Abstract
Introduction
Studies suggest that many emergency department (ED) visits and hospitalizations for patients with cancer may be preventable. The Centers for Medicare & Medicaid Services has implemented changes to the hospital outpatient reporting program that targets acute care in-treatment patients for preventable conditions. Oncology urgent care centers aim to streamline patient care. Our cancer center developed an urgent care center called the direct referral unit in 2011.
Methods
We abstracted visits to our adjacent hospital ED and direct referral unit from January 2014 to June 2018. Patient demographics, cancer and visit diagnoses, visit charges, and 30-day therapy utilization were assessed.
Results
An analysis of 13 114 visits demonstrated that increased direct referral unit utilization was associated with decreased monthly ED visits (P < .001). Common direct referral unit visit diagnoses were dehydration, nausea and vomiting, abdominal pain, and fever. Patients receiving active cancer treatment more frequently presented to the direct referral unit (P < .001). The average charges were $2221 for the direct referral unit and $10 261 for the ED.
Conclusion
The association of decreased ED visits with increased direct referral unit utilization demonstrates the potential for urgent care centers to reduce acute care visits. Many patients presented to our direct referral unit with preventable conditions, and these visits were associated with considerable cost savings, supporting its use as a cost-effective method to reduce acute care costs.
In 2022, more than 18 million people with a history of invasive cancer were alive, with an estimated 1.9 million new cases diagnosed (1,2). Compared with the general population, oncology patients have higher rates of emergency department (ED) visits and inpatient admissions (3,4). Acute hospital care is a chief driver of regional spending variation and out-of-pocket expenses in Medicare beneficiaries with cancer (5,6). Moreover, recent studies suggest that a substantial proportion of ED visits and hospitalizations for patients with cancer are potentially preventable (7,8).
Given the high rates of utilization, reducing unplanned acute care represents a potential strategy for cost containment in oncology, particularly given the cost savings demonstrated in the general population (9-11). The Centers for Medicare & Medicaid Services (CMS) has proposed new payment and health-care delivery models designed through the institution of the Oncology Care Model. One major aim of this program was to reduce avoidable acute care visits (12,13). CMS has also added an oncologic quality measure to its hospital outpatient reporting program—OP-35—which measures the risk-adjusted rates of ED visits and inpatient admissions for patients who have received chemotherapy within the previous 30 days. Specifically, it targets preventable conditions, including anemia, dehydration, neutropenia, diarrhea, pain, emesis, pneumonia, fever, and sepsis (14,15).
Oncology-specific urgent care centers have recently become more prevalent at cancer centers nationally (16-26). Studies have demonstrated decreased ED utilization in association with oncology urgent care centers; nonetheless, the impact on cost and hospitalizations remains controversial (20,23,27). Although substantial operating costs are associated with running an outpatient oncology-specific urgent care center, the ability to divert patients from the ED and avoid hospitalization is thought to reduce overall spending. Gould Rothberg et al. (23) established a cancer-specific urgent care center using Oncology Care Model funding and were able to recover expenses related to the center through regular outpatient billing practices.
Our institution, a comprehensive cancer center, developed an urgent care center called the direct referral unit in July 2011. We conducted a single-institution retrospective study examining health-care utilization following the creation of the direct referral unit to further explore the role that urgent care centers play in acute care.
Methods
Patient population
We identified patients at our institution with visits to the direct referral unit or our associated ED, which serves as a hub for admission to the institution. The direct referral unit is a 4-bed urgent care center primarily open on weekdays from 8:00 am to 5:00 pm, with Saturday hours available until 2017. The direct referral unit is staffed by internal medicine–certified hospitalists and nurse practitioners who are familiar with common cancer complications and treatment toxicities. A patient’s primary oncologist, radiation oncologist, or surgeon can refer their patients to the direct referral unit pending bed availability and discussion with direct referral unit staff. Alternatively, patients can be sent to the unit from an office visit if additional care is needed. In the direct referral unit, patients can undergo laboratory testing, imaging, and basic treatment (fluids, antibiotics), and they can be admitted directly to the hospital, if necessary. Although there are no specific criteria for acceptance to the direct referral unit, higher-acuity patients potentially requiring urgent intervention (cardiac catheterization, stroke intervention, etc) are generally triaged to the ED. Patients who remain in the unit after hours are discharged to home or directly admitted to the cancer hospital for further acute care; transfers to the ED are uncommon. If no direct referral unit bed is available or the patient presents after hours, the patient is referred to our associated hospital ED for evaluation.
All patients in the study received chemotherapy, immunotherapy, radiation therapy, or surgery at Fox Chase Cancer Center between January 2014 and June 2018 and had a minimum of 3 outpatient visits. Patients with hematologic malignancies were excluded because they do not readily utilize the direct referral unit and are generally admitted to a different affiliate hospital. Data were obtained by querying the electronic health record system—either Epic (Epic Systems Corporation, Verona, WI) or Soarian (Oracle Cerner, North Kansas City, MO) through the institution’s data warehouse. Patient demographics, including age, sex, race, ethnicity, and insurance provider, were also abstracted. Sex, race, and ethnicity were all self-reported. “Other” under race or ethnicity referred to patients who did not identify with other listed race and ethnicity categories. Cancer diagnoses were obtained by linking the dataset to the national tumor registry using the International Statistical Classification of Diseases, 0-3 coding. The study was approved by the institutional review board of our institution (No. 19-9003), with a waiver of patient consent.
