- Email: [email protected]
Overuse of preoperative laboratory coagulation testing and ABO blood typing: a French national study
British Journal of Anaesthesia, 119 (6): 1186–93 (2017) doi: 10.1093/bja/aex268 Advance Access Publication Date: 9 October 2017 Quality and Patient Safety
QUALITY AND PATIENT SAFETY
Overuse of preoperative laboratory coagulation testing and ABO blood typing: a French national study H. Beloeil1,*, D. Ruchard2, N. Drewniak2 and S. Molliex3 1
Poˆle anesthe´sie-re´animation, CHU Rennes, Inserm Numecan CIC 1414, Universite´ Rennes 1, 35033 Rennes Cedex, France, 2Hospitalisation Department, National Health Insurance, CNAMTS, 50 avenue du Pr Andre´ Lemierre, 75986 Paris Cedex 20, France and 3Departement d’Anesthesie-Reanimation, CHU Hopital Nord, Universite´ de Saint-Etienne, 42055 Saint-Etienne Cedex 2, France *Corresponding author. E-mail: [email protected]
Abstract Background: Following publication of guidelines on routine preoperative tests, the French Society of Anaesthesiology and Intensive Care (SFAR), in association with French national public health insurance, conducted a survey to evaluate adherence to guidelines and the economic consequences. Methods: Using the French Hospital Discharge Database and National Health Insurance Information system, tests performed during the 30 days before surgery were analysed for two situations: (1) standard laboratory coagulation tests and ABO blood typing in children able to walk and scheduled for tonsillectomy/adenoidectomy; and (2) ABO blood typing in adults before laparoscopic cholecystectomy, thyroidectomy, lumbar discectomy or breast surgery. Guidelines do not recommend any preoperative tests in these settings. Results: Between 2013 and 2015, a coagulation test was performed in 49% of the 241 017 children who underwent tonsillectomy and 39% of the 133 790 children who underwent adenoidectomy. A similar pattern was observed for ABO blood typing although re-operation rates for bleeding on the first postoperative day were very low (0.12–0.31% for tonsillectomy and 0.01– 0.02% for adenoidectomy). Between 2012 and 2015, ABO blood typing was performed in 32–45% of the 1 114 082 patients who underwent one of the four selected procedures. The transfusion rate was very low (0.02–0.31%). The mean cost for the four procedures over the 4 yr period was e5 310 000 (SD e325 000). Conclusions: Standard laboratory coagulation tests and ABO blood typing are still routinely prescribed before surgery and anaesthesia despite current guidelines. This over-prescription represents a high and unnecessary cost, and should therefore be addressed. Key words: preoperative tests; haemostasis; blood typing
Routine testing is part of preoperative assessment designed to: (1) reduce the risks associated with the procedure and/or anaesthesia; (2) detect unsuspected conditions on the basis of patient interviews and/or clinical examination that could modify surgical/anaesthetic management; (3) provide a reference for postoperative assessment; and/or (4) evaluate the risk of postoperative
complications with an independent predictive value. Previous studies have shown that systematic preoperative testing of unselected patients provides only a minor contribution to these objectives. When Blery and colleagues1 published their study on preoperative tests before cataract surgery in 1986, more than 20 studies had already shown that preoperative testing could be
Editorial decision: July 1, 2017; Accepted: July 16, 2017 C The Author 2017. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved. V
For Permissions, please email: [email protected]
1186
Overuse of preoperative laboratory coagulation testing and ABO blood typing
Editor’s key points • Preoperative testing is frequently overused against the
recommendations of national guidelines. • The frequency of preoperative standard laboratory coag-
ulation testing and blood typing were retrospectively analysed in two large French databases for six common surgical procedures. • Both standard coagulation tests and blood typing were performed in 32–49% of patients, despite very low rates of bleeding and transfusion, leading to high unjustified costs.
reduced without increasing postoperative morbidity, and 60% of tests could be eliminated without affecting patient safety.2 This situation has not changed over the last 30 yr. Recent studies continue to report over-requesting of preoperative tests— reflecting the lack of clear guidelines or consensus with level 1 evidence concerning preoperative testing.3 4 In 2012, because of the absence of regulatory standards, the French Society of Anaesthesiology and Intensive Care (SFAR)5 issued guidelines based on an extensive review of the literature in order to rationalize and consequently reduce preoperative tests that are predominantly prescribed in France by anaesthesiologists at preoperative consultations. The guidelines were endorsed by 17 surgical and medical scientific societies that have promoted their use. To encourage implementation of these guidelines, SFAR organized local and national educational programmes in 2012 and thereafter. Physicians’ observations of these guidelines and the efficacy of the implementation strategy (including continuing education meetings with tailored interventions and dissemination of printed educational materials) to improve professional practices remain unknown and need to be evaluated. We therefore decided to study a large national cohort of patients scheduled for surgical procedures in which testing is generally not recommended.5 We retrospectively studied French nationwide medical and administrative databases to assess routine haemostasis tests and blood typing in two different clinical settings: (1) children able to walk and scheduled for tonsillectomy/adenoidectomy; and (2) adults before four surgical procedures (i.e. laparoscopic cholecystectomy, thyroidectomy, lumbar discectomy or breast surgery). These procedures were chosen because they are performed routinely and do not require routine preoperative testing. We hypothesized that preoperative testing rates would decrease in France following the implementation of these guidelines.
Methods Data source We performed a retrospective study using two French nationwide datasets linked by an anonymous and unique patient identifier: (1) The French Hospital Discharge database (PMSI), which contains medical procedures for all patients admitted to public and private hospitals identified by their code according to the Classification Commune des Actes Me´dicaux [(CCAM), i.e. French medical classification for clinical procedures)] and discharge diagnoses coded according to the International Classification of Diseases, 10th revision (ICD-10 codes) and registered as principal, related or associated diagnoses; and (2) The
|
1187
French National Health Insurance information system (SNIIRAM), which collects all individualized healthcare claims reimbursed by French National Health Insurance covering more than 96% of the French population. This database records all reimbursements for patient care and comprises internal and external quality control processes. The database comprehensively records all outpatient prescriptions, services and procedures performed, together with their date of prescription and implementation. This database also contains patient data such as long-term diseases (LTDs) coded according to ICD-10, which are eligible for 100% health insurance coverage, prescribed medications identified according to Anatomical Therapeutic Classification (ATC) code and laboratory investigations identified from the French national laboratory test coding table (NABM: nomenclature des actes de biologie me´dicale). This linkage has been previously used to conduct large-scale epidemiological studies.6 7 The French Data Protection Authority (CNIL: Commission Nationale de l’informatique et des Liberte´s) has approved data collection from the SNIIRAM by the National Health Insurance (CNAMTS).8
Study population, variables of interest and outcomes Two patient populations were evaluated. Specific surgical procedures were identified according to their CCAM code. The index date was the date of the first hospitalization for the type of surgery considered. Group 1: all patients 2 yr and <18 yr of age scheduled for tonsillectomy and/or adenoidectomy between 2013 and 2015 were included. All prescriptions for haemostasis tests [prothrombin time (PT), activated partial thromboplastin time (aPTT) and ABO blood typing with testing for irregular agglutinins and Rh phenotype determination] prescribed or performed within 30 days before hospitalization for surgery were collected from the SNIIRAM databases. This 30 day period corresponds to the validity period for preoperative anaesthesiology consultations recommended by SFAR. Patients in whom comorbidities and medications deemed to be risk factors for bleeding or impaired haemostasis were identified by hospitalization discharge/LTD diagnoses and medications dispensed at least once before the index date were excluded (Supplementary data, Table S1). Re-operation for bleeding on the first postoperative day and hospital re-admission within 7 days were collected as outcome measures. Group 2: all patients 18 yr of age scheduled for laparoscopic cholecystectomy, thyroidectomy (excluding associated tracheal or laryngeal resections and extra-cervical approach), lumbar discectomy (excluding associated laminectomy or interbody fusion) or breast tumour surgery (excluding radical mastectomy, mastectomy with flap reconstruction or dermal autograft) between 2012 and 2015 were included. These four surgical procedures were selected because of a transfusion rate 1%9 regardless of patient medical conditions. All ABO blood typing, testing for irregular agglutinins and Rh phenotype determinations prescribed or performed within 30 days before hospitalization for surgery were collected from the SNIIRAM databases. Red blood cell transfusions during the first two postoperative days were collected as outcome measures. According to guidelines, these tests should not be performed in these two patient populations.5 The time period of this study was before/during implementation of guidelines for the four surgical procedures and only during implementation of guidelines for adenoidectomy and tonsillectomy.
