Abstract
Purpose:
Non-specific presenting symptoms make multiple myeloma (MM) one of the most difficult hematologic cancers to diagnose, increasing the risk of late diagnosis and emergency presentation (EP). We examined the frequency of EP in MM, estimated associations between EP and patient-level factors, and MM symptom-related healthcare utilization in the year before diagnosis.
Methods:
We studied Medicare beneficiaries age ≥66 years diagnosed with MM between 2008–2017. Emergency presentation (EP) was defined as diagnosis preceded by a recent ED visit, subclassified as: inpatients (EP-IP) for subsequently admitted patients and outpatients (EP-OP) for those discharged for follow-up care.
Results:
Among 20,437 patients, 52% were male, 75% non-Hispanic White (NHW), and the median age was 76. Overall, 37% were diagnosed through involvement with the ED (EP-IP: 30%; EP-OP: 7%). In iteratively adjusted models, oldest age, female sex, non-Hispanic Black, Hispanic, Asian American, Native Hawaiian or Other Pacific Islander race and ethnicity, unmarried status, dual eligibility for Medicaid and Medicare, lowest county SES, rurality, frailty, and higher healthcare use in the year pre-diagnosis were associated with higher frequency of ED involvement in diagnosis. Symptom-related healthcare utilization was comparable among EP and non-EP groups in the year preceding MM diagnosis but markedly increased for the EP group in the month before diagnosis.
Conclusions:
More than one-third of older US adults with MM are diagnosed through EP. Several socio-demographic, clinical and healthcare utilization factors are associated with EP. Similarities in healthcare utilization across EP status in the months preceding diagnosis suggest missed opportunities for earlier detection of MM.
Keywords: Emergency presentation, Multiple myeloma, SEER-Medicare, Older adults
Background
Only four percent of multiple myeloma (MM) cases are diagnosed at early disease stages [1] due to the broad, non-specific symptom signature of the condition and lack of a screening modality. Although the manifestation of prodromal symptoms may prompt primary care or specialist visits, diagnosis may still be delayed due to the non-specific nature of presenting complaints which result in multiple encounters. There are several documented factors responsible for diagnostic delays for MM, including patient (e.g., lack of awareness about related symptoms), physician (e.g., atypical and non-specific symptoms which chiefly encompass musculoskeletal pain, and pursuing only commonly encountered diagnoses) and system-related delays (e.g., limitations in the diagnostic lab procedure) [2,3]. In addition, MM disproportionately affects Black patients in terms of incidence and access to care, and delays are more common in these individuals [4]. Delays before diagnosis are associated with more complicated disease [5] as well as increased end organ damage and worse progression-free survival [2].
Another consequence of such diagnostic lag is that MM may be discovered during an emergency healthcare encounter. The time and resources for the detailed relay of prognosis, follow-up and treatment plans are not readily available in such settings. This emergency diagnostic pathway is not unique to MM but represents a broader challenge across cancer types. However, the number of pre-diagnostic, primary care consultations has been shown to be highest for MM compared to other cancers, indicating missed opportunities for earlier diagnosis [6]. Compared to other hematologic cancers, MM often has the longest time from symptom onset and medical help-seeking to diagnosis [7,8]. MM is diagnosed via lab testing with protein electrophoresis and immunofixation but being relatively rare, it is not diagnosed and treated on a regular basis. The symptom-specific diagnostic criteria of MM also includes features of end organ damage consisting of hypercalcemia, renal insufficiency, anemia and lytic bone lesions (CRAB criteria) [9]. Patients may present to primary care with non-specific manifestations of these features such as pain and fatigue [1]. The ambiguous nature of these symptoms contributes to delays in diagnosis and increases the potential for diagnosis in the emergency department [10].
Emergency department (ED) involvement in cancer diagnosis (termed emergency presentation or EP) is associated with poor health outcomes and disparities in care [11–16]. People who have an EP cancer diagnosis tend to be older, have less access to consistent primary care, lower socioeconomic resources [17,18], clinically advanced disease and poor survival [19–21]. Most of the existing evidence on EP in hematologic malignancies come from international studies and research on this topic is limited in the United States [18].
Understanding the issues that lead up to a delayed diagnosis of MM in the US starts with identifying what proportion of individuals are diagnosed via the ED route. Assessing EP for MM would reveal whether existing disparities seen in several other cancer contexts, are at play in the prevalence of ED-involved diagnosis of MM. Findings from such an investigation offer important evidence of the burden of MM as related to a suboptimal route of diagnosis. To address this gap, the objectives of this study were to estimate the prevalence of ED involvement in the diagnostic episode of MM among Medicare-enrolled patients over a recent 10-year period and to examine the association with patient and healthcare characteristics. We also described symptoms reported in the year before MM diagnosis in emergency and non-emergency settings.
Methods
We analyzed the Surveillance, Epidemiology and End Results (SEER) program-Medicare database [22]. Our study population included Medicare beneficiaries aged ≥66 years with a new diagnosis of MM (2008–2017). We excluded beneficiaries who had another primary cancer diagnosis, had MM as the secondary cancer or were diagnosed at autopsy or death certificate. To ensure valid estimation of the Charlson Comorbidity Index score and health care utilization in the year before diagnosis, we also excluded anyone not continuously enrolled in fee-for-service Medicare parts A/B, or who had evidence of managed care plan enrollment in the year before diagnosis. This study was reviewed as exempt by The University of North Carolina Institutional Review Board (No. 22–2998).
We defined the primary outcome Emergency Presentation (EP) as an episode of care that includes ED involvement in an initial diagnosis of MM. ED involvement in the initial diagnosis was defined as whether a patient visited the ED in the 30 days before their claims-based index date. This index date was derived using a validated algorithm, defined as the date of the earliest myeloma-specific International Classification of Diseases (ICD) code in the SEER month of diagnosis ± 1 month [23,24]. ICD codes sought from inpatient and outpatient claims were 203.0–203.02 and C90. If no claims were available, the index date was imputed as the 15th day of the SEER month of diagnosis. We categorized EP into three mutually exclusive groups: patients diagnosed during an inpatient hospital admission that originated from an ED visit (EP-IP), patients with an ED visit followed by one or more outpatient specialist encounters that led to diagnosis (EP-OP) and patients diagnosed without any ED interaction (non-EP) (Fig. 1). This EP classification scheme was based on methods described in our prior published work [18].
Fig. 1.

Diagnostic pathways of Multiple Myeloma (MM) in the US. This figure illustrates three distinct pathways through which patients were diagnosed with MM. Emergency Presentation-Inpatient (EP-IP): When patients initially present to the Emergency Department (ED) (Step 1) and are subsequently admitted for inpatient evaluation (Step 2). The inpatient evaluation leads to MM diagnosis. Emergency Presentation-Outpatient (EP-OP): When patients present to the ED (Step 1) and are evaluated and discharged from outpatient without admission (Step 3) with subsequent evaluation leading to MM diagnosis. Non-Emergency Presentation (NON-EP): When patients visit primary care physicians or specialty clinics (Step 4) and evaluation leads to diagnosis of MM without any involvement of ED. Arrows indicate diagnostic pathways: Red arrows indicate EP-IP, orange arrows indicate EP-OP and green arrows indicate NON-EP. Icons of “Emergency Room”, “Primary Care”, “Specialist Care”, “Inpatient” and “Outpatient” indicate point of care. Sample Sizes: Total number of patients diagnosed with MM: 20,437; EP-IP: 5,991, EP-OP: 1,493, NON-EP: 12,953. Data Source: SEER-Medicare (2008–2017)
Demographic variables included age at diagnosis, sex, race and ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic of all races, Asian American and Native Hawaiian or other Pacific Islander, American Indian or Alaskan Native, mixed race and other or unknown race) and year of diagnosis. Social factors include marital or domestic partner status, Medicaid eligibility at the SEER month of diagnosis (none, partial, full) [25], and county socioeconomic status (SES) quintiles [26]. We used the Yost index to measure county SES status [26]. Clinical characteristics included a claims-based Charlson Comorbidity Index (CCI) and a validated claims-based frailty index [27,28]. Both the comorbidity index and the frailty index were ascertained from inpatient, outpatient and carrier claims, with the frailty index also incorporating skilled nursing facility, home health agency, and durable medical equipment claims. Geographic factors include SEER region and county metropolitan status. For county metropolitan status, we used the Rural-Urban Continuum Codes [29] developed by the United States Department of Agriculture (USDA) where counties are classified as metropolitan (counties in metro areas with population fewer than 250,000 to 1 million or more), urban non-metropolitan (urban population of 2500 to 20,000 or more regardless if adjacent or non-adjacent to metro areas), completely rural and unknown. Healthcare utilization history in the year before diagnosis was characterized by the number of ED visits, inpatient admissions, evaluation and management (E&M) outpatient visits (CPT codes 99201–99409 sourced from carrier, outpatient, nursing facility, and home health claims) and receipt of any cancer screening (yes/no) for screen-detectable cancers (breast, cervical, prostate, lung, and colorectal). Codes used to identify screening services included HCPCS G0101–0106, G0202, G0120–0121, G0297, G0328, G0464 [30]. E&M visits represent outpatient healthcare encounters in which a clinician assesses the patient’s health status and plans appropriate clinical management [31]. Diagnosis year as a time-related characteristic was categorized into five intervals spanning 2008 to 2017 (Supplementary Table 1).
