Motorcycle ambulances for referral of obstetric emergencies in rural Malawi: Do they reduce delay and what do they cost?

1. Introduction 

Achieving the fifth Millennium Development Goal by reducing maternal mortality and morbidity remains a great challenge for many low resourced countries, particularly in sub-Saharan Africa where little overall progress has been observed and maternal mortality ratios are even increasing in some countries [1]. The main causes of maternal mortality are puerperal sepsis, hemorrhage, obstructed labor, ruptured uterus, complications of abortion, and eclampsia [2], [3].

Besides the direct and indirect obstetric causes there are many underlying factors that contribute to high maternal mortality in developing countries; one of these is delay in reaching an emergency obstetric care (EmOC) facility after a complication has occurred. This delay can take different forms such as delay in deciding to seek care, delay in reaching a health facility for EmOC, and delay in receiving adequate care at the health facility [4]. Cultural, socioeconomic, geographical, and health system factors play a role in these delays. A large percentage of maternal deaths could be avoided with better referral [5]. Poor transport for the referral of women with obstetric emergencies is one of the avoidable factors.

Ensuring access to skilled attendance at birth for all women and timely access to quality EmOC for women with obstetric complications are increasingly recognized as priority interventions needed to reduce maternal and neonatal mortality and morbidity [6], [7], [8], [9]. Availability of EmOC facilities and services, distance from an EmOC facility, and availability and cost of transport are crucial factors determining timely access to EmOC.

To reach EmOC services one of the vital needs is the availability and accessibility of suitable and affordable transport. Most attempts to improve access to EmOC focus on establishing a minimum acceptable number of EmOC facilities, improving quality of care in such facilities, and mobilizing communities to enable women to utilize these services. However, alleviating transport problems to access EmOC facilities is a relatively neglected area. Poor transport arrangements and transport costs for referral of women with obstetric emergencies contribute to the second type of delay. Transport costs, including transport of emergencies, also put a strain on limited district health budgets, making emergency ambulance services difficult to sustain. Where free ambulance services are not available the cost of emergency transportation to a referral hospital is a barrier to accessing EmOC for poor mothers and their families in remote rural areas.

Various options are currently being implemented to improve rural transport and communication systems with little documentation of effectiveness [10], [11]. Since rapid transportation to an EmOC facility is of prime importance in obstetric emergencies, car ambulances seem an obvious solution. However, car ambulances are costly in terms of purchase, fuel, maintenance, and repair. They are usually stationed at referral hospital level and it is too expensive to provide rural health centers (HC) with these ambulances. Moreover, they are not always available when needed and are not easily accessible for health centers without radio or telephone. Some nongovernmental organizations (NGOs) and safe motherhood projects have introduced bicycle ambulances, either at health centers or at village level [12], but for transport to a distant referral hospital they still cause considerable delay. Maternity waiting homes are another option for improving access to obstetric care [13]. In some programs arrangements and fees for emergency referral have been negotiated with local transporters [14], [15], but vehicles may not be available near some health centers.

To investigate strategies to improve referral for women with obstetric complications the Safe Motherhood Project in the southern region of Malawi initiated an operational research and pilot project in which motorcycle ambulances, designed by the Ranger Production Company (www.eranger.com) and pilot tested by a Zimbabwean NGO (“Riders for Health” www.riders.org), were provided to 3 rural health centers for transporting obstetric emergencies to the district hospital. The present paper reports the impact of motorcycle ambulances on referral time and their costs compared with a 4WD car ambulance.

2. Methods 

The objectives of this operational research were to assess referral time, costs, acceptability, and feasibility of motorcycle ambulances placed at 3 remote rural health centers for referral of obstetric emergencies to the district hospital, and to compare the referral delays and costs with those of a car ambulance stationed at the district hospital. In the latter situation, car ambulances were called out to the health centers to collect emergency cases.

Three motorcycle ambulances, consisting of a 250 cc Yamaha motorcycle with sidecar (Fig. 1), which could carry 2 adults, were stationed at 3 remote rural health centers (Makanjira, Mase, and Phirilongwe) in Mangochi district, Malawi. At each health center a Health Surveillance Assistant (HSA) was selected as the rider. HSAs are government-paid community health workers. The 3 riders were trained over 2 weeks to ride the motorcycle ambulance and on simple maintenance. They were also instructed on data recording in logbooks. According to the Ministry of Health's policy, referral of patients by motorcycle ambulance was free of charge, as was referral by a government car ambulance.

View Large Image Download to PowerPoint

Figure 1. Motorcycle ambulance.

