Abstract
Background:
Receipt of GCT is associated with improved prognosis in foregut cancers. Studies show that patients living in areas of high socioeconomic deprivation have worse healthcare outcomes, however its effect on GCT in foregut cancers has not been evaluated. We studied the impact of the Area Deprivation Index (ADI) as a barrier to GCT.
Study Design:
A single-institution retrospective review of 498 foregut cancer patients (gastric, pancreatic, and hepatobiliary adenocarcinoma) from 2018–2022 was performed. GCT was defined based on National Comprehensive Cancer Network guidelines. Area deprivation index (ADI), a validated measure of neighborhood socioeconomic disadvantage was divided into terciles (low, medium, and high) with high ADI indicating the most disadvantage.
Results:
328/498 patients (66%) received GCT: 66%, 72% and 60% in pancreatic, gastric and hepatobiliary cancers, resp. Median (IQR) time from symptoms to work-up was 6 weeks (3–13), and from oncology appointment to treatment was 4 weeks (2–10). 46% were diagnosed in the Emergency Department (ED). On univariable analyses, age ≥75 years (p=0.001), Black race (p<0.0001), ED diagnosis (p=0.003), no prior cancer screening (p=0.004), high ADI (p<0.0001), ≥ 3 ED visits after diagnosis (p=0.03) were associated with non-receipt of GCT. Additionally, times from symptoms to work up; from diagnosis to oncology appointment and from oncology appointment to treatment were ≥6, ≥4 and ≥4weeks, resp. and were also associated with non-receipt of GCT(p<0.001).
On multivariable analyses, age ≥75 years [OR 0.39 (95% CI 0.18–0.87)], Black race [OR 0.52 (95% CI 0.31–0.86)], high ADI [0.25(0.14–0.48)], ≥6 weeks from symptoms to work-up [0.44(0.27–0.73)] and ≥4 weeks from oncology appointment to treatment [0.63 (0.36–0.98)] were all independently associated with non-receipt of GCT.
Conclusions:
Residence in an area of high deprivation predicts non-receipt of GCT. This is due to multiple individual and system level barriers. Identifying these barriers and developing effective interventions, including community outreach and collaboration, leveraging telehealth, and increasing oncologic expertise in under-served areas may improve access to GCT.
Keywords: Area deprivation, disparities, care equity, foregut cancers
INTRODUCTION
Foregut cancers account for 20% (over 110,000) of new cancer deaths in the United States each year (1). The incidence of foregut cancers is projected to increase substantially from 2010–2030, such that gastric, hepatobiliary and pancreas cancer are projected to have an increased incidence of 67%, 59% and 55% respectively (2). Both the burden of disease and cancer-related mortality are higher in southern states (3). Surgical resection remains the only potentially curative treatment in resectable disease, however the majority of patients present with metastatic or locally advanced disease precluding resection, where chemotherapy is the mainstay of treatment (4). The National Cancer Comprehensive Network (NCCN) clinical practice guidelines that reflect evidence-based, expert consensus-driven management, have been used to establish standard of care for various cancers (5). Treatment that is adherent to these guidelines i.e., guideline concordant treatment (GCT) has been shown to improve patient outcomes, however up to 70% of patients with foregut cancers do not receive GCT (6–10).
Social determinants of health negatively impact cancer care and outcomes (6). One third of survival disparities are thought to be mediated by socioeconomic disparities in curative-intent surgery for poor-prognosis cancers such as pancreatic cancer (6). Neighborhood deprivation impacts health through several mechanisms, including a dearth of medical expertise and health care resources (11), ineffective or disjointed referral systems (11), medical mistrust (12, 13), and transportation barriers (14–16).
Safety-net hospitals play a vital role in providing care to vulnerable populations. While studies based on national databases have suggested that safety-net hospitals may provide equivalent access to surgical resection, chemotherapy and radiation therapy in pancreatic cancer (17), other studies have shown decreased access to liver surgery for hepatocellular cancer, and variable quality of care and outcomes in hepatopancreatobiliary and other cancers (18, 19). While these studies evaluate zip code-based socioeconomic status, they do not capture the various domains and facets that comprise neighborhood disadvantage. Additionally, national database studies may not have enough granular information to allow for a nuanced evaluation of timeliness of access to care, and appropriate reasons for non-receipt of certain aspects of care, such as non-receipt of curative-intent surgery due to progression while on neoadjuvant chemotherapy. In this context, we sought to study the impact of neighborhood deprivation using a validated measure of neighborhood disadvantage (Area Deprivation Index; ADI) on access to timely GCT at a safety-net hospital that serves a large minority population in the Deep South.
METHODS
Study Site and Population
A retrospective review of electronic health records was performed to identify patients with pancreatic, gastric and hepatobiliary adenocarcinoma who were treated at a tertiary-care facility in southern Alabama from January 2018 to December 2022. The facility serves as the safety-net hospital and main tertiary-referral center for central and southern Alabama and the adjoining areas of the Gulf coast including southeast Mississippi and portions of Louisiana and northwest Florida. The region served by the facility is home to 5.6 million people, with a large rural and racial minority population (20).
We identified patients with the specified cancers, as defined by International Classification of Disease (ICD) codes C25.0-C25.3, C25.7- C25.9 (pancreatic), C16.0- C16.9 (gastric) and C22.1, C22.7-C22.9, C23, C24.0, C24.1, C24.8, C24.9 (hepatobiliary). Individual patient charts were reviewed to include only those with histology consistent with adenocarcinoma. This study was approved by the Institutional Review Board at The University of South Alabama.
