The Pharmacoeconomics of Managing Acute Agitation in the Emergency Department

What Do We Know and How Do We Approach It?

Esther W Chan; David McD Taylor; Jonathan C Knott; Danny Liew; David CM Kong


Expert Rev Pharmacoeconomics Outcomes Res. 2012;12(5):589-595. 

In This Article

Pharmacoeconomic Analyses of Drug Alternatives for the Management of Acute Agitation in the ED: The Considerations

Modes of Pharmacoeconomic Analyses Four approaches to pharmacoeconomic analysis (cost–benefit analysis [CBA], cost–effectiveness analysis [CEA], cost–utility analysis and cost-minimization analysis [CMA]) may be employed to compare alternative drugs for the management of acute agitation in the ED.[24–26] These methods are not limited to the comparison of drugs but are also used for economic analyses of other healthcare strategies.[27] The types of evaluation differ with respect to their measurement of outcomes.

In CBA, the effects of two or more alternative interventions are measured and analyzed only in monetary units, allowing for direct comparison of net costs. The intervention that leads to the least net cost is favored.[28] Relatively few CBAs in the mental health field have been conducted. This is due to the difficulty in valuing outcomes in monetary terms.[28] In the mental healthcare setting, the dollar value of long-term health outcomes measures such as the quality-adjusted life year (QALY) can be estimated using 'ranges of conventional estimates' to estimate the 'monetized value' of the patient's health status.[15]

With respect to CEA, a single effectiveness measure in natural units, such as, mmHg for blood pressure reduction or life years gained by antipsychotic use, is considered.[28] In its strictest application, the outcome measure is common to the alternatives under comparison, but achieved to different degrees (i.e., a difference in effectiveness or 'net effectiveness'[24]). The incremental cost–effectiveness ratio (ICER) is generated by dividing net effectiveness by net costs.[28] As such, the study result is measured in dollars per unit health gain (i.e., cost per unit of effect[16,25]).

Cost–utility analysis is similar to CEA except that the impact of an intervention takes into account health-related QoL. A universal measure of health-related QoL is utility,[24,29] assigning a numerical value (adjustment factor) to any period of time lived by people with a burdensome disease or condition (perfect health assigned a utility of 1.0, death assigned a utility of zero). The number of years lived by a person with a particular disease/condition is multiplied by the relevant utility to derive QALYs.[28] For example, assume that a particular mental illness is associated with a utility of 0.9, a person living for 1 year with mental illness is then assumed to have lived 0.9 QALYs. Numerous methods and scales have been used to determine utility and commonly involve asking patients to prioritize health outcomes according to their preferences.[25] Similar to CEA, the study result is measured in ICERs. As utility is commonly measured in chronic health states, there is limited application of cost–utility analysis in the analysis of acute conditions such as agitation in the ED setting.

CMA is chosen when it has been proven beforehand that alternative therapies are equally effective.[26] Results from clinical studies are used to demonstrate the equivalence of effectiveness. As the clinical outcomes of different options are equivalent, the choice is then made based on the least costly option.[24] CMA is commonly only justifiable in situations where the therapies being compared embody a near-identical technology or are drugs from the same pharmacologic class.[25] In a double-blind randomized clinical trial (RCT), Knott et al. compared midazolam with droperidol for the management of acute agitation in the ED.[30] Time to sedation and the proportions of patients sedated at 5 and 10 min were not significantly different between midazolam and droperidol. A CMA comparing the use of these two agents for the management of acute agitation in the ED was then conducted to determine the less expensive option.[6,31]

Findings from pharmacoeconomic analyses allow informed choices to be made to attain the greatest efficiency.[25] Efficiency is achieved when the greatest clinical outcome (i.e., output) is attained with the expenditure of existing resources (i.e., input) or when the desired clinical outcome is achieved using the least resources.[25,28]

Use of Decision Analysis in Pharmacoeconomic Evaluations Decision-analytic (decision-tree) analysis is a useful decision-making modeling technique that can form the methodical basis for pharmacoeconomic analysis comparing alternative sedating agents used in the management of acute agitation in the ED. Decision analysis depicts a decision-making process, portraying an average scenario in the course of progression of the specific clinical condition.[26] Branches of the decision tree represent the possible courses that management could take, depicted with the probability of the occurrence (Figure 1).[24] In Figure 1, decision nodes (represented by a square) depict the decision of whether to use one alternative drug over another. Chance nodes (represented by circles) depict possible subsequent events. Probabilities are assigned to each branch extending from chance nodes. Triangles describe the point at which a decision analysis evaluation ends.

Figure 1.

Hypothetical decision tree for managing acute agitation.

In the decision-analytic model, health outcomes and/or costs are assigned to each end of the decision tree branches and weighted by the probability of each branch. The cumulative, weighted health outcomes and/or costs of the interventions being assessed can then be compared.[26] For example, using the hypothetical scenario depicted in Figure 1, assume that the cost of Drug A and B (both sedating drugs) are $10 and $20, respectively. If 100 patients are given each drug, 70 of those given Drug A would be successfully sedated at a cost of $1000, while 90 of those given Drug B would be successfully sedated at a cost of $2000. Compared with Drug A, 20 extra patients would be successfully sedated at a net cost of $1000. Therefore, the ICER is $50 per extra sedation.

The decision-analytic approach was used for pharmacoeconomic modeling of lorazepam, midazolam and propofol for sedation in the critical care setting.[17] In that study, decision analysis models were constructed for each sedative, incorporating the duration of sedation, institutional costs with drug administration, personnel time and adverse events.

Integration of Pharmacoeconomic Studies in Randomized Clinical Trials Patient-level data are fundamental to decision-analytic models used in pharmacoeconomic evaluations. Both clinical probability (i.e., outcome) and economic data (i.e., cost) are required to fill the decision-analytic model.[24] The more robust the data, the more confidently it can be relied upon for decision making. With this consideration, pharmacoeconomic analyses are often supplementary to, or integrated within, clinical trials ('piggyback' analysis[25]). Economic evaluation data collected prospectively as part of a clinical study may reduce the overall cost of conducting the study and enhance the internal validity of the pharmacoeconomic study results.[25] However, such evaluations are not always planned at the time that a clinical trial is conducted. In the psychiatry setting, pharmacoeconomic studies have compared alternative drugs in the management of agitation tested in RCTs.[7,12,14–16] However, although no study of this nature has been published in the ED setting, several are useful in providing clinical probability data.[30,32]

The RCT is the preferred design in clinical research and is considered most desirable for pharmacoeconomic evaluations.[28] The RCT serves as a key source of data,[33] but it may be necessary to utilize other sources of information (e.g., observational components and expert opinion) to provide important patient-level data upon which to build decision-analytic models. These alternative sources are of relatively lesser quality.[24]