Implementing Critical Care Billing on a Pediatric Hospital Medicine Service

Appropriate billing and coding significantly impact institutional reimbursement and operating margins, which can influence hospital finances, resource allocation, staffing, and the quality of patient care. Multiple studies have associated increased institutional financial stability with decreased readmission rates and length of stay, as well as improved patient-reported experiences and fewer adverse safety events.^1–5 

Critical care billing generates more relative value units (RVUs) and higher reimbursement compared with standard evaluation and management (E/M) billing. Physicians outside of the intensive care unit (ICU) can bill for critical care if both the patient and treatment provided meet the criteria set by the Centers for Medicare and Medicaid Services (CMS).^6–8 As more critically ill children are cared for by pediatric hospital medicine (PHM), it is essential for pediatric hospitalists to accurately capture the services provided and seek the corresponding reimbursement.^9,10 Failure to bill in cases where critical care has been delivered represents a significant financial loss with opportunity for improvement, as critical care can be delivered regardless of location.^6–8 While several prior initiatives within adult and emergency medicine have demonstrated success in improving clinician billing and documentation practices, there are no published data, to our knowledge, on improving these efforts within PHM.^11–14 Promoting the importance of critical care billing and broadening the understanding of this topic is therefore vital for pediatric hospitalists.

We previously investigated the existing billing practices of PHM physicians at our institution who treated critically ill patients admitted in status asthmaticus requiring continuous albuterol and identified a significant gap in clinician billing practices. The study revealed a substantial number of potential missed opportunities in physician RVUs and reimbursement for patients eligible for critical care billing on initial attending physician evaluation.⁹ This led to the systematic implementation of a quality improvement (QI) effort to initiate and improve critical care billing in eligible PHM patients at our institution.

The primary aim of our current QI initiative was to increase the proportion of critical care codes professionally billed by PHM attending physicians in a subset of critically ill patients meeting study eligibility criteria (receiving ≥5 L of high-flow nasal cannula [HFNC] or continuous albuterol) from a baseline mean of 21% to more than 50% in a 6-month period.

The initiative was conducted at a tertiary care academic children’s hospital for patients admitted to the PHM service from November 2023 through April 2024. Our institution comprises 150 pediatric inpatient hospital beds, with 56 total beds assigned to 4 general care units and 12 assigned to the pediatric ICU (PICU). On average, our PHM service consists of 20 clinical attending physicians rotating on day service weekly, with 1 evening shift in-house per week from 4 pm to midnight. A typical PHM census consists of 20 to 30 patients, and our service admits more than 2000 patients per year. Each clinical team comprises 2 or 3 resident physicians, an attending physician, and occasionally, an advanced practice provider (APP). APPs admit pediatric patients during evening and overnight shifts and can bill independently for PHM patients in general care on the calendar day of admission; however, critical care billing by APPs had not yet been established before this study.

Patients requiring up to 15 L of HFNC or continuous albuterol can be managed on the inpatient general care units, as well as patients who are tracheostomy dependent but not ventilator dependent, and those requiring established continuous or bilevel positive airway pressure. PHM clinicians file charges per patient encounter, and our compliance department audits these at random to ensure billing and documentation meet established guidelines. Our hospital has a small, group RVU–based incentive program, and our PHM division has consistently exceeded RVU benchmarks and earned incentive pay during the last several years. There was no change in incentive pay with the initiation of critical care billing. Our electronic health record (EHR) is Epic (Epic Systems).

Patients aged 0 to 18 years admitted to the PHM service who received HFNC and/or continuous albuterol with an encounter billed by a PHM attending physician during admission were included in the analysis. Patients without an E/M code applied or those billed by an APP on admission were excluded. Eligibility for critical care billing was determined by the CMS definition, which states, “critical illness or injury acutely impairs one or more vital organ systems such that there is a high probability of imminent or life-threatening deterioration in the patient’s condition.”⁸ Our QI group chose the application of HFNC and/or continuous albuterol as markers for potential critical illness. To more reliably capture the critical care services provided for patients meeting these criteria, we limited our study population to patients requiring at least 5 L of HFNC or continuous albuterol on attending physician evaluation. We additionally excluded patients weaned to intermittent albuterol or less than 5 L of HFNC either before or at the time of attending physician evaluation, as patients receiving less than 5 L of HFNC are often transitioned to simple nasal cannula at our institution. For eligible patients evaluated by a PHM attending physician and receiving multiple days of critical care services, each calendar day was included as a separate encounter and potential opportunity for critical care billing.