We generated a list of known MM symptoms and corresponding ICD9/10 codes through review of published literature (Koo et al., 2018) [32], UpToDate [33] and the American Cancer Society [34] which describe common presenting features of multiple myeloma. Symptom categories included back pain, chest pain, dyspnea, bone pain, joint pain, fracture, nosebleeds, weight loss, nausea, spinal cord compression, peripheral neuropathy, recurrent chest infection, and hyperviscosity; test abnormalities included anemia, thrombocytopenia, leukopenia, hypercalcemia, and bone weakness (from osteoporosis or plasmacytoma). We queried all inpatient outpatient, and ED claims for the corresponding ICD codes to capture all symptom-related healthcare encounters in the year before MM diagnosis (Supplementary Table 2).
Statistical Analysis
Descriptive statistics were used to examine the distribution of demographic, socioeconomic, geographic, health status, and healthcare utilization characteristics overall and stratified by EP category (non-EP, any EP, EP-IP, EP-OP). We described frequency of symptom-related healthcare during the 0–30 days and 31–365 days before diagnosis.
We used Poisson regression with robust standard errors to obtain valid confidence intervals for prevalence ratios when modeling binary outcomes, as recommended in prior methodological work by Spiegelman et al [35]. The models used non-EP as the reference group to estimate prevalence ratios (PRs) and 95% confidence intervals for associations between key patient characteristics and EP status. Prior to fitting the models, we checked for collinearity between the Charlson comorbidity index and the Kim frailty index using correlation coefficients and variance inflation factors. In the data, there was no evidence of substantial multicollinearity in these variables, with a VIF < 2 (Supplementary Table 3). To avoid over-interpretation of coefficients from a single fully adjusted (kitchen-sink) model (i.e., treating mutually adjusted associations as independent effects), we used a sequential/domain-based modeling strategy. Models were fit separately for estimating associations with demographic, social, geographic, clinical and healthcare utilization factors. Model 1 adjusted for demographic (sex, age, race and ethnicity, year of diagnosis) variables; Model 2 included demographic and social (marital status, Medicaid eligibility, county SES) variables; Model 3 included demographic and geography (county rurality, SEER region) variables; Model 4 included demographic, social, geography, and clinical (CCI, frailty) variables; Model 5 included model 4 variables and healthcare utilization variables. The distribution of missing values was examined for each variable and described. For variables with a high proportion of missingness, the missing indicator category was retained and incorporated into the models. Variables with minimal missingness were handled using a complete case approach. All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC).
Results
66,575 patients were diagnosed with multiple myeloma in all SEER regions between 2008–2017. After applying inclusion and exclusion criteria, the final analytic cohort included 20,437 adults diagnosed with multiple myeloma (Supplementary Fig. 1); 52% were male and the median age at diagnosis was 76 years old. Overall, 37% of patients had an EP (30% EP-IP; 7% EP-OP). EP-IPs were older (21% aged ≥ 85 years vs 14%) and more often non-Hispanic Black (17% vs 13%) compared to non-EPs. EP-IP and EP-OP patients more often resided in lower SES counties (19–21% in the lowest quintile vs 14% non-EP). EP-IP patients had higher comorbidity, with 31% having ≥ 3 comorbidities (vs. 20% non-EP) and 29% classified as mildly or moderately to severely frail (vs 13% of non-EPs). Full dual Medicare–Medicaid eligibility was nearly twice as prevalent among EP-IP patients (18%) vs. non-EPs (10%). The vast majority of patients had a claims-based index date but for 10.8% the date was imputed as the 15th of the SEER month. EP-IPs were similarly distributed across all years of diagnosis (Supplementary Table 4), while EP-OPs were more common in 2016–2017 than 2008–2009 (22% vs 17%). EP-OPs had more inpatient stays than non-EP in the prior year (28% vs 23%), and lower rates of any cancer screening than non-EP (32% vs 34%). High use of evaluation and management services was observed among EP-OP patients (56% vs 55% in Non-EP) (Table 1).
Table 1.
Characteristics of SEER-Medicare patients diagnosed with multiple myeloma (2008–2017)
| Characteristics | Overall N (%) | EP-IP N (%) | EP-OP N (%) | NON-EP N (%) | P-value |
|---|---|---|---|---|---|
| All Participants | 20,437 (100) | 5,991 (30) | 1,493 (7) | 12,953 (63) | |
| Demographics | |||||
| Sex | <0.0001 | ||||
| Male | 10,636 (52) | 3,014 (50) | 721 (48) | 6,901 (53) | |
| Female | 9,801 (48) | 2,977 (50) | 772 (52) | 6,052 (47) | |
| Age at Diagnosis | <0.0001 | ||||
| 66–69 years | 3,853 (19) | 1,053 (18) | 277 (19) | 2,523 (19) | |
| 70–74 years | 4,949 (24) | 1,304 (22) | 363 (24) | 3,282 (25) | |
| 75–79 years | 4,463 (22) | 1,272 (21) | 312 (21) | 2,879 (22) | |
| 80–84 years | 3,799 (19) | 1,119 (19) | 263 (18) | 2,417 (19) | |
| ≥85 years | 3,373 (17) | 1,243 (21) | 278 (19) | 1,852 (14) | |
| Race and Ethnicity | <0.0001 | ||||
| Non-Hispanic White | >15,257 (75) | >4,242 (71) | >1,130 (76) | >9,877 (76) | |
| Non-Hispanic Black | 2,833 (14) | 1,015 (17) | 184 (12) | 1,634 (13) | |
| Hispanic | 1,366 (7) | 459 (8) | 113 (8) | 794 (6) | |
| Asian American, Native Hawaiian or Other Pacific Islander | 745 (4) | 243 (4) | 44 (3) | 458 (4) | |
| American Indian or Alaskan Native | 90 (0.5) | 21 (<0.1) | 11 (1) | 58 (0.5) | |
| Mixed/Other/Unknown Race | >142 (1) | 12 (<0.1) | <11 (<0.1) | 123 (<0.1) | |
| Year of Diagnosis | <0.0001 | ||||
| 2008–2009 | 4,067 (20) | 1,204 (20) | 253 (17) | 2,610 (20) | |
| 2010–2011 | 4,149 (20) | 1,219 (20) | 288 (19) | 2,642 (20) | |
| 2012–2013 | 4,088 (20) | 1,214 (20) | 298 (20) | 2,576 (20) | |
| 2014–2015 | 4,061 (20) | 1,195 (20) | 322 (22) | 2,544 (20) | |
| 2016–2017 | 4,072 (20) | 1,159 (19) | 332 (22) | 2,581 (20) | |
| Social Factors | |||||
| Married or domestic partner | <0.05 | ||||
| Yes | 7,868 (38) | 2,032 (34) | 610 (41) | 5,226 (40) | |
| No | 5,897 (29) | 2,037 (34) | 488 (33) | 3,372 (26) | |
| Unknown | 6,672 (33) | 1,922 (32) | 395 (26) | 4,355 (34) | |
| Medicaid Eligibility | <0.0001 | ||||
| None | 17,169 (84) | 4,638 (77) | >1,201 (81) | >11,316 (87) | |
| Partial Dual | 659 (3) | 221 (4) | 62 (4) | 376 (3) | |
| Full Dual | 2,579 (13) | 1,108 (18) | 221 (15) | 1,250 (10) | |
| Missing | 30 (<0.1) | 24 (<0.1) | <11 (<0.1) | <11 (<0.1) | |
| County SES (Quintile) | <0.0001 | ||||
| Q1 (Lowest) | 3,302 (16) | 1,130 (19) | 312 (21) | 1,860 (14) | |
| Q2 | 3,715 (18) | 1,146 (19) | 290 (19) | 2,279 (18) | |
| Q3 | 3,903 (19) | 1,161 (19) | 299 (20) | 2,443 (19) | |
| Q4 | 4,130 (20) | 1,160 (19) | 257 (17) | 2,713 (21) | |
| Q5 (Highest) | 4,430 (22) | 1,140 (19) | 272 (18) | 3,018 (23) | |
| No Claims | 957 (5) | 254 (4) | 63 (4) | 640 (5) | |
| Geography | |||||
| County Metropolitan Status | <0.0001 | ||||
| Metropolitan | >17,514 (86) | >5,255 (88) | >1,133 (77) | >11,116 (86) | |
| Non-metropolitan urban | 2,613 (13) | 657 (11) | 310 (21) | 1,646 (13) | |
| Rural | 295 (1) | 76 (1) | 39 (3) | 180 (1) | |
| Unknown | <11 (<0.1)* | <11 (<0.1)* | <11 (<0.1)* | <11 (<0.1)* | |
| SEER Region | <0.0001 | ||||