Outcomes
The primary outcome was receipt of guideline-concordant therapy (GCT), as defined based on National Comprehensive Cancer Network guidelines. Secondary outcomes were timely access to oncologic care i.e., time from symptoms to start of work-up, time from diagnosis to being seen by relevant oncologic expert, time from oncology appointment to complete staging, and time from oncology appointment to start of treatment.
GCT varied based on type of cancer and stage. For Stage IV disease it was defined as receipt of systemic therapy, unless there was documentation of clinical status that precluded receipt, or documentation of a palliative care consult and patient desire to not receive therapy and proceed with hospice care. In patients with pancreatic adenocarcinoma, for Stage I and II disease, GCT was defined as receipt of chemotherapy and curative-intent surgical resection, unless there was documentation of comorbidities that precluded resection, or progression of disease on chemotherapy. For Stage III pancreatic adenocarcinoma, GCT was defined as receipt of chemotherapy, with or without curative-intent surgical resection or radiation therapy. In patients with non-metastatic gastric adenocarcinoma, GCT was defined as receipt of endoscopic or surgical resection (for T1a and T1b tumors without nodal or other involvement), or chemotherapy and curative-intent surgical resection, unless there was documentation of comorbidities that precluded resection, or progression of disease on chemotherapy. In patients with non-metastatic cholangiocarcinoma, GCT varied based on location and surgeon assessment of resectability: for intra-hepatic and extrahepatic cholangiocarcinoma determined to be resectable, GCT was defined as curative-intent surgical resection (if R0 resection were achieved prior to 2020) or curative-intent surgical resection and systemic therapy (2020 onwards, and in all patients without an R0 resection); for perihilar cholangiocarcinoma, GCT was defined as curative-intent surgical resection with or without radiation with or without systemic therapy, or systemic therapy with or without radiation. For patients with resectable gallbladder adenocarcinoma, GCT was defined as curative-intent surgical resection (if R0 resection were achieved prior to 2020) or curative-intent surgical resection and systemic therapy (2020 onwards, and in all patients without an R0 resection). For unresectable gallbladder adenocarcinoma, GCT was defined as systemic therapy with or without radiation.
Exposure of Interest
The primary exposure of interest was the Area Deprivation Index (ADI), a validated neighbourhood-level composite index, based on 17 specific measures captured in the American Community Survey (21). The specific ACS measures that make up the composite index include factors within the theoretical domains of income, education, employment and housing quality (22). The ADI allows for ranking of neighborhoods based on socioeconomic disadvantage at the state or national level. The specific ADI score reflects neighborhood deprivation relative to the remainder of the region of interest (state or nation). The ADI at the state level is a decile ranking at the block group level from 1 to 10. The deciles are constructed by ranking ADI from low to high and grouping the neighborhoods into bins corresponding to each 10% range of the ADI. Group 1 is the lowest ADI, indicating the lowest level of disadvantage within the state, and group 10 is the highest ADI indicating the highest level of disadvantage within the state. The ADI at the national level is a percentile ranking from 1 to 100, with group 1 indicating the lowest level of disadvantage within the nation, and group 10 indicating the highest level of disadvantage within the nation. Given the relative poverty and socioeconomic status of Alabama and the surrounding catchment area compared to the nation (23), we elected to use state level data. The state ADI score was calculated using the ADI mapping atlas and patient addresses (22).
The 2020 ADI was used for this study, which uses the ACS data for 2020, representing a 5-year period from 2016–2020. The ADI state rankings range from 1 to 10, with higher numbers represented increased disadvantage. For the purpose of this study, state deciles were grouped into tertiles (24, 25): low deprivation (low ADI; least disadvantaged), intermediate deprivation and high deprivation (high ADI; most disadvantaged).
Covariates
Covariates included sociodemographic factors, access-related and cancer-specific variables. Sociodemographic variables included age (<50 years, 50–74 years, ≥75 years), race (White, Black, Other; since patients of other minority races were of very small numbers, these were analyzed as a group), relationship status (single, married, divorced/separated, and unknown), and residence in a rural location. The health and access-related variables abstracted included insurance status (private insurance, Medicare, Medicaid, other and uninsured status), body mass index (BMI), comorbidity status, which was assessed using the Charlson Comorbidity Index (CCI), presence of a primary care physician (PCP), prior screening behavior (age- and guideline- appropriate colonoscopy and mammogram screening), tobacco use, alcohol use, where the cancer was diagnosed (in the outpatient setting or in the Emergency Department (ED) without pre-existing work-up initiated by patient’s physician in the outpatient setting), ED utilization after establishment of cancer diagnosis. Cancer-specific variables included cancer type, tumor characteristics (clinical and pathological stage), cancer-specific staging imaging, additional staging procedures (including endoscopic staging and diagnostic laparoscopy when appropriate), curative-intent surgical treatment, systemic therapy, and radiation therapy. GCT was determined by review of individual medical records to account for adherence to NCCN type- and stage- appropriate guidelines and treated as a dichotomous variable. Consensus regarding definition of GCT was achieved by two surgical oncologists.
Statistical Analysis
Descriptive statistics were used to compare patient characteristics with chi-square tests for categorical variables and t-test for continuous variables. We compared patients who received GCT compared with those who did not, to assess for factors associated with receipt of GCT. Multivariable logistic regression was used to assess factors independently associated with receipt of GCT. Multivariable logistic regression was used to assess factors (listed above) independently associated with receipt of GCT. Variables with p values lower than 0.05 were considered significant. Statistical analyses were performed using Stata, version 16.0 (StataCorp, College Station, TX, USA).
