Diagnosis of chronic disease during admission for emergency general surgery: a portal to healthcare

Abstract

Background

Although nearly half of the US population has a chronic disease, many remain undiagnosed, leading to significant morbidity and mortality. Sociodemographic factors affect access to preventative healthcare, increasing rates of undiagnosed chronic disease. We hypothesize that emergency general surgery (EGS) is an important access point into the healthcare system and sought to characterize factors impacting the new diagnosis of chronic disease during admission for EGS.

Methods

This was a Level III retrospective cohort study conducted at a single, academic institution. Patients undergoing EGS intervention, including colectomies, cholecystectomy, hernia repair, and peptic ulcer surgeries, were identified during 2018–2019. Univariate descriptive statistics and bivariate analyses were conducted, with χ² tests for categorical variables and Mann-Whitney U tests for continuous variables. We finally conducted a multivariable logistic regression to identify important factors related to the diagnosis of a new chronic disease.

Results

A total of 978 patients were included. Of these, 42.7% received a new diagnosis of chronic disease during their EGS admission. The most common diagnoses were gastroesophageal reflux disease (n=120), obesity (n=116), type 2 diabetes (n=60), and hypertension (n=48). No significant associations were found with sociodemographic factors or prior healthcare visits. Length of stay was significantly longer for those with new diagnoses (p<0.001).

Conclusions

Hospital admissions for EGS present a critical opportunity to identify undiagnosed chronic conditions, particularly in patients with limited healthcare access. Length of stay was associated with an increased likelihood of diagnosis. These findings suggest that emergency surgical care can serve as a gateway to preventive care. This study provides Level III evidence of the role of emergency general surgery in chronic disease diagnosis.

Level of evidence

III

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Chronic diseases are often underdiagnosed in the U.S., especially among individuals with poor access to primary care, and their delayed recognition contributes to worse health and surgical outcomes.

WHAT THIS STUDY ADDS

• We found that over 40% of patients undergoing emergency general surgery were newly diagnosed with a chronic disease during hospitalization, highlighting the frequency of missed diagnoses even among insured populations.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings underscore the potential for emergency surgical encounters to serve as a critical point for chronic disease detection and intervention, expanding the role of acute care surgeons in upstream prevention.

Introduction

More than half of adults in the USA suffer from at least one chronic disease.¹ Chronic diseases, such as cardiovascular disease, obesity, and type 2 diabetes, place a significant burden on both individuals and the healthcare system. They lead to higher costs,² increased mortality,³ decreased quality of life and functional status,⁴ and increased utilization of healthcare services.⁵ Chronic diseases are the leading cause of death and disability in the USA, contributing to more than 85% of the nation’s US$4.1 trillion annual healthcare expenses.²

Timely and accurate diagnosis of chronic diseases is essential due to the significant medical and surgical implications of undiagnosed conditions.⁶ Patients with undiagnosed diabetes exhibit higher rates of chronic kidney disease and retinopathy,⁷ whereas those with undiagnosed obstructive sleep apnea (OSA) face an increased risk of cardiovascular disease.⁸ From a surgical standpoint, individuals with multiple comorbidities are more likely to encounter poor outcomes during emergency general surgery (EGS).⁹ Additionally, undiagnosed OSA is an independent risk factor for postoperative complications.¹⁰

Numerous initiatives have been launched to address the growing burden of chronic diseases worldwide. The WHO launched the Innovative Care for Chronic Conditions initiative in 2002, and the Centers for Disease Control and Prevention (CDC) has developed a National Center for Chronic Disease Prevention and Health Promotion to enhance the prevention and management of chronic diseases.^{11 12} Despite these efforts, many chronic diseases remain undiagnosed.¹³ In the developed world, it is estimated that 90% of cases of depression, osteoarthritis, and OSA go undiagnosed.¹⁴ The rate of underdiagnosis for type 2 diabetes is predicted to be between one-quarter and one-third,¹⁵ a trend that has persisted for almost two decades.¹⁶

