Abstract
This cross-sectional study characterizes seasonal variations in the proportion of cancer diagnoses in the US across the calendar year and discusses the implications of these patterns.
Introduction
Timely diagnosis of and treatment initiation for cancer are critical for optimizing survival outcomes. Delays in diagnosis, even when measured in weeks, have been shown to result in a stage migration.1 Studies suggest that seasonal calendar–related events, including holidays and vacation seasons, are associated with reduced health care delivery and accessibility.2 Variations in monthly diagnostic patterns, whether due to patient behavior, scheduling availability, health care access, or other seasonal factors, may represent overlooked factors for patients seeking cancer care. We aimed to characterize monthly variation in the proportion of cancer diagnoses, stage at diagnosis, and net survival across the calendar year in the US as a hypothesis-generating examination of potential health-system influences on diagnostic timing. Our study may lead to quality improvement initiatives with the goal of identifying factors of monthly variations in cancer care.
Methods
This cross-sectional study did not require institutional review board approval in accordance with the National Program of Cancer Registries (NPCR) and Surveillance, Epidemiology, and End Results (SEER) policies due to the use of the deidentified versions of NPCR and SEER datasets. All analyses were conducted in compliance with the SEER data use agreement and reporting followed the STROBE reporting guideline. We utilized the combined NPCR-SEER dataset (2001-2021), which represents the entire US population for incidence, to calculate monthly proportions of overall diagnoses, stage at diagnosis and the 10 most common cancer sites. For survival analyses, we utilized the SEER Research Plus 22 Registries Database (2001-2021) that covers approximately 42% of the US population. Analyses were limited to 2001 to 2019 to avoid COVID-19 pandemic–related confounding. The stage was classified using the SEER Merged Summary Stage (localized, regional, and distant). We used the Pohar-Perme method to calculate net relative survival to minimize bias associated with classical overall or cause-specific survival methods.3 The selection was further age-standardized per the International Cancer Survival Standard-1, thereby meeting the CONCORD and SEER standards of reporting. Analyses were conducted using SEER*Stat version 8.4.5 (Information Management Services) on April 20, 2025. Significance was defined as a 2-sided P < .05. Further detail can be found in the eMethods in Supplement 1.
Results
We identified 30 184 124 new cancer cases between 2001 and 2019. The highest number of diagnoses occurred in January (2 660 525 diagnoses [8.81%]), followed by June (2 595 361 diagnoses [8.60%]) and October (2 589 319 diagnoses [8.58%]), while the lowest number occurred in December (2 342 132 [7.76%]) (Table). These patterns persisted across individual cancer sites, clustering around the same monthly peaks and troughs. Melanoma and prostate cancer showed a consistent rise in diagnoses from late spring through summer. The proportion of distant-stage diagnoses was highest in January (651 795 diagnoses [24.49%]) and lowest in October (605 412 diagnoses [23.38%]). The 5-year survival ranged from worst in December (64.70%; 95% CI, 64.60%-64.90%) to best in October 66.10% (95% CI, 66.00%-66.30%) (Figure). The monthly proportion of localized-stage diagnoses correlated positively with 5-year net survival, including breast (r = 0.95; P < .001), prostate (r = 0.89; P < .001), colorectal (r = 0.89; P < .001), and lung (r = 0.81; P = .001) cancers. These patterns remained stable across the study period since 2001.
Table. Monthly Variation in Cancer Diagnosis, Stage at Diagnosis, and Net Survival in the US, 2001-2019a.