RESULTS
Patient Characteristics
The study identified 498 patients who met study criteria: 286 (57.43%) with pancreatic adenocarcinoma, 106 (21.29%) with gastric adenocarcinoma, and 106 (21.29%) with hepatobiliary adenocarcinoma. Two- third (n=328; 65.86%) of the patients received GCT, which varied based on underlying cancer type: 66.08% in pancreatic adenocarcinoma (n=189), 71.69% in gastric adenocarcinoma (n=76), and 59.43% in hepatobiliary adenocarcinoma (n=63).
Overall, 274 patients (55.02%) were male; 327 (65.55%) were White, 145 (29.12%) were Black, and 26 (5.22%) were of other races. Most patients (57.03%, n=284) were insured through Medicare, followed by private insurance (25.70%, n=128), Medicaid (6.63%, n=33) and VA insurance (6.63%, n=33). 20 patients (4.02%) were uninsured. Seventy-two patients (14.46%) lived in rural locations. Over 40% (n=207) of patients lived in low deprivation areas, 30% (n=147) lived in intermediate deprivation areas, and 29% (n=144) lived in high deprivation areas.
Sociodemographic and Access-related Variables by ADI Terciles
Black patients were more likely to live in areas of high deprivation compared with White patients (56.55% vs. 16.82%, p <0.0001), as were unmarried patients (36.40% vs. 22.59%. p=0.001). Patients living in areas of high deprivation were more likely to have Medicaid insurance (51.52% vs. 21.21%) and less likely to have Medicare (27.11% vs. 45.07%) or private insurance (23.44% vs. 42.97%), but the percentage of uninsured patients was similar across ADI categories (p= 0.02). Patients living in areas of high deprivation were less likely to have a PCP (21.52% vs. 48.82%, p <0.0001) and less likely to have had prior cancer screening (16.09% vs. 57.47%, p <0.0001). Patients living in a high deprivation area were also more likely to have their work-up performed in an ED (49.34% vs. 25.11%, p <0.0001) and have increased ED utilization after diagnosis (≥3 ED visits; 29.9% vs. 6.8%, p <0.0001). Of the 671 individual ED visits after diagnosis, 395 (58.8%) were for reasons that may have been prevented with improved healthcare access (Supplementary Table 1). The most frequent reasons for ED utilization overall included poor pain control (118 visits [17.6%]), nausea/vomiting (75 visits [11.2%]), dehydration (72 visits [10.7%]), and symptomatic anemia (38 visits [5.7%]). Patient sociodemographic and access-related variables based on ADI tertiles are summarized in Table 1.
Table 1:
Sociodemographic and access-related variables based on Area Deprivation Index (ADI) tertiles
| Variable | Low Deprivation | Intermediate Deprivation | High Deprivation | p-value |
|---|---|---|---|---|
| Age | ||||
| <55 years | 34 (16.4%) | 20 (13.6%) | 25 (17.4%) | 0.07 |
| 55–74 years | 118 (57%) | 103 (70.1%) | 94 (65.3%) | |
| 75+ years | 55 (26.6%) | 24 (16.3%) | 25 (17.4%) | |
| Gender | ||||
| Male | 115 (55.6%) | 82 (55.8%) | 77 (53.5%) | 0.91 |
| Female | 92 (44.4%) | 65 (44.2%) | 67 (46.5%) | |
| Race | ||||
| White | 172 (83.1%) | 100 (68%) | 55 (38.2%) | <0.0001 |
| Black | 23 (11.1%) | 40 (27.2%) | 82 (56.9%) | |
| Other | 12 (5.8%) | 7 (4.8%) | 7 (4.9%) | |
| Marital Status | ||||
| Married | 128 (61.8%) | 81 (55.1%) | 61 (42.4%) | 0.001 |
| Unmarried | 79 (38.2%) | 66 (44.9%) | 83 (57.6%) | |
| Insurance Status | ||||
| Medicare | 128 (61.8%) | 79 (53.7%) | 77 (53.5%) | 0.02 |
| Private | 55 (26.6%) | 43 (29.3%) | 30 (20.8%) | |
| Medicaid | 7 (3.4%) | 9 (6.1%) | 17 (11.8%) | |
| Uninsured | 5 (2.4%) | 5 (3.4%) | 10 (6.9%) | |
| VA | 12 (5.8%) | 11 (7.5%) | 10 (6.9%) | |
| Presence of PCP | ||||
| No | 13 (6.5%) | 33 (22.6%) | 62 (43.1%) | <0.0001 |
| Yes | 186 (93.5%) | 113 (77.4%) | 82 (56.9%) | |
| Work-up by | ||||
| ED | 57 (27.5%) | 58 (39.4%) | 112 (77.8%) | <0.0001 |
| PCP | 150 (72.6%) | 89 (60.6%) | 32 (22.2%) | |
| Prior Cancer Screening | ||||
| No | 107 (51.7%) | 101 (68.7%) | 116 (80.6%) | <0.0001 |
| Yes | 100 (48.3%) | 46 (31.3%) | 28 (19.4%) | |
| Stage at Diagnosis | ||||
| 1 | 23 (11.3%) | 18 (12.2%) | 17 (11.9%) | 0.42 |
| 2 | 31 (15.2%) | 19 (12.9%) | 28 (19.6%) | |
| 3 | 63 (30.9%) | 52 (35.4%) | 47 (32.9%) | |
| 4 | 87 (42.7%) | 58 (39.5%) | 51 (35.7%) | |
| ED visits after diagnosis | ||||
| <3 | 193 (93.2%) | 128 (87.1%) | 101 (70.1%) | <0.0001 |
| ≥3 ED visits | 14 (6.8%) | 19 (12.9%) | 43 (29.9%) | |
Cancer-specific and Treatment-related Variables by ADI Terciles
There was no difference in stage at presentation, or in percentage of metastatic disease based on ADI (35.66% vs. 42.65% in high vs. low deprivation areas, p= 0.42). Twenty-five patients (5%) did not have complete staging imaging; patients in high deprivation areas were more likely to not have complete staging imaging compared to those in low deprivation areas (13.2% vs. 1%, p <0.0001). Compared to patients living in low deprivation areas, patients living in high deprivation areas were less likely to receive GCT (45.1% vs. 76.8%, p <0.0001), less likely to receive chemotherapy (70% vs. 93.6%, p <0.0001) and less likely to receive curative-intent surgical resection when appropriate i.e., for non-metastatic, non-locally advanced disease without documentation of progression or comorbidities precluding surgical resection (42.4% vs. 64.1%, p= 0.002).
