Health worker effectiveness and retention in rural Cambodia

Introduction: A decade after health sector reform, public health services in rural Cambodia remain under-utilised for multiple reasons related to financial, structural and personnel factors. Ineffectiveness of rural public health services has led to a significant increase in private providers, often the same people who staff public facilities. Public health clinics are often portrayed as low quality, with long waiting times and unexpected costs; in contrast, private clinics are seen to provide more convenient health care. Several strategies, including contract management and health equity funds, have been introduced to improve public sector performance and encourage utilization; these efforts are ongoing. However, the feasibility of these strategies remains in question, particularly in terms of cost-effectiveness and sustainability. Methods: In this article the strategies of and barriers met by health workers who remain in rural areas and deliver public health services are elucidated. Ethnographic research conducted in 2008 with health providers involved in treating tuberculosis patients in Kampong Speu Province, Cambodia is drawn on. Participants were recruited from the provincial health department, provincial hospital and four health centres. Data collection involved in-depth interviews, participation in meetings and workshops aimed at health workers, and observation of daily activities at the health facilities. Data were transcribed verbatim, imported into NVivo software (www.qsrinternational.com) for management, and analysed using a grounded theory approach. Results: Primary healthcare service delivery in rural Cambodia was reliant on the retention of mid-level of health staff, primarily midwives and nurses. Its performance was influenced by institutional characteristics relating to the structure of the health system. Personal factors were impacted on by these structural issues and affected the performance of health staff. Institutional factors worked against the provision of high-quality public health services, and included the fragmentation of service delivery and


Introduction
A qualified and motivated health workforce is a key component for health system strengthening and achieving the UN Millennium Development Goals (MDGs) 1 . Recommendations suggest a minimum ratio of 2.5 health workers per 1000 people to achieve MDGs for health 2 .
Nonetheless, most countries experience health staff shortages and maldistribution. This problem is aggravated in poor countries, where resources for producing health workers are scarce and retaining them in rural and remote areas is challenging. Health workers have tended to be concentrated in urban areas, while large rural populations remain under-served 2,3 . The scale of this problem varies considerably: Nepalese data show marked differences in physician-to-population ratios in urban (1:1000) and rural (1:41 000) areas 3 . In India, 74% of physicians practise in urban areas, where only 26% of the population live 4 .
Rural health staff retention is associated with several factors: financial rewards, career development, continuing education, work environment, resource availability, recognition/ appreciation, and social factors (education for children, political stability, and rural upbringing) 3,5 . Economic incentives are used to motivate and retain health workers, but other strategies, such as improved working conditions, supervision and management, and education opportunities, have also influenced staff motivation and retention 6,7 .
As in other low income countries, Cambodia faces challenges to ensure healthcare delivery to people in rural and remote areas because of the unequal distribution of doctors and increasing shortages of midwives 8 . Health staff distribution is heavily skewed toward urban areas, with approximately 54% of physicians employed in the capital city, Phnom Penh, where only 9.3% of the population live 8 ( Table 1).

The Cambodian public health system
Prolonged civil war devastated the Cambodian health system. Fewer than 50 of the 600 doctors who practised medicine before 1975 survived 'the killing fields', health facilities were damaged, and medical equipment was ruined 10,11 . After post-Khmer Rouge remodelling of the health system, the 'Health Coverage Plan' was launched to address health reform in 1995. This involved building new facilities in areas where none existed, and transforming existing hospitals and commune clinics into health centres.
Concurrently, capacity in managing local health services was developed at provincial and operational district (OD) levels.
Cambodia was divided into 77 ODs, each serving 100 000-200 000people and containing one referral hospital (RH) and several health centres 12 . The health centre is the closest public health facility to the population, with each serving 8000-12 000people. It provides basic preventive and curative services through the Minimum Package of Activities (MPA), including the treatment of common diseases (malaria, diarrhoea, sexually transmitted diseases), TB, leprosy, and minor injuries; child immunization; antenatal care and family planning; and refers cases to the hospital if necessary 12 . In addition to the MPA, the RH provides a Complementary Package of Activities (CPA), including treatment of complicated TB cases, referred cases, and medical, surgical and obstetrical emergency cases; in addition to surgery, maternal and child health, provision of X-ray, ultrasound and laboratory services, and rehabilitation services 13 .
Despite these reforms, health services in rural areas have yet to attain government benchmarks for health provision. The profound shortage of qualified health staff has contributed to poor performance. Two-thirds of the health centres in the country were unable to deliver full MPA because of shortages of midwives and nurses, while the upgrade of some RHs to fully implement the CPA was delayed because of shortages of qualified physicians 13,14 .
The root cause of these shortages is staff concentration in urban areas, strongly related to inadequate and irregular salaries, personal feelings about security, poor working conditions, shortages of drugs and medical supplies, and inadequate management of the rural government facilities 8,14 . Staff shortages in rural areas have also led to a rise in the informal health sector. Approximately 10 000 traditional birth attendants, and some thousands of untrained individuals, including traditional healers, work parallel with the formal health system in Cambodia; the majority are unlicensed 9 . When sick, most patients consult a private provider in their community, often a drug-seller or private practitioner whose convenience is related to proximity, availability and flexibility of payment by instalment [15][16][17] . Patients often perceive government services as low quality, with long waiting lists and unexpected high costs from informal payments 18,19 . These perceptions lead to delayed health-seeking from the public health centres.
Several factors contribute to poor performance and inhibit access to public health facilities in Cambodia. Financial incentive strategies, such as contract management and health equity funds, were introduced to address these, to specifically improve performance and encourage the utilization of the public sector [19][20][21][22][23] . However, little is known of how health workers struggle with institutional and financial challenges to remain in rural public health services.
In this article, the various barriers that limit access to rural public health services are elaborated, and the way rural health workers maintain public health service delivery to the population they serve, while meeting their family's economic needs is elucidated.

