WHEN we were anesthesia residents in Palo Alto and New Haven two decades ago, patients on the morning of the surgery would often say, “I’m worried I won’t wake up after the operation.” Except for the sickest patients undergoing complex surgery, we do not hear these concerns anymore. The public now believes that anesthesia is extraordinarily safe from catastrophic outcomes, including major organ dysfunction and death. In 2006, many patients are concerned about pain and postoperative nausea or vomiting (PONV). This is especially true for the majority of patients undergoing less invasive surgery.
What is the value of interventions directed toward such transient outcomes? In this issue of Anesthesiology, Kalkman et al. 1add to the growing scientific literature supporting not only the patient’s desire to avoid pain and PONV, but the monetary valuation of relief from pain and PONV. Their 808-patient study found that the median willingness to pay (out of pocket) for a pill that completely prevents pain after surgery amounted to $35 (interquartile range, 7–69), whereas patients were willing to pay $17 (interquartile range, 7–69) for a pill that completely prevents PONV.
Upon hearing such results, practitioners often ask, “How can those dollar numbers be useful? Willingness to pay will depend on the patient’s wealth or income, and patients are rarely asked to pay for health care.”
The goal of this editorial is to explain why willingness-to-pay studies are done. To some extent, the answer is straightforward: Economic evaluations such as the cost–utility analyses traditionally used to assess chronic care are not directly transferable to acute conditions such as pain and nausea because these typically only last for a few hours or days.
Our second goal is to challenge the specialty of anesthesiology to also emphasize the patient’s desire to eliminate pain and PONV from clinical practice in the same way safety was emphasized during the past several decades.
Although the willingness-to-pay research methodology is conceptually simple and intuitive, it remains controversial. Its theoretical underpinning is from welfare economics, which aims to maximize social welfare by examining the economic activities of individuals, as opposed to communities or societies. This branch of economics assumes individuals are the best judges of their own welfare and are able to measure welfare adequately either in monetary terms or as a preference.
In a free market, shopping decisions are based on the commodity’s price and the individual’s reservation price, or the maximum perceived worth of that commodity. However, this does not occur in healthcare systems with third-party payors because patients are shielded from the true costs of care. With willingness to pay, benefits (including nonhealth attributes such as convenience) are valued in monetary terms so that outcomes are comparable across all interventions.
In general, willingness-to-pay data are useful in three settings:
Helping a company set the price for a new service or intervention (e.g. , drug or device) that is paid fully by the patient (e.g. , an over-the-counter medicine).
When a health system has decided to offer a new service or intervention but wants to charge the patient some amount of money to recover some of the costs. As an example, this was done in Tanzania to determine what rural populations were willing to pay for cataract surgery.2
Offer guidance in selecting which rival interventions are preferred by patients as criteria for payment (coverage) by the payor. For example, diabetic patients may prefer inhaled insulin over subcutaneous insulin.3
The willingness-to-pay values of $35 for analgesia and $17 for PONV relief reported in the study1are lower than those in previous studies performed in North America and may be related in part to income differences. The annual median household income for the Dutch patients studied was approximately $25,000. In contrast, 66 patients studied at Stanford in 2001, with incomes roughly twice as high, reported being willing to pay $110 to avoid pain and $100 for PONV relief. Also, the methodology used at Stanford was different, because these patients were asked how much money they were willing to pay to prevent an outcome that would hypothetically last for 6 h postoperatively.4Outcome descriptions were created using patients’ own words.
As far back as 1992, Fred Orkin, M.D., published an abstract reporting that patients were willing to pay $15 out of pocket to avoid any immediate postanesthetic side effect and $50 to avoid emetic episodes. A separate study of 80 patients in North Carolina from 2001 determined $56 as the median willingness to pay to avoid PONV.5A Canadian study found that 86% of patients were willing to pay at least $50 and as much as $200 for patient-controlled analgesia after surgery.6
Certainly how you ask the questions (e.g. , having patients fill out a paper questionnaire in their living room while trying to think about pain vs. being interviewed at the bedside in the postanesthesia care unit while experiencing pain) or whether the issue is defined as prevention or treatment may affect the willingness-to-pay answers. So how should we interpret these dollar amounts? First, they support the importance of reducing the incidence of common but less serious side effects of surgery and anesthesia. A survey of 56 anesthesiologists around the United States revealed that the five items considered most important to patients were incisional pain, nausea, vomiting, preoperative anxiety, and discomfort from intravenous line insertion.7Patients concurred.8
Because burgeoning healthcare costs are straining budgets in all countries, there is pressure to define the value of interventions (e.g. , PONV prophylaxis, processed electroencephalographic signal monitoring, antiinfective-coated central venous catheters, continuous nerve blocks) by assessing their costs and benefits, including long-term health and complication risks. As anesthesiologists, we care for a wide spectrum of patients and procedures, from patients with complex medical conditions undergoing major cardiovascular procedures to healthy individuals undergoing routine surgery.
Economic analyses do not necessarily require that less money should be spent. Typically, an economic evaluation of a new medical intervention would take the form of a cost–utility analysis. A cost–utility analysis measures the incremental net cost of performing an intervention (expenditures for the intervention minus savings from future healthcare costs) and compares it with the marginal, or incremental, benefit obtained.
