How to dissect surgical journals: xiii economic analyses

ANZJSurg.com
How to dissect surgical journals: XIII – Economic analyses*ans_56471.3
We come here to serve god, and also make money. example, pharmaceutical companies and manufacturers of surgical devices. Johnson-Lans3 commented that ‘oligopolistswill often try to differentiate their products from those of their Economic analyses link the use of resources with clinical outcomes.
competitors. They are likely to devote an even higher propor- At the core are interactions between patients, health-care providers tion of their budgets to advertising and selling costs. This and health funds.1 This means that we are traversing red seas created behaviour tends to erect barriers to new firms attempting to by bloody competition rather than blue skies. Everybody wants to enter the market, since it develops brand loyalty to the firm’s Investigators usually assume that they are dealing with a ‘closed Always look for the loser in economic analyses system’ that is frozen in time. The end result is a ‘snapshot’ of theexperience of a particular health-care facility. This may make inter- Studies that carefully attempt to unravel these interactions are esting reading, but it does not promote confidence in the inevitable rare. More often, authors use economic analyses to push a conviction generalizations and predictions that follow. The wide conclusions or boost a thin manuscript. Some of the reasons why readers of drawn from cost evaluations are often discordant with their narrow surgical journals should be wary of economic analyses: (1) Costs are elastic, for example, buyer-groups have bargaining power, employees can negotiate salaries, third-party payers An economist is someone who knows the price of everything and can squeeze health-care providers and governments inter- vene. It is a bit iffy to assume that costs will hold still for more (2) Both assets and consequences degrade over time – but at what Doctors are people who know the price of nothing and think they rate? How do you value clinical benefit and harm? Discount- ing is the term that economists use for devaluation over time.
Despite being subjective and highly variable, it is generallyheld to be between 3% and 5%.
(3) It is impossible to predict the implications of uncommon Estimating costs
major adverse events or ‘black swan events’, for example, It requires a lot of effort to determine the monetary value of benefits.
mass litigation related to silicon implants.
There are two issues. First, the nature of the items to be costed, and, (4) Demands for care are not static. They are influenced by social second, the quantity of the resources consumed or saved and their and economic factors; for example, recipients of workers unit cost. It is important that these items are documented because compensation benefits tend to have longer rehabilitation times.
(5) Insurance can promote ‘adverse selection’, that is, attract Let us look at the cost of an operation. The salient costs relate to patients with poor health. Askerlof2 pointed out that people rent (hospital stay, operating room time), professional fees (sur- with cars that are ‘lemons’ will tend to trade them in earlier geons, nurses, anaesthetists and paraclinical services) and consum- than owners of reliable cars, which tends to reduce both the ables (surgical supplies, drugs, disposable items). The upstream price and quality of used cars. Health-care insurers try to costs embrace all of the preoperative events. The downstream costs overcompensate for this phenomenon by ‘beneficial selec- include all of the future costs attributable to the procedure including tion’, that is, by insuring healthy individuals. Nevertheless, secondary procedures (such as removing a plate from a bone), reha- insured patients tend to overuse services.
bilitation and productivity losses. So, working out the cost of an (6) Savings can be hard to realize. For instance, long-acting anti- operation is not a trivial exercise.
biotics reduce the costs of administration, including staff An additional issue is that new interventions ride on the back of time. However, unless this is realized by reducing the cost of new technology. Besides the cost of new equipment, there is also the salaries, then that is only a potential or ‘opportunity’ saving.
cost of training staff. So initially, new technology can be ‘cost- (7) Health care does not work as a free economy. It is serviced by increasing’ – the cost of equipment is high, staff are on a learning oligopolies with a restricted number of competitors, for curve and the demand for interventions can increase. For instance, *The 16 articles in this series are being made available on ANZJSurg.com the popularity of laparoscopic surgery has resulted in an increase in the prevalence of cholecystectomies: fewer patients with gallstones 2011 The AuthorANZ Journal of Surgery 2011 Royal Australasian College of Surgeons are being managed by ‘masterly inactivity’. In general, fee-for- colectomy. The data probably represent an accurate snapshot of the service systems promotes innovation, while managed care can slow in-hospital costs of open and minimally invasive colectomy at the down the diffusion of new technology.
time of the study. However, inspection of the itemized hospital costsshows that patients undergoing open colectomy had greater variationin the length of stay in hospital and higher non-operating room costs.
