Cost-effectiveness and cost-benefit analyses are analytical techniques that can help physicians and policy makers by comparing health care practices and technologies in terms of their relative economic efficiencies in providing health benefits (1). Cost-effectiveness analysis compares the net monetary costs of a health care intervention with some measures of clinical outcome or effectiveness, such as increased longevity, expressed most often as life-years saved. Whereas other health care outcomes, e.g., dollar per episode of postoperative nausea and vomiting (PONV) avoided, can be calculated with a cost-effectiveness analysis, there is no standard against which to measure the value obtained. Cost-effectiveness analysis is also paternalistic, i.e., someone who manages the delivery of health care (neither patients nor their doctors) will decide what is worthwhile. In a cost-benefit analysis, the costs of a health care intervention are assessed the same way as in cost-effectiveness analysis, but measures of clinical outcomes or effectiveness are typically converted into monetary units (2). The patients will determine what monetary value they place on their therapy and what they want to fund from their own resources, their insurance, or tax dollars.
Many new drugs have been introduced in anesthesia, often replacing older and less expensive drugs. Much has been written on the benefits of these newer drugs in terms of physiological and pharmacological responses of patients. Little is known of the value patients put on the benefits of these new drugs. PONV is an unpleasant experience that occurs often after surgery. The results of several studies suggest that patients not only rank the absence of PONV as being important (3) but also rank it as more important than an earlier discharge from an ambulatory surgical unit (4).
We hypothesized that patients were willing to pay (WTP) for an antiemetic that would eliminate PONV and that the WTP amount was related to the patient’s demographic characteristics.
After obtaining IRB approval and signed informed patient consent before anesthesia, we recruited 80 consecutive patients presenting for elective day surgery using general anesthesia. Patients were asked to complete an interactive computer questionnaire on WTP (see Appendix 1) in the postanesthetic care unit (PACU) after they met institutional PACU discharge criteria. Potential participants were not told in advance that a computer would be used to administer the questionnaire to avoid bias against those who might have an aversion to the use of computers. Those who did not understand the English language were excluded from participation. A research nurse familiarized the patients with the interactive questionnaire in the preoperative holding area before surgery.
A working group of a statistician, a computer programmer, and a staff member from the Departments of Anesthesia and Economics and Health Policy formulated the questionnaire, which was tested for ease of use on departmental volunteers before deployment. The interviews consisted of three parts. The first set of questions introduced the participants to the format of the questionnaire and ensured that the participants understood the concept of WTP. A hypothetical scenario was presented to assess whether each patient understood the WTP concept. The “chance of nausea and vomiting” numbers were randomly generated by the computer. If their answers were appropriate, (i.e., they rejected the new drug because it had both an increased incidence of PONV and a larger cost compared with the routine drug), then they were deemed to have understood the concept and were allowed to proceed to the next part of the questionnaire. If they did not get this right, the concept was explained to them, and they had to correctly answer that question before continuing.
The second set of questions established the value patients placed on the avoidance of PONV and the amount they were WTP for a hypothetical antiemetic that would completely prevent it. The third set of questions established demographic characteristics, health functional status, income, education levels, and other factors that might affect the response.
Patients were not compensated financially for participation to avoid a disproportionate share of patients with smaller income who might be more willing to participate on that basis. A “play money” bias was eliminated by noting that the drug would be an out-of-pocket expense. This intentional deception was IRB approved and noted in the informed consent (without noting which part of the interview might include some “minor false statements”). We evaluated the play money effect by also evaluating how much this therapy would be valued if someone else (insurance companies) were paying. Starting point bias was eliminated by randomly varying the cost of a new antiemetic between US$5 and US$50 (an a priori reasonable range and consistent with IV drug acquisition costs for antiemetics and other anesthetics).
Patients were first acquainted with the type of questions asked. The computer then asked the patients whether they would choose a new antiemetic that would completely abolish the incidence of PONV or the routine antiemetic, which would still allow a chance of PONV. If the respondent chose the routine antiemetic, the random price x of the new drug was decreased by x/8 (minimal cost was set at US$2) and the question reiterated up to eight times, with the cost reduced proportionally after each question. If the respondent chose the new antiemetic, the price of the new antiemetic was increased by x/8 and the question reiterated, again up to eight times. The process continued until the price of the new antiemetic was enough to make the respondent reverse the preference. After eight iterations, patients were asked the amount they would be WTP for the new antiemetic. A maximum ceiling on the value of the antiemetic was set at US$500, with no minimum.