| Monthc | No. (%) | Net survival, % (95% CI)b | ||||
|---|---|---|---|---|---|---|
| Diagnoses (N = 30 184 124)d | Localized stagee | Distant stagee | 1-y | 3-y | 5-y | |
| January | 2 660 525 (8.81) | 1 132 857 (42.58) | 651 795 (24.49) | 79.30 (79.20-79.40) | 69.10 (69.00-69.30) | 64.90 (64.80-65.00) |
| February | 2 349 728 (7.78) | 1 037 104 (44.13) | 557 894 (23.74) | 79.10 (79.00-79.20) | 69.20 (69.10-69.40) | 65.20 (65.10-65.40) |
| March | 2 570 705 (8.51) | 1 133 078 (44.07) | 613 511 (23.86) | 79.00 (78.90-79.10) | 69.20 (69.10-69.30) | 65.10 (65.00-65.30) |
| April | 2 520 457 (8.35) | 1 118 838 (44.39) | 597 296 (23.69) | 79.30 (79.20-79.40) | 69.40 (69.30-69.50) | 65.30 (65.20-65.40) |
| May | 2 549 240 (8.44) | 1 132 908 (44.44) | 603 945 (23.69) | 79.30 (79.20-79.40) | 69.50 (69.40-69.60) | 65.40 (65.30-65.60) |
| June | 2 595 361 (8.60) | 1 124 012 (43.30) | 625 650 (24.10) | 79.50 (79.40-79.60) | 69.50 (69.30-69.60) | 65.30 (65.20-65.50) |
| July | 2 466 183 (8.17) | 1 084 514 (43.97) | 593 061 (24.04) | 78.90 (78.80-79.00) | 69.10 (69.00-69.20) | 65.10 (64.90-65.20) |
| August | 2 553 953 (8.46) | 1 133 106 (44.36) | 607 401 (23.78) | 79.20 (79.20-79.30) | 69.50 (69.40-69.60) | 65.40 (65.30-65.60) |
| September | 2 409 557 (7.98) | 1 065 518 (44.22) | 573 931 (23.81) | 79.10 (79.00-79.20) | 69.30 (69.20-69.40) | 65.40 (65.20-65.50) |
| October | 2 589 319 (8.58) | 1 160 835 (44.83) | 605 412 (23.38) | 79.70 (79.60-79.80) | 70.10 (70.00-70.30) | 66.10 (66.00-66.30) |
| November | 2 370 840 (7.85) | 1 057 923 (44.62) | 556 882 (23.48) | 79.30 (79.20-79.40) | 69.80 (69.70-70.00) | 65.90 (65.80-66.10) |
| December | 2 342 132 (7.76) | 1 022 179 (43.64) | 566 844 (24.20) | 78.30 (78.20-78.40) | 68.70 (68.60-68.80) | 64.70 (64.60-64.90) |
Incidence and stage data are from the National Program of Cancer Registries and Surveillance, Epidemiology, and End Results (SEER) combined dataset, covering approximately 97% of the US population (total diagnoses, 30 184 124). Net survival is from the SEER Research Plus 22-Registries Database, covering 41.9% of the US population (9 395 033 individuals). Analyses were limited to 2001 to 2019. A total of 206 097 records with a missing month of diagnosis (of which 40 538 were localized and 61 413 were distant stage) and 27 records with invalid values were excluded from the monthly distribution.
Net survival was estimated with the Pohar-Perme estimator and age-standardized to the International Cancer Survival Standard-1 standard.
Peak and lowest months, respectively, of diagnosis for the 10 most common cancer sites are as follows: (October: 383 039 diagnoses [8.88%]; February: 333 204 diagnoses [7.72%]), lung and bronchus (January: 359 656 diagnoses [8.80%]; November: 318 507 diagnoses [7.80%]), prostate (January: 378 113 diagnoses [9.40%]; December: (307 158 diagnoses [7.63%]), colorectal (January: 240 580 diagnoses [8.68%]; December: 213 280 diagnoses [7.70%]), urinary bladder (June: 116 732 diagnoses [8.74%]; December: 102 578 diagnoses [7.68%]), melanoma of the skin (June: 123 397 diagnoses [9.48%]; December: 90 524 diagnoses [6.95%]); non-Hodgkin lymphoma (January: 109 215 diagnoses [8.83%]; February: 96 150 diagnoses [7.78%]), kidney and renal pelvis (October: 90 815 diagnoses [8.74%]; February: 80 867 diagnoses [7.79%]), uterine (October: 80 857 diagnoses [8.95%]; December: 69 801 diagnoses [7.72%]), and pancreatic (January: 70 444 diagnoses [8.77%]; February: 62 292 diagnoses [7.75%]). Total diagnoses by site were 4 312 855 for breast, 4 083 046 for lung and bronchus, 4 022 712 for prostate, 2 771 554 for colorectal, 1 334 559 for urinary bladder, 1 301 396 for melanoma, 1 235 840 for non-Hodgkin lymphoma, 1 038 447 for kidney and renal pelvis, 903 506 for uterine, and 803 445 for pancreatic. Together these 10 sites accounted for 21 807 360 diagnoses (72.24% of all cancers).