Methods
The data used in this article derive from a larger ethnographic study conducted in 2008 with 32 tuberculosis   In this article, data from the 10 health workers involved in TB control activities is drawn on. Participants were purposively recruited from health centres, the provincial hospital, and the provincial TB program; only health workers who had been involved in TB control activities were invited to participate. Data were collected using exploratory descriptive methods by the first author, and were derived from 10 in-depth interviews, participation in meetings and workshops, and observation of daily activities at select health centres and the hospital. An iterative approach was employed in data collection, whereby a theme that arose in one interview was used to inform subsequent interviews 24,25 . This was also used to determine data saturation and, therefore, to determine the sample size 26 . Data were imported into NVivo software (www.qsrinternational.com) for management and were analysed using a grounded theory approach 27,28 . Data analysis commenced by carefully reading individual transcripts, and identifying the themes and subthemes which arose. Themes were cross-checked by all authors. A check was made for themes that emerged across transcripts, which allowed the identification of commonalities 25 . Themes were identified inductively 24 and were used in developing the theory when approximately half of the participants spoke about them 29 . The main themes that arose related to: • how rural health staff maintained health service delivery and family economic status • barriers and incentives for retaining staff as government health workers undertaking TB control activities • the impact of challenges of rural health workers on provision of public health services.
Rigour and validity were ensured by methodological triangulation and involving all authors in data analysis 30  decision-making about working in rural areas. These included work responsibilities and motivation, and issues around financial needs. Many of these factors overlapped, and each is now discussed.

Institutional factors
Service delivery and structure: Each health centre attempted to achieve quantitative targets set by the provincial health department. The main objectives of the National TB Control Program (NTP) for 2006-2010 were to attain a smear-positive TB case detection rate of over 70% and a cure rate of greater than 85% 31 . To reach these targets, each health centre was required to diagnose at least two patients each month. This implied variable workloads, with many patients screened in some months and few in others; achieving targets was influenced significantly by staff motivation and supervision. In 2006, Kampong Speu achieved a 73% smear-positive TB case detection rate and 89% cure rate 32 .
While health centres provided the basic MPA, they lacked facilities to diagnose many conditions. All X-ray, ultrasound or pathology tests, including for TB, were undertaken at the RH, often located some distance from the health centre. This contributed to increased workloads for participants, who had to travel to follow-up test results: Informal payments mainly resulted from patient-provider negotiations ensuring that TB patients did not have to attend the health centre daily during the first 2 months of treatment.
Such practices were consistent with Khmer culture, where small gifts express gratitude and respect.
Although they described frustration and discouragement from low salaries and poor supplies, health staff advocated for the provision of public health services, particularly for the poor: Our service is much cheaper than private services and beneficial for poor people. Our user-fees are low [and] set up according to the living standards of our local people. Government employment also conferred status, which was recognised and appreciated by community members, and facilitated access to other government services (eg school, police). In addition, a government job provided employees with opportunities to update their knowledge through participating in training or workshops:

