Assessing the Frequency of Deferrable Thyroid Nodule Biopsies to De-Escalate Low-Value Care

Abstract

Objective:

Thyroid nodules are common, yet fewer than 1 in 10 harbors malignancy. When present, thyroid cancer is typically indolent with excellent survival. Therefore, patients who are not candidates for thyroid cancer treatment due to comorbid disease may not require further thyroid nodule evaluation. The goal of this study was to determine the rate of deferrable thyroid nodule biopsies in patients with limited life expectancy.

Methods:

We identified patients who underwent thyroid fine-needle aspiration (FNA) between 2015–2018 at our institution. The primary outcome was the number of deferrable FNAs, defined as FNAs performed in patients who died within two years after biopsy. Secondary outcomes included cytologic Bethesda score, procedure costs, and final diagnosis on surgical pathology. Multivariable logistic and Cox proportional hazards regressions were used to evaluate factors associated with FNA in patients with limited life expectancy.

Results:

A total of 2,565 FNAs were performed. Most patients were female (79%), and 37 (1.5%) patients died within 2 years. Non-thyroid specialists were significantly more likely to order deferrable FNAs (OR 4.13, p<0.001). Of the patients who died within two years, most (78%) had a concomitant diagnosis of non-thyroid cancer, and four went on to have thyroid surgery (Bethesda scores: 3, 4, 4, and 6). Spending associated with deferrable FNAs and subsequent surgery totaled over $98,000.

Conclusions:

Overall, the rate of deferrable thyroid nodule biopsies was low. However, there is an opportunity to reduce low-value biopsies in patients with a concurrent non-thyroid cancer by partnering with oncology providers.

Introduction

Thyroid nodules are very common, with a prevalence of over 60% in the general population.¹ As such, thyroid nodules are frequent incidental findings on imaging or physical exam. While less than 7% of thyroid nodules harbor cancer, a fine needle aspiration (FNA) biopsy may be indicated if there is clinical or radiologic suspicion for malignancy.^2–4 Treatment of thyroid cancer is primarily surgical, with radioactive iodine therapy indicated for a subset of high-risk cancers.⁵ The prognosis for the majority of well-differentiated thyroid cancers is excellent, with a five-year survival rate greater than 98%.^{6, 7}

While the detection and diagnosis of thyroid cancer is important, the frequency at which benign thyroid nodules are incidentally identified and evaluated is responsible for significant healthcare utilization without clear benefit to patients. As a result, the United States Preventative Services Taskforce recommends against screening thyroid exams in asymptomatic patients to avoid over-diagnosis and overtreatment.⁸ Thyroid nodule evaluation may involve multiple healthcare visits for laboratory testing, imaging, and biopsy, which place a significant economic, time, and psychological burden on the patient.⁹ If surgery is ultimately recommended, this also exposes patients to procedural risks. Therefore, it is important to consider the patient’s clinical context when choosing to further evaluate incidental thyroid nodules.

Because thyroid cancers are typically indolent with very high survival rates, patients with limited life expectancy or competing medical comorbidities in particular may not benefit from further evaluation of a thyroid nodule.^{7, 10} In fact, active surveillance of low-risk, small papillary thyroid cancer has been increasingly accepted as a safe alternative to surgical management for many patients, suggesting that deferring evaluation of thyroid nodules in patients with other serious medical problems may often be appropriate.^{11, 12} To that end, a recent white paper from the American College of Radiology explicitly stated that patients with serious comorbidities or limited life expectancy should not have further evaluation of an incidentally discovered thyroid nodule unless clinically warranted or specifically requested.¹⁰

Therefore, as healthcare costs continue to climb in the United States, identifying opportunities to defer thyroid nodule biopsies in patients who are unlikely to benefit from them represents one way to de-escalate low-value care and mitigate unnecessary costs while optimizing patient outcomes.¹³ To examine current practice at our own institution and promote more judicious thyroid nodule management, the goal of this study was to evaluate the rate of deferrable thyroid nodule biopsies in patients with limited life expectancy.

