An Accountable Care Organization Maintains Access for Appendicitis During the COVID-19 Pandemic

Abstract

Introduction

It has been reported that pediatric patients experienced a delay in treatment for acute appendicitis during the pandemic resulting in increased rates of complicated appendicitis. We investigated the association of the COVID-19 pandemic and the incidence and severity of acute appendicitis among pediatric Medicaid patients using a population-based approach.

Methodology

The claims database of Partners For Kids, a pediatric Medicaid accountable care organization in Ohio, was queried for cases of acute appendicitis from April–August 2017-2020. The monthly rate of acute appendicitis/100,000 covered lives was calculated each year and compared over time. Rates of complicated appendicitis were also compared. Diagnosis code validation for classification as complicated or uncomplicated appendicitis was performed for patients treated at our hospital.

Results

During the study period, 465 unique cases of acute appendicitis were identified. Forty percent (186/465) were coded as complicated. No significant difference in the incidence of acute appendicitis cases was observed across the 4 years, either in an overall comparison or in pairwise comparisons (p>0.15 for all). The proportion of acute appendicitis cases that were coded as complicated did vary significantly over the 4-year study period (p=0.005); this was due to this proportion being significantly higher in 2018 than in either 2019 (p=0.005 vs. 2018) or 2020 (p=0.03 vs. 2018).

Conclusions

The COVID-19 pandemic was not associated with reduced access to treatment for acute appendicitis among patients in a pediatric Medicaid accountable care organization. This suggests that an accountable care organization may promote continued healthcare access for their covered population during an unexpected crisis.

Introduction

Acute appendicitis is a common disease with the potential to affect all age groups. It is often classified into two categories based on disease severity - complicated and uncomplicated. The precise definition of complicated appendicitis is often a source of debate among surgeons. Generally, it is understood to be the more severe disease manifestation, with perforation of the appendix and its sequelae as the hallmark of the classification. Uncomplicated, on the other hand, can be characterized as disease confined to the appendix, without perforation. Within the pediatric population, acute appendicitis makes up the majority of acute abdominal pain hospitalizations at 20-30%(1). Since the 1990s however, several European studies have observed a decline in the incidence of acute appendicitis in the pediatric population (2, 3, 4), with one study showing this trend to be true for both uncomplicated appendicitis and complicated appendicitis(2).

In early 2020, public awareness of the first cases of COVID 19 disease in the Unites States led to the implementation of public health regulations aimed at slowing the spread of disease. One such measure came in the form of public advisory to remain at home except for essential travel. While these measures were effective in “bending the curve,” or decreasing the surge of patients requiring hospitalization, unanticipated trends began to emerge as well. Those trends included a decrease in admissions for urgent/emergent disease processes that typically have minimal fluctuations from year to year. One such urgent disease process was acute appendicitis. In the US and abroad, clinicians noted a decrease in the incidence of acute appendicitis presentations to healthcare centers and an increase in the incidence of severe disease, presumably from delayed presentation. Most studies investigating this phenomenon are single institution, or small multi-institutional studies, with very few focused on pediatric populations (5, 6, 7, 8, 9, 10). Available population studies were either not pediatric specific or focused on the immediate impact of the COVID-19 pandemic by evaluating the pre and post effect of government issued “states of alarm” in 2020 (11, 12, 13, 14, 15). The purpose of this study was to better characterize the association of the COVID-19 pandemic and the incidence and severity of acute appendicitis among pediatric patients at the population level, in comparison to trends seen in years prior to the pandemic. We hypothesized that the incidence of appendicitis was lower and the proportion of cases that were complicated was higher in 2020 compared to the years prior.