1188
|
Beloeil et al.
Statistics Statistical analysis was performed using XLSTAT (Paris, France) V3 software. For each type of surgical procedure and each preoperative test, patients with preoperative testing were compared with patients without testing by v2 test to determine whether the preoperative testing decreased over the time period considered. The relative risks (RR) and confidence interval (95% CI) for the various outcomes according to the presence or the absence of preoperative testing were also determined.
Results The study population scheduled for tonsillectomy and adenoidectomy in 2013, 2014 and 2015 is presented in Table 1. When both procedures were performed at the same time, they were listed as tonsillectomy. The haemostasis test rate before adenoidectomy and tonsillectomy was high in a population in which these tests are not recommended by SFAR guidelines5 (Table 2). Haemostasis tests were performed in 49% of children before tonsillectomy and in 39% of children before adenoidectomy. This rate decreased significantly between 2013 and 2015. A similar pattern was observed for ABO blood typing, testing for irregular agglutinins and Rh phenotype determination. The preoperative testing rate was higher for tonsillectomy than for adenoidectomy, for which preoperative testing rates continued to decline between 2013 and 2015. The percentage of reduction of preoperative testing between 2013 and 2015 was between 1.8% and 4.2% regardless of the surgical procedure. The average annual cost of unnecessary preoperative testing over the 3 yr period was e1 758 000 (SD e97 000) and e895 000 (SD e71 000) for tonsillectomy and adenoidectomy, respectively. The reoperation rate for bleeding on the first postoperative day was very low during the study period, ranging from 0.12% to 0.31% for tonsillectomy and was even lower for adenoidectomy (0.01– 0.02%) (Table 3). The re-operation rate was higher in patients scheduled for tonsillectomy without preoperative testing. An additional search was performed in patients who were reoperated for bleeding on the first postoperative day following tonsillectomy according to the presence or absence of preoperative testing. During the study period considered (2013–5), none of these patients received vasopressin or vasopressin analogues
or blood coagulation factors defined by their ATC codes or presented related or principal diagnoses of coagulation defects or haemorrhagic conditions defined by their ICD-10 codes from the day of surgery or during the following year. The hospital readmission rate within 7 days was higher following tonsillectomy than adenoidectomy, but was statistically independent of preoperative testing (Table 3). A total of 1 114 082 patients underwent one of the four selected surgical procedures in France between 2012 and 2015 (Table 4). The ABO blood typing, testing for irregular agglutinins and Rh phenotype determinations rates ranged from 28% to 52% of patients scheduled for surgery (Table 5). A higher preoperative testing rate was observed for lumbar discectomy than for the other three procedures and the lowest rate was observed for breast tumour surgery. The preoperative testing rate decreased from 2012 to 2014, but increased again in 2015 reaching values higher than those observed in 2012 for lumbar discectomy (þ1.9%, þ4.1% and þ1.9% for ABO blood typing, testing for irregular agglutinins and Rh phenotype determination, respectively). The annual cost of this unnecessary preoperative testing over the 4 yr period was e2 330 000 (e143 000); e852 932 (e68 000), e724 098 (e26 000) and e1 400 000 (e107 000) for laparoscopic cholecystectomy, thyroidectomy, lumbar discectomy or breast tumour surgery, respectively. The cumulative cost for the four surgical procedures over the 4 yr period was e5 310 000 (e325 000). Red blood cell transfusion rate during the first 48 postoperative hours was extremely low (<1%) regardless of surgical procedure (Table 6). Red blood cell transfusion rate after cholecystectomy was significantly higher in patients in whom preoperative testing was not performed.
Discussion This study confirmed a high rate of unnecessary preoperative testing based on more than 100 000 patients undergoing adenoidectomy or tonsillectomy and more than 270 000 patients undergoing one of the four selected surgical procedures. These results are in accordance with the literature reporting overprescription of preoperative testing.10 This is the first study on this subject based on 4 yr data from two national databases, constituting a large cohort of patients following publication of
Table 1 Study population for tonsillectomy and adenoidectomy in 2013, 2014 and 2015. The sum of the excluded patients related to ATC codes for medications and ICD-10 codes for long-term diseases and for comorbidities is greater than the number of patients actually excluded because the same patient sometimes presented several reasons for exclusion. ATC, Anatomical Therapeutic Classification; ICD10, International Classification of Diseases, 10th revision
Tonsillectomy Patients included, n (%) Patients excluded, n (%) Exclusion related to ATC codes for medications, n Exclusion related to ICD-10 codes for long-term diseases, n Exclusion related to ICD-10 codes for comorbidities, n Adenoidectomy Patients included, n (%) Patients excluded, n (%) Exclusion related to ATC codes for medications, n Exclusion related to ICD-10 codes for long-term diseases, n Exclusion related to ICD-10 codes for comorbidities, n
2013
2014
2015
60 816 (93.7) 4096 (6.3) 3546 160 621
62 729 (94.3) 3796 (5.7) 3206 168 619
56 656 (94.4) 3385 (5.6) 2830 175 637
45 826 (93.9) 2971 (6.1) 2741 72 213
44 639 (94.2) 2724 (5.8) 2523 52 209
42 325 (94.6) 2409 (5.4) 2190 53 198
Overuse of preoperative laboratory coagulation testing and ABO blood typing
|
1189
Table 2 Comparisons between patients with preoperative testing and patients without preoperative testing over the time period considered were analysed with v2 tests. Rh, rhesus; PT, prothrombin time; aPTT, activated partial thromboplastin time Tonsillectomy PT n testing rate (%) aPTT n testing rate (%) ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%) Adenoidectomy PT n testing rate (%) aPTT n testing rate (%) ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%)
2013
2014
2015
P-value
29 318 48.2
28 147 44.9
25 137 44.4
<0.0001
33 687 53.2
32 047 51.1
29 110 51.4
<0.0001
30 681 50.4
29 312 46.7
26 761 47.2
<0.0001
28 904 47.5
27 977 44.6
25 674 45.2
<0.0001
30 359 49.9
29 173 46.5
26 476 46.7
<0.0001
17 154 37.4
15 876 35.6
14 378 34.0
<0.0001
20 140 43.9
18 394 41.2
16 823 39.7
<0.0001
16 310 35.6
14 813 33.2
13 579 32.1
<0.0001
14 710 32.1
13 607 30.5
12 592 29.8
<0.0001
16 170 35.3
14 753 33.0
13 474 31.8
<0.0001
Table 3 Re-operation rate for bleeding at D0 and hospital re-admission within 7 days following tonsillectomy and adenoidectomy in 2013, 2014 and 2015. RR, relative risk; 95% CI, 95% confidence interval; D0, day 0 Tonsillectomy Re-operation for bleeding at D0 Preoperative testing – (%) Preoperative testing þ (%) RR (95% CI) Hospital re-admission within 7 days Preoperative testing – (%) Preoperative testing þ (%) RR (95% CI) Adenoidectomy Re-operation for bleeding at D0 Preoperative testing – (%) Preoperative testing þ (%) RR (95% CI) Hospital re-admission within 7 days Preoperative testing – (%) Preoperative testing þ (%) RR (95% CI)
2013
2014
2015
0.28 0.15 0.56 (0.39–0.80)
0.31 0.13 0.42 (0.28–0.61)
0.24 0.12 0.50 (0.33–0.76)
1.72 1.85 1.07 (0.95–1.21)
1.96 1.78 0.91 (0.81–1.02)
2.00 2.07 1.03 (0.92–1.16)
0.01 0.01 0.84 (0.08–9.31)
0.02 0.01 0.26 (0.03–2.13)
0.01 0.01 0.66 (0.06–6.30)
0.22 0.18 0.82 (0.53–1.26)
0.16 0.19 1.20 (0.76–1.90)
0.27 0.20 1.39 (0.92–2.10)
1190
|
Beloeil et al.
