Differential Ultrasound Rates Mirror Sex Disparities in Thyroid Cancer

Abstract

Background:

Expanding ultrasound use has increased the detection of thyroid cancer. Incidence has always been higher among females, a disparity that has grown over time. The sex difference in thyroid cancer is understudied in the context of diagnostic testing, particularly among privately insured adults in whom thyroid cancer is most common. We evaluated the association between thyroid ultrasound, fine needle aspiration biopsy (FNAB), and cancer incidence by sex in a large, integrated health system.

Methods:

This longitudinal retrospective cohort study included Kaiser Permanente of Washington enrollees aged 18 and over who underwent thyroid ultrasound from 1997 to 2019. Data included electronic billing claims for patients linked to tumor registry diagnoses. We estimated (1) annual overall ultrasound, FNAB, and cancer incidence rates; (2) the proportion of ultrasound requiring FNAB; and (3) cancer diagnoses per FNAB. A Poisson model with offset determined the relationship between sex and the proportion of ultrasound requiring FNAB adjusting for patient and sociodemographic characteristics.

Results:

A total of 33,589 patients underwent ultrasound (78% females; mean age 56). Ultrasound rates per 100,000 covered lives, defined as insured individuals per year, increased five-fold among males (111.11–490.97) and >four-fold among females (382.27–1331.14) between 1997 and 2019. FNAB rates also increased over time (rates per 100,000: 174.09–430.37 in females vs. 58.38–189.13 in men). Overall, FNAB rates per ultrasound changed little over time, and FNAB per ultrasound was greater in males compared with females (Adj rate ratio = 1.06 [confidence interval 1.01–1.11]). Cancer incidence was higher in females over the study period, but cancer incidence per FNAB was similar between sexes (both 0.06, p = 0.4).

Conclusions:

Sex disparities in thyroid ultrasound rates are stark and are a likely driver of sex disparities in thyroid cancer incidence. Interestingly, ultrasound-triggered FNAB was more common in males and changed little over time, challenging the prevailing understanding that females have much higher rates of thyroid cancer. Although the population-based differences between sexes for FNAB and cancer were large, the differences among people who had ultrasound were small.

Introduction

Thyroid cancer incidence increased over 200% (5.7–13.1 per 100,000) from 1992 to 2018 with no corresponding change in mortality rate.^1,2 Papillary thyroid cancers (PTCs) represent 85% of all thyroid cancers, and small PTCs (<2 cm) have contributed most to the increased incidence (4.4% annual increase from 1974 to 2013).³ Small PTCs are largely indolent and have a low likelihood of becoming symptomatic or reducing lifespan.^4
–6 Females have disproportionately contributed to the rising thyroid cancer incidence and are diagnosed with small PTCs at 4.4 times the rate observed in men.^7
–9 One explanation for the significant rise is the over-detection of small thyroid nodules and cancer from increased use and availability of imaging, particularly ultrasound.^10
–12

Patterns of thyroid imaging and cancer incidence in federally insured populations in the United States and Canada support this explanation. Haymart et al.¹³ evaluated thyroid cancer imaging and incidence in a longitudinal study of Medicare beneficiaries and found that the 20% annual increase in the use of thyroid ultrasound resulted in more than 6000 excess diagnoses from 2003 to 2013. The analysis also showed that ultrasound use was disproportionately higher among females and patients with comorbidities. Similarly, Zevallos et al.¹⁴ demonstrated that the doubling of thyroid cancer incidence from 2000 to 2012 was associated with a five-fold increase in thyroid ultrasound use and a seven-fold increase in fine needle aspiration biopsy (FNAB) in an overwhelmingly male Veterans Affairs (VA) population. Although these results suggest a role for imaging use in thyroid cancer incidence among Medicare and primarily male VA patients, the majority of thyroid cancers are diagnosed in females aged 46–60 years.¹⁵ The sex disparity in thyroid cancer incidence has been understudied in the context of diagnostic testing patterns, particularly in privately insured adult populations, where thyroid cancer is most common.^16
–18

We therefore analyze the association between thyroid imaging and cancer incidence by sex in a large, privately insured population over 22 years. We evaluated patterns of thyroid ultrasound, FNAB, and cancer incidence over time among members of Kaiser Permanente of Washington (KPWA) from 1997 to 2019. Assuming that a small sex difference in underlying thyroid cancer risk has stayed relatively stable over the study period,^7,18 we had two hypotheses: (1) the rate of FNAB per ultrasound performed would remain relatively stable over time and would not differ between males and females and (2) differential use of diagnostic imaging over time is associated with the large sex difference in thyroid cancer incidence.

