The cost of the district hospital: a case study in Malawi.

The cost of the district hospital: a case study in Malawi A.J. Mills,1 J. Kapalamula,2 & S. Chisimbi2 Described is an analysis of the cost to the Ministry of Health of providing district health services in Malawi, with particular emphasis on the district hospital. District resource allocation patterns were assessed by carefully disaggregating district costs by level of care and hospital department. A strikingly low proportion of district recurrent costs was absorbed by salaries and wages (27-39%, depending on the district) and a surprisingly high proportion by medical supplies (24-37%). The most expensive cost centre in the hospital was the pharmacy. A total of 27-39% of total recurrent costs were spent outside the hospital and 61-73% on hospital services. The secondary care services absorbed 40-58% of district recurrent costs. Unit costs by hospital department varied considerably by district, with one hospital being consistently the most expensive and another the cheapest. A total of 3-10 new outpatients could be treated for the average cost of 1 inpatient-day, while 34-55 could be treated for the average cost of 1 inpatient. The efficiency of hospital operations, the scope for redistributing resources districtwide, and the costing methodology are discussed. Introduction Very few in-depth investigations have been carried out on the economics of hospitals in developing countries, despite their significance in terms of sec- tor expenditure (1, 2). The present study reviewed the allocation of Ministry of Health resources within six districts in Malawi, with particular emphasis on the proportion absorbed by the hospital and on the unit costs for hospital departments. The objectives were to develop policies and guidelines to increase the efficiency of hospital operations and to improve the intrasectoral allocation of health sector resources. Health care in the study districts was the respon- sibility of a district health officer (DHO), who held a budget that covered the non-salary expenditure of all government facilities. Ministry of Health services were largely provided free. The responsibility for the day-to-day running of the hospital was usually del- egated to a chief or senior clinical officer and to the sister-in-charge and hospital secretary. A district health inspector was responsible for preventive ac- tivity, reporting both to the DHO and the regional health inspector. A number of staff based at the dis- trict hospital had districtwide responsibilities, most notably the DHO, district health inspector, maternal I Reader, London School of Hygiene and Tropical Medicine, and London School of Economics and Political Science. Requests for reprints should be sent to Dr Mills at London School of Hygiene and Tropical Medicine, Keppel Street, London WC1 E 7HT, England. 2 Ministry of Health, Lilongwe, Malawi. Reprint No. 5391 and child health (MCH) coordinators, and tuberculo- sis, leprosy and ophthalmic assistants. The Ministry of Health services in all districts were complemented by mission facilities, ranging from general hospitals similar to the ministry's own district hospitals to simple dispensaries, and by district council units. These facilities were not examined in the present study but clearly are important in planning the over- all provision of district health services. Materials and methods Because of physical constraints, districts had to be purposively rather than randomly selected. The main aim of the selection process was to choose a group of districts containing hospitals that reflected the range of types of district hospital in Malawi in terms of size, services offered, staffing, age of buildings, and geographical location. Total district costs were compiled from several sources, as described below. Expenditure controlled at the district level With the exception of salary costs, most of the recur- rent expenditure of districts was identifiable at the level of the individual districts. A particular difficul- ty arose in calculating the costs of drugs and medical supplies, since annual expenditure was not necessari- ly a good reflection of the value of the drugs used (due to changes in stock levels and donated items). To overcome this and to estimate the cost of drugs according to hospital department, we examined a selection of both the duplicate books used by hospi- Bulletin of the World Health Organization, 71 (3/4): 329-339 (1993) © World Health Organization 1993 329 A.J. Mills et al. tal wards and departments to order drugs (usually daily) from the pharmacy and the monthly requisi- tion forms from health centres. The prices charged by the central medical store were applied to the quantities of drugs used to obtain the total cost; this was then converted to a cost per day and hence an annual cost. Donated drugs were therefore costed at the prices charged by the medical store for equiva- lent drugs. Vaccine costs were estimated on the basis of the quantities used and price CIF (cost-insurance- freight) Malawi. The annual amount of X-ray film used (by film size) was noted from departmental records and its cost allocated to the wards and outpatient