Table 4 Study population for the four selected surgical procedures in 2012, 2013, 2014 and 2015. See text for exclusions
Laparoscopic cholecystectomy Partial or total thyroidectomy Lumbar discectomy Breast tumour surgery Total
2012
2013
2014
2015
112 935
113 662
115 433
117 585
47 622
46 033
45 051
42 817
33 605 86 738 280 900
32 442 85 637 277 774
31 710 85 959 278 153
30 601 86 252 277 255
Table 5 Preoperative ABO blood typing, testing for irregular agglutinins and Rh phenotype determination for laparoscopic cholecystectomy, partial or total thyroidectomy, lumbar discectomy and breast tumour surgery in 2012, 2013, 2014 and 2015. Rh, Rhesus. Comparisons between patients with preoperative testing and patients without preoperative testing over the time period considered were analysed with v2 tests 2012
national guidelines and an intensive implementation strategy; therefore, providing major support to the results. The statistically significant decrease in preoperative testing rates during the study period was not clinically meaningful. Outcomes were not modified in patients without preoperative testing except for the re-operation rate in patients scheduled for tonsillectomy, which was higher in these patients. In accordance with the literature, perioperative events were not different in patients with or without preoperative testing.11 In minor procedures such as cataract surgery, morbidity and mortality are not reduced by routine preoperative tests.11 The incidences of re-operation for bleeding after adenoidectomy, re-admission within 7 days after adenoidectomy and tonsillectomy, and transfusion within 48 h after the four selected surgical procedures were very low and not different between patients with or without preoperative tests. Standard laboratory coagulation tests and ABO blood typing are still over-prescribed,4 10 more than any other tests for a number of reasons, including defensive medicine to prevent complaints or criticism. Personal experience also constitutes an important factor.12 13 Physicians with personal histories of perioperative bleeding continue to order preoperative testing despite guidelines. However, a blood test is not necessarily beneficial to the patient because of the risk of false positives for statistical, pre-analytical and analytical reasons, thereby triggering further follow-up testing and inappropriate patient management. Further investigation can delay surgery and increase the number of unnecessary tests for the patient. Systematic preoperative standard laboratory coagulation tests fail to predict the risk of bleeding,14 15 16 17 An abnormal test could be associated with a non-haemorrhagic disease. For example, a deficit in factor XII increases aPTT with no clinical consequences. In the general population, aPTT results follow a normal distribution: 2.5% of the patients have an aPTT higher than two standard deviations from the mean with no increased bleeding risk. In asymptomatic patients, the prevalence of congenital clotting factor deficiencies associated with bleeding risk is approximately 1 in 40 000. If routine aPTT could detect all asymptomatic patients with an increased risk (1/40 000), the probability of a randomly increased aPTT (1/40) would be 1000 times greater than the probability of detecting increased aPTT associated with haemorrhagic disease. The predictive value of aPTT is only 0.1% in a symptomatic male patient. In our study, re-operations for bleeding after tonsillectomy were more frequent in patients without preoperative tests. Considering the incidence of asymptomatic patients with an increased bleeding risk because of congenital deficiencies of coagulation factors (1/40 000),18 this situation would concern an average of one
Laparoscopic cholecystectomy ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%) Partial or total thyroidectomy ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%) Lumbar discectomy ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%) Breast tumour surgery ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%)
2013
2014
2015
P-value
44 182 42 997 40 516 43 054 <0.0001 39.1 37.8 35.1 36.6
52 951 52 218 50 315 54 300 <0.0001 46.9 45.9 43.6 46.2
43 895 42 751 40 376 42 610 <0.0001 38.9 37.6 35.2 36.2
17 511 16 728 14 786 14 824 <0.0001 36.8 36.3 32.8 34.6
20 104 19 547 17 668 17 842 <0.0001 42.2 42.5 39.2 41.7
17 271 16 580 14 733 14 663 <0.0001 36.3 36.0 32.7 34.2
14 387 14 141 13 335 13 674 <0.0001 42.8 43.6 42.1 44.7
16 036 15 951 15 166 15 839 <0.0001 47.7 49.2 47.8 51.8
14 320 14 102 13 330 13 623 <0.0001 42.6 43.5 42.0 44.5
29 249 27 411 24 003 25 225 <0.0001 33.7 32.0 27.9 29.2
33 364 31 883 29 891 31 481 <0.0001 38.5 37.2 34.8 36.5
28 420 26 632 23 891 24 966 <0.0001 32.8 31.1 27.8 28.9
patient each year among those with preoperative testing and those without preoperative testing in our study. This does not explain the difference in the re-operation rate for bleeding between patients with and without preoperative testing (45 vs
Overuse of preoperative laboratory coagulation testing and ABO blood typing
|
1191
Table 6 Red blood cells transfusion rate within 48 h after laparoscopic cholecystectomy, partial or total thyroidectomy, lumbar discectomy and breast tumour surgery in 2012, 2013, 2014 and 2015. RR, relative risk; CI 95%, 95% confidence interval
Laparoscopic cholecystectomy Transfusion rate at within 48 hours after Preoperative testing – (%) Preoperative testing þ (%) RR (CI 95%) Partial or total thyroidectomy Transfusion rate at within 48 hours after Preoperative testing – (%) Preoperative testing þ (%) RR (CI 95%) Lumbar discectomy Transfusion rate at within 48 hours after Preoperative testing – (%) Preoperative testing þ (%) RR (CI 95%) Breast tumour surgery Transfusion rate at within 48 hours after Preoperative testing – (%) Preoperative testing þ (%) RR (CI 95%)
2012
2013
2014
2015
0.28 0.16 0.57 (0.43–0.75)
0.31 0.14 0.46 (0.35–0.62)
0.25 0.12 0.46 (0.33–0.64)
0.27 0.14 0.53 (0.40–0.71)
0.13 0.08 0.65 (0.35–1.22)
0.10 0.03 0.32 (0.12–0.81)
0.07 0.04 0.55 (0.21–1.46)
0.06 0.03 0.52 (0.17–1.54)
0.09 0.02 0.25 (0.07–0.86)
0.06 0.06 1.07 (0.43–2.65)
0.04 0.06 1.54 (0.58–4.11)
0.04 0.05 1.05 (0.36–3.03)
0.15 0.14 0.93 (0.63–1.37)
0.14 0.16 1.13 (0.78–1.66)
0.15 0.16 1.09 (0.75–1.59)
0.16 0.15 0.92 (0.62–1.34)
88 in 2013, 36 vs 108 in 2014 and 30 vs 77 in 2015, respectively). Moreover, from the day of surgery and during the following year, none of the re-operated patients received treatment for or was diagnosed with a coagulation defect. This difference could therefore be explained by a surgical haemostasis problem rather than an undetected haemostatic disease. Another risk associated with systematic preoperative haemostatic tests is falsenegative results: a normal aPTT does not guarantee the absence of coagulopathy. The sensitivity of aPTT is highly dependent on the severity of the factor deficiency. The sensitivity of aPTT to detect haemophilia is 100% in the presence of a severe deficiency but only 90% in the case of moderate deficiency (16–30%). It can be as low as 48% in von Willebrand disease. Factor XIII deficiency is not detected by any routine tests. Studies have shown that haemostatic tests can be useful when preceded by bleeding history questionnaire. However, no specific questionnaire for the assessment of bleeding risk before surgery has been validated. SFAR guidelines recommend that the following items be included in the questionnaire: tendency for prolonged/ unusual bleeding (epistaxis, small cuts), tendency to develop ecchymoses/bruising/haematomas, prolonged bleeding after tooth extraction, major bleeding after surgery (especially circumcision or tonsillectomy), family history, and in women, menorrhagia or postpartum haemorrhage. When a thorough physical examination and interview are performed, the addition of blood tests does not modify patient management15 and is therefore of little value. Our results show that the re-operation rate for bleeding after tonsillectomy was lower than that reported in the literature, in which post-tonsillectomy bleeding rates are 3–5%.19 Blood typing before surgery is appropriate in patients likely to require blood transfusion. In agreement with European guidelines,20 SFAR does not recommend blood typing when there is a low risk of transfusion (i.e. <1%).5 Our results showed that blood typing was performed in more than 50% of patients before surgical procedures associated with very low transfusion
rates. In a validated algorithm to help physicians avoid unnecessary type and screen orders before surgery, Dexter and colleagues21 showed that blood typing is not necessary when the transfusion rate is lower than 5%. However, we observed a high rate of unjustified tests despite our choice of surgical procedures associated with a very low risk of transfusion. Unjustified preoperative testing represents a very high cost for society. Shein22 estimated the cost of these tests to be $150 million per year in the USA in 1996. In our study, the annual cost for haemostasis tests and ABO blood typing for just six surgical procedures was over e7 million for 370 000 patients out of the 10 million patients undergoing surgical procedures. The potential savings are therefore considerable. Our study highlights the inconsistency between guidelines and practice10 23 as well as the lack of effectiveness of the implementation strategy developed following publication of guidelines. Despite quality processes involving continuing medical evaluation, test rates have not decreased.24 25 Audits comprising feedback are also widely used to improve practice and compliance with guidelines. However, their effectiveness remains controversial, as Ivers and colleagues26 reported a significant increase in healthcare professional compliance with desired practice but insignificant improvement in patient outcome. Tailored intervention strategies have been suggested.27 The latest meta-analysis on the subject included 32 studies, and reported a modest benefit to improve professional practices and healthcare outcomes.28 Computerized systems could also help improve performance,29 but regardless of the technique, no significant improvement has yet been observed.24 30 This study has several of limitations. Data analysis was retrospective and data were derived from administrative databases with inherent limitations concerning data collection and coding. The main strength of this study concerns the use of the SNIIRAM database with data for more than 96% of the French population in 2013. The impact of bias related to population groups insured by other schemes as a result of their occupation
1192
|
Beloeil et al.
or sector of activity is probably low. However, the algorithms used in this study cannot take into account preoperative tests performed on inpatients in the public healthcare system. This study therefore underestimates preoperative testing rates, although these inpatient tests probably represent only a small proportion of all tests performed. Furthermore, this database records reimbursements for patient care and comprises internal and external quality control processes ensuring exhaustiveness of the database. However, exclusions were performed from healthcare utilization data and cannot take into account subjects with diseases not managed by a hospital stay or a specific LTD or medication prescription. Outcome misclassification could constitute another limitation because the external validity of the various outcome measures used in this study has not been previously assessed in the French PMSI database. However, principal related or associated hospital discharge diagnoses were used to define outcomes thereby limiting the risk of underestimation.
