Does the Current American Society of Anesthesiologists Physical Status Classification Represent the Chronic Disease Burden in Children Undergoing General Anesthesia?

Abstract

In 1940, the American Society of Anesthesiologists physical status (ASA-PS) classification system was conceived as a system to standardize the categorization of operative risk.1 After detailed reflection, research, and discussion on the topic, those developing the ASA-PS system considered it best to classify and grade patients by physical status. Over time, a 6-point scale emerged ranging from a healthy to deceased patient: (I) healthy person, (II) person with mild systemic disease, (III) person with severe systemic disease, (IV) person with severe systemic disease that is a constant threat to life, (V) moribund person who is not expected to survive without the operation, and (VI) declared brain-dead person whose organs are being removed for donor purposes.2 Although developed primarily for perioperative physical status in adult patients, the ASA-PS is universally applied to pediatric patients with potential for the current ASA-PS classification to underestimate the chronic disease burden in children undergoing general anesthesia. The use of the ASA-PS in children is challenging due to the subjectivity of classifications and their limited descriptions. Those attributes have resulted in ASA-PS scoring variation across anesthesia providers, both pediatric and adult.3,4 ASA-PS scoring in children is also challenging due to the heterogeneous array of rare, childhood chronic conditions that have substantial differences with regard to severity and complexity. Multiple studies have reported discrepancies in the ASA-PS scoring between providers in both academic and nonacademic hospitals.3–5 The limited studies on ASA-PS use in children report less inter-rater reliability than in adult patients.5,6 Despite this finding, ASA-PS scoring is embedded in preanesthetic care worldwide as a standard component of perioperative health assessment for patients of all ages. To best match a child’s chronic disease profile with ASA-PS categories, a novel system is proposed to inform perioperative risk stratification. This is not intended to alter the classification for acute injury or illness which is clearly defined in the current ASA-PS categories but to more explicitly consider the impact of complex and chronic diseases. Children have clinical vulnerability especially during the perioperative period, which further emphasizes the need for a modification of the current ASA-PS classification with uniform definitions and examples to accurately and consistently classify and predict perioperative care needs and outcomes when applied to pediatric patients. Little has been documented regarding the chronic disease profiles across ASA-PS categories in children undergoing anesthesia. To assess the prevalence of chronic diseases represented in the ASA-PS classification system for children undergoing surgery and related procedures under anesthesia, 4564 pediatric patients undergoing surgery and related procedures under general anesthesia between May 1, 2017 and December 31, 2017 in a tertiary-care, freestanding children’s hospital were reviewed. The number, prevalence, and complexity of chronic conditions across ASA-PS categories were compared to distinguish areas of incongruent assignment of ASA-PS category with respect to chronic disease burden. CURRENT STATE AND ASSESSMENT OF CHRONIC DISEASE The ASA-PS scores were defined as I = completely healthy; II = mild systemic disease; III = severe systemic disease; IV = severe systemic disease that is a constant threat to life; and V = moribund patient (who is not expected to live 24 hours with or without surgery).7 A score was assigned to each patient and documented preoperatively by the attending anesthesiologist responsible for each patient. Scoring was based on the assessment of general medical health by review of the electronic health record and in-person history and physical examination. The Agency for Healthcare Research and Quality (AHRQ) Chronic Condition Indicator (CCI) system and Feudtner Complex Chronic Conditions (CCC) were used to assess the chronic disease burden of the populations of patients in each assigned ASA-PS classification group.8–10 Both systems use International Classification of Diseases (ICD) diagnosis codes. The ICD codes were documented preoperatively through health care claims for outpatient visits, emergency department visits, and prior hospitalizations. ICD diagnosis codes were assigned by clinicians (eg, physicians, advance practice nurses, physician assistants) during outpatient clinic visits, emergency department visits, and hospitalizations ahead of the elective surgery encounter studied. The AHRQ CCI system is an open-source, publicly available diagnosis classification scheme that classifies approximately 14,000 (ICD) diagnosis codes as chronic or not. Using the CCI designation, a chronic condition is defined as a condition that lasts 12 months or longer and has one or both of the following effects: (1) it places limitations on self-care, independent living, and social