Methods

Study Population:

This is a retrospective cohort study of all patients who underwent an outpatient thyroid nodule FNA at a high-volume, urban, academic institution between March 2015 and December 2018. Current Procedural Terminology codes 10005 and 10022 were used to identify thyroid FNA procedures, which are performed by both endocrinologists and radiologists at our institution.

Variables:

Patient demographic and clinical data were extracted from the electronic health record, including age, sex, race, date of FNA, date of death (if applicable), as well as specialty and degree of the provider who ordered the FNA. Endocrinologists, otorhinolaryngologists, and endocrine surgeons were classified as thyroid specialists based upon health system clinical practice; all other providers were classified as non-thyroid specialists. Elixhauser comorbidity scores were calculated using the International Classification of Diseases, Version 10 codes documented in each patient’s problem list and past medical history at the time of the FNA.¹⁴ Nodule size and Bethesda scores were abstracted from FNA cytology reports.¹⁵ If there were multiple Bethesda scores for a single patient (i.e., from multiple FNAs), the highest score was selected for inclusion in the analysis. Surgical pathology results were also collected for patients who subsequently underwent surgical resection. Data on molecular testing was not available for review.

Outcomes:

The primary outcome was the rate of deferrable thyroid nodule biopsies in patients with limited life expectancy, defined as biopsies performed in patients who were retrospectively identified as having expired within two years of the date of the FNA. Secondary outcomes included Bethesda score, presence of thyroid cancer on final surgical pathology, and associated healthcare costs.

Statistical Analysis:

Descriptive statistics were calculated. Categorical variables were reported as frequencies with percentages, and continuous variables were reported as medians with interquartile ranges or means with standard deviations, as appropriate. Chi-square tests, Wilcoxon rank sum tests, and Student’s t-tests were used to compare characteristics between subgroups. Multivariable logistic regression was used to evaluate variables associated with patient mortality within two years of FNA, including age, sex, Elixhauser comorbidity score, and ordering provider specialty. Cox proportional hazard regression analysis was also performed to investigate the relationship between ordering provider and patient life expectancy. Analyses were completed in Stata/IC, version 12.1 (Stata Corp, College Station, TX) and RStudio, version 2023.03.0 (R Development Core Team, Vienna, Austria). Our university’s Institutional Review Board deemed this investigation to be part of a quality improvement initiative and therefore exempt from full review.

Cost Analysis:

Publicly available data regarding total charges for FNA, total thyroidectomy, and thyroid lobectomy within our healthcare system were obtained. Additionally, negotiated reimbursement amounts for FNA and surgical services from 66 different insurance payors were obtained. To estimate the overall cost burden of the deferrable thyroid nodule biopsies and their subsequent management during the study period, the total charges and median reimbursement amounts for FNA, total thyroidectomy, and thyroid lobectomy were multiplied by the number of patients who utilized these services and also expired within two years of biopsy. Furthermore, for each patient in this group who received indeterminate cytology (Bethesda 3 or 4) on FNA, $3,600 was added to the total estimated spending based on the Medicare reimbursement rate for molecular testing, as it is standard practice at our institution to send molecular testing on indeterminate cytology.¹⁶

Results

Study Cohort:

A total of 2,565 outpatient FNAs were performed on 2,166 unique patients during the study period. The majority of patients were female (79%) and non-Hispanic White (66%). The mean patient age was 54.3 ± 14.4 years, and 24% of the patients had an Elixhauser comorbidity score of ≥3, as shown in Table 1.

Table 1.

Characteristics of patients who underwent thyroid nodule fine needle aspiration, 2015–2018.