2. 3.
4.
5. 6.
7.
Conclusions Standard laboratory coagulation tests and ABO blood typing before surgery and anaesthesia are still performed routinely even in clinical situations in which they are not medically justified. In this context, these tests are unnecessary and have no clinical impact. Compliance of physicians with guidelines and the effectiveness of the implementation strategy developed by the French Society of Anaesthesiology remain very limited, raising the question of the capacity of physicians to change their behaviour in the absence of incentives.
8.
9.
10.
Authors’ contributions Helped conduct the study, analysing the data and writing the manuscript: H.B. and S.M. Helped conduct the study: D.R. Helped analysing the data and conducted the statistical analysis: N.D. All authors have seen the original study data, reviewed the analysis of the data and approved the final manuscript.
11.
12.
13.
Supplementary material Supplementary material is available at British Journal of Anaesthesia online.
14.
Declaration of interest This work was supported by the Socie´te´ Franc¸aise d’Anesthe´sie et de Re´animation (SFAR). National guidelines concerning Preoperative testing were published by the SFAR. H.B. and S.M. were members of the working group which elaborated these guidelines and involved as experts in the local and national educational programmes sponsored by the SFAR.
Funding
15.
16.
17.
Funding was solely institutional.
References 1.
Blery C, Charpak Y, Szatan M, et al. Evaluation of a protocol for selective ordering of preoperative tests. Lancet 1986; 1: 139–41
18. 19.
Roizen MF. The compelling rationale for less preoperative testing. Can J Anaesth 1988; 35: 214–8 Benarroch-Gampel J, Sheffield KM, Duncan CB, et al. Preoperative laboratory testing in patients undergoing elective, low-risk ambulatory surgery. Ann Surg 2012; 256: 518–28 Bernard R, Benhamou D, Beloeil H. [Routine preoperative testing: Impact of implementation of local recommendations in a teaching hospital]. Ann Fr Anesth Reanim 2010; 29: 868–73 Molliex S, Pierre S, Blery C, Marret E, Beloeil H. [Routine preinterventional tests]. Ann Fr Anesth Reanim 2012; 31: 752–63 Maura G, Blotiere PO, Bouillon K, et al. Comparison of the short-term risk of bleeding and arterial thromboembolic events in nonvalvular atrial fibrillation patients newly treated with dabigatran or rivaroxaban versus vitamin K antagonists: a French nationwide propensity-matched cohort study. Circulation 2015; 132: 1252–60 Fagot JP, Blotiere PO, Ricordeau P, Weill A, Alla F, Allemand H. Does insulin glargine increase the risk of cancer compared with other basal insulins? A French nationwide cohort study based on national administrative databases. Diabetes Care 2013; 36: 294–301 Roussel R, Chaignot C, Weill A, et al. Use of fibrates monotherapy in people with diabetes and high cardiovascular risk in primary care: a French Nationwide Cohort Study Based on National Administrative Databases. PloS One 2015; 10: e0137733 Puel C, Ducharne T, Mialon A, et al. [Surgical risk of transfusion in a French Universitary Hospital]. Ann Fr Anesth Reanim 2012; 31: 132–40 Bryson GL, Wyand A, Bragg PR. Preoperative testing is inconsistent with published guidelines and rarely changes management. Can J Anaesth 2006; 53: 236–41 Schein OD, Katz J, Bass EB, et al. The value of routine preoperative medical testing before cataract surgery. Study of Medical Testing for Cataract Surgery. N Engl J Med 2000; 342: 168–75 Lippi G, Favaloro EJ, Franchini M. Dangers in the practice of defensive medicine in hemostasis testing for investigation of bleeding or thrombosis: part I–routine coagulation testing. Semin Thromb Hemost 2014; 40: 812–24 Toker A, Shvarts S, Perry ZH, Doron Y, Reuveni H. Clinical guidelines, defensive medicine, and the physician between the two. Am J Otolaryngol 2004; 25: 245–50 Asaf T, Reuveni H, Yermiahu T, et al. The need for routine pre-operative coagulation screening tests (prothrombin time PT/partial thromboplastin time PTT) for healthy children undergoing elective tonsillectomy and/or adenoidectomy. Int J Pediatr Otorhinolaryngol 2001; 61: 217–22 Bonhomme F, Ajzenberg N, Schved JF, et al. Pre-interventional haemostatic assessment: Guidelines from the French Society of Anaesthesia and Intensive Care. Eur J Anaesthesiol 2013; 30: 142–62 Levy JH, Szlam F, Wolberg AS, Winkler A. Clinical use of the activated partial thromboplastin time and prothrombin time for screening: a review of the literature and current guidelines for testing. Clin Lab Med 2014; 34: 453–77 Baron DM, Metnitz PG, Fellinger T, Metnitz B, Rhodes A, Kozek-Langenecker SA. Evaluation of clinical practice in perioperative patient blood management. Br J Anaesth 2016; 117: 610–6 Janvier G. [Preoperative evaluation of hemorrhagic risk]. Ann Fr Anesth Reanim 1998; 17(Suppl 1): 2s–5s Mitchell RM, Parikh SR. Hemostasis in tonsillectomy. Otolaryngol Clin North Am 2016; 49: 615–26
Overuse of preoperative laboratory coagulation testing and ABO blood typing
20. De Hert S, Imberger G, Carlisle J, et al. Preoperative evaluation of the adult patient undergoing non-cardiac surgery: guidelines from the European Society of Anaesthesiology. Eur J Anaesthesiol 2011; 28: 684–722 21. Dexter F, Ledolter J, Davis E, Witkowski TA, Herman JH, Epstein RH. Systematic criteria for type and screen based on procedure’s probability of erythrocyte transfusion. Anesthesiology 2012; 116: 768–78 22. Schein OD. Assessing what we do. The example of preoperative medical testing. Archiv Ophthalmol 1996; 114: 1129–31 23. van Gelder FE, de Graaff JC, van Wolfswinkel L, van Klei WA. Preoperative testing in noncardiac surgery patients: a survey amongst European anaesthesiologists. Eur J Anaesthesiol 2012; 29: 465–70 24. Sigmund AE, Stevens ER, Blitz JD, Ladapo JA. Use of preoperative testing and physicians’ response to professional society guidance. JAMA Intern Med 2015; 175: 1352–9 25. Forsetlund L, Bjorndal A, Rashidian A, et al. Continuing education meetings and workshops: effects on professional
26.
27.
28.
29.
