NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Acute Care (UK). Preoperative Tests: The Use of Routine Preoperative Tests for Elective Surgery. London: National Collaborating Centre for Acute Care (UK); 2003 Jun. (NICE Clinical Guidelines, No. 3.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Preoperative Tests

Preoperative Tests: The Use of Routine Preoperative Tests for Elective Surgery.

Show details

Appendix 5Economics of Routine Preoperative Testing

1.1. Introduction

Preoperative testing represents a major drain on health service resources internationally. It has been estimated that the annual cost of preoperative evaluation in the USA was as much as $30bn in the 1980s1 with diagnostic testing being a substantial component. The figure for the UK is unknown. The cost is likely to be substantially less in the UK, given that salaries, overheads and the frequency of testing are all lower and that the US figure may reflect charges rather than costs. However, with 6.2 million admissions in England containing at least one surgical procedure in the financial year 2000/2001 (source: Hospital Episode Statistics: http://www.doh.gov.uk/hes/) the annual cost of ‘routine’ preoperative testing is likely to be at least in the 10s of £m. It is the subject of the economic component of this review to try to assess whether this is money that is well spent or whether the UK NHS could increase health gain by redeploying its valuable resources. Certainly a particular test is likely to be good value for money (ie cost-effective) for some groups of patients and less good value for others.

1.2. Background

1.2.1. Health economics and clinical guidelines

The explicit use of economic evaluation in clinical guideline development is a recent, but international, phenomenon. In the USA, the Committee on Clinical Practice Guidelines has recommended that every clinical guideline include cost information for alternative patient management strategies.2 In the UK, the remit of the National Institute for Clinical Excellence (NICE) is to produce national clinical guidelines that address cost-effectiveness as well as clinical effectiveness.

The reasoning behind the application of economic criteria to clinical guidelines is that no health system anywhere in the world has enough resources to provide every potentially beneficial preventative, diagnostic, curative and palliative procedure. Therefore, there is a need to redeploy resources to those procedures where the potential health gain is greatest. This requires abandoning practices that are relatively poor value for money.

There is a well-developed methodological literature for assessing the relative cost-effectiveness (value for money) of different health care procedures2–6. There is still some debate over some of the specific methods of economic evaluation in health care but essentially there are six steps to evaluating the relative efficiency of any procedure.

  1. Identify the target group (eg male preoperative patients aged 50 years), the procedure to be evaluated (eg preoperative resting ECG) and its alternative strategy (eg no preoperative resting ECG).
  2. Identify all the important health and resource outcomes that are likely to differ between the procedure and its alternative.
  3. Measure the differences in identified health and resource outcomes.
  4. Estimate the value of the health gain and the value of the resource use. [Resource use is valued in terms of its monetary value, its economic cost. Health gain is sometimes valued in monetary terms, but more often a nonpecuniary measure such as the quality-adjusted life-year (QALY) is used].
  5. Estimate the ratio of net health gain to net resource cost (eg the cost per QALY gained) and compare this with the ratios estimated for other commonly used health programmes to assess its relative efficiency. The estimation of net health gain and net cost requires some kind of model (such as a decision analysis) to combine probability and outcome information.
  6. Consider the robustness of the cost-effectiveness estimate in terms of statistical precision and generalisability to other settings.

Ideally one would repeat each of these steps for each procedure considered within the guideline (and within each procedure, for each relevant patient subgroup). This would allow us to see for which group of patients the procedure is good value for money.

1.2.2. Health and resource outcomes for routine preoperative testing

Step 1 – Eleven tests were identified for inclusion in the systematic review (Chapter 3): chest x-ray, resting ECG, full blood count (FBC), haemostasis, renal function, random glucose, urine dipstick, sickle cell, pregnancy, blood gases and lung function.

Patient subgroups identified for the purposes of the consensus process were defined by the following characteristics:

  • Gender (pregnancy test only)
  • Age group
  • Presence of a particular (common) comorbidity
  • Grade of surgery
  • Ethnic group (sickle cell test only)

Step 2 – A general scheme for the health and resource outcomes of preoperative testing is shown in Table 1. The level of each component will differ between tests and patient subgroups. Description of some of the components (eg changes to surgical practice arising from the test and complications occurring) can be found for each test under the relevant section of the systematic review (Chapter 3). However, quantitative measurement of the change in resource use resulting from these clinical outcomes is not available anywhere in the published literature.

TABLE 1. Health and resource consequences of preoperative testing.

TABLE 1

Health and resource consequences of preoperative testing.

Some components are of less importance than others are. For example, the iatrogenic effects of testing should (hopefully) be small (perhaps negligible) compared with the health gain from avoiding complications. Testing has both positive and negative effects on health; likewise there are both positive and negative effects on resource use (see Table 1). For some of the components we cannot even say which direction the change in outcome will be. Hence it is impossible to say a priori whether or not the test is beneficial overall and whether or not it is cost saving overall without having some kind of measurement and valuation of the component parts.

Steps 3 & 4 – In the literature there has been some measurement of A1 and A2 (Table 1), but only in terms of number of patients affected not in terms of the health gain per patient (see main systematic review Chapter 3). The cost of items B2 and B4 appear in the literature; evidence is presented in Sections 1.4.1 and 1.4.2. There is no evidence for any of the other components in the literature.

Steps 5 & 6 – In Section 1.4.3 below, an attempt is made to combine all the available evidence to try to assess the cost-effectiveness of the different preoperative tests.

1.2.3. Current practice in England and Wales

There are no widely accepted clinical guidelines on preoperative testing. However, the Oxford Handbook of Clinical Medicine, currently in its fifth edition, is perhaps a reasonable indicator of typical current practice in the UK. The recommendations are essentially as follows:

Routine testing – ‘most patients’

  • Urine and electrolytes
  • Blood glucose

Age indications

Other indications

  • Chest x-ray: diagnosis/pathology/symptoms of cardiorespiratory disease
  • Resting ECG: poor exercise tolerance or history of heart disease, hypertension or rheumatic fever
  • Haemostasis: history of liver disease, massive blood loss or use of heparin/warfarin
  • Blood glucose – diabetes
  • Sickle cell test: origins in Africa, West Indies, Mediterranean and other malarial regions (including most of India)

Anecdotal evidence would suggest that the adherence to these guidelines is variable between institutions and, in particular, the sickle cell test is rarely performed on potential surgical candidates.

