|Clinical and economic effectiveness of an inpatient anticoagulation service
|Mamdani M M, Racine E, McCreadie S, Zimmerman C, O'Sullivan T L, Jensen G, Ragatzki P, Stevenson J G
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.
Implementing a pharmacist-managed anticoagulation service for inpatient treatment of deep vein thrombosis (DVT) or pulmonary embolism (PE). A 24-hour, 7-day/week pharmacist-managed anticoagulation service was established to provide dosing and monitoring of unfractionated heparin (UFH) and warfarin therapies at the request of a physician. The Raschke weight-based UFH dosing nomogram and a modified Fennerty flexible induction warfarin nomogram were approved by the institutions' Pharmacy and Therapeutics and Medical Executive Committees. Pharmacists could deviate from the nomograms based on professional judgement. Once consulted, they could adjust UFH and warfarin dosages and order appropriate monitoring tests. Before implementation, all clinical pharmacists were trained through a certification process in UFH and warfarin therapeutics and dosing strategies.
Type of intervention
Treatment and secondary prevention.
Economic study type
Patients admitted for treatment of DVT or PE and who received intravenous UFH. Exclusion criteria were as follows: thrombolytic therapy, LMWH, subcutaneous adjusted-dose UFH, thrombectomy or embolectomy, less than 24 hours of intravenous heparin administration, delay of greater than 48 hours from hospital admission to start of heparin therapy, discharge against medical advice, and events unrelated to anticoagulation therapy that would unnecessarily prolong hospital stay by more than 48 hours.
Hospital. The economic analysis was carried out in the USA.
Dates to which data relate
Effectiveness and resource use data corresponded to patients admitted to the study institutions between June 1996 and April 1997. The price year was 1997.
Source of effectiveness data
The evidence for the final outcomes was based on a single study.
Link between effectiveness and cost data
Costing was conducted prospectively on the same patient sample as that used in the effectiveness analysis.
Power calculations were used to determine the sample size: with an alpha of 0.05 and power of at least 90%, 50 patients per study group were required to detect expected differences in primary clinical and economic end points. A total of 50 patients, chosen as a result of reviewing a total of 92 charts, and with a mean (SD) age of 60.4 (19.1) were in the usual care group and 50 (chosen as a result of reviewing a total of 83 charts) and with a mean (SD) age of 56.7 (18.3) were in the pharmacist-managed care group.
This was a prospective cohort study, carried out in two centres. The duration of the follow-up appears to have been until discharge. Loss to follow-up was 2 patients in the usual care group and 1 patient in the pharmacist-managed group. Patient data were collected sequentially from medical charts as a convenience sample until 50 evaluable patients in each group were identified. All clinical data were collected with a standard form to allow for consistent, objective information.
Analysis of effectiveness
The principle used in the analysis of effectiveness appears to have been treatment completers only. The primary therapeutic end point was the time elapsed between starting UFH therapy and surpassing the therapeutic threshold, which was defined as an activated partial thromboplastin time (aPTT) of 48 seconds. Other clinical outcomes were number of aPTT values examined, percentage of aPTT values which were subtherapeutic/therapeutic/supratherapeutic, percentage of patients surpassing therapeutic threshold within 48 hours of starting heparin, proportion of patients in whom warfarin was begun within 2 days of heparin, and adverse events (major and minor bleeding). Kaplan-Meier time-to-event analysis was conducted to assess differences between groups with respect to the time required to surpass the therapeutic threshold and to obtain an aPTT within the therapeutic range. The study groups were comparable in terms of baseline demographic data. Multiple regression analysis was used to adjust for potential confounders (determinants of the clinical outcomes).
Median (interquartile range) time to first therapeutic aPTT was 25.3 (29.3) hours in the usual care group versus 23.6 (24.4) hours in the pharmacist-managed group, (p=0.14). 489 aPTT values were examined in the usual care group compared to 476 in the pharmacist-managed group, (p<0.001). 21.3% of aPTT values were subtherapeutic compared to 15.8% in the pharmacist-managed group, (p=0.03). 41.5% were therapeutic 41.5% versus 47.7% (pharmacist-managed), (p=0.05), and 37.2% were supratherapeutic 37.2% versus 36.6% (pharmacist-managed), (p=0.83). 92% of patients surpassed the therapeutic threshold within 48 hours of starting heparin versus 98% (pharmacist-managed), (p=0.36). 63% of the usual care patients had warfarin within 2 days of heparin versus 82% of pharmacist-managed patients, (p=0.05). Major bleeding did not occur in any patient receiving pharmacist-managed care but did occur in two (4%) receiving usual care. No statistically significant differences were observed between groups with respect to minor bleeding complications (8% each group).
No significant differences between groups were noted for the primary therapeutic end point. A greater proportion of therapeutic aPTT values, earlier start of warfarin therapy relative to UFH therapy, faster response to non-therapeutic aPTT values, and shorter hospitalisation were significantly associated with the pharmacist-managed group.