depart- ment, according to the origin of patients X-rayed and the film size (both these were routinely noted in the X-ray department ledger). Vehicle logbooks were analysed for 1 month in order to allocate vehicle costs to user departments. Costs of salaries Salaries were paid centrally; district salary costs could only be calculated for two broad groupings: senior staff and junior staff. Since the cost apportion- ment required more detailed data, salary costs were estimated from staff numbers by type and grade and average salaries (mid-point of the salary scale). The total actual costs were used to cross-check this pro- cedure: the actual and estimated expenditures were in general extremely close. To allow for employment benefits, we added 10% to the annual salary cost of pensionable staff. Subsidized housing represented an important category of employment benefit and was treated as a capital rather than a staff-related cost; however, its significance as an employment benefit is discussed below. Other recurrent costs Specialists from the two central hospitals made fairly regular visits to district hospitals. These visits were not taken into account in the study since they were fairly similar for most hospitals, it was difficult to estimate accurately the time involved and the amount of work done, and the sums involved were probably small. A few items, e.g., most postal charges, were paid centrally and not charged to individual districts. Such charges were not included and were probably, in any case, an insignificant proportion of total dis- trict expenditure. Capital costs Capital costs comprised hospital buildings and equipment, rural facilities and equipment (rural hospitals, health centres, dispensaries, maternity units), staff housing, vehicles and motorbikes. Such costs were in general based on the estimated replace- ment cost of a standard unit (hospital, house, vehicle, etc.), unadjusted for location. Costs were annualized, assuming a lifespan of 30 years for buildings, 10 years for equipment, 3 years for linen and bedding, 10 years for ambulances, 6 years for a car, and 4 years for a motorbike. A real rate of interest of 3% was used in calculating annual costs to allow for the opportunity cost of capital purchases. The ultimate purpose of the analysis was to allo- cate all costs to departments that provided direct patient care, i.e., the wards, outpatient department, matemal and child health department, and rural fa- cilities. Diagnostic and support departments such as the operating theatre, X-ray facility, and laboratory were considered to provide services to these direct patient care departments. The methodology adopted, therefore, was slightly different from the standard hospital costing methods used in the USA (3), where costs are primarily allocated to "revenue-producing" departments and no clear distinction is made be- tween direct patient care departments and diagnostic and support departments. The Ministry of Health accounting system did not distinguish between hospital and non-hospital expenditure. We therefore used an estimation pro- cedure to apportion costs between not only hospital departments but also the hospital and rural facilities. This procedure involved three stages: in the first, expenditure by line item was grouped into a smaller number of categories; in the second, costs were allo- cated to cost centres; and in the third, the costs of cost centres were stepped down to the departments providing direct patient care. The cost centres are listed in Table 1, and the criteria used to allocate costs to them are shown in Table 2. Most are self-explanatory, with the excep- tion of staff costs. The majority of staff could be directly allocated to cost centres on the basis of their duties or a rota. The remaining staff were either asked how they spent their time or this was esti- mated by the DHO or matron. In the step-down pro- cedure, the order is important since the costs are assigned from a particular cost centre to subsequent cost centres and not to those that precede it. A more elaborate apportionment procedure was not consider- ed worthwhile, however, since the costs to be appor- tioned were estimates. The criteria whereby costs were distributed to direct service departments are shown in Table 3. The cost analysis used activity statistics for some of the apportionment procedures and for the production of unit costs. Unfortunately, few hospitals paid much attention to inpatient statistics, which were frequently incomplete. The information avail- WHO Bulletin OMS. Vol 71 1993330 District hospital costs in Malawi Table 1: List of cost centres used In the study Administration of district Maintenance of buildings and grounds for district Transport Domestic Laundry Kitchen Pharmacy Laboratory X-ray department Operating theatre Male ward Female ward Paediatric ward Tuberculosis ward Hospital outpatient department Hospital maternal and child health department Environmental health department at hospital Rural hospitals Health centres and primary health care able, other sources of data (e.g., ward admission books and daily inpatient census records), and the advice of district staff were used to estimate missing data or to