Recommendations were also made for a few other specific tests, not considered in this review, including liver function, thyroid function, HIV and cross-matching.

The recommendations in other medical textbooks do differ from these. For example the Oxford Handbook of Surgery recommends the use of ECG in all patients over the age of 50 (rather than 65). However, the Oxford Handbook of Clinical Medicine is more influential in the UK medical education system and is therefore more likely to reflect current practice. Having an approximate definition of current practice is important when it comes to estimating the cost impact of the guidelines – see Section 1.3.4.

1.3. Methods

The Health Technology Assessment (HTA) programme’s systematic review of routine preoperative testing did not investigate cost and cost-effectiveness.7 Indeed, after thorough searching, we did not come across a comprehensive review of this subject in the published literature. Hence the systematic literature review that follows (see Section 1.3.2) may be useful. However, such a review is unlikely to capture all of the resource and health implications of preoperative testing strategies that would be relevant to the NHS.

The economic review presented in this chapter has four components:

  • estimation of unit costs of the tests under consideration;
  • a review of the literature around the economics of preoperative testing;
  • simple economic modelling of the cost-effectiveness of preoperative testing in England and Wales; and
  • simple economic modelling of the cost impact of preoperative testing in England and Wales.

1.3.1. Unit costs of tests

A dual methodology was used to collect unit cost figures:

  • review of the literature of the last six years; and
  • collection of cost data from a small sample of hospital laboratories.

For each test, the upper and lower estimates of unit cost (from both methods combined) were noted and the mid-point calculated. These figures were incorporated into the consensus documents.

Literature review

A search was carried out to find any costing or economic information regarding the selected tests (Table A1.i, Appendix 1). This was not restricted to the surgical context, as the cost of the test should be identical or similar regardless of the setting (with the exception of point-of-care testing). It is worth noting that any search for published unit costs is likely to be relatively insensitive because a unit cost is usually only a small component of an economic evaluation and hence is unlikely to get a mention in the abstract or Medical Subject Headings (MeSH®). The search was initially limited to studies conducted in the UK NHS because staff costs and overheads vary considerably between health systems. Given the rapidly developing technology in the diagnostic field the search was limited to the years 1995–2001. In addition to the databases searched for the main systematic review (Chapter 3), two specific health economic databases were searched:

Both databases include studies from the UK and overseas and both have relatively complex and comprehensive strategies for screening the medical and economic literature. The latter database reviews only full economic evaluations (ie those that systematically consider both cost and health effect), whereas the former has a broader remit and reviews all identified economic analyses. Abstracts and/or database reviews of the papers found were reviewed by the health economist and were discarded if they appeared not to contain a unit cost for any of the tests under study. Costs extracted were inflated to April 2001 prices using the health component of the Retail Prices Index.

Given the low number of relevant UK studies found, data were also collected from overseas studies. These costs were converted to pound sterling using GDP purchasing power parities for the relevant year and then inflated. Most of the overseas costs pertained to the USA. After converting charges to costs using ratios from the US Government’s Health Care Financing Administration, the estimates were in the region of five to ten times higher than those estimated for the UK, as would be expected. Consequently all non-UK unit cost estimates were excluded on the grounds of noncomparability.

Primary data collection

Six district general hospital laboratories around England and Wales were approached for unit cost information. Three hospitals responded, supplying almost complete information as requested: Luton and Dunstable Hospital, South Tyneside Hospital and Sunderland City Hospital. Unit costs were also available from research recently conducted at Central Middlesex Hospital.

The chief scientific officers in both the haematology and biochemistry laboratories filled in a questionnaire. They were asked to specify what components were included in their cost estimates and how the estimates were calculated. Additional information was also collected, such as model of equipment used; volume of tests performed etc. The data were collected using a study form (see Annex at the end of this chapter). The form was developed after detailed discussion with staff at one of the centres (Luton and Dunstable Hospital).

For the urinalysis dipstick, costs were extracted from the British National Formulary.

1.3.2. Review of preoperative evaluation costing studies

Using the same search strategy as for the main systematic review in Chapter 3, but with an additional filter to locate costing information (Table A1.i, Appendix 1), a search was performed on the databases searched in the main review plus the two health economic databases referred to above. Abstracts of papers found were reviewed by the health economist and were discarded only if:

  • they appeared not to contain any economic data; or
  • if their focus was not preoperative testing.

Relevant references in the bibliographies of reviewed papers were also identified and reviewed. Unlike the extraction of unit costs, overseas studies were included. This was justified because

  • there were very few UK studies;
  • the studies contain, in addition to unit costs, resource use data, which does not vary between health systems as much as unit costs do; and
  • study of overseas methods might be useful for the development of our own cost analysis (see Section 1.3.3).

As with the main review formal differentiation of study quality was not carried out because all studies were case series. This meant that methodological quality was consistently poor across all studies reported. The data summarised for each study include country, surgical setting, sample size, incremental cost, incremental cost per patient, incremental cost-effectiveness and cost comparison made. In some cases incremental cost or cost-effectiveness was not presented in the paper, but could be calculated from evidence that was presented. Some studies looked at the cost of preoperative testing as part of an evaluation of preoperative evaluation clinics. These studies were summarised separately.

1.3.3. Modelling of cost-effectiveness of preoperative testing for England and Wales

For each test a very simple decision analytic model was constructed like the one represented by the decision tree in Figure 1. A decision analysis simply calculates an overall outcome, for example cost, as the sum of all the individual outcomes, each weighted by the probability of that individual outcome occurring. The costs of the tests themselves were estimated from the literature and from a small sample of NHS Trusts (see Section 1.3.1). However, as noted in Table 1, the overall ‘incremental’ cost of testing to the NHS also includes certain costs arising as a consequence of testing (B2–B9) and there may be costs incurred by the patient and their families (C1–C5). An approximate cost of further diagnostic testing (B2) was estimated by assuming that it consisted of one extra outpatient appointment for all those patients with an initial positive test. This cost is clearly tentative as the real cost is unknown and varies according to the test taken, and we know that for a proportion of tests the results are not read. The mean cost of a surgical outpatient appointment was extracted from the NHS Reference Costs 2000 database.