Unfortunately, timely access to primary care services, where chronic diseases are often diagnosed, continues to be an issue throughout the USA.¹⁷ Over 40 million Americans do not have a consistent source of ongoing healthcare, leading to increased utilization of primary and emergent care services among insured and uninsured patients.18,20 Since patients are not receiving appropriate care to reduce disease burden, they present with more complex and severe surgical conditions, consequently worsening their prognosis. For example, patients are more likely to require emergency surgical intervention if they have not received sufficient primary care before presenting to the emergency department.²¹ This is significant, considering that there is up to a fivefold increase in mortality when a procedure is performed emergently rather than electively.²² Furthermore, with the increasing prevalence and poor maintenance of chronic diseases, patients are presenting with a greater number of poorly controlled conditions, which confers greater surgical risk, as previously described.

We think this highlights emergency care, particularly emergency surgical care, as a critical entry point into the healthcare system for both insured and uninsured patients. These trends in healthcare utilization suggest that undiagnosed chronic diseases are often first discovered in the emergency department setting. However, there is limited data describing the identification of new chronic disease diagnoses in patients presenting for EGS. Therefore, in this study, we hypothesize that EGS is an important access point for patients to enter the healthcare system, and we will characterize the factors contributing to the new diagnosis of chronic diseases during these hospital admissions. We aim to underscore the important role that acute care surgery can play in primary and secondary prevention of the long-term effects of chronic health conditions.

Methods

This study was approved by the Colorado Multiple Institutional Review Board (COMIRB) 18–1210. This is part of a larger study that seeks to examine emergent surgical procedures that have an elective option when patients are insured or insurance-eligible to investigate why healthcare coverage is not sufficient to ensure access to timely care. The CPT codes of interest were 43xxx, 44xxx, 47xxx, and 49xxx, representing procedures that may be performed either electively or emergently, including colectomies, cholecystectomies, hernias (including abdominal wall, inguinal, and intra-abdominal), and peptic ulcer surgeries. Surgical procedures like appendectomies were excluded, as they do not have an elective treatment. We included patients at our local level 1 trauma center from 2018 to 2019 (to avoid the biases of access during COVID-19), over 18 years old, who had insurance or were insurance-eligible, and had undergone an emergency general surgery procedure that had an elective treatment. We excluded prisoners or wards of the state.

Three authors (CGV, DMG, CH) reviewed patient charts to analyze when these operations were performed emergently. The larger database includes over 3000 patients who underwent both emergent and elective procedures. This study focuses solely on the emergent cases, with 978 records meeting the inclusion criteria. In our system, these patients present to our main emergency department from home, one of our free-standing emergency departments around the metro area, or from one of our connected outpatient clinics.

The primary outcome variable was the diagnosis of a new chronic disease during the surgical encounter of interest. This was defined as any chronic condition added to the patient’s problem list in the electronic health record (EHR) during the emergency encounter dates. The authors followed the CDC definition of chronic disease, which includes conditions lasting 1 year or more and requiring ongoing medical attention. Diagnoses that could be acute or chronic, such as anemia, were excluded if the coding suggested an acute process. In cases where the distinction was unclear, team members used their clinical judgment. The timing of the diagnosis was analyzed based on when it was formally recorded in the problem list. The variable was measured by counting new diagnoses in the encounter (range: 1–8) and then dichotomized as a “Yes” or “No” to compare groups. Most patients with new diagnoses had only one new condition.

Demographic information encompassed nominal categories for gender, race, ethnicity, insurance type categorized as Private, Medicaid, Medicare, or Other (such as state assistance programs), and None, and surgical procedure (eg, hernia, stomach, cholecystectomy, and intestines/colon). Age was measured continuously. Race was classified using US Census categories, with categories containing low counts (less than 10) collapsed into one “Other” category to protect confidentiality (ie, Asian, Native American, and Alaska Native). The primary language of patients was categorized as English, Spanish, or Other. We also identified whether they needed a translator (yes or no) and whether they had a primary care provider (PCP), and we recorded continuous values for the number of prior emergency department (ED) and PCP visits for their primary concern. Additionally, these variables were dichotomized into 0 visits versus >1 visit for comparison.