Mangochi district with a population of about 600 000 was chosen because it had poor maternal health indicators (high maternal mortality, low institutional delivery rate), it is vast and with difficult and diverse geography, and it was a target district of the Safe Motherhood project. The 3 health centers were chosen because of their remoteness, varied geographical features, and high number of reported maternal deaths. Makanjira is a government health center on the eastern side of Lake Malawi, 110 km from the referral hospital; Mase is a faith-based health center on the eastern side of Lake Malombe, 30 km from the referral hospital; and Phirilongwe is a government health center in the western part of Mangochi, 60 km from the referral hospital (Fig. 2). The connecting roads are dirt roads that are difficult to navigate during the rainy season. The 3 health centers provided the full range of preventive and curative primary health care services such as outpatient treatment, prenatal, delivery and postnatal care, and under-5 clinics, including immunizations. Family planning services were only available in the 2 government health centers, but not in Mase HC, which is operated by the Roman Catholic Church.

View Large Image Download to PowerPoint

Figure 2. Map of Mangochi District, Malawi.

Data were collected over a 12-month period from October 2001 to September 2002 and collection tools included:

Quantitative data were analyzed using Excel (Microsoft Corp, Redmond, WA, USA). Calculations of referral time and median and mean referral delay were performed using either the spreadsheet or a scientific calculator.

Ethical approval was given by the National Health Sciences Research Committee. The research was discussed and planned with the District Health Management Team of Mangochi district. Prior to the research the project was explained and discussed with maternity staff, district transport officers, and drivers from the district hospital, as well as community leaders and health workers in the 3 study areas including traditional birth attendants (TBAs). Written informed consent was obtained from all women to be referred using a consent form with a standard text explaining the research; they were also given the option to be referred with the motorcycle ambulance or to call for a car ambulance from the district hospital.

3. Results 

During the 1-year study period a total of 194 patients were referred by motorcycle ambulance, of whom 68 (35%) were obstetric emergencies. No patient refused transport by motorcycle ambulance. Table 1 shows the obstetric cases referred for all means of transport. Nonobstetric referrals were mainly pediatric emergencies (such as severe anemia, malaria, pneumonia or meningitis, and injuries), but also included adult patients.

Table 1. Obstetric cases referred to Mangochi District Hospital by means of transport, 2001–2002

	Obstetric cases referred by motorcycle ambulance	No.	Obstetric cases referred by other means of transport	No.
	Obstructed/prolonged labor:	23	Obstructed/prolonged labor	10
	Complications of abortion	12	Complications of abortion	10
	Previous cesarean section	5	Previous cesarean section	3
	Transverse lie (1 with cord prolapse)	3	Breech (1 with IUD, 1 draining liquor)	3
	Antepartum hemorrhage	3	Retained placenta	2
	Primigravida <18 years (1 short stature not in labor, 2 in labor)	3	Twins (1 retained second twin, 1 with transverse lie)	2
	Intrauterine death (IUD)	2	Primigravida below <18 years (in labor)	2
	Breech	2	Intrauterine death	2
	Draining liquor	2	Eclampsia	1
	Retained placenta	1	Transverse lie	1
	Postpartum hemorrhage	1	Antepartum hemorrhage	1
	Eclampsia	1	Draining liquor	1
	Ruptured uterus	1	Cord presentation	1
	Puerperal sepsis	1	Cord prolapse	1
	Retained second twin	1	“Poor maternal condition”	1
	Face presentation	1	Severe backache in pregnancy	1
	High blood pressure and anemia	1	Stillbirth+anemia (Hb check)	1
	Severe malaria and pneumonia in pregnancy	1	Reason for referral not specified	1
	Severe anemia in pregnancy	1
	Psychiatric case in labor	1
	Cervical prolapse after delivery	1
	VVF and paralyzed leg after delivery	1
	Total	68		44

On average, 16 patients per month were referred by motorcycle ambulance, of which 5.7 were obstetric cases, or an average of 2 obstetric cases and 4 nonobstetric cases per month per health center. However, it is important to note that these figures do not represent the total number of obstetric referrals; some referrals were transported by other means such as private vehicles, bicycles, motorcycles, or the ambulance from the district hospital. Table 2 presents the total number of obstetric referrals by means of transport for each health center. Not included in Table 2 is one woman with postpartum hemorrhage who died at Makanjira HC shortly after arrival from home. Also excluded are 11 obstetric cases referred by TBAs and collected by motorcycle ambulance from their villages and taken to the health center; the objective of the research was to assess referral time from health center to referral hospital. Overall, 61% of all obstetric referrals were transported by motorcycle ambulance.