FIGURE 1. Cost-effectiveness model.

FIGURE 1

Cost-effectiveness model.

The NHS reference cost database8 contains accounting cost data from every NHS hospital trust. Each trust reports an average cost per hospital episode, categorised by type of visit (eg outpatient, elective inpatient etc), clinical specialty and Healthcare Resource Group (HRG). The NHS Reference Costs 2000 database contains information for 69.4 million hospital episodes amounting to 88% of annual expenditure on services by NHS hospitals. Accounting practices do vary between hospitals but the costs should reflect the full cost of the service (including direct, indirect and overhead costs), as described in the NHS Costing Manual.9

The health outcomes and the remaining potential cost components were considered too difficult to quantify, even approximately, using the available evidence. Usable evidence would require specifically designed prospective studies of each test.

The systematic review did not find any evidence of changes to health outcomes. Some studies provided enough evidence to allow the calculation of the proportion of tests that resulted in a change in management. Hence cost-effectiveness for each test was calculated in terms of the incremental cost per change in management. To estimate the probability of a change in management, data on the following were taken from the main systematic review (Chapter 3):

  • abnormal test result rates (for each test this was the average of all relevant studies that included only ASA 1 and 2 patients weighted by study size); and
  • positive predictive value (for each test this was the average of all relevant studies weighted by study size).

Comparisons between tests would have to be very cautious given that the typical change in patient management and the resulting health outcome are likely to vary greatly between the tests. Sensitivity analyses were conducted to test the sensitivity of the results to the model parameters.

  • For the unit costs, the range was used.
  • For the cost of an outpatient visit, the range was used.
  • For the probabilities, the most extreme estimates from the literature review were taken (except where the most extreme estimate was zero –in this case the lowest estimate above zero was used).

1.3.4. Modelling the cost impact of the new preoperative testing guidelines

The cost of implementing the guidelines proposed in this document (Chapter 6) was calculated by estimating the expected number of each test that would be indicated by the guidelines and multiplying these numbers by the unit costs (Section 1.3.1).

The number of surgical procedures

Data on all the elective hospital Finished Consultant Episodes (FCEs) in England in 2000/2001 that contained at least one surgical procedure (n=4.7 million FCEs) were obtained from the Hospital Episode Statistics (HES) section of the Department of Health. The data were categorised by procedure code (OPCS4) and five year age bands. Emergency and maternity episodes were not included.

Severity of surgical procedures

Three surgical research fellows filled in a survey. They were given a list of the summary groups of procedures (HES Table 4) and asked to grade each one according to the severity of surgery (related to the physiological stress involved). The responders felt that some categories were too broad. For each of these cases the broad category was broken down into their three-digit OPCS4 codes (for example we omitted BD1 excision of breast and replaced it with two separate categories: B27 Total excision of breast and B28 Other excision of breast). After all three responders had completed the questionnaire, differences between responses were noted and the responders were asked to reach a consensus on each category. The resulting scheme is presented in Table 2. Applying this grading system to the HES data allowed the breakdown of FCEs by age and severity of surgery as in Table 3. The severity grading system covered 57% of surgical procedures. The remaining 43% were then allocated to each of the severity categories so as to keep the proportions of each grade the same within each age band (Table 4).

TABLE 4. FCEs with surgical operations (after allocating previously ungraded FCEs).

TABLE 4

FCEs with surgical operations (after allocating previously ungraded FCEs).

TABLE 2. Surgical procedures by severity grading.

TABLE 2

Surgical procedures by severity grading.

TABLE 3. FCEs with surgical operations.

TABLE 3

FCEs with surgical operations.

Number of patients with comorbidity

The guideline outlined in Chapter 6 recommends testing by severity score, age and evidence of comorbidity. Using more HES data, the number of FCEs were categorised according to whether they had one of the three comorbidities. All of the nonprimary diagnosis fields were searched for the ICD-10 codes presented in Table 5. It was assumed that the incidence of comorbidity would be age-related but would be independent of severity of surgery, a logistically necessary simplification. Hence the age-specific incidence of each comorbidity was multiplied with the age-specific relative frequency of each severity category, giving the number of FCEs presented in Table 6.

TABLE 5. Definition of comorbidities (ICD-10 codes).

TABLE 5

Definition of comorbidities (ICD-10 codes).

TABLE 6. FCEs with surgical operations (after allocating previously ungraded FCEs).

TABLE 6

FCEs with surgical operations (after allocating previously ungraded FCEs).

The costs of testing

The annual cost of testing was calculated simply by considering the categories of patients where testing is recommended, finding the estimated numbers of patients in these categories (Tables 4 and 6) and multiplying them by the mid-point estimates of the unit costs. Our estimates omit the cost of further diagnostic testing, which we tentatively estimated in Section 1.3.3. Also omitted were the cost components B3–B9 and C1–C5 identified in Table 1 for which evidence is not currently available.

For a number of patient categories the Guideline Development Group (GDG) could not come to a consensus on whether or not to test. For the purpose of costing the guideline we took two extreme scenarios. In the first scenario, we assume that in the areas that a consensus had not been achieved, the test is ALWAYS carried out (the ‘broad guideline’). In the alternative scenario, we assume that for these grey areas the test is NEVER carried out (the ‘narrow guideline’).

Finally to get number of tests for England and Wales instead of just England, all the figures were adjusted up by a factor of 5.9%. (The populations of England and Wales are 50.0 and 2.9 million, respectively; Source: ONS estimate for mid-2000.)

The cost impact

To estimate the cost impact of the guideline we need to know the number of routine preoperative tests being carried out at present. We do not know the frequency of testing currently so the current system was estimated by taking a set of existing guidelines, the Oxford Handbook of Clinical Medicine (see Section 1.2.3). The following, slightly stylised, definition of these guidelines used was:

  • FBC, renal function and random blood glucose tests for all patients.
  • chest x-ray and ECG for all patients over 65.
  • chest x-ray for those with respiratory comorbidity.
  • ECG and chest x-ray for those with cardiovascular comorbidity.

The cost impact was estimated to be approximately equal to the cost of these guidelines subtracted from the cost of the guidelines presented in Chapter 6.