Measure of benefits used in the economic analysis
No summary benefit measure was identified in the economic analysis, and only separate clinical outcomes were reported.
Costs were not discounted due to the short time frame of the cost analysis. Quantities were reported separately from the costs. Some cost items were reported separately. Cost analysis covered the costs of hospitalization and pharmacist intervention time. The perspective adopted in the cost analysis was not explicitly reported. The source of cost data was the study institution's central finance department using a cost accounting system. The pharmacist's salary was used to assess the economic impact of the service in the sensitivity analysis. Furthermore, the opportunity costs of pharmacist intervention time was estimated. The estimation of pharmacist time was based on a questionnaire. Since clinical pharmacists were routinely involved in cost-saving interventions and were paid an annual salary irrespective of these interventions, an opportunity cost was incurred. The price year was 1997.
Statistical analysis of costs
It appears that the Mann-Whitney test was used to detect differences between the groups in terms of costs. Bivariate and multivariate regression analyses were conducted to examine potential determinants of hospitalisation cost.
One-way and two-way sensitivity analyses were performed on the cost and effect parameters. A 10,000-iteration Monte Carlo stimulation using Latin hypercube sampling was conducted to assess total variability in economic data.
Estimated benefits used in the economic analysis
Total hospital costs were significantly lower for the intervention group than for the usual care group ($1,594 and $2,014, respectively, p=0.04). The opportunity cost of pharmacist time was $82. The overall net savings per patient was $338. Results of the Monte Carlo simulation were consistent with the original analyses and indicated a mean net savings/patient of $340 (95% CI for the mean: $313 - 368).
Synthesis of costs and benefits
Costs and benefits were not combined.
Despite the limitations, the study findings suggest potential clinical and economic benefits associated with a pharmacist-managed anticoagulation service relative to usual care.
CRD COMMENTARY - Selection of comparators
The strategy of using usual care was regarded as the comparator. You, as a database user, should consider whether this is a widely used health technology in your own setting.
Validity of estimate of measure of effectiveness
The internal validity of the effectiveness results can not be reasonably assured due to the issues of bias, confounding, and chance associated with not employing a blinded randomised design, as acknowledged by the authors. Furthermore, it was noted that the study results might have been significantly confounded by a cross-contamination effect. However, the study groups were comparable in terms of baseline demographic data and all efforts were made to avoid selection bias by following a priori exclusion criteria. Furthermore, bivariate and multivariate analyses were performed to identify the potential determinants of the clinical outcomes. It was acknowledged that it was difficult to assess the effects of physicians' perception of the difficulty of consulted cases (cases referred to the pharmacists) on the study results. The study sample appears to have been representative of the study population. However, it was acknowledged that differences in less common outcomes, such as major bleeding, may not have been apparent due to the small sample.
Validity of estimate of measure of benefit
The authors did not derive a summary measure of health benefit. The study was therefore a cost-consequences analysis.
Validity of estimate of costs
The validity of the cost analysis is likely to have been enhanced by the following positive features of the cost calculations. Most quantities were reported separately from the costs, adequate details of the methods of cost estimation were given, the price year was specified, the opportunity cost of the pharmacist's time was included and statistical analyses were performed on resource consumption and cost data. However, the cost analysis had the following limitations. The estimation of pharmacist's time was based on a questionnaire rather than on time-motion studies, the perspective adopted in the cost analysis was not explicitly reported, and the effects of alternative procedures on a set of indirect costs were not addressed.
The authors' conclusion appears to be justified given a set of simple and probabilistic sensitivity analyses performed to address uncertainties and variability associated with the data. The issue of generalisability to other settings was addressed by acknowledging that the generalisability of the results was dependent on the institution, disease states treated, and nomograms used. Some comparisons were made with other studies. Regarding the issue of the representativeness of the study sample of the study population, it was acknowledged that the study results pertained to patients diagnosed with DVT or PE and may not be extrapolated to those with other disorder
Implications of the study
Larger samples are necessary to detect important differences in less common outcomes, such as major bleeding. Efforts must be made to improve nomogram compliance and enhance earlier nomogram implementation through the anticoagulation services.
Mamdani M M, Racine E, McCreadie S, Zimmerman C, O'Sullivan T L, Jensen G, Ragatzki P, Stevenson J G. Clinical and economic effectiveness of an inpatient anticoagulation service. Pharmacotherapy 1999; 19(9): 1064-1074
Subject indexing assigned by NLM
Anticoagulants /economics /therapeutic use; Cohort Studies; Female; Heparin /therapeutic use; Hospital Costs; Humans; Length of Stay /economics; Male; Middle Aged; Partial Thromboplastin Time; Pharmacy Service, Hospital; Pulmonary Embolism /drug therapy; Time Factors; Venous Thrombosis /drug therapy; Warfarin /therapeutic use
Date bibliographic record published
Date abstract record published