30.
|
1193
practice and health care outcomes. Cochrane Database Syst Rev 2009; CD003030 Ivers N, Jamtvedt G, Flottorp S, et al. Audit and feedback: effects on professional practice and healthcare outcomes. Cochrane Database Syst Rev 2012; CD000259 Leung BC, Nazeer S, Smith M, McRae D. Reducing unnecessary preoperative testing in elective ENT surgery: clinical and financial implications. J Perioper Pract 2015; 25: 225–9 Baker R, Camosso-Stefinovic J, Gillies C, et al. Tailored interventions to address determinants of practice. Cochrane Database Syst Rev 2015; CD005470 Garg AX, Adhikari NK, McDonald H, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA 2005; 293: 1223–38 Flamm M, Fritsch G, Seer J, Panisch S, Sonnichsen AC. Nonadherence to guidelines for preoperative testing in a secondary care hospital in Austria: the economic impact of unnecessary and double testing. Eur J Anaesthesiol 2011; 28: 867–73 Handling editor: Hugh C Hemmings Jr
QUALITY AND PATIENT SAFETY
Overuse of preoperative laboratory coagulation testing and ABO blood typing: a French national study H. Beloeil1,*, D. Ruchard2, N. Drewniak2 and S. Molliex3 1
Poˆle anesthe´sie-re´animation, CHU Rennes, Inserm Numecan CIC 1414, Universite´ Rennes 1, 35033 Rennes Cedex, France, 2Hospitalisation Department, National Health Insurance, CNAMTS, 50 avenue du Pr Andre´ Lemierre, 75986 Paris Cedex 20, France and 3Departement d’Anesthesie-Reanimation, CHU Hopital Nord, Universite´ de Saint-Etienne, 42055 Saint-Etienne Cedex 2, France *Corresponding author. E-mail: [email protected]
Abstract Background: Following publication of guidelines on routine preoperative tests, the French Society of Anaesthesiology and Intensive Care (SFAR), in association with French national public health insurance, conducted a survey to evaluate adherence to guidelines and the economic consequences. Methods: Using the French Hospital Discharge Database and National Health Insurance Information system, tests performed during the 30 days before surgery were analysed for two situations: (1) standard laboratory coagulation tests and ABO blood typing in children able to walk and scheduled for tonsillectomy/adenoidectomy; and (2) ABO blood typing in adults before laparoscopic cholecystectomy, thyroidectomy, lumbar discectomy or breast surgery. Guidelines do not recommend any preoperative tests in these settings. Results: Between 2013 and 2015, a coagulation test was performed in 49% of the 241 017 children who underwent tonsillectomy and 39% of the 133 790 children who underwent adenoidectomy. A similar pattern was observed for ABO blood typing although re-operation rates for bleeding on the first postoperative day were very low (0.12–0.31% for tonsillectomy and 0.01– 0.02% for adenoidectomy). Between 2012 and 2015, ABO blood typing was performed in 32–45% of the 1 114 082 patients who underwent one of the four selected procedures. The transfusion rate was very low (0.02–0.31%). The mean cost for the four procedures over the 4 yr period was e5 310 000 (SD e325 000). Conclusions: Standard laboratory coagulation tests and ABO blood typing are still routinely prescribed before surgery and anaesthesia despite current guidelines. This over-prescription represents a high and unnecessary cost, and should therefore be addressed. Key words: preoperative tests; haemostasis; blood typing
Routine testing is part of preoperative assessment designed to: (1) reduce the risks associated with the procedure and/or anaesthesia; (2) detect unsuspected conditions on the basis of patient interviews and/or clinical examination that could modify surgical/anaesthetic management; (3) provide a reference for postoperative assessment; and/or (4) evaluate the risk of postoperative
complications with an independent predictive value. Previous studies have shown that systematic preoperative testing of unselected patients provides only a minor contribution to these objectives. When Blery and colleagues1 published their study on preoperative tests before cataract surgery in 1986, more than 20 studies had already shown that preoperative testing could be
Editorial decision: July 1, 2017; Accepted: July 16, 2017 C The Author 2017. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved. V
For Permissions, please email: [email protected]
1186
Overuse of preoperative laboratory coagulation testing and ABO blood typing
Editor’s key points • Preoperative testing is frequently overused against the
recommendations of national guidelines. • The frequency of preoperative standard laboratory coag-
ulation testing and blood typing were retrospectively analysed in two large French databases for six common surgical procedures. • Both standard coagulation tests and blood typing were performed in 32–49% of patients, despite very low rates of bleeding and transfusion, leading to high unjustified costs.
reduced without increasing postoperative morbidity, and 60% of tests could be eliminated without affecting patient safety.2 This situation has not changed over the last 30 yr. Recent studies continue to report over-requesting of preoperative tests— reflecting the lack of clear guidelines or consensus with level 1 evidence concerning preoperative testing.3 4 In 2012, because of the absence of regulatory standards, the French Society of Anaesthesiology and Intensive Care (SFAR)5 issued guidelines based on an extensive review of the literature in order to rationalize and consequently reduce preoperative tests that are predominantly prescribed in France by anaesthesiologists at preoperative consultations. The guidelines were endorsed by 17 surgical and medical scientific societies that have promoted their use. To encourage implementation of these guidelines, SFAR organized local and national educational programmes in 2012 and thereafter. Physicians’ observations of these guidelines and the efficacy of the implementation strategy (including continuing education meetings with tailored interventions and dissemination of printed educational materials) to improve professional practices remain unknown and need to be evaluated. We therefore decided to study a large national cohort of patients scheduled for surgical procedures in which testing is generally not recommended.5 We retrospectively studied French nationwide medical and administrative databases to assess routine haemostasis tests and blood typing in two different clinical settings: (1) children able to walk and scheduled for tonsillectomy/adenoidectomy; and (2) adults before four surgical procedures (i.e. laparoscopic cholecystectomy, thyroidectomy, lumbar discectomy or breast surgery). These procedures were chosen because they are performed routinely and do not require routine preoperative testing. We hypothesized that preoperative testing rates would decrease in France following the implementation of these guidelines.
Methods Data source We performed a retrospective study using two French nationwide datasets linked by an anonymous and unique patient identifier: (1) The French Hospital Discharge database (PMSI), which contains medical procedures for all patients admitted to public and private hospitals identified by their code according to the Classification Commune des Actes Me´dicaux [(CCAM), i.e. French medical classification for clinical procedures)] and discharge diagnoses coded according to the International Classification of Diseases, 10th revision (ICD-10 codes) and registered as principal, related or associated diagnoses; and (2) The
|
1187
French National Health Insurance information system (SNIIRAM), which collects all individualized healthcare claims reimbursed by French National Health Insurance covering more than 96% of the French population. This database records all reimbursements for patient care and comprises internal and external quality control processes. The database comprehensively records all outpatient prescriptions, services and procedures performed, together with their date of prescription and implementation. This database also contains patient data such as long-term diseases (LTDs) coded according to ICD-10, which are eligible for 100% health insurance coverage, prescribed medications identified according to Anatomical Therapeutic Classification (ATC) code and laboratory investigations identified from the French national laboratory test coding table (NABM: nomenclature des actes de biologie me´dicale). This linkage has been previously used to conduct large-scale epidemiological studies.6 7 The French Data Protection Authority (CNIL: Commission Nationale de l’informatique et des Liberte´s) has approved data collection from the SNIIRAM by the National Health Insurance (CNAMTS).8
Study population, variables of interest and outcomes Two patient populations were evaluated. Specific surgical procedures were identified according to their CCAM code. The index date was the date of the first hospitalization for the type of surgery considered. Group 1: all patients 2 yr and <18 yr of age scheduled for tonsillectomy and/or adenoidectomy between 2013 and 2015 were included. All prescriptions for haemostasis tests [prothrombin time (PT), activated partial thromboplastin time (aPTT) and ABO blood typing with testing for irregular agglutinins and Rh phenotype determination] prescribed or performed within 30 days before hospitalization for surgery were collected from the SNIIRAM databases. This 30 day period corresponds to the validity period for preoperative anaesthesiology consultations recommended by SFAR. Patients in whom comorbidities and medications deemed to be risk factors for bleeding or impaired haemostasis were identified by hospitalization discharge/LTD diagnoses and medications dispensed at least once before the index date were excluded (Supplementary data, Table S1). Re-operation for bleeding on the first postoperative day and hospital re-admission within 7 days were collected as outcome measures. Group 2: all patients 18 yr of age scheduled for laparoscopic cholecystectomy, thyroidectomy (excluding associated tracheal or laryngeal resections and extra-cervical approach), lumbar discectomy (excluding associated laminectomy or interbody fusion) or breast tumour surgery (excluding radical mastectomy, mastectomy with flap reconstruction or dermal autograft) between 2012 and 2015 were included. These four surgical procedures were selected because of a transfusion rate 1%9 regardless of patient medical conditions. All ABO blood typing, testing for irregular agglutinins and Rh phenotype determinations prescribed or performed within 30 days before hospitalization for surgery were collected from the SNIIRAM databases. Red blood cell transfusions during the first two postoperative days were collected as outcome measures. According to guidelines, these tests should not be performed in these two patient populations.5 The time period of this study was before/during implementation of guidelines for the four surgical procedures and only during implementation of guidelines for adenoidectomy and tonsillectomy.
1188
|
Beloeil et al.