If not an accurate estimate of the actual cost impact, this should at least indicate the difference in cost between the new guidelines and one set of existing guidelines. However, even this cost difference does not include the broader and longer term cost consequences (Table 1). A range was calculated for the cost impact using the ranges for the unit costs.

Pregnancy testing and sickle cell testing

These two tests were treated separately both because they were not related to severity of surgery and because there were no obvious differences between these guidelines and those of the Oxford handbook. The number of surgical FCEs relating to women between the ages of 15 and 50 years was taken from HES data. Data on the breakdown of the general population (for Great Britain 2000–01) were taken from the official statistics. The incidence of surgical operations was assumed to be the same for these groups as for the rest of the population (adjustments were not made for age).

1.4. Results

1.4.1. Unit costs of tests

Unit cost data was collected from four laboratories and extracted from 16 articles relating to the UK NHS.10–25 Data were also extracted from a further 34 overseas studies (28 from USA). However, as expected the unit costs appeared to be quite different in the overseas papers (in the case of the USA, the reported costs/charges were up to ten times the cost estimated in UK studies) and therefore all non-UK studies were excluded on the grounds of noncomparability. Initially 1437 abstracts were reviewed to see if they were likely to contain unit costs.

Table 7 presents the range and mid-point estimates of unit cost for each test based on the UK cost estimates. Chest x-ray and ECG were, not surprisingly, considerably more expensive than laboratory tests. The urinalysis dipstick was cheapest of all. Full blood count had the broadest relative range and ECG the broadest absolute range.

TABLE 7. Unit costs of tests.

TABLE 7

Unit costs of tests.

Components of cost estimates

The cost estimates, both collected directly, or taken from the literature, essentially include the following components:

  • cost of consumables, eg x-ray film, chemical reagents, testing kit, etc;
  • laboratory staff time; and
  • capital equipment costs, eg laboratory analysers, etc.

The exceptions are the Sickledex test, pregnancy test and urinalysis dipstick, where the cost estimates include only the cost of the kit itself. Calculating overheads for diagnostic tests is a difficult task and is not carried out consistently in all institutions. The laboratories approached did not include nondepartmental overheads in their estimates, although this component may have been included in some of the estimates from the literature. Hence these unit costs are underestimates inasmuch as they do not necessarily include overheads nor do they include the cost of the clinicians’ time in ordering and interpreting these results. These omissions are unlikely to affect the estimates greatly in absolute terms (as the clinician time involved will be small for most instances of testing), however, for the urinalysis dipstick the difference will be proportionately quite high, as the cost of the kit itself is minimal.

Also excluded were the economic costs associated with testing incurred by the patients themselves. If patients are given an additional appointment for the purposes of testing then the patient cost might be fairly substantial. If, however, the tests are carried out while patients are attending the hospital for some other reason, perhaps as part of the normal work-up for the surgery, then the incremental private cost of testing is likely to be negligible.

Cost savings from elimination of unwarranted tests

If the number of the tests were to be reduced, the proportion of the cost that would be saved in the short term varies between the tests.

  • For the kits (pregnancy, urinalysis and Sickledex), the full cost of the kit would be recovered.
  • For the other pathology tests, the reagent costs would be recovered in the short term. Many laboratories now purchase their equipment on the basis of ‘reagent contracts’, such that there is no fixed cost for the equipment but laboratories pay a mark-up on the reagents they purchase. In this case the capital cost as well as the reagent cost is recovered in the short term. Also in the short term, laboratory staff time will be freed up.
  • For ECG and chest x-ray, consumable costs will be recovered in the short term and staff time will be freed up, but capital costs will only be recovered in the long term if at all. Although these capital cost savings may not be realised financially, they should still be considered to be opportunity costs as they may allow the use of the facilities for additional patients.

1.4.2. Review of preoperative evaluation costing studies

Cost analyses of preoperative testing

We identified 13 papers that had conducted formal or informal cost analyses of preoperative testing. A further six studies had considered the cost of preoperative testing in the context of evaluating preoperative evaluation clinics; these are reported separately. The characteristics of the 13 studies are summarised in Table 8.

TABLE 8. Economic analyses of preoperative testing – study characteristics.

TABLE 8

Economic analyses of preoperative testing – study characteristics.

All the studies were coming from the perspective of seeking to reduce preoperative testing, hence total cost savings (at the sample, hospital or national level) or costs saved per patient were the outcomes used. Three studies used charges instead of economic costs.26–28

The studies were heterogeneous in the following respects:

  • tests being evaluated;
  • target population (age, type of surgery etc);
  • collection of data (prospectiveness, consecutiveness);
  • health service setting (country, health financing system, specialist ordering the test, timing of test etc); and
  • cost measures employed.

In addition they varied as to the testing strategies being compared. The comparisons made were as follows:

  1. Routine testing versus indicated testing;26–29
  2. Observed current practice versus indicated testing;10,28,30–32
  3. Observed current practice versus not testing;26,33,34
  4. Routine testing versus no testing;15,27,35–37 and
  5. Observed change in practice over time.28

In these studies ‘observed current practice’ was different to routine testing, inasmuch as for each test not every patient was tested. ‘Indicated testing’ varied between studies. They included specific clinical indications ascertained from physical examination or case history, as well as age, gender and occasionally some other sociodemographic variable. It would seem that only Kaplan et al30 did not include age as an indication. Comparison 1, best answers the theoretical question about what is the incremental cost of routine testing. However comparison 2 may give a more realistic estimate of cost savings actually achievable, given that it is quite rare for every patient to receive every test at a given institution. Comparison 3 may give an accurate estimate of cost savings but only if not testing really is a clinically acceptable option. Likewise comparison 4 is relevant if routine testing is in current practice. The ‘not testing’ option is more acceptable if the population is narrowly defined (eg only ASA grade 1 patients) and/or the study is concentrating on a single test. This was the case for all those studies that conducted comparisons 3 and 4.

The comparison chosen was related to the methodology taken (or vice versa). For example, a study calculating the cost savings of not testing compared with routine testing, only requires knowledge of the unit cost of the test and the size of the target population. One comparing current practice with indicated testing must measure the prevalence of the test in a sample population and must identify in which cases that test was indicated according to a specific protocol.

