound
  responses.
•  
• Direct Standardization with Survey Weights
• Standardization with applied survey weights
  allowed for valid cross-country comparisons of
  tobacco daily, occasional, and non-smoking rates.
• Survey weights from the WHS, under the variable
  pweight, were used. Pweight is the inverse
  of the probability that an individual is selected
  for the survey.
• The WHO Standard Population is used to directly
  standardize the tobacco data. WHO Standard
  Population is a representation of the worlds
  population by age range which, when applied,
  gives countries identical population distribution
  by age and gender.

Weighted Age-Sex Standardized Smoking Rate by
Country
Figure 1 Daily, occasional, and non-smoking
rates standardized by age and sex across
countries using the WHO Standard Population.
REFERENCES  World health survey. (2002).
Retrieved Jan. 19, 2005, from www3.who.int/whs
 Mackay, J., Mensah, G. A., Mendis, S.,
Greenlund, K. (2004). The atlas of heart disease
and stroke. Geneva, Switzerland World Health
Organization.
ACKNOWLEDGMENTS The WHO Chronic Diseases Health
Promotion Unit Jackie Lippe, Regina Guthold,
Tomoko Ono, Kate Strong, Ju Yang St. Olaf College
MSCS Dept Julie Legler
Fig. 2 Gender specific daily smoking rates
standardized using the WHO Standard Population.
Weights (pweights) were applied to smoking rates
as provided by the WHS survey data.
Fig. 4 Gender specific non-smoking smoking
rates standardized using the WHO Standard
Population. Weights (pweights) were applied to
smoking rates as provided by the WHS survey data.
Fig. 3 Gender specific occasional rates
standardized using the WHO Standard Population.
Weights (pweights) were applied to smoking rates
as provided by the WHS survey data.