Timeliness of Care
Median (Interquartile range [IQR]) time from symptoms to start of workup was 5.9 (2.8–12.6) weeks. Time from diagnosis to first oncology appointment was 3.9 (1.3–9.6) weeks, time from oncology appointment to complete staging was 1.4 (0.1– 5.9) weeks and time from oncology appointment to start of treatment (either systemic or surgical) was 3.9 (1.8– 8.5) weeks. Patients living in areas of high deprivation were more likely to have longer times from symptoms to work-up, from diagnosis to first oncology appointment and from oncology appointment to start of treatment (Table 2, Figure 1).
Table 2:
Timeliness of Care based on Area Deprivation Index (ADI) Tertiles
| Low ADI | Intermediate ADI | High ADI | p-value | |
|---|---|---|---|---|
| Time from symptoms to start of work-up in weeks | ||||
| Median (IQR) | 2.9 (1.7–4.4) | 6 (3.9–10.6) | 14.6 (10–21.4) | <0.0001 |
| Time from diagnosis to oncology appointment in weeks | ||||
| Median (IQR) | 2 (0.7– 3.6) | 2.7 (1.3– 4.9) | 5 (3–8.4) | <0.0001 |
| Time from oncology appointment to complete staging in weeks | ||||
| Median (IQR) | 0 (−1.4 −1.4) | 0.1 (−1.4–2.1) | 0.1 (−1.7–3.4) | 0.16 |
| Time from oncology appointment to start of treatment in weeks | ||||
| Median (IQR) | 2.1 (1.1– 3.9) | 3.3 (2.1– 5.7) | 4.7 (3.3–7.3) | <0.0001 |
Figure 1:
Timeliness of Care based on Area Deprivation Index (ADI) terciles.
Receipt of Guideline Concordant Treatment (GCT)
On univariable analysis, younger age (72.1% vs. 51.9%, p= 0.003), White race (72.5% vs. 51%, p < 0.0001), being married (73% vs. 57.5%, p < 0.0001), presence of a PCP (69% vs. 53.7%, p= 0.002), prior cancer screening (74.1% vs. 61.4%, p= 0.004), lower CCI score (61% vs. 73%, p=0.008) and having one’s work-up initiated in the outpatient setting instead of the ED (71.6% vs. 59%, p= 0.003) were associated with receipt of GCT. Metastatic disease was associated with a higher receipt of GCT (75% vs. 60.7%, p= 0.001). ADI was also associated with receipt of GCT (76.8% vs. 70.7% vs. 45.1% in low, medium, and high deprivation areas, p < 0.0001) (Figure 2). High ED utilization (3+ ED visits) after diagnosis was associated with lower receipt of GCT (55.3% vs. 67.8%, p= 0.03). Increased time (≥6 weeks) from symptoms to start of work up was associated with lower receipt of GCT (54.1% vs. 77%, p < 0.0001), as was increased time (≥4 weeks) from diagnosis to oncology appointment (55.1% vs. 71.9%, p < 0.0001), and increased time (≥4 weeks) from oncology appointment to start of treatment (49.6% vs. 79%, p < 0.0001). Insurance status (p= 0.06), residence in a rural location (p= 0.3), and time from oncology appointment to staging (p= 0.58) were not associated with GCT.
Figure 2:
Map of Alabama (2A) and the counties surrounding the health system (2B) showing receipt of treatment.
*GCT: Guideline-Concordant Treatment; ADI: Area Deprivation Index
On multivariable analysis, age ≥75 years (Odds ratio [OR] 0.35; 95% confidence interval [CI] 0.15–0.85), Black race (OR 0.56; 95% CI 0.31– 0.86), BMI ≥40 (OR 0.33; 95% CI 0.17– 0.64), being unmarried (OR 0.68; 95% CI 0.14– 0.87), and living in an area of high deprivation (OR 0.25; 95% CI 0.16– 0.48) were independently associated with non-receipt of GCT. Additionally, ≥6 weeks from symptoms to start of work up (OR 0.47; 95% CI 0.25– 0.89), ≥4 weeks from diagnosis to oncology appointment (OR 0.76; 95% CI 0.46– 0.93) and ≥4 weeks from oncology appointment to treatment (OR 0.30; 95% CI 0.19– 0.50) were predictive of non-receipt of GCT. The presence of metastatic disease was independently associated with increased receipt of GCT (OR 1.40; 95% CI 1.1– 2.3) (Table 3).