Methods

This study underwent review and received approval by the Institutional Review Board HRP-580 (STUDY00001577) with a waiver of consent. The study population consisted of children in Partners For Kids (PFK), our hospital’s pediatric Medicaid accountable care organization, which covers approximately 470,000 individuals annually. Individuals residing across 47 counties in central, southeastern, and western Ohio who are aged 0-19 years are covered by PFK if they are eligible for Medicaid due to being from a low-income household. Individuals aged 0-21 years are covered by PFK if they are eligible for Medicaid due to a disability. Approximately 95% of all children in PFK’s geographic region, who are enrolled in Medicaid, are part of PFK. The PFK claims database was queried by ICD-10-CM codes for unique eligible patients for each month of April – August for each of the years 2017, 2018, 2019, and 2020. The time frame of April - August was chosen to correspond to the period right after the state-wide shut down went into effect, to the time when many restrictions were lifted. The same time period was compared across years to account for annual changes in appendicitis presentation. Fourteen appendicitis diagnosis codes were found in the PFK database, corresponding to the 5 original ICD-10-CM codes established in 2015 and the 9 additional ICD-10-CM codes added in October 2018 (See Table.1 for code descriptions). Based on surgeon interpretation of the code description and the described intended classification changes made by the Center for Medicare and Medicaid Services and the Center for Disease Control (CDC) 2017 ICD-10 Coordination and Maintenance Committee Meeting, K35.2, K35.3, K35.20, K35.21, K35.32, and K35.33 were deemed complicated appendicitis and K35.80, K35.89, K37, K35.30, K35.31, K35.8, K35.890, and K35.891 were designated as uncomplicated appendicitis. A sub-analysis was then conducted to validate these code designations via manual chart review of those patients in the PFK database who were treated for appendicitis at our institution. Complicated appendicitis was classified as having any of the following described in the operative report: hole visualized in the appendix, extraluminal fecalith, well-formed abscess in the abdominal cavity, or frank stool in the abdominal cavity. If a patient was treated non-operatively during the index presentation, those with perforated appendicitis with an abscess or phlegmon noted on imaging that was not amenable to surgical intervention were classified as complicated appendicitis. All others were deemed uncomplicated.

Table 1.

ICD10 Acute Appendicitis Codes and Descriptions With 2017 Modifications

ICD 10 Code	Description
COMPLICATED APPENDICITIS
K35.2	Acute appendicitis with generalized peritonitis. Includes: Appendicitis (acute) with generalized (diffuse) peritonitis following rupture or perforation of appendix.
K35.20	Acute appendicitis with generalized peritonitis, without abscess. (Acute) appendicitis with generalized peritonitis not otherwise specified (NOS).
K35.21	Acute Appendicitis with generalized peritonitis, with abscess
K35.3	Acute appendicitis with localized peritonitis
K35.32	Acute appendicitis with perforation and localized peritonitis, without abscess. (Acute) appendicitis with perforation NOS
	Perforated appendix NOS
	Ruptured appendix (with localized peritonitis) NOS
K35.33	Acute appendicitis with perforation and localized peritonitis, without abscess. (Acute) appendicitis with (peritoneal) abscess NOS
	Ruptured appendix with localized peritonitis and abscess.
UNCOMPLICATED APPENDICITIS
K35.30	Acute appendicitis with localized peritonitis, without perforation or gangrene. Acute appendicitis with localized peritonitis NOS
K35.31	Acute appendicitis with localized peritonitis and gangrene, without perforation
K35.8	Other and unspecified acute appendicitis
K35.80∗	Unspecified acute appendicitis
K35.89	Other acute appendicitis
K35.890	Other acute appendicitis without perforation or gangrene
K35.891	Other acute appendicitis without perforation, with gangrene. (Acute) appendicitis with gangrene NOS
K37∗	Unspecified appendicitis

ICD10 codes taken from the 2017 ICD-10 Coordination and Maintenance Committee Meeting and The Web's Free 2021 ICD-10-CM/PCS Medical Coding Reference (icd10data.com).

^∗ICD10 code not included in the 2017 ICD-10 Coordination and Maintenance Committee Meeting but in the original ICD 10 acute appendicitis codes. NOS = not otherwise specified.