ED and urgent care visits
The primary outcome of the study was utilization of the ED following the institution of the direct referral unit. Secondary outcomes included inpatient admission rates, 30-day health-care and therapy utilization, and visit charges. We compared visit rates to the ED and direct referral unit on monthly and yearly bases to assess trends at both the population and patient levels. Each ED and direct referral unit visit was associated with a unique visit ICD diagnosis code. International Classification of Diseases, Ninth Revision (ICD-9) was used until October 1, 2015, after which we transitioned to ICD-10. Two authors (C.D. and P.W.) reviewed the visit codes for duplicates and compiled a list of the most common visit diagnoses to assess trends.
Inpatient admissions
We analyzed hospitalization rates from the ED and direct referral unit to the inpatient hospital at the cancer center. The facilities at our institution include a 100-bed inpatient unit, where our patient population is preferentially admitted. Hospitalization rates for the study did not include admissions to outside institutions. We also determined the average length of stay (LOS) for all inpatient admissions.
Visits or admissions within 30 days
To assess the impact of the ED and direct referral unit on later acute-care episodes, we identified patients with an index visit to the ED or direct referral unit who then had another ED or direct referral unit visit or an inpatient admission to the cancer center within 30 days of their original visit. Inpatient admission within 30 days of an ED or direct referral unit visit did not include patients who were admitted the day of the index visit. This data point was calculated as the percentage of total visits to assess the differences. We also assessed receipt of chemotherapy, immunotherapy, and radiation therapy within 30 days of the index visit to assess active treatment patients. Specifically, this metric quantified the number of visits to the ED or direct referral unit that occurred within 30 days of treatment.
Health-care cost
The average charges associated with ED and direct referral unit visits from January 2014 to June 2018 were calculated. These values reflected the recorded cost of these visits rather than the payment responsibility of individual patients, given the variability in insurance coverage and plans. Insurance-negotiated charges for the direct referral unit are often less than negotiated charges for the ED (28).
Statistical analysis
We characterized the data using means (SD), medians, and proportions. Tests for statistical significance were 2-sided, and significance was defined as a P value less than .05. There was no primary independent variable. We investigated relationships between ED and direct referral unit utilization in an exploratory manner, including the following covariates: age, sex, race, ethnicity, year, cancer diagnosis, insurance, and day of visit. The primary outcome was utilization of the direct referral unit and ED at both the visit and person levels. For examination of demographics between those who utilized the ED vs the direct referral unit, we used χ2 and t tests for unadjusted differences. For adjusted differences, we used random-effects logistic regressions, in which a binary ED vs direct referral unit indicator was the response and the demographic and clinical characteristics were entered as covariates (either continuous or categorical via dummy indicators). In these regressions, the random intercept accounted for repeated measures within patients.
Next, we modeled the relationship between ED and direct referral unit utilization temporally in 2 ways at the population level and 2 ways at the patient level. First, we examined the population-level monthly counts with a Poisson regression, not accounting for yearly effects, and a Poisson regression accounting for yearly effects via dummy indicators and clustering within year using cluster-corrected SEs. We examined with and without year because year was a key confounder of inferences. The cluster-corrected SEs accounted for extra Poisson variation. Cluster-corrected SEs are more computationally stable for generalized linear models than random-effects models.
Second, we modeled the yearly number of ED visits (dependent variable) vs direct referral unit visits (independent variable) using a Poisson regression at the individual level in which we do not account for year and one in which we included year via dummy indicator variables, with cluster-corrected SEs to account for repeated measurements within years. As a sensitivity analysis, we also used cluster-corrected SEs to account for repeated patient visits (ie, clustering by patient) (29).
We used adjusted random-effects logistic and linear regressions of in-patient admission and hospital utilization outcomes (dependent variables), with an indicator variable of ED vs direct referral unit (independent variable). A random intercept accounted for repeated measures within an individual, and we adjusted for age, sex, race, year, cancer diagnosis, insurance status, and day of week in the models. We used the random-effects models to calculate marginal means and proportions. We used Stata MP version 15 (StataCorp LP, College Station, TX) for all analyses. All data were obtained from electronic health record data from the institution’s data warehouse.