Table 3:
Variables associated with Receipt of Guideline Concordant Treatment (GCT)
| Univariable Analysis | Multivariable Analysis | ||||
|---|---|---|---|---|---|
| Variable | Non-GCT | GCT | p-value | OR (95% CI) | p-value |
| Age | |||||
| <55 years | 22 (27.9%) | 57 (72.1%) | 0.003 | Ref | |
| 55–74 years | 98 (31.1%) | 217 (68.9%) | 1.18 (0.57–2.43) | 0.64 | |
| 75+ years | 50 (48.1%) | 54 (51.9%) | 0.35 (0.15–0.85) | 0.02 | |
| Sex | |||||
| Male | 85 (31%) | 189 (69%) | 0.10 | ||
| Female | 85 (37.9%) | 139 (62.1%) | |||
| Race | |||||
| White | 90 (27.5%) | 237 (72.5%) | <0.0001 | Ref | |
| Black | 71 (49%) | 74 (51%) | 0.56 (0.31–0.86) | 0.01 | |
| Other | 9 (34.6%) | 17 (65.4%) | 1.01 (0.34–2.99) | 0.64 | |
| Marital Status | |||||
| Married | 73 (27%) | 197 (73%) | <0.0001 | Ref | |
| Unmarried | 97 (42.5%) | 131 (57.5%) | 0.68 (0.14–0.87) | 0.03 | |
| Insurance Status | |||||
| Medicare | 111 (39.1%) | 173 (60.9%) | 0.06 | Ref | |
| Private | 37 (28.9%) | 91 (71.1%) | 1.35 (0.75–2.44) | 0.3 | |
| Medicaid | 11 (33.3%) | 22 (66.7%) | 2.13 (0.77–5.8) | 0.14 | |
| Uninsured | 4 (20%) | 16 (80%) | 1.93 (0.93–7.63) | 0.07 | |
| VA | 8 (24.2%) | 25 (75.8%) | 2.72 (0.25–1.88) | 0.07 | |
| Presence of PCP | |||||
| No | 50 (46.3%) | 58 (53.7%) | 0.002 | Ref | |
| Yes | 118 (31%) | 263 (69%) | 0.99 (0.5–1.97) | 0.7 | |
| Work-up by | |||||
| ED | 93 (41%) | 134 (59%) | 0.003 | Ref | |
| PCP | 77 (28.4%) | 194 (71.6%) | 0.71 (0.4–1.25) | 0.23 | |
| Prior Cancer Screening | |||||
| No | 125 (38.6%) | 199 (61.4%) | 0.004 | Ref | |
| Yes | 45 (25.9%) | 129 (74.1%) | 1.85 (1.12– 3.03) | 0.01 | |
| Metastatic Cancer | |||||
| No | 117 (39.3%) | 181 (60.7%) | 0.001 | Ref | |
| Yes | 49 (25%) | 147 (75%) | 1.40 (1.1– 2.3) | 0.01 | |
| Charlson Comorbidity Index (CCI) | |||||
| CCI mild-moderate | 51 (26.98%) | 138 (73.02%) | 0.008 | Ref | |
| CCI severe | 119 (38.52%) | 190 (61.48%) | 0.72 (0.25–2.65) | 0.51 | |
| Area Deprivation Index | |||||
| Low deprivation | 48 (23.2%) | 159 (76.8%) | <0.0001 | Ref | |
| Intermediate deprivation | 43 (29.3%) | 104 (70.7%) | 0.68 (0.40–1.15) | 0.15 | |
| High deprivation | 79 (54.9%) | 65 (45.1%) | 0.25 (0.16–0.48) | 0.02 | |
| Rural county | |||||
| No | 142 (33.33%) | 284 (66.67%) | 0.3 | ||
| Yes | 28 (38.9%) | 44 (61.1%) | |||
| ED visits after diagnosis | |||||
| <3 ED visits | 136 (32.2%) | 286 (67.8%) | 0.03 | Ref | |
| ≥3 ED visits | 34 (44.7%) | 42 (55.3%) | 0.62 (0.32–1.20) | 0.16 | |
| Time: Symptoms to start of work-up | |||||
| ≤6 wks | 56 (23%) | 188 (77%) | <0.0001 | Ref | |
| >6 wks | 107 (45.9%) | 126 (54.1%) | 0.47 (0.25–0.89) | 0.02 | |
| Time: Diagnosis to Oncology Appointment | |||||
| ≤4 wks | 90 (28.1%) | 230 (71.9%) | <0.0001 | Ref | |
| >4 wks | 80 (44.9%) | 98 (55.1%) | 0.76 (0.46– 0.93) | 0.03 | |
| Time: Oncology Appointment to Staging | |||||
| ≤2wks | 122 (32.4%) | 255 (67.6%) | 0.58 | ||
| >2 wks | 32 (33.2%) | 64 (66.7%) | |||
| Time: Oncology Appointment to Treatment | |||||
| ≤4wks | 58 (21%) | 218 (79%) | <0.0001 | Ref | |
| >4 wks | 112 (50.4%) | 110 (49.6%) | 0.30 (0.19–0.50) | <0.0001 | |
DISCUSSION
This study found that a substantial percentage of patients with foregut cancers do not receive GCT, even after being evaluated at a tertiary care academic center. Neighborhood disadvantage is independently associated with non-receipt of GCT, even after accounting for individual sociodemographic, comorbidity, and access to care measures. Additionally, patients living in areas of high deprivation had significantly longer time to access care across the cancer care continuum, with longer times from symptoms to start of work-up, longer times from cancer diagnosis to appointment with oncologic experts, and longer times from oncology appointment to start of treatment.