All statistical analyses were conducted using Stata statistical software: release 18 (College Station, Texas, USA).²³ P values <0.05 were considered statistically significant. We performed univariate demographic statistics (table 1) and bivariate analyses to check for associations (table 2). χ² tests of association were used for categorical variables, and Mann-Whitney U tests were used for continuous variables when comparing two groups. A multivariable logistic regression was performed using the diagnosis of a new chronic disease as the dependent variable, with no diagnosis considered a failure (coded as 0). The logistic regression included three models. In the logistic regression, Model 1 included only insurance, Model 2 added demographic variables, and Model 3 incorporated hospital and medical history variables (table 3).

Table 1. Descriptive statistics and patient demographics.

N=978	n (%)
Diagnosed with a new chronic disease
No	562 (57.5)
Yes	416 (42.5)
Number of new diagnoses, median (IQR)	1.00 (1.00–2.00)
Age, median (IQR)	48.00 (34.00–64.00)
Sex assigned at birth
Female	540 (55.2)
Race
White	488 (51.7)
Black	153 (16.2)
Other	302 (32.0)
Ethnicity
Hispanic or Latino	289 (30.4)
Primary language
English	794 (83.1)
Spanish	123 (12.9)
Other*	39 (4.1)
Type of insurance
Private	190 (19.4)
Medicaid	221 (22.6)
Medicare	247 (25.3)
Other (including state assistance)	22 (2.2)
None	298 (30.5)
Length of stay, mean (SD)	7.72 (12.20)
Type of operation
Hernia†	85 (8.7)
Stomach	33 (3.4)
Cholecystectomy	519 (53.1)
Intestines/colon	341 (34.9)
Prior ED visits for chief concern
0	690 (74.8)
≥1	232 (25.2)
Prior PCP visits for chief concern
0	758 (88.2)
≥1	101 (11.8)

^*Over 20 languages represented in the dataset including Arabic, Vietnamese, Karen, Somali, and several others.

^†Classified by primary procedure code—this may have included bowel resection.

AIC, Akaike Information Criterion; ED, emergency department; PCP, primary care provider.

Table 2. Bivariate analysis of diagnosis of chronic disease.

	No, N (%)	Yes, N (%)	P value
n=978	562 (57.5)	416 (42.5)
Age	45.50 (33.00–63.00)	51.00 (37.00–65.00)	0.002
Sex assigned at birth
Female	321 (57.1)	219 (52.6)	0.164
Race
White	273 (50.3)	215 (53.8)	0.401
Black	95 (17.5)	58 (14.5)
Other	175 (32.2)	127 (31.8)
Ethnicity
Hispanic or Latino	175 (31.9)	114 (28.4)	0.244
Primary language
English	453 (82.2)	341 (84.2)	0.453
Spanish	77 (14.0)	46 (11.4)
Other*	21 (3.8)	18 (4.4)
Type of insurance
Private	107 (19.0)	83 (20.0)	0.123
Medicaid	127 (22.6)	94 (22.6)
Medicare	128 (22.8)	119 (28.6)
Other	12 (2.1)	10 (2.4)
None	188 (33.5)	110 (26.4)
Length of stay, mean (SD)	6.18 (8.7)	9.81 (15.5)	<0.001
Type of operation
Hernia†	50 (8.9)	35 (8.4)	0.091
Stomach	17 (3.0)	16 (3.8)
Cholecystectomy	316 (56.2)	203 (48.8)
Intestines/colon	179 (31.9)	162 (38.9)
Prior ED visits for chief concern
0	394 (73.9)	296 (76.1)	0.453
≥1	139 (26.1)	93 (23.9)
Prior PCP visits for chief concern
0	426 (86.4)	332 (90.7)	0.053
≥1	67 (13.6)	34 (9.3)

^*Over 20 languages represented in the dataset including Arabic, Vietnamese, Karen, Somali, and several others.

^†Classified by primary procedure code—this may have included bowel resection.

ED, emergency department; PCP, primary care provider.

Table 3. Logistic regression for odds of new diagnosis.