Table 2. Obstetric referrals to the district hospital by means of transport

	Health center	Obstetric cases referred by motorcycle ambulance	Obstetric cases referred by other means	Total
	Makanjira HC	13 (37%)	22 (63%)	35 (100%)
	Mase HC	28 (74%)	10 (26%)	38 (100%)
	Phirilongwe HC	27 (69%)	12 (31%)	39 (100%)
	Total	68 (61%)	44 (39%)	112 (100%)

3.1. Referral delay 

Only obstetric referrals were included in the calculations of the mean and median duration of the motorcycle ambulance trips from the 3 health centers to Mangochi. Duration of transportation was calculated from the time of departure to the time of arrival at the district hospital. At all 3 sites the delay between calling the rider and actual departure from the health center seldom exceeded 15 minutes, although some women who presented at night were transported the next morning. The mean and median duration of travel to Mangochi district hospital for each health center is shown in Table 3.

Table 3. Mean and median duration of transportation of obstetric referrals by motorcycle ambulance by health center

	Health center	Mean duration of referral	Median duration of referral	Range in minutes
	Makanjira HC	5 h	4 h 32 min	2 h 45 min to 7 h 26 min
	Mase HC	1 h 28 min	1 h 25 min	1 h to 2 h 10 min
	Phirilongwe HC	2 h 4 min	2 h	1 h 10 min to 3 h 15 min

Poor quality of data recording in the car ambulance logbooks at the district hospital did not allow for complete reliable and detailed data on the number and duration of obstetric referrals by car ambulances. The research project failed to get referral hospital staff involved in accurately recording information on received obstetric referrals that were not transported by motorcycle ambulance and information was obtained through interviews. Interviews with drivers and the transport officer from the district hospital and recording of personal travel time to the sites by the researchers revealed that a car ambulance takes at least 3 hours to reach Makanjira HC (4 hours during the rainy season if the road is passable), or 6 hours and 15 minutes to and from Mangochi assuming a 15-minute delay at the health center. For Mase, a car ambulance takes a minimum of 20 minutes to reach Mase HC, and 55 minutes to make the round trip. To reach Phirilongwe HC it takes at least 1 hour, or 2 hours and 15 minutes for the round trip. In the rainy season when road conditions are worse travel time is longer and Mase and Makanjira often cannot be reached by car. However, delay also occurs at district level after receiving a request for an ambulance and before it departs. When an ambulance call is received at the district hospital, a midwife in the maternity ward informs the transport officer who in turn arranges for the ambulance dispatch. This takes at least 15 minutes, but sometimes much longer; interviewees estimated an average of 30 minutes for the normal delay between an incoming call and the actual departure of the ambulance.

In reality, the situation leading to delay in referring obstetric emergencies at health center level by car ambulance or other means of transport was more complex. In Makanjira, in the case of daytime emergencies, someone from the health center had to go to the post office to phone Mangochi district hospital and ask for an ambulance, which was estimated to take 15 minutes. At night when the post office is closed, contact was made through the radio at Makanjira police station with the police station in Mangochi and a message had to be sent to the district hospital from Mangochi police station, which usually led to considerable delay. It was reported that messages were regularly received 1 or 2 days later.

Mase HC had neither means of transport nor means of communication with Mangochi, no public transport was available, and motorized transport was very rare on the road which passes Mase. This meant that someone had to go to Mangochi by bicycle to ask for an ambulance. It took 3 hours by bicycle to reach Mangochi. If a patient was carried by bicycle to Mangochi it took up to 6 hours to reach the district hospital.

The situation in Phirliongwe was similar to Mase. Phirilongwe HC had neither transport nor means of communication, no public transport was available and motorized transport was rare. When health workers from Phirilongwe HC wanted to call Mangochi to ask for an ambulance someone had to go 20 km by bicycle to Chilipa HC which had radio-communication; this took 2 hours.

Based on this information, using a motorcycle ambulance at Makanjira reduced referral time at least by 2.5 hours but usually by between 4.5 hours to more than 7 hours; this is a reduction of at least 35%. At Mase, referral time was reduced from 4.5–6 hours to 1.5 hours, a reduction of at least 68% to 76%. At Phirilongwe, referral time was reduced from at least 4.75 hours to 2 hours, a reduction of at least 58%. In general, referral delay was reduced by 2–4.5 hours or at least 35% to 76%, depending on the site.