Results
Patient population and visits
Patient characteristics at the visit level are summarized in Table 1. Many patients had multiple visits. Visits to the ED and direct referral unit were more common in women (direct referral unit = 55%, ED = 58%), and the majority of patients were White (direct referral unit = 68%, ED = 66%) and non-Hispanic (direct referral unit = 99%, ED = 98%). In a full model, African American and Hispanic patients were more likely to utilize the ED than the direct referral unit (odds ratio [OR] = 1.63 compared with White, 95% confidence interval [CI] = 1.36 to 1.96, P < .001; OR = 4.42 relative to non-Hispanic, 95% CI = 1.51 to 12.90, P < .001). The visit use per day of the week is presented in Table 1. Direct referral unit utilization was most frequent on Mondays (22%), whereas ED use occurred most often on Sundays (17%). An average of 137 patients were seen in the direct referral unit per month (range = 83-192). On average, patients in the sample had 1.01 (SD = 1.63) ED visits and 1.30 (SD = 1.55) direct referral unit visits. Patients were more likely to visit the direct referral unit multiple times; 38.6% of the sample had 1 direct referral unit visit, 15.4% had 2 unit visits, and 14.4% had 3 or more unit visits, whereas 33.8% had 1 ED visit, 11.3% had 2 ED visits, and 9.6% had 3 or more ED visits. Commercial insurance (48%) use was most common, followed by Medicare (35%) and Medicaid (15%). Gastrointestinal (GI) (27%), thoracic (16%), and breast (12%) malignancies were most frequently observed. Medicaid patients were more likely to use the ED than the direct referral unit in the full model (OR = 1.36 vs Medicare, 95% CI = 1.09 to 1.70, P = .007; OR = 1.28 compared with commercial insurance, 95% CI = 1.05 to 1.55, P = .01). Common visit diagnoses are listed in Supplementary Table 1 (available online). The most common visit diagnoses to the direct referral unit were dehydration, nausea and vomiting, abdominal pain, fever, shortness of breath, fatigue, diarrhea, cellulitis/rash, constipation, and anemia. In the ED, the most common diagnoses included abdominal/pelvic symptoms, urinary tract disorders, back/limb pain, fever, dehydration, nausea and vomiting, chest pain, shortness of breath, cellulitis/rash, and pneumonia.
Table 1.
Patient demographics and visit trends
| Emergency department | Direct referral unit | Total | ||
|---|---|---|---|---|
| (n = 5737) | (n = 7377) | (N = 13 114) | P | |
| Age at index visit, y | .16 | |||
| Mean (SD) | 62.5 (14.2) | 62.8 (12.2) | 62.7 (13.1) | |
| No. (% nonmissing) | 5737 (100.0) | 7377 (100.0) | 13114 (100.0) | |
| Sex, No. (%) | <.001 | |||
| Female | 3326 (58.0) | 4036 (54.7) | 7362 (56.1) | |
| Male | 2411 (42.0) | 3341 (45.3) | 5752 (43.9) | |
| Race, No. (%) | <.001 | |||
| Asian | 221 (3.9) | 383 (5.2) | 604 (4.6) | |
| African American | 1116 (19.5) | 1028 (13.9) | 2144 (16.3) | |
| White | 3803 (66.3) | 5024 (68.1) | 8827 (67.3) | |
| Othera | 597 (10.4) | 942 (12.8) | 1539 (11.7) | |
| Ethnicity, No. (%) | <.001 | |||
| Non-Hispanic | 5615 (97.9) | 7360 (99.8) | 12975 (98.9) | |
| Hispanic | 122 (2.1) | 17 (0.2) | 139 (1.1) | |
| Index visit year, No. (%) | <.001 | |||
| 2014 | 961 (16.8) | 1900 (25.8) | 2861 (21.8) | |
| 2015 | 1185 (20.7) | 1728 (23.4) | 2913 (22.2) | |
| 2016 | 1251 (21.8) | 1655 (22.4) | 2906 (22.2) | |
| 2017 | 1510 (26.3) | 1448 (19.6) | 2958 (22.6) | |
| 2018 | 830 (14.5) | 646 (8.8) | 1476 (11.3) | |
| Cancer diagnosis, No. (%) | <.001 | |||
| Lip, oral cavity, pharynx, head, neck (C00–C14, C30-C32) | 232 (4.7) | 288 (4.0) | 520 (4.3) | |
| Digestive organs (C15-C26) | 1185 (24.3) | 2065 (28.7) | 3250 (26.9) | |
| Respiratory system and intrathoracic organs (C33-C39) | 666 (13.6) | 1205 (16.7) | 1871 (15.5) | |
| Breast (C50) | 698 (14.3) | 769 (10.7) | 1467 (12.1) | |
| Genitourinary (C51-58, C60-63, C64-68) | 1483 (30.4) | 1938 (26.9) | 3421 (28.3) | |
| Secondary, endocrine, skin, soft, bone (C43-C44, C45-C49, C76-C80, C73-C75, C40-C41) | 299 (6.1) | 467 (6.5) | 766 (6.3) | |
| Independent (primary) multiple sitesb (C97) | 322 (6.6) | 465 (6.5) | 787 (6.5) | |
| Number missing | 852 | 180 | 1032 | |
| Day of week, No. (%) | <.001 | |||
| Monday | 926 (16.1) | 1607 (21.8) | 2533 (19.3) | |
| Tuesday | 728 (12.7) | 1422 (19.3) | 2150 (16.4) | |
| Wednesday | 707 (12.3) | 1337 (18.1) | 2044 (15.6) | |
| Thursday | 685 (11.9) | 1247 (16.9) | 1932 (14.7) | |
| Friday | 798 (13.9) | 1371 (18.6) | 2169 (16.5) | |
| Saturday | 905 (15.8) | 392 (5.3) | 1297 (9.9) | |
| Sunday | 988 (17.2) | 1 (0.0) | 989 (7.5) | |
| Insurance carrier, No. (%) | <.001 | |||
| Medicaid | 883 (18.2) | 1006 (13.6) | 1889 (15.5) | |
| Medicare | 1591 (32.9) | 2619 (35.5) | 4210 (34.5) | |
| Private insurance | 2265 (46.8) | 3563 (48.3) | 5828 (47.7) | |
| Miscellaneous | 103 (2.1) | 188 (2.5) | 291 (2.4) | |
| Number missing | 895 | 1 | 896 |
“Other” refers to participants who did not identify with other listed race or ethnicity categories.