	Model 1		Model 2		Model 3
	OR	95% CI	OR	95% CI	OR	95% CI
Insurance (private ref.)
Medicaid	0.954	(0.645 to 1.41)	1.06	(0.706 to 1.59)	1.21	(0.770 to 1.89)
Medicare	1.199	(0.820 to 1.75)	1.03	(0.669 to 1.59)	1.06	(0.666 to 1.70)
Other	1.074	(0.443 to 2.61)	1.24	(0.496 to 3.07)	1.83	(0.601 to 5.57)
None	0.754	(0.521 to 1.09)	0.77	(0.518 to 1.13)	0.83	(0.543 to 1.25)
Age			1.01	(0.999 to 1.02)	1.01	(0.996 to 1.02)
Sex (female ref.)			1.11	(0.850 to 1.45)	1.04	(0.768 to 1.39)
Race (white ref.)
Black			0.76	(0.522 to 1.12)	0.78	(0.517 to 1.17)
Other			1.22	(0.771 to 1.92)	1.25	(0.737 to 2.13)
Ethnicity (Hispanic ref.)
Not Hispanic or Latino			1.233	(0.772 to 1.97)	1.25	(0.718 to 2.16)
Prior ED visits CC (0 ref.)
≥1 visits					1.11	(0.786 to 1.57)
PCP visits for CC (0 ref.)
≥1 visits					0.64	(0.401 to 1.02)
Operation (hernia ref.)
Stomach					1.10	(0.43 to 2.80)
Cholecystectomy					1.3	(0.74 to 2.27)
Intestines/colon					1.18	(0.67 to 2.08)
Length of stay					1.02	(1.01 to 1.04)*
Intercept	0.78	(0.58 to 1.03)	0.37	(0.18 to 0.78)*	0.30	(0.10 to 0.87)†
Number of observations	978		943		817
AIC	1336.64		1290.24		1120.42
Area under the curve	0.55		0.57		0.59

Dependent variable: Receiving a new diagnosis during encounter for emergency surgery: Model 1: Insurance at admission; Model 2: Including demographic variables; Model 3: Adding in surgery type, and visits to ED and PCP for same visit as well as length of stay.

^*p<0.01.

^†*p<0.05.

CC, chief concern; ED, emergency department; PCP, primary care provider; ref., reference.

Results

We included a total of 978 patients who underwent emergency surgery at our level 1 trauma center between 2018 and 2019. Table 1 shows the preoperative demographic and baseline clinical data of the patient cohort. The median age was 48 (IQR 34–64). Females comprised 55.2% of the patient cohort. The majority of patients were white (51.7%) and black (16.2%), whereas 32% were classified as other, and 30% of patients were identified as Hispanic. Additionally, 69.5% had insurance prior to their hospital visit, whereas 30.5% did not. Among those who had insurance, the majority were covered by Medicare (25.3%) or Medicaid (22.6%).

Overall, 42.5% of patients had a new diagnosis of chronic disease during their hospital admission for emergency surgical care. The number of new diagnoses ranged from 1 to 8. Gastroesophageal reflux disease (GERD) was the most diagnosed chronic disease (n=120), followed by obesity (n=116), type 2 diabetes (n=60), and primary hypertension (n=48).

Table 2 presents the bivariate analysis comparing patients with and without a new chronic disease diagnosis. There was no statistically significant difference in sex, race, ethnicity, or primary language between groups, and insurance types were similar. There was no significant difference in the number of prior ED visits or PCP visits between patients who did and did not receive a new chronic disease diagnosis. However, patients who received a new diagnosis were significantly older (median age 51 (37–65) vs 45 (33–63); p=0.002), and a new chronic disease diagnosis was associated with a longer length of hospital stay. Patients with a new diagnosis had a mean length of stay of 9.8 days (SD 15.5), compared with 6.18 days (SD 8.74) for those without a new diagnosis (p<0.001). Surgical procedure type was not significantly associated with the likelihood of receiving a new diagnosis (p=0.09). Dichotomous (table 2) and multiple category analysis of ED visits (not shown) showed a similar pattern. When disaggregating the ED visits to 0, 1–2, and >2 visits to the ED and/or PCP, these associations did not change.

Findings from the logistic regression analysis (table 3) were consistent with the bivariate analysis. Each year increase in age was associated with a 1.01 times higher likelihood of receiving a new chronic disease diagnosis. Length of stay remained a significant predictor with each additional hospital day increasing the odds of a new diagnosis by 2% (OR=1.02, p<0.01, table 3).