This study expands on the literature in several ways. To the best of our knowledge, this is the first study to evaluate the impact of neighborhood disadvantage on specific access to care measures for foregut cancer care. This study is set in a Deep South state with a significant minority population: 26% African-American compared to the national average of 13% (26). The health system serves as the safety-net hospital for the county, which has a high minority population (55% African American) and high rates of poverty, income inequality and segregation (23, 26), and the cancer institute serves a high percentage of minority and underinsured patients: 37% of patients who received cancer care at the cancer institute identified as Black and 78% had insurance coverage other than private insurance (including 14% uninsured). Notably, this study defined GCT based on in-depth review of electronic health records, and therefore provides sufficient granularity including that of comorbidities, physician documented discussions, and a more nuanced assessment of tumor biology and disease progression, which allows for assessment of aspects of GCT not available in national databases. The analysis also controls for race and other sociodemographic variables, as well as certain measures of access to care, and is therefore able to capture the impact of neighborhoods themselves on access to high-quality cancer care.
Abdominal foregut cancers as a group are aggressive malignancies associated with poor overall survival (1). Biological aggressiveness resulting in early metastases, relative chemoresistance, technically complex operations and chemotherapy regimens all contribute to challenges to the delivery of care and overall poor prognosis. While still far from ideal, current chemotherapy regimens, improvements in surgical technique, and regionalization of complex surgical oncologic care has resulted in improved outcomes in patients who are able to access this care. Stage-specific standard of care treatment i.e., GCT has been shown to be associated with improved cancer-specific outcomes across multiple cancer types (27–30). However, GCT in the case of foregut cancers requires coordinated and highly specialized multidisciplinary care, often delivered at least partly in tertiary care centers, with multimodal therapy being recommended in a large proportion of patients with local and locoregional disease. This poses additional access barriers to patients. Identifying these specific barriers at the personal, interpersonal, provider, institution and health system level is vital to develop strategies and interventions to mitigate these barriers. In addition to neighborhood deprivation, other factors independently predictive of GCT in this study included marital status, prior cancer screening, and timeliness of care. Marital status has been shown to be associated with improved outcomes in prior studies (31–33), and likely represents increased social support, as well as self-efficacy (34) and greater financial resources (35). Prior relevant cancer screening and timeliness of care may represent measures of access to the healthcare system. Patients with metastatic disease had higher rates of GCT than patients without metastatic disease in this study, likely due to how GCT was defined: receipt of systemic therapy in the case of metastatic disease, which was analyzed as a binary variable.
Timeliness of care is a pivotal factor in the delivery of patient-centric, high-quality cancer care, that may impact treatment options and outcomes. Delays in time to treatment can be associated with progression of disease, severely limiting treatment options, and resulting in worse outcomes. Delays in time to treatment have been shown to be associated with worse outcomes in various cancers including pancreatic cancer (36–40), however they have not been reliably shown in gastric cancer (41, 42). In our study, delays in time to care across various parts of the cancer care continuum were predictive of non-receipt of GCT. In addition to much longer intervals between symptom onset to diagnosis, patients living in areas of high deprivation also had longer time intervals from diagnosis to referral to oncologic expertise, and from oncology appointments to start of treatment, demonstrating that access to care continues to be a barrier even after patients have been evaluated at academic centers. Surrogates for an established relationship with the healthcare system such as presence of a PCP, workup initiated or performed by a PCP, and lower ED utilization after cancer diagnosis were associated with receipt of GCT on univariable, however, were not independently predictive of GCT on multivariable analysis. Nonetheless, this data offers insight into potential targets for intervention.
The results of the study serve as a stark reminder of the critical need for academic health systems to do better in addressing health disparities. While the data reflect the many structural reasons contributing to unequal cancer care, they also underscore shortcomings within our own healthcare institutions. Even after patients from disadvantaged neighborhoods are able to access our health systems, they are much less likely to receive timely care, or indeed guideline-concordant care, compared to patients living in less disadvantaged neighborhoods. The longer intervals between diagnosis, oncology appointments and treatment represent missed opportunities for early intervention, and potentially curative treatment in some patients. Almost 60% of the ED visits in this study were for reasons that may have been preventable, at least to some extent, with improved outpatient health care access, representing additional opportunities to improve the quality of care with measures such as the creation of infusion clinics, involvement of palliative care, improved symptom control, and more primary care access.
This study has several limitations that must be considered when its results are interpreted. This is a single institution study in a southern state, and as such, its results may not be generalizable to other areas of the country. GCT is nuanced and may be difficult to identify in a retrospective fashion, however, this study defines GCT based on detailed review of electronic health records, with consensus between surgical oncologists regarding whether a particular course of treatment was GCT or not. To the extent possible, resectability was defined as comprehensively as possible, taking into consideration anatomic, biological, and patient condition related factors. In addition to insurance coverage, other access to care measures were evaluated such as presence of a PCP, location of work-up, appropriate cancer screening, ED utilization, however these may not capture all access related variables. While this study captures all patients who were seen at all locations of the hospital system, it does not capture patients who may have not received any care for their foregut cancer, as well as patients who received all of their cancer care outside this health system. While 14% of the patients treated at the cancer center were uninsured overall, this percentage was much lower (4%) in this study and may represent a group of patients that are not being referred to the hospital system for any of their care. As such, it may underestimate the percentage of patients receiving non-GCT, perhaps from a sense of nihilism around foregut cancers, lack of oncologic or medical expertise in the community, physician bias, patient lack of resources, health literacy, self-advocacy, or a combination of all the above. This study only captures access and adherence to GCT and does not analyze its impact on overall survival given that the study period extends through the year 2022, however this will be evaluated in the future.