| 1-East | 8,231 (40) | 2,610 (44) | 466 (31) | 5,155 (40) | |
| 2-South | 3,967 (19) | 1,086 (18) | 320 (21) | 2,561 (20) | |
| 3-Midwest | 1,864 (9) | 577 (10) | 186 (12) | 1,101 (8) | |
| 4-West | 6,375 (31) | 1,718 (29) | 521 (35) | 4,136 (32) | |
| Clinical Characteristics | |||||
| Charlson Comorbidity Index | <0.0001 | ||||
| 0 | <8,231 (40) | 1,975 (33) | > 647(42) | 5,604 (43) | |
| 1 | 4,308 (21) | 1,170 (20) | 340 (23) | 2,798 (22) | |
| 2 | 2,953 (14) | 898 (15) | 205 (13) | 1,850 (14) | |
| ≥3 | 4,623 (23) | 1,830 (31) | 296 (21) | 2,497 (20) | |
| No Claim | >322 (2) | 118 (2) | <11 (<0.1) | 204 (2) | |
| Kim Frailty Index | <0.0001 | ||||
| Robust (<0.15) | 6,515 (32) | 1,363 (23) | 358 (24) | 4,794 (37) | |
| Pre-frail (0.15–0.24) | 10,217 (50) | 2,904 (48) | 820 (55) | 6,493 (50) | |
| Mildly frail (0.25–0.34) | 2,852 (14) | 1,243 (21) | 243 (16) | 1,366 (11) | |
| Moderate to Severely frail (≥0.35) | 853 (4) | 481 (8) | 72 (5) | 300 (2) | |
| Healthcare Utilization in Year Prior to Diagnosis | |||||
| Count of ED visits | <0.0001 | ||||
| 0 | 11,837 (58) | 2,781 (46) | 677 (45) | 8,379 (65) | |
| 1 | 4,603 (22) | 1,453 (24) | 376 (25) | 2,774 (21) | |
| 2 | 2,023 (10) | 810 (14) | 180 (12) | 1,033 (8) | |
| ≥3 | 1974 (10) | 947 (16) | 260 (18) | 767 (6) | |
| In patient Days | <0.0001 | ||||
| No | 15,864 (78) | 4,776 (80) | 1,081 (72) | 10,007 (77) | |
| Yes | 4,573 (22) | 1,215 (20) | 412 (28) | 2,946 (23) | |
| Evaluation and Management Visits | <0.0001 | ||||
| None | 851 (4) | 405 (7) | 40 (3) | 406 (3) | |
| Low Use (1–3 visits) | 1,711 (8) | 793 (13) | 119 (8) | 799 (6) | |
| Moderate Use (4–10 visits) | 7,391 (36) | 2,284 (38) | 499 (33) | 4,608 (36) | |
| High Use (11 or more visits) | 10,484 (51) | 2,509 (42) | 835 (56) | 7,140 (55) | |
| Any Cancer Screening | <0.0001 | ||||
| Yes | 6,303 (31) | 1,372 (23) | 479 (32) | 4,452 (34) | |
| No | 14,134 (69) | 4,619 (77) | 1,014 (68) | 8,501 (66) | |
Categories have been combined (coarsened) to meet CMS minimum-cell-size policy (no cells with n = 1–10)
Associations between patient characteristics and emergency presentation
Modeling (Table 2, Model 1) showed that older age groups compared to age group of 66–69, and non-Hispanic Black (NHB), Hispanic and Asian American, Native Hawaiian or Other Pacific Islander (AANHOPI) racial groups compared to non-Hispanic White were associated with greater likelihood of EP overall (PRolder: 1.33; 95% CI:1.25, 1.41; PRNHB: 1.23; 95% CI:1.17, 1.29; PRHispanic: 1.22; 95% CI:1.14, 1.30; and PRAANHOPI: 1.10; 95% CI: 1.01, 1.21, vs Non-Hispanic White). Being female was associated with EP overall; the association was stronger for EP-OP (PR:1.19; 95% CI: 1.08, 1.31) than EP-IP (PR:1.04; 95% CI:1.00, 1.09).
Table 2.
Association between patient sociodemographic, clinical, and tumor characteristics and ED involvement among SEER-Medicare patients diagnosed with multiple myeloma (2008–2017)
| Characteristic | Prevalence Ratio (95% CI) | ||
|---|---|---|---|
| Any EP vs. Non-EP (N=7,484 vs. 12,953) | EP-IP vs. Non-EP (N=5,991 vs. 12,953) | EP-OP vs. Non-EP (N=1,493 vs. 12,953) | |
| Demographics 1 | |||
| Sex | |||
| Male | Ref. | Ref. | Ref. |
| Female | 1.06 (1.02, 1.10) | 1.04 (1.00, 1.09) | 1.19 (1.08, 1.31) |
| Age (years) | |||
| 66–69 | Ref. | Ref. | Ref. |
| 70–74 | 0.98 (0.93, 1.04) | 0.97 (0.91, 1.04) | 1.01 (0.87, 1.17) |
| 75–79 | 1.03 (0.98, 1.10) | 1.05 (0.98, 1.12) | 0.99 (0.85, 1.16) |
| 80–84 | 1.07 (1.00, 1.13) | 1.09 (1.02, 1.17) | 1.00 (0.85, 1.18) |
| ≥85 | 1.33 (1.25, 1.41) | 1.40 (1.30, 1.49) | 1.32 (1.13, 1.54) |
| Race and Ethnicity | |||
| Non-Hispanic White | Ref. | Ref. | Ref. |
| Non-Hispanic Black | 1.23 (1.17, 1.29) | 1.31 (1.24, 1.39) | 0.98 (0.85, 1.17) |
| Hispanic | 1.22 (1.14, 1.30) | 1.25 (1.16, 1.36) | 1.21 (1.01, 1.45) |
| Asian American, Native Hawaiian or Other Pacific Islander | 1.10 (1.01, 1.21) | 1.17 (1.06, 1.30) | 0.85 (0.64, 1.13) |
| American Indian or Alaskan Native | 1.03 (0.79, 1.36) | 0.91 (0.64, 1.31) | 1.55 (0.90, 2.66) |
| Year of Diagnosis | |||
| 2008–2009 | Ref. | Ref. | Ref. |
| 2010–2011 | 1.01 (0.96, 1.07) | 1.00 (0.94, 1.07) | 1.11 (0.94, 1.30) |
| 2012–2013 | 1.03 (0.98, 1.09) | 1.02 (0.95, 1.08) | 1.17 (1.00, 1.38) |
| 2014–2015 | 1.05 (0.99, 1.11) | 1.02 (0.95, 1.09) | 1.27 (1.09, 1.49) |
| 2016–2017 | 1.03 (0.97, 1.09) | 0.99 (0.93, 1.06) | 1.30 (1.11, 1.52) |
| Social Factors 2 | |||
| Married or Domestic Partner | |||
| Yes | Ref. | Ref. | Ref. |
| No | 1.14 (1.09, 1.19) | 1.19 (1.12, 1.25) | 1.04 (0.92, 1.17) |
| Unknown | 0.99 (0.94, 1.03) | 1.04 (0.98, 1.09) | 0.76 (0.67, 0.86) |
| Medicaid Eligibility | |||
| None | Ref. | Ref. | Ref. |
| Partial Dual | 1.15 (1.05, 1.26) | 1.15 (1.03, 1.28) | 1.30 (1.02, 1.65) |
| Full Dual | 1.41 (1.34, 1.48) | 1.48 (1.40, 1.56) | 1.49 (1.29, 1.72) |
| County SES (Quintile) | |||
| Q1 (Lowest) | 1.22 (1.15, 1.29) | 1.19 (1.11, 1.27) | 1.62 (1.39, 1.89) |
| Q2 | 1.14 (1.07, 1.20) | 1.13 (1.06, 1.21) | 1.27 (1.10, 1.48) |
| Q3 | 1.13 (1.07, 1.19) | 1.12 (1.05, 1.20) | 1.26 (1.08, 1.46) |
| Q4 | 1.06 (1.00, 1.12) | 1.07 (1.00, 1.14) | 1.01 (0.87, 1.18) |
| Q5 (Highest) | Ref. | Ref. | Ref. |
| Geography 3 | |||
| County Metropolitan Status | |||
| Metropolitan | Ref. | Ref. | Ref. |
| Urban, non-metropolitan | 1.06 (1.00, 1.12) | 0.95 (0.88, 1.02) | 1.61 (1.43, 1.82) |
| Rural | 1.14 (0.99, 1.32) | 1.02 (0.84, 1.23) | 1.82 (1.35, 2.46) |
| SEER-Region | |||
| West | Ref. | Ref. | Ref. |
| South | 0.99 (0.94, 1.05) | 1.02 (0.95, 1.09) | 0.92 (0.80, 1.06) |
| Midwest | 1.16 (1.08, 1.23) | 1.18 (1.09, 1.27) | 1.19 (1.01, 1.40) |
| East | 1.06 (1.02, 1.11) | 1.14 (1.08, 1.20) | 0.77 (0.69, 0.87) |
| Clinical Characteristics 4 | |||
| Charlson Comorbidity Index | |||
| 0 | Ref. | Ref. | Ref. |
| 1 | 0.94 (0.90, 0.99) | 0.96 (0.90, 1.02) | 0.87 (0.76, 0.98) |
| 2 | 0.91 (0.86, 0.97) | 0.96 (0.90, 1.03) | 0.71 (0.61, 0.83) |
| ≥3 | 0.96 (0.91, 1.01) | 1.04 (0.98, 1.11) | 0.62 (0.53, 0.72) |
| No Claim | 1.37 (1.15, 1.55) | 1.59 (1.37, 1.85) | Low |
| Kim Frailty Index | |||
| Robust (<0.15) | Ref. | Ref. | Ref. |
| Pre-frail (0.15–0.24) | 1.39 (1.32, 1.46) | 1.39 (1.30, 1.47) | 1.74 (1.54, 1.98) |
| Mildly frail (0.25–0.34) | 1.88 (1.76, 2.01) | 1.95 (1.81, 2.11) | 2.55 (2.14, 3.04) |
| Moderate to Severely frail (≥0.35) | 2.27 (2.10, 2.46) | 2.42 (2.21, 2.65) | 3.35 (2.59, 4.33) |
| Healthcare Utilization in Year Prior to Diagnosis 5 | |||
| Count of ED Visits | |||
| 0 | Ref. | Ref. | Ref. |
| 1 | 1.12 (1.07, 1.18) | 1.04 (0.99, 1.10) | 1.58 (1.39, 1.80) |
| 2 | 1.26 (1.19, 1.34) | 1.17 (1.10, 1.26) | 1.99 (1.69, 2.36) |
| ≥3 | 1.43 (1.35, 1.52) | 1.31 (1.23, 1.41) | 3.39 (2.89, 3.98) |
| Inpatient Days | |||
| No | Ref. | Ref. | Ref. |
| Yes | 1.55 (1.48, 1.62) | 1.88 (1.78, 1.98) | 0.77 (0.68, 0.88) |
| Evaluation and Management Visits | |||
| None | 1.89 (1.75, 2.04) | 2.22 (2.05, 2.41) | 1.18 (0.89, 1.56) |
| Low Use (1–3 visits) | 1.84 (1.74, 1.94) | 2.10 (1.98, 2.24) | 1.43 (1.20, 1.71) |
| Moderate Use (4–10 visits) | 1.34 (1.29, 1.40) | 1.48 (1.41, 1.55) | 1.07 (0.96, 1.20) |
| High Use (≥ 11 visits) | Ref. | Ref. | Ref. |
| Any Cancer Screening | |||
| Yes | 0.90 (0.86, 0.94) | 0.85 (0.81, 0.90) | 1.02 (0.91, 1.13) |
| No | Ref. | Ref. | Ref. |
Prevalence ratios estimated using generalized linear models with log link and modified Poisson distribution. Model 1 variables included demographic exposures: sex, age group, race and ethnicity and year of diagnosis.