compile alternative statistics. The statistics on admissions were probably more accurate than those on inpatient-days. Outpatient statistics, when available, recorded only new outpatients and not the total number of visits. Where possible, estimates of total visits were also obtained. Results Table 4 shows the distribution of recurrent and total district costs, by broad input category. There are two striking features: the low proportion of recurrent costs absorbed by salaries and wages (27-39%); and the high proportion (46-57%) absorbed by the annual value of capital, of which buildings and equipment were the most significant and transport very small. As a proportion of total recurrent costs, the costs for drugs and medical stores accounted for 24-37% and vehicle-associated costs for 11- 18%. Food, the only other significant item of recur- rent expenditure, comprised 3-7% of total recur- rent costs. When costs were distributed by cost centre, the pharmacy was the most costly department, absorbing 25-38% of total recurrent costs. Administration accounted for 5-9%, maintenance for under 4% (in four districts for <2%), and kitchen costs for 5-9%. The medical support departments apart from the pharmacy-laboratories, X-ray facility and operating theatre-had low direct costs. Table 5 shows total recurrent costs distributed to direct service departments and Table 6 the unit recurrent costs per bed, inpatient, inpatient-day, new outpatient, and outpatient. Deficiencies in hospital statistics precluded calculating all these unit costs for each district. Table 2: Criteria for allocation of costs to cost centres Cost item Allocated to: Recurrent costs Domestic supplies Food Vehicle-associated Heating, lighting Hospitality Staff-associated Maintenance, buildings Maintenance, office equipment Maintenance, plant Postage, printing Public transport (i.e., travel warrants for patients) Medical supplies Subsistence, transport claims Telephone, telex Water, sanitation Bicycle maintenance Uniforms (staff and patients) Salaries and wages Capital costs Buildings Equipment Staff houses Vehicles Motorbikes Domestic cost centre Kitchen cost centre Transport cost centre In proportion to likely energy use of cost centres Administration cost centre In proportion to salary costs by cost centre Maintenance cost centre Administration cost centre According to plant location Administration cost centre Wards in proportion to in- patients Pharmacy 10% to administration, 90% to health centres and PHC Administration cost centre In proportion to likely use by cost centres Health centres and PHC 88% to wards in proportion to inpatients, 1.5% to OPD, 0.5% to MCH, 10% to health centres According to distribution of staff time by cost centre Hospital, in proportion to floor space by cost centre; to health centres etc., directly Hospital equipment list broken down by cost centre; health centre equipment estimated For hospital houses, in proportion to salary costs Transport Transport a PHC = primary health care; OPD = outpatient department; MCH = maternal and child health. WHO Bulletin OMS. Vol 71 1993 331 A.J. Mills et al. Table 3: Criteria for the distribution of the cost of cost centres to direct service departments Cost centre Distribution criteria Administration Direct costs of cost centre Maintenance 30% to health centres; 70% to hospital distributed on the basis of floor area Transport Shared between administration, wards, and health centres and PHC based on the analysis of logbooks: the administration share was allocated in proportion to the direct costs of cost centres following the transport cost centre; the ward share to wards in proportion to inpatients Domestic Floor area of hospital after health centre supplies had been subtracted Laundry Share of inpatient-days by ward Kitchen Share of inpatient-days by ward Pharmacy Estimation of actual use of drugs and medical supplies Laboratory Estimated allocation of time between inpatients and outpatients, and to wards in proportion to inpatients X-ray Actual use of X-ray department by OPD and wards Operating theatre To maternity department in relation to Caesarian sections as a proportion of total major (weighting = 1) and minor (weighting = 0.5) operations; remainder in proportion to male and female surgical inpatients (if distinguished separately) or male and female inpatients a PHC = primary health care; OPD = outpatient department. The distribution of recurrent costs in Table 5 gives an accurate indication of the resources ab- sorbed by rural services (27-39%), as opposed to the district hospital. While the hospital accounted for 61-73% of the district recurrent costs, part of this was spent on primary health care services for the local population; the secondary care services pro- vided by the hospital consumed 40-58% of district recurrent costs. There was a considerable variation in the relative importance of the different wards, depending on the hospital. In Salima and Kasungu hospitals, the matemity ward had the largest share, in Chiradzulu and Rumphi, the children's ward, and in Mulanje and Nkhata Bay, the male ward. However, the differences were not particularly large except where there were major variations in