An interprofessional QI team was formed to identify strategies to successfully improve critical care billing in PHM. The team included a PHM fellow and APP as co-leads, 1 PICU and 5 PHM attending physicians, 4 resident physicians, 2 medical billing and coding specialists, and a pediatric billing revenue advisor.

Critical care billing for patients receiving HFNC or continuous albuterol was first initiated on the general care inpatient units in September 2023 after an educational session was provided to PHM faculty by the project leader (SNR). This included the reception of HFNC or continuous albuterol as a threshold for physicians to consider critical care billing and determine whether patients met the criteria as defined by CMS on assessment. Additional education was provided on the application of a time-based vs age-based global critical care code. Faculty were instructed that a time-based critical care code for patients aged 6 years or more required at least 30 minutes of critical care and that for patients for whom less than 30 minutes of critical care was provided, the application of an E/M code was recommended (eg, an 11-year-old patient requiring continuous albuterol for whom only 20 minutes of critical care time was provided by the attending physician).¹⁵ Simultaneously, critical care billing codes were uploaded into the EHR under a “Charge Capture Quick List” and grouped accordingly via patient age or time and initial or subsequent encounter (Supplemental Figure 1). Additionally, a standardized EHR template was developed and shared with PHM physicians that included the required documentation to support critical care billing (Supplemental Figure 2).

The active QI intervention period to standardize and improve critical care billing in PHM spanned from November 2023 through April 2024. Interventions were applied using Plan-Do-Study-Act (PDSA) cycle methodology and informed by a key driver diagram to systematically organize the theory of change (Figure 1).¹⁶ The first PDSA cycle consisted of a PHM reference guide for critical care billing with weekly EHR secure messaging reminders sent to the rotating on-service PHM physicians. Both the reference guide and chat reminders included guidance for when to consider critical care billing in eligible patients as well as when to bill an age-based vs time-based critical care code.

Targeted PHM clinician reeducation occurred in January 2024 and focused primarily on double-billing errors identified through an increase in insurance denials for patients transferred from the PHM service to the ICU on the same calendar day, as a global age-based critical care code could only be applied once per calendar day.¹⁵ A solution that was mutually agreed upon between the two services for same-day patient transfers consisted of the receiving service billing a time-based critical care code and the transferring service billing an age-based global code. This mirrored a similar process for patients transferred between the neonatal ICU and PICU previously established at our institution.

A billing algorithm was developed to standardize clinician billing practices and provide a visual representation for clinicians regarding same-day PHM/ICU transfers as well as additional guidance on when to bill an age-based global vs time-based critical care code per CMS regulations (Supplemental Figure 3).

Clinician auditing was retrospectively implemented by 2 QI members (SNR and MM) with bimonthly emails sent to PHM physicians identifying barriers and potential missed opportunities for critical care billing through the intervention period. Clinician auditing additionally evaluated the application of a time-based critical care code or age-based global code to standardize coding and documentation practices. Visual reminder cards were provided to faculty to post on their workstations as an additional intervention at the end of the month (Supplemental Figure 4).

The final intervention included a reminder in the PHM attending physician EHR attestation with a disappearing text for all clinicians to consider critical care billing in eligible patients. The critical care billing documentation template was further updated to include guidance on when to consider the application of a time-based vs age-based critical care code. This consisted of a disappearing text reminder that an age-based global critical care code could only be applied if the patient remained critically ill until midnight on the calendar day of the PHM attending physician evaluation¹⁵ (Supplemental Figure 5).

Data were collected for 4 months after the conclusion of the intervention period to evaluate the effort’s sustainability. Feedback from our institution’s billing and coding specialists regarding insurance denials was reviewed on a quarterly basis.

The primary outcome measure was the percentage of critical care codes billed in eligible PHM patients requiring at least 5 L of HFNC or continuous albuterol. A secondary outcome measure assessed the number of RVUs and estimated reimbursement generated from the initiation of critical care billing in PHM. This was compared with standard E/M RVUs and reimbursement to evaluate the financial differences and potential implications associated with these two practices. The process measure evaluated the percentage of patient charts containing the required documentation to support critical care billing in eligible critically ill patients. The balancing measure assessed the number of insurance denials during the study period.

