The patients and their treatment

We have been performing the Norwood sequence of procedures at Rikshospitalet University Hospital since 1999 (Table 1). For the purposes of this study, we analysed 7 consecutive patients with hypoplasia of the left heart admitted to the Hospital between January 2003 and September 2003. They included four patients undergoing the first stage of the Norwood sequence, two in the second stage, and one patient in the third stage (Table 2). We calculated the costs for each of the three stages of the Norwood sequence.

Table 1. Population of patients at the Rikshospitalet University Hospital.

Table 2. Characteristics of patients.

Operative procedure. Access was via standard median sternotomy. Arterial cannulation via the patent duct, and single atrial cannulation was performed. After establishing satisfactory extracorporeal circulation, the ventilation was discontinued, cooling was started, and the duct ligated centrally and divided. After division of the pulmonary trunk at its bifurcation, the distal part was closed with a piece of pulmonary homograft patch. A 3.5- or 4-millimetre Goretex tube, depending on the size of the patient, was then anastomosed in end-to-side fashion to the right pulmonary artery. We have summarised the precise anatomical diagnosis for each patient in Table 3. When a rectal temperature over 18 degrees centigrade was reached, the arterial pump was stopped, the brachiocephalic arteries were snared, the descending aorta cross-clamped, and cardioplegia was begun, using St. Thomas solution, 10 millilitres per kilogram, given via the arterial cannula. A classical first stage Norwood procedure, augmenting the arch with a pulmonary homograft, was then performed during circulatory arrest, and an atrial septostomy was done. Cardiopulmonary bypass was then reinstituted, rewarming started, and the proximal end of the shunt was joined to the brachiocephalic artery. After reperfusion and rewarming, the rectal temperature being over 34 degrees centigrade, the patient was weaned from bypass, using mild inotropic support with epinephrine at 0.05 micrograms per kilogram per minute, and ventilatory support with a concentration of inspired oxygen of 0.3. The sternum was routinely closed, and the patient transferred to the intensive care unit.

Table 3. Anatomic diagnosis and surgical procedures.

Postoperative management

The children were closely monitored in the cardiac intensive care unit for 5 to 7 days. The essential goal of therapy was to balance the pulmonary and systemic circulations in order to provide a sufficient supply of oxygen to the tissues. Signs of systemic circulatory failure, such as lactic acidosis, low systemic blood pressure, or low urinary output, were treated by circulatory support. In such cases, mild inotropic support was continued, combined with vasodilation using nitroprusside. In most cases, the children were extubated during the first 48 postoperative hours. In more complicated cases, the need for ventilatory support was extended, and this was combined with additional medication to provide inotropic support. Enteral feeding was started as soon as possible, and arterial and venous lines were removed when circulatory stability was achieved.

The financing of health care services in Norway is changing. Instead of global budgeting, more of the hospitals' financing is based on activity in relation to the number of specific treatments and procedures undertaken. The concept of diagnosis-related groups plays an increasingly important role in hospital financing, as the central authorities in Norway have decided that a proportion of the total revenue received by the hospital should be derived from activity according to such diagnosis-related groups.

We raised the question as to whether the present reimbursement based on diagnosis-related groups and regular financial budgeting covers the true costs incurred by the hospital in managing patients undergoing the Norwood sequence of procedures. A prospective methodology was developed to measure the costs related to the care of individual patients. The patients were closely observed and the related data collected during their stay in hospital.

Analysis of costs

The basic principle of analysis of costs is that of collating the resources used by an individual patient, and integrating them into the total in-hospital cost of care.⁶ We collected cost data prospectively, and calculated the total costs for each patient. A prototype database was developed which enabled us to collect data for the analysis.

We used a bottom-up microcosting method, which initially focuses on activity at the level of the patient. Cost per patient is accumulated from prospective records of the utilization of the drivers of cost, such as personnel resources, medication, and blood products. The calculation is based initially on all related or direct activities, and includes the costs of whole time equivalent staff, theatre utilization, and the cost of pharmaceuticals. The subsequent allocation of indirect costs is achieved by predefining key variables, such as the length of stay and the number of discharged patients. Microcosting is the process of determining cost through time and motion studies. The utilization of each resource used in the production of special services is recorded. This method takes account of all resources, such as costs of staff and overheads. A cost is allocated to each service by multiplying the resources used by the unit cost to obtain a measure of the total cost. This approach preserves information about the variability inherent in the medical treatment of individuals, while making it possible to estimate average resources used at the levels of the clinical department and the hospital.⁷

Direct registration of the drivers of cost. Personnel resources: The most important cost in health care resources is that of the time of the professional staff. This cost usually forms the largest segment of the running costs of a hospital, and includes clinicians, nurses, and other health care professionals. We have determined staff costs with reference to salary rates and other labour costs, and have correlated the number of beds with the number of whole time equivalent nurses in post. This information was used in developing four different categories of resource intensivity, and the unit resources were assigned to one of these four categories. On the basis of information gained from daily observation, direct information from the responsible nurse, and registration forms, we determined which levels of nursing care the patient received during each working shift. The number of directly dedicated nurses involved in the care of the patient during each working shift was the major determinant of costs. The four categories of nursing intensity, and of cost, were

• high intensive care, where there was one whole time equivalent staff nurse per patient at each shift, so the nurse-to-patient ratio was one to one;
• light intensive care, where there was a nurse-to-patient ratio of 0.5 to one or more;
• intermediate care, where the ratio was greater than or equal to 0.25 to one; and
• ordinary care, where the ratio was less than 0.25 to one nurse per patient.

The cost method involved two data sets: one which included patient-related or direct costs, and another which included non-patient-related or indirect costs. The basic principle of the study was that the costs should be recorded at the level of the individual patient. A method of prospective registration was applied to collect all data directly related to the patient during the period of hospitalization. These data were divided into the following cost items:

• hospitalization and basic medical service, including medical and nursing staff;
• pharmaceuticals and blood products;
• procedures such as operating theatre.

Indirect costs were estimated from the costs of the clinical service departments, for instance radiology, laboratory services, pharmaceuticals, and the non-clinical service departments such as transport, housekeeping and kitchen services. The assignment of indirect costs was performed through a set of predefined allocation keys, such as the number of discharged patients and length of stay in the unit. Capital costs were not included in the calculations. To obtain the cost per discharged patient, the operating costs of the clinical department were divided by the number of discharged patients corrected for patient mix for that unit in a given year. Costs related to treatment in other hospitals, and at outpatient clinics, were not included in the analysis.