When the decision to refer occurred after working hours is of particular interest because of the potential extra delay. For 20 patients referred by motorcycle ambulance the decision to refer was made between 5PM and 6AM. The median delay in these cases between decision to refer and departure was 2 hours 12 minutes (range, 30 minutes to 13 hours 10 minutes). For 8 patients the delay was less than 2 hours and for 6 patients it was more than 8 hours.

3.2. Costs 

The purchase price of a motorcycle ambulance was US $1965, excluding the shipping cost from South Africa to Malawi. Other investment costs included the purchase of a chain and padlock, rain suit and gumboots for the rider, and 2 crash helmets for each motorcycle ambulance. Per motorcycle these costs were US $163. With these costs included, one motorcycle ambulance costs US $2128. The Toyota land cruiser ambulance costs US $48 325. Both prices are duty free.

The mean total annual operating costs during the research period were US $508 per motorcycle ambulance, of which US $293 on maintenance and repair, US $197 on fuel, US $10 on insurance, and US $7 on a road license.

The total annual operating cost of the new car ambulance in Mangochi district for the year 2001 was US $12095, including US $6362 on maintenance and repair, US $4831 on fuel, and US $902 on insurance. However, government vehicles do not pay a road license fee. If we add US $44 for this to the annual operation costs of the car ambulance, total running costs would reach US $12139.

The car ambulance drove 72 680 km in 2001, while the mean distance for a motorcycle ambulance was 10504 km (range, 5619–14387 km). The operating costs (fuel, maintenance, and repair) for the car ambulance came to US $0.17/km, of which US $0.07 was for fuel and US$ 0.09 was for maintenance and repair. Based on recordings in the logbooks, fuel efficiency varied from 20–32 km per liter for the different motorcycle ambulances. For the motorcycle ambulance the cost of fuel was US$ 0.02/km and maintenance costs were US $0.03/km. Table 4 shows a comparison of specific operating costs between a motorcycle ambulance and a Toyota land cruiser car ambulance.

Table 4. Comparison of specific operating costs between a motorcycle ambulance and a car ambulance

		Motorcycle ambulance	Car ambulance
		(Yamaha AG 200+sidecar)	(Toyota land cruiser)
		US$	US$
	Comprehensive insurance (1 year)	10.45	901.65
	Road license (per year)	7.40	44.40
	Fuel consumption
	Per year	197.45	4830.70
	Per 10 km	0.20	0.70
	Km/liter	25 kpm	7 kpm
	Repair and maintenance
	1st year	508.20	6362.20
	2nd year	646.75	0.90
	Per km	0.30
	Tires (2004 prices)
	Rear tire	46.95	109.80
	Front tire	45.65	109.80

4. Discussion 

Not all obstetric referrals were by motorcycle ambulance. Sometimes it could not be used because the ambulance, the rider, or fuel was not available, or because there was more than one patient in need of transport; where possible the district hospital was then requested to send the car ambulance. Thus, placing motorcycle ambulances at rural health facilities does not eliminate the need for a car ambulance at district level, but it does reduce the demand for this ambulance and the cost of referral.

In Makanjira 63% of obstetric referrals were by other means of transport, while at the 2 other health centers most cases were referred by motorcycle ambulance. This was because during the study period the district health office failed to provide fuel for the motorcycle ambulances because of budget constraints. This risk was foreseen and discussed with community leaders and representatives prior to introduction of the motorcycle ambulances. In Mase and Phirilongwe communities were willing and able to pay for fuel if it was not available, but in Makanjira this option was rejected by the chief of the area who indicated that this is the government's responsibility. Thus, in Makanjira, patients relied on other means of transport for more than 3 months. The feasibility of motorcycle ambulances for referral of obstetric emergencies depends on supporting health systems, which must ensure availability of fuel, maintenance and repair, as well as on willingness of communities to pay for fuel if the health sector is unable to.

Mase HC referred many nonobstetric cases, including nonemergencies, because the health center had no clinician and all patients were treated by the nurse–midwife. The situation worsened after the nurse–midwife died and patients were treated by patient attendants (a health worker with little training who assists health professionals with tasks such as registering and weighing patients, dressing wounds, and dispensing drugs) or HSAs, who lack clinical skills and referred more patients for treatment. This also meant that no skilled health professional was available to authorize referrals based on clinical criteria.

Delays at night resulted from delay in getting consent from relatives for referral, decision to wait until the next morning, delay in mobilizing the rider, and sometimes reluctance of the rider to travel at night (uncertainty about receiving a night-out allowance in some instances played a role in this).

