Multicentre aneurysm screening study (MASS): cost effectiveness analysis of screening for abdominal aortic aneurysms based on four year results from randomised controlled trial

Table A1

Table A2
 

Appendix

Additional detail of the costing methods

Measurement of costs

Adopting a health service perspective we calculated patient specific costs related to abdominal aortic aneurysms up to four years by applying specifically calculated unit costs to the patient based recording of screening events (invitation for screening, re-invitation to non-attenders, attendance at screening clinic, and attendance at follow up/recall clinic) and surgery (assessment of suitability, elective surgery for aneurysm repair, and emergency surgery for aneurysm rupture). We did not included costs for aneurysm ruptures that led to death without the patient being admitted to hospital for attempted emergency surgery. All the unit costs are presented on the price base of financial year 2000-1, adjusted where necessary to that date with the hospital and community health service inflation index.^w1 Actual staff salaries were used with standard estimates of average staff hours worked.^w1 Unless otherwise indicated, unit cost information was provided by the finance departments of each of the four centres.

Costs associated with screening

Screening clinic costs were based on the standard resourcing of screening for this trial and the trial experience of patient throughput. Each clinic team undertook 25 three hour clinics per month, with 20 patients invited to each clinic. The costs for initial invitation include clerical staff time; consumables such as postage and stationary; the cost of obtaining patient details from the Family Health Services Authority (FHSA); and office space and equipment. The re-invitation cost simply excludes the payment to the FHSA.

The cost per clinic attendance allowed for the time of each member of clinic staff (and their travel costs, as staff were hospital based but provided screening sessions in primary care locations), disposables, maintenance costs for screening equipment, charge for clinic rooms, and an annual equivalent charge for the capital cost of the equipment (assuming a useful life of five years and a discount rate of 6%).

All recall scans for monitoring of the aorta involved the costs of ultrasonography, with only a proportion seeing a consultant (7.6%, specifically those that were attending for their third three monthly scan or for whom the frequency of their recall scanning was altered from yearly to quarterly). The recall scan cost was therefore calculated by taking the average of the cost of an ultrasound scan at each of the four trial centres and allowing for a proportion of the cost of a follow up surgical outpatient attendance (again averaged across the study’s four centres). We ignored the costs of any opportunistic screening that may have occurred in the control group.

Costs associated with surgery

For each of the four MASS trial surgical centres we calculated a standard assessment for surgery cost based on the standard local procedure and investigations included for such a consultation. A patient specific record of assessments was available for screened patients: the same average number was assumed for each elective patient undergoing surgery in the control arm.

To obtain reliable estimates of the costs of elective and emergency procedures, we collected patient specific resource use data for cohorts of consecutive patients whose primary cause of admission was for aortic aneurysm repair and who were men aged ³ 65 years at each of the four MASS trial surgical centres. To ensure sufficient cases to provide adequately precise cost estimates, these consecutive cohorts included patients from within the trial and others receiving aortic aneurysm surgery who met the inclusion criteria for the trial but who were not involved in the screening study itself. Detailed information on resources used by 360 elective repairs and 217 emergency ruptures was collected and is presented by centre in table A1. Costs were calculated to include any related hospital readmissions during the 12 months after surgery to allow for the costs of complications of surgery. We used a "bottom up" approach to the costing of surgery. Unless otherwise stated we used centre specific unit costs to cost patient specific resource use. These unit costs are presented in table A2.

All patients undergoing an emergency aneurysm repair would have required an emergency ambulance to travel to hospital and would have entered the hospital through the accident and emergency department. We used a fixed unit cost for emergency transport^w1 and the cost of an attendance at accident and emergency from each of the four hospitals.

We costed time spent in intensive care, high dependency units, and general surgical wards before and after surgery using the appropriate hospital bed day costs inclusive of hospital overheads but exclusive of drugs, bloods, and non-pathological investigations (which were costed separately).

Theatre duration in minutes was collected from each patient’s anaesthetic sheets. To cost the use of the theatre per se, finance departments provided a fixed cost per hour for an empty theatre (inclusive of hospital overheads, building charges, and routine theatre capital equipment). This cost was applied to the duration in theatre of each of the patients in the surgery sample. Three of the four consultant surgeons provided information on medical staff routinely present during a typical elective and an emergency repair. Staffing levels were assumed not to vary between patients. For each member of staff, the mid-point from their annual salary scale plus "on costs" (employers National Insurance and pension contributions) was used to calculate a fixed staff cost per minute, which was applied to each patient’s observed theatre time. The average staff cost per minute (for an elective and an emergency repair) across the three hospitals that provided the information was used to calculate theatre staff costs for patients operated on in the fourth hospital.