CONCLUSION
This study found that neighborhood disadvantage is independently associated with non-receipt of GCT in patients with foregut cancers, even when other sociodemographic variables are controlled for. It is imperative that we investigate the individual, interpersonal, and structural barriers that contribute to this inequitable access and develop effective interventions to ensure more equitable access to high-quality cancer care. Addressing these barriers requires not only an acknowledgement of the structural reasons that exist in broader society, but also critically reviewing the barriers that exist within our own healthcare systems that perpetuate unequal care. Improving community outreach, leveraging telehealth to increase collaboration with primary care providers and local hospitals in the community, and streamlining referral processes may improve referral of patients to tertiary care centers. Once patients have been seen at the health center, improved outpatient access to care including referral to primary care physicians, involvement of palliative care when appropriate, infusion clinic access and improved symptom control for patients receiving active cancer care may help increase GCT and decrease ED utilization.
Supplementary Material
Support:
This work was supported by the National Institutes of Health [Award Number 1K23MD018383–01]
Footnotes
Meeting: Southern Surgical Association 135th Annual Meeting, Hot Springs, VA, December 2023
REFERENCES
- 1.Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2022. CA: A Cancer Journal for Clinicians. 2022;72(1):7–33. [DOI] [PubMed] [Google Scholar]
- 2.Smith BD, Smith GL, Hurria A, Hortobagyi GN, Buchholz TA. Future of cancer incidence in the United States: burdens upon an aging, changing nation. J Clin Oncol. 2009;27(17):2758–65. [DOI] [PubMed] [Google Scholar]
- 3.Cancer Facts & Figures 2020. Atlanta, Ga: American Cancer Society; 2020. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/annual-cancer-facts-and-figures/2020/cancer-facts-and-figures-2020.pdf. American Cancer Society. 2020. [Google Scholar]
- 4.Meredith K. Management of foregut malignancies and hepatobiliary tract and pancreas malignancies. Journal of Gastrointestinal Oncology. 2018;9(5):878–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Benson AB 3rd, Brown E. Role of NCCN in Integrating Cancer Clinical Practice Guidelines into the Healthcare Debate. Am Health Drug Benefits. 2008;1(1):28–33. [PMC free article] [PubMed] [Google Scholar]
- 6.Swords DS, Mulvihill SJ, Brooke BS, Firpo MA, Scaife CL. Size and Importance of Socioeconomic Status-Based Disparities in Use of Surgery in Nonadvanced Stage Gastrointestinal Cancers. Ann Surg Oncol. 2020;27(2):333–41. [DOI] [PubMed] [Google Scholar]
- 7.Swords DS, Mulvihill SJ, Brooke BS, Skarda DE, Firpo MA, Scaife CL. Disparities in utilization of treatment for clinical stage I-II pancreatic adenocarcinoma by area socioeconomic status and race/ethnicity. Surgery. 2019;165(4):751–9. [DOI] [PubMed] [Google Scholar]
- 8.Frohman HA, Martin JT, Le AT, Dineen SP, Tzeng CD. Failure to operate on resectable gastric cancer: implications for policy changes and regionalization. J Surg Res. 2017;214:229–39. [DOI] [PubMed] [Google Scholar]
- 9.Idrees JJ, Merath K, Gani F, Bagante F, Mehta R, Beal E, et al. Trends in centralization of surgical care and compliance with National Cancer Center Network guidelines for resected cholangiocarcinoma. HPB (Oxford). 2019;21(8):981–9. [DOI] [PubMed] [Google Scholar]
- 10.Bilimoria KY, Bentrem DJ, Ko CY, Stewart AK, Winchester DP, Talamonti MS. National failure to operate on early stage pancreatic cancer. Ann Surg. 2007;246(2):173–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bhatt J, Bathija P. Ensuring Access to Quality Health Care in Vulnerable Communities. Acad Med. 2018;93(9):1271–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ahern MM, Hendryx MS. Social capital and trust in providers. Soc Sci Med. 2003;57(7):1195–203. [DOI] [PubMed] [Google Scholar]
- 13.Jaiswal J, Halkitis PN. Towards a More Inclusive and Dynamic Understanding of Medical Mistrust Informed by Science. Behav Med. 2019;45(2):79–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Obrochta CA, Parada H Jr., Murphy JD, Nara A, Trinidad D, Araneta MRH, Thompson CA. The impact of patient travel time on disparities in treatment for early stage lung cancer in California. PLoS One. 2022;17(10):e0272076. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Guidry JJ, Aday LA, Zhang D, Winn RJ. Transportation as a barrier to cancer treatment. Cancer Pract. 1997;5(6):361–6. [PubMed] [Google Scholar]
- 16.Kirkwood MK, Bruinooge SS, Goldstein MA, Bajorin DF, Kosty MP. Enhancing the American Society of Clinical Oncology workforce information system with geographic distribution of oncologists and comparison of data sources for the number of practicing oncologists. J Oncol Pract. 2014;10(1):32–8. [DOI] [PubMed] [Google Scholar]
- 17.Dhar VK, Hoehn RS, Kim Y, Xia BT, Jung AD, Hanseman DJ, et al. Equivalent Treatment and Survival after Resection of Pancreatic Cancer at Safety-Net Hospitals. J Gastrointest Surg. 2018;22(1):98–106. [DOI] [PubMed] [Google Scholar]
- 18.Farooq A, Paredes AZ, Merath K, Hyer JM, Mehta R, Sahara K, et al. How Safe Are Safety-Net Hospitals? Opportunities to Improve Outcomes for Vulnerable Patients Undergoing Hepatopancreaticobiliary Surgery. J Gastrointest Surg. 2020;24(11):2570–8. [DOI] [PubMed] [Google Scholar]
