Patient Requests for Tests and Treatments Impact Physician Management of Hypothyroidism

Abstract

Background: Levothyroxine is one of the most commonly prescribed medications in the United States. Although prior research focused on over- and undertreatment and patient dissatisfaction with thyroid hormone, little is known about physician-reported barriers to managing thyroid hormone therapy. In addition, the impact of patient requests for tests and treatments on hypothyroidism management remains unexplored.

Methods: We randomly surveyed physician members of the Endocrine Society, American Academy of Family Practice and American Geriatrics Society. Respondents were asked to rate barriers to management of thyroid hormone therapy. We conducted multivariable logistic regression analyses to determine correlates with physician report of the most commonly reported barriers, including patient requests.

Results: Response rate was 63% (359/566). Almost half of the physicians reported that patient requests for tests and treatments were somewhat to very likely to being a barrier to appropriate management of thyroid hormone therapy (46%). Endocrinologists (odds ratio [OR] = 2.29 [95% confidence interval, CI 1.03–5.23], compared with primary care physicians) and physicians with more than 25% of patients on thyroid hormone therapy per year (OR = 1.90 [CI 1.05–3.46], compared with those with <25% patients per year) were more likely to report patient requests as a barrier. Physicians with more years in practice were less likely to do so (11–20 years: OR = 0.44 [CI 0.21–0.89]; >20 years: OR = 0.24 [CI 0.12–0.46], compared with ≤10 years). Physician-reported patient requests included requests for preparations other than synthetic thyroxine (52%), adjusting thyroid hormone dose based on symptoms when biochemically euthyroid (52%), maintaining thyrotropin level below the reference range (32%), and adjusting dose according to serum T3 level (21%). Physicians who reported receiving patient requests for the former three unconventional practices were more likely to execute them (p < 0.001, p = 0.014, p < 0.001, respectively).

Conclusions: Physicians reported patient requests for tests and treatments as a common barrier to appropriate thyroid hormone management. In some scenarios, physician adherence to patient requests may be a driver for inappropriate care and lead to harm. Understanding physician-reported barriers to thyroid hormone management and factors associated with physician perception that patient requests are a barrier is key to improving patient care.

Introduction

Hypothyroidism is prevalent in the United States, affecting nearly 10% of the population. It often requires lifelong treatment with exogenous thyroid hormone (1,2). Levothyroxine, the most commonly used synthetic form of thyroid hormone, has been one of the most frequently prescribed medications in the United States for the past several years (3).

Despite guidelines addressing appropriateness of thyroid hormone therapy (4,5), prior studies have shown that over- and undertreatment, as well as inappropriate thyroid hormone use, are common and may place some patients at risk for harm (6–9). Our prior work showed that up to one-fourth of physicians inappropriately use suppressive doses of thyroid hormone to achieve shrinkage of a goiter or thyroid nodule in euthyroid patients (6). In addition, recent studies have demonstrated that a significant proportion of patients with hypothyroidism remain dissatisfied with their treatment and report impaired quality of life compared with controls (10,11).

It is unclear whether patient pressure and physician fulfillment of patient requests are drivers for inappropriate care in the management of hypothyroidism, and whether they play a role in patient satisfaction. Understanding physician barriers to the management of thyroid hormone therapy, including the impact of patient requests, will facilitate improved patient care and provide insight into the development of targeted strategies to streamline treatment in these patients.

To elucidate physician-perceived barriers to appropriate thyroid hormone management in clinical practice, we conducted a nationwide survey among a diverse group of providers, including primary care physicians and specialists. We hypothesized that both physician and patient characteristics and behaviors influence decision making in the management of thyroid hormone therapy.

Materials and Methods

Study population

We conducted a nationwide survey of physician members of the Endocrine Society, the American Academy of Family Practice, and the American Geriatrics Society. Physicians were randomly selected from the societies' active membership lists. To enhance response rate, we used a modified Dillman method of survey administration (12). This method included: (i) an initial mailing of an introductory letter, the survey instrument, a postage-paid return envelope, and a small monetary gift; (ii) a postcard reminder; and (iii) a second survey instrument with postage-paid return envelope to nonresponders. In addition, follow-up telephone encounters were conducted for nonrespondents. Physicians completed the surveys in 2018.

As shown in Figure 1, of the 600 physicians surveyed, 34 physicians were ineligible because they were deceased, ill, retired, not treating patients with thyroid disorders, or had an incorrect mailing address. Of the remaining 566 eligible physicians who were mailed a survey, 359 responded, resulting in a 63% response rate (359/566). Data from the survey were de-identified and recorded by using a double-entry method to ensure <1% error.

FIG. 1.

Flow diagram of survey respondents.

The study was granted exemption by the University of Michigan Institutional Review Board.

Survey design and measures

The survey instrument was developed and reviewed for content validity by a diverse group of experts, which included endocrinologists, primary care physicians, geriatricians, and survey methodologists. It was then piloted in a multidisciplinary group of providers at the University of Michigan before survey administration.