Materials and Methods

Study population

We conducted a retrospective cohort study among enrollees of KPWA, an integrated health care delivery system providing medical coverage and/or care to more than 600,000 people in Washington State. This study obtained institutional review board’s (IRB) approval with a waiver of consent to analyze records from KPWA and the University of Wisconsin (UW) (UW Health Science IRB # 2020-0439).

We defined a cohort of adults aged 18 years or older who underwent thyroid ultrasound with or without FNAB (Current Procedural Terminology [CPT] codes 76536, 88170, 88171, 10021, 10022, 60001; International Classification of Diseases, 10th Revision, Procedure Coding System BG44ZZZ) during 1997–2019 and had continuous enrollment in the health plan for 12 months prior to the ultrasound. We limited the cohort to people who resided in one of the 13 counties of Western Washington, which are covered by the Seattle-Puget Sound Surveillance Epidemiology and End Results (SEER) cancer registry in order to link to cancer diagnoses. SEER is a population-based tumor registry that collects data on cancer diagnoses, characteristics of treatment, and outcomes for people who reside in a specific geographic area. KPWA sends a monthly case file to the local SEER registry with diagnosis and procedure codes that might indicate a cancer diagnosis. Codes from billing and claims data identify individuals who are KPWA members but were diagnosed outside of the health system. The file includes standard identifiers (name, date of birth, race/ethnicity, sex, Social Security number, medical record number, and billing date). The SEER site uses this file to determine which of its cases are KPWA members based on the identifiers.

We sought to understand patterns of care for new thyroid cancers only, excluding people with a prior thyroid cancer diagnosis. Our analysis was thus designed to capture incident rather than prevalent users of thyroid ultrasound. As such, we excluded 1494 people with a history of thyroidectomy or lobectomy as indications for ultrasound in this population are systematically different (e.g., surveillance for recurrence; Supplementary Fig. S1). Patients concurrently undergoing radiation, chemotherapy, or endocrine treatment for cancer of another site were also excluded (N = 3640) to avoid conflating thyroid cancer treatment with treatment for other conditions. We also excluded people who were pregnant at the time of the ultrasound (N = 292) owed to their potential for nonstandard treatment courses. If a patient had ultrasounds in two nonconsecutive years along with other qualifying characteristics, they could be counted more than once during the study period as a separate episode of care. To calculate rates of imaging, FNAB (CPT-4 codes: 60001, 60300, 60100, 10021, 10022, 76942, 88170, 88171 or CPT 10005 with an associated diagnosis code C73, D34, E04.1, E04.2, E04.9, E05.20, E06.9, E07.9), and cancer incidence (International Classification of Diseases for Oncology, 3rd Edition [ICD-O-3], C73), we used a covered lives dataset as the denominator containing counts of all adults with 18 months of continuous enrollment in KPWA by age and sex for each calendar year from 1997 to 2019.

Measures

We extracted data on demographic characteristics and health care utilization from the Virtual Data Warehouse (VDW).¹⁹ The VDW consists of standardized data tables with information from clinical and administrative data sources, including electronic health records, tumor registries, and health care claims. We captured counts of all inpatient, outpatient, and emergency department visits in the 12 months before the thyroid ultrasound, as well as the occurrence of any additional thyroid imaging or FNAB, to ensure eligibility. Sociodemographic information included age, sex from the health record, race/ethnicity, and insurance type (Kaiser in-network coverage vs. optional out-of-network coverage) measured at the patient level at the time of the thyroid ultrasound. Area-level socioeconomic deprivation was measured at the census tract level using the publicly available 2018 Area Deprivation Index (ADI), which was developed from the Health Resources & Services Administration American Community Survey and allows for rankings of neighborhoods by socioeconomic disadvantage in a region of interest.^20,21 Charlson Comorbidity Index was calculated at an individual level using previously developed methods to classify diagnosis codes in the 12 months before the thyroid ultrasound.^22,23

Subsequent diagnosis of thyroid cancer within 12 months of the ultrasound, tumor size, staging, and histological type (papillary: 8050, 8260, 8340, 8341, 8342, 8343, 8344, 8350, 8450; follicular: 8290, 8330, 8331, 8332, 8335, 8337, 8339; medullary: 8345, 8510, 8347; or anaplastic: 8021) was extracted through linkage with SEER. The absence of an ICD-O-3 code after ultrasound and/or FNAB for thyroid cancer was considered a noncancer biopsy result.