“Multiple primary” refers to patients who have 2 independent malignancies at different sites.
The 2014-2018 cohort contained 5657 patients and 13 114 visits, including 7377 direct referral unit visits and 5737 visits to the ED (Table 1). Overall, ED utilization increased over time, consistent with trends in the local area (30). For the population-level analysis adjusted by year, there was no difference between ED and direct referral unit visits, but we found that the monthly number of ED visits was inversely related to the monthly number of direct referral unit visits (Figure 1) (relative risk [RR] = 0.995, 95% CI = 0.993 to 0.997, P < .001), suggesting that an increase in direct referral unit visits leads to a decrease in ED visits. When we examined yearly counts within individuals, we noted a strong inverse relationship between the yearly number of visits to the direct referral unit with the number to the ED (RR = 0.74, 95% CI = 0.69 to 0.79, P < .001) when not accounting for year. This relationship held when adjusting for year as both a fixed-effects and robust cluster-level variable (RR = 0.75, 95% CI = 0.66 to 0.85, P < .001). The inferences were the same when we instead accounted for clustering within patients (RR = 0.75, 95% CI = 0.69 to 0.81, P < .001).
Figure 1.
Population-level monthly visit counts. ED = emergency department.
Inpatient admission and LOS
Table 2 summarizes inpatient admissions and LOS for patients admitted to the hospital from their index direct referral unit or ED visits from 2014 to 2018. Overall, there was no difference in hospitalization rates from the ED or direct referral unit in unadjusted rates as there were 1256 inpatient admissions (22%) from the ED and 1622 admissions (22%) from the direct referral unit (P = .91). After adjusting for covariates and patient-level clustering, however, inpatient admission was more common from the ED than from the direct referral unit, with marginal adjusted probabilities of 27% (95% CI = 26% to 29%) in the ED and 22% (95% CI = 21% to 23%) (P ≤ .001). Also, LOS for those admitted did not differ after adjustment (with marginal adjusted means of 5.5 days, 95% CI = 5.1 to 5.9, and 5.6 days, 95% CI = 5.2 to 5.9; P = .84).
Table 2.
Differences between the ED and direct referral unit
| Health-care/treatment utilization patterns and average visit charges | |||||
|---|---|---|---|---|---|
| Unadjusted ED, No. (%) | Unadjusted direct referral unit, No. (%) | Adjusteda ED, % (95% CI) | Adjusteda direct referral unit, % (95% CI) | Adjusted P | |
| Inpatient admission | <.001 | ||||
| Yes | 1256 (22) | 1622 (22) | 27 (26 to 29) | 22 (21 to 23) | |
| No | 4481 (78) | 5755 (78) | 73 (71 to 74) | 78 (77 to 79) | |
| Length of stay, mean (SD) [median], d | 5.55 (6.29) [4] | 5.49 (6.27) [4] | 5.48 (5.08 to 5.88) | 5.54 (5.19 to 5.89) | .84 |
| Chemotherapy within 30 d | <.001 | ||||
| Yes | 1404 (24) | 3094 (42) | 31 (20 to 33) | 39 (38 to 41) | |
| No | 4333 (76) | 4283 (58) | 69 (67 to 80) | 61 (59 to 62) | |
| Immunotherapy within 30 d | <.001 | ||||
| Yes | 545 (9) | 1098 (15) | 11 (10 to 12) | 14 (13 to 15) | |
| No | 5192 (91) | 6279 (85) | 89 (88 to 90) | 86 (85 to 87) | |
| Radiation therapy within 30 d | .42 | ||||
| Yes | 424 (7) | 781(11) | 10 (9 to 11) | 11 (10 to 11) | |
| No | 5313 (93) | 6596 (89) | 90 (89 to 91) | 89 (89 to 90) | |
| ED/direct referral unit repeat visit within 30 d | .004 | ||||
| Yes | 1333 (23) | 2017 (27) | 22 (21 to 24) | 25 (24 to 26) | |
| No | 4404 (77) | 5360 (73) | 78 (76 to 79) | 75 (74 to 76) | |
| Inpatient admission within 30 d | <.001 | ||||
| Yes | 1378 (24), 331 readmissions | 1784 (24), 364 readmissions | 27 (26 to 29) | 22 (21 to 24) | |
| No | 4359 (76) | 5593 (76) | 73 (71 to 74) | 78 (76 to 79) | |
| Average charges per visit, mean (SD) [median], $ | 10 261 (6884) [9651] | 2221 (4433) [1354] | 11 521 (11 327 to 11 716) | 1930 (1732 to 2129) | <.001 |
Adjusted for age, sex, race, year, cancer diagnosis, insurance status, and day of week. CI = confidence interval; ED = emergency department.
