Secondary Logo

Journal Logo

Preoperative assessment

American Society of Anesthesiologists’ Physical Status system: a multicentre Francophone study to analyse reasons for classification disagreement

Cuvillon, Philippe; Nouvellon, Emmanuel; Marret, Emmanuel; Albaladejo, Pierre; Fortier, Louis-Philippe; Fabbro-Perray, Pascale; Malinovsky, Jean-Marc; Ripart, Jacques

Author Information
European Journal of Anaesthesiology: October 2011 - Volume 28 - Issue 10 - p 742-747
doi: 10.1097/EJA.0b013e328348fc9d



The American Society of Anesthesiologists’ (ASA) Physical Status is a grading system designed originally (1940–1941) to standardise physical status categories for statistical studies and for hospital records so that uniform interpretation would be possible.1–3 This classification system is used worldwide in clinical practice, in textbooks, in clinical research and for economic evaluation of anaesthesia services. Although ASA grading is a simple and easy classification (only six categories in the latest version; Table 1), it is a well known source of disagreement between assessors.4,5 Several clinical studies designed to explain why there is so much discrepancy among assessors have suggested the following.

  1. Country (cultural) and language differences influence the judgement of assessors. Studies conducted in North America,3,6 Europe (Great Britain, Finland),4,5 Australia and South Africa7 have reported considerable variation in the ASA classification allocation, except for specialists who have been trained in multiple centres worldwide.7 Moreover, the experience of practitioners (junior vs. senior), employment in teaching hospitals (vs. non-teaching hospitals) and pattern of practice (private vs. public) influence the accuracy of ASA grading: the Kappa score (Table 1) was 0.21–0.4.5,7–9
  2. Although classification disagreements have been reported in all categories (ASA 1–4), ASA grades 2 and 3 have been involved in the most significant allocation disagreement. This was demonstrated (retrospectively) with questionnaires containing five to 10 scenarios mailed to anaesthesiologists who were asked to assign each patient to an ASA grade.6–8 The extremities of age, anaemia, myocardial disease, obesity3 or obstetric scenarios6 were associated with significant disagreement, probably because these factors, including the associated physiological disturbances, were not included in the ASA scoring definitions.
  3. In daily practice, it is also not clear what the ASA grade of a patient should be if two or more coexisting diseases, all with physiological disturbances, are present, particularly if each has a different severity. In addition, the ASA grading system has not been evaluated well in patients with some conditions (e.g. sleep apnoea, metabolic disorders). Most case scenario studies have included active members of different anaesthetic associations who may not represent the routine anaesthetic population and situation.
Table 1
Table 1:
Description of the American Society of Anesthesiologists physical status classification and guidelines for the interpretation of the Kappa coefficient

Because few blinded evaluations have been conducted in routine practice or in specialised clinical situations, this prospective cross-over study was designed to determine the degree of agreement of ASA gradings performed routinely by accredited specialist anaesthesiologists (primary end-point) and to characterise whether specific categories of disease or physiological disturbance influenced ASA grading disagreements.


Following approval from the Canadian institutional ethics committee (Jocelyne Tessier MD, May 2004, n° 4052, Maisonneuve-Rosemont, Montréal, Canada), this study was conducted in nine French and Canadian institutions (public or private), performing more than 15 000 anaesthetic procedures per year, using a similar ASA grade classification as that shown on the website physicalstatus.htm (Table 1). In each institution, a local investigator was selected before the study period. Other than that investigator, no other anaesthesiologist from each institution was aware of data gathering for the duration of the study.

Inclusion criteria

During a 2-week period, all anaesthesia records of patients scheduled for surgery with a staff anaesthesiologist in attendance were included. Emergency procedures and anaesthesia records completed by residents or by the local investigator were excluded.

Management of the records

At the end of the 2-week study period, the local investigator collected all of the anaesthesia records from his/her centre. Exclusion criteria were checked at this time. If the anaesthesia record was suitable for inclusion, it was duplicated and names of the patient and the anaesthesiologist were concealed. All duplicate records were mailed to the independent statistics division (Fig. 1; Institut Universitaire de Recherche Clinique and Département Informatique Médicale, Nîmes, France).

Fig. 1
Fig. 1:
no caption available.