We abstracted data from the EHR weekly including patient demographics, primary insurance type, the application of HFNC and/or continuous albuterol during admission, and critical care code used by a PHM attending. The patient’s order history (including the start and stop times for continuous albuterol and/or HFNC) and use of an age-based global vs time-based critical care code were manually reviewed by 2 QI team members (SNR and MM) during each eligible patient encounter and compared with documentation within the EHR. This dual-review process assessed patients who met inclusion criteria (requiring continuous albuterol and/or HFNC during admission) and evaluated the documentation as well as the timing of respiratory weaning and/or discontinuation of interventions within the EHR to ensure patients receiving at least 5 L of HFNC or continuous albuterol on attending physician evaluation were included in the analysis. We additionally reviewed the required written documentation within the EHR to support the level of critical care billing applied. Supporting documentation in the attending physician attestation consisted of an assessment of the patient’s critical illness, the physician’s direct management indicating the delivery of critical care, and the inclusion of critical care time when appropriate.^8,15 

To assess the impact of interventions, measures were tracked prospectively via statistical process control charts (P-charts) in Microsoft Excel and analyzed using established rules for special cause variation.^17,18 To evaluate areas for improvement, Pareto charts eliciting barriers as well as potential missed opportunities for eligible critical care billing encounters were created at multiple points throughout the study period and reviewed to inform future interventions.

To assess the secondary outcome measure, the number of RVUs and associated estimated reimbursement for the critical care code applied by the attending physician were evaluated and compared with a standard inpatient “high” initial or subsequent E/M Current Procedural Terminology (CPT) code (99223 or 99233, respectively). RVUs were assessed based on the 2019 American Medical Association rules (Table 1).¹⁵ Estimated average reimbursement was provided by our institution’s revenue cycle advisor (Table 2).

This project was undertaken as a QI initiative, and as per our institutional guideline for determining human subject research, it did not meet the definition of research according to 45 CFR §46.

Of the eligible encounters billed for critical care, most patients had Medicaid insurance (55%), self-identified as white/Caucasian (56%), were not Hispanic or Latino (72%), were male (64%), and had a median age of 13 months. Within these encounters, 144 of 216 (67%) patients were aged 29 days 2 to years, 40 of 216 (18%) were aged 2 to 5 years, and 32 of 216 (15%) were aged 6 to 20 years. During admission, 183 of 216 (85%) patients received HFNC, and 48 of 216 (22%) received continuous albuterol.

The baseline period consisted of 9 data points (n = 90 eligible patient encounters) after the initiation of critical care billing in PHM from September until the end of November 2023, when the active QI interventions were first implemented.

After the study’s first intervention at the end of November 2023, we detected special cause variation and the mean centerline for the primary outcome measure shifted from 21% to 59% (Figure 2). After additional interventions, special cause variation was again observed, and the mean centerline shifted from 59% to 74%, with improvements sustained to September 2024.

For the percentage of charts containing documentation to support critical care billing in eligible patients, we detected special cause variation after the second cycle of interventions with an upward shift in the baseline mean from 31% to 70% (Figure 3). We did not observe additional special cause variation, but improvements were similarly sustained to September 2024.

From September 2023 to September 2024, a total of 2092 RVUs and $222 934 in estimated reimbursement were generated using critical care codes for patients who met study eligibility criteria. In comparing the financial differences between critical care billing and standard “high” inpatient E/M billing, use of standard “high” inpatient E/M billing codes would have generated 709 RVUs and $55,051 in total estimated reimbursement. This illustrates that critical care billing in eligible patients led to a 3-fold increase in the total number of RVUs and a 4-fold increase in the estimated total reimbursement generated when compared with standard E/M billing practices.

During the study period, there were 5 (<1%) insurance denials for critical care billing out of a total of 826 (15.3%) insurance denials from 5407 charges by the PHM division. Critical care denials were primarily due to double-billing errors identified between the PHM/ICU services, in which an age-based global critical care code was used twice on the same calendar day. This led to the development of a third PDSA cycle targeting provider reeducation in January 2024 and a standardized billing process for same-day PHM/ICU transfers that was mutually agreed upon by the two services as previously described, with no additional insurance denials identified after February 2024.

Through an interprofessional QI initiative, our team systematically implemented critical care billing in eligible PHM patients and successfully tripled the use of critical care billing codes and doubled the number of charts containing the necessary documentation to support critical care billing. Our QI initiative effectively implemented critical care billing on a PHM service and significantly increased physician RVUs as well as institutional reimbursement, based on available estimates.