Health-care utilization within 30 days
Table 2 lists patients with an index visit to the ED or direct referral unit who subsequently had another visit or inpatient admission within 30 days. At the visit level, patients were more likely to have an additional ED or direct referral unit visit within 30 days after initially visiting the direct referral unit (P < .004). After adjustment, more patients from the ED were admitted to the cancer center’s inpatient ward within 30 days of an index visit (27% admitted within 30 days of the index visit to the ED compared with 22% from the direct referral unit, P < .001). Additionally, patients presenting to the direct referral unit were more likely to have received chemotherapy or immunotherapy within 30 days (P < .001) (Table 2). When adjusted, chemotherapy use was most common among the 3 modalities (direct referral unit = 39%, ED = 31%). After adjustment, there was no difference in radiation therapy use within 30 days between the 2 settings.
Health-care cost
The average unadjusted charge for a direct referral unit visit was $2221 and for an ED visit $10 261 (P < .001). This finding reflects a difference of $8040 between the care settings. The difference widened after adjustment (ED = $11 521, direct referral unit = $1930; P < .001).
Discussion
Implementing the direct referral unit at our institution was associated with decreased ED utilization. For the population-level monthly analysis, a clear inverse relationship between ED and direct referral unit monthly visits was demonstrated, suggesting that direct referral unit utilization is associated with decreased ED usage. Although this relationship was not statistically significant when adjusting for year, adjusting for year diluted this trend. When we examined yearly trends within individuals, an additional strong inverse relationship was seen between direct referral unit and ED visits, suggesting that patients are substituting ED visits for the direct referral unit.
The ED visit–level data obtained did not distinguish between cancer and non–cancer-related ED visits; thus, the number of cancer-specific ED visits may even be lower than that noted in this study. Our results support the findings of Hong et al. (20), who found a statistically significant reduction in weekday ED visits after the institution of their urgent care center. Likewise, Gould Rothberg et al. (23) also found an association between use of a cancer-specific urgent care center and decreased ED utilization. Although our study pertains to a single institution, Hong et al. assessed ED visits from a regional database that yielded a greater number of ED visits for analysis. In addition, the Hong study included only ED visits among patients with cancer on weekdays within 180 days of diagnosis, whereas our study assessed ED utilization at any time during a patient’s treatment course and included weekends. With no time-specific criteria, it is possible that our analysis captured a potentially sicker population at a higher risk of requiring acute care. Despite these needs, we still found that providing an urgent care center can decrease ED utilization, which supports further expansion of the oncology urgent care model. Larger national studies involving multiple cancer centers are needed to better assess ED utilization trends.
Our study confirms the findings of prior studies that have demonstrated that comprehensive urgent care for oncology patients represents a promising strategy to decrease acute care utilization (16-26), just as it does in a more general population (10,11). Given the rapid rise in treatment modalities with new toxicities as well as improved oncology patient survival, one would expect acute care needs to increase over time. Thus, an urgent care alternative for select patients can reduce ED utilization moving forward. Given the trends in ED utilization, the current direct referral unit model with limited weekday hours may be insufficient to address future patient volumes. ED use was most common on Sundays, a day the direct referral unit was not open, which also suggests a need to expand the direct referral unit’s hours.
Although they represented a minority of the visits, African American and Hispanic patients were more likely to seek care in the ED rather than the direct referral unit, suggesting that racial disparities may have also contributed to ED utilization in our study. Lash et al. (3) previously showed that African American race and ethnicity is a predictive factor of increased ED utilization among patients with cancer. Patients with Medicaid insurance were more likely to present to the ED, consistent with prior studies supporting higher use in underserved populations (31). Further efforts should focus on increasing oncology urgent care access in this patient population, and studies will be needed to assess the effects on care utilization.
We also found major differences between the ED and direct referral unit in our financial analysis. Direct referral unit visits were associated with considerable unadjusted charge savings of more than $8000 compared with ED visits. We did not have full access to itemized cost data to be able to analyze the differences in charges, but the reasons for this cost differential are likely multifaceted. For one, services provided in the ED are often more expensive than the same services provided in the direct referral unit, and EDs have facility fees that urgent care centers do not charge (28,32-34). Additionally, more comprehensive and expensive testing could have been ordered in the ED, but this cost savings may be a function of more efficient, targeted care provided to patients in the direct referral unit. The staff in the direct referral unit are familiar with oncologic complications and treatment toxicities; their expertise informs the care they provide. Direct referral unit clinicians also have direct access to each patient’s oncology care team, allowing for collaborative decisions on testing or imaging necessity, streamlined communication, and avoidance of unnecessary testing. It is also possible that a sicker cohort of patients who need more urgent and costly interventions may preferentially utilize the ED. ED visits among patients with cancer cause financial burden (6); thus, these cost savings strongly underscore the importance and mission of embedded urgent care centers. At the population level, these results highlight the potential for urgent care centers to facilitate cost containment and reduction in the cancer patient population.
In our study, patients with GI (27%), thoracic (15%), and breast (12%) malignancies had the highest number of visits in both settings. These findings support those of Rivera et al. (4), who found that patients with breast and lung malignancies had more frequent ED utilization. The higher prevalence of GI, thoracic, and breast cancer in the United States likely contributed to our results but is unlikely to explain the totality of the utilization patterns. Patients with GI and thoracic cancer, in particular, are at high risk of complications and treatment toxicity, contributing to their high acute care needs. Further studies should explore direct interventions in these high-risk populations to reduce health-care utilization.