The results of the studies are summarised in Table 9. The largest estimate of potential cost saving was $190 per patient.29 Narr et al33 estimated a potential cost saving achievable in the USA of between $3bn and $4bn. Macario et al28 found that expenditure on preoperative testing was already declining by 1987, but that the reduction in test ordering was only a fraction of what could be achieved if a move to indication-only testing were to take place. They reported that there had been a reduction in indicated testing, as well as a reduction in nonindicated testing. Routine testing was by definition more costly than either of its comparators. Likewise, observed practice was by definition more costly than not testing. Those studies comparing current practice with indicated practice all found potential for cost saving. The South African study31 found that indicated testing would imply less use of chest x-rays but more use of ECG, however the cost savings attributable to the former more than offset the additional costs of the latter. An additional unpublished study, not included in the table, found potential cost savings of moving from current practice to indication-only testing of £21,000–£28,000 for a particular district general hospital in England (personal communication: John Carlisle).

TABLE 9. Economic analyses of preoperative testing – cost of routine testing.

TABLE 9

Economic analyses of preoperative testing – cost of routine testing.

With one exception, all studies considered only the cost of the test itself in the calculation of incremental cost. Hoare36 included the costs attributable to lost theatre time and the cost of following up positive test results in terms of extra clinic visits. They attributed £50 for waste of theatre time and another £50 for an extra clinic visit for each of the occasions (10/372) when surgery was delayed due to a positive test result.

Cost-effectiveness of preoperative testing

Of course the lowest cost strategy need not be the best value for money. Routine testing could, in theory, be good value for money (ie cost-effective) if there is a relatively substantial health gain.

Of the 13 cost studies identified, six provided some kind of estimate of cost-effectiveness (Tables 8 and 9). The measure of effectiveness varied between the studies as follows:

  • Number of clinically significant abnormal test results;27,30,31,35
  • Number of clinically significant abnormal test results that changed treatment;26
  • Number of complications averted;34,35 and
  • Number of lives saved.30

By definition, these studies found that testing did detect clinically significant surgical risk factors as well as increasing costs (even in ASA grade 1 patients).

Only the study that estimated the number of lives saved30 can be compared with other interventions and at $4.2m per life-saved this is considerably less cost-effective than a lot of publicly funded health care interventions. However, their calculation of effectiveness is questionable. They assume that surgical mortality for patients with an abnormal test result is only 1 in 500 and that acting on the test results prevents half of these deaths. They do not support this assumption with evidence.

The estimates of cost per complication averted, from Turnbull and Buck34 (various tests) and from Archer et al35 (chest x-ray), would represent good value for money if the complication averted were death. The less serious the complication, the less cost-effective is the test.

The cost of preoperative assessment clinics

We identified eight papers that had conducted formal or informal cost analyses of preoperative evaluation clinics. One study was excluded because it only considered the cost of the clinic itself and did not estimate the incremental cost savings. This was an important omission given that one of the main reasons for establishing such a clinic is to reduce unnecessary expenditure. The characteristics of the remaining studies are summarised in Table 10.

TABLE 10. Cost analyses of preoperative evaluation clinics – study characteristics.

TABLE 10

Cost analyses of preoperative evaluation clinics – study characteristics.

All seven studies compared the cost of preoperative evaluation in an anaesthetist-led outpatient clinic with the cost of surgeon-led preoperative evaluation after inpatient admission. Each study compared two patient cohorts apart from:

  • Pollard et al38 who made a before and after comparison of financial records (a top-down costing approach compared with the bottom-up costing method of the other studies); and
  • France et al39 who, after calculating the cost of preoperative testing in Belgium using a cohort of patients, applied the 59.3% reduction in cost estimated by Fischer.40 This method is only likely to be accurate if the testing norms in Belgium are similar to those observed by Fischer before the introduction of the preoperative evaluation clinic.

Not every study stated the timing of the clinic relative to surgery, and there was some disparity between those that did. Two studies saw all of their patients within the two weeks before surgery, whereas in MacDonald et al41 patients were seen within three months of surgery.

Four studies39,40–43 measured only the cost of preoperative testing (Table 11). McDonald et al41 also measured the other running costs of the outpatient clinic, as did Boothe et al44 who also considered the cost of operating theatre time and time in hospital. Pollard et al38 estimated the cost of time in hospital but not the cost of preoperative testing.

TABLE 11. Cost analyses of preoperative evaluation clinics – cost components measured.

TABLE 11

Cost analyses of preoperative evaluation clinics – cost components measured.

The studies were also heterogeneous in the following respects:

  • Preoperative tests being included (although those that did measure them included a whole battery of tests);
  • target population (age, type of surgery etc);
  • collection of data (prospectiveness, consecutiveness);
  • health service setting (country, health financing system, specialist ordering the test, timing of test etc); and
  • cost measures employed.

The results of the studies are summarised in Table 12. All studies measuring preoperative testing found a cost saving associated with reduced testing in the preoperative evaluation clinic arm, with the exception of MacDonald et al41 who only measured laboratory costs in that arm. The three studies that considered other cost components, all found overall cost savings with the introduction of the preoperative evaluation clinic. The largest estimate of potential cost saving from reduced preoperative testing was $112 per patient.40 Boothe et al44 estimated an overall cost saving of Can$366 per patient.

TABLE 12. Cost analyses of preoperative evaluation clinics – cost savings.

TABLE 12

Cost analyses of preoperative evaluation clinics – cost savings.

Only one study has attempted to estimate the cost savings associated with fewer surgical cancellations,44 however other studies have measured the change in the number of cancellations and these have been summarised by Fischer.40 Estimates range between 20% and 88% (see Table 13), so clearly the potential for cost saving in this area could be quite substantial.

TABLE 13. Reduction in day-of-surgery cancellations attributable to preoperative evaluation clinics.

TABLE 13

Reduction in day-of-surgery cancellations attributable to preoperative evaluation clinics.

1.4.3. Cost-effectiveness of preoperative testing in England and Wales

Table 14 shows estimates of the cost per change in management. On the basis of these cost per change in management figures, pregnancy testing, urine dipsticks and full blood count appear to be the most cost-effective for the asymptomatic patient; haemostasis, renal function and chest x-ray the least cost-effective. Interestingly these estimated cost-effectiveness rankings are almost identical to those of Robbins and Mushlin37 published more than twenty years earlier, despite very different absolute estimates of the cost per case (see Table 15). The only anomaly is urinalysis, which drops down to fifth ranking if just protein is analysed, but moves up to first place if both bacteriuria and chronic renal disease are included. Given the uncertainty about unit costs and detection rates the overall correlation might be largely spurious.