At the statistics centre, all records were numbered. ASA grading assessments (‘initial ASA’) from all the records were stored and concealed. None of the anaesthesia records was excluded, even if the ASA grade was missing. At this step, each anaesthesia record was randomly returned to one of the nine institutions for reassessment (using a random number generator). Anaesthesia records were arbitrarily pooled in samples of 30. All samples were mailed to the institution indicated by the randomisation procedure.

In the centres, for each blinded sample of 30 records, an accredited specialist anaesthetist with more than 5 years of experience was asked by the local investigator to assess the ASA grade for each patient (‘final ASA’). If the local investigator noted that the final ASA grade was different from the initial ASA grade, the anaesthesiologist was asked to indicate the disease present on the anaesthesia record justifying this status. All the anaesthesia records were then mailed back to the statistics division for analysis.

Studied parameters

The primary end-point was the evaluation of agreement between the two ASA grades for each patient. Data collected were the initial ASA and final ASA grades, demographic data of the patient (age, weight, height), type of surgery, disease justifying the final ASA grade if the two grades were different and the type of centre (private vs. public).

Statistical analysis

Data are presented as number (%) or mean ± SD (standard deviation) as indicated.

The level of agreement between the two measurements of the ASA grade was calculated by using the weighted Kappa coefficient (95% confidence limits; Table 1). First, it was calculated for the entire study population. Then it was calculated in different strata of records to search for effects of the centre and type of anaesthesia provider (private or public). An overall Kappa coefficient and a test for equal Kappa were calculated to assess overall agreement. McNemar's test of symmetry was used for non-parametric data (type of centre, pattern of practice, disease). A P value less than 0.05 was considered statistically significant. Statistical analyses were performed using SAS/STAT software, version 8.1 (SAS Institute, Cary, North Carolina, USA).


Over the study period, 1554 anaesthesia records were collected in nine centres (six public and three private). An ASA grade was found in 1384 anaesthesia records (90%). For the second evaluation, no file was excluded (1554 records were sent from the statistics centre for a second evaluation), but a reassessment was impossible on 37 anaesthesia records because they were unreadable or poorly completed. Thus, ‘first’ and ‘final’ ASA grades were usable for analysis in 1357 anaesthesia records (197 anaesthesia records were excluded from analysis because the initial ASA grade, the final ASA grade or both were missing; Fig. 1). Characteristics of the usable anaesthesia records and the excluded records are shown in Table 2. After the first evaluation, the distribution of ASA grades was as follows: ASA 1, 571; ASA 2, 591; ASA 3, 177; and ASA 4, 18 (Table 3). After the second evaluation, the distribution was as follows: ASA 1, 583; ASA 2, 520; ASA 3, 223; and ASA 4, 31 (Table 3). No ASA 5 (or, obviously, ASA 6) grades were recorded. After the second evaluation, 19% of patients had a higher physical status grading and 15% had a lower grading (P = 0.035; Table 4). In the second evaluation, 1.25% of patients assessed initially as ASA 1 were graded as ASA 3 (P > 0.05). Conversely, no patient initially assessed as ASA 3 or ASA 4 was reassessed as ASA 1.

Table 2
Table 2:
Details of the patients in the included and excluded records (total patients: 1554)
Table 3
Table 3:
Contingency table of the 1357 anaesthesia records with initial and final American Society of Anesthesiologists grades
Table 4
Table 4:
Description of the subgroups of patients in whom initial and final American Society of Anesthesiologists grades were present and exceeded American Society of Anesthesiologists 1

The degree of agreement between the two grades evaluated by the weighted Kappa index was 0.53 (0.49–0.56; Table 3). There was no effect on agreement of the centre of collection (P > 0.25) or the centre of reassessment (P > 0.58). The degree of agreement was not influenced by the type of anaesthesia provider: private [Kappa index 0.52 (0.46–0.58)] or public [0.53 (0.49–0.57)] (P > 0.76).

The nature of diseases justifying a status other than ASA 1 during the final ASA evaluation and ASA grades of this subgroup of patients are shown in Table 4. Multiple coexisting diseases were identified in 219 (15%) patients. Patients with multiple diseases, obesity, allergy, sleep apnoea, obstructive lung disease, renal insufficiency or hypertension gave rise to the greatest disagreement (Table 4). There was good agreement between the initial and final ASA grades in patients who had only diabetes mellitus, or cardiac (congestive or coronary) disease, or stroke, without any coexisting disease [Kappa 0.61–0.65 (good agreement)]. Kappa was less than 0.4 when associated with coexisting disease (Table 4).