Analysis

The analytical dataset contained a row of data for each year, sex, and age-group (176 rows total for 22 years, 2 sexes, and 4 age-groups). Descriptive characteristics are reported as means and proportions calculated for each year and then averaged; the minimum value represents the value for the specific year where the minimum was observed, and the maximum value similarly represents the maximum value across all years. Diagnostic testing and cancer incidence are presented as unadjusted rates per 100,000 covered lives. A time trend analysis estimated annual rates of ultrasound and FNAB utilization and cancer incidence for the total KPWA population. To do this analysis, we first estimated the rate of growth for annual ultrasound use as a function of calendar year using the covered lives data set as the denominator. Next, for the cohort of patients with a thyroid ultrasound, a similar time trend analysis estimated the FNAB rates with covered lives as the denominator and then for patients with ultrasound only. Finally, we calculated thyroid cancer incidence among all covered lives and among patients who underwent FNAB over time. Joinpoint regression analysis with autoregressive correlation determined inflection points for ultrasound and FNAB rate changes during the study period.

We used a Poisson model with an offset term to evaluate the relationship between sex and FNAB rate, controlling for age-group, insurance type, comorbidity, and regional ADI tertile. The offset term represented the number of ultrasounds for the calendar year by sex and age-group. Pooled rate ratios are presented by sex and cancer histology. Analyses were performed using Python 3 for figures,^24
–26 R v.4 for modeling,²⁷ STATA version 17.0 SE,²⁸ for descriptive statistics, and Joinpoint 4.9.1.0.^29,30

Results

In total, 33,589 people underwent 39,412 thyroid ultrasounds from 1997 to 2019. The number of covered individuals in KPWA grew over time from 235,692 (45% male, 55% female) in 2017 to 311,521 in 2019 (46% male, 54% female). Of ultrasounds, 7045 were simultaneous ultrasound-guided FNAB with initial ultrasound, and 15,947 (40%) had an FNAB either concurrently or following ultrasound. Overall, 865 (2.6%) of patients undergoing ultrasound were diagnosed with thyroid cancer.

Sociodemographic characteristics of patients who received ultrasound are summarized in Table 1. A majority of people who received an ultrasound were female (78%). On average, 70% of patients who underwent ultrasound were White. Eighty-two percent of patients primarily received health insurance and care within the KPWA network, and 25% had insurance coverage for care outside the network. Sociodemographic characteristics were similar between males and females.

Table 1.

Cohort Descriptive Characteristics, Mean Proportion per Year (n = 33,589)

	Overall				Female				Male
Variable	Proportion	SD	Min	Max	Proportion	SD	Min	Max	Proportion	SD	Min	Max
Age (mean)	55.51	2.23	52.88	57.66	55.68	0.91	52.67	57.35	58.64	1.11	56.38	60.68
Age 18–34	0.1	0.01	0.07	0.13	0.1	0.01	0.08	0.13	0.07	0.03	0.01	0.13
Age 35–49	0.25	0.05	0.19	0.35	0.27	0.05	0.2	0.39	0.19	0.04	0.12	0.28
Age 50–64	0.37	0.04	0.29	0.42	0.37	0.04	0.29	0.42	0.37	0.05	0.27	0.5
Age 65+	0.29	0.03	0.24	0.34	0.26	0.03	0.21	0.32	0.37	0.04	0.31	0.45
Female	0.78	0.02	0.75	0.81	—	—	—	—	—	—	—	—
Male	0.22	0.02	0.19	0.25	—	—	—	—	—	—	—	—
KPWA non-network	0.25	0.05	0.19	0.38	0.24	0.05	0.19	0.37	0.26	0.07	0.15	0.41
White race/ethnicity	0.7	0.04	0.63	0.81	0.71	0.05	0.63	0.83	0.7	0.04	0.63	0.77
Non-White race/ethnicity	0.3	0.04	0.19	0.37	0.29	0.05	0.17	0.37	0.3	0.04	0.23	0.37
Top tertile Area Deprivation Index	0.67	0.02	0.62	0.7	0.67	0.02	0.6	0.7	0.67	0.05	0.54	0.75
Middle tertile Area Deprivation Index	0.28	0.02	0.23	0.31	0.28	0.03	0.22	0.32	0.27	0.04	0.18	0.39
Lowest tertile Area Deprivation Index	0.05	0.03	0.03	0.11	0.05	0.03	0.03	0.11	0.05	0.03	0.02	0.12
Charlson: 1 or more comorbidity	0.44	0.08	0.29	0.54	0.41	0.09	0.24	0.54	0.48	0.05	0.38	0.58