interactions; and (2) it results in the need for ongoing intervention with medical products, services, and special equipment.11,12 For infants younger than 12 months of age, the condition has to be perceived as lasting 12 months or longer to qualify. The AHRQ CCI system consolidates ICD codes into clinically meaningful, mutually exclusive diagnosis categories (eg, asthma, epilepsy, and gastroesophageal reflux disease).11,12 The classification systems were purposely developed to include complex and rare, genetic syndromes. The AHRQ CCI system captures all coexisting conditions. For example, a child with cystic fibrosis who also has codes for gastroesophageal reflux, failure to thrive, and diabetes would show in the system as having respiratory, gastrointestinal, and endocrine organ systems affected. Children with single ventricle physiology (eg, hypoplastic left heart syndrome) with a tracheostomy, parenteral nutrition dependence, and cholestasis would show in the system as having cardiac, respiratory, and digestive organ systems affected. All chronic condition categories map exclusively to 1 of 25 overarching groups that are organized by organ system. The AHRQ CCI system has been adapted for pediatric use by Berry et al.9,13,14 The AHRQ CCI system (1) determines whether a patient’s diagnosis is a chronic condition and (2) identifies the specific organ system primarily affected by each chronic condition. Both systems rely on ICD diagnosis codes to identify and classify chronic conditions. All of the codes were evaluated by means of a modified Delphi method/expert opinion engaging a panel of Boston Children’s Hospital senior clinicians. Chronic conditions such as hearing loss, allergic dermatitis, and cataracts were eliminated if they were not thought to be physiologically relevant to anesthetic risk. Congenital conditions that do not contribute to anesthetic risk such as congenital deafness or club feet were not included. Associated conditions were classified with their own codes when present. Feudtner CCCs were integrated with the AHRQ CCI to designate pediatric chronic conditions that are considered complex.8 The CCCs are life-limiting, childhood health conditions that are associated with severe limitations in function as well as high morbidity and mortality.8,15 Organ system-based paired examples of CCCs and non-CCCs include cystic fibrosis and asthma (respiratory), inflammatory bowel disease and constipation (gastroenterologic), and attention-deficit hyperactivity disorder and cerebral palsy (neurologic/behavioral). Included in the CCC system are assistive medical technologies, including gastrostomy, tracheostomy, and cerebrospinal fluid ventricular shunt. The CCCs correlate strongly with the risk of postoperative complications in children.16 CHARACTERISTICS OBSERVED The medical records of 4564 patients who were admitted to an inpatient unit postoperatively were reviewed (Supplemental Digital Content 1, Table 1, https://links.lww.com/AA/C629). ASA-PS I: “Completely Healthy Patients”Figure.: The percentages of children with the chronic condition attribute for each ASA-PS score. The chronic conditions are anesthesia relevant; they do not include mild chronic conditions (eg, hearing loss). The blue shaded boxes distinguish areas of potential underclassification of ASA-PS based on the chronic condition attribute. ASA-PS indicates American Society of Anesthesiologists physical status.Forty-five percent had ≥1 chronic condition, and 17.1% had ≥2 chronic conditions. The most common chronic conditions were asthma (4.7%, n = 23), cleft palate (4.7%, n = 23), and congenital face/neck anomalies (3.0%, n = 15) (Supplemental Digital Content 2, Table 2, https://links.lww.com/AA/C630). Twenty-nine percent had ≥1 CCCs. The most common CCCs were scoliosis (10.2%, n = 50), cardiac dysrhythmias (4.3%, n = 21), and congenital anomaly of skull (4.1%, n = 20). One percent (n = 7) were assisted with medical technology, and the most common technologies were enterostomy (n = 3) and an orthopedic prosthetic device (n = 3) (Figure, ASA I). ASA-PS II: “Mild Systemic Disease” Seventy-eight percent had ≥1 chronic conditions, and 50.3% had ≥2 chronic conditions. The most common chronic conditions were heart valve disorders (14.6%, n = 316), asthma (13.0%, n = 282), and obesity (7.4%, n = 159) (Supplemental Digital Content 3, Table 3, https://links.lww.com/AA/C631). Sixty-one percent had ≥1 CCCs. The most common CCCs were cardiac dysrhythmias (13.7%, n = 297), scoliosis (10.6%, n = 228), and central nervous system disorders (7.4%, n = 160). Seventeen percent (n = 377) were assisted with medical technology, and the most common technologies were enterostomy (8.1%, n = 175) and mechanical ventilator dependence (2.7%, n = 59) (Figure, ASA II). ASA-PS III: “Severe Systemic Disease” Ninety-eight percent had ≥1 chronic conditions, and 88.5% had ≥2 chronic conditions. The most common chronic conditions were heart valve disorders (35.1%, n = 638), asthma (23.0%, n = 418), and epilepsy (18.5%, n = 336) (Supplemental Digital Content 4, Table 4, https://links.lww.com/AA/C632). Ninety-three percent had ≥1 CCCs. The most common CCCs were cardiac