Data were summarized by calendar month, and linear regression models were fitted to estimate mean monthly rates of appendicitis per 100,000 covered lives and to compare these rates by calendar year. Analogous models were fit to evaluate the proportion of patients with complicated appendicitis. Whether these outcomes varied by year was evaluated overall and in pairwise comparisons after adjustment for multiple comparisons using the Tukey method. SAS Enterprise Guide version 8.1 was used for all statistical analyses.

Results

Between April and August of 2017-2020, there were 465 unique cases of acute appendicitis among children in PFK (Table.2 ). The median age and interquartile range (IQR) of children in the study cohort for 2017-2020 was as follows: 12 [IQR 8-16], 12 [IQR 8-15], 12 [IQR 8-15], and 13 [IQR 9-15]. The age distribution did not vary by year (p=0.79). Forty percent (N=186) of these acute appendicitis cases were complicated. Of the 465 patients, 233 (50%) were treated for appendicitis at our institution. Validation of the diagnosis codes in the claims data through review of the medical records of these 233 patients revealed that the diagnosis codes we used to define complicated appendicitis actually represented complicated appendicitis 51.3% (95% CI 42.3-60.2) of the time, and the diagnosis codes we used to define uncomplicated appendicitis represented cases of uncomplicated appendicitis 99.1% (95% CI 95.3-99.9) of the time. When considering just the years 2019 and 2020, when the newer ICD-10-CM diagnosis codes for appendicitis were being used, we found that that the diagnosis codes we used to define complicated appendicitis actually represented complicated appendicitis 75.6% (95% CI 60.7-86.2) of the time, and the diagnosis codes we used to define uncomplicated appendicitis represented cases of uncomplicated appendicitis 98.6% (95% CI 92.6-99.8) of the time.

Table 2.

Cohort Demographics

Characteristic	All patients (n = 465)	2017 (n = 120)	2018 (n = 131)	2019 (n = 116)	2020 (n = 98)	P-Value
Age	12 (8-15)	12 (8-16)	12 (8-15)	12 (8-15)	13 (9-15)	0.79
Male	282 (60.7)	69 (57.5)	86 (65.7)	62 (53.5)	65 (66.3)	0.13
Race/Ethnicity						<0.001a
White	275 (59.1)	64 (53.3)	72 (55.0)	70 (60.3)	69 (70.4)
Black/African American	62 (13.3)	9 (7.5)	19 (14.5)	24 (20.7)	10 (10.2)
Asian	7 (1.5)	0 (0)	1 (0.8)	2 (1.7)	4 (4.1)
Native American	7 (1.5)	1 (0.8)	4 (3.1)	1 (0.9)	1 (1.0)
Other or Unknown	114 (24.5)	46 (38.3)	35 (26.7)	19 (16.4)	14 (14.3)
Metropolitan resident	292 (62.8)	71 (59.2)	91 (69.5)	75 (64.7)	55 (56.1)	0.16

Median (IQR) or n (column %) are shown. P values are from Mann Whitey U tests, chi square tests, or Fisher exact tests as appropriate.

^ap = 0.09 if the patients of “other or unknown” race are excluded

We did not find a significant difference in the incidence of acute appendicitis cases across the 4 years, either in an overall comparison or in pairwise comparisons (p>0.15 for all). However, there was a marginally significant linear decline in the incidence over the study period (p=0.08) (Figure 1 ). In the overall study cohort, the proportion of acute appendicitis cases that were complicated did vary significantly over the 4-year study period (p=0.005); this was due to this proportion being significantly higher in 2018 than in either 2019 (p=0.005 vs. 2018) or 2020 (p=0.03 vs. 2018) (Figure 2 a). However, when only the 233 patients treated at our hospital were included, we found that the proportion of acute appendicitis cases that were truly complicated did not vary significantly over the 4-year study period, either in an overall comparison or in pairwise comparisons (p>0.60 for all) (Figure 2b).

Figure 1.