This is the first cross-over blinded study to evaluate the consistency of ASA grading in a large population. In this current practice multicentre study, we confirmed that the degree of agreement between anaesthesiologists for ASA grading is moderate (overall Kappa 0.53). Moreover, we confirmed and identified specific disorders associated with frequent disagreement about ASA grading: obesity, allergy, chronic arterial hypertension, obstructive sleep apnoea, chronic obstructive lung disease and multiple coexisting diseases.

The calculated agreement in this study is stronger than those previously reported (Kappa index ranging from 0.33 to 0.47).1,7,8 The methodology used in this study is very different from that used in previous studies.1,4–8 In these studies, the ASA classification system was evaluated on 10 clinical case scenarios. In some studies, the case scenarios were theoretical, or focused on specific patient groups, such as paediatric patients. The scenarios were mailed to a large number of anaesthesiologists for ASA grading. The agreement was calculated among all gradings collected (interindividual agreement) or between gradings collected and those of the authors. Conversely, we chose to include a large number of anaesthesia records. ASA grades were re-assessed by a small number of anaesthesiologists and agreement was calculated between the initial and the final grades. We chose this methodology because 10 scenario ‘patients’ are not representative of a general population, and it may be hazardous to extrapolate the coherence of ASA grade assessment from a small sample. Accuracy of ASA grading was evaluated by measuring the level of agreement between the initial and final ASA grades. No expert was assigned in this study, because we assumed that all accredited specialist anaesthesiologists are sufficiently expert to undertake ASA grading. Similarly, in case scenarios, Haynes and Lawler4 did not propose a correct answer for each case because of the difficulty in achieving a complete consensus for each case. Furthermore, the anaesthesiologists were not aware of the study when they made the initial ASA grade assessment (‘current practice measurement’). This methodology may emphasise a potential disagreement between the two measurements. We observed that after the second evaluation, 19% of patients had a greater ASA grade; 1.25% of patients evaluated initially as ASA 1 were given grades of ASA 3 or 4 at the second assessment, and 70 patients evaluated as ASA 3 were graded ASA 4 after the second evaluation. Underestimation of the ASA grade may lead to erroneous anaesthesia management or the use of an inappropriate clinical pathway.9 The disagreement about ASA grades may be due to the low specificity of the classification system, particularly for recent or ‘new’ disorders such as obesity or sleep apnoea (Table 4). Recently, National Surgical Quality Program (NSQIP) risk factors (59 variables) were used to predict the ASA grade for 5878 patients.2 In that study, the NSQIP risk factor predictions of ASA grade agreed with the anaesthesiologists’ assessment in 67.6% of the cases, were within one grade in 99.1% of cases and disagreed by two or more grades in only 0.9% of cases. In paediatric practice, an objective system-based approach (Neurological, Airway, Respiratory, Cardiovascular and Other: NARCO) correlated with perioperative risk more closely than the ASA Physical Status classification system.10 Multiple subclass assessment systems such as NSQIP and NARCO provide better prediction of outcome (and, therefore, risk), but may be difficult to apply in routine use because of their complexity.10

In this study, a number of diseases were associated with greater variability of ASA grades (Table 4). For example, in four initial records, patients who presented with diabetes mellitus, treated with large doses of insulin, were graded ASA 3, but were graded as ASA 4 in the second evaluation. In three cases, patients with asymptomatic obesity (BMI 35–40 kg m−2) were considered initially to be ASA 2 or 3, but were assessed as ASA 1 at the second evaluation. These differences were probably related to the study design. The initial evaluation was performed after direct assessment of the patient, but the second evaluation was entirely a paper analysis. This emphasises that the ASA classification system is entirely an individual subjective interpretation.9 Haynes and Lawler4 believe that such differences may also be related to the nature of the surgery; patients undergoing a minor procedure (with few anticipated surgical complications) are often given a lower ASA grade than expected, even if suffering from a serious disease. In these cases, the ASA grading system is being used to quantify surgical risk, not physical status.11

Our data showed that patients with coexisting diseases correlated significantly with ASA grade discordance (Table 4). This demonstrates that complex diseases (e.g. sleep apnoea, obesity and diabetes mellitus) result in varying individual subjective interpretations.9 To improve the ASA Physical Status classification system, some authors have proposed modified grading systems which might improve prediction of outcome (and risk).12 For example, Barbeito et al.6 proposed the addition of a modifier for pregnancy (G for gravid). Other situations might also take into account age or genetic variations of various diseases, and be included in large prospective studies for analysis.