Model 2 includes model 1 and married or domestic partner, Medicaid eligibility and county SES (quintile)
Model 3 includes model 1 and county metropolitan status and SEER-Region
Model 4 includes model 2, 3 and Charlson comorbidity index and Kim frailty index
Model 5 includes model 4 and count of ED visits, inpatient days, Evaluation and management visits, and any cancer screening
Model 2 shows being unmarried/unpartnered was associated with EP compared to married patients (PR: 1.14; 95% CI:1.09, 1.19). Those living in the lowest quintile of county SES had higher prevalence of EP (PR: 1.22; 95% CI:1.15, 1.29) with higher likelihood for EP-OP (PR: 1.62; 95% CI:1.39, 1.89) than EP-IP (PR: 1.19; 95% CI:1.11, 1.27). Being dual eligible for Medicaid and Medicare was strongly associated with EP overall (PR: 1.41; 95% CI:1.34, 1.48) compared to non-dual enrollees.
Model 3 shows that county geography was associated with EP overall [(PR: 1.06; 95% CI:1.00, 1.12) for non-metropolitan urban and (PR: 1.14; 95% CI:0.99, 1.32) for rural, vs metropolitan counties], as well as with EP-OP [(PR: 1.61; 95% CI:1.43–1.82) for non-metropolitan urban and (PR: 1.82; 95% CI:1.35, 2.46) for rural, vs metropolitan counties].
Model 4 shows that Charlson comorbidity was not associated with EP; however, frailty was: overall, when compared to the minimally frail group, pre-, mildly and moderately to severely frail patients had a higher likelihood of any EP (PR: 2.27; 95% CI:2.10, 2.46) (Table 2).
Model 5 shows that a higher frequency of ED visits in the year prior to diagnosis was associated with a greater likelihood of EP (PR:1.43; 95% CI: 1.35, 1.52 for ≥3 visits, vs no ED visit). Patients with no or low E&M consults had a significantly higher prevalence of EP compared with those with high use (≥11 visits), with PRs of 1.89 (95% CI:1.75, 2.04) and 1.84 (95% CI: 1.74, 1.94) respectively. Inpatient stay in the year before diagnosis was associated with any EP (PR: 1.55; 95% CI:1.48, 1.62). This association was stronger for EP-IP (PR: 1.88; 95% CI:1.48, 1.62) while for EP-OP, prior inpatient stay was protective (PR: 0.77; 95% CI:0.68, 0.88) (Table 2). The full estimates for all models in each sequential fitting are shown in Supplementary Tables 5a, b and c.
Symptom related healthcare utilization patterns
In the 30 days before diagnosis, the most frequent symptom related claims were for anemia (46%), back pain (25%), chest pain (19%), bone weakness (18%), and dyspnea (17%). Among patients with any EP, nearly all symptoms were more common than in non-EP patients, particularly for anemia (57% vs 40%, p<0.001), back pain (38% vs 17%), chest pain (37% vs 9%), dyspnea (31% vs 9%), fracture (29% vs 10%), and hypercalcemia (27% vs 5%). Among patients with an EP, EP-IP patients had higher frequencies of anemia, hypercalcemia, and dyspnea-related claims, whereas EP-OP patients had higher frequencies of bone weakness and joint pain.
Comparing the preceding 31–365 day and the 0–30 day windows, overall claims (among all patients) associated with anemia increased from 42% to 46%, bone weakness from 12% to 18%, while hypercalcemia tripled from 4% to 13% (all p<0.001). Among any EP patients, the increase was greater; anemia from 39% to 57%, chest pain from 35% to 37%, and fracture from 14% to 29% (Table 3). The distribution of CRAB symptoms (related to hypercalcemia, renal insufficiency, anemia and lytic bone lesions) varied; EP-IP patients were the most likely to have ≥2 CRAB symptoms (43%) and non-EP patients were the least likely (13%) (Supplementary Table 6). The number of symptom-related encounters was similar between patients with any EP and non-EP patients 31–365 days before diagnosis, but EP patients experienced markedly higher number of symptoms in the 30 days leading up to diagnosis (Figs. 2 and 3, Supplemental Fig. 2, 3 and 4).
Table 3.
Distribution of symptom related healthcare utilization in the year prior to multiple myeloma diagnosis among patients in SEER-Medicare (2008–2017)
| 0–30 Days Prior Diagnosis | 31–365 Days Prior Diagnosis | P-value** | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Overall N (%) | Any EP N (%) | EP-IP N (%) | EP-OP N (%) | NON-EP N (%) | Overall N (%) | Any EP N (%) | EP-IP N (%) | EP-OP N (%) | NON-EP N (%) | |
| 20,437 (100) | 7,484 (37) | 5,991 (30) | 1,493 (7) | 12,953 (63) | 20,437 (100) | 7,484 (37) | 5,991 (30) | 1,493 (7) | 12,953 (63) | <0.001 |
| 9,483 (46) | >4,259 (57) | >3,555 (60) | 704 (47) | >5,213 (40) | 8,528 (42) | 2,870 (38) | 2,298 (38) | >558 (38) | 5,658 (44) | <0.001 |
| 5,102 (25) | 2,865 (38) | 2,287 (38) | 578 (39) | 2,237 (17) | 6,475 (32) | 2,551 (34) | 1,999 (33) | 552 (37) | 3,924 (30) | 0.004 |
| 3,917 (19) | 2,736 (37) | 2,242 (37) | 494 (33) | 1,181 (9) | 6,549 (32) | 2,611 (35) | 2,035 (34) | 576 (39) | 3,938 (30) | <0.001 |
| 3,738 (18) | 1,620 (22) | 1,199 (20) | 421 (28) | 2,118 (16) | 2,514 (12) | 763 (10) | 572 (10) | 191 (13) | 1,751 (14) | <0.001 |
| 3,485 (17) | 2,350 (31) | 2,049 (34) | 301 (20) | 1,135 (9) | 5,795 (28) | 2,273 (30) | 1,821 (30) | 452 (30) | 3,522 (27) | <0.001 |
| 3,401 (17) | 2,163 (29) | 1,759 (29) | 404 (27) | 1,238 (10) | 2,518 (12) | 1,078 (14) | 865 (14) | 213 (14) | 1,440 (11) | <0.001 |
| 3,009 (15) | 1,588 (21) | 1,233 (21) | 355 (24) | 1,421 (11) | 6,293 (31) | 2,281 (31) | 1,784 (30) | 497 (33) | 4,012 (31) | <0.001 |
| 2,716 (13) | 2,041 (27) | 1,873 (31) | 168 (11) | 675 (5) | 706 (4) | 259 (4) | 206 (3) | 53 (4) | 447 (4) | <0.001 |
| 1,747 (9) | 1,076 (14) | 986 (17) | 90 (6) | 671 (5) | 1,033 (5) | 341 (5) | 270 (5) | 71 (5) | 692 (5) | <0.001 |
| 1,643 (8) | 920 (12) | 792 (13) | 128 (9) | 723 (6) | 1,591 (8) | 603 (8) | 471 (8) | 132 (9) | 988 (8) | 0.34 |
| 937 (5) | 703 (9) | 578 (10) | 125 (8) | 234 (2) | 1,528 (8) | 618 (8) | 478 (8) | 140 (9) | 910 (7) | <0.001 |
| 770 (4) | 250 (3) | 190 (3) | 60 (4) | 520 (4) | 686 (3) | 178 (2) | 138 (2) | 40 (3) | 508 (4) | 0.025 |
| 623 (3) | 190 (3) | 142 (2) | 48 (3) | 433 (3) | 530 (3) | 128 (2) | 89 (2) | 39 (3) | 402 (3) | 0.006 |
| 531 (3) | 236 (3) | 176 (3) | 60 (4) | 295 (2) | 441 (2) | 130 (2) | 86 (1) | 44 (3) | 311 (2) | 0.004 |
| 408 (2) | 286 (4) | 252 (4) | 34 (2) | 122 (1) | 95 (1) | 35 (1) | 28 (1) | <11 (0.1) | 60 (1) | <0.001 |
| 307 (2) | 146 (2) | 120 (2) | 26 (2) | 161 (1) | 1,822 (9) | 629 (8) | 493 (8) | 136 (9) | 1,193 (9) | <0.001 |
| 260 (1) | 176 (2) | 134 (2) | 42 (3) | 84 (1) | 555 (3) | 218 (3) | 164 (3) | 54 (4) | 337 (3) | <0.001 |
| 11 (<0.1)* | <11 (<0.1)* | <11 (<0.1)* | 0 (0) | <11 (<0.1)* | 56 (<0.1) | 20 (<0.1) | 19 (<0.1) | <11 (<0.1)* | 36 (<0.1) | <0.001 |
Categories have been combined (collapsed) to meet CMS minimum-cell-size policy (no cells with n = 1–10)
P-values compare symptom prevalence between the 0–30-day and 31–365-day pre-diagnosis periods based on the overall counts and were calculated using Pearson χ2 tests.