Characteristic	N (%)
Sex
Male	459 (21)
Female	1707 (79)
Race/ethnicity
Non-Hispanic White	1432 (66)
Non-Hispanic Black	489 (23)
Hispanic or Latino	35 (2)
Asian or Pacific Islander	91 (4)
Other, Unknown, or Patient Declined	119 (5)
Elixhauser comorbidity score
0	699 (32)
1	571 (26)
2	377 (17)
≥3	519 (24)

Patients with Limited Life Expectancy:

A total of 37 (1.5%) patients died within two years of biopsy and were categorized as having limited life expectancy. Of these patients, the average age was 62.9 ± 11.4 years, and 68% were female. Additionally, 29 of these patients (78%) had a diagnosis of a non-thyroid cancer identified in their medical record at the time of FNA, which was ultimately implicated as the direct or indirect cause of death according to clinical documentation in 28 of the 29 cases. On multivariable logistic regression analysis, increasing age (OR 1.05, 95% CI 1.02–1.08, p=0.001) and having a non-thyroid specialist ordering provider (OR 4.00, 95% CI 2.08–7.69, p<0.001) were significant predictors of FNA performed in patients with limited life expectancy (Table 2).

Table 2.

Multivariable regression analysis of factors associated with death within two years of thyroid nodule fine needle aspiration.

Characteristic	Odds Ratio (95% CI)	Coefficient	p-value
Age, each additional year	1.05 (1.02 – 1.08)	0.05	0.001*
Elixhauser score, each additional point	0.95 (0.76 – 1.15)	−0.05	0.609
Sex
Male	Reference	Reference	Reference
Female	0.56 (0.28 – 1.17)	−0.40	0.269
Race/ethnicity
Non-Hispanic White	Reference	Reference	Reference
Non-White or Unknown	1.18 (0.59 – 2.29)	0.21	0.551
Ordering provider specialty
Thyroid specialist	Reference	Reference	Reference
Non-thyroid specialist	4.00 (2.08 – 7.69)	1.37	<0.001*

^*indicates p<0.05.

Ordering Provider Specialty:

On Cox proportional hazard regression analysis, FNAs ordered by non-thyroid specialists were significantly associated with limited life expectancy (HR 4.13, 95% CI 2.18–7.82, p<0.001). FNAs ordered by medical oncologists and radiation oncologists were associated with the highest rate of patient death within two years of biopsy (HR 18.96, 95% CI 8.58–41.90, p<0.001), as shown in Figure 1. Of the FNAs ordered by providers in medical and radiation oncology, 71% were ordered by individuals with an MD or DO degree, while the remainder were ordered by nurse practitioners and physician assistants.

Figure 1. Vital status two years after thyroid nodule fine needle aspiration, stratified by ordering provider specialty.

Of the FNAs ordered within the study period, 77% were ordered by a thyroid specialist: 66% by endocrinologists (n=1694), 9% by otorhinolaryngologists (n=225), and 2% by endocrine surgeons (n=52). The remaining 23% were ordered by non-thyroid specialists: 2% by medical and radiation oncologists (n=63), 1.6% by other surgical specialists (n=41), 0.4% by obstetricians/gynecologists (n=10), and 19% by other medical specialists (n=480). FNAs ordered by medical and radiation oncologists were associated with the highest rate of patient death within two years of FNA, at 19.1%.

Cytology and Surgical Pathology:

Bethesda scores for FNA cytology were available for 2,313 FNAs (90%). For the entire cohort, benign nodules (Bethesda 2) were the most common finding, detected in 1,159 subjects (50%). Among patients with indeterminate FNA results, cytologic diagnoses were Bethesda 3 for 404 (17%), Bethesda 4 for 365 (16%) and Bethesda 5 for 68 (3%) FNAs. The overall rate of malignancy (Bethesda 6) was 11% (n=254). For the 37 patients with limited life expectancy, none received a non-diagnostic (Bethesda 1) result. Benign cytology (Bethesda 2) was observed in 13 subjects, while an additional 14 subjects had indeterminate cytology (Bethesda 3 or 4). Malignancy (Bethesda 6) was identified in eight patients (Figure 2).

Figure 2. Distribution of Bethesda scores from thyroid nodule fine needle aspiration for (A) the total cohort, and (b) subjects with limited life expectancy.