Power calculation revealed that 80 patients would be adequate to detect a significant (P < 0.05) relationship between WTP value and the presence or absence of nausea with α = 0.9 and β = 0.2. Descriptive statistics were calculated and reported. Univariate and multivariate linear regressions were fitted to relate WTP to possible predictors. The confidence intervals and P values of the various covariates were presented. The distribution of WTP was nonnormal (P value for the Shapiro-Wilk test of normality was 0.0001). Hence, the rank analysis of variance was adopted. Kendal’s τ b correlation was used to assess redundancy in the explanatory variables. A P value <0.05 was considered statistically significant.
Eighty patients (32 men, 48 women) completed the questionnaire. The median age was 46 years, with a median (25%–75%) household income of US$52,500 (US$24,725–US$96,750). Patients were WTP US$56 (US$26–US$97; median, 25%–75%) for an antiemetic that would reduce the incidence of PONV to zero. Patients who did not develop PONV were WTP US$40 (US$20–$US95; median, 25%–75%;P < 0.05). Patients who developed nausea (n = 21, 26%) were WTP US$73 (US$44–US$110), and those who actually experienced emesis (n = 9, 11%) were WTP the larger sum of US$100 (US$61–US$200; median, 25%–075%). P < 0.05 among the groups. When the cost of the antiemetic was to be borne by an insurance payment, the WTP amount allotted by the patients increased by an average of 68%. One patient, who did not develop PONV, had a WTP borne by insurance value above the $500 cutoff and was hence capped at $500.
Seventy-six percent of patients considered avoiding postoperative nausea and 78% considered avoiding postoperative vomiting important (patient response ≥50 mm on a 0–100-mm visual analog scale). A number of covariates significantly increased the patients’ WTP amount and the patients’ WTP amount to be borne by insurance. These included nausea or vomiting in the PACU, larger patient’s income, previous history of PONV, the importance the patient placed on avoiding PONV, and increasing age (Table 1).
This study elucidated the value that patients place on eliminating PONV. Patients having surgery under general anesthesia value avoidance of PONV and are WTP US$56–US$100 to avoid experiencing these unpleasant side effects.
Health care costs consume more than 13% of the American gross domestic product. Although direct anesthesia costs are a very small component of this US$1 trillion annual cost in the United States, pressure by hospital administrators and payers to cut anesthesia-related costs has intensified (5). However, there is a paucity of data establishing the monetary value associated with the avoidance of common undesirable postoperative side effects from the patient’s perspective. Because traditional cost-effectiveness is defined as dollars per year of life saved, therapies (including most of what an anesthesiologist does) that do not prolong life but ameliorate pain, suffering, anxiety, and other intangibles have not been considered by traditional cost-effectiveness analyses. Short-term amelioration of pain and suffering does not save years of life. Although quality of life adjustments are possible via cost-utility analysis (by using some sort of survey methodology), these are appropriate for long-term disease effects, not the short-term quality of life issues, such as acute pain or recovery from anesthetic effects, that anesthesiologists routinely address. There are no benchmarks to judge whether the cost of a therapy that affects this short-term quality of life is worthwhile. We sought to define a benchmark for the most common anesthetic side effect—PONV.
PONV is one of the most common complications after anesthesia, with an incidence of between 20% and 80%(6–8). Patients consider PONV one of the most undesirable postoperative symptoms, and it is one of the most common reasons for poor patient satisfaction rating in the postoperative period (9). Macario et al. (3) quantified patients’ preferences for postoperative outcomes before surgery. PONV was among the 10 most undesirable outcomes after surgery. Indeed, patients allocated the largest amount (US$30) to avoid PONV out of a total of US$100 they were allowed to spend to avoid all complications.
WTP is based on simple supply-demand theory in economics. A consumer will pay up to a certain amount for the benefits of consuming a particular good or service, but above maximum WTP no longer feels that the benefits obtained outweigh the opportunity cost of purchasing the item (10). WTP has been extensively used as a tool in economic evaluation of alternative land use, but it has only recently been applied to the economic evaluation of health care (11). The WTP technique has been used to value interventions in screening, prevention, treatment, and service provision. This method gives valid and consistent results when valuing health care interventions (12).