Another key finding of this study was the visit patterns of treatment patients. We found that patients on active treatment with chemotherapy or immunotherapy within 30 days preferentially presented to the direct referral unit for evaluation as opposed to the ED. Overall toxicity tends to be higher in patients receiving active treatment (35-37); therefore, it is not surprising that visits were more common in this population. These results highlight the direct referral unit’s role as an effective triage center to manage treatment-related toxicities and complications in a customized and targeted fashion. Notably, many of the most common visit diagnoses to the direct referral unit (dehydration, nausea and vomiting, fever, anemia) align with the CMS list of preventable conditions under OP-35 (14,15). With patients on treatment presenting in higher numbers to the direct referral unit, there is great potential for the direct referral unit to improve the quality of care for preventable conditions.
One goal of the study was to assess differences in health-care utilization patterns between patients in oncology urgent care centers vs the ED. Our prestudy assumption was that patients presenting to the ED would have higher rates of utilization. Although the unadjusted comparison showed no differences, the adjusted results showed a higher admission rate from the ED than from the direct referral unit (adjusted P < .001) but no difference in LOS (P = .84). The higher adjusted admission rate from the ED compared with the direct referral unit may be a result of the ED seeing higher-acuity patients, though it may also be a demonstration of the direct referral unit’s ability to effectively triage patients and provide cancer-specific care, thus decreasing the rate of admission. Factors that can affect LOS are multifaceted (38-40), and the similar LOS between the ED and direct referral unit may not be a reflection of triage capabilities or effectiveness. We also found that repeat visits to the ED or direct referral unit were more common following an initial direct referral unit visit (P = .004) and that inpatient admission within 30 days was higher after an initial visit to the ED (P < .001). The increased repeat visits in the direct referral unit population may be due to inadequate initial diagnostic workup in the direct referral unit or the need for additional testing. Many of these visits as well as the increased admissions after an initial ED visit may also be due to disease progression or ongoing severe toxicities from treatment, underscoring the high care needs of this patient population. With our limited dataset, we were unable to determine what drives inpatient admissions and 30 day repeat visits. Thus, a future study investigating visit reason, acuity level or admission beyond ICD-9 and ICD-10 codes, and its relation to the cancer diagnosis would help elucidate these nuances.
Although our study has several important findings, we acknowledge certain limitations. We examined data from a single academic institution that may not reflect the general oncology population. Our original intent was to compare ED utilization both before and after institution of the direct referral unit; however, this was not possible because of issues with data capture after a change in our electronic health record system. Patient staging, treatment intent (curative vs palliative), clinical visit severity, and early patient death data were also not available for comparison. Data about patients who sought care outside our institution were not captured, and because of the deidentified and retrospective nature of the study, we do not know how often the direct referral unit was functioning at capacity and had to divert patients elsewhere. Both these factors may have affected the results. In addition, the majority of patients in our study were insured; thus, the findings may not be generalizable to an uninsured population. It is also important to note the possibility that some patients presented with symptoms unrelated to their oncologic diagnoses. Missing or inaccurate data for patients with dual diagnoses may have complicated the interpretation of information from the data warehouse or tumor registry. Finally, although we found a notable difference in charges between direct referral unit and ED visits, it is important to note that these charges do not reflect the bill the patients receive but rather the amount charged to the insurance company. Moreover, we did not have access to itemized charges to elucidate more specific reasons for the cost differential.
In our study, the establishment of an oncologic urgent care center at our institution decreased overall ED utilization and was associated with substantial cost savings. We found the direct referral unit to be an effective triage unit for our patients on active treatment, specifically for preventable conditions. Racial disparities may affect care-seeking patterns, and this requires further investigation. Our findings highlight the importance of urgent care centers in providing oncologic-specific care outside the ED. Further study of such urgent care centers should examine health-care utilization and patient outcomes on a national scale.
Supplementary Material
Acknowledgements
The funder had no role in the design of the study; the collection, analysis, or interpretation of the data; or the writing of the manuscript and decision to submit it for publication.
Prior presentations: Please note that a related abstract was presented at the American Society for Clinical Oncology Quality Care Symposium and the National Comprehensive Cancer Network Annual Meeting, but it did not include the entirety of the data analysis, which is included in the current article.
Contributor Information
Christopher D’Avella, Department of Hematology/Oncology, Fox Chase Cancer Center, Philadelphia, PA, USA; Department of Medicine, Division of Hematology-Oncology, University of Pennsylvania, Perelman School of Medicine, Philadelphia, PA, USA.
Peter Whooley, Beth Israel Deaconess Medical Center, Department of Medical Oncology, Boston, MA, USA.
Emily Milano, Children’s Hospital of Philadelphia, Philadelphia, PA, USA.
Brian Egleston, Department of Biostatistics, Fox Chase Cancer Center, Philadelphia, PA, USA.
James Helstrom, Division of Anesthesiology, Department of Surgical Oncology, Fox Chase Cancer Center, Philadelphia, PA, USA.
Kenneth Patrick, Department of Medicine, Fox Chase Cancer Center, Philadelphia, PA, USA.
Martin Edelman, Department of Hematology/Oncology, Fox Chase Cancer Center, Philadelphia, PA, USA.