Among the 2,565 FNAs performed during the study period, 1,159 subjects (50%) were found to have a Bethesda score of 2, 404 (17%) had a Bethesda score of 3, 365 (16%) had a Bethesda score of 4, 68 (3%) FNAs had a Bethesda score of 5, and 254 (11%) of patients were found to have malignancy (Bethesda 6). For the 37 patients with limited life expectancy, 14 subjects had indeterminate cytopathology (Bethesda 3 or 4), and malignancy (Bethesda 6) was identified in eight patients.

Of 2,166 unique patients in the study cohort, 629 (29%) underwent surgery. Of these, 430 (68%) patients were found to have surgical pathology consistent with thyroid carcinoma. Of the 37 patients with limited life expectancy, four (11%) underwent surgery. The cytologic diagnoses for these patients were Bethesda 3, 4, 4, and 6. Per the preoperative notes from these four patients, only one patient reported compressive symptoms related to their nodule, which included mild throat pain and intermittent dysphagia. On final surgical pathology, three of these patients (75%) had carcinoma present (Figure 3). For these patients, the average tumor size was 5.4 ± 0.72 cm, and final surgical pathology revealed follicular carcinoma, papillary carcinoma, and papillary carcinoma with anaplastic carcinomatous change.

Figure 3. Surgical pathology results for patients who underwent surgical management of biopsied thyroid nodules.

Of 2166 patients in the total cohort, 629 (29%) underwent surgical management. Of these, 430 (68%) had evidence of cancer on final surgical pathology. Among 37 patients with limited life expectancy, four underwent surgery; three (75%) had cancer diagnoses on final surgical pathology.

Procedure Cost Analysis:

At our institution, total charges for a single thyroid FNA, total thyroidectomy, and thyroid lobectomy were $744, $82,451, and $59,879, respectively. Therefore, the cumulative charges for the 37 deferrable FNAs during the study period was $27,528. Of the patients with limited life expectancy who underwent surgery, two received total thyroidectomies and two received lobectomies. Thus, the procedure charges associated with potentially deferrable surgery in this cohort totaled $284,660.

The median reimbursement amounts across data available for major insurance payors for a single thyroid FNA, total thyroidectomy, and thyroid lobectomy were $435, $9,621, and $6,203, respectively. Fourteen of the patients in the limited life expectancy cohort had FNAs associated with Bethesda 3 or 4 cytology, thus $3,600 was added to the calculated cost for each intermediate FNA to represent the Medicare reimbursement rate for molecular testing. Therefore, the estimated cost to insurance payors was $66,497 for deferable FNAs, and $31,648 for potentially deferable surgeries, during the study period.

Discussion

Reducing low-value testing and treatment represents an important healthcare priority for patients and payors in the United States. Given the high prevalence of thyroid nodules in the general population, de-escalating the evaluation of incidental nodules in patients with limited life expectancy may result in meaningful improvements in resource stewardship and patient experience. In this study, we sought to investigate the rate of deferrable thyroid biopsies in patients with limited life expectancy at our institution and identify opportunities to improve performance. We found that 1.5% of FNAs conducted during the study period were in patients who died within the subsequent two years. Of these patients who underwent deferrable biopsies, almost 80% had a concurrent cancer diagnosis at the time of their FNA, and only 11% of FNAs in these patients were followed by thyroid resection. Thus, despite the inconvenience to patients and healthcare spending associated with these deferrable biopsies, many of these FNAs appeared to provide negligible clinical benefit. When life expectancy is limited, patients may be best served by selectively proceeding with FNA specifically in cases of large nodule size or presence of compressive symptoms, when surgical management is more likely to provide a meaningful improvement in short-term quality of life.

