A 24-hour laboratory-based service was established in the City Hospita
l, Birmingham, England, in 1987 for semi-quantitative serum human chor
ionic gonadotrophin (hCG) assay and specific clinical criteria for req
uests were agreed. However, a 309% increase in workload between 1988 a
nd 1992 prompted a detailed clinical audit of this service. During a r
epresentative month only 35% of 142 requests conformed to the agreed c
riteria. Those for routine diagnosis of pregnancy, threatened miscarri
age, and 'abdominal pain screen' comprised 49% of total requests and s
eemed inappropriate. No pregnancy was diagnosed by 'abdominal pain scr
een'. Adhering to agreed clinical criteria, therefore, would not jeopa
rdise patient care and could halve the cost of the service. As a resul
t of this audit the established criteria for serum hCG estimation were
reinforced and re-issued to all relevant clinical teams. Additionally
, to improve delivery of service and reduce costs a semi-quantitative
urine hCG assay was introduced in the Accident and Emergency Departmen
t for use by approved staff. Records in this department showed no incr
ease in hCG requests during the six months after the service change, c
ompared with those in the preceding half year. However, theoretical co
st savings of pound 5173 per annum were eroded by an estimated 10% bec
ause 208 (26%) tests were unaccounted for. The reasons for this appare
nt 'wastage' are unclear, and highlight the difficulty in auditing a b
iochemical test performed outside the laboratory.