TABLE 14. Estimated cost per change in management in ASA grade 1 and 2 patients.

TABLE 14

Estimated cost per change in management in ASA grade 1 and 2 patients.

TABLE 15. Comparison with Robbins and Mushlin.

TABLE 15

Comparison with Robbins and Mushlin.

Table 16 shows that the results, in terms of cost per change in management, were sensitive to the estimates of the model parameters. In particular, the model was highly sensitive to the broad range of estimates of the probability of a positive test result and the positive predictive value.

TABLE 16. Cost per change in management – sensitivity analysis.

TABLE 16

Cost per change in management – sensitivity analysis.

Even if the estimates of cost per change in management were relatively precise, it would still not be clear which tests are cost-effective (ie good value for money) and which are not. To properly assess cost-effectiveness, we would need to know how often a change in management affects patient outcomes and what these outcomes are. If a life was saved in every ten changes of management, then it is likely that all of these tests would be considered cost-effective (Table 17). The tests could also be cost-effective, if they were to lead to substantial improvements in quality of life but no improvement in life expectancy. On the other hand, if for example a life was saved in every 10,000 tests and there was no substantial improvement in patient quality of life then none of the tests are likely to be cost-effective in nonindicated patients (Table 17).

TABLE 17. Estimated cost per life saved in ASA grade 1 and 2 patients.

TABLE 17

Estimated cost per life saved in ASA grade 1 and 2 patients.

1.4.4. The cost impact of these preoperative testing guidelines

Tables 18 to 20 show the annual number of tests for England associated with these guidelines and those of the Oxford Handbook of Clinical Medicine. It would appear that the expected number of routine preoperative tests associated with these NICE guidelines are 3.2m (0.7 per patient), with an additional 13.6m tests (2.9 per patient) up to the discretion of clinicians for those areas where the guidelines were inconclusive (the broad guideline). For each test the narrow NICE guideline represents fewer tests than the Oxford Handbook. The broad guideline, however, represents fewer of some (chest x-ray, FBC, renal function and blood glucose) and more of others (urine, haemostasis, blood gases, lung function and ECG).

TABLE 18. Recommended tests – NICE guideline (narrow).

TABLE 18

Recommended tests – NICE guideline (narrow).

TABLE 19. Recommended tests – NICE guideline (broad).

TABLE 19

Recommended tests – NICE guideline (broad).

TABLE 20. Recommended tests – Oxford Handbook.

TABLE 20

Recommended tests – Oxford Handbook.

Table 21 shows the costs of the tests for England and Wales. The tests recommended in this guideline would cost approximately £35.6m compared with an estimated cost of £130.9m associated with the guidelines contained in the Oxford Handbook. However, in the unlikely event that tests were carried out in all those cases where this guideline could not make a recommendation, then the cost of testing could be as much as £138.5m. Testing for pregnancy could cost another £2.0m and the sickle cell test possibly £0.8m

TABLE 21. Cost of routine preoperative testing in England and Wales.

TABLE 21

Cost of routine preoperative testing in England and Wales.

The comparison with the Oxford Handbook suggests that the NICE guidelines could potentially save tens of £m but this would depend on the current situation in Trusts across the country and this we do not know. Anecdotal evidence would suggest that sickle cell testing is not common at present. This would represent an additional cost. None of these calculations take into account the broader resource consequences in terms of subsequent further diagnostic testing and changes to surgical procedures. Neither can the precise effect of this change in practice on quality of care and health outcomes be determined. The magnitude of costs and cost savings were quite sensitive to the unit costs used, as represented by the sensitivity intervals in Table 21.

1.5. Discussion

The literature review appears to show that there is potential for substantial cost savings when preoperative testing is reduced. Naturally the extent of potential cost savings depends, among other factors, on one’s starting point. This varies not just between countries but also between and within institutions. In England and Wales, the current situation is not very clear. The magnitudes of cost savings as estimated in the literature are unlikely to be accurate for the NHS. In particular, the results of those studies conducted overseas are inapplicable. Our own cost impact analysis suggested that the guidelines contained in this document could potentially reduce testing costs in England and Wales, when compared with an alternative set of guidelines. However, current practice across the country is unclear and therefore the magnitude and even the direction of the change in cost are uncertain. Furthermore, the reduction of testing might not save money overall. For example, testing might lead to a reduction in the number of (risky) surgical procedures carried out; it might reduce litigation costs and the resource consequences of diagnosing chronic conditions are uncertain. Quantitative evidence for these resource outcomes and for the net health gain associated with testing is nonexistent.

A model of the cost-effectiveness of routine preoperative testing was constructed for England and Wales. This went further than the evidence in the literature, because it used unit cost estimates that are more suitable for the NHS and because it included an approximate estimate of the costs of further investigations. However, the results were not robust to the variability in its parameters (especially those taken from the systematic review) and the model omitted a number of potentially important health and resource outcomes.

The effect on patient outcomes (in terms of morbidity and mortality) of these interventions has not been measured, hence all estimates of cost-effectiveness have been based on intermediate outcomes or have been entirely speculative (or both). There are iatrogenic effects associated with some tests. One would hope that these risks are outweighed by the health gain associated with testing but again there is no quantitative evidence to support this assertion.

In conclusion, there is no good evidence that routine preoperative testing is or is not cost-effective. In particular the evidence base is lacking in terms of:

  • the quality of evidence for the number of cases detected;
  • the health outcomes associated with detecting a case; and
  • resources used (and their cost) as a consequence of detecting a case.

The context of testing may have important resource implications. A number of studies have found that anaesthetist-led preoperative evaluation clinics can save substantially on resource use. The literature suggests that valuable health service resources could be saved if:

  • staff responsible for ordering tests are those that are best informed about the utility of testing (be they surgeons or anaesthetists);
  • wherever possible tests should be conducted in advance of the day of surgery to avoid last-minute cancellations and to ensure optimal use of operating theatres (perhaps in a dedicated preoperative evaluation clinic); and
  • staff should check that the test has not already been recently ordered.