Fig. 2.

Count of distinct billed symptoms for healthcare encounters at various time periods prior to an emergency department-involved diagnosis of multiple myeloma versus Non-EP (Graph A), and diagnosis associated with an inpatient admission or outpatient care (EP-IP and EP-OP) versus Non-EP (Graph B)
Fig. 3.

Proportion of patients with cumulative billed symptom encounters in time points (months) leading up to the diagnosis multiple myeloma in non-emergency presenters (A) and emergency presenters (B)
Discussion
To our knowledge, no prior US studies have examined EP in older adults with MM. We observed that more than one third of MM patients were diagnosed via EP. We also found that socioeconomically vulnerable and frail patients with lack of outpatient service utilization and multiple ED visits in the year prior to diagnosis were more likely to be EPs, pointing to different barriers to timely diagnosis. Healthcare utilization for symptoms such as anemia, hypercalcemia, dyspnea, bone-related issues increased in the month before diagnosis compared to 31–365 days before diagnosis. Symptom escalation close to diagnosis could indicate missed opportunities for earlier MM detection. The occurrence of EP may represent gaps in access to primary and specialist care as well as acute care use among severe MM cases. We examined EP stratified into two distinct care pathways that could encompass these scenarios [23]. The EP-IP group having more severe symptoms indicates greater clinical acuity, while the lack of post-ED hospitalization for the EP-OP group likely reflects patients presenting to the ED due to socioeconomic barriers to accessing preventive care. EP-OP patients lived more in rural areas and had greater number of ED visits prior to diagnosis, supporting the notion that the ED serves as a main source of care in this population. These findings emphasize that the drivers of EP are heterogeneous and should be taken into account for future studies.
In one UK investigation of MM patients diagnosed from 2012 to 2013, the prevalence of EP was 28% [20]. The Tackling EArly Morbidity and Mortality in Multiple Myeloma (TEAMM) trial which explored diagnostic pathways in MM observed that 29% of participants were diagnosed via ED consultation, or an acute medical or surgical admission [19]. The prevalence of EP identified in our study is most comparable to another UK study which observed that 37% of patients with MM were emergency presenters between 2006–2008 [11,36]. This similarity with our findings is striking given the differences in healthcare systems, racial/ethnic distributions and social drivers of health between the UK and the US. The equivalence of these numbers suggests a biological underpinning of MM that is consistent across different populations, making it difficult to diagnose. Prior research has shown that EP patients are more likely to have CRAB symptoms at time of diagnosis. Howell et al. reported that among MM patients diagnosed in EDs, 56.5% had at least 2 CRAB symptoms [20]. Patients who were diagnosed via an ED route in the TEAMM trial had a higher burden of renal impairment, anemia and bone disease than other routes to diagnosis [19]. We found that over 40% of EPs had 2 or more CRAB symptoms and identified chest and back pain as the most prevalent symptoms among patients diagnosed through the ED, especially in the month leading up to diagnosis. This is consistent with recent studies of patients with MM in the UK [7,37]. The prevalence of EP-IP was similar across the years, while the EP-OP route to diagnosis increased slightly in later years. In future studies, investigating changes in EP prevalence over time may reveal whether improvements in primary care management and greater awareness of MM have minimized this diagnostic pathway.
We observed that EP was more common among patients who were older, frailer, unpartnered, and living in lower SES counties, as well as patients who were not non-Hispanic White. This emphasizes the importance of social drivers of ED involvement in the MM diagnostic pathway rather than disease biology alone. The finding that older age and moderate-to-severe frailty were significantly associated with EP of MM is consistent with prior studies [15,18,38]. Frailty is associated with falls and increases with age, which may result in emergency care among older adults [39]. We observed that the EP-IP diagnostic pathway was more common among older adults. Inpatient admissions are more common for older patients with a high comorbidity burden [40]. The observation that being unpartnered was associated with EP has been similarly observed in older adults with solid tumors [18]. Partnership confers multiple advantages for timely cancer diagnosis including insurance access, higher education, income and interpersonal support [41–43]. The association between low SES and EP status is an extensively documented finding that reflects limited access to healthcare resources that could promote timely cancer diagnosis [17,18,44,45]. The observed disparities in EP risk for non-Hispanic Black, AANHOPI and Hispanic patients compared to non-Hispanic White patients, are findings that have been observed in patients with other cancers in the SEER-Medicare population [18,44], in Michigan [45] and in Indiana [17]. The higher proportion of non-Hispanic Black and AANHOPI patients being represented in the EP-IP group contributes to the broad literature on disparate inpatient hospitalization rates for racial/ethnic minoritized patients with cancer [46–48]. These factors are associated with worse cancer outcomes and less receipt of guideline-concordant treatment, emphasizing the compounding effect of EP across the cancer continuum.
Our findings suggest that the type and quality of pre-diagnostic care shapes the diagnostic setting and, given the established associations between EP and poor outcomes, ultimately impacts patient prognosis. An earlier study of UK patients who were emergency presenters showed that patients with diagnostically challenging cancers (including MM) had frequent general practitioner encounters prior to the ED visit that resulted in their cancer diagnosis [49]. We found that evaluation and management visits and prior cancer screening were associated with lower EP prevalence while frequent prior ED use was associated with an increased EP prevalence. Cancer screening may reflect an established primary care relationship, which provides more opportunities for MM detection. Outpatient E&M visits emphasize the intentional time and effort spent in assessment and management of patients, rather than acute tests and procedures [31]. Longitudinal engagement across multiple non-ED healthcare settings provides opportunities for recognizing and linking symptoms that could lead to earlier MM diagnosis. Frequent prior ED visits preceding EP suggest missed diagnostic opportunities as EDs focus on acute management and are not designed to recognize evolving patterns suggestive of MM precursor stages. The observed strong association between ED visits and EP-OP could indicate a population that seeks out care in the ED regularly and for whom severity of symptoms did not lead to hospitalization [50,51]. These findings may reflect gaps in primary care access, with patients using EDs for ongoing ambulatory care despite their structural limitations for detecting slowly progressive conditions like MM.
The temporal symptom escalation pattern (uptick of symptoms in the month before diagnosis) we observed suggests that EPs experience acute symptom crises that drive them to emergency care, whereas non-EP patients may have chronic, persistent symptoms that are gradually worked up in outpatient settings. These patterns relate to both biology and care access differences. The findings imply that symptom-related diagnoses are documented during healthcare encounters throughout the year prior to MM diagnosis but escalate in the final weeks for those who present emergently. These distinct trajectories suggest that symptom education may help patients recognize when to seek urgent versus routine care for ambiguous symptoms. The higher prevalence of documented anemia, back pain, chest pain, dyspnea and hypercalcemia in EP patients in the month before diagnosis reflects manifestations of CRAB criteria [52]. CRAB includes clinical presentations that are non-specific to MM and may be overlooked or attributed to other conditions. Escalation of these clinical manifestations may lead to an EP. Healthcare institutions that solely focus on timely cancer diagnosis could potentially be feasible intermediaries between primary care and oncology as they can help differentiate MM from other conditions that present with these symptoms. These institutions, often referred to as Undiagnosed Cancer Clinics, include multidisciplinary teams that examine symptomatic patients who do not have a confirmed diagnosis of cancer and can reduce wait times between abnormal test results to definitive diagnosis [53–55]. The EP-IP group had slightly higher frequencies of thrombocytopenia, which is associated with high-risk cytogenetics and poor survival [56]. Because inpatient admissions are associated with severe disease cases, the higher frequencies of symptoms in EP-IP patients likely reflect acute clinical presentations. Interestingly, bone weakness and joint pain were more frequently observed in the EP-OP group. This suggests that musculoskeletal complaints have lower acuteness and are less likely to trigger admission.