The minimum and maximum columns represent the value for the year when that variable was least or greatest, respectively. For the non-White race/ethnicity, 0.14 was unknown, 0.08 was Asian, 0.04 was Black, 0.02 was multiple races/ethnicities, 0.01 was Native American, and 0.01 was other race/ethnicity.

KPWA, Kaiser Permanente of Washington; SD, standard deviation.

Table 2 summarizes annual ultrasound, FNAB, and thyroid cancer incidence rates. The mean annual ultrasound rate was 587 per 100,000 covered lives (839 female; 286 male). The mean annual FNAB rate was 225 per 100,000 (311 female; 123 male). Overall pooled cancer incidence was 12.4 per 100,000 (18 female; 8 male), with papillary being the most common type at 11.5 per 100,000 (16 female; 7 male), followed by follicular, medullary, and anaplastic. Only four tumors were anaplastic over the entire study period, making inference for this group infeasible.

Table 2.

Diagnostic Test and Tumor Rates per Year per 100K Covered Lives

	Overall				Female				Male
Variable	Mean	SD	Min	Max	Mean	SD	Min	Max	Mean	SD	Min	Max
Ultrasound rate	587.24	224.73	260.09	936.37	839.41	309.92	382.27	1331.14	286.41	124.19	111.11	490.97
FNAB rate	225.29	64.44	131.39	316.53	311.11	84.77	174.09	430.37	122.87	41.76	58.38	189.13
Cancer incidence	13.47	6.41	2.97	23.08	18.25	9.07	2.32	32.08	7.77	3.87	1.71	14.3
Anaplastic incidence	0.163	0.271	0	0.82	0.273	0.46	0	1.5	0.034	0.164	0	0.78
Medullary incidence	0.274	0.251	0	0.75	0.42	0.49	0	1.37	0.101	0.268	0	0.89
Follicular incidence	0.799	0.716	0	2.75	0.988	1.18	0	5.05	0.572	0.624	0	1.8
Papillary incidence	11.52	5.93	2.12	20.98	15.65	8.1	1.54	29.51	6.58	3.73	0.842	14.3

The minimum and maximum columns represent the value for the year when that variable was least or greatest, respectively.

FNAB, fine needle aspiration biopsy.

Figure 1 shows longitudinal trends in ultrasound, FNAB, and cancer incidence. The annual rate of thyroid ultrasounds increased from 613 per 100,000 in 1997 to 2917 per 100,000 in 2019. Overall, females underwent 3 times more thyroid ultrasounds, 3 times the rate of FNABs, and had 2.4 times the cancer incidence rates than those observed in males. In addition, the absolute difference in the number of ultrasounds between sexes increased over time (Fig. 1A). Although overall FNAB per 100,000 increased substantially over time among KPWA members, by 949 per hundred thousand for females and 360 for males (Fig. 1B), the number of ultrasounds that led to a biopsy decreased over time (Fig. 1C). This suggests that the rate of suspicious findings on ultrasound that led to a biopsy recommendation decreased over the study period. Moreover, rates of FNAB per ultrasound were higher in males for the majority of years studied. Cancer incidence in females was 1.3 times higher than among males (5.00 vs. 3.98 per hundred thousand) in 1997 and increased to 3.4 times higher in 2019 (18.98 vs. 5.60; Fig. 1D). Across all years, the rate of all thyroid cancer diagnoses per FNAB was not consistently different between sexes (Fig. 1E).

FIG. 1.

Rates of thyroid ultrasound, biopsy, and cancer over time in the Kaiser Permanente Washington population.