bed numbers between the wards (for example, the small matemity ward in Mulanje). The unit recurrent costs (Table 6) varied con- siderably between the hospitals. The most expensive hospital was consistently Mulanje, and the cheapest, Rumphi: in terms of cost per bed, Mulanje was 75% more expensive than Rumphi; in terms of cost per inpatient, 83% more expensive; and in terms of cost per inpatient-day, 234% more expensive. These dif- ferences should, however, be interpreted with cau- tion since they are highly sensitive to the accuracy of the activity statistics used. Mulanje reported a bed occupancy rate of 76%, which was implausibly low. The range of unit recurrent costs is probably better represented by excluding Mulanje, whereupon the cost per inpatient varied between K (kwacha) 45.42 Table 5: Distribution of recurrent costs, by direct service department, in the six study hospitals, 1987-88 % distribution in: Departmenta Chiradzulu Kasungu Mulanje Nkhata Bayb Rumphi Salima Male ward 11.4 11.7 12.6 13.3 9.5 12.0 Female ward 9.4 11.7 8.1 9.3 10.7 11.9 Children's ward 13.8 7.7 9.7 11.0 12.7 13.1 Maternity 11.9 14.2 6.3 10.1 10.3 13.9 Tuberculosis ward 11.5 2.4 3.6 5.2 1.4 4.5 Poliomyelitis annex - - - - 3.9 - All wards 58.0 47.7 40.3 48.9 48.5 55.4 OPD 10.8 11.2 18.2 10.6 14.7 14.3 MCH/district hospital 2.3 2.8 1.7 1.8 2.4 2.0 Environmental health/district 1.5 1.3 1.0 2.1 2.0 0.9 Rural hospitals - 5.1 - 12.2 16.2 - Health centres and PHC 27.4 31.9 38.8 24.4 16.2 27.4 All non-wards 42.0 52.3 59.7 51.1 51.5 44.6 Total 100.0 100.0 100.0 100.0 100.0 100.0 a OPD = outpatient department; MCH = maternal and child health; PHC = primary health care. b The urban health centre at Mzuzu was included under rural hospital category. WHO Bulletin OMS. Vol 71 1993332 District hospital costs in Malawi 1N- CD N Nm 0 o N- Cq 0 N- 00)C~~~) N 0 L N- 00 aco(' 0 N- 0 C N '- 0 N LO) 0 (0q CO (0 N-~ N 0 CO N coC) 0 mNa C 0 Nocm 0 0 N9 0 qC LON 0 N- N C-0 - 0) CO ( (0 0)CO'- 0 N-~~,I C) co E (0-C (0 CO C/) co 0) 0 0 >, C.) .0 C 4- CI N - C' N- ' 75 lt 1, n 0 P -0), =30(0 N-NOM C') Oc U) CO N- NM (000CCIO C6L6 L6 ' o5 0. .0 C U) 0 U~~~~~~~~~~~~C 0~~~~~~ C/)~~~~~~U 0:3~~~ .0 0 0) a) ~ ~ C 0 L) U ) 0) 0Ic 0o U) 0) Cu " H- CQLL >IICO: WHO Bulletin OMS. Vol 71 1993 N- OD 0 (0 CO 0) CO 0 CO a)6 CO) C') (0 N- 0 0 N 0 CM 0) 0 6) 0) 0 0i 0 0 0 0 0 0 0 0>6 CM N-(0 0~ Nl-~ (0 0 Oc Clf) (0(0 N-(0 N\ 0 NM N- (0 C\, 'C C\O CY) 'C\ N-~ C'f) (0 N- CY) 0 co N- (0 N- co(0 N- 6\ U) E 0) 0 N- 'Cf LO(0l 6Y (0 NM 0 (0 cc N- CO N- 0 -1- CU 0 C C c 0 0 3 13 0 0 0 0 0 0 0 0 0 0 C1 cu C E 0) 0 CO (0 0 N- CO 0 (0 0 N- 0 N Ul) 0 c Nl-cc (0 CO 0 (0 N- C,, (0 N- (0 (0 LO (.0 C,,) CO C,) CY) CO CY) NC\ CM (0 C,) CY) 0> 0> 0i 0> 0 U) ( U) CY 0 U) CO 0 co U) C ) 0 0( Cus .~~ _ -~~ -~~ .2- a). e 0Cu ~~~~ CTL 0Ca) HO Wr CD V) 0 C. 0 a U) V3 co) 0- (0 (0f N- 0q 0~ (0 N- (0 N- N- (0 (0 OD 0 co C') 0. E CE cu COco z cu 0) (0 CO) CO 0) Nl- co 0! (0) L-('0 04 N- (0 (0 LO (0 6q 0 0 0 0- 0 0) 0) 0 0 0 6 0 06! N- 0f N; (0 (0 0 NI 0= (0 0 (0 0 N 0 0 0 N 0 0 0 6 0) 0 0i 0 0 0 0 0 0 0 0 (.0 0) 0 (0 CO 0 0 N- 0) CO CO0 (0 cc 0 0 (0 N- 0 0O (0 0 (0 cc CO N- (0 0 0; 0 0 0) 0) 0 0) 0 0 0) 0 0 0 cc 00 cc CC) cc 0 co(0 0~ CO CD cc N-~ CM Ict (0) CY) (.0 co r'C (0 CM (0(0 LO N- 0) (0 CO N- cc N- (0CR NC (0 c1)(0 N (0 0) LO N 0 (0) N LO N\ U) co 0 0 H 333 A.J. Mills et al. Table 6: Unit recurrent costs in kwacha (K) (1US$ = K 2.21) In the six study hospitals, 1987-88 Recurrent costa Chiradzulu Kasungu Mulanje Nkhata Bayb Rumphi Salima Per bed per ward Male ward Female ward Children's ward Maternity ward Tuberculosis ward Poliomyelitis annex All wards Per inpatient by ward Male ward Female ward Children's ward Maternity ward Tuberculosis ward Poliomyelitis annex All wards Per inpatient-day by ward Male ward Female ward Children's ward Maternity ward Tuberculosis ward Poliomyelitis annex All wards Per new outpatiente Hospital OPD Rural hospital No. 1 No. 2 No. 3 Health centres All 2091 2036 5962 3143 2104 2680 54.41 52.69 29.77 44.90 303.87 50.26 4.37 4.14 5.65 5.74 8.94 5.43 0.91 2906 2900 3986 4347 1007 3046 72.85 49.41 33.37 38.56 139.84 47.10 11.40 8.05 8.16 9.64 5.81 8.98 0.91 3305 3051 3533 11 053 1351 3239 100.16 79.39 74.14 63.04 136.09 83.06 12.72 13.22 11.12 21.49 5.44 11.75 2.48 1.10 0.76 0.79 1.88 1.42 NAd 2936 2053 5557 2942 1433 2704 78.35 58.34 38.59 43.97 125.13 55.18 10.87 8.90 8.98 12.21 6.91 9.64 1342 1515 3309 3034 1634 2053 1846 53.55 48.86 38.54 59.33 166.18 52.64 45.42 3.65 3.56 3.87 4.63 6.65 2.92 3.52 2715 2690 2976 3151 1082 2540 70.13 67.76 57.37 56.86 131.71 64.88 6.07 7.99 8.53 8.95 4.39 7.31 1.22 1.14 1.53 0.92 c 0.90 - - 1.40 - - 1.15 - 1.43 1.13 1.17 1.14 1.13 1.27 a OPD = outpatient department. b Estimated cost per outpatient visit was K 0.64 (Kasungu OPD); K 0.57 (Mulanje OPD). c Urban health centre at Mzuzu was included under rural hospital category. d NA = not available. (US$ 20.55) in Rumphi and K 64.88 (US$ 29.36) in Salima, and the cost per inpatient-day between K 3.52 (US$ 1.59) in Rumphi and K 9.64 (US$ 4.36) in Nkhata Bay. The average hospitalwide unit cost disguises considerable differences between the various wards. In terms of cost per inpatient, after the tuberculosis