Jessica Bauman, Department of Hematology/Oncology, Fox Chase Cancer Center, Philadelphia, PA, USA.
Data availability
The data that support the findings of this study are available on request from the corresponding author (C.D.). The data are not publicly available due to privacy restrictions as informed consent was waived for this particular study.
Author contributions
Christopher D’Avella, MD (Conceptualization; Data curation; Formal analysis; Methodology; Project administration; Writing—original draft; Writing—review & editing), Peter Whooley, DO, MBA (Conceptualization), Emily Milano, MS (Data curation), Brian Egleston, MPP, PhD (Data curation; Formal analysis; Methodology), James Helstrom, MD, MBA (Conceptualization), Kenneth Patrick, MD (Writing—review & editing), Martin Edelman, MD (Writing—original draft; Writing—review & editing), Jessica Bauman, MD (Conceptualization; Supervision; Validation; Writing—original draft; Writing—review & editing).
Funding
This project was funded in part by National Institutes of Health/National Cancer Institute grant No. P30CA006927 (Fox Chase Cancer Center support grant) for our data analysis.
Conflicts of interest
No authors have relevant disclosures or conflicts of interest.
References
- 1. American Cancer Society. Cancer Facts & Figures 2023. Atlanta: American Cancer Society; 2023. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/annual-cancer-facts-and-figures/2023/2023-cancer-facts-and-figures.pdf. Accessed January 4, 2024.
- 2. Siegel RL, Miller KD, Jemal A.. Cancer statistics, 20220. CA Cancer J Clin. 2022;68(1):7-30. [DOI] [PubMed] [Google Scholar]
- 3. Lash RS, Bell JF, Reed SC, et al. A systematic review of emergency department use among cancer patients. Cancer Nurs. 2017;40(2):135-144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Rivera DR, Gallicchio L, Brown J, Liu B, Kyriacou DN, Shelburne N.. Trends in adult cancer–related emergency department utilization. JAMA Oncol. 2017;3(10):e172450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Brooks GA, Li L, Uno H, Hassett MJ, Landon BE, Schrag D.. Acute hospital care is the chief driver of regional spending variation in Medicare patients with advanced cancer. Health Aff. 2014;33(10):1793-1800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Narang AK, Nicholas LH.. Out-of-pocket spending and financial burden among Medicare beneficiaries with cancer. JAMA Oncol. 2017;3(6):757-765. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Brooks GA, Abrams TA, Meyerhardt JA, et al. Identification of potentially avoidable hospitalizations in patients with GI cancer. J Clin Oncol. 2014;32(6):496-503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Panattoni L, Fedorenko C, Greenwood-Hickman MA, et al. Characterizing potentially preventable cancer- and chronic disease–related emergency department use in the year after treatment initiation: a regional study. JOP. 2018;14(3):e176-e185. [DOI] [PubMed] [Google Scholar]
- 9. Handley NR, Schuchter LM, Bekelman JE.. Best practices for reducing unplanned acute care for patients with cancer. JOP 2018;14(5):306-313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Allen L, Cummings J, Hockenberry J.. The impact of urgent care centers on nonemergenct emergency department visits. Health Serv Res. 2021;56(4):721-730. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Carlson L, Raja A, Dworkis D, et al. Impact of urgent care openings on emergency department visits to two academic centers within an integrated health system. Ann Emerg Med. 2020;75(3):382-391. [DOI] [PubMed] [Google Scholar]
- 12. Services, Center for Medicare and Medicaid Services. The Oncology Care Model. https://innovation.cms.gov/initiatives/Oncology-Care/. Accessed October 4, 2018.
- 13. Mendenhall MA, Dyehouse K, Hayes J, et al. Practice transformation: early impact of the oncology care model on hospital admissions. JOP. 2018;14(12):e739-e745. [DOI] [PubMed] [Google Scholar]
- 14. Services, Center for Medicare and Medicaid Services. Hospital Outpatient Reporting Program.https://www.cms.gov/Medicare/Quality-Initiatives-PatientAssessmentInstruments/HospitalQualityInits/HospitalOutpatientQualityReportingProgram. Accessed October 4, 2018.
- 15. Centers for Medicare and Medicaid Services. Admissions and Emergency Department (ED) Visits for Patients Receiving Outpatient Chemotherapy. Oncology Urgent Care and Health Care Utilization. https://www.cms.gov/Medicare/Quality-Initiatives-Patient-AssessmentInstruments/HospitalQualityInits/Measure-Methodology.html. Accessed October 4, 2018.
- 16. Ayers A. New urgent care models help cancer patients. J Urgent Care Med. 2018. https://www.jucm.com/new-urgent-care-models-help-cancer-patients/. Published 2017. Accessed October 4, 2018. [Google Scholar]
- 17. Andrews M. Urgent Care Sites Cater To Cancer Patients, Letting Them Check Some Worries At Door. Kaiser Health News. https://khn.org/news/urgent-care-sites-cater-tocancer-patients-letting-them-check-some-worries-at-door/. Published 2017. Accessed October 4, 2018.