Because MM has a heterogeneous symptom presentation, warning signs are not easily recognized or linked to the condition until symptoms worsen. Early recognition of signs is challenging in emergency environments where there is little time for systematic disease detection. There is no widely accepted screening protocol for MM due to lack of published evidence on screening effectiveness [57]. All cases of MM originate from monoclonal gammopathy of undetermined significance (MGUS) which sometimes progresses to smoldering MM, before symptomatic MM [58–60]. However, not all patients with MGUS eventually develop MM [61,62], making it challenging to identify which cases will progress. While mounting evidence examines the feasibility of early screening and intervention for MM and its precursors [63], differentiating patients at highest risk of developing active MM remains a critical step for early detection. Awareness of non-specific signs and symptoms that should raise suspicion is important to ensure prompt presentation to appropriate care pathways and timely referrals rather than delayed diagnosis through emergency presentation. One ideal diagnostic pathway could be direct referral to hematology/oncology from primary care or monoclonal gammopathy testing if CRAB symptoms are observed. As primary care increasingly focuses on chronic disease management, targeted diagnostic efforts may arise from clinical experiences that raise suspicion for malignancy. These results illustrate that MM follows varying presentation trajectories based on distinct symptom phenotypes that could inform prediction of diagnostic pathways.
Study strengths and limitations
While this study uses nationally representative data to quantify EP for patients with MM, there are limitations. We relied on claim-based index dates which may not exactly define the date of MM diagnosis; however, this approach may reduce misclassification of pre-diagnostic healthcare utilization more than would reliance on registry data alone [64]. Claims data inherently lack clinical detail (e.g., lab testing results) that might further explain decisions made in the ED related to inpatient admission or outpatient referral. Severity of presentation, especially cancer stage, may explain EP; however, we lacked stage data. Race and ethnicity data are based on submitted medical records and might be misclassified due to variations in the coding conventions of different reporting facilities [65]. These findings may not generalize beyond the older U.S. adult Medicare population. The SEER-Medicare dataset and, consequently, our study, is restricted to patients 66 or older; though the majority (60%) of incident cases with MM is concentrated in this age group [66], many cases occur in younger patients. It is not possible to infer from our data the proportion of younger patients who would have been diagnosed as emergencies. Furthermore, the investigation could not capture pathways to diagnosis for patients without health insurance, who may have had substantial ED involvement with their diagnosis. Research from countries with national health systems with free point-of-need care indicates a U-shaped association between age and risk of emergency diagnosis in MM patients, with both older and younger patients being at higher risk [21,67]. However, the generalizability of those findings in the U.S. setting is uncertain, given differences in insurance coverage for younger patients. Further research in U.S. populations should aim to address this question. We also excluded beneficiaries who were not continuously enrolled in Medicare FFS (including those enrolled in Medicare Advantage plans, which are not available in SEER-Medicare). However, this approach enhances internal validity in the measurement of EP prevalence.
Conclusion and implications
The EP of cancer is an important measure that is predictive of morbidity and mortality and other downstream health effects. Our study adds to the growing body of evidence that calls for efforts towards (1) improved health care access for individuals with social identities that limit their access to quality health care, (2) streamlining the early diagnosis process for MM, which remains a cancer that is hard to detect at an early stage, and (3) promoting the awareness of symptoms that could be warning signs for MM diagnosis so as to precipitate timely clinical and laboratory evaluation. Further research is needed to investigate patient health care trajectories that contribute to the delayed diagnosis of MM in the ED. These may include documented multiple visits for the same symptom, referral pathways, and patterns of presenting symptoms. Addressing these gaps would improve our understanding of patient groups that may benefit from intervention strategies that minimize the risk of EP.
Supplementary Material
Acknowledgements:
This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the National Cancer Institute; Information Management Services (IMS), Inc.; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database. The collection of cancer incidence data used in this study was supported by the California Department of Public Health pursuant to California Health and Safety Code Section 103885; Centers for Disease Control and Prevention’s (CDC) National Program of Cancer Registries, under cooperative agreement NU58DP007156; the National Cancer Institute’s Surveillance, Epidemiology and End Results Program under contract HHSN261201800032I awarded to the University of California, San Francisco, contract HHSN261201800015I awarded to the University of Southern California, and contract HHSN261201800009I awarded to the Public Health Institute. The ideas and opinions expressed herein are those of the author(s) and do not necessarily reflect the opinions of the State of California, Department of Public Health, the National Cancer Institute, and the Centers for Disease Control and Prevention or their Contractors and Subcontractors. We also acknowledge the support we received from members of Thompson lab.
Funding:
This research was supported by a grant from the NIH National Cancer Institute (R01CA264176).
Footnotes
Competing Interests: The authors have no relevant financial or non-financial interests to disclose.
Data Availability:
The data analyzed during the current study are not publicly available due to protections for individual privacy. Investigators may obtain a Data Use Agreement directly from the National Cancer Institute to access the SEER-Medicare data upon approval of a specific research project. Further information is available at https://healthcaredelivery.cancer.gov/seermedicare/obtain/.
References
- [1].National Cancer Institute. Myeloma - Cancer Stat Facts [Internet]. SEER. https://seer.cancer.gov/statfacts/html/mulmy.html [Accessed June 25, 2025]. [Google Scholar]
- [2].Zorlu T, Kayer MA, Okumus N, Ulaş T, Dal MS, Altuntas F (2025) Challenges, Difficulties, and Delayed Diagnosis of Multiple Myeloma. Diagnostics (Basel) 15: 1708. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Black GB, Boswell L, Harris J, Whitaker KL (2023) What causes delays in diagnosing blood cancers? A rapid review of the evidence. Prim Health Care Res Dev 24: e26. 10.1017/S1463423623000129 e26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Bhutani M, Blue BJ, Cole C, Badros AZ, Usmani SZ, Nooka AK, Bernal-Mizrachi L, Mikhael J (2023) Addressing the disparities: the approach to the African American patient with multiple myeloma. Blood Cancer J 13: 189. 10.1038/s41408-023-00961-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Kariyawasan CC, Hughes DA, Jayatillake MM, Mehta AB (2007) Multiple myeloma: causes and consequences of delay in diagnosis. QJM 100: 635–40. 10.1093/qjmed/hcm077 [DOI] [PubMed] [Google Scholar]
- [6].Lyratzopoulos G, Neal RD, Barbiere JM, Rubin GP, Abel GA (2012) Variation in number of general practitioner consultations before hospital referral for cancer: findings from the 2010 National Cancer Patient Experience Survey in England. The Lancet Oncology 13: 353–65. 10.1016/S1470-2045(12)70041-4 [DOI] [PubMed] [Google Scholar]