Discussion

Our study demonstrates that emergency surgical admissions can be a critical opportunity for diagnosing previously undetected chronic diseases. Notably, over 40% of our patients received a new chronic disease diagnosis during their hospital admission—regardless of sex, race, ethnicity, or insurance type—supporting external data that link undiagnosed conditions with emergent surgical care.^{13 21} Although previous studies have found associations between sociodemographic factors (eg, younger age, minority status, limited healthcare access) and undiagnosed conditions,724,26 our findings indicate broader issues of healthcare access in the USA.²⁷

The expansion of Medicaid eligibility under the Affordable Care Act increased insurance coverage for an additional 16 million people during the following decade²⁸; however, significant barriers—such as inconvenient clinic hours, provider shortages, transportation challenges, and long wait times—persist.²⁷ Consequently, even with insurance, Americans face obstacles to accessing primary care, which increases the likelihood of undiagnosed chronic diseases. This is consistent with our findings, which showed no difference in the type of insurance between patients with and without new chronic disease diagnoses. In fact, those diagnosed with a new chronic disease were more likely to be insured. This suggests that, independent of insurance status, many patients encounter challenges in accessing primary care, as evidenced by the high proportion of first-time chronic disease diagnoses in the acute care setting. These findings highlight additional social determinants of health that render certain populations more vulnerable and underscore that insurance alone is insufficient for equitable care. Increased efforts are needed to address other modifiable factors, such as the availability of childcare, access to paid time off, and transportation concerns.

Our data also highlights a unique opportunity for primary (interventions that prevent disease before it occurs) and secondary prevention (early disease detection to reduce severity) in the emergency surgical setting. Although current post-discharge healthcare navigation primarily addresses tertiary prevention (occurring after a disease has already been diagnosed and aiming to prevent further complications), this leaves a significant gap for primary and secondary prevention, where the role of the acute care surgery service model could be incredibly informative. Acute care surgeons frequently make initial diagnoses of severe diseases—such as colon cancer, where up to one in six patients may present to the emergency surgical care setting previously undiagnosed.²⁹ This not only increases the likelihood that the cancer will be detected at an advanced stage but also raises the patient’s risk for emergency hospital admissions, obstruction, or perforation compared with if they had been up to date with colorectal cancer screening in a primary care setting.³⁰ Although it would be ideal for all patients to have access to timely primary care and not present in such a severe condition, this is not the reality given the limitations in our healthcare system. Therefore, we wish to highlight the preventative model proposed by Peck et al³¹ and other literature that demonstrates the importance of a collaborative approach in addressing the growing burden of chronic disease throughout the USA.³² The acute care surgery team, incorporating providers and social work support, has a unique opportunity to engage patients who might otherwise lack adequate healthcare access. For instance, by screening patients with a hemoglobin A1c level or ensuring that a patient has a primary care appointment scheduled at discharge, the acute care surgeon can make a substantial contribution to reducing a patient’s chronic disease burden and improving surgical outcomes.

In this study, GERD and obesity were the two most diagnosed conditions. Documenting these conditions in the outpatient setting is challenging. For instance, the clinical presentation of GERD varies widely and may not correlate with confirmatory diagnostic tests, which themselves rely on multiple, often cumbersome modalities that are difficult to obtain outside of a hospital setting.³³ Similarly, obesity—the second most common diagnosis in our study—presents its own challenges. Historically underdiagnosed, studies indicate that fewer than 2% of patients meeting clinical criteria for obesity receive a formal diagnosis.³⁴ This underdiagnosis may be partly due to the recent recognition of obesity as a chronic disease and the fact that patients presenting for emergency general surgery, who often lack consistent outpatient care,^{13 21} may not have had the opportunity for a formal obesity diagnosis.³⁵ Furthermore, several of the conditions most frequently identified in our study—GERD, obesity, and type 2 diabetes—are either symptomatic or readily observable by clinicians. In addition, hypertension, the fourth most common diagnosis, is routinely monitored through daily vital sign assessments in the inpatient setting. These factors underscore why these conditions may have been more likely to be diagnosed in our study, where comprehensive evaluations and routine monitoring facilitate the identification of conditions that might otherwise be undiagnosable without consistent outpatient care.