Seasonal factors such as the weather and condition of roads did not affect referral time greatly. Occasional long delays were the result of breakdown, such as a puncture. Motorcycle ambulances were able to transport patients throughout the year, even in the rainy season when the roads to Makanjira and Mase were not passable for other vehicles.

A serious limitation of the study was the inability to collect quantitative information on referral time of obstetric patients transported by car ambulance or other means of transport. This information could only be obtained through interviews. Although it had been agreed and expected that car ambulance drivers would record data in logbooks and maternity staff in the district hospital would complete referral forms from referred patients not transported by motorcycle ambulance, this did not materialize. For similar future research it is recommended to pay a ward clerk for follow up on referrals and recording of information or to post a research assistant in the hospital to do so.

The inclusion of some nonemergency cases may have resulted in some bias influencing median referral time as the rider may have driven more slowly than for emergencies. However, this is unlikely because riders in general were not able to differentiate between urgent and less urgent cases and consider all obstetric cases as emergencies, particularly since most women were in labor or fairly sick. It was observed that return journeys without patients or trips where no patients were transported usually took more time than trips with patients.

Capital costs for purchasing a motorcycle ambulance are much less than for a car ambulance and the price of a single 4WD Toyota land cruiser equals that of 22 motorcycle ambulances and their related costs. Including the shipping cost of a motorcycle ambulance, estimated at 20% of the purchase price, the price of one car ambulance would buy 19 motorcycle ambulances.

Average operating costs for a motorcycle ambulance in the first year of use (US $508) were 24 times cheaper than for the car ambulance including road license (US $12139). However, the car ambulance was much more intensely used, for a wide range of purposes, and travels more kilometers. If we take into consideration the difference in kilometers travelled by calculating operational costs per 10000 km, which was roughly the mean annual distance travelled by a motorcycle ambulance, operating costs of a car ambulance were 4 times higher than for a motorcycle ambulance. Costs of fuel consumption per kilometer are 3.5 times more for a car ambulance than for a motorcycle ambulance.

Costs for maintenance and repair of the motorcycle ambulances during the first year were low because the motorcycles only drove 10504 km on average in the first year. The highest distance was 14387 km for the Phirilongwe motorcycle. This means that during the first year not much was spent on spare parts. Routine replacement of several spare parts starts at 12000 km and 20000 km, such as replacement of the chain drive, front sprocket, cylinder head gasket, air cleaner, fit valve and brake liners after 12000 km. So operating costs, which mainly include spare parts and fuel are relatively low in the first year and start to increase in subsequent years. This issue did not apply to the car ambulance as it travelled far more kilometers (72680 km) in the first year because it was also used to pick up patients from other health centers in the district as well as for various other purposes, such as taking staff to meetings or workshops, transportation of medical supplies and the bodies of patients who died in the district hospital. In view of these additional expenses, the operating costs of a motorcycle ambulance were estimated at US $647 for the second and subsequent years, which is still 18 times less than the running costs for the car ambulance in the first year.

Additional costs of the motorcycle ambulance transport system were allowances for the riders, such as allowances when they had to spend the night in Mangochi. These expenses are not included in the above mentioned operating costs of the motorcycle ambulances, nor are extra allowances paid to drivers of the car ambulance, drivers' salaries, or opportunity costs to compensate for riders' time used for transportation of patients. These night-outs varied between 0 and 4 nights per month per rider, with an average of 1 night per month. The issue of payment of allowances led to disagreement between the riders and the administration of the district hospital, which was resolved by paying a fixed rate of US $14 per month to each rider.

An advantage of motorcycle ambulances as observed from the logbooks is that they are much less likely than car ambulances to be misused for nonhealth-related purposes; this also contributes to reduction of transport expenses.

Another advantage of motorcycle ambulances at health centers is that they are on site and can depart with little delay, while a car ambulance from the district hospital first has to travel to the referring health center then back to the referral hospital.

5. Conclusions 

Motorcycle ambulances reduce the delay in referring women with obstetric complications from remote rural health centers to the district hospital, particularly under circumstances where health centers have no access to other transport or means of communication to call for an ambulance.

They are also a relatively cheap and effective option for referral of patients in developing countries, particularly in rural areas with little or no public transport. Nineteen motorcycle ambulances can be bought for the price of one Toyota land cruiser car ambulance. Operating costs compare in a similar way. Motorcycle ambulances also potentially help reduce costs for women and their families to access EmOC, although this was not the subject of this study.