- 19.Sarkar R, Guss ZD, Lopez C, Brandel M, Murphy JD. Impact of hospital safety-net burden on oncology patterns of care and outcomes. Journal of Clinical Oncology. 2018;36(15_suppl):6567-. [Google Scholar]
- 20.United States Census Bureau. 2020. Census Data. [Google Scholar]
- 21.Singh GK. Area deprivation and widening inequalities in US mortality, 1969–1998. Am J Public Health. 2003;93(7):1137–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kind AJH, Buckingham WR. Making Neighborhood-Disadvantage Metrics Accessible - The Neighborhood Atlas. N Engl J Med. 2018;378(26):2456–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Fontenot K, Semega J, M K. Income and Poverty in the United States. U.S. Department of Commerce Economics and Statistics Administration, U.S. Census Bureau. 2018. [Google Scholar]
- 24.Borrelli S, Chiodini P, Caranci N, Provenzano M, Andreucci M, Simeon V, et al. Area Deprivation and Risk of Death and CKD Progression: Long-Term Cohort Study in Patients under Unrestricted Nephrology Care. Nephron. 2020;144(10):488–97. [DOI] [PubMed] [Google Scholar]
- 25.Stankowski TJ, Schumacher JR, Hanlon BM, Tucholka JL, Venkatesh M, Yang DY, et al. Barriers to breast reconstruction for socioeconomically disadvantaged women. Breast Cancer Res Treat. 2022;195(3):413–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. .CensusScope. www.CensusScope.org. Social Science Data Analysis Network, University of Michigan. www.ssdan.net. [Google Scholar]
- 27.Hamad A, DePuccio M, Reames BN, Dave A, Kurien N, Cloyd JM, et al. Disparities in Stage-Specific Guideline-Concordant Cancer-Directed Treatment for Patients with Pancreatic Adenocarcinoma. J Gastrointest Surg. 2021;25(11):2889–901. [DOI] [PubMed] [Google Scholar]
- 28.Ju MR, Wang SC, Mansour JC, Polanco PM, Yopp AC, Zeh HJ 3rd, Porembka MR. Disparities in Guideline-Concordant Treatment and Survival Among Border County Residents With Gastric Cancer. JCO Oncol Pract. 2022;18(5):e748–e58. [DOI] [PubMed] [Google Scholar]
- 29.Ahmed HZ, Liu Y, O’Connell K, Ahmed MZ, Cassidy RJ, Gillespie TW, et al. Guideline-concordant Care Improves Overall Survival for Locally Advanced Non-Small-cell Lung Carcinoma Patients: A National Cancer Database Analysis. Clin Lung Cancer. 2017;18(6):706–18. [DOI] [PubMed] [Google Scholar]
- 30.Javid SH, Varghese TK, Morris AM, Porter MP, He H, Buchwald D, et al. Guideline-concordant cancer care and survival among American Indian/Alaskan Native patients. Cancer. 2014;120(14):2183–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kravdal O The impact of marital status on cancer survival. Soc Sci Med. 2001;52(3):357–68. [DOI] [PubMed] [Google Scholar]
- 32.Baine M, Sahak F, Lin C, Chakraborty S, Lyden E, Batra SK. Marital status and survival in pancreatic cancer patients: a SEER based analysis. PLoS One. 2011;6(6):e21052. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Jin JJ, Wang W, Dai FX, Long ZW, Cai H, Liu XW, et al. Marital status and survival in patients with gastric cancer. Cancer Med. 2016;5(8):1821–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.In: Adler NE, Page AEK, editors. Cancer Care for the Whole Patient: Meeting Psychosocial Health Needs. The National Academies Collection: Reports funded by National Institutes of Health. Washington (DC)2008. [PubMed] [Google Scholar]
- 35.Stack S, Eshleman JR. Marital Status and Happiness: A 17-Nation Study. Journal of Marriage and Family. 1998;60(2):527–36. [Google Scholar]
- 36.Gobbi PG, Bergonzi M, Comelli M, Villano L, Pozzoli D, Vanoli A, Dionigi P. The prognostic role of time to diagnosis and presenting symptoms in patients with pancreatic cancer. Cancer Epidemiol. 2013;37(2):186–90. [DOI] [PubMed] [Google Scholar]
- 37.Polverini AC, Nelson RA, Marcinkowski E, Jones VC, Lai L, Mortimer JE, et al. Time to Treatment: Measuring Quality Breast Cancer Care. Ann Surg Oncol. 2016;23(10):3392–402. [DOI] [PubMed] [Google Scholar]
- 38.Gamboa AC, Rupji M, Switchenko JM, Lee RM, Turgeon MK, Meyer BI, et al. Optimal timing and treatment strategy for pancreatic cancer. J Surg Oncol. 2020;122(3):457–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Neal RD, Tharmanathan P, France B, Din NU, Cotton S, Fallon-Ferguson J, et al. Is increased time to diagnosis and treatment in symptomatic cancer associated with poorer outcomes? Systematic review. Br J Cancer. 2015;112 Suppl 1(Suppl 1):S92–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hanna TP, King WD, Thibodeau S, Jalink M, Paulin GA, Harvey-Jones E, et al. Mortality due to cancer treatment delay: systematic review and meta-analysis. BMJ. 2020;371:m4087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Brenkman HJF, Visser E, van Rossum PSN, Siesling S, van Hillegersberg R, Ruurda JP. Association Between Waiting Time from Diagnosis to Treatment and Survival in Patients with Curable Gastric Cancer: A Population-Based Study in the Netherlands. Ann Surg Oncol. 2017;24(7):1761–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Han KT, Kim W, Kim S. Does Delaying Time in Cancer Treatment Affect Mortality? A Retrospective Cohort Study of Korean Lung and Gastric Cancer Patients. Int J Environ Res Public Health. 2021;18(7). [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.