Intended for healthcare professionals


Increased mortality among Dutch development workers

BMJ 1995; 311 doi: (Published 18 November 1995) Cite this as: BMJ 1995;311:1343
  1. Erik J Schouten, medical officeraa,
  2. Martien W Borgdorff, epidemiologista
  1. aHealth Net International, 1017 MB Amsterdam, The Netherlands
  2. aRoyal Tropical Institute, 1092 AD Amsterdam, The Netherlands
  1. Correspondence to: Dr Schouten.
  • Accepted 16 August 1995

Early this century Africa was known as the “white man's grave.” Little is known about health risks of expatriates in developing countries today. We compared the mortality of development workers who were sent out by Dutch development organisations with that of the general population in the Netherlands, adjusted for age and sex.

Subjects, methods, and results

Records of three large Dutch development organisations were reviewed. From 1984 to 1994 over 6500 development workers and spouses spent a total of 15144 years abroad. About 75% of them lived in sub-Saharan Africa. Causes of death were classified according to the International Classification of Diseases: mortality caused by traffic accidents (ICD 800-848), other injuries (ICD 880-959 and 980-989), homicide (ICD 960-969), and all other causes (ICD 001-799, 849-879, and 930-959).1 Age and sex specific mortality rates in the Dutch population were used to calculate the “expected” number of deaths in the study population. The observed number of deaths was divided by the expected number to obtain the standardised mortality ratio (SMR).

The table shows that mortality of development workers was 1.9 times that of the Dutch population, corresponding to an increase from an expected mortality of 1.1 to an observed mortality of 2.1 per 1000. The standardised mortality ratio in women was significantly higher than that in men (ratio of standardised mortality ratios 2.4; 95% confidence interval 1.1 to 5.1). A high standardised mortality ratio was found for traffic accidents, particularly in women. The observed increased risk for other injuries and homicide was not significant. Mortality in development workers from other causes was similar to that in the Dutch population. Mortality from AIDS, however, accounted for 3 (9%) out of 32 deaths.

Those with higher levels of education have a lower mortality.2 Mortality is reduced by 10% for men and 5% for women for each year of educational attainment in the age group 20-44 years (Anton Kunst, personal communication 1995). On average development workers and their spouses had 3.5 more years of education than the Dutch population. This difference would reduce the expected mortality by 31% for men and 16% for women. After correction for education the standardised mortality ratio would be 2.6 (95% confidence interval 1.8 to 3.7) for the total population of development workers and spouses, 2.0 for men (95% confidence interval 1.2 to 3.2), and 4.0 for women (2.2 to 6.6).


Dutch development workers had a mortality almost double that of the general Dutch population. The true increase in mortality was probably higher because of a healthy cohort effect (medical selection) and mortality after the end of a contract attributable to infections acquired abroad, but leading to death later, such as hepatitis, malaria, or HIV infection.

In a study among Dutch expatriates returning from sub-Saharan Africa 4 out of 1122 men and 1 out of 846 women were found to have HIV infection, which had probably been acquired abroad, giving an estimated incidence rate of 0.7/1000 person-years.3 As the observed mortality was 2.1 per 1000 during the contract period, HIV infection might increase mortality attributable to working in developing countries by one third.

Standardised mortality rates (SMR) in Dutch development workers from 1984 to 1994 by cause of death

View this table:

Traffic accidents were also an important cause of death for Peace Corps Volunteers4 and American missionaries.5 Among missionaries, however, the overall standardised mortality ratio was 0.5 because of a strongly reduced mortality from other causes of death. In particular, mortality attributable to cardiovascular diseases was much reduced, which was explained by a healthy cohort effect, better diet, and more physical exercise. In the relatively young Dutch expatriates cardiovascular diseases were not an important cause of death and the mean duration of contract is too short to expect major benefits from a change of lifestyle.

Efforts to reduce excess mortality among expatriates should be undertaken. In particular, development organisations and workers should explore measures to reduce the risk of traffic accidents and of HIV infection. Additional studies are needed on the health risks to expatriate workers and their families. Major areas not covered by the present study are mortality after return, mortality in children, and morbidity.


We thank Clarice Tjon, Joep Bremmers, and Jorien de Kort for helping in collecting data; Anton Kunst for commenting on an earlier draft of this paper; and Wil Dolmans for his important role in getting the study started.


  • Funding None.

  • Conflict of interest None.