dysrhythmias (32.1%, n = 583), scoliosis (19.4%, n = 352), and central nervous system disorders (19.3%, n = 351). Fifty-five percent (n = 999) were assisted with medical technology, and the most common technologies were enterostomy (29.8%, n = 542) and ventilator dependence (14.7%, n = 268) (Figure, ASA III). ASA-PS IV: “Severe Systemic Disease That Is a Constant Threat to Life” All patients (100%) had ≥1 chronic conditions, and 97.7% had ≥2 chronic conditions. The most common chronic conditions were heart valve disorder (56.6%, n = 73), respiratory insufficiency (28.7%, n = 37), and asthma (27.1%, n = 35). Ninety-nine percent had ≥1 CCCs (Supplemental Digital Content 5, Table 5, https://links.lww.com/AA/C633). The most common CCCs were cardiac dysrhythmias (53.5%, n = 69), congenital heart anomaly (43.4%, n = 56), and cardiac conduction disorder (42.6%, n = 55). Seventy-six percent (n = 99) were assisted with medical technology, and the most common technologies were enterostomy (50.4%, n = 65) and ventilator dependence (43.4%, n = 56) (Figure, ASA IV). Chronic Disease Burden Across ASA-PS Categories In general, chronic disease burden increased with ASA-PS score. As ASA-PS increased from I to IV, the percentage of patients with no chronic condition decreased from 54.7% to 0.0%, the percentage of patients with multiple chronic conditions increased from 17.1% to 97.5%, the percentage of patients with a CCC increased from 29.7% to 99.2%, and the percentage of patients with technology assistance increased from 1.4% to 76.7% (P < .001 for all). Despite the findings above, several incongruent discoveries of chronic disease burden across ASA-PS were observed. For example, across the systemic disease ASA-PS categories II, III, and IV, there were 515 (12.5%) patients without any chronic disease. In ASA-PS I (healthy patients), there were 223 (45.3%) patients with a chronic disease. In ASA-PS categories I (healthy patient) and II (mild systemic disease), there were 1469 (55.3%) patients with CCCs, which are associated with the highest morbidity and mortality risk of all chronic conditions. Included in the patients with CCCs who were scored ASA I or II there were 384 (14.5%) who were assisted with major medical technology (eg, enterostomy) due to a CCC that caused significant physiologic and/or functional impairment. SHORTCOMINGS OF THE CURRENT ASA-PS CLASSIFICATION IN PEDIATRIC PATIENTS It is suggested that in the populations of children classified by current ASA-PS, there were distinguishable areas of incongruent ASA-PS assignment with respect to chronic disease burden. Most patients (85.3%; n = 3891) had at least 1 organ system affected by an anesthesia relevant chronic condition and 71.6% (n=3294) had at least 1 anesthesia relevant CCC (Table 1). In certain situations, the assignment may have underrepresented the disease burden, and in others, the assignment may have overrepresented the burden. Leveraging patients’ chronic disease information with the AHRQ CCI and Feudtner CCCs may be useful to consider when identifying processes to optimize the use of ASA-PS in children and align with the chronic disease burden of children undergoing anesthesia for surgical and related procedures.Table 1.: Clinical Characteristics of Patient Undergoing Preanesthetic EvaluationThe subjectivity of ASA-PS assignment to pediatric patients may have contributed substantially to the main findings. Examples of pediatric chronic diseases and coexisting conditions are not present in current definitions and instructions for the use of ASA-PS; therefore, anesthesia providers relied heavily on their perceptions of health and chronic disease when assigning ASA-PS classification. Derived predominately for stratification of adult patients, the ASA-PS scoring system is particularly challenging when applied to the pediatric population. Most of the health conditions and behaviors provided in the ASA-PS system documentation (eg, myocardial infarction, alcohol dependence, smoking, etc) do not apply to children.7 The chronic health conditions which are more prevalent in the pediatric population and the burden of risk that they carry are not well defined within the ASA-PS classification, which may result in ambiguity and subjectivity of the scores. In the context of the increasingly medically complex population of patients receiving care at our institution and at children’s hospitals throughout the United States, it is possible that the providers, who are constantly exposed to high acuity patients, applied a skewed view that minimized the complexity and the severity of systemic disease. This may explain why so many pediatric patients in the current study with multiple and CCCs were classified with the “healthy” ASA I-PS. Better education regarding the role of medical comorbidities in determining the ASA-PS classification in children that are different than those considered in adults might be of value. The incongruences in chronic disease burden and ASA-PS classification observed have immediate implications for clinical practice. The ASA-PS is currently used in several tools that assess perioperative risk, including the Pediatric Risk assessment score17 