Joinpoint analysis is summarized in Figure 2 (Supplementary Tables S1, S2, S3 and S4). The annual percent change (APC) in ultrasound rate was 8.04 from 1997 to 2009 for females. After 2009, the APC dropped significantly to 3.50. Among men, the 1997–2009 APC was 8.78 and changed to 5.36 after 2009. For FNAB rates, Joinpoint analysis indicated four discrete trends for females: the APC was −4.55 from 1997 to 2001, 8.89 from 2002 to 2009, 1.44 from 2009 to 2017, and −15.50 after 2017. Among men, only three discrete periods were noted: the APC was 7.99 from 1997 to 2008, 3.33 from 2008 to 2017, and −16.01 after 2017.

FIG. 2.

(A) Joinpoint regression analysis of male and female thyroid ultrasound rates from 1997 to 2019. (B) Joinpoint regression analysis of male and female biopsy rates per 100,000 covered lives from 1997 to 2019.

The results of the Poisson model with offset predicting FNAB rate (FNAB among individuals with ultrasound) are summarized in Table 3. Adjusting for comorbid conditions, high ADI, in-network status, race/ethnicity, and age, the effect of sex on FNAB persists (see Fig. 1 for univariate). Males who had an ultrasound had 1.06 times more biopsies [confidence interval 1.01–1.11], although this may not represent a clinically meaningful difference. Having a high comorbidity burden and being of age 50–64 were also associated with greater FNAB rates. Age 18–34 and the proportion of patients out of the KPWA network for the year were associated with lower FNAB rates. The year was also negatively associated with FNAB rates, with FNAB rates decreasing by 0.98 times per year among people with prior thyroid ultrasounds; this indicates that, over time, fewer people received biopsies after ultrasound.

Table 3.

Poisson Model Predicting Biopsy Rate over Time

Variables	Adj. rate ratio	CI
Male sex	1.060	1.010	1.111^*
Age 18–34	0.798	0.738	0.862^**
Age 50–64	1.081	1.013	1.154^*
Age 65+	1.039	0.906	1.191
Year	0.979	0.976	0.983^**
KPWA non-network	0.378	0.282	0.508^**
One or more comorbidity	1.842	1.286	2.639^**
Top tertile of ADI	0.373	0.097	1.415

p = 0.05.

p = 0.001.

Ref. Cat. female sex, age 35–49, KPWA in network, no comorbidity, mid and lowest ADI tertile.

ADI, Area Deprivation Index; CI, confidence interval.

The female-to-male diagnosis rate ratios across all years (female/male) were 2.78 for PTC <2 cm, 1.69 for PTC >2 cm, and 1.47 for all other thyroid cancers. In other words, the majority of the sex difference in thyroid cancers is driven by small PTCs, and as the clinical significance and mortality risk of the tumor type increases, the ratio approaches 1.

Discussion

Sex differences in thyroid cancer are explained in part by differential rates of diagnostic ultrasound. In other words, although females develop more thyroid cancer, the sex difference in incidence is mirrored by diagnostic imaging patterns.^8,14,31 We observed a pattern where sex differences in ultrasound, FNAB, and cancer diagnosis rates increased over time. The sex-specific and overall incidence trends in our data are consistent with incidence trends in SEER data. Specifically, increases in thyroid cancer incidence have been driven by greater diagnoses of papillary and small tumors, which have leveled off over the last few years.^17,32,33 However, when we examined biopsies as a percentage of ultrasounds performed and cancers as a percentage of biopsies, the sex difference disappeared. Males consistently underwent FNAB at slightly higher rates than females over time, and cancer diagnoses following FNAB were not consistently different between male and female patients who received FNAB. Furthermore, as ultrasounds increased over time, the FNAB rate per ultrasound decreased. This effect persisted in adjusted analysis.

There are likely provider and patient behaviors underlying this phenomenon. Previous research shows that (1) physicians with overall high testing rates have patient panels with higher thyroid cancer rates⁵ and (2) a third of physicians would order a thyroid ultrasound without indication.^11,34 These findings suggest that increased thyroid cancer incidence is likely a supply-side, provider-driven effect resulting from providers ordering more ultrasounds over time.^5,11,34 Patient-level behavior also affects incidence. For example, females seek care more and have a higher lifetime prevalence of metabolic thyroid disease and thus may be uniquely predisposed to excess imaging studies through exposure to the health care system.¹⁸ Data from Korea, which implemented widespread commercial thyroid cancer screening during the early 2000s, show that females received ultrasounds at a higher rate than males and that screening drives excessive, unnecessary thyroid cancer diagnosis through increased papillary tumor diagnosis.⁴