Benefits
The health centre is where the activities of all national programs gather to transfer to community members. I think I have a great opportunity to learn new things… We need to have knowledge on common health problems in the community.
Opportunities for professional improvement were important, providing education for individual staff to perform their work more effectively; however, unequal opportunities to access professional improvement may have de-motivated those with fewer opportunities.
In addition, recognised qualifications and work experience in the public system was important for staff pursuing private work. Seniority was attractive: community people preferred to go to private clinics to seek help from senior health providers who had worked at public services: I've worked at the hospital for quite long, so that many patients know me and refer their friends to my private clinic. In private practice, we have to ensure quality and continuously upgrade our services and equipment otherwise clients don't want to use our service.
Staff built their networks, reputation and capacity at government facilities, then applied these for personal benefit in private practice. Government jobs also offered flexible working environments, without strict regulations regarding absence or punctuality, and staff could set their own working hours. This supported the institutionalization of parallel jobs, even though it adversely limited provision of public services.
Strategies were in place to compensate staff on especially low incomes: people were paid allowances to motivate work performance or to participate in training or workshops.
Although the rates were small, participants regarded these opportunities as a kind of motivation to retain them in public facilities: The NTP provides $US32 to follow-up patients, supervise sputum preparation, and provide health education. RHAC [an NGO] provides $US16 for health education. This amount is small but it can help us to supply fuel to travel to villages. More importantly, this budget never been cut or missed.
To go to villages for child immunization, staff get around $US1-2, a low amount not good for a living, but is an incentive for their work.
Financial incentives were important. Performance-based incentives, combined with supportive monitoring, motivated staff, encouraged them to take on increased responsibility, and might have been the most practical way to ensure quality staff performance and retention. However, social reasons were also important. Family acceptance into the local community was crucial in participants' decisions to remain in rural areas. Working in the public health service meant that health workers were able to keep their families together: all had a rural upbringing and were well accommodated by their local communities, even those originally from a different community.

Discussion
The delivery of primary health care services in rural Cambodia relied on the retention of mid-level health providers (midwives and nurses). Although low salaries contributed to staff dissatisfaction, other financial strategies motivated staff to work in rural areas and undertake TB activities. These included: job security, recognition and status in society, capacity building opportunities, flexible working environments, and allowances. This echoes research in Vietnam, Africa and Nepal which found that health workers were motivated through non-financial incentives such as appreciation and support from managers and colleagues, job and income stability, and provision of professional development opportunities 3,37,38 .
In Cambodia, the opportunity to hold parallel jobs -where health staff occupied positions in both public and private health services -played a central role in keeping health workers in rural areas 7,39-41 , yet also undermined their satisfaction with working in the public health system. Parallel jobs enabled health workers to earn a sufficiently high family income, and was widely accepted as necessary to meet economic demands. However, this created a conflict of interest, where health workers, either intentionally or not, diverted clients from public services to their private clinics 8 .
Although the providers did not see it as problematic, this promoted increasing use of unregulated private facilities. This was exacerbated by poor access to and irregular supplies in public services, which drove community members to seek help from private providers, and contributed to high costs of health care and delayed access to appropriate care. Although TB diagnosis and treatment were provided free of charge at public clinics, patients usually chose private providers and only presented to public services when their health had deteriorated significantly.
Institutional barriers impacted significantly on rural health staff, hampering performance and contributing to poor utilisation of health services, particularly among the poor.
The provision of basic care at public health centres was strongly influenced by staff motivation and input from higher levels: supplies, allowances to support activities, and appropriate supervision. Addressing these issues by providing secured supplies of drugs and equipment, implementing regular monitoring and supporting outreach activities would give staff motivation to achieve provincial targets.
Strategies to improve retention should balance the potential of non-financial benefits, financial opportunities, and staff workload. Institutional reform may not work, if the financial needs of health workers are addressed inadequately. The recent government decision to cease all salary incentive schemes from January 2010 potentially jeopardizes the current public health system 42 , unless alternative plans to ensure adequate financial benefits for public health workers are implemented. Ultimately, this decision will increase the health burden for people, particularly the poor, placing adequate health care further out of reach. This concern may be reflected through the reinstatement of salary incentives provided through donors until the issuance of a new government incentive modality 43 .

Limitations and implications
Several issues related to sampling and sample size may limit the generalisability of the study findings. First, our study only recruited staff actively involved in TB-control activities, who received regular governmental support (drugs, facilities). In consequence, it may include an overrepresentation of motivated health staff. Second, the small sample size may limit the generalisability of the findings; however, our use of theoretical saturation in determining sample size suggests that our findings may be relevant for other health workers in this province.

Conclusion
The data from this study provides important insights into how personal and institutional factors influence the motivation and retention of health workers in rural areas of Kampong Speu Province, Cambodia. The findings reported here also highlight an important direction for future research, within Cambodia and elsewhere, in order to understand whether issues around parallel jobs and public health system performance are applicable in other contexts. This research has implications for the delivery of health care to people in rural and remote areas, where access to services is already limited, and provides important insights into redressing the social and structural barriers affecting the provision of quality, affordable health care.