- [7].Howell DA, Smith AG, Jack A, Patmore R, Macleod U, Mironska E, Roman E (2013) Time-to-diagnosis and symptoms of myeloma, lymphomas and leukaemias: a report from the Haematological Malignancy Research Network. BMC Blood Disord 13: 9. 10.1186/2052-1839-13-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Koshiaris C, Oke J, Abel L, Nicholson BD, Ramasamy K, Van den Bruel A (2018) Quantifying intervals to diagnosis in myeloma: a systematic review and meta-analysis. BMJ Open 8: e019758. 10.1136/bmjopen-2017-019758 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Rajkumar SV, Dimopoulos MA, Palumbo A, Blade J, Merlini G, Mateos M-V, Kumar S, Hillengass J, Kastritis E, Richardson P, Landgren O, Paiva B, Dispenzieri A, Weiss B, LeLeu X, Zweegman S, Lonial S, Rosinol L, Zamagni E, Jagannath S, Sezer O, Kristinsson SY, Caers J, Usmani SZ, Lahuerta JJ, Johnsen HE, Beksac M, Cavo M, Goldschmidt H, Terpos E, Kyle RA, Anderson KC, Durie BGM, Miguel JFS (2014) International Myeloma Working Group updated criteria for the diagnosis of multiple myeloma. The Lancet Oncology 15: e538–e48. 10.1016/S1470-2045(14)70442-5 [DOI] [PubMed] [Google Scholar]
- [10].Shephard EA, Neal RD, Rose P, Walter FM, Litt EJ, Hamilton WT (2015) Quantifying the risk of multiple myeloma from symptoms reported in primary care patients: a large case–control study using electronic records. Br J Gen Pract 65: e106–e13. 10.3399/bjgp15X683545 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Elliss-Brookes L, McPhail S, Ives A, Greenslade M, Shelton J, Hiom S, Richards M (2012) Routes to diagnosis for cancer – determining the patient journey using multiple routine data sets. Br J Cancer 107: 1220–6. 10.1038/bjc.2012.408 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].McPhail S, Elliss-Brookes L, Shelton J, Ives A, Greenslade M, Vernon S, Morris EJA, Richards M (2013) Emergency presentation of cancer and short-term mortality. Br J Cancer 109: 2027–34. 10.1038/bjc.2013.569 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Pham TM, Gomez-Cano M, Salika T, Jardel D, Abel GA, Lyratzopoulos G (2019) Diagnostic route is associated with care satisfaction independently of tumour stage: Evidence from linked English Cancer Patient Experience Survey and cancer registration data. Cancer Epidemiol 61: 70–8. 10.1016/j.canep.2019.04.011 [DOI] [PubMed] [Google Scholar]
- [14].Herbert A, Abel GA, Winters S, McPhail S, Elliss-Brookes L, Lyratzopoulos G (2019) Are inequalities in cancer diagnosis through emergency presentation narrowing, widening or remaining unchanged? Longitudinal analysis of English population-based data 2006–2013. J Epidemiol Community Health 73: 3–10. 10.1136/jech-2017-210371 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].McPhail S, Swann R, Johnson SA, Barclay ME, Abd Elkader H, Alvi R, Barisic A, Bucher O, Clark GRC, Creighton N, Danckert B, Denny CA, Donnelly DW, Dowden JJ, Finn N, Fox CR, Fung S, Gavin AT, Gomez Navas E, Habbous S, Han J, Huws DW, Jackson CGCA, Jensen H, Kaposhi B, Kumar SE, Little AL, Lu S, McClure CA, Møller B, Musto G, Nilssen Y, Saint-Jacques N, Sarker S, te Marvelde L, Thomas RS, Thomas RJS, Thomson CS, Woods RR, Zhang B, Lyratzopoulos G, Filsinger B, Forster K, May L, Morrison DS, Thomas AF, Warlow JL, You H (2022) Risk factors and prognostic implications of diagnosis of cancer within 30 days after an emergency hospital admission (emergency presentation): an International Cancer Benchmarking Partnership (ICBP) population-based study. The Lancet Oncology 23: 587–600. 10.1016/S1470-2045(22)00127-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Thompson CA, He Y, Soppe SE, Zhang X, Mullins MA, Barclay ME, Pettit N, Kurian AW, Dillon EC, Pruitt S, Lyratzopoulos G (2026) Mortality after emergency presentation: evidence from a cohort of 929 378 Medicare beneficiaries with common cancers diagnosed 2008–2017. JNCI: Journal of the National Cancer Institute. 10.1093/jnci/djag236 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Pettit N, Sarmiento E, Kline J (2022) Disparities in outcomes among patients diagnosed with cancer in proximity to an emergency department visit. Sci Rep 12: 10667. 10.1038/s41598-022-13422-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Thompson CA, Sheridan P, Metwally E, Peacock Hinton S, Mullins MA, Dillon EC, Thompson M, Pettit N, Kurian AW, Pruitt SL, Lyratzopoulos G (2024) Emergency department involvement in the diagnosis of cancer among older adults: a SEER-Medicare study. JNCI Cancer Spectrum 8. 10.1093/jncics/pkae039 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Atkin C, Iqbal G, Planche T, Pratt G, Yong K, Wood J, Raynes K, Low E, Higgins H, Neal RD, Dunn J, Drayson MT, Bowcock S, Group TTM, Investigators T (2021) Diagnostic pathways in multiple myeloma and their relationship to end organ damage: an analysis from the Tackling Early Morbidity and Mortality in Myeloma (TEAMM) trial. British Journal of Haematology 192: 997–1005. 10.1111/bjh.17044 [DOI] [PubMed] [Google Scholar]
- [20].Howell D, Smith A, Appleton S, Bagguley T, Macleod U, Cook G, Patmore R, Roman E (2017) Multiple myeloma: routes to diagnosis, clinical characteristics and survival – findings from a UK population-based study. British Journal of Haematology 177: 67–71. 10.1111/bjh.14513 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Rasmussen LA, Vedsted P, Jensen H, Frederiksen H, El-Galaly TC, Kristensen IB, Virgilsen LF (2024) Multiple myeloma: unplanned diagnostic pathways and association with risk factors and survival – a nationwide register-based cohort study in Denmark. BMC Cancer 24: 998. 10.1186/s12885-024-12706-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Enewold L, Parsons H, Zhao L, Bott D, Rivera DR, Barrett MJ, Virnig BA, Warren JL (2020) Updated Overview of the SEER-Medicare Data: Enhanced Content and Applications. JNCI Monographs 2020: 3–13. 10.1093/jncimonographs/lgz029 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Soppe SE, Peacock Hinton S, Dillon EC, Pruitt SL, Lyratzopoulos G, Barclay ME, Mullins MA, Kurian AW, Pettit N, Thompson M, Thompson CA (2026) Emergency cancer diagnosis in older adults: patterns, subgroups, and implications for health-care quality metrics. JNCI Cancer Spectrum 10. 10.1093/jncics/pkag008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Virgo K, Lin C (2014) Diagnosis Date Agreement Between SEER and Medicare Claims Data: Impact on Treatment. Medical Care 52: 32–7. 10.1097/MLR.0000000000000004 [DOI] [PubMed] [Google Scholar]
- [25].Centers for Medicare & Medicaid Services. Medicare-Medicaid Dual Eligibility Code – January | ResDAC [Internet]. https://resdac.org/cms-data/variables/medicare-medicaid-dual-eligibility-code-january [Accessed July 22, 2025].
- [26].Yost K, Perkins C, Cohen R, Morris C, Wright W (2001) Socioeconomic status and breast cancer incidence in California for different race/ethnic groups. Cancer Causes & Control 12: 703–11. 10.1023/A:1011240019516 [DOI] [PubMed] [Google Scholar]
- [27].Klabunde CN, Potosky AL, Legler JM, Warren JL (2000) Development of a comorbidity index using physician claims data. Journal of Clinical Epidemiology 53: 1258–67. 10.1016/S0895-4356(00)00256-0 [DOI] [PubMed] [Google Scholar]
- [28].Kim DH, Schneeweiss S, Glynn RJ, Lipsitz LA, Rockwood K, Avorn J (2017) Measuring Frailty in Medicare Data: Development and Validation of a Claims-Based Frailty Index. The Journals of Gerontology: Series A 73: 980–7. 10.1093/gerona/glx229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].National Cancer Institute. Rural-Urban Continuum Code - SEER Datasets [Internet]. https://seer.cancer.gov/seerstat/variables/countyattribs/ruralurban.html [Accessed July 22, 2025].
- [30].Centers for Medicare & Medicaid Services. Medicare Claims Processing Manual. Chapter 18 - Preventive and Screening Services. https://www.cms.gov/regulations-and-guidance/guidance/manuals/downloads/clm104c18.pdf [Accessed July 14, 2026].
- [31].Centers for Medicare & Medicaid Services. Documentation Guidelines for Evaluation and Management Services. 1997. Accessed September 8, 2025. https://www.cms.gov/outreach-and-education/medicare-learning-network-mln/mlnedwebguide/downloads/97docguidelines.pdf
- [32].Koo MM, Hamilton W, Walter FM, Rubin GP, Lyratzopoulos G (2018) Symptom Signatures and Diagnostic Timeliness in Cancer Patients: A Review of Current Evidence. Neoplasia 20: 165–74. 10.1016/j.neo.2017.11.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Laubach JP. Multiple myeloma: Clinical features, laboratory manifestations, and diagnosis. UpToDate. 2025. https://www.uptodate.com/contents/multiple-myeloma-clinical-features-laboratory-manifestations-and-diagnosis [Accessed October 20, 2025].
- [34].American Cancer Society. Signs and Symptoms of Multiple Myeloma. https://www.cancer.org/cancer/types/multiple-myeloma/detection-diagnosis-staging/signs-symptoms.html [Accessed June 25, 2025].
- [35].Spiegelman D, Logan R, Grove D (2011) Regression Calibration with Heteroscedastic Error Variance. Int J Biostat 7. doi: 10.2202/1557-4679.1259 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [36].National Cancer Intelligence Network. Routes to Diagnosis: Exploring Emergency Presentations (NCIN Data Briefing). 2012.
