This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.

The provision of a clinical pharmacy service in a geriatric primary care clinic was studied.

Primary prevention.

Cost-effectiveness analysis.

The patient population comprised men attending a Veterans Affairs Medical Centre in the USA. No detailed characteristics of the population were provided.

The setting was primary care. The economic study was carried out in the USA.

The study took place over 8 months from early 1998. This was the time period during which the effectiveness and resource evidence were collected. The dates for the prices used were not specified, but could be assumed to relate to the same period.

The effectiveness data came from a single study.

The costing data came from the same patient sample as that used in the effectiveness analysis.

A power calculation was conducted to determine the sample size. The sample population comprised patients attending the clinic over a 2-month period. Patients for the intervention group were selected if they were taking 5 or more agents per day, or were receiving drugs for chronic diseases. The control group comprised other patients seen by the team over the period. The intervention group made 106 visits to the clinic while the control group made 144. No information was provided on the number of visits excluded.

The study was a prospective, non-randomised controlled trial, which was conducted at one Veterans Affairs Medical Centre. The follow-up was for 6 months. No information on loss to follow-up was provided. There was no blinding for the assessment of the outcomes.

It was not stated whether the clinical analysis was based on intention to treat or on treatment completers only. The clinical outcomes reported included:

the types of drug-related problems identified;

the categories of drugs discontinued;

the categories of drugs begun;

the percentage of the pharmacist-recommended changes in drug therapy that were accepted;

the mean number of drugs per patient at the start and finish of the study; and

positive, neutral or negative effects of changes in therapy.

A positive outcome was the achievement of a therapeutic goal, for example, a reduction in blood pressure. A negative outcome was a worsening of the patient's disease state or an adverse drug reaction after change in drug therapy. No information on the comparability of the treatment and control groups was provided. No account was taken of potential confounders.

The types of drug-related problems identified (out of 220) were:

drug use without indication, 50.9%;

untreated indication, 11.4%;

adverse drug reaction, 9.5%;

overdose, 9.5%;

improper drug selection, 7.7%;

duplication, 3.6%;

laboratory monitoring necessary, 3.6%;

subtherapeutic dosage, 3.2%; and

drug interaction, 0.5%.

The categories of discontinued drugs (out of 112) were:

psychotropic, 28%;

cardiovascular, 20%;

topical, 18%;

gastrointestinal, 14%; and

other, 20%.

The categories of drugs begun (out of 25) were:

cardiovascular, 56%;

central nervous system, 16%;

musculoskeletal, 8%;

respiratory, 8%;

genitourinary, 8%; and

gastrointestinal, 4%.

The percentage of pharmacist-recommended changes in drug therapy that were accepted was 98.6%.

The mean number of drugs per patient at the end of the study was 7.2 for the treatment group and 7.0 for the control group, (p=0.0007).

The mean change in the number of drugs per patient was -3.4 for the treatment group and -0.4 for the control group, (p=0.0001).

The effects of changes in drug therapy (out of 220) were positive for 47.5%, neutral for 52%, and negative for 0.5%.

Pharmacy services in a geriatric primary care clinic led to a statistically significant decrease in the mean number of drugs per patient. These services also achieved positive or neutral clinical outcomes in almost all cases.

There was no summary measure of benefit, although it might have been assumed that the treatment was unequivocally more effective than the control. This study therefore constitutes a cost-consequences analysis.

No discounting was carried out as the costs were only incurred over a 1-year period. The resource quantities and the unit costs were not reported separately for the drugs. The costs were reported on a per annum basis for the pharmacist's salary at 5 hours per week (12.5% of $58,000), drugs (costs saved, $18,260; cost incurred, $3,080) and additional laboratory monitoring ($142). It was assumed that the costs related to the date of the study, i.e. 1998. The costs were not reflated as it was inappropriate. No distinction was made between the average and marginal costs. All the costs common to both alternatives were excluded.

No statistical analysis of the costs was performed.

The indirect costs were not estimated.

US dollars ($).

No sensitivity analysis was conducted.

See the 'Effectiveness Results' section.

The costs were reported in terms of cost changes between the intervention and control.

The cost-savings from drug discontinuation were $18,260.

Additional costs due to laboratory monitoring were $142.

Additional costs due to new therapy were $3,080.

Additional costs due to professional services were $7,250.

The total additional costs were $10,472.

The net annual cost-savings ($18,260 minus $10,472) were $7,788.

Not applicable due to the cost-consequences approach adopted.

Pharmacy services in a geriatric ambulatory clinic decreased the mean number of drugs per patient, and achieved positive or neutral clinical outcomes. They also resulted in net cost-savings to the clinic.

No explicit justification was given for the comparator. However, as a new service was being assessed, it is assumed that the normal situation was the primary care setting studied without the provision of pharmacy services.

The authors acknowledged that the estimate of effectiveness (mainly the drug reduction per patient) of the intervention is likely to have been biased: more patients on larger numbers of drugs were selected for the intervention group than the control group. The authors suggested that one can assume that no reduction in the treatment group would have occurred without the intervention of a pharmacist. However, this is a strong assumption, which also implies the negation of the need for the supposed control group. Finally, the measure of effectiveness in terms of the change in clinical status was poorly defined and not open to scrutiny.

There was no summary measure of benefit.

The costing was very limited, in that only changes in the costs due to the introduction of the pharmacy service were included in the analysis. The costs and the quantities were not reported separately for the drugs, although the source of unit costs was given.

The article was poorly edited. The results quoted in the text referred to the wrong figures. The figures referred to in the text did not exist. The authors pointed out a number of reasons why the extent of the drug reductions achieved was likely to be inflated. First, the intervention groups were those with larger numbers of drugs per patient. Second, for these, some reduction may have occurred without pharmacy service provision. Third, the number of drug-related problems are likely to taper over time as visits from patients recur.

Consequently, one can only assume that the extent of the annual cost-savings estimated are only obtained where pharmacy services are used selectively, and for the first year of introduction of the service. Clearly, they also only relate to the particular setting of the study, which included only male geriatric patients in the USA.

The authors claim that the results demonstrated the value of pharmacists in multidisciplinary patient care, and consequently they decided to introduce a full time pharmacist to their team. It should be clear from the criticisms made, particularly in terms of the likelihood of bias in effectiveness and the lack of costing, that there is insufficient evidence to support this.

None stated.