Journal name: International Journal of Public Health

Article Title: Examining the relationship between human resources and mortality: the effects of methodological choices.

First author: Pablo Villalobos Dintrans, Msc., Harvard T.H. Chan School of Public Health

Second / Corresp.author: Claire Chaumont, Msc. , Harvard T.H. Chan School of Public Health,

Annex: Summary of studies

Reference / Year / Sample / Source / Unit / Dependent variable / Independent variable / Controls / Method / Results (coefficient on human resources) / Implications / Google Scholar citation count
(As reviewed on September 11,2016)
Anand and Bärnighausen (2004) / 2004 / Initial sample: 198
Final sample: 117
Final sample (including poverty): 83 / * 2004 WHO dataset, Estimates of Health Personnel
* World Development Indicators
* Human Development Report 2003 / Country / Maternal mortality, infant mortality, under-five mortality / Human resource density (Physician, nurses, midwives) / Gross national income per person, poverty, female adult literacy / OLS (cross-section) / Maternal mortality: -0.474
Infant mortality: (-0.235, -0.212)
Under-five mortality: (-0.260, -0.231) / "… investing in human resources for health should be explicitly considered as part of a strategy to achieve the MDG…" / 517
Ananthakrishnan et al. (2010) / 1997-1999 / 67 / * Pennsylvania cancer registry
* American Medical Association / County (Pennsylvania) / Colorectal cancer (CRC) incidence rate, age-adjusted CRC-related mortality rate / Human resource density (primary care physicians) / Population density, proportion of patients of black race or Hispanic ethnicity in the county, proportion of residents over the age of 25 who are high school graduates, socioeconomic index / Ecological analysis (correlations) / Non-metropolitan: (-0.41, -0.22)
Metropolitan: (-0.05, 0.13) / "Recent concerns have been raised about the potential shortage of physicians...our study supports these concerns"
"… increase in physician supply may need to be targeted to non- metropolitan counties" / 31
Castillo-Laborde (2011) / Different years for different variables / Initial sample: 193
Final sample: 125 / * World Health Statistics 2009
* Global Health Atlas
* World Economic Outlook Database
* World Development Indicators
* WHO Health Statistics and Health Information Systems / Country / Disability-adjusted life years (DALYs) / Human resource density (Physician, nurses, midwives) / Health expenditure/ GDP, GDP per capita, % rural population with access to clean water, GINI coefficient, income share by the lowest 10% / OLS (cross-section) / DALY 0: (-0.036, -0.013)
DALY 1: (-0.151, -0.103)
DALY 2: (0.0253, 0.0259)
DALY 3: (-0.020, -0.012) / "The findings of the study have implications not only for health and health policy, but also for research. They represent supporting evidence of the importance of health workers for health" / 16
Farahani et al. (2009) / 2000 (cross-section)
1960-2000 (panel) / # countries: 92 (cross-section), 99 (panel) / * World Bank’s World Development Indicators
* Penn World Table
* Barro–Lee educational attainment dataset / Country / Infant mortality rate / Physicians / GDP, years of education / OLS (cross-section, panel), fixed effects (panel), GMM (panel) / Infant mortality; (-0.302; -0.116) / "Over time, investment in programs that increase the number of health care professionals can play an important role in helping to achieve the Millennium Development Goal of reducing infant mortality" / 2
Ricketts and Holmes (2007) / 1996-2000 / 3,070 counties (49 states) / * 2002 Area Resource File (ARF) (U.S. Bureau of Health Professions) / County (US) / Mortality rate, disease-specific mortality rate (heart, cancer) / Physicians / Per capita income, education, unemployment, age, race, poverty / Fixed effects / All causes: (-0.457, 0.409)
Heart: (-0.188, -0.0017)
Cancer: (-0.049, 0.0215) / "There are similar differences in the distribution of physicians, but their distribution is not clearly related to mortality patterns… these relationships are not consistent across the United States" / 49
Robinson and Wharrad (2001) / Different years for different variables / 155 / * UNICEF
* WHO
* World Bank / Country / Maternal mortality / Human resource density (physician, nurses), births attended by professionals, GNP per capita, female literacy / GNP per capita / Multivariate analysis / Maternal mortality: (-0.921, -0.241) / "… maternal deaths are substantially reduced when ahigh proportion of births are attended by health professionals" / 79
Speybroeck et al. (2006) / 2006 / Initial sample: 191
Final sample: 166 / * WHO dataset / Country / Skilled birth attendants, Measles coverage, Infant mortality, Under five mortality, Maternal mortality / Human resource density / GDP, Income poverty, female literacy / OLS / Maternal mortality: (-0.5825, -0.5719)
Infant mortality: (-0.1175, -0.1010)
Under-five mortality: (-0.1540, -0.1447) / "Combined health worker density is also strongly associated with health outcomes" / 70
Starfield et al (2005) / 1996-2000 / 3,075 counties / * National Center for Health Workforce Analysis / County (US) / Mortality rate, disease-specific mortality rate (heart, cancer) / Primary and specialist supply / Per capita income, education, unemployment, age, race, poverty / Multivariate analysis / All causes: (-0.0352, 0.051)
Heart: (-0.0171, -0.0031)
Cancer: (-0.039, 0.0053) / "... increasing the specialist supply is not justifiable… there is already considerable evidence that increasing the supply of primary care physicians would have a beneficial impact on the health of the population" / 263