The disappearance of the sex differences in biopsy and cancer incidence after imaging suggests that the higher incidence of thyroid cancer in females is owed in part to differences in ultrasound referrals rather than a much higher underlying rate of cancer. This is further supported by the incidence rate ratios, which demonstrate that PTCs <2 cm largely drive the increase in diagnosis in females; for tumor types with higher morbidity and mortality, the sex-based rate ratios are closer to one, although a small sex difference is still apparent. Two centimeters is the threshold indicated in the American Thyroid Association (ATA) guideline for biopsy of nodules absent suspicious features and, without clinical data from provider notes, represents the best available criterion for indolence. These findings are consistent with previously published rates in a recent meta-analysis where there was no sex difference in thyroid cancer incidence in decedents, and analysis of SEER data demonstrated that the sex incidence rate ratio approached 1 as tumor lethality increased.⁷

The results of our Joinpoint analysis of FNAB rates suggest that these tests are responsive to guidelines with each inflection point matching a guideline release from the ATA³⁵ or the American College of Radiology.^36,37 This is consistent with previous findings demonstrating that PTC incidence trends are responsive to changes in guidelines.³⁸ Notably, although FNAB rates responded to changes in guidelines, ultrasound rates, which are not subject to guidelines due to their low risk, have continued to rise in both sexes unchecked.¹² In addition, FNAB rates in females changed more than rates in males at each guideline implementation, suggesting that responsiveness to guideline changes may occur more readily in populations with greater overuse of diagnostic imaging. Therefore, the responsiveness of FNAB rates to changes in guidelines may provide a potential policy lever to reduce ultrasound rates. It is, however, important to note that we only saw a decrease over 2 years (2017–2019), and we do not know if this trend will continue.

The limitations of our study warrant mention. KPWA is a single-managed care system in the Pacific Northwest and is uniquely incentivized to control costs. A major limitation of our analysis is that we did not have access to the indication for ultrasound, which would allow for stronger inference regarding appropriate utilization of imaging. Practice patterns outside KPWA likely vary more widely. Also, because KPWA is a large urban group practice where guideline implementation is supported by electronic health records and greater financial resources, guidelines may be implemented more readily than other practice types.^39,40 A proportion of the population studied received care outside the KPWA network. Although we may not have captured everything that occurred outside the network, we should have captured the majority given that KPWA paid for these procedures as an insurer. Additionally, this study was observational, and although longitudinal data lend robustness to our results, we cannot infer causality. In addition, we used the 2018 ADI, which only covers 2014–2018. Although we looked back 18 months for any previous thyroid cancer, an individual could have had a diagnosis before that time that we did not capture leading to the inclusion of a small number of recurrences. The small sample and high year-to-year variability in tumor counts made it difficult to effectively summarize longitudinal trends in cancer incidence, and we were, therefore, unable to model incidence trends. Previous work demonstrates that the KPWA population is generalizable to working-aged adults in the United States, but the population is likely more socioeconomically homogeneous with somewhat lower rural representation.⁴¹

Overall, our longitudinal analysis of thyroid ultrasounds in a large private health system demonstrates that the sex difference in thyroid cancer is consistent with the sex difference in imaging rather than a large underlying difference in cancer incidence. Because thyroid ultrasound is a procedure with few inherent risks, ultrasound rates have continued to increase, even in years when FNAB rates have been subject to more stringent, evidence-based guidelines. As a result, females, who have only a slightly higher thyroid cancer incidence than males, likely bear an undue burden of avoidable cancer diagnosis and treatment. The potential for widespread bias in health care practice to create observable differences in cancer epidemiology that do not reflect true underlying trends is an essential area of further research.

Footnotes

Authors’ Contributions

Conceptualization: S.F.-T., D.O.F., L.D., B.H., M.V., E.J.A.B., and A.Y.C. Data curation: E.J.A.B., S.F.-T., M.V., C.K., and R.D. Formal analysis: S.F.T., M.V., C.K., and N.A. Funding acquisition: S.F.-T., D.O.F., L.D., B.H., E.J.A.B., and N.A. Critical editing of the article: all authors.

Author Disclosure Statement

The authors have no conflicts of interest to disclose.