Incidence of Acute Appendicitis. The average monthly incidence of acute appendicitis /100,000 covered lives over the 4-year study period is depicted. There was no statistically significant difference in the incidence over the study period.

Figure 2.

a: Yearly Proportion of Complicated Appendicitis in all Patients: The proportion of appendicitis patients with a claim having a diagnosis code for complicated appendicitis over the 4-year study period is depicted. There was a significant difference overall, which was due to 2018 being significantly higher than 2019 (p=0.005 vs. 2018) and 2020 (p=0.03 vs. 2018). Figure 2b: Yearly Proportion of Complicated Appendicitis in Patients Treated at our Tertiary Children’s Hospital: The proportion of appendicitis patients treated at our institution with complicated appendicitis, according to chart review, over the 4-year study period is depicted.

Discussion

There was no significant association of the COVID-19 pandemic and the incidence and severity of acute appendicitis in a pediatric Medicaid population in Ohio. No significant differences in the overall incidence of acute appendicitis were identified from April-August 2020 compared to the same time frame in 2017, 2018, and 2019. With regard to the proportion of cases that were complicated, we found that 2020 did not have a statistically significant higher proportion of complicated appendicitis compared to the same time frame in the three prior years. Lastly, we found that while the accuracy of the ICD-10-CM coding system improved with the modification that took place in October 2018, there remains significant overlap in the verbiage of the descriptions that allows for the usage of several of the codes for both complicated and uncomplicated appendicitis.

The findings of our study may in part be explained by the nature of PFK as an accountable care organization (ACO). Accountable care organizations are a relatively new business model in which participating physician groups or hospitals and physician groups function to coordinate the healthcare of a population of patients through shared savings and quality improvement (16). The shared savings are based upon benchmarks of previous costs for the population of patients cared for by the participating members of the ACO(16). Through this coordination of care, multiple studies have demonstrated statistically significant improved timely access to care and coordination with the primary care physician (17, 18) compared to non-ACOs as well as pre-and post-ACO implementation. In the PFK ACO model, a tertiary pediatric children’s hospital partners with over 2,100 individual providers in the central and southeastern regions of Ohio to bridge the gap and provide a safety net of high-quality care to the resource poor pediatric Medicaid population in this area. Looking at our results in this context, while the overall rate of complicated appendicitis is on par with some of the higher rates of complicated appendicitis among pediatric Medicaid populations (19), the maintenance at that rate despite the pressure of the global pandemic likely speaks to a well-established referral network between those community providers and our tertiary children’s hospital. PFK has served the Ohio Medicaid community for over 20 years and has worked to improve the quality of care its patients receive (20). Its long-established presence and ease of access within this ACO may be reflected in the observed results.

In comparing our results to the literature, only a few studies analyzing the same question from a pediatric population perspective (21, 22) were encountered. In the study by Schäfer et al, the primary endpoint was the incidence of complicated appendicitis during the 2-month statewide shut down of 2020 compared to the same time period in 2018 and 2019. Their results demonstrated a significant increase in the rate of complicated appendicitis (p = 0.035). This study was conducted abroad (Germany) and included all major pediatric surgery institutions within the state of Bavaria. While Germany is known to be a dual public-private health insurance system, the article gives no mention as to the insurance status of the children included in this study, making a direct comparison to our study difficult. The differences in observed outcomes, may highlight the utility of an accountable care organization type structure, to help facilitate access to the healthcare system and prompt referral to a surgery center for treatment of acute appendicitis.

In the study by Theodorou et al, the primary outcome was the rate of complicated appendicitis and its management. The study was conducted in the United States and included four tertiary children’s hospitals in the state of California. They evaluated a 6-month period of shelter in place orders in 2020 to the same time period the year prior and found the rate of complicated appendicitis to be unchanged. Unfortunately, a sub-analysis based on the type of patient insurance (public or private) was not conducted, but two-thirds of their patients had public insurance. It may be that similar to our study, with 100% of participants having public insurance, their patients were also part of one or several accountable care organizations and heavily swayed results, or simply, that the pediatric population in the U.S. did not experience a delay in care, regardless of insurance type.