Several biases may have been caused by the design of this study. First, during the initial ASA grade evaluation, anaesthesiologists were not aware of the study and many data were missing on anaesthesia records, particularly regarding weight and height. For 240 patients (17% of the population), it was impossible to detect potential obesity or malnutrition at the second evaluation. Furthermore, ASA grades were missing on 170 anaesthesia records, more frequently in public centres (13.3%) than in private centres (0.7%), probably because, in France, the ASA grade influences financial reimbursement. Second, the final assessment was not performed by the attending practitioner, but by accredited specialist anaesthesiologists in other centres. These evaluations were based on blinded files, and important clinical factors apparent to the anaesthesiologist who made the initial assessment may have been overlooked. To reduce this bias, we used three methods that were not performed in previous studies:3–6 multicentre evaluations, Kappa scoring using a weighted coefficient and the use only of experienced practitioners for the second evaluation. Arbitrarily, 5 years of experience was used to select these practitioners. This could introduce another bias but some experience was necessary to analyse specific diseases that caused disagreement. Although they had 5 years of experience, these accredited specialist anaesthesiologists were not ‘ASA experts’ as in other retrospective studies3–6 and the second evaluation was performed using blinded records.

In conclusion, this study demonstrated the poor consistency of application of the ASA Physical Status classification system on a large population and identified several diseases which were associated with frequent disagreement. In the future, a graded score with modifiers should reduce the subjectivity and disagreement associated with the ASA grading system, but without producing excessive complexity.


The study received financial support from Association pour la Recherche en Anesthésie, Nîmes, France.


1. Jacqueline R, Malviya S, Burke C, Reynolds P. An assessment of interrater reliability of the ASA physical status classification in pediatric surgical patients. Paediatr Anaesth 2006; 16:928–931.
2. Davenport DL, Bowe EA, Henderson WG, et al. National Surgical Quality Improvement Program (NSQIP) risk factors can be used to validate American Society of Anesthesiologists Physical Status Classification (ASA PS) levels. Ann Surg 2006; 243:636–641.
3. Owens WD, Felts JA, Spitznagel EL Jr. ASA physical status classifications: a study of consistency of ratings. Anesthesiology 1978; 49:239–243.
4. Haynes SR, Lawler PG. An assessment of the consistency of ASA physical status classification allocation. Anaesthesia 1995; 50:195–199.
5. Ranta S, Hynynen M, Tammisto T. A survey of the ASA physical status classification: significant variation in allocation among Finnish anaesthesiologists. Acta Anaesthesiol Scand 1997; 41:629–632.
6. Barbeito A, Muir HA, Gan TJ, et al. Use of a modifier reduces inconsistency in the American Society of Anesthesiologists physical status classification in parturients. Anesth Analg 2006; 102:1231–1233.
7. Mak PH, Campbell RC, Irwin MG. The ASA Physical Status Classification: inter-observer consistency. American Society of Anesthesiologists. Anaesth Intensive Care 2002; 30:633–640.
8. Aronson WL, McAuliffe MS, Miller K. Variability in the American Society of Anesthesiologists Physical Status Classification Scale. AANA J 2003; 71:265–274.
9. Sidi A, Lobato EB, Cohen JA. The American Society of Anesthesiologists’ physical status: category V revisited. J Clin Anesth 2000; 12:329–334.
10. Malviya S, Voepel-lewis T, Chiravuri D, et al. Does an objective system-based approach improve assessment of perioperative risk in children? A preliminary evaluation of the ‘NARCO’. Br J Anaesth 2011; 21:352–358.
11. Wolters U, Wolf T, Stützer H, Schrôder T. ASA classification and perioperative variables as predictors of postoperative outcome. Br J Anaesth 1996; 77:217–222.
12. Higashizawa T, Koga YT. Modified ASA Physical Status (7 grades) may be more practical in recent use for preoperative risk assessment. Internet J Anesthesiol 2007; 15:1–5.

anaesthesia; American Society of Anesthesiologists’ status; classification; preoperative evaluation

© 2011 European Society of Anaesthesiology