Statistics Statistical analysis was performed using XLSTAT (Paris, France) V3 software. For each type of surgical procedure and each preoperative test, patients with preoperative testing were compared with patients without testing by v2 test to determine whether the preoperative testing decreased over the time period considered. The relative risks (RR) and confidence interval (95% CI) for the various outcomes according to the presence or the absence of preoperative testing were also determined.
Results The study population scheduled for tonsillectomy and adenoidectomy in 2013, 2014 and 2015 is presented in Table 1. When both procedures were performed at the same time, they were listed as tonsillectomy. The haemostasis test rate before adenoidectomy and tonsillectomy was high in a population in which these tests are not recommended by SFAR guidelines5 (Table 2). Haemostasis tests were performed in 49% of children before tonsillectomy and in 39% of children before adenoidectomy. This rate decreased significantly between 2013 and 2015. A similar pattern was observed for ABO blood typing, testing for irregular agglutinins and Rh phenotype determination. The preoperative testing rate was higher for tonsillectomy than for adenoidectomy, for which preoperative testing rates continued to decline between 2013 and 2015. The percentage of reduction of preoperative testing between 2013 and 2015 was between 1.8% and 4.2% regardless of the surgical procedure. The average annual cost of unnecessary preoperative testing over the 3 yr period was e1 758 000 (SD e97 000) and e895 000 (SD e71 000) for tonsillectomy and adenoidectomy, respectively. The reoperation rate for bleeding on the first postoperative day was very low during the study period, ranging from 0.12% to 0.31% for tonsillectomy and was even lower for adenoidectomy (0.01– 0.02%) (Table 3). The re-operation rate was higher in patients scheduled for tonsillectomy without preoperative testing. An additional search was performed in patients who were reoperated for bleeding on the first postoperative day following tonsillectomy according to the presence or absence of preoperative testing. During the study period considered (2013–5), none of these patients received vasopressin or vasopressin analogues
or blood coagulation factors defined by their ATC codes or presented related or principal diagnoses of coagulation defects or haemorrhagic conditions defined by their ICD-10 codes from the day of surgery or during the following year. The hospital readmission rate within 7 days was higher following tonsillectomy than adenoidectomy, but was statistically independent of preoperative testing (Table 3). A total of 1 114 082 patients underwent one of the four selected surgical procedures in France between 2012 and 2015 (Table 4). The ABO blood typing, testing for irregular agglutinins and Rh phenotype determinations rates ranged from 28% to 52% of patients scheduled for surgery (Table 5). A higher preoperative testing rate was observed for lumbar discectomy than for the other three procedures and the lowest rate was observed for breast tumour surgery. The preoperative testing rate decreased from 2012 to 2014, but increased again in 2015 reaching values higher than those observed in 2012 for lumbar discectomy (þ1.9%, þ4.1% and þ1.9% for ABO blood typing, testing for irregular agglutinins and Rh phenotype determination, respectively). The annual cost of this unnecessary preoperative testing over the 4 yr period was e2 330 000 (e143 000); e852 932 (e68 000), e724 098 (e26 000) and e1 400 000 (e107 000) for laparoscopic cholecystectomy, thyroidectomy, lumbar discectomy or breast tumour surgery, respectively. The cumulative cost for the four surgical procedures over the 4 yr period was e5 310 000 (e325 000). Red blood cell transfusion rate during the first 48 postoperative hours was extremely low (<1%) regardless of surgical procedure (Table 6). Red blood cell transfusion rate after cholecystectomy was significantly higher in patients in whom preoperative testing was not performed.
Discussion This study confirmed a high rate of unnecessary preoperative testing based on more than 100 000 patients undergoing adenoidectomy or tonsillectomy and more than 270 000 patients undergoing one of the four selected surgical procedures. These results are in accordance with the literature reporting overprescription of preoperative testing.10 This is the first study on this subject based on 4 yr data from two national databases, constituting a large cohort of patients following publication of
Table 1 Study population for tonsillectomy and adenoidectomy in 2013, 2014 and 2015. The sum of the excluded patients related to ATC codes for medications and ICD-10 codes for long-term diseases and for comorbidities is greater than the number of patients actually excluded because the same patient sometimes presented several reasons for exclusion. ATC, Anatomical Therapeutic Classification; ICD10, International Classification of Diseases, 10th revision
Tonsillectomy Patients included, n (%) Patients excluded, n (%) Exclusion related to ATC codes for medications, n Exclusion related to ICD-10 codes for long-term diseases, n Exclusion related to ICD-10 codes for comorbidities, n Adenoidectomy Patients included, n (%) Patients excluded, n (%) Exclusion related to ATC codes for medications, n Exclusion related to ICD-10 codes for long-term diseases, n Exclusion related to ICD-10 codes for comorbidities, n
2013
2014
2015
60 816 (93.7) 4096 (6.3) 3546 160 621
62 729 (94.3) 3796 (5.7) 3206 168 619
56 656 (94.4) 3385 (5.6) 2830 175 637
45 826 (93.9) 2971 (6.1) 2741 72 213
44 639 (94.2) 2724 (5.8) 2523 52 209
42 325 (94.6) 2409 (5.4) 2190 53 198
Overuse of preoperative laboratory coagulation testing and ABO blood typing
|
1189
Table 2 Comparisons between patients with preoperative testing and patients without preoperative testing over the time period considered were analysed with v2 tests. Rh, rhesus; PT, prothrombin time; aPTT, activated partial thromboplastin time Tonsillectomy PT n testing rate (%) aPTT n testing rate (%) ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%) Adenoidectomy PT n testing rate (%) aPTT n testing rate (%) ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%)
2013
2014
2015
P-value
29 318 48.2
28 147 44.9
25 137 44.4
<0.0001
33 687 53.2
32 047 51.1
29 110 51.4
<0.0001
30 681 50.4
29 312 46.7
26 761 47.2
<0.0001
28 904 47.5
27 977 44.6
25 674 45.2
<0.0001
30 359 49.9
29 173 46.5
26 476 46.7
<0.0001
17 154 37.4
15 876 35.6
14 378 34.0
<0.0001
20 140 43.9
18 394 41.2
16 823 39.7
<0.0001
16 310 35.6
14 813 33.2
13 579 32.1
<0.0001
14 710 32.1
13 607 30.5
12 592 29.8
<0.0001
16 170 35.3
14 753 33.0
13 474 31.8
<0.0001
Table 3 Re-operation rate for bleeding at D0 and hospital re-admission within 7 days following tonsillectomy and adenoidectomy in 2013, 2014 and 2015. RR, relative risk; 95% CI, 95% confidence interval; D0, day 0 Tonsillectomy Re-operation for bleeding at D0 Preoperative testing – (%) Preoperative testing þ (%) RR (95% CI) Hospital re-admission within 7 days Preoperative testing – (%) Preoperative testing þ (%) RR (95% CI) Adenoidectomy Re-operation for bleeding at D0 Preoperative testing – (%) Preoperative testing þ (%) RR (95% CI) Hospital re-admission within 7 days Preoperative testing – (%) Preoperative testing þ (%) RR (95% CI)
2013
2014
2015
0.28 0.15 0.56 (0.39–0.80)
0.31 0.13 0.42 (0.28–0.61)
0.24 0.12 0.50 (0.33–0.76)
1.72 1.85 1.07 (0.95–1.21)
1.96 1.78 0.91 (0.81–1.02)
2.00 2.07 1.03 (0.92–1.16)
0.01 0.01 0.84 (0.08–9.31)
0.02 0.01 0.26 (0.03–2.13)
0.01 0.01 0.66 (0.06–6.30)
0.22 0.18 0.82 (0.53–1.26)
0.16 0.19 1.20 (0.76–1.90)
0.27 0.20 1.39 (0.92–2.10)
1190
|
Beloeil et al.
