Jump to comment:

• Quantitative analyses, ethical quandaries, and policy debates: A rejoinder
  Alok Bhargava
  Published on: 5 August 2013

• Published on: 5 August 2013

  Quantitative analyses, ethical quandaries, and policy debates: A rejoinder

  • Alok Bhargava

  Javier Hidalgo's response[1] to my commentary[2] was unsatisfactory and is likely to mislead the readership of JME. First, biomedical journals often discourage authors from citing unpublished studies. After reading Hidalgo's response, one can see the wisdom of that rule. He quotes several incorrect assertions made by Michael Clemens in 2007 in an unpublished paper[3] about my article with Frederic Docquier.[4]

  Se...

  Show More

  Javier Hidalgo's response[1] to my commentary[2] was unsatisfactory and is likely to mislead the readership of JME. First, biomedical journals often discourage authors from citing unpublished studies. After reading Hidalgo's response, one can see the wisdom of that rule. He quotes several incorrect assertions made by Michael Clemens in 2007 in an unpublished paper[3] about my article with Frederic Docquier.[4]

  Second, Hidalgo repeats the misleading quote from page 6 of Clemens[3] in his response: "Bhargava and Docquier note that the fraction of physicians abroad has a positive and significant effect on the number of adult deaths due to AIDS in general, while - interestingly--- it has a negative and significant effect on AIDS deaths in countries where HIV prevalence is low." In fact, I had spelled out how the net effect of physician emigration rate on adult deaths due to AIDS is computed in non-linear models[2] and stated that: "Thus, the net effect at the sample midpoint was positive; net effect was very close to 0 (-0.005) when computed at the start of the sample in 1991." Thus, Clemens is wrong about the negative effect being significant-- it is not statistically different from zero at the start of the sample. In fact, to show that the negative net effect is statistically significant for countries with low HIV prevalence rates, one would have to evaluate its standard error that in turn depends on standard errors of two other estimated coefficients and their sample covariance.[2] Such computations could not have been performed by Clemens since the covariance between the estimated coefficients was not reported by Bhargava and Docquier.[4] Further, the net effect was positive at the sample midpoint where the expectation should be evaluated in such models. This is because the net effect is a function of the random variable HIV prevalence rates that varies across countries and over time. Thus, evaluating the net effect at the sample midpoint is a reasonable approximation for its mathematical expectation.

  Third, I had made several points about the importance of micro data that are essential for "integrating the epidemiological evidence from developing countries with the results from analyses of country-level data". One of the problems in policy oriented research is that some authors may not be familiar with quantitative analyses necessary for extracting information from longitudinal data sets covering heterogeneous individuals over time. While it is easy to base one's claims on analyses of country-level data, the ethical quandary is whether to incorporate the findings from elaborate micro studies that may support or contradict the assertions. For example, I had cited our study in South Africa[5] showing the benefits of uptake of healthcare services for AIDS patients' CD4 cell counts and quality of life indicators. There are numerous vacancies for nurses and physicians in South African clinics because many staff members have left for more lucrative venues in OECD countries. For a policy debate to be constructive, it is essential that the available evidence be examined. It would have been helpful if the micro evidence especially from sub-Saharan African countries received greater prominence in this discussion.

  Finally, Clemens has claimed that our definition of emigrating physicians is "problematic"- a claim that Hidalgo repeats. Because the data agencies in OECD countries use three definitions, we compared the alternative definitions of emigrating physicians in a subsequent paper[6] and found this issue to be unimportant. In fact, correlations between bilateral stocks of emigrating physicians on the basis of countries of birth and countries of training were very high for countries such as the U.S. (0.98), France (0.97), and Canada (0.94). Moreover, developing countries are justified in recovering the investments made in educating the emigrating physicians.

  In summary, policy debates surrounding the emigration of physicians need to be conducted in a scholarly manner. This can be achieved to some degree by conducting several empirical or analytical studies and summarizing the findings for a policy readership. As I had noted previously, "the policy of regularly recruiting physicians from developing countries runs contrary to the ethos of technology transfer". Hidalgo's response avoids addressing many issues that are important from a policy viewpoint.

  REFERENCES

  1. Hidalgo, J. Defending the active recruitment of health workers: a response to commentators. J. Med Ethics. Published online first: 31 may 2013. doi: 10.1136/medethics-2013-101325

  2. Bhargava, A. Physician emigration, population health and public policies. J Med Ethics. Published online first: 26 January 2013. doi:10.1136/medethics-2012-101235.

  3. Clemens, M. Do visas kill? Health effects of African health professional emigration. Center for Global Development Working Paper Number 114 2007:1-47.

  4. Bhargava, A., Docquier, F. HIV pandemic, medical brain drain and economic development. World Bank Econ Rev 2008: 22:345-66.

  5. Bhargava, A., Booysen, F. Healthcare infrastructure and emotional support are predictors of CD4 cell counts and quality of life indices of patients on anti-retroviral treatment in Free State Province, South Africa. AIDS Care 2010:22: 1-9.

  6. Bhargava, A., Docquier, F., Moullan, Y. Modeling the effects of physician emigration on human development. Econ Human Biol 2011:9: 172-83.

  Conflict of Interest:

  None declared

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top