The cost of innovations depends on the stage of their life cycle A possible explanation is that patients undergoing open colectomystayed in hospital longer because of complications (the authors made Weisbrod4 coined the term ‘halfway technologies’ for innovations no mention of complications and failed to study patients once they that are still evolving. These issues can be controversial, especially left the hospital). Based on the contents of this article, rather than when territorial boundaries are breached.
preconceived notions, the data do not support the conclusion thatminimally invasive colectomy ‘is cost-effective and results in sig-nificant savings to the health-care system’. Unfortunately, the litera- Cost-benefit analyses
ture is full of this type of extravagant extrapolation.
Benefits include ‘soft’ but important outcomes that estimate theimpact of an intervention. This is difficult because patients vary in The incremental cost-effectiveness ratio
their perception of adversity according to their personalities and pastexperiences. There is a need to be cautious about attempts to express Economists work around the margins. When comparing two surgical overall health status as a number. Current efforts include the estima- interventions, it is possible to estimate the benefit that is gained from tion of quality of life, the calculation of benefit as a ‘health year a marginal increase in the cost. This is called the incremental cost- equivalents’ and opinions about trade-offs (the amount that patients effectiveness ratio (ICER). The numerator of the ICER is the mar- suggest that they are prepared to pay for an outcome). Other studies ginal difference of the mean cost of each intervention, and the look at more concrete outcomes. Cost-consequences analysis relates denominator is the marginal mean difference of the effectiveness).
to specific events, for example, limbs salvaged, avoidance of acancer, years of survival.
Cost-utility refers to the general usefulness of an intervention. In the best studies, it tests the ‘day-to-day’ robustness of cost-benefits Bell et al.5 noted that cost-effectiveness analyses tend to report in large numbers of patients. In lesser studies, it is just an alternative ‘positive’ or ‘negative’ results but not intermediate results. They proposed three explanations for these findings. First, it may reflect Benefits are sometimes expressed as the cost of quality-adjusted the truth because manufacturers do not bring economically unattrac- life-years (QALYs). QALYs are based on the utility score of various tive interventions to market. Second, analysts and journals may not degrees of health – 0 represents death and 1 represents perfect be interested in interventions with mid-range cost-effectiveness health. The results of cost-benefit analyses are usually expressed as ratios. Third, there may be biases that prevent studies with unfavour- the ‘cost per QALY’. This enables comparisons to be made between able ratios from reaching the journals. There is a concern that the different interventions. If hip replacement for arthritis represents a base unit for costs, then kidney transplants and breast screening areabout 5 to 6 times as much, and neurosurgery for malignant tumours Reporting economic analyses in clinical trials
But there are traps. An absurd situation arose when the impact of Surgical investigators sometimes piggyback an economic evaluation photodynamic therapy for macular degeneration was evaluated in onto a clinical trial. This approach offers the prospect of high-quality QALYs. The argument was that the blind have poor QALYs, and information about effectiveness and costs to be measured simulta- therefore a low claim for limited resources. The National Institute neously. However, some authors use analyses of costs to bolster for Clinical Excellence in the United Kingdom solved this problem by ruling that photodynamic therapy for macular degeneration I scanned surgical trials published over 3 years (2007–2009) in 26 journals – the 21 surgical journals mentioned in my previous litera-ture survey8 plus five medical journals (BMJ, Lancet, N. Engl. J.
Med., Ann. Intern. Med., and JAMA). Ten percent of the trials (55/ Cost-effectiveness analyses
Cost-effectiveness studies compare the costs of alternative interven- (1) Forty-four percent (24/55) were formal studies (they men- tions when the clinical effectiveness can be measured as a single tioned costs in the methods and results sections). Most were variable, for example, survival time, the rate of a successful inter- studies of cost-efficiency (19/24). Only 33% (8/24) men- vention. So, unlike cost-benefit analyses, there is no need to attach a tioned costs in the title, but 88% (21/24) included comments monetary value on the outcome. The term ‘cost-minimization analy- about costs in the abstract. The commonest descriptor used in sis’ is sometimes used when comparing the cost of interventions that these studies was ‘cost analysis’. There was wide variation in the sophistication of the measurements of cost.
As in all financial documents, there is a need to read the small (2) Fifty-six percent (31/55) just made comments about costs, print. Salloum et al.7 evaluated the costs of minimally invasive usually in the discussion section (24/31). It is of interest that ANZ Journal of Surgery 2011 Royal Australasian College of Surgeons in 18 of these 24 articles, the comments were restricted to • Comprehensive economic analyses are rare – they usually either the penultimate or the last paragraph. In one-half of describe the experience of a particular health-care facility at these articles (9/18), the authors made useful comments about the cost of specific items or just raised the issue of cost.