Table 4 Study population for the four selected surgical procedures in 2012, 2013, 2014 and 2015. See text for exclusions
Laparoscopic cholecystectomy Partial or total thyroidectomy Lumbar discectomy Breast tumour surgery Total
2012
2013
2014
2015
112 935
113 662
115 433
117 585
47 622
46 033
45 051
42 817
33 605 86 738 280 900
32 442 85 637 277 774
31 710 85 959 278 153
30 601 86 252 277 255
Table 5 Preoperative ABO blood typing, testing for irregular agglutinins and Rh phenotype determination for laparoscopic cholecystectomy, partial or total thyroidectomy, lumbar discectomy and breast tumour surgery in 2012, 2013, 2014 and 2015. Rh, Rhesus. Comparisons between patients with preoperative testing and patients without preoperative testing over the time period considered were analysed with v2 tests 2012
national guidelines and an intensive implementation strategy; therefore, providing major support to the results. The statistically significant decrease in preoperative testing rates during the study period was not clinically meaningful. Outcomes were not modified in patients without preoperative testing except for the re-operation rate in patients scheduled for tonsillectomy, which was higher in these patients. In accordance with the literature, perioperative events were not different in patients with or without preoperative testing.11 In minor procedures such as cataract surgery, morbidity and mortality are not reduced by routine preoperative tests.11 The incidences of re-operation for bleeding after adenoidectomy, re-admission within 7 days after adenoidectomy and tonsillectomy, and transfusion within 48 h after the four selected surgical procedures were very low and not different between patients with or without preoperative tests. Standard laboratory coagulation tests and ABO blood typing are still over-prescribed,4 10 more than any other tests for a number of reasons, including defensive medicine to prevent complaints or criticism. Personal experience also constitutes an important factor.12 13 Physicians with personal histories of perioperative bleeding continue to order preoperative testing despite guidelines. However, a blood test is not necessarily beneficial to the patient because of the risk of false positives for statistical, pre-analytical and analytical reasons, thereby triggering further follow-up testing and inappropriate patient management. Further investigation can delay surgery and increase the number of unnecessary tests for the patient. Systematic preoperative standard laboratory coagulation tests fail to predict the risk of bleeding,14 15 16 17 An abnormal test could be associated with a non-haemorrhagic disease. For example, a deficit in factor XII increases aPTT with no clinical consequences. In the general population, aPTT results follow a normal distribution: 2.5% of the patients have an aPTT higher than two standard deviations from the mean with no increased bleeding risk. In asymptomatic patients, the prevalence of congenital clotting factor deficiencies associated with bleeding risk is approximately 1 in 40 000. If routine aPTT could detect all asymptomatic patients with an increased risk (1/40 000), the probability of a randomly increased aPTT (1/40) would be 1000 times greater than the probability of detecting increased aPTT associated with haemorrhagic disease. The predictive value of aPTT is only 0.1% in a symptomatic male patient. In our study, re-operations for bleeding after tonsillectomy were more frequent in patients without preoperative tests. Considering the incidence of asymptomatic patients with an increased bleeding risk because of congenital deficiencies of coagulation factors (1/40 000),18 this situation would concern an average of one
Laparoscopic cholecystectomy ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%) Partial or total thyroidectomy ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%) Lumbar discectomy ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%) Breast tumour surgery ABO blood typing n testing rate (%) Testing for irregular agglutinins n testing rate (%) Rh phenotype determination n testing rate (%)
2013
2014
2015
P-value
44 182 42 997 40 516 43 054 <0.0001 39.1 37.8 35.1 36.6
52 951 52 218 50 315 54 300 <0.0001 46.9 45.9 43.6 46.2
43 895 42 751 40 376 42 610 <0.0001 38.9 37.6 35.2 36.2
17 511 16 728 14 786 14 824 <0.0001 36.8 36.3 32.8 34.6
20 104 19 547 17 668 17 842 <0.0001 42.2 42.5 39.2 41.7
17 271 16 580 14 733 14 663 <0.0001 36.3 36.0 32.7 34.2
14 387 14 141 13 335 13 674 <0.0001 42.8 43.6 42.1 44.7
16 036 15 951 15 166 15 839 <0.0001 47.7 49.2 47.8 51.8
14 320 14 102 13 330 13 623 <0.0001 42.6 43.5 42.0 44.5
29 249 27 411 24 003 25 225 <0.0001 33.7 32.0 27.9 29.2
33 364 31 883 29 891 31 481 <0.0001 38.5 37.2 34.8 36.5
28 420 26 632 23 891 24 966 <0.0001 32.8 31.1 27.8 28.9
patient each year among those with preoperative testing and those without preoperative testing in our study. This does not explain the difference in the re-operation rate for bleeding between patients with and without preoperative testing (45 vs
Overuse of preoperative laboratory coagulation testing and ABO blood typing
|
1191
Table 6 Red blood cells transfusion rate within 48 h after laparoscopic cholecystectomy, partial or total thyroidectomy, lumbar discectomy and breast tumour surgery in 2012, 2013, 2014 and 2015. RR, relative risk; CI 95%, 95% confidence interval
Laparoscopic cholecystectomy Transfusion rate at within 48 hours after Preoperative testing – (%) Preoperative testing þ (%) RR (CI 95%) Partial or total thyroidectomy Transfusion rate at within 48 hours after Preoperative testing – (%) Preoperative testing þ (%) RR (CI 95%) Lumbar discectomy Transfusion rate at within 48 hours after Preoperative testing – (%) Preoperative testing þ (%) RR (CI 95%) Breast tumour surgery Transfusion rate at within 48 hours after Preoperative testing – (%) Preoperative testing þ (%) RR (CI 95%)
2012
2013
2014
2015
0.28 0.16 0.57 (0.43–0.75)
0.31 0.14 0.46 (0.35–0.62)
0.25 0.12 0.46 (0.33–0.64)
0.27 0.14 0.53 (0.40–0.71)
0.13 0.08 0.65 (0.35–1.22)
0.10 0.03 0.32 (0.12–0.81)
0.07 0.04 0.55 (0.21–1.46)
0.06 0.03 0.52 (0.17–1.54)
0.09 0.02 0.25 (0.07–0.86)
0.06 0.06 1.07 (0.43–2.65)
0.04 0.06 1.54 (0.58–4.11)
0.04 0.05 1.05 (0.36–3.03)
0.15 0.14 0.93 (0.63–1.37)
0.14 0.16 1.13 (0.78–1.66)
0.15 0.16 1.09 (0.75–1.59)
0.16 0.15 0.92 (0.62–1.34)
88 in 2013, 36 vs 108 in 2014 and 30 vs 77 in 2015, respectively). Moreover, from the day of surgery and during the following year, none of the re-operated patients received treatment for or was diagnosed with a coagulation defect. This difference could therefore be explained by a surgical haemostasis problem rather than an undetected haemostatic disease. Another risk associated with systematic preoperative haemostatic tests is falsenegative results: a normal aPTT does not guarantee the absence of coagulopathy. The sensitivity of aPTT is highly dependent on the severity of the factor deficiency. The sensitivity of aPTT to detect haemophilia is 100% in the presence of a severe deficiency but only 90% in the case of moderate deficiency (16–30%). It can be as low as 48% in von Willebrand disease. Factor XIII deficiency is not detected by any routine tests. Studies have shown that haemostatic tests can be useful when preceded by bleeding history questionnaire. However, no specific questionnaire for the assessment of bleeding risk before surgery has been validated. SFAR guidelines recommend that the following items be included in the questionnaire: tendency for prolonged/ unusual bleeding (epistaxis, small cuts), tendency to develop ecchymoses/bruising/haematomas, prolonged bleeding after tooth extraction, major bleeding after surgery (especially circumcision or tonsillectomy), family history, and in women, menorrhagia or postpartum haemorrhage. When a thorough physical examination and interview are performed, the addition of blood tests does not modify patient management15 and is therefore of little value. Our results show that the re-operation rate for bleeding after tonsillectomy was lower than that reported in the literature, in which post-tonsillectomy bleeding rates are 3–5%.19 Blood typing before surgery is appropriate in patients likely to require blood transfusion. In agreement with European guidelines,20 SFAR does not recommend blood typing when there is a low risk of transfusion (i.e. <1%).5 Our results showed that blood typing was performed in more than 50% of patients before surgical procedures associated with very low transfusion
rates. In a validated algorithm to help physicians avoid unnecessary type and screen orders before surgery, Dexter and colleagues21 showed that blood typing is not necessary when the transfusion rate is lower than 5%. However, we observed a high rate of unjustified tests despite our choice of surgical procedures associated with a very low risk of transfusion. Unjustified preoperative testing represents a very high cost for society. Shein22 estimated the cost of these tests to be $150 million per year in the USA in 1996. In our study, the annual cost for haemostasis tests and ABO blood typing for just six surgical procedures was over e7 million for 370 000 patients out of the 10 million patients undergoing surgical procedures. The potential savings are therefore considerable. Our study highlights the inconsistency between guidelines and practice10 23 as well as the lack of effectiveness of the implementation strategy developed following publication of guidelines. Despite quality processes involving continuing medical evaluation, test rates have not decreased.24 25 Audits comprising feedback are also widely used to improve practice and compliance with guidelines. However, their effectiveness remains controversial, as Ivers and colleagues26 reported a significant increase in healthcare professional compliance with desired practice but insignificant improvement in patient outcome. Tailored intervention strategies have been suggested.27 The latest meta-analysis on the subject included 32 studies, and reported a modest benefit to improve professional practices and healthcare outcomes.28 Computerized systems could also help improve performance,29 but regardless of the technique, no significant improvement has yet been observed.24 30 This study has several of limitations. Data analysis was retrospective and data were derived from administrative databases with inherent limitations concerning data collection and coding. The main strength of this study concerns the use of the SNIIRAM database with data for more than 96% of the French population in 2013. The impact of bias related to population groups insured by other schemes as a result of their occupation
1192
|
Beloeil et al.
or sector of activity is probably low. However, the algorithms used in this study cannot take into account preoperative tests performed on inpatients in the public healthcare system. This study therefore underestimates preoperative testing rates, although these inpatient tests probably represent only a small proportion of all tests performed. Furthermore, this database records reimbursements for patient care and comprises internal and external quality control processes ensuring exhaustiveness of the database. However, exclusions were performed from healthcare utilization data and cannot take into account subjects with diseases not managed by a hospital stay or a specific LTD or medication prescription. Outcome misclassification could constitute another limitation because the external validity of the various outcome measures used in this study has not been previously assessed in the French PMSI database. However, principal related or associated hospital discharge diagnoses were used to define outcomes thereby limiting the risk of underestimation.