1.6. Acknowledgements

For assistance in collecting data from hospital laboratories we would like to thank Reinhold Gruen, Andrew Hutchings and Jenny Roberts (London School of Hygiene and Tropical Medicine), Richard Gray (South Tyneside Health Care NHS Trust), Colin Smith (City Hospitals Sunderland NHS Trust), David Mills and Richard Ball (Luton and Dunstable Hospitals Trust).

Annex: Unit costs – data collection form

Annex Table 1Unit cost of preoperative tests

Type of testName & type of capital equipment used (or if kit used instead, name of kit)Volume of tests (total number carried out by the lab in one year)Total cost per patient tested (£)
excluding overheadsincluding overheads (if known)
Renal function tests (U, Cr, Na, K)
Glucose tests
Urine analysis (dipstick)
Full haemoglobin count
Haemostasis
Sickle solubility test
Hb electrophoresis
Pregnancy test
Liver function tests (please specify)
Thyroid function tests
Blood viscosity test
Theophylline test
Calcium test
Blood gases test

Annex Table 2Components of unit cost

For each test, which items were included in the estimate of unit cost recorded in Annex Table 1? (please tick)*

Type of testReagentsCapital equipmentStaff timeQuality control (Int & Ext)MaintenanceOther consumables
Renal function tests (U, Cr, Na, K)
Glucose tests
Urine analysis (dipstick)
Full haemoglobin count
Haemostasis
Sickle solubility test
Hb electrophoresis
Pregnancy test
Liver function tests (please specify)
Thyroid function tests
Blood viscosity test
Theophylline test
Calcium test
Blood gases test

Annex Table 3Breakdown of unit cost

If information available, for each test please give a break down of the cost per test:

Type of testCost per patient tested (£)
ReagentsCapital equipmentStaff timeQuality control (Int & Ext)MaintenanceOther consumablesTotal cost per patient tested (excluding overheads)
Renal function tests (U, Cr, Na, K)
Glucose tests
Urine analysis (dipstick)
Full haemoglobin count
Haemostasis
Sickle solubility test
Hb electrophoresis
Pregnancy test
Liver function tests (please specify)
Thyroid function tests
Blood viscosity test
Theophylline test
Calcium test
Blood gases test