• Cost-benefit analyses seek to determine the monetary value of However, in the other one-half of these articles, the authors made assertions about cost-efficiency in the absence of any • Results can be expressed as the cost of QALYs.
• The ICER calculates the marginal benefit that is gained from a The tendency to make unfounded ‘last-minute’ assertions about costs is a concern. For example, Miura et al.9 evaluated post- • Cost-effectiveness studies compare the costs of alternative tonsillectomy pain in children. Costs were not listed among the interventions when the clinical effectiveness can be measured secondary outcomes nor were they mentioned in the methods or as a single variable, for example, survival time.
results section, yet the last line of the discussion stated that ‘Because • Cost-consequences analyses relate to specific events for it is simple, safe, tolerated, and of low cost, it is our opinion that example, limbs salvaged, avoidance of a cancer, years of topical sucralfate is an important tool in adjuvant treatment of post- tonsillectomy pain’. Although sucralfate might be cheap to buy, the • Cost-minimization analyses compare the cost of interventions costs of administering it as an irrigation at the end of a procedure – while the patient is still anaesthetized – and as a mouth rinse, four • The cost of innovations depends on the stage of their life cycle.
times a day for five days, cannot be ignored. Furthermore, the • Look for sensitivity analyses that estimate the robustness of the qualifying phrase ‘it is our opinion’ was not included in the abstract, which read ‘Because it is simple, safe, tolerated, and low cost, it is References
an important tool as adjuvant treatment of post-tonsillectomy pain’.
1. Brazier JE, Johnson AG. Economics of surgery. Lancet. 2001; 358:
Evaluating economic analyses
2. Askerlof G. The market for ‘lemons’: qualitative uncertainty and the Economic analyses should be valid, reliable and relevant. Ask the market mechanism. Q. J. Econ. 1970; 84: 488–500.
following questions when evaluating an economic analysis10,11: 3. Johnson-Lans S. A Health Economics Primer. Boston: Addison-Wesley, 4. Weisbrod BA. The health care quadrilemma: an essay on technological change, insurance, quality of care, and cost containment. J. Econ. Lit. Have all the indirect costs and benefits been included? 1991; 29: 523–52.
Have the numbers of items been declared separately from 5. Bell CM, Urbach DR, Ray JG et al. Bias in published cost effectiveness studies: systematic review. BMJ. 2006; 332: 699–703.
6. Harford T. The Undercover Economist. London: Abacus, 2006.
Could my patients expect similar health outcomes? 7. Salloum RM, Bulter DC, Schwartz SI. Economic evaluation of mini- Have patient and caregiver costs been estimated? mally invasive colectomy. J. Am. Coll. Surg. 2006; 202: 269–74.
8. Hall JC. How to dissect surgical journals: III – the contents of surgical journals. ANZ J. Surg. 2010; 80: 658–60.
Were all the relevant clinical strategies compared? 9. Miura MS, Saleh C, de Andrade M, Assmann M, Ayres M, Neto JFL.
Has there been an allowance for long-term consequences and Topical sucralfate in post-adenotonsillectomy analgesia in children: a double-blind randomizd clinical trial. Otolaryngol. Head Neck Surg. 2009; 141: 322–8.
10. Drummond MF, Richardson WS, O’Brien BJ, Levine M, Heyland D.
Users’ guides to the medical literature XIII. How to use an article on Sensitivity analysis is an exploration of ‘what ifs’. It estimates the economic analysis of clinical practice. A. Are the results of the study robustness of the analysis; for example, is laparoscopic surgery still valid. JAMA. 1997; 277: 1552–7.
of good value if the cost of equipment rises by 10%? Are the fees for 11. O’Brien BJ, Heyland D, Richardson WS, Levine M, Drummond MF.
anaesthesia the same as for your operating lists? What allowances Users’ guides to the medical literature XIII. How to use an article on might be made for the bargaining power of purchasers? Needless to economic analysis of clinical practice. B. What are the results and say, great care must be exercised before the results of economic will they help me in caring for my patients. JAMA. 1997; 277:
analyses are allowed to influence clinical actions.
Key points
• Economic analyses link the use of resources with clinical The University of Western Australia, Perth, Western Australia, Australia • Costs are elastic and are unlikely to hold still for more than 1 2011 The AuthorANZ Journal of Surgery 2011 Royal Australasian College of Surgeons

Source: http://www.anzjsurg.com/SpringboardWebApp/userfiles/anzjs/file/Dissecting%2013.pdf