Funding Information

This research was supported by the Cancer Surveillance System of the Fred Hutchinson Cancer Research Center, which is funded by contract nos.: N01-CN-67009, N01-PC-35142, N01-PC-2010–00029, HHSN261201300012I, N01 PC-2013-00012, HHSN261201800004I, and PC-2018-00004 from the SEER Program of the National Cancer Institute (NCI) with additional support from the Fred Hutchinson Cancer Research Center and the state of Washington. The work was also funded by 3R01CA251566-02S1 from National Institutes of Health (NIH)/NCI (D.O.F. and S.F.-T.) and R01CA251566 from NIH/NCI.

Supplementary Material

Supplementary Figure S1

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Table S4

References

Surveillance, Epidemiology, and End Results (SEER) Program. SEER*Stat Database: Incidence - SEER Research Data, 8 Registries, Nov 2021 Sub (1975–2020) - Linked To County Attributes - Time Dependent (1990–2020) Income/Rurality, 1969–2020 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, released April 2023, based on the November 2022 submission. Available from: www.seer.cancer.gov [Last accessed: November 2023 ].

Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov). SEER*Stat Database: Mortality - All COD, Aggregated With State, Total U.S. (1969-2020) <Katrina/Rita Population Adjustment>, National Cancer Institute, DCCPS, Surveillance Research Program, September 2022. Underlying mortality data provided by NCHS. Available from: www.cdc.gov/nchs [Last accessed: November 2023 ].

Limaiem

, Rehman

, Anastasopoulou

, et al. Papillary thyroid carcinoma. In: StatPearls [Internet]. StatPearls Publishing: Treasure Island (FL); 2023. https://www.ncbi.nlm.nih.gov/books/NBK536943/

Decallonne

, Van den Bruel

, Macq

, et al. The impact of regional variation in clinical practice on thyroid Cancer diagnosis: a national population-based study. Eur Thyroid J, 2020; 9(1):32–39.

Hall

, Irish

, Groome

, et al. Access, excess, and overdiagnosis: the case for thyroid cancer. Cancer Med, 2014; 3(1):154–161.

Hall

, Irish

, Griffiths

, et al. Explaining the variation in surgical practice for differentiated thyroid cancer in Ontario, Canada. JAMA Otolaryngol Head Neck Surg, 2019; 145(10):949–954.

LeClair

, Bell

, Furuya-Kanamori

, et al. Evaluation of gender inequity in thyroid cancer diagnosis: differences by sex in US thyroid cancer incidence compared with a meta-analysis of subclinical thyroid cancer rates at autopsy. JAMA Intern Med, 2021; 181(10):1351–1358.

Lincango-Naranjo

, Solis-Pazmino

, El Kawkgi

, et al. Triggers of thyroid cancer diagnosis: a systematic review and meta-analysis. Endocrine, 2021; 72(3):644–659.

Mitchell

, Livingston

, Chang

, et al. Trends in thyroid cancer demographics and surgical therapy in the United States. Surgery, 2007; 142(6):823–828; discussion 828.e1.

10.

Davies

, Welch

. Current thyroid cancer trends in the United States. JAMA Otolaryngol Head Neck Surg, 2014; 140(4):317–322.

11.

Pasternak

. When do physicians actually order neck ultrasound for evaluation of thyroid nodules? Clin Thyroidol, 2021; 33(11):481–483.

12.

Edwards

, Iñiguez-Ariza

, Singh Ospina

, et al. Inappropriate use of thyroid ultrasound: a systematic review and meta-analysis. Endocrine, 2021; 74(2):263–269.

13.

Haymart

, Banerjee

, Reyes-Gastelum

, et al. Thyroid ultrasound and the increase in diagnosis of low-risk thyroid cancer. J Clin Endocrinol Metab, 2019; 104(3):785–792.

14.

Zevallos

, Hartman

, Kramer

, et al. Increased thyroid cancer incidence corresponds to increased use of thyroid ultrasound and fine‐needle aspiration: a study of the Veterans Affairs health care system. Cancer, 2015; 121(5):741–746.

15.

Rahbari

, Zhang

, Kebebew

. Thyroid cancer gender disparity. Future Oncol, 2010; 6(11):1771–1779.

16.

Sosa

, Hanna

, Robinson

, et al. Increases in thyroid nodule fine-needle aspirations, operations, and diagnoses of thyroid cancer in the United States. Surgery, 2013; 154(6):1420–1427.

17.

Kim

, Durr

, Darbinian

, et al. Ten-year thyroid cancer incidence in an integrated healthcare delivery system. Perm J, 2021; 25.

