Health Policy www.thelancet.com Vol 371 February 23, 2008 675 Salaries and incomes of health workers in sub-Saharan Africa David McCoy, Sara Bennett, Sophie Witter, Bob Pond, Brook Baker, Jeff Gow, Sudeep Chand, Tim Ensor, Barbara McPake Public-sector health workers are vital to the functioning of health systems. We aimed to investigate pay structures for health workers in the public sector in sub-Saharan Africa; the adequacy of incomes for health workers; the management of public-sector pay; and the fiscal and macroeconomic factors that impinge on pay policy for the public sector. Because salary differentials affect staff migration and retention, we also discuss pay in the private sector. We surveyed historical trends in the pay of civil servants in Africa over the past 40 years. We used some empirical data, but found that accurate and complete data were scarce. The available data suggested that pay structures vary across countries, and are often structured in complex ways. Health workers also commonly use other sources of income to supplement their formal pay. The pay and income of health workers varies widely, whether between countries, by comparison with cost of living, or between the public and private sectors. To optimise the distribution and mix of health workers, policy interventions to address their pay and incomes are needed. Fiscal constraints to increased salaries might need to be overcome in many countries, and non-financial incentives improved. Introduction The pay and income of health workers affect health care and health systems in many ways. Pay and income have been described as hygiene factors 1 that affect motivation, performance, morale, and the ability of employers to attract and retain staff. When pay is low in absolute terms, health workers will moonlight to supplement their incomes by providing health-care services privately, engaging in other income-earning activities, extracting informal fees from their patients, or seeking per-diem payments by attending workshops and seminars. 2 The wider earning power of health workers depends on the context in which they work; richer urban settings generally provide opportunities for private practice, whereas rural regions provide opportunities to supplement pay with non-financial income such as locally grown food. Health workers are also affected by relative differences in pay and income. Relatively low pay can cause dissatisfaction and loss of motivation, and cause migration towards higher earning jobs. The size of the pay differential between different types of health worker (eg, doctors and nurses) can also affect morale, working relationships, and the available mix of cadres. Differences in pay and income can therefore affect both retention within countries and distribution of health workers, whether between urban and rural areas or between the public and private sector. Pay for health workers is also an important determinant of overall health expenditure. In 2006 in Ghana, for example, when health worker pay and emoluments went 35% over budget, they absorbed 76% of government spending on health; this left only 6% of the government budget for non-wage recurrent expenditure once capital expenditure had been spent. 3 Availability of data Policy debate and discussion about health-worker salaries and incomes in countries with low and middle incomes is constrained by insufficient data. In theory, data on public-sector pay should be readily available from government databases, but in practice, such data are inaccurate, incomplete, unclear, and out of date. 4–6 Pay structures are often complex, consisting of a mix of salary, various allowances, periodic bonuses, overtime payments, and other forms of remuneration such as per diems. Most data exclude non-wage benefits (such as employers’ contributions to pensions, medical insurance, and housing); unofficial payments or gifts from patients; and other income from private sources. Because data collection is not standardised or consistent, cross-country comparisons and analyses of longitudinal trends are difficult. Cross-country comparisons are also hindered by differences in job descriptions, roles, and levels of training. Another difficulty is that data about the average salary of a particular type of health worker might not reflect variations in pay for subsets of workers within that grade (eg, different types of nurses). Data on health workers who are employed in the private sector are more scarce than those for government staff, notwithstanding concerns about the brain drain of health workers from the public sector. 7,8 The heterogeneity and lack of regulation that characterise this sector in sub-Saharan Africa complicate data collection. Even if the commercial private sector is excluded, scores of non-profit organisations employ health workers under different terms and conditions of service. Because of the poor quality of routinely collected data, we searched recent surveys of health-worker pay and income in sub-Saharan Africa. We searched Medline, and the International Monetary Fund (IMF), World Bank, International Labour Organization (ILO), and WHO websites. We also consulted expert researchers and WHO staff. We obtained public-sector data from two surveys by the Initiative for Maternal Mortality Programme Assessment (IMMPACT) in Burkina Faso 9 and Ghana, 10 and from two by the World Bank in Zambia 11 and Nigeria, 12 in collaboration with country governments (table 1). We have also drawn on three other sources of secondary data from Zambia, 13 Ethiopia, 14 and Malawi. 15 Lancet 2008; 371: 675–81 See Editorial page 623 See Comment page 632 Centre for International Health and Development, University College London, UK (D McCoy DrPh, S Chand MFPH); Alliance for Health Systems and Policy Research, WHO, Geneva, Switzerland (S Bennett PhD); IMMPACT, University of Aberdeen, UK (S Witter MA, T Ensor PhD); World Health Organization, Geneva, Switzerland (B Pond MD); Health GAP, Northeastern University School of Law, Boston, MA, USA (Prof B Baker JD); Queen Margaret University, Edinburgh, UK (B McPake PhD); and Health Economics and HIV/AIDS Research Division, University of Kwazulu-Natal, Durban, South Africa (J Gow PhD) Correspondence to: David McCoy, Centre for International Health and Development, University College London, 30 Guilford Street, London WC1N 1EH, UK d.mccoy@ucl.ac.uk