- [37].Cranfield BM, Abel GA, Swann R, Moore SF, McPhail S, Rubin GP, Lyratzopoulos G (2023) Pre-Referral Primary Care Blood Tests and Symptom Presentation before Cancer Diagnosis: National Cancer Diagnosis Audit Data. Cancers (Basel) 15: 3587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [38].Mitchell ED, Pickwell-Smith B, Macleod U (2015) Risk factors for emergency presentation with lung and colorectal cancers: a systematic review. BMJ Open 5: e006965. 10.1136/bmjopen-2014-006965 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [39].Wildes TM, Fiala MA (2018) Falls in older adults with multiple myeloma. European Journal of Haematology 100: 273–8. 10.1111/ejh.13009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [40].Fillmore NR, DuMontier C, Yildirim C, La J, Epstein MM, Cheng D, Cirstea D, Yellapragada S, Abel GA, Gaziano JM, Do N, Brophy M, Kim DH, Munshi NC, Driver JA (2021) Defining Multimorbidity and Its Impact in Older United States Veterans Newly Treated for Multiple Myeloma. J Natl Cancer Inst 113: 1084–93. 10.1093/jnci/djab007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [41].Jansen J, van Weert JCM, Wijngaards-de Meij L, van Dulmen S, Heeren TJ, Bensing JM (2010) The role of companions in aiding older cancer patients to recall medical information. Psycho-Oncology 19: 170–9. 10.1002/pon.1537 [DOI] [PubMed] [Google Scholar]
- [42].Stewart SJ, Roberts L, Brindle L (2021) Romantic partner involvement during oncology consultations: A narrative review of qualitative and quantitative studies. Patient Education and Counseling 104: 64–74. 10.1016/j.pec.2020.08.018 [DOI] [PubMed] [Google Scholar]
- [43].Powell DS, Reed NS, Oh E, Samus QM, Deal JA, Wolff JL (2023) The role of care partners in medical visits of older adults with hearing loss and dementia: A national study. Journal of the American Geriatrics Society 71: 909–15. 10.1111/jgs.18151 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [44].Pruitt SL, Davidson NO, Gupta S, Yan Y, Schootman M (2014) Missed opportunities: racial and neighborhood socioeconomic disparities in emergency colorectal cancer diagnosis and surgery. BMC Cancer 14: 927. 10.1186/1471-2407-14-927 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [45].Sikka V, Ornato JP (2012) Cancer diagnosis and outcomes in Michigan EDs vs other settings. The American Journal of Emergency Medicine 30: 283–92. 10.1016/j.ajem.2010.11.029 [DOI] [PubMed] [Google Scholar]
- [46].Dickman SL, Gaffney A, McGregor A, Himmelstein DU, McCormick D, Bor DH, Woolhandler S (2022) Trends in Health Care Use Among Black and White Persons in the US, 1963–2019. JAMA Network Open 5: e2217383–e. 10.1001/jamanetworkopen.2022.17383 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [47].Perry LM, Walsh LE, Horswell R, Miele L, Chu S, Melancon B, Lefante J, Blais CM, Rogers JL, Hoerger M (2021) Racial Disparities in End-of-Life Care Between Black and White Adults With Metastatic Cancer. Journal of Pain and Symptom Management 61: 342–9.e1. 10.1016/j.jpainsymman.2020.09.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [48].Smith AK, Earle CC, McCarthy EP (2009) Racial and Ethnic Differences in End-of-Life Care in Fee-for-Service Medicare Beneficiaries with Advanced Cancer. Journal of the American Geriatrics Society 57: 153–8. 10.1111/j.1532-5415.2008.02081.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- [49].Abel GA, Mendonca SC, McPhail S, Zhou Y, Elliss-Brookes L, Lyratzopoulos G (2017) Emergency diagnosis of cancer and previous general practice consultations: insights from linked patient survey data. Br J Gen Pract 67: e377–e87. 10.3399/bjgp17X690869 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [50].Stube CW, Ljungberg AS, Borton JA, Chadha K, Kelleran KJ, Lerner EB (2024) Why Do Patients Opt for the Emergency Department over Other Care Choices? A Multi-Hospital Analysis. West J Emerg Med 25: 921–8. 10.5811/westjem.18647 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [51].Wu N, Woloski JR (2024) Emergency Department Versus Primary Care Use: A Patient Perspective. PRiMER 8. 10.22454/PRiMER.2024.526921 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [52].Ludwig H, Kainz S, Schreder M, Zojer N, Hinke A (2023) SLiM CRAB criteria revisited: temporal trends in prognosis of patients with smoldering multiple myeloma who meet the definition of ‘biomarker-defined early multiple myeloma’—a systematic review with meta-analysis. eClinicalMedicine 58: 101910. 10.1016/j.eclinm.2023.101910 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [53].Ugarte S, D’Avella C, Gabriel PE, Ciunci CA, Roy M, Schwartz A, Boyk K, Walker S, Tindall J, Dougherty DW (2024) The cancer diagnostic clinic: A rapid access clinic to improve outcomes for patients with a suspected cancer diagnosis. JCO Oncology Practice 20: 77–. 10.1200/OP.2024.20.10_suppl.77 [DOI] [Google Scholar]
- [54].Vaidya P, Cutting ER, Kato S, Mitchell WM, Heyman BM, Gold KA (2023) The UC san diego health suspicion of cancer clinic. Journal of Clinical Oncology 41: e18528–e. 10.1200/JCO.2023.41.16_suppl.e18528 [DOI] [Google Scholar]
- [55].Triglianos T, Sorah JD, Cao F, Bunch J, Collichio FA, Allard M, Medley C, Basch E, Sanoff HK (2023) Rapid care transition: Implementation of an undiagnosed cancer clinic. JCO Oncology Practice 19: 80–. 10.1200/OP.2023.19.11_suppl.80 [DOI] [Google Scholar]
- [56].Charalampous C, Goel U, Kapoor P, Binder M, Buadi F, Dingli D, Dispenzieri A, Fonder A, Gertz M, Gonsalves W, Hayman S, Hobbs M, Hwa YL, Kourelis T, Lacy M, Leung N, Lin Y, Warsame R, Kyle RA, Rajkumar V, Kumar SK (2024) Association of Thrombocytopenia With Disease Burden, High-Risk Cytogenetics, and Survival in Newly Diagnosed Multiple Myeloma Patients Treated With Novel Therapies. Clinical Lymphoma Myeloma and Leukemia 24: e329–e35. 10.1016/j.clml.2024.05.020 [DOI] [PubMed] [Google Scholar]
- [57].Rögnvaldsson S, Thorsteinsdóttir S, Kristinsson SY (2024) Screening in Multiple Myeloma and Its Precursors: Are We There Yet? Clinical Chemistry 70: 128–39. 10.1093/clinchem/hvad148 [DOI] [PubMed] [Google Scholar]
- [58].Landgren O, Kyle RA, Pfeiffer RM, Katzmann JA, Caporaso NE, Hayes RB, Dispenzieri A, Kumar S, Clark RJ, Baris D, Hoover R, Rajkumar SV (2009) Monoclonal gammopathy of undetermined significance (MGUS) consistently precedes multiple myeloma: a prospective study. Blood 113: 5412–7. 10.1182/blood-2008-12-194241 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [59].Weiss BM, Abadie J, Verma P, Howard RS, Kuehl WM (2009) A monoclonal gammopathy precedes multiple myeloma in most patients. Blood 113: 5418–22. 10.1182/blood-2008-12-195008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [60].Kyle RA, Greipp PR (1980) Smoldering Multiple Myeloma. New England Journal of Medicine 302: 1347–9. doi: 10.1056/NEJM198006123022405 [DOI] [PubMed] [Google Scholar]
- [61].Kyle RA, Larson DR, Therneau TM, Dispenzieri A, Kumar S, Cerhan JR, Rajkumar SV (2018) Long-Term Follow-up of Monoclonal Gammopathy of Undetermined Significance. New England Journal of Medicine 378: 241–9. doi: 10.1056/NEJMoa1709974 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [62].Turesson I, Kovalchik SA, Pfeiffer RM, Kristinsson SY, Goldin LR, Drayson MT, Landgren O (2014) Monoclonal gammopathy of undetermined significance and risk of lymphoid and myeloid malignancies: 728 cases followed up to 30 years in Sweden. Blood 123: 338–45. 10.1182/blood-2013-05-505487 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [63].Rögnvaldsson S, Love TJ, Thorsteinsdottir S, Reed ER, Óskarsson JÞ, Pétursdóttir Í, Sigurðardóttir GÁ, Viðarsson B, Önundarson PT, Agnarsson BA, Sigurðardóttir M, Þorsteinsdóttir I, Ólafsson Í, Þórðardóttir ÁR, Eyþórsson E, Jónsson Á, Björnsson AS, Gunnarsson GÞ, Pálsson R, Indriðason ÓS, Gíslason GK, Ólafsson A, Hákonardóttir GK, Brinkhuis M, Halldórsdóttir SL, Ásgeirsdóttir TL, Steingrímsdóttir H, Danielsen R, Dröfn Wessman I, Kampanis P, Hultcrantz M, Durie BGM, Harding S, Landgren O, Kristinsson SY (2021) Iceland screens, treats, or prevents multiple myeloma (iStopMM): a population-based screening study for monoclonal gammopathy of undetermined significance and randomized controlled trial of follow-up strategies. Blood Cancer J 11: 94. 10.1038/s41408-021-00480-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- [64].Soppe SE, Peacock Hinton S, Halula JC, Lund JL, Baggett CD, Pruitt SL, Mullins MA, Dillon EC, Barclay ME, Thompson M, Pettit N, Lyratzopoulos G, Thompson CA (2025) Registry versus claims-based index dates for studies of cancer diagnosis in administrative data. Cancer Causes & Control 36: 539–50. 10.1007/s10552-024-01953-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [65].Codden RR, Sweeney C, Ofori-Atta BS, Herget KA, Wigren K, Edwards S, Carter ME, McCarty RD, Hashibe M, Doherty JA, Millar MM (2024) Accuracy of patient race and ethnicity data in a central cancer registry. Cancer Causes Control 35: 685–94. 10.1007/s10552-023-01827-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [66].National Cancer Institute. SEER*Explorer. https://seer.cancer.gov/explorer/index.html [Accessed August 8, 2026].
- [67].Abel GA, Shelton J, Johnson S, Elliss-Brookes L, Lyratzopoulos G (2015) Cancer-specific variation in emergency presentation by sex, age and deprivation across 27 common and rarer cancers. Br J Cancer 112: S129–S36. 10.1038/bjc.2015.52 [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 analyzed during the current study are not publicly available due to protections for individual privacy. Investigators may obtain a Data Use Agreement directly from the National Cancer Institute to access the SEER-Medicare data upon approval of a specific research project. Further information is available at https://healthcaredelivery.cancer.gov/seermedicare/obtain/.