Lastly, the accuracy of the ICD-10-CM coding system to represent complicated vs. uncomplicated appendicitis bears further discussion. The introduction of the ICD-10-CM coding system was made with the purpose of improving the accuracy of diagnoses. However, it has previously been demonstrated with regard to acute appendicitis, that this transition carried with it an inherent bias toward complicated appendicitis, with a significant rise in the number of cases coded as complicated appendicitis in the pediatric population(23) comparing pre- and post-transition. Tian et. al conducted their study utilizing the Pediatric Health Information System (PHIS) database, which is made up of 49 different children’s hospitals and contains inpatient, ambulatory surgery, emergency department, and observation patient data, and compared the rates of complicated appendicitis pre- and post- transition. They utilized the original five ICD-10 codes for acute appendicitis and designated them according to the Agency for Healthcare Research and Quality 2018 categorization with codes K35.2 and K35.3 as complicated appendicitis and codes K35.80, K35.89, and K37 as uncomplicated appendicitis. Given the database nature of their study they could not quantify the degree to which ICD codes K35.2 and K35.3 may be overestimating complicated appendicitis but noted a significant increase in complicated appendicitis from pre- to post-ICD 10 implementation.

The use of these two codes for both complicated and uncomplicated appendicitis was recognized by the Centers for Disease Control and Prevention (CDC) as well, and in the 2017 ICD-10 Coordination and Maintenance Committee Meeting, the issue was addressed. The outcome of that meeting was changes in the language of code descriptions to better characterize when a code was intended to characterize a perforation in the appendix, and removal of the word peritonitis from some descriptions, which clinically does not always signify perforation. While additional codes were also added to improve clarity, the original codes were not removed, thus the ambiguity persists. Additionally, the term peritonitis, described as generalized or focal, is still present, and the discussion attached with the meeting notes implies they intended to use generalized peritonitis to indicate perforated appendicitis. Utilizing both the explanation of what they intended and the code descriptions themselves, we arrived at the code designations utilized in this study. As we demonstrated, the accuracy improves to 75% as more centers utilized the updated codes, but we still found a 25% error rate through our own internal code validation. While this improvement is substantial, a large margin of error remains, highlighting the fact that caution should be exercised when conducting population level database studies that have inherent discrepancies due to the limitations in code accuracy that cannot be verified without the granular data, such as the operative report. There remains room for improvement in the code descriptions utilized in the 2018 revisions, and perhaps the term peritonitis has no role as an indicator of appendiceal perforation.

This study has several limitations inherent to its design. As a database study it is subject to misclassification and limited granularity of the data. This limitation was partially mitigated through institutional chart review of over half of the patients in the PFK database. Additionally, these findings may be unique to the Medicaid pediatric population of central/southeastern Ohio in which the study was conducted, and specifically, Medicaid patients that are part of an accountable care organization.

Conclusion

Overall, this study adds unique insight to the literature on the association of the COVID-19 pandemic and the rates and severity of acute appendicitis in the pediatric Medicaid population. Similar to another U.S. multi-center study with a high percentage of patients on public insurance, no differences were found in the incidence or severity of acute appendicitis in 2020 compared to pre-pandemic years. In this study, the ability of an accountable care organization to maintain access to the healthcare system likely played a role in the observed findings. Additionally, this study highlights the continued challenge of database studies utilizing ICD-10 diagnosis codes for appendicitis that have inherent ambiguity in their descriptions. While significant improvement was made in the accuracy of acute appendicitis codes with the 2018 modifications, the coding system may benefit from further revisions and complete removal of terminology that can be applicable for both complicated and uncomplicated appendicitis, such as peritonitis.

Declarations of Interest

none

Disclosures

The authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data Statement

Data available upon request