ward, which was costly because of the long length of stays, the male ward was the most expensive, and the children's and maternity wards the cheapest. In terms of cost per inpatient-day, the male and female wards tended to be cheaper than the children's and materni- ty wards. The cost of the tuberculosis ward was to some extent determined by whether rifampicin (a very expensive drug) was supplied during the period analysed. Excluding the tuberculosis ward, maternity was the most expensive in terms of cost per inpatient- day in five of the six hospitals. Outpatient statistics were incomplete and com- parisons were therefore more difficult. Also, while the cost per hospital outpatient reflected the cost of the outpatient department alone, the cost per out- WHO Bulletin OMS. Vol 71 1993 District hospital costs in Malawi patient in a health centre or rural hospital included the cost of other activities (e.g., maternity ward for many health centres and general inpatient activities for rural hospitals) that could not readily be indi- vidually assigned. The cost per new hospital out- patient varied between K 0.91 (US$ 0.41) for Chi- radzulu and Kasungu to K 2.48 (US $ 1.12) for Mulanje, with the other hospitals in the range K 1.14-1.53 (US$ 0.52-0.69). The cost for Mulanje is probably an overestimate, because of overesti- mates of outpatient drugs. On the whole, those hos- pitals that provided cheaper inpatient care also pro- vided cheaper outpatient care, though Rumphi's extremely low inpatient unit costs were not repli- cated in the outpatient department. Allowing for the distortion in the cost per health centre outpatient and rural hospital outpatient, it seems likely that health centres and rural hospitals provided outpatient care at a slightly lower cost than the district hospital. It should have been possible to estimate the unit costs of the various cost centres, particularly the laboratory and operating theatre. Unfortunately, however, neither of these kept records detailed enough to quantify their activities or to weight them for their intensity of resource use. The cost of the kitchen was divided by the number of inpatient-days to produce a unit cost (Table 7). This varied consider- ably between hospitals, as did the cost of the food component. The variation primarily arose because of differences in expenditures on food relative to the number of inpatient-days, some hospitals such as Rumphi and Salima being relatively economical, and also because of the cost of water and heating. The addition of annualized capital costs to recur- rent costs only altered slightly the ranking of hospi- tals in terms of unit costs; this is not unexpected, since capital costs were based on a common replace- ment cost for a standard hospital and adjusted only for the number of beds. The difference in the unit costs of outpatients and inpatients illustrates the scope for increasing efficiency by replacing inpatient care by outpatient Table 7: Cost hospitals of providing food in the six study Cost of the kitchen per inpatient-day (in kwacha) Food as a % of cost Chiradzulu 0.80 80 Kasungu 1.48 73 Mulanje 1.95 66 Nkhata Bay 2.04 71 Rumphi 0.52 63 Salima 0.74 56 care if the characteristics of patients permitted. The ratios of the recurrent cost per new outpatient, per inpatient-day, and per inpatient, respectively, were as follows in the six hospitals: - Chiradzulu, 1: 6: 55; - Kasungu, 1: 10 : 52; - Mulanje, 1: 5: 34; - Nkhata Bay, 1:8 : 45; - Rumphi, 1: 3: 40; and - Salima, 1 : 5 : 42. Discussion Mix of inputs The cost of manpower was low, partly because of the provision of subsidized housing to staff. Also, the rates of pay were below their true market level, both relative to the private sector in Malawi and interna- tionally. Since the prices paid for imported inputs were similar to international levels, the proportion of the total expenditure taken up by salaries and wages was lower than in countries with a different price structure of inputs. The low salary and wage share was also a reflection of the mix of staff. Medical officers consti- tuted only 3-6% of overall staff costs in the six dis- tricts; technical and technical assistant grades, 25-33%; and nurses, 21-37%. There was thus heavy reliance on technical, auxiliary and nursing staff. In contrast to many African countries, staff were provided with the resources with which to work. For example, there was little evidence, on the whole, that inputs of pharmaceuticals were in short supply, although shortages and overstocking of certain items occurred because of poor management at the hospital and central medical stores. Transport services did not appear to be constrained by lack of money for fuel; rather, the availability of vehicles in operating order was the main problem. The annual value of capital represented an extremely high proportion of total costs. Probably capital (especially imported building supplies and equipment) was expensive relative to recurrent inputs, not least because salaries and wages were low. However, further study is required to determine whether the capital costs used were overestimates. Unit costs Unit costs are an important guide to efficiency, but need to be interpreted