Violent conflicts pose a challenge to human civilisations, human health and health systems [1–3]. Epidemiological studies indicate that war ranks among the top-ten causes of death worldwide [4–6]. Populations affected by armed conflict experience severe public health consequences mediated by population displacement, food scarcity, and the collapse of basic health services, which together often give rise to complex humanitarian emergencies [7, 8]. Conflict has both direct and indirect effects on people's health and on the overall health system [8]. Armed conflicts can also cause the displacement of people and an increase in infectious diseases [2, 9].

Nepal recently emerged from a decade-long violent conflict (1996 to 2006). This violent conflict had an effect on both the population's health and the health care system[10–12].It led to over 13,000 fatalities[13], the disappearance of at least 1,200 people [10, 14], the disablement of thousands of people, and the internal displacement of many more [14, 15]. Over 1,000 health posts in rural areas were destroyed [16], more than a dozen health workers had been killed and many others were harassed, kidnapped, threatened and prosecuted by the warring factions [14, 17, 18]. The conflict aggravated the already poor health services as one third of Nepal's health centres is in rural areas (where some of the fighting was heaviest) and often operates without health staff [19–21]. Torture and sexual-abuse related to insurgency were also prominent [11, 22, 23], and the conflict also hindered health programmes implemented by non-governmental organisations [24, 25].

The Maoist rebels put restrictions on field staff mobility and both the security forces and rebels tried to stop public gatherings focused on health-related awareness. Furthermore, the Maoists objected to the implementation of the Community Drug Programme (CDP) by opposing the minimal fees associated with it.

Amidst the civil war, Nepal appeared to have made improvements in its human development index, life expectancy and child and maternal health indicators [18, 21, 27]. Some of the publicly available datasets suggest that Nepal has made considerable progress on certain key health indicators, however for few other indicators the progress seems to have stagnated. In recognition of its progress on reduction in maternal and child mortality rate and improvement on other health indicators, the US (United States) Government recently commended Nepal under its Global Health Initiative Programme. Nepal is only one of eight countries receiving this award. Nepal's progress seems puzzling as it contradicts our common understanding that civil conflict is an impediment for improving the health services. It raises a question whether the progress was real, and if so, what could have contributed to achieve this progress? The possible hypotheses could be that Nepal's violent conflict: i) worsened health indicators; ii) improved health indicators; or iii) had a mixed effect, i.e. improvement in some and stagnation or deterioration in other indicators. This paper analyses Nepal's main health indicators before and after the conflict and offers some possible explanations for the observed changes.

This paper is based on secondary analysis, in which "data collected by one researcher are re-analysed by another investigator usually to test new research hypotheses" [28]. Thus secondary analysis uses data which have already been collected and the research question might, or might not, have formed part of the remit of the original study design. In this paper we draw upon data from three main sources: i) demographic health surveys 1996, which were called Nepal Family Health Survey (NFHS) in 1996 [29] and Nepal Demographic Health Survey (NDHS) in 2006[30]; ii) a study led by the first author under the auspicious of Nepal Health Research Council (NHRC) in 2006 [27]; and iii) data from the Ministry of Health and Population (MOHP) and similar sources. The NFHS 1996 used household questionnaire and women interviews while the NDHS 2006 used household interviews and separate interviews with women and men. This paper compares health indicators based on the household interviews (particularly demographic characteristics, water, sanitation, nutritional status of children) and women interviews (e.g. education, marriage, childbirth, family planning, fertility, maternity care, immunisation, awareness of HIV/AIDS) in order to address gender biases whilst comparing the 1996 and 2006 data.

There is a possibility that problems occurred during the data collection of the various surveys. The insecurity due to conflict made the survey data collection less reliable [31] since (a) parts of the country was not under Government control; and (b) Census enumerators might have been afraid to approach people whom they believed to be Maoist sympathisers as Census enumerators were working for the Government. Some of this may also have occurred during the data collection for the studies used in our secondary analysis.

The study conducted by the NHRC in 2006 covered 800 women with children under the age of two, 40 health service providers, 145 key informants, 104 exit clients at the service outlets and 400 focused group discussion participants from across 10 districts representing all five regions of Nepal [27]. The methods and tools of this study does not seem compatible to the NFHS 1996 and NDHS 2006. Moreover, the sample size of the NHRC study is relatively small. The results of the NHRC study however give a comparative picture on 6 out of 19 indicators included in Table 2. It offers data for supplementation to the NDHS 2006. The qualitative data from the NHRC study (2006) are used to supplement the analysis of the changes over time (where available and appropriate).