Tang et al. (13), in an efficacy and cost-benefit analysis of administering prophylactic antiemetic to women having gynecological laparoscopy, asked patients how much they would be WTP out of their own pocket to prevent PONV if they were to undergo the same laparoscopic operation in the future. They provided a choice of US$0, US$50, US$100, US$250, or another dollar amount. They found that patients were WTP an average of US$117 ± US$82 (mean ± sd). The authors, however, did not examine the factors that could affect the WTP value, e.g., income, education, insurance payer, etc. It is conceivable that the larger WTP value compared with our results may reflect the higher socioeconomic status of that population. Also, their results may have been affected by providing ceiling values, anchor ranges, and starting point biases that were all addressed in this study through initial value randomization and value iteration. Our initial values were limited to US$5–US$50, so the fact that patients were never presented with a value as high as $250 because of the iterative value generation (x +x/8, maximum iterative value of US$129) may have predisposed this patient group to a lower value compared with the results of Tang et al. In addition, there was no cap on their maximum values. Because the data in our study were right skewed, we placed a cap on a maximum amount of $500 and reported the median and interquartile ranges. Diez (14) used a structured questionnaire to evaluate parents’ WTP for a reduction in postoperative emesis in their children. The median value was found to be £50 (US$75; 95% CI, £20–£80), which was very similar to this study.
Previous studies have established the validity and reliability of WTP estimates in other health care applications (15). However, cost-benefit analysis is not often performed in anesthesiology. When severe morbidity and mortality outcomes are extremely rare, it is important to assess the economics of treating less tangible but more common outcomes, such as pain, sore throat, PONV, and anxiety. Who in this world can be a better judge of the value of those interventions than those who have experienced the problem? We have demonstrated that a majority of patients associated the absence of PONV as desirable and equated that with a monetary value, as we might predict. Those who developed nausea, vomiting, or both on the day of surgery had a more vivid realization of how unpleasant these symptoms were and hence, as would be expected, attached a larger sum to avoiding PONV in the future. In addition, we have shown that the WTP amount correlated significantly with a number of covariates, such as income levels (a surrogate for marginal utility of income, which is expected to affect WTP), the importance of avoiding PONV, and a history of PONV. In this study, we addressed only the issue of WTP for an effective antiemetic. It would be interesting to ask patients their WTP to avoid other unpleasant symptoms related to anesthesia and surgery.
WTP is difficult to ascertain by simply asking a single question (16). Asking closed-end questions (suggesting one value to respondents, who can either accept or decline it) involves larger costs and resources, takes more time, and may limit responses around the initial starting value (starting point bias). Our questionnaire addressed these issues by providing a varied starting point, in the reasonable range of US$5–US$50. This was designed to reduce starting point bias (i.e., if you suggest an initial value of $5, it is less likely that $50 will be seen as reasonable, and vice versa.). However, after patients in our survey went through the eight iterations, they were asked to state the amount they were WTP, allowing an open-ended question to supplement the closed-ended questions previously iterated. We also found no correlation between the final WTP value and the initial suggested dollar amount. An interactive computer program has several advantages over a paper questionnaire. The problem of interviewer bias is avoided, and it may promote more honest revelation of preferences. Subjects are better able to recall information provided in a multimedia format compared with receiving the same information in written format alone (17).
We used a hypothetical antiemetic that would totally prevent PONV in the questionnaire. Such a drug admittedly does not exist. However, we wanted to present a relatively straightforward scenario that patients could easily comprehend. Patients have difficulty in understanding various hypothetical efficacy rates (10). It is probable that the WTP amounts we have demonstrated may be more than might occur in real life, when the efficacy of the drug would not be 100% (certainty equivalent factor). However, the results from this study can still be used in future cost-benefit analysis of any particular antiemetic therapy.
Determination of WTP may be sought from different samples of populations. The issue of whether WTP estimates differ among those with and without the symptoms is critically important. It is known that particular patients’ preferences tend to overvalue a particular health state or health benefit if it is a personal feature of their lives. However, a previous study of preferences for health states (cancer) found a high degree of correspondence between ratings provided by patients and by the general public (18). By using the “greatest good” argument to justify a society perspective for an economic analysis, it is suggested that a representative community sample be used to value a disease state or value of a drug. However, in this scenario, patients who are undergoing surgery have the risk of developing PONV and hence have a stake in avoiding these unpleasant symptoms, so societal value may not be relevant. Hence, we interviewed patients who were having surgery rather than a sample of community at large, in which immediate postoperative considerations would not have received the thought and attention necessary to make value judgments.