Bibliography

1.
Marcello PW, Roberts PL. “Routine” preoperative studies. Which studies in which patients. Surg Clin North Am. 1996;76:11–23. [PubMed: 8629194]
2.
Institute of Medicine. Guidelines for clinical practice: from development to use. Washington DC: National Academy Press; 1992.
3.
Drummond MF, O’Brien B, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes. Second Edition. Oxford: Oxford Medical Publications; 1997.
4.
Gold MR, Siegel JE, Russell LB, Weinstein MC, editors. Cost-effectiveness in health and medicine. New York City: Oxford University Press; 1996.
5.
National Institute for Clinical Excellence. Guidance for manufacturers and sponsors. London: National Institute for Clinical Excellence; 2001.
6.
Weinstein MC, Siegel JE, Gold MR, Kamlet MS, Russell LB. Recommendations of the Panel on Cost-effectiveness in Health and Medicine. JAMA. 1996;276:1253–1258. [PubMed: 8849754]
7.
Munro J, Booth A, Nicholl J. Routine preoperative testing: a systematic review of the evidence. Health Technol Assess. 1997;1(12) [PubMed: 9483155]
8.
NHS Executive. Reference Costs. London: Department of Health; 2000. 2001.
9.
NHS Executive. NHS Costing Manual. Leeds: Department of Health; 2001.
10.
Callaghan LC, Edwards ND, Reilly CS. Utilisation of the pre-operative ECG. Anaesthesia. 1995;50:488–490. [PubMed: 7618658]
11.
Fenwick EA, Briggs AH, Hawke CI. Management of urinary tract infection in general practice: a cost-effectiveness analysis. Br J Gen Pract. 2000;50:635–639. [PMC free article: PMC1313774] [PubMed: 11042915]
12.
Hayes PD, Lloyd AJ, Lennard N, Wolstenholme JL, London NJ, Bell PR, et al. Transcranial Doppler-directed Dextran-40 therapy is a cost-effective method of preventing carotid thrombosis after carotid endarterectomy. Eur J Vasc Endovasc Surg. 2000;19:56–61. [PubMed: 10706836]
13.
Hunt MT, Woodhouse CR. Cost-effectiveness of investigations for invasive bladder cancer. J R Soc Med. 1987;80:143–144. [PMC free article: PMC1290718] [PubMed: 3106633]
14.
Karnon J, Zeuner D, Ades AE, Efimba W, Brown J, Yardumian A. The effects of neonatal screening for sickle cell disorders on lifetime treatment costs and early deaths avoided: a modelling approach. J Public Health Med. 2000;22:500–511. [PubMed: 11192278]
15.
Livesey JR. Are haematological tests warranted prior to tonsillectomy. J Laryngol Otol. 1993;107:205–207. [PubMed: 8509696]
16.
Molyneux EM, Robson WJ. A dipstick test for urinary tract infections. J Accid Emerg Med. 1995;12:191–193. [PMC free article: PMC1342478] [PubMed: 8581245]
17.
Murray ET, Fitzmaurice DA, Allan TF, Hobbs FD. A primary care evaluation of three near patient coagulometers. J Clin Pathol. 1999;52:842–845. [PMC free article: PMC501598] [PubMed: 10690176]
18.
Pearson M, Layton C. Value of the chest radiograph before cardiac catheterisation in adults. Br Heart J. 1994;72:505. [PMC free article: PMC1025632] [PubMed: 7857730]
19.
Preston A, O’Donnell T, Phillips CA. Screening for urinary tract infections in a gynaecological setting: validity and cost-effectiveness of reagent strips. Br J Biomed Sci. 1999;56:253–257. [PubMed: 10795368]
20.
Ratcliffe J, Ryan M, Tucker J. The costs of alternative types of routine antenatal care for low-risk women: shared care vs care by general practitioners and community midwives. J Health Serv Res Policy. 1996;1:135–140. [PubMed: 10180859]
21.
Roberts JA, Upton PA, Azene G. Escherichia coli O157:H7; an economic assessment of an outbreak. J Public Health Med. 2000;22:99–107. [PubMed: 10774911]
22.
Robertson JF, Whynes DK, Dixon A, Blamey RW. Potential for cost economies in guiding therapy in patients with metastatic breast cancer. Br J Cancer. 1995;72:174–177. [PMC free article: PMC2034118] [PubMed: 7599049]
23.
Scott A, Currie N, Donaldson C. Evaluating innovation in general practice: a pragmatic framework using programme budgeting and marginal analysis. Fam Pract. 1998;15:216–222. [PubMed: 9694178]
24.
Underwood SR, Godman B, Salyani S, Ogle JR, Ell PJ. Economics of myocardial perfusion imaging in Europe—the EMPIRE Study. Eur Heart J. 1999;20:157–66. [PubMed: 10099913]
25.
Zeuner D, Ades AE, Karnon J, Brown J, Dezateux C, Anionwu EN. Antenatal and neonatal haemoglobinopathy screening in the UK: review and economic analysis. Health Technol Assess. 1999;3(11) [PubMed: 10516927]
26.
Adams JG, Weigelt JA, Poulos E. Usefulness of preoperative laboratory assessment of patients undergoing elective herniorrhaphy. Arch Surg. 1992;127:801–804. [PubMed: 1524480]
27.
Kettler RE. The cost of preoperative pregnancy testing. Anesth Analg. 1996;83:439–440. [PubMed: 8694348]
28.
Macario A, Roizen MF, Thisted RA, Kim S, Orkin FK, Phelps C. Reassessment of preoperative laboratory testing has changed the test-ordering patterns of physicians. Surg Gynecol Obstet. 1992;175:539–547. [PubMed: 1448735]
29.
Velanovich V. How much routine preoperative laboratory testing is enough. Am J Med Qual. 1993;8:145–151. [PubMed: 8219877]
30.
Kaplan EB, Sheiner LB, Boeckmann AJ, Roizen MF, Beal SL, Cohen SN, et al. The usefulness of preoperative laboratory screening. JAMA. 1985;253:3576–81. [PubMed: 3999339]
31.
Sommerville TE, Murray WB. Information yield from routine pre-operative chest radiography and electrocardiography. S Afr Med J. 1992;81:190–196. [PubMed: 1738905]
32.
Wattsman TA, Davies RS. The utility of preoperative laboratory testing in general surgery patients for outpatient procedures. Am Surg. 1997;63:81–90. [PubMed: 8985077]
33.
Narr BJ, Hansen TR, Warner MA. Preoperative laboratory screening in healthy Mayo patients: cost-effective elimination of tests and unchanged outcomes. Mayo Clin Proc. 1991;66:155–159. [PubMed: 1899710]
34.
Turnbull JM, Buck C. The value of preoperative screening investigations in otherwise healthy individuals. Arch Intern Med. 1987;147:1101–1105. [PubMed: 3592875]
35.
Archer C, Levy AR, McGregor M. Value of routine preoperative chest x-rays: a meta-analysis. Can J Anaesth. 1993;40:1022–1027. [PubMed: 8269561]
36.
Hoare TJ. Pre-operative haemoglobin estimation in paediatric ENT surgery. J Laryngol Otol. 1993;107:1146–1148. [PubMed: 8289005]
37.
Robbins JA, Mushlin AI. Preoperative evaluation of the healthy patient. Med Clin North Am. 1979;63:1145–1156. [PubMed: 529881]
38.
Pollard JB, Zboray AL, Mazze RI. Economic benefits attributed to opening a preoperative evaluation clinic for outpatients. Anesth Analg. 1996;83:407–410. [PubMed: 8694327]
39.
France FH, Lefebvre C. Cost-effectiveness of preoperative examinations. Acta Clin Belg. 1997;52:275–286. [PubMed: 9489121]
40.
Fischer SP. Development and effectiveness of an anesthesia preoperative evaluation clinic in a teaching hospital. Anesthesiology. 1996;85:196–206. [PubMed: 8694365]
41.
MacDonald JB, Dutton MJ, Stott DJ, Hamblen DL. Evaluation of pre-admission screening of elderly patients accepted for major joint replacement. Health Bull (Edinb) 1992;50:54–60. [PubMed: 1612896]
42.
Power LM, Thackray NM. Reduction of preoperative investigations with the introduction of an anaesthetist-led preoperative assessment clinic. Anaesth Intensive Care. 1999;27:481–488. [PubMed: 10520388]
43.
Starsnic MA, Guarnieri DM, Norris MC. Efficacy and financial benefit of an anesthesiologist-directed university preadmission evaluation center. J Clin Anesth. 1997;9:299–305. [PubMed: 9195353]
44.
Boothe P, Finegan BA. Changing the admission process for elective surgery: an economic analysis. Can J Anaesth. 1995;42:391–394. [PubMed: 7614645]
45.
Azzam FJ, Padda GS, DeBoard JW, Krock JL, Kolterman SM. Preoperative pregnancy testing in adolescents. Anesth Analg. 1996;82:4–7. [PubMed: 8712424]
46.
Velanovich V. The value of routine preoperative laboratory testing in predicting postoperative complications: a multivariate analysis. Surgery. 1991;109:236–243. [PubMed: 2000554]
47.
Velanovich V. Preoperative laboratory screening based on age, gender, and concomitant medical diseases. Surgery. 1994;115:56–61. [PubMed: 8284762]
48.
Haucotte G, Lefèbvre CH, Bauwens P, Vandermeersch E. Etude BIOMED, Evaluation du bilan préopératoire. Rapport au Ministère de la Santé Publique et de l’Environnement; Brussels: 1996.
49.
Hand R, Levin P, Stanziola A. The cause of cancelled elective surgery. Qual Asur Utilization Rev. 1990;5:2–6. [PubMed: 2136658]
50.
Macarthur A, Macarthur C, Bevan J. Preoperative assessment clinic reduces day surgery cancellations [abstract] Anesthesiology. 1991;75:A11091.
51.
Department for Transport, Local Government and the Regions. Highway economics note 1: 2000. London: Department for Transport, Local Government and the Regions; 2000.
Copyright © 2003, National Collaborating Centre for Acute Care.
Bookshelf ID: NBK48483

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...