cautiously since the available indicators of output, such as bed-days, ignore any quality differences between hospitals. Often, how- WHO Bulletin OMS. Vol 71 1993 335 A.J. Mills et al. ever, the differences in unit costs between the hospi- tals were so great that quality was unlikely to account for all the variation. Moreover, the interpre- tation of certain unit costs, for example, the cost of food per inpatient-day, is less affected by quality considerations. The variation in unit recurrent costs between the six hospitals requires comment. Chiradzulu is a small hospital without a doctor or X-ray equipment and therefore referred many cases elsewhere; not surpris- ingly, it had low unit costs. Rumphi, which is the referral hospital for the northem region and was staf- fed by two doctors, had unit costs that were almost as low. Differences in the type of hospital therefore did not inevitably lead to differences in unit costs. Because a significant proportion of hospital costs are fixed, the intensity with which resources are used is likely to influence strongly unit costs. The first step in elucidating the unit costs is therefore to examine bed occupancy rates and lengths of stay (Table 8). The two hospitals with the highest bed occupancy rates, Chiradzulu and Rumphi, had the lowest unit costs. Salima was operating at around 100% capacity and Nkhata Bay and Mulanje were lower (although there were discrepancies in the stat- istics). Lengths of stay varied considerably, being lowest overall in Kasungu and Nkhata Bay and highest in Rumphi; this suggests scope for reducing the length of stay, notably in Rumphi but also in Salima, Mulanje, and Chiradzulu. Compared with other countries, however, the lengths of stay were low. It is very unlikely that case-mix differences would account for all the variation in length of stay, especially between Kasungu, Mulanje and Salima, although Rumphi did serve a regional role. Although bed occupancy rates and lengths of stay account for some of the unit cost variation between hospitals, particularly the low cost per inpatient-day in Rumphi, they do not explain all of it; there was a clear difference in the efficiency with which hospitals used resources that was independent of occupancy rates or length of stay. It is also impor- tant to distinguish those costs that could be locally controlled and those over which the hospital had no influence. Examples of the latter were the source of heat for the kitchen and laundry (mains electricity or wood) and the technical efficiency of the equipment; and of the former, drugs and food. Detailed analysis of supplies of drugs to outpatient departments and wards indicated considerable variations in practices between hospitals, which were probably not related simply to the characteristics of patients: the cost of Table 8: Bed occupancy rates and length of stay statistics for the six study hospitals Male Female Children's Maternity Tuberculosis Poliomyelitis All Hospital ward ward ward ward ward annex wards Chiradzulu BOR (%) 131 135 289 150 64 - 135 LOS (days) 12.4 12.7 5.3 7.8 34.0 - 9.3 Kasungub BOR (%)c 69.8 98.7 133.8 100.0 47.5 - 88.0 LOS (days) 6.4 6.1 4.1 4.0 24.1 - 5.2 Mulanje BOR (%)c 71.2 63.2 87.0 140.9 68.0 - 75.5 LOS (days) 7.9 6.0 6.7 2.9 25.0 - 7.1 Nkhata Bay BOR (%)c 74.0 63.2 169.6 66.0 58.5 - 76.9 LOS (days) 7.2 6.6 4.3 3.6 18.1 - 5.7 Rumphi BOR (%) 100.6 116.6 234.1 179.5 67.4 192.3 143.6 LOS (days) 14.7 13.7 10.0 12.8 25.0 18.0 12.9 Salima BOR (%) 122.6 92.3 95.6 96.4 91.4 - 98.2 LOS (days) 11.6 8.5 6.7 6.4 30.0 - 9.2 a BOR = bed occupancy rate; LOS = length of stay. b Official statistics except maternity, where 100% BOR and 4-day c Probably underestimated. ,LOS were assumed. WHO Bulletin OMS. Vol 71 1993336 District hospital costs in Malawi drugs per new outpatient varied between K 0.55 (US$ 0.25) and K 1.84 (US$ 0.83); in one hospital, supplies of tablets and injections were strictly controlled, but large amounts of bandages were regu- larly supplied; some hospitals used large quantities of intravenous fluids and others less; and use of anti- biotics and injectables varied enormously. The find- ings on the cost of food and kitchens indicated substantial differences that could not be explained by regional price variations. Districtwide distribution of resources The distribution of resources districtwide was deter- mined to a considerable extent by the physical infra- structure of rural facilities. This helps to explain, for example, why health centres and primary health care absorbed 27% of district recurrent costs in Chiradzulu (which had few rural units under its control) and 39% in Mulanje (which had many rural units). There was a fair amount of local flexibility, however, since a considerable number of staff and other resources were based at the hospital to serve the district as a whole. Two indicators of the effort expended in dis- trictwide activities are the proportion of time the DHO spent on district supervision and the proportion of vehicle mileage devoted to health centre and pri- mary health care activities. The former varied be- tween 9% and 25%: most DHOs spent one day per week on districtwide activities with a further 30-40% of