To put this into concrete terms, imagine that you plan to go to a major league sporting event. Suppose one year before you have ever considered buying tickets to the game, and without reference to any particular situation, you are asked by someone on the street if you would pay $5 for a beer and $3 for a soft drink. Most rational people would say no. Flash forward to a hot, crowded sports stadium. You and your little boy are quite thirsty, and he wants a soft drink. That causes you to assess your marginal utility of income at that time for that purpose, and you end up WTP eight times what you would pay in the grocery store. You choose to spend your money that way, at that time, in that context. That is the real world. That is the essence of all rational economic transactions, and that is the essence of this study’s valuation method.
In summary, this study demonstrates that it is possible to establish patients’ value for a particular therapy. WTP estimates relates to income, previous history of PONV, PONV in the PACU, and importance of avoiding PONV in the expected direction. Patients value not having PONV and are willing to pay US$56–US$100 out of their own pocket for a totally effective antiemetic.
We appreciate the assistance of Rhonda Dear, RN, and Lee McClurkin, RN, with data collection and Angela Rogers for manuscript preparation.
WILLING TO PAY FOR AVOIDANCE OF POSTOPERATIVE NAUSEA AND VOMITING TABLETABLETABLE
1. Mark DB. Economic analysis: methods and endpoints. In: Califf RM, Mark DB, Wagner G, eds. Acute coronary care. 2nd ed. St. Louis: Mosby-YearBook, 1995: 167–82.
2. Detsky AS, Naglie IG. A clinician’s guide to cost-effectiveness analysis. Ann Intern Med 1990; 113: 147–54.
3. Macario A, Weinger M, Carney S, Kim A. Which clinical anesthesia outcomes are important to avoid: the perspective of patients. Anesth Analg 1999; 89: 652–8.
4. Philip BK. Patients’ assessment of ambulatory anesthesia and surgery. J Clin Anesth 1992; 4: 355–8.
5. Smith S, Freeland M, Heffler S. Trends: the next ten years of health spending—what does the future hold? Health Aff 1998; 17: 128–40.
6. Gan TJ, Collis R, Hetreed M. Double-blind comparison of ondansetron, droperidol and saline in the prevention of postoperative nausea and vomiting. Br J Anaesth 1994; 72: 544–7.
7. Gan TJ, Ginsberg B, Grant AP, Glass PS. Double-blind, randomized comparison of ondansetron and intraoperative propofol to prevent postoperative nausea and vomiting. Anesthesiology 1996; 85: 1036–42.
8. Watcha MF, White PF. Postoperative nausea and vomiting: its etiology, treatment, and prevention [review]. Anesthesiology 1992; 77: 162–84.
9. Myles PS, Williams DL, Hendrata M, et al. Patient satisfaction after anesthesia and surgery: results of a prospective survey of 10811 patients. Br J Anaesth 2000; 84: 6–10.
10. O’Brien BJ, Novosel S, Torrance G, Streiner D. Assessing the economic value of a new antidepressant: a willingness-to-pay approach. Pharmacoeconomics 1995; 8: 34–45.
11. Hawksworth C. Willingness to pay for “better” day surgery drugs. J One Day Surg 1996/1997;6:3–4.
12. Appel LJ, Steinberg EP, Powe NR. Risk reduction from low osmolality contrast media: what do patients think it is worth? Med Care 1990; 28: 324–37.
13. Tang J, Wang BG, White PF, et al. The effect of timing of ondansetron administration on its efficacy, cost-effectiveness, and cost-benefit as a prophylactic antiemetic in the ambulatory setting. Anesth Analg 1998; 86: 274–82.
14. Diez L. Assessing the willingness of parents to pay for reducing postoperative emesis in children. Pharmacoeconomics 1998; 13: 589–95.
15. Sloan FA, Viscusi WK, Chesson HW, et al. Alternative approaches to valuing intangible health losses: the evidence for multiple sclerosis. J Health Econ 1998; 17: 475–97.
16. Miedzybrodzka Z, Semper J, Shakely P. Stepwise or couple antenatal carrier screening for cystic fibrosis: women’s preferences and willingness to pay. J Med Genet 1995; 32: 282–3.
17. Goldstein M, Clarke A, Michelson D, et al. Developing and testing a multimedia presentation of a health-state description. Med Decis Making 1994; 14: 336–44.
18. Balaban D, Fagi P, Goldfarb N, Nettler S. Weights for scoring the quality of well-being instruments among rheumatoid arthritics: a comparison of general population weights. Med Care 1986; 24: 973–80.