De-escalation of care is an increasingly important topic across multiple cancer subtypes. Appropriate de-escalation incorporates screening, evaluation and treatment in patients for whom diagnosis or treatment of any discovered malignancy is not likely to confer quality of life or survival benefit. For example, the United States Preventive Services Task Force does not recommend routine screening mammography or colonoscopy for most individuals after age 75 due to lack of survival benefit.^17–20 For incidentally identified thyroid nodules, de-escalation of evaluation and treatment is similarly intuitive. In population-based studies, fewer than 5% of thyroid nodules harbor malignancy, and five-year survival for well-differentiated thyroid cancer without distant metastasis surpasses 98%.^{2, 3, 6, 21} This exceedingly favorable outlook regarding the likelihood and consequences of thyroid nodules harboring malignancy thus shifts the risk-benefit balance especially in patients whose medical comorbidities make them much more likely die with, rather than from, thyroid cancer. The exception to this overarching paradigm shift is when patients present with malignancies that threaten their remaining quality of life. In the case of thyroid cancer, when large tumors cause symptomatic compression and/or invasion of nearby structures, the potential relief that can be provided by surgery often justifies additional evaluation and surgical referral, despite the patient’s limited life expectancy.

In our study, the overall rate of appropriate FNA use was high; however, our finding that 1.5% of patients died within two years of biopsy suggests that there is an opportunity to improve our stewardship of thyroid nodule evaluation. The patients who underwent these deferrable biopsies differed from the remainder of the cohort in that they were older, and the majority had a concurrent diagnosis of non-thyroid cancer. Furthermore, we identified that the FNAs for these patients were most often being ordered by specialists in oncology. These findings suggest that many of these thyroid nodules were identified on staging and/or surveillance imaging ordered for the purpose of assessing patients’ primary malignancy. Prior work has demonstrated that thyroid nodules with avidity for ¹⁸F-fluorodeoxyglucose are incidentally discovered on approximately 1.6% of positron emission tomography (PET) studies, and these nodules have an increased risk of malignancy, at 20–35%, compared to nodules detected by other modalities.^{22, 23} Thus, oncology providers understandably face the dilemma of how to manage these higher-risk nodules in patients with active or recently treated non-thyroid cancers.

Yet the positive downstream effects of reducing deferrable thyroid nodule biopsies are numerous. With regard to healthcare spending, we estimated that the total reimbursements for these biopsies and subsequent surgical procedures during the study period amounted to over $98,000. Importantly, this estimate does not include the costs of outpatient visits, patient and family travel time and missed work, or postoperative hospitalizations, and therefore is a significant underestimate of the total costs associated with deferrable biopsies.²⁴ Beyond these global cost concerns, the financial burden related to thyroid nodule evaluation can be significant for individual patients.⁹ Co-payments for FNA and the associated outpatient appointments may be challenging for patients on fixed incomes, and when surgery is recommended, cost-sharing may become increasingly demanding. In fact, one study found that 28% of patients with thyroid cancer faced material financial hardship after treatment.²⁵ Additionally, deferrable thyroid nodule biopsies may lead to substantial yet preventable psychological burden. In one qualitative study of patients who received a diagnosis of thyroid cancer but elected not to pursue treatment – as several patients with limited life expectancy in our cohort did – participants described significant psychological distress associated with their decision.²⁶ Furthermore, the time and transportation requirements associated with additional appointments for thyroid nodule evaluation may impose substantial strain on patients and their caregivers, which may be further exacerbated if a patient is already encumbered by another life-limiting illness.²⁷

While the proportion of patients with limited life expectancy who went on to undergo thyroid resection after FNA was appropriately lower in this study than the rate for the overall cohort, these potentially low-value operations represent another important reason to de-escalate evaluation of nodules in the first place. In addition to the more extensive time and financial burden imposed by surgery, thyroidectomy exposes the patient to anesthetic and surgical risks, including neck hematoma, hypoparathyroidism, and vocal cord paralysis, which may be less well-tolerated in patients with significant underlying comorbidities.¹¹ Additionally, patients who undergo total thyroidectomy will need to take thyroid hormone supplementation daily. For these reasons, the recognition of specific patients who may not stand to gain improvement in overall prognosis or quality of life from workup and treatment of their thyroid nodules is important in mitigating unnecessary costs, anxiety, inconvenience, and risk. However, as noted in the recent recommendations from the American College of Radiology, these potential benefits of de-escalation must be balanced by an appropriate, personalized assessment of the nodule’s threat to a patient’s health and quality of life.¹⁰ For large, symptomatic nodules or those with characteristics suspicious for aggressive pathology, proceeding with FNA may still be warranted despite a patient’s overall limited prognosis.