In our study, length of stay (LOS) was significantly different between patients who received a new diagnosis of chronic disease and those who did not. Patients with a new diagnosis of chronic disease had an average LOS about 3 days longer than those who did not receive a new diagnosis, and the longer the stay, the more likely they were to receive a diagnosis. This is consistent with data from other studies that have demonstrated that increased LOS in the hospital is associated with more diagnoses of chronic disease on discharge.³⁶ Although our study only supports an association and not a causal relationship, it is understandable that more diagnoses of chronic diseases are discovered during longer hospitalizations, as patients are likely to have undergone more laboratory study work, diagnostic studies, and had more opportunities to share subjective symptoms. However, it is also possible that patients with more chronic diseases require longer hospitalization stays due to their comorbidities and associated poor health. Although causality cannot be analyzed, our data still adds to the body of literature implicating chronic diseases with increased morbidity and healthcare costs.

The study’s strengths include a diverse patient population and the inclusion of data from the Medicaid expansion period, which allows us to investigate the impact of other factors outside of insurance coverage amid potential disparities in social determinants of health. Limitations of the present study include its retrospective nature and reliance on a single-center dataset, which limits the generalizability of findings. The method used to analyze new diagnoses during the study period was based on the date of diagnosis recorded in the EHR for our hospital, which could potentially misclassify a prior diagnosis as new, although these are entered and reviewed by a team of nurse analysts after each hospitalization to assure that no diagnoses are missed for hospital benchmarking. Our study did not incorporate a review of medication reconciliation data or prior prescriptions to confirm whether patients were already receiving pharmacologic treatment for the newly documented chronic disease, as these medications cannot be ordered in our system without the diagnosis and are reviewed by the analyst team, so these should be covariates. We are currently auditing our data to confirm that this is true. This limitation may result in an overestimation of newly diagnosed conditions, as some patients may have been receiving treatment prior to formal documentation in the problem list. Additionally, diagnoses made at institutions outside our healthcare system were not included, potentially underestimating prior diagnoses. This would also impact the power of the measure for previous primary care and emergency department visits, although most of our EGS patients come from the area in close proximity to our hospital and are treated primarily in our system. Our system has also recently gained more access to other health systems’ data, which is now populating directly to our EHR, so we will be able to refine how we look at this.

Another limitation is in defining certain conditions included in the CDC definitions of chronic disease. For example, the diagnosis of anemia is particularly challenging, as it can be unclear whether it is acute anemia secondary to acute blood loss or chronic anemia due to a blood disorder or other chronic disease, although the procedures we examined are not typically associated with large blood loss. We are already planning an additional study that examines laboratory studies and interventions, such as transfusion and hemoglobin A1C, to better refine how we diagnose these conditions. The availability of data was restricted to the years 2018–2019, which does not fully capture longitudinal trends. Our future research aims to address this limitation by extending the data collection period and incorporating qualitative and mixed methods techniques.

Another important limitation is the relatively short length of stay of some of the EGS conditions. We found that increasing length of stay was associated with diagnosing these chronic conditions, which is not surprising because we had more opportunity to identify these conditions through laboratory studies and monitoring. We are likely missing diagnoses in patients who are discharged quickly, thus potentially underestimating the extent of the problem. We are also currently reviewing our data to assess if we can analyze if those who did not have insurance obtained it afterward as a result of the admission, or if those paid through emergency Medicaid converted to regular Medicaid.

Conclusion

In a large sample of patients undergoing emergency general surgery at a single institution we found that 42.7% were diagnosed with a new chronic disease during admission, and increased LOS was associated with an increased number of diagnoses. Americans are experiencing barriers that are prohibiting access to timely primary care services and are increasingly presenting for emergency surgical care as a result. Thus, emergency surgical care has become an important portal to healthcare. We think that this offers an opportunity for health systems to leverage the expertise of trauma/acute care surgeons in all stages of prevention and intervention to improve healthcare outcomes.