While several efforts have previously highlighted the importance of critical care billing and its optimization in hospitalized adults or in emergency medicine settings, our study demonstrates success in implementing and improving critical care billing for pediatric inpatients.^11–14 As children’s hospitals throughout the country face increasing financial strain with the divestment of resources from pediatric care, this initiative could serve as an important opportunity for other institutions and PHM divisions to improve financial performance and increase operating margins.^1,19 Furthermore, our study could be expanded upon to include additional eligible pediatric conditions and interventions, recognizing the variability in patients cared for by pediatric hospitalists.⁹ 

Improvements in our study’s measures largely coincided with PDSA cycles, emphasizing specific interventions that drove its success. The development of clinical decision support tools within the EHR, the establishment of an interprofessional QI team, and provider auditing with feedback were key to the success of our initiative. The creation of a PHM reference guide with weekly reminders to rotating on-service physicians as the first intervention led to special cause variation for our primary outcome measure. Although this required significant individualized effort through weekly messaging, it was paired with higher-reliability interventions to sustain the QI effort. These included workflow changes within the EHR, such as updating the critical care billing documentation tool with a disappearing text reminder for all PHM physicians to consider critical care billing in eligible patients in their attestation. Similar studies evaluating workflow interventions have demonstrated the effectiveness of higher-reliability interventions in QI studies, reinforcing that EHR changes may contribute to changes in practice and improve outcomes.^20,21 Furthermore, sustained improvement and special cause variation after conclusion of the individualized messaging highlights the effectiveness of the EHR changes applied and higher-reliability interventions used.

Engaging an interprofessional team as key stakeholders in our QI effort was also essential to ensuring clinician practice change and to successfully integrate PDSA cycles into provider workflow. By including ICU and PHM attending physicians, as well as billing and coding experts from our institution, we determined ways to standardize provider billing practices and documentation, discussed billing errors and mitigation strategies, and established a mutually agreed-upon process for patients transferred between the ICU/PHM services.

Through clinician auditing, we evaluated barriers and provided individualized feedback on potential missed opportunities for critical care billing, allowing for specific and targeted QI interventions that not only sustained the initial improvement but also led to an additional upward shift demonstrating further improvements in our primary outcome measure. Clinician auditing served as a reminder and an opportunity for improvement for clinicians for whom critical care billing was a cultural change from established billing practices. Individualized feedback seemed to help decrease the number of missed potential opportunities and improved documentation, demonstrating the effectiveness of clinician feedback as noted by other billing initiatives.¹¹ 

The QI effort was conducted at a single-site academic children’s hospital with billing and coding experts, QI specialists, and clinicians able to support its implementation, with institutional billing practices that may limit its generalizability to other centers.⁹ Additionally, critical care billing was influenced by the attending physician’s individual assessment, whereby some encounters were intentionally not billed for critical care, despite patients meeting study eligibility criteria as elicited by attending physician feedback (eg, a patient requiring 5 L of HFNC on attending physician evaluation but not considered critically ill on assessment and intentionally billed an E/M code). While this may have reduced the results of our measures, we opted to include these encounters to capture all potentially eligible patients and why our goal for this initiative was not 100%. Furthermore, our reimbursement projections were based on estimated averages per our institution’s revenue cycle advisor, which could either increase or decrease these estimates and may limit generalizability at other centers where funding models and payer mix vary.⁹ 

An interprofessional QI initiative to systematically implement critical care billing in eligible PHM patients at our institution was associated with an improvement in critical care billing and documentation, thereby significantly increasing physician RVUs and institutional reimbursement. Pairing higher-reliability EHR workflow interventions with clinician education, individualized auditing and clinician feedback, and visual reminders led to the optimization of critical care billing. Future directions include expanding the implementation of critical care billing to APPs at our institution and for additional applicable diagnoses. Our QI initiative could serve as an important intervention for other PHM divisions to consider to accurately capture the services provided and seek the corresponding reimbursement.⁹ 

We acknowledge Sarah Bartz for her assistance in data collection as well as Dr Natalia Paciorkowski for her resources on critical care billing.

CMS

Centers for Medicare and Medicaid Services

CPT

Current Procedural Terminology

E/M

evaluation and management

HFNC

high-flow nasal cannula

ICU

intensive care unit

PHM

Pediatric Hospital Medicine

PICU

pediatric intensive care unit

QI

quality improvement

RVU

relative value unit