Table 2 presents the key health indicators before the start of the violent conflict in 1996 and immediately after denouncement of violence by the Maoists in 2006.The data are presented into two sub-headings; health outcomes demonstrating improvement and health outcomes that remained stagnant or even worse during the decade-long conflict.

From the point of view of the impact of the conflict, the data available from the two DHSs suggest more of a positive than of a negative impact on the health outcomes. The comparative data on 19 MDG-related indicators show that 16 out of 19 indicators had improved to such a level that MDG would be likely to be achieved by 2015. While two indicators-reductions in neonatal mortality and improvement in skilled attendance at birth had increased at a slower pace, hence the related MDGs are unlikely to be achieved. One indicator, the percentage of undernourished children under three years old worsened in 2006 compared to the reference year 1996. Most of these findings on the trend of progress are compatible to the trends of health indicators shown in the MDG Progress Report published by Nepal's National Planning Commission in 2010 [32]. According to this report "Nepal is likely to meet the targets on reducing under five mortality by two-thirds, reduce the maternal mortality ratio by three quarters, halt and reverse the spread of HIV/AIDS, halt and reverse the incidence of malaria and other major diseases and halve proportion of population without sustainable access to improved water source. It is potentially likely to meet the targets on achieving universal access to treatment for HIV/AIDS for all those who need it. However, the report reiterates that Nepal is unlikely to meet the targets of achieving universal access to reproductive health and halving proportion of population without sustainable access to improved sanitation"[32].

Contrary to evidence from other conflicts [8, 34–37] as well as from Nepal [38–40] of a negative impact of conflict on the health of populations, we found that in Nepal progress has been made in most health indicators. There does not appear much literature on what made it possible to achieve such progress despite a decade-long armed conflict. The discussion below explores the key drivers contributing to the better than expected changes in people's health status in a period of civil unrest and armed violence.

A second explanation is that the former rebels put pressure on the health care providers in their 'base areas' or the contested areas to attend regularly at clinics in order to ensure consistent drug supplies and treatment [42]. As a result, the government was under pressure to supply appropriate health staff and supplies. In spite of the security threat, 78% of staff positions in hospital, 75% in primary health care centres (PHCCs), 96% in health posts and 90% in sub-health posts were filled during the conflict [27].

The example of improved coordination despite the conflict in Nepal was also found during conflicts in East Timor [43] and Mozambique [37] where improved coordination amongst the key stakeholders helped increase utilisation of health services by the local population. In Nepal, it encouraged inclusive, people-based and transparent humanitarian programmes at the local level. Exemption of user fees to poor and disadvantaged populations and provision of citizen charters (agarics adapter) at service outlets could be taken as examples [27]. It also recognised the role of civil society and the local community groups in these health development activities.

Fourthly, building on the lessons from the protracted conflict, Nepal's public health system adopted a number of health improvement approaches and programmes. Some of the key policies focused on disadvantaged groups including dalits, women, disabled and elderly people, whilst helping to increase coverage of the health programmes in more remote and underserved areas. The policies also included the establishment of emergency funds and community drugs schemes and handing over the government ownership of the health facilities to the local communities [27].

Fifthly, Nepal strived to maintain a visible, sustained and adequate provision of health services at all levels from the centre to the community. There has been a substantial increase in the number of health care institutions, from 1,098 in 1991 to 4,552 in 2007/2008 [45]. The Government health facilities, such as health posts, sub-health posts, primary health care centres and outreach clinics provided basic community-based services, mostly free of charge. Nepal implemented many popular programmes such as the community-based integrated management of childhood diseases (CB-IMCI); community-based newborn care package(CB-NCB), community drug programme (CDP); direct observation treatment system (DOTS) for treatment of tuberculosis; HIV and AIDS prevention and control programmes; rural water supply and sanitation programme (RWSSP) and a food security programme. These initiatives helped increase access to and utilisation of the available health services [27, 32].