For such cost–utility studies, quality of life must be quantified, i.e. , converted into units that can be compared among different health conditions, which are typically chronic (e.g. , renal failure requiring dialysis, heart failure). Utilities are numerical ratings or “preference weights” of the desirability of health states that reflect a person’s preferences, on a linear scale from 0.00 (death) to 1.00 (perfect health). Preference values for health states are commonly obtained using valuation techniques such as the standard gamble, the time trade-off, or the visual analog scale.
Quality-adjusted life years (QALYs) reflect both quality of life and duration of survival. QALYs are obtained by multiplying the utility value for a given health condition by its duration. For example, an additional year of survival for an individual in perfect health (with a utility of 1.0) is considered equivalent to a patient having two additional years but with a health state with a utility of 0.5. The QALYs equal 1 in both cases.
The incremental cost–utility ratio is equal to (C2− C1)/(QALY2− QALY1), or incremental costs divided by incremental QALYs. The output of a cost–utility analysis is reported as $/QALY. Medical interventions are considered to be cost effective when they produce health benefits at a cost comparable to that of other commonly accepted treatments. Although no consensus has been reached, many commonly used interventions cost less than $50,000 per QALY gained. These are considered cost effective and should be performed. Those that cost $50,000–100,000 per QALY are questionable, and those above $100,000 per QALY are not considered cost effective. Prevention of a perioperative myocardial infarction, renal failure, or major infection may fall within this framework and be deemed cost effective. As anesthesiologists, we continue to study and implement new strategies to reduce these complications.
However, many anesthesia side effects are so short lasting that it is not possible for an intervention with transient effect to have an incremental cost–utility ratio smaller than the commonly accepted threshold of $50,000. An example follows to illustrate this important notion. Let us assume that the surgery health state without any anesthesia has a utility weight of 0.1 (on the 0–1 scale). If the surgery lasted an hour, the general anesthetic that allows the patient to be analgesic, amnestic, and unconscious (assume a utility to the patient of 1) would result in a utility gain of 0.9 (1 minus 0.1), multiplied by its duration in years (0.000114 yr = 1 h/8,760 h in a year), for a total QALY gain of 0.0001. That is, the QALY gain is the difference in QALY weight (1.0 vs. 0.1) multiplied by the number of years for which the health state is experienced.
Based on a threshold of $50,000, the maximum amount that the society should be willing to pay for an anesthetic for an hour-long surgery requiring general anesthesia should be $5. However, most people in the United States would readily place the value of an anesthetic to be much more than that, in part because patients perceive that major complications are averted. This example indicates that cost–utility methodology commonly used for chronic diseases can lead to less than optimal economic decisions for acute illness.
Because translating the value of healthcare benefits (e.g. , decreased pain and suffering) into monetary terms is tricky, the need arises for willingness-to-pay values. These then theoretically could feed into a cost–benefit analysis (a second, different type of economic assessment), which forces an explicit decision about whether the benefit is worth the cost.
Contingent valuation involves querying individuals directly regarding the maximum amount they are willing to pay to have the commodity in question or the minimum amount they would be willing to accept in compensation for being deprived of it. Willingness to pay is a primary tool that health economists have for valuing the effect of acute conditions on patients. Outside of anesthesia, and for purposes of illustration, patients reported a willingness to pay of $1,200 for benefits of ultrasound in an uncomplicated pregnancy (e.g. , reassurance of normal fetal anatomy and growth),9$100 per month for a new treatment of advanced non–small-cell lung cancer,10and $50 to avoid the minor side effects of intravenous contrast dye using nonionic agents.11
In most industries, the quality of the product is assessed by the customer, in our case the patient. The quality of medical decisions, patient satisfaction, and clinical outcomes can be improved by eliciting patient preferences. For example, the choice of an opiate to relieve postoperative pain may actually reduce the quality of the recovery period of a postoperative patient who considers nausea more objectionable than pain. In this patient, a less emetogenic, nonopioid analgesic may provide the patient with the postoperative outcome that he or she desires. For many patients undergoing outpatient surgery, it may also justify prescribing a new brand-name pharmaceutical that is more efficacious than the current standard, or convincing industry to develop new agents that the patient would be willing to pay for out of pocket.
In the same way the specialty of anesthesiology focused on developing drugs, devices, knowledge, and protocols to dramatically reduce catastrophic events related to anesthesia in the operating room, postanesthesia care unit, and intensive care units, anesthesiologists are now challenged to bring similar quality improvement efforts to pain and nausea, and practically eliminate those as well. Continuous peripheral nerve blocks reduce pain compared with opioids for certain surgeries,12but not all eligible patients get them. Why not? Multimodal PONV prevention based on a priori risk scores also reduce PONV.13But not all high-risk PONV patients get prophylaxis. The current article by Kalkman et al. 1further highlights the importance of pain and PONV treatment from the patient’s perspective.
* Department of Anesthesia, Stanford University School of Medicine, Stanford, California. † Department of Anesthesiology and Critical Care, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania. firstname.lastname@example.org