time being spent on district administra- tion. The DHO thus found it difficult to avoid spend- ing the great majority of his time in the hospital. Vir- tually no other senior staff were involved in district supervision. Analysis of the vehicle logbooks indicated that only about 20% of mileage was associated with pri- mary health care and that the great majority was devoted to the transport of inpatients. The require- ment to have vehicles available to transport patients was a major factor that precluded their use for other activities, particularly primary health care. Conclusions The cost analysis permitted conclusions to be drawn under the following headings: the efficiency of hos- pital operations; scope for redistribution of resources districtwide; and the costing methodology and scope for more frequent cost analysis. Efficiency of hospital operations The study identified areas where improvements in efficiency might be feasible, for example, through better drug purchasing, stock control, distribution, and prescribing. There was scope for economies in some hospitals, particularly for food purchasing and vehicle running costs. In contrast, the sums spent on maintenance appeared to be inadequate. Although capital costs were high relative to recurrent costs, there was little evidence that buildings were unduly lavish. Building standards had improved in recent years, new hospitals being more spacious than old ones, permitting greater efficiency. Many of the problems in the hospitals resulted from organizational weaknesses, particularly in rela- tion to the pharmacy, supplies, and transport. Since the pharmacy was the biggest single manager of resources in the whole district, there is a strong case for strengthening its management. Scope for reducing inpatient numbers and the lengths of stay was relatively limited, although in view of the low cost of diagnostic services compared with the total hospital costs, it should be investigated whether improved laboratory services would permit better standards of care and shorter stays. Most district hospitals were limited to a very basic range of haematological, biochemical and parasitological tests, other tests being referred to Blantyre or Lilongwe (50-60 cases per week in the case of Salima). The outpatient department was considered to be overused for minor ailments, and in highly populated districts such as Kasungu and Mulanje was viewed as a drain on hospital resources, especially staff time. One way of tackling this would be to charge nominal fees to discourage those with minor complaints. Further investigation of health centre costs is required, since they are an alternative source of care for outpatients and maternity cases. In some hospi- tals maternity care was one of the more expensive types of inpatient care, accounting for 10-15% of recurring costs, suggesting that further decentraliza- tion of maternity care might be desirable. Redistribution of district resources In recent years there has been a notable increase in the number of staff based at the district hospital who have districtwide responsibilities. Availability of transport is vital to the efficient use of their time: without it, they are presumably underemployed unless, as with MCH services, there are sufficient hospital-based activities to occupy them. The diffi- culty posed by transport was not simply a shortage of vehicles but also one of maintenance and priori- ties: district outreach activities received lower priori- ty for transport than administrative activities or transport of patients. There was a similar problem with regard to the DHO's time: hospital business WHO Bulletin OMS. Vol 71 1993 337 A.J. Mills et al. tended to displace primary health care despite good intensions to the contrary. There is no easy solution to changing priorities, which must include increasing staff motivation. Vehicles have been assigned to par- ticular activities and this may help as will improving vehicle maintenance. Provision of vehicles unsuit- able for patient transport (e.g., motorbikes) could be increased, provided maintenance is ensured. Acknowledgements We are grateful to the staff of the six districts visited and to the officials of the Ministry of Health who transcribed information for us. The fieldwork was funded by the World Bank (Hospital Work Programme) and the final writing up by Overseas Development Administration support to the London School of Hygiene and Tropical Medicine Health Economics and Financing Programme. Costing methodology Our analysis has shown that detailed costing of hospital operations is feasible in a country such as Malawi, despite the inclusion of rural facility and hospital expenditures in a single district budget, and the poor availability of information on district salaries and wages. Indeed, the greatest problem lay not in the accounts but in obtaining accurate activity statistics. Such costing can lead to useful conclusions on district resource allocation pattems and hospital efficiency. This type of analysis cannot be performed routinely in Malawi with the current accounting system, but marginal changes could facilitate the more regular production of unit costs that would be useful in planning