and the American College of Surgeons National Surgical Quality Improvement Program.18 Under- and overclassification of ASA-PS might compromise the accuracy of assignment for perioperative risk in children. In addition, ASA-PS is currently used by hospitals for payor reimbursement for perioperative health services. Greater reimbursement occurs for patients with higher ASA-PS classification, especially with classifications III, IV, and V. Billing for anesthesia services is based on units of time, procedure and diagnosis codes, duration of anesthesia, and modifiers. One additional billing unit is added for ASA-PS 3 patients, 2 additional units for ASA-PS 4, and 3 additional units for the ASA-PS 5 patients. If patients are incorrectly under classified with regard to ASA score, reimbursement for anesthesia services may be compromised.19 One study projects that adoption of Medicare rates by other payers will lead to significant decreases in total payments to academic anesthesiology. The estimated decrease would range from 21% to 37%, depending on payment policies for uninsured patients.20 For instance, 17% of ASA II patients in our study required medical technology assistance and therefore by definition should be an ASA III physical status (Figure). More accurate classification of patients may have the potential to offset some of the decreased revenue. Hospitals and health care systems, especially those treating complex patients, should focus their efforts on improving their cost structures through more accurate assignment of ASA-PS classification. It is possible that the open-source chronic diagnosis classification systems such as the AHRQ CCI and Feudtner CCCs could be applied in real time to existing clinical data to produce a chronic disease profile for each pediatric patient ahead of their surgical or related procedure. Adding pediatric-specific CCCs with examples might further standardize information to help define the current existing categories. This profile could be augmented with clinical decision support to help accurately assign ASA-PS classification. The proposed Pediatric ASA-PS definitions should help to resolve inherent subjectivity in differentiating patients with “mild systemic disease,” “severe systemic disease,” and “severe systemic disease that is a constant threat to life” (Table 2). For example, the system might prompt an anesthesiology provider to consider the classification of children assisted with medical technology (eg, tracheostomy) automatically as ASA-PS III or higher. Such prompts could potentially optimize matching between ASA-PS classification and actual chronic disease burden. The proposed Pediatric ASA-PS classification would be based on the existing ASA-PS classification which is already widely used to determine resource allocation, reimbursement for anesthesia services, statistical data collection and analysis, and predicting perioperative risk.1–4 The improved accuracy of a pediatric ASA-PS classification has the potential to translate into improved assessment of preoperative risk in pediatric patients, recognition for the need for specialized expertise for care optimization, and improved reimbursement rates for anesthesia services.Table 2.: Proposed Pediatric ASA-PS Score ExamplesProfiling the chronic disease burden of pediatric patients using administrative diagnosis codes may not capture the true severity or functional limitations that occur as a result of the chronic condition. By highlighting areas where there may be incongruences between this burden and ASA-PS, subsequent efforts can move forward to generate additional information to perioperative clinicians to best guide them on how to accurately assign ASA-PS in children. It is not suggested that this be the entire basis for ASA-PS scoring but 1 component in addition to acute health issues which may be present. The proposed pediatric ASA-PS will require testing and validation; however, these efforts may ultimately lead to pediatric-specific guidance on the accurate assignment of ASA-PS in infants and undergoing anesthesia for surgery and related procedures. DISCLOSURES Name: Izabela Leahy, RN, BSN, MS. Contribution: This author helped design the study/generate the idea, manage and analyze the data, review the literature, and compose and edit the manuscript. Name: Jay G. Berry, MD, MPH. Contribution: This author helped design the study/generate the idea, manage and analyze the data, review the literature, and compose and edit the manuscript. Name: Connor J. Johnson, BS. Contribution: This author helped manage and analyze the data and compose and edit the manuscript. Name: Charis Crofton, BA. Contribution: This author helped manage and analyze the data and compose and edit the manuscript. Name: Steven J. Staffa, MS. Contribution: This author helped design the study/generate the idea, manage and analyze the data, review the literature, and compose and edit the manuscript. Name: Lynne Ferrari, MD. Contribution: This author helped design the study/generate the idea, manage and analyze the data, review the literature, and compose and edit the manuscript. This manuscript was handled by: James A. DiNardo, MD, FAAP.