18.

Lam

, Davies

, Sawka

. Women and thyroid cancer incidence: overdiagnosis versus biological risk. Curr Opin Endocrinol Diabetes Obes, 2022; 29(5):492–496.

19.

Ross

, Ng

, Brown

, et al. The HMO research network virtual data warehouse: a public data model to support collaboration. EGEMS (Wash DC), 2014; 2(1):1049.

20.

Kind

AJH

, Buckingham

. Making neighborhood disadvantage metrics accessible: the neighborhood atlas. N Engl J Med, 2018; 378(26):2456–2458; doi: 10.1056/NEJMp1802313

21.

University of Wisconsin School of Medicine and Public Health. Area Deprivation Index. 2018. Available from: https://www.neighborhoodatlas.medicine.wisc.edu/

22.

Charlson

, Pompei

, Ales

, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis, 1987; 40(5):373–383.

23.

Deyo

, Cherkin

, Ciol

. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol, 1992; 45(6):613–619.

24.

Harris

, Millman

, van der Walt

, et al. Array programming with NumPy. Nature, 2020; 585(7825):357–362; doi: 10.1038/s41586-020-2649-2(Publisher link)

25.

McKinney

, others . Data structures for statistical computing in Python. In: Proceedings of the 9th Python in Science Conference. 2010; pp. 51–6.

26.

Hunter

. Matplotlib: a 2D graphics environment. Comput Sci Eng, 2007; 9(3):90–95.

27.

R Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing: Vienna, Austria; 2021. Available from: https://www.R-project.org/

28.

StataCorp. Stata Statistical Software: Release 18. StataCorp LLC: College Station, TX; 2023.

29.

Joinpoint Regression Program, Version 5.0.2 - May 2023. Statistical Methodology and Applications Branch, Surveillance Research Program, National Cancer Institute.

30.

Kim

, Fay

, Feuer

, et al. Permutation tests for Joinpoint regression with applications to cancer rates. Statist Med, 2000; 19(3):335–351. (correction: 2001;20:655).

31.

Kitahara

, Schneider

. Epidemiology of thyroid cancer. Cancer epidemiology, biomarkers & prevention. Cancer Epidemiol Biomarkers Prev, 2022; 31(7):1284–1297.

32.

Lim

, Devesa

, Sosa

, et al. Trends in thyroid cancer incidence and mortality in the United States, 1974-2013. JAMA, 2017; 317(13):1338–1348.

33.

Megwalu

, Moon

. Thyroid cancer incidence and mortality trends in the United States: 2000–2018. Thyroid, 2022; 32(5):560–570.

34.

Chen

, Reyes-Gastelum

, Radhakrishnan

, et al. Physician-reported misuse of thyroid ultrasonography. JAMA Surg, 2020; 155(10):984–986.

35.

Cooper

, Doherty

, Haugen

, et al. American Thyroid Association (ATA) Guidelines Taskforce on Thyroid Nodules and Differentiated Thyroid Cancer. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer: The American Thyroid Association (ATA) guidelines taskforce on thyroid nodules and differentiated thyroid cancer. Thyroid, 2009; 19(11):1167–1214.

36.

Horvath

, Majlis

, Rossi

, et al. An ultrasonogram reporting system for thyroid nodules stratifying cancer risk for clinical management. J Clin Endocrinol Metab, 2009; 94(5):1748–1751.

37.

Tessler

, Middleton

, Grant

, et al. ACR thyroid imaging, reporting and data system (TI-RADS): white paper of the ACR TI-RADS committee. J Am Coll Radiol, 2017; 14(5):587–595.

38.

Pereira

, Williams

, Hallanger Johnson

, et al. Thyroid cancer incidence trends in the United States: association with changes in professional guideline recommendations. Thyroid, 2020; 30(8):1132–1140.

39.

, Jeffs

, Barwick

, et al. Organizational contextual features that influence the implementation of evidence-based practices across healthcare settings: a systematic integrative review. Syst Rev, 2018; 7(1):72–79.

40.

Davino-Ramaya

, Krause

, Robbins

, et al. Transparency matters: Kaiser Permanente’s national guideline program methodological processes. Perm J, 2012; 16(1):55–62.

41.

Davis

, Voelkel

, Remmers

, et al. Comparing Kaiser Permanente members to the general population: implications for generalizability of research. Perm J, 2023; 27(2):87–98.