- 18. Ruegg TA. A nurse practitioner-led urgent care center-meeting the needs of the patient with cancer. Clin J Oncol Nurs. 2013;Aug 117(4):E52-7. [DOI] [PubMed] [Google Scholar]
- 19. Sivendran S, Holliday R, De la Torre K, Newport KB.. Impact of a nurse practitioner-staffed, symptom-management clinic on emergency department utilization in a large community oncology cancer institute. J Clin Oncol 2014;32(suppl 31):58. doi: 10.1200/jco.2014.32.31_suppl.58 [DOI] [Google Scholar]
- 20. Hong AS, Froehlich T, Clayton Hobbs S, Lee SJC, Halm EA.. Impact of a cancer urgent care clinic on regional emergency department visits. J Oncol Pract. 2019;15(6):e501-e509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Klotz A, Martin S, Atoria CL, Lipitz Snyderman AN, Groeger JS.. The effect of observation units in cancer care on hospital admissions. J Clin Oncol 2014;32(suppl 15):e17633. doi: 10.1200/jco.2014.32.15 [DOI] [Google Scholar]
- 22. Parikh A, Berizzi D, Tsao C-K, Smith CB.. Characterization of sick visits at an enhanced oncology urgent care center. J Clin Oncol 2018;36(suppl 30):285. doi: 10.1200/jco.2018.36.30_suppl.285 [DOI] [Google Scholar]
- 23. Gould Rothberg BE, Canavan ME, Mun S, et al. Impact of a dedicated cancer urgent care center on acute care utilization. J Clin Oncol Oncol Pract 2022;18(1):e129-e136. [DOI] [PubMed] [Google Scholar]
- 24. Liang K-L, Tackett S, Myers S, et al. An oncology urgent care clinic for the management of immune-related adverse events: a descriptive analysis. Curr Oncol 2022;29(6):4342-4353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Fleming C, Kelly D.. Oncology urgent care clinics: understanding utilization and best practices. AdvancedPractitioner.com. 2022;13(3):265-269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Decker K, Lambert P, Galloway K, et al. Predictors for urgent care clinic and emergency department visits for individuals with cancer. Curr Oncol. 2021;28(3):1773-1789. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Wang B, Mehrotra A, Friedman A.. Urgent care centers deter some emergency department visits but, on net, increase spending. Health Affairs. 2021;40(4):587-595. [DOI] [PubMed] [Google Scholar]
- 28.Temple Health. Hospital Pricing Information. https://www.templehealth.org/patients-visitors/billing-financial/hospital-pricing-information. Accessed November 28, 2023.
- 29. Williams RL. A note on robust variance estimation for cluster-correlated data. Biometrics. 2000;56(2):645-646. [DOI] [PubMed] [Google Scholar]
- 30. Pennsylvania Department of Health D of HI. Data from Annual Health Questionnaire: Emergency Services Capability and Utilization. https://www.health.pa.gov/topics/HealthStatistics/HealthFacilities/HospitalReports/Pages/hospital-reports.aspx. Published 2020. Accessed June 2020.
- 31. Mortensen K, Song PH.. Minding the gap: a decomposition of emergency department use by Medicaid enrollees and the uninsured. Med Care. 2008;46(10):1099-1107. [DOI] [PubMed] [Google Scholar]
- 32. Ho V, Metcalfe L, Dark C, et al. Comparing utilization and costs of care in freestanding emergency departments, hospital emergency departments, and urgent care centers. Ann Emerg Med 2017;70(6):846-857.e3. [DOI] [PubMed] [Google Scholar]
- 33. Mehrotra A, Liu H, Adams JL, et al. Comparing costs and quality of care at retail clinics with that of other medical settings for 3 common illnesses. Ann Intern Med. 2009;151(5):321-328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Baker LC, Baker LS.. Excess cost of emergency department visits for nonurgent care. Health Aff (Millwood). 1994;13(5):162-171. [DOI] [PubMed] [Google Scholar]
- 35. Panzone J, Welch C, Morgans A, et al. Association of race with cancer-related financial toxicity. J Clin Oncol Oncol Pract 2022;18(2):e271-e283. [DOI] [PubMed] [Google Scholar]
- 36. Rico JF, Caterino JM, Stephens JA, et al. Variables associated with admission rates among cancer patients presenting to emergency departments: a CONCERN group study. Emerg Cancer Care 2023;2(1):1-12. [Google Scholar]
- 37. Pittman N, Hopman W, Mates M.. Emergency room visits and hospital admission rates after curative chemotherapy for breast cancer. J Oncol Pract. 2015;11(2):120-125. [DOI] [PubMed] [Google Scholar]
- 38. Siddique S, Tipton K, Leas N, et al. Interventions to reduce hospital length of stay in high-risk populations a systematic review. JAMA Network Open. 2021;4(9):e2125846. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Morris J, Gould Rothberg B, Prsic E, et al. Outcomes on an inpatient oncology service after the introduction of hospitalist comanagement. J. Hosp. Med. 2023;18(5):391-397. [DOI] [PubMed] [Google Scholar]
- 40. Stone K, Zwiggelaar R, Jones P, Mac Parthaláin N.. A systematic review of the prediction of hospital length of stay: Towards a unified framework. PLOS Digital Health. 2022;1(4):e0000017. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author (C.D.). The data are not publicly available due to privacy restrictions as informed consent was waived for this particular study.