The percentage is the proportion of all diagnoses occurring in the given month.
The percentage is the proportion of that month’s diagnoses classified at the specified stage. Stage was defined by the SEER Merged Summary Stage.
Figure. Five-Year Age-Adjusted Net Relative Survival of Top 10 Cancer Sites by Month of Diagnosis From 2001 to 2019.
Discussion
This cross-sectional study provides foundational evidence of consistent monthly variation in cancer diagnosis across the US. The difference between diagnosed cases in January and December was 1%, a small proportion but a substantial absolute difference at the population scale, corresponding to approximately 300 000 cases. This finding may reflect either lower diagnostic activity in December or increased clinical activity in January, consistent with reported postholiday spikes.2,4,5 There are multiple plausible explanations for these findings, many of which can be overlapping or simultaneously contributing. Health care access and patient behavior may shift around holidays, translating into variable clinical outcomes across calendar months.6,7 Fluctuations in workforce and specialist availability may lengthen time to specialist assessment and diagnosis, while referral and scheduling patterns further shape diagnostic timing. Insurance-related factors may also contribute, such as calendar resets, activation, and deductibles.6 The October peak in several cancers, particularly breast, may partly reflect awareness campaigns, and the spring-summer rise in melanoma may relate to greater sun exposure and seasonal skin surveillance. Our dataset is on a population scale and thus can be highly heterogeneous between health care systems and regions. These findings should be interpreted as associations rather than evidence of any single causal mechanism. Quality improvement initiatives should identify the factors driving these monthly variations. These observations may extend beyond cancer care and warrant evaluation across other specialties. If replicated across health care systems, calendar-month variation could serve as a systems-level metric for assessing the temporal consistency of health care delivery.
The study has several limitations. It is observational and descriptive, and causal inference cannot be confirmed; no statistical adjustment was performed for normal variation over time in diagnoses, stage, or survival. We did not assess race and ethnicity, socioeconomic status, or geography, which may interact with both calendar month and outcomes. SEER summary staging lacks the granularity of American Joint Committee on Cancer classification and includes hematologic malignant tumors not staged by anatomic extent, limiting inferences about stage migration hypothesis. Delays in data reporting or misclassification of diagnosis month are possible but unlikely to fully explain the magnitude of these patterns. Future research should identify the factors driving these patterns to inform workforce planning, resource allocation during lower-activity months, payor considerations, and policy aimed at optimizing outcomes.
eMethods.
Data Sharing Statement
References
- 1.Crosby D, Bhatia S, Brindle KM, et al. Early detection of cancer. Science. 2022;375(6586):eaay9040. doi: 10.1126/science.aay9040 [DOI] [PubMed] [Google Scholar]
- 2.Nilbert M. Striving to excel—effects from vacation periods on cancer incidence and outcome. Acta Oncol. 2023;62(2):101-102. doi: 10.1080/0284186X.2023.2180662 [DOI] [PubMed] [Google Scholar]
- 3.Roche L, Danieli C, Belot A, et al. Cancer net survival on registry data: use of the new unbiased Pohar-Perme estimator and magnitude of the bias with the classical methods. Int J Cancer. 2013;132(10):2359-2369. doi: 10.1002/ijc.27830 [DOI] [PubMed] [Google Scholar]
- 4.Lambe M, Blomqvist P, Bellocco R. Seasonal variation in the diagnosis of cancer: a study based on national cancer registration in Sweden. Br J Cancer. 2003;88(9):1358-1360. doi: 10.1038/sj.bjc.6600901 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wikén I, Andersson TM, Radkiewicz C. Seasonal effects on cancer incidence and prognosis. Acta Oncol. 2023;62(2):103-109. doi: 10.1080/0284186X.2023.2178325 [DOI] [PubMed] [Google Scholar]
- 6.Roychoudhuri R, Robinson D, Coupland V, Holmberg L, Møller H. Season of cancer diagnosis exerts distinct effects upon short- and long-term survival. Int J Cancer. 2009;124(10):2436-2441. doi: 10.1002/ijc.24213 [DOI] [PubMed] [Google Scholar]
- 7.Gao F, Machin D, Khoo KS, Ng EH. Seasonal variation in breast cancer diagnosis in Singapore. Br J Cancer. 2001;84(9):1185-1187. doi: 10.1054/bjoc.2000.1654 [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
eMethods.
Data Sharing Statement