and management. Minor changes could be introduced in routine information systems that would permit health centre and hospital costs to be separated. Analysis of intra-hospital costs would nevertheless remain a problem. However, our findings suggest that overall unit costs (not distinguished by department) per outpatient, per inpatient, and per inpatient-day give a reasonable indication of the rela- tive costliness of different hospitals and thus could be used to monitor efficiency, supplemented by readily calculable unit costs such as the cost of food per inpatient-day. This would simplify the drug analysis, and make it feasible to use pharmacy tally cards to note the issue of supplies for inpatient and outpatient use. Use of a prospective rather than a retrospective analysis would enable temporary changes to be made in information systems to facilitate the study and would encourage better recording of activity statistics. The results of the study are relevant also to other countries. Detailed examination of the allocation of resources within hospitals; especially of staff time and drugs, is a valuable exercise. Investigation of the distribution of resources between inpatients and out- patients is particularly important because of its varia- tion in different hospitals; it should not be appor- tioned using an assumed ratio of cost per outpatient to cost per inpatient-day. It is also important to examine the hospital's role in the district health system and the extent to which its resources and skills serve the district as a whole. Resume Cout de l'hopital de district: une 6tude de cas au Malawi L'article presente les resultats d'une 6tude sur les ressources allou6es par le Ministere de la Sante dans six districts du Malawi, I'accent portant en particulier sur la proportion absorbee par l'h6pital de district et sur la d6termination des coOts uni- taires pour les services de l'h6pital de district. L'objectif etait de choisir les options propres a ameliorer l'efficacite de l'h6pital et I'allocation des ressources a l'interieur de chaque secteur. Six districts correspondant a une fourchette d'h6pitaux de district diff6rents ont ete choisis pour l'etude. Les donn6es sur l'investissement initial et les coOts renouvelables ont 6te obtenues ou evalu6es en cas de besoin. Ces chiffres ont ensuite ete ventil6s par centre de coOts (services hospitaliers et services externes), repartis entre les services dispensant des soins directs aux malades, et les coOts unitaires ont finalement ete calcul6s. Les principaux r6sultats concernent les coOts unitaires et la repartition des coOts comme suit: par categorie de ressources utilisees, par centre de coOts et par service fournissant des soins directs. Dans les six districts, les salaires n'absor- bent qu'une part 6tonnamment faible des coOts renouvelables (27-39%) et les medicaments et magasins de fournitures medicales une part 6ton- namment elev6e (24-37%); le principal centre de coOts de l'h6pital, en termes de ressources ger6es, est la pharmacie: de 27 a 39% de tous les coOts renouvelables sont d6penses en dehors de l'h6pital et de 61 a 73% sont consacr6s aux services dispens6s par l'h6pital de district; les ser- vices de soins secondaires de l'h6pital absorbent de 40 a 58% des coOts renouvelables au niveau du district et les coOts unitaires accusent des 6carts sensibles entre les h6pitaux de district, I'un d'eux etant systematiquement le plus cher et un autre systematiquement le moins cher. De 3 a 8 nouveaux malades en consultation externe peu- vent etre soign6s pour le coOt moyen d'une jour- n6e d'hospitalisation tandis que de 34 a 55 nou- WHO Bulletin OMS. Vol 71 1993338 District hospital costs in Malawi veaux malades en consultation externe peuvent etre soign6s pour le coOt moyen d'un sejour a I'h6pital. Les aspects qu'il convient de noter sont les suivants: la faible proportion des coOts absorb6s par les personnels, les 6carts sensibles entre les coOts unitaires et la mesure dans laquelle les per- sonnels hospitaliers peuvent s'acquitter d'activit6s s'6tendant a 1'ensemble du district. Des 6conomies sont apparues possibles concernant les ressources utilis6es, notamment les m6dicaments et les denr6es alimentaires, mais les h6pitaux sont en general bien utilis6s et la dur6e des s6jours relativement breve en regard des normes internationales. Accroitre le r6le de I'h6pital de district dans les activites mises en ceuvre a 1'echelle du district suppose des ressources et une plus grande moti- vation. Des am6liorations pourront etre obtenues si des efforts sont faits tout a la fois pour inciter les personnels a travailler en dehors de l'h6pital, pour mettre a leur disposition des moyens de transport (comme des velomoteurs) Ia ou le deplacement des malades n'est pas indiqu6 et pour les encourager a entretenir les v6hicules. References 1. Mills, A. The economics of hospitals in developing countries. Part I. Health policy and planning, 5: 107-117 (1990). 2. Mills, A. The economics of hospitals in developing countries. Part II. Health policy and planning, 5: 203-218 (1990). 3. Cost-finding and rate-setting for hospitals. Chicago, IL, American Hospital Association, 1968. WHO Bulletin OMS. Vol 71 1993 339