In response to the findings from this study, we have created a thyroid nodule evaluation pathway at our institution which includes embedded language in the electronic order for thyroid nodule FNA that recommends deferral in patients with limited life expectancy. Such decision support tools have demonstrated meaningful improvements in appropriate care across a range of healthcare settings.^28–31 As a next step at our institution, we propose the development of a partnership between thyroid specialists and oncologists to reduce the frequency of deferrable thyroid nodule evaluation in patients with concurrent cancer. Specifically, the discovery of an incidental nodule on PET scan can trigger a multidisciplinary case review, such that the patient’s clinical context can be discussed prior to proceeding with FNA. Importantly, to be effective, such interventions must not place excessive burden on providers. Leveraging the electronic health record to provide clinical decision support can avoid many of the challenges associated with expecting oncologists to alter their daily workflows.³¹ For example, a best practice alert based on keywords in the PET scan report, or standard language added to the radiology report itself, can remind oncology providers that an incidental thyroid nodule should be referred for case review prior to further evaluation.

It is important to note that, while the decision to pursue further workup or intervention of a thyroid nodule can be aided by evidenced-based guidelines, a nuanced conversation between the provider and patient is always warranted. While many patients may view unnecessary visits and diagnostic testing as particularly burdensome, others may find relief in having definitive management. In one study which surveyed over 200 patients with thyroid cancer undergoing active surveillance, 32% reported worrying to the extent that their mood was affected, and 14% reported their anxiety sometimes affected their ability to function.³² These aspects are important to address when recommending specific care plans to patients in order to optimize quality of life on an individualized basis.

There are several limitations to this study. Due to its retrospective nature, this study is inherently limited by ascertainment and selection bias. Specifically, we were unable to account patients who were not referred for FNA, those who declined further evaluation, or those who might have undergone surgery or additional management at another institution. Thus, providers may be appropriately deferring thyroid nodule evaluation more often than we could measure; yet, there are clearly still important opportunities for improvement. Furthermore, while patients were retrospectively identified as having died within two years of FNA, it is possible that the information available to clinicians at the time of thyroid nodule evaluation belied their poor prognosis. Additionally, our cost analysis was limited to the costs of the FNAs and subsequent surgical procedures, as we did not have access to data on other associated costs such as outpatient visits and postoperative hospitalizations. Lastly, this is a single institution study, and therefore results may not be readily generalizable.

Conclusions

In patients with limited life expectancy or competing medical comorbidities, the routine evaluation and treatment of thyroid nodules may be of limited clinical benefit, while burdening patients and healthcare resources. In our cohort, although overall rates of appropriate thyroid nodule biopsy were high, FNAs ordered by medical and radiation oncologists were most likely to be associated with patient death within two years. De-escalation of thyroid nodule evaluation and treatment in patients with a concurrent non-thyroid cancer diagnosis represents an opportunity to prevent low-value healthcare and minimize unnecessary procedures close to the end of life.

Highlights.

• We assessed appropriate use of thyroid nodule evaluation at our academic center.

• Patients with limited life expectancy underwent 1.5% of thyroid nodule biopsies.

• These patients rarely underwent thyroid surgery regardless of biopsy results.

• Procedure charges associated with low-value biopsies totaled nearly $285,000.

• Multidisciplinary collaboration may help reduce the frequency of low-value biopsies.

Clinical Relevance.

Thyroid nodules are common and rarely harbor aggressive cancer. Therefore, deferring further evaluation of thyroid nodules in patients with significant comorbidities represents an opportunity to de-escalate low-value care. In this study, we examined characteristics of low-value thyroid nodule care at our institution and describe initiatives to promote more judicious management.