2. 3.
4.
5. 6.
7.
Conclusions Standard laboratory coagulation tests and ABO blood typing before surgery and anaesthesia are still performed routinely even in clinical situations in which they are not medically justified. In this context, these tests are unnecessary and have no clinical impact. Compliance of physicians with guidelines and the effectiveness of the implementation strategy developed by the French Society of Anaesthesiology remain very limited, raising the question of the capacity of physicians to change their behaviour in the absence of incentives.
8.
9.
10.
Authors’ contributions Helped conduct the study, analysing the data and writing the manuscript: H.B. and S.M. Helped conduct the study: D.R. Helped analysing the data and conducted the statistical analysis: N.D. All authors have seen the original study data, reviewed the analysis of the data and approved the final manuscript.
11.
12.
13.
Supplementary material Supplementary material is available at British Journal of Anaesthesia online.
14.
Declaration of interest This work was supported by the Socie´te´ Franc¸aise d’Anesthe´sie et de Re´animation (SFAR). National guidelines concerning Preoperative testing were published by the SFAR. H.B. and S.M. were members of the working group which elaborated these guidelines and involved as experts in the local and national educational programmes sponsored by the SFAR.
Funding
15.
16.
17.
Funding was solely institutional.
References 1.
Blery C, Charpak Y, Szatan M, et al. Evaluation of a protocol for selective ordering of preoperative tests. Lancet 1986; 1: 139–41
18. 19.
Roizen MF. The compelling rationale for less preoperative testing. Can J Anaesth 1988; 35: 214–8 Benarroch-Gampel J, Sheffield KM, Duncan CB, et al. Preoperative laboratory testing in patients undergoing elective, low-risk ambulatory surgery. Ann Surg 2012; 256: 518–28 Bernard R, Benhamou D, Beloeil H. [Routine preoperative testing: Impact of implementation of local recommendations in a teaching hospital]. Ann Fr Anesth Reanim 2010; 29: 868–73 Molliex S, Pierre S, Blery C, Marret E, Beloeil H. [Routine preinterventional tests]. Ann Fr Anesth Reanim 2012; 31: 752–63 Maura G, Blotiere PO, Bouillon K, et al. Comparison of the short-term risk of bleeding and arterial thromboembolic events in nonvalvular atrial fibrillation patients newly treated with dabigatran or rivaroxaban versus vitamin K antagonists: a French nationwide propensity-matched cohort study. Circulation 2015; 132: 1252–60 Fagot JP, Blotiere PO, Ricordeau P, Weill A, Alla F, Allemand H. Does insulin glargine increase the risk of cancer compared with other basal insulins? A French nationwide cohort study based on national administrative databases. Diabetes Care 2013; 36: 294–301 Roussel R, Chaignot C, Weill A, et al. Use of fibrates monotherapy in people with diabetes and high cardiovascular risk in primary care: a French Nationwide Cohort Study Based on National Administrative Databases. PloS One 2015; 10: e0137733 Puel C, Ducharne T, Mialon A, et al. [Surgical risk of transfusion in a French Universitary Hospital]. Ann Fr Anesth Reanim 2012; 31: 132–40 Bryson GL, Wyand A, Bragg PR. Preoperative testing is inconsistent with published guidelines and rarely changes management. Can J Anaesth 2006; 53: 236–41 Schein OD, Katz J, Bass EB, et al. The value of routine preoperative medical testing before cataract surgery. Study of Medical Testing for Cataract Surgery. N Engl J Med 2000; 342: 168–75 Lippi G, Favaloro EJ, Franchini M. Dangers in the practice of defensive medicine in hemostasis testing for investigation of bleeding or thrombosis: part I–routine coagulation testing. Semin Thromb Hemost 2014; 40: 812–24 Toker A, Shvarts S, Perry ZH, Doron Y, Reuveni H. Clinical guidelines, defensive medicine, and the physician between the two. Am J Otolaryngol 2004; 25: 245–50 Asaf T, Reuveni H, Yermiahu T, et al. The need for routine pre-operative coagulation screening tests (prothrombin time PT/partial thromboplastin time PTT) for healthy children undergoing elective tonsillectomy and/or adenoidectomy. Int J Pediatr Otorhinolaryngol 2001; 61: 217–22 Bonhomme F, Ajzenberg N, Schved JF, et al. Pre-interventional haemostatic assessment: Guidelines from the French Society of Anaesthesia and Intensive Care. Eur J Anaesthesiol 2013; 30: 142–62 Levy JH, Szlam F, Wolberg AS, Winkler A. Clinical use of the activated partial thromboplastin time and prothrombin time for screening: a review of the literature and current guidelines for testing. Clin Lab Med 2014; 34: 453–77 Baron DM, Metnitz PG, Fellinger T, Metnitz B, Rhodes A, Kozek-Langenecker SA. Evaluation of clinical practice in perioperative patient blood management. Br J Anaesth 2016; 117: 610–6 Janvier G. [Preoperative evaluation of hemorrhagic risk]. Ann Fr Anesth Reanim 1998; 17(Suppl 1): 2s–5s Mitchell RM, Parikh SR. Hemostasis in tonsillectomy. Otolaryngol Clin North Am 2016; 49: 615–26
Overuse of preoperative laboratory coagulation testing and ABO blood typing
20. De Hert S, Imberger G, Carlisle J, et al. Preoperative evaluation of the adult patient undergoing non-cardiac surgery: guidelines from the European Society of Anaesthesiology. Eur J Anaesthesiol 2011; 28: 684–722 21. Dexter F, Ledolter J, Davis E, Witkowski TA, Herman JH, Epstein RH. Systematic criteria for type and screen based on procedure’s probability of erythrocyte transfusion. Anesthesiology 2012; 116: 768–78 22. Schein OD. Assessing what we do. The example of preoperative medical testing. Archiv Ophthalmol 1996; 114: 1129–31 23. van Gelder FE, de Graaff JC, van Wolfswinkel L, van Klei WA. Preoperative testing in noncardiac surgery patients: a survey amongst European anaesthesiologists. Eur J Anaesthesiol 2012; 29: 465–70 24. Sigmund AE, Stevens ER, Blitz JD, Ladapo JA. Use of preoperative testing and physicians’ response to professional society guidance. JAMA Intern Med 2015; 175: 1352–9 25. Forsetlund L, Bjorndal A, Rashidian A, et al. Continuing education meetings and workshops: effects on professional
26.
27.
28.
29.
30.
|
1193
practice and health care outcomes. Cochrane Database Syst Rev 2009; CD003030 Ivers N, Jamtvedt G, Flottorp S, et al. Audit and feedback: effects on professional practice and healthcare outcomes. Cochrane Database Syst Rev 2012; CD000259 Leung BC, Nazeer S, Smith M, McRae D. Reducing unnecessary preoperative testing in elective ENT surgery: clinical and financial implications. J Perioper Pract 2015; 25: 225–9 Baker R, Camosso-Stefinovic J, Gillies C, et al. Tailored interventions to address determinants of practice. Cochrane Database Syst Rev 2015; CD005470 Garg AX, Adhikari NK, McDonald H, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA 2005; 293: 1223–38 Flamm M, Fritsch G, Seer J, Panisch S, Sonnichsen AC. Nonadherence to guidelines for preoperative testing in a secondary care hospital in Austria: the economic impact of unnecessary and double testing. Eur J Anaesthesiol 2011; 28: 867–73 Handling editor: Hugh C Hemmings Jr