Sixthly, there was a functional community support system including the Health Facility Management Committees, mothers groups, Female Community Health Volunteers (FCHVs) and Traditional Birth Attendants (TBAs) for the mobilisation of local communities. One study showed that one-thirds of women were member of local women's groups, and that 43% members of the health facility management committees were from lower socio-economic groups such as Janajatis and dalits[27]. However, motivation and performance of these groups were often questionable in terms of their voluntariness as opposed to their desire for economic incentives, including the coping strategy in the context of the political conflict [46].

Seventhly, the UN (United Nations) and various international non-governmental organisations (INGOs) contributed for increasing the coverage and effectiveness of the health services in Nepal. They implemented conflict-sensitive development programmes whilst keeping a low profile [47]. Nonetheless, in the absence of clear government policy and elected representatives, coordination between the government, development partners and the community people appeared to be poor [27].

Increase in access to education and communication could have supported positive changes in health outcomes. During the decade of 1996-2006, adult literacy increased from 34% in 1996 to 79% in 2006 [29, 30]. The primary school enrollment rate increased from 57% to 73%. In 1996, only 7% of all households had a radio and television, which increased to 28% in 2006 [33].

Ninthly, Nepal achieved a steady economic growth and substantial reduction in poverty. Between 1995/96 and 2005/6, the percentage of the population living below the poverty line (US$1/day) decreased from 42% to 31%, and the absolute poverty dropped by one percentage points per year over the past couple of years. This somehow seems to contradict the economic explanation on the causation of conflict that underdevelopment and poverty fuels conflict [48–50]. However, a 2005 regional poverty profile shows that Nepal has varying regional deprivation levels. During 2003-2004, Kathmandu had the lowest level of poverty (3%) while the other urban and rural areas had higher poverty levels i.e. 9.6% and 34.6% respectively [51]. The Nepal Living Standard Survey (NLSS II), 2003/2004 also reveals discrepancies in the distribution of poverty by development regions. It is lowest in the Central Development Region (27%) and highest in the Mid-western Development Region (45%), which is considered as the epicentre of the Maoist insurgency [52].

Economic inequality was reported between (a) the centre and the periphery; (b) the 'haves and have-nots; (c) different castes; and (e) people with different levels of education. For instance, in Kathmandu the average gross domestic product (GDP) was almost four times higher than that of some rural regions [52].

The increase in government's health sector budget, though only a small percentage change, might have helped towards achieving these health outcomes. The share of health sector budget increased from 5.99% in 1995/96 to 6.41% in 2005/2006 [32]. Moreover, the share of foreign aid of total government expenditure increased from 17.96% (2001/2002) to 19.88% in 2005/2006 and its contribution in Nepal's development expenditure increased from 58.07% to 74.45% [32]. Similarly, the share of foreign aid to GDP in the same period increased from 3.13% to 3.37% [32].These inputs would have contributed to the positive changes in the health indicators.

In spite of the violent conflict, Nepal made progress in 16 out of 19 health indicators over the period 1996-2006. The indicators of universal access to reproductive health, halving proportion of population without sustainable access to improved sanitation and proportion of underweight children has remained stagnant. We have outlined nine possible factors that help explain this phenomenon of seemingly improved health outcomes in a time of war. It is, of course, very likely that a combination of these nine factors interacted to create the positive environment in Nepal, despite, or perhaps because of its internal conflict.

The lessons from Nepal are that in order to ensure functional delivery of health services and improvement in health outcomes during conflict, the warring sides should adopt a strategy of coexistence and the international community should continue and increase their support to strengthen the health sector with a principle of 'do-no-harm' and impartiality and the government should implement conflict-sensitive measures and improve coordination amongst the key actors. Moreover, the overall national economic and social context should be conducive to bridging divides, and finally the government should work to fulfill its commitment towards the national policies and programmes and international instruments. It is equally important to reform the health services by building on Nepal's experience and consider the positive transformations that can occur as a result of conflict.

As this was the first comparative study that examined the health outcomes before and after the conflict and presented available evidences to explore the reasons for the positive changes, this paper provides general trend of health indicators overtime. Future studies should try to differentiate between conflict affected and peaceful areas and look at the conflict attributes that generate positive and negative consequences for the health services. Perhaps a little more focus is needed on the positive aspects as most of the studies conducted elsewhere portray negative consequences of conflict and ignore the transformation that occurs as a result of conflict.