Three of the four consultant surgeons identified the quantities of specific consumables used during elective and emergency aneurysm repairs in their centre. Again these quantities were assumed not to vary between patients. Two of the centres provided unit costs for each of the items on the list. For the remaining two centres we used the mean of these two sets of costs. The cost of the specific type of graft (straight or bifurcated) inserted during surgery was included. Detailed data on patient readmissionsto theatre related to aortic aneurysms during the initial inpatient stay were also available and were costed in the same way as the initial theatre stay.

Detailed data on drug use covering the time in theatre and the inpatient hospital stay were collected from the notes of a subsample of 60 patients (emergency and elective) at one centre. Drugs were then costed, allowing for normal wastage (for example, of part vials) and for tax, using the British National Formulary.^w2 A linear regression model was used to predict hospital drug costs for all patients in the surgery sample, regressing drug cost data against hours spent in intensive care and time spent in theatre.

Regression model

Log of drug costs=3.816+0.089 SQRT(ITU hours)+0.004735 theatre duration

Adjusted R²=0.607 (t=7.669) (t=4.888)

To allow for variable discounts received by hospitals when purchasing pharmaceuticals we complied a list of drugs identified as high volume (from the notes of the sub-sample of 60 patients) and sent it to the pharmacy managers of each hospital. Hospital prices were then compared to the British National Formulary and a mean centre specific discount calculated. Patient specific drug costs (estimated using the regression) were then adjusted to take account of this discount. Pharmacy overhead costs were incorporated into the analysis by increasing each patient’s drug cost by an amount equivalent to what their hospital’s 2000-1 pharmacy overheads were as a percentage of 2000-1 total drug expenditure.

We obtained patient specific data on blood products issued to patients during their inpatient hospital stay from each hospital’s blood bank. Costs and handling charges were used to cost products according to whether they were used, wasted, or returned unused to the blood bank. The number and type of non-pathology investigations each patient underwent was obtained from patient notes. Unit costs for each type of investigation were obtained from hospital finance departments and applied to each patient’s investigation resource use data.

For patients discharged to other hospitals, speciality specific bed day costs for the relevant NHS trusts, were obtained from the trust financial returns^w3 and were multiplied by each patient’s length of stay data. Details of any admissions or readmissions related to aortic aneurysms in the 12 months after surgery were obtained from patient’s notes and were costed using the relevant bed day costs plus the cost of any further surgery. Expert advice was sought about the number of postdischarge follow up appointments patients would receive at a general surgical outpatient clinic. The costs of routine postdischarge follow up appointments were obtained from hospital finance departments and applied to each patient (depending on whether they were still alive at each clinic time point).

We calculated mean costs for each type of surgery by centre and for the whole sample. Because of the non-normal distribution of costs we used non-parametric bias-adjusted bootstrap methods to estimate confidence intervals around the mean resource use and unit costs.^w4

w1. Netten A, Rees T, Harrison G. Units costs of health and social care. Canterbury: Personal Social Services Research Unit, University of Kent at Canterbury, 2001.

w2. British Medical Association, Royal Pharmaceutical Society of Great Britain. British national formulary. BMA, RPS, 1998 (No 35).

w3. Chartered Institute of Public Finance and Accountancy. The health service financial database and comparative tool. Croydon: Institute of Public Finance, 2000.

w4. Barber JA, Thompson SG. Analysis of cost data in randomised trials: an application of the non-parametric bootstrap. Stat Med 2000;19:3219-36.
 
 

Table A1 (Posted as supplied by author)
 

(n) denotes number

NA Not applicable* Assumed each elective patient had one pre-surgery assessment

† Assumes each emergency patient required emergency ambulance transport to hospital

‡ Assume all emergency patients entered via A&E. Three elective patients arrived at A&E with back pain but were awaiting an elective AAA repair and remained in hospital

§ All but two elective patients (who went into ITU at Centre A) were admitted to general surgical wards. A small number of emergency patients were admitted to ITU or general surgical wards

¶ Three patients were admitted for AAA but underwent other procedures prior to AAA surgery hence increasing their pre-surgery LOS (1 x amputation of left big toe,

1 x left inguinal hernia repair, 1 x sigmoid colectomy and splenectomy)

Theatre staff -- 1=Consultant Surgeon, 2=SpR, 3=Sister, 4= Staff Nurse, 5=Other Nurse (middle grade), 6=Grade A Nurse, 7=Consultant Anaesthetist, 8=SpR (anaesthetics), 9=ODA

** These patients appears to have had two grafts during the initial AAA operation

†† This patient was readmitted to surgery 5 times in total

‡‡ Applies to inpatient stay at operating centre only
 

Table A2 (Posted as supplied by author)

Bold type indicates where a cost is an average from other trial hospitals

NA Not applicable