The Referral system: a neglected element in the health district concept. Eight years fiels research in rural Niger.

Samenvatting

Between 1996 and 2003, the referral system was studied in two rural districts in Niger where a GTZ-financed project aimed improving the local health system. Already early in the process it was found that the referral system was particularly weak: women were dying from delivery complications at the peripheral health centres because of lack of transport. The rural hospital in Ouallam district, catering in theory for more than 250,000 people was virtually empty at any one time.From literature could be deduced that the referral system was hardly subject of research in developing countries, though it is a cornerstone of the two-tiered health district concept. The fact that it does not seem to function anywhere suggests though that things are more complex than initially assumed.The study design for a complex social reality such as a referral system has to combine different research tools such as direct observation, interviews and focus group discussions, and experimental intervention.A nation-wide study on unmet obstetric needs confirmed that the referral system was extremely weak in Niger. The performance of the two national maternities –0.9% of all deliveries needed a major obstetric intervention to save the mother’s life- served as a benchmark to evaluate the performance in the rest of the country. Obstetric patient files in all the hospitals of Niger were explored to be compared with the national maternity’s benchmark. More than 75% of the women (90% in the rural areas) in need of a major obstetric intervention to save their own life did not benefit from such intervention. The major reasons were the failing referral system and the scarcity of proper hospital care in the country. No wonder that Niger has an extremely high maternal mortality: 1600 maternal deaths for every 100,000 life births estimated by WHO.Individual questionnaires for patients and health staff and focus group discussions in villages provided a useful insight in the determinants of the referral system. This information was completed with direct observation data and monitoring of routine district indicators. Triangulation of data strengthened the conclusions.The referral system was found to be determined by different factors like distance, transport availability and affordability, technical skills of the health centre staff and socio-cultural barriers.For the first time, the intuitive relation between hospital credibility and acceptability of referral proposals for the patients was demonstrated. Emergency referrals only increased significantly at the moment the hospital was upgraded, though the ambulance service was already operational a few years earlier.The relational aspects of referrals were studied in detail. The relation between the health centre staff and the patient was characterised by an authoritarian staff, often fearing to loose prestige when referring a patient and a submissive patient taken by fear for death and for the unknown when a referral was proposed. The combination of the two, in addition of the economic barrier to referral for the rural poor made that health centres were hardly referring. Referral rates in the rural areas initially did not exceed 0.5% of all new patients whereas in urban areas 5% was not exceptional.Referral rates are difficult to interpret because there is no such thing as a universally accepted good or ideal referral rate. To overcome this problem, last-year medical students replaced temporarily the nurses in some of the rural health centres to maximise good clinical practice. They strictly applied the clinical guidelines normally followed in Niger’s health centres which resulted in 2.5% referral proposals among the nearly 4,000 new patients they consulted. This figure served as a benchmark to evaluate actual referrals in rural areas. At least 50% of the patients in need of referral either did not receive a referral proposal or did not respect it. This led inevitably to avoidable mortality and underused rural hospitals. Especially children seemed to pay a high tribute.Observation and monitoring of referral patterns showed that a number of serious pathologies were hardly or too late referred. Severe malnutrition in young children was one of such neglected diseases. Initially more than half the maternal deaths in Ouallam district took place at the health centre level without referring to the district hospital. Patients were also kept under observation at the health centres to avoid referral. The quality of this activity was studied, and the hypothesis that keeping patients under observation could avoid referrals was verified. In absence of any guidelines and confronted with complex clinical syndromes, the health centre staff was not apt to deal with such situations. The hospitalisations at health centre level rather led to delayed referral or were so unsystematic that it did not provide any benefit for the patient. Some patients with obvious obstructive abdominal syndrome were treated with up to 7 different drugs during 36 hours before they were referred. Operational instructions for patients kept under observation at the health centre should be introduced urgently, especially in the integrated child care programme.A typical intervention study on the organisation of an emergency referral system in the districts showed in how far the referral system was subject to change and which additional factors needed to be addressed to make it function. In this experimental study, radio communication between the rural health centres and the district hospital was established in combination with an ambulance service. The number of useful emergency referrals increased from less than 20 patients per year to more than 450 in a district of 250,000 inhabitants. A cost-effectiveness analysis showed that the individual cost to transfer a patient remained high but that the costs per inhabitant for an emergency referral service turned out to be very low and could be easily absorbed into health insurance schemes.The referral system is an important part within the itinerary of a patient in need of hospital care. The 3-stage model developed by Thaddeus and Maine and which is used in many studies to analyse bottlenecks to access appropriate care, has been modified and tested empirically. Patients in need of hospital care actually pass through 5 different stages that, though closely intertwined, represent specific potential barriers for hospital utilisation. It shows among other things why the fight against maternal mortality –one of the millennium development goals- has booked very little progress so far. It sheds a new light on the importance of the referral system and how it can be dealt with both for its hardware and software aspects. Operational indicators could be identified to qualify these specific bottlenecks.The referral problem should not be detached from the general health care system. Though important and often particularly neglected, it is but a sub-system of the global district health system. Isolated corrective measures to improve the referral system do not have any impact on people’s health if they are not integrated in the management of the global system. The health district concept remains a valuable model to develop solid and efficient health services as long as the referral system gets the necessary attention.

Datum prijs	29 jun. 2006
Originele taal	English
Prijsuitreikende instantie	Vrije Universiteit Brussel
Begeleider	Anne-Marie Depoorter (Promotor), Wim Van Lerberghe (Promotor), Freddy Louckx (Jury), Tony Mets (Jury), Jacques De Grève (Jury), Abdallah Bchir (Jury), Marleen Temmerman (Jury) & Wim Van Damme (Jury)