Covariates

Physicians were asked about their specialty, practice setting, years in practice, volume of patients on thyroid hormone replacement therapy, thyroid hormone preparations used, and their knowledge of existing guidelines on hypothyroidism. Physician specialty was categorized as primary care, including those reporting family practice or internal medicine as their specialty, endocrinology, and geriatrics.

Thyroid hormone preparations used included synthetic thyroxine, desiccated preparations, combination T4/T3 therapy, and over-the-counter dietary supplements and nutraceuticals, such as “thyroid support” or iodine supplements. Physicians were asked to rate how often they recommend these preparations to treat their patients by using a 5-point Likert scale as follows: almost always, often, sometimes, seldom, and never. For the purpose of the analyses, thyroid hormone preparations used were categorized as sometimes to almost always prescribing synthetic thyroxine only versus prescribing any thyroid hormone preparation, including at least one of the following: desiccated thyroid, T4/T3 combination, and dietary supplements.

Physician-perceived barriers to appropriate thyroid hormone management

Information on physician-perceived barriers regarding appropriate thyroid hormone replacement therapy was also obtained. Surveyed physicians were asked to rate barriers to the appropriate management of thyroid hormone therapy in their patients, including the following: clinic visit time constraints, multiple providers managing thyroid hormone medication, polypharmacy, concern about patient dissatisfaction, physician concern for cost, lack of clinical decision support tools to determine whether a patient would benefit or not from thyroid hormone therapy, uncertainty about how involved the patient wants to be in the decision-making process, unawareness of guidelines, uncertainty about patient preferences, and patient relying on physician for all recommendations.

Potential patient-level barriers listed included patient nonadherence, patient requests for tests and treatments, lack of patient education and patient concern for cost. A 5-point Likert scale was used with the following options: very likely, likely, somewhat likely, unlikely, and not at all likely. The scale was subsequently dichotomized as somewhat to very likely versus unlikely to not at all likely for the analyses.

Physician-reported patient requests

Surveyed physicians were asked to rate how often their patients request the following: use of preparations other than synthetic T4 alone to treat hypothyroidism, maintaining target thyrotropin (TSH) below normal when treating hypothyroidism, adjusting thyroid hormone dose to treat symptoms suggestive of hypothyroidism when euthyroid, and adjusting thyroid hormone dose according to serum-free T3. A 5-point Likert scale with the following options was used: almost always, often, sometimes, seldom, and never. The scale was dichotomized as sometimes to almost always versus seldom to never for the analyses.

Physician-reported practice patterns

Surveyed physicians were also asked to rate how often they execute the following when managing thyroid hormone management: use of preparations other than synthetic T4 alone to treat hypothyroidism, maintaining target TSH below normal when treating hypothyroidism, adjusting thyroid hormone dose to treat symptoms suggestive of hypothyroidism when euthyroid, and adjusting thyroid hormone dose according to serum-free T3. A 5-point Likert scale with the following options was used: almost always, often, sometimes, seldom, and never. The scale was dichotomized as sometimes to almost always versus seldom to never for the analyses.

Physician perspectives on potential interventions to optimize thyroid hormone management

Surveyed physicians were asked to rate how likely they think a physician-directed intervention, such as a clinical decision support tool, would reduce overtreatment, reduce harm to patients, increase overall adherence of physicians to evidence-based guidelines, and reduce undertreatment. In addition, surveyed physicians were asked to rate how likely they think a patient-directed intervention, such as a decision aid, would reduce overtreatment, reduce harm to patients, increase overall adherence of physicians to evidence-based guidelines, and reduce undertreatment. Finally, we inquired about their willingness to use such tools to facilitate the shared decision-making process. A 5-point Likert scale with the following options was used: very likely, likely, somewhat likely, unlikely, not at all likely. The scale was dichotomized as somewhat to very likely versus unlikely to not at all likely for the analyses.

Statistical analyses

We first generated summary descriptive statistics by using frequencies and percentages. The chi-square test of association was used to determine the most appropriate classifications for the distribution of categorical variables as appropriate. Then, using Pearson's chi-square tests, we performed univariate analyses of physician characteristics associated with the three most commonly reported factors as being somewhat to very likely barriers by physicians, namely patient nonadherence, patient requests for tests and treatments, and multiple providers managing thyroid hormone.

Multivariable logistic regression analyses were subsequently performed to evaluate the independent effect of physician characteristics on reporting of each of the three aforementioned barriers to appropriate thyroid hormone management. Covariates used in the model included volume of patients on thyroid hormone therapy, type of thyroid hormone therapy, physician specialty, practice setting, years in practice, and having read the guidelines. Finally, the chi-square test was used to determine the impact of physician report of patient requests for tests and treatments on practice patterns.

All statistical analyses were performed by using R version 3.5.2. A 95% confidence interval [CI] not including the null value, and a p-value of <0.05 were considered statistically significant.

Results

Physician characteristics are shown in Table 1. Of the 359 respondents, 128 (36%) were primary care physicians, 114 (32%) were endocrinologists, and 113 (32%) were geriatricians. Most physicians were in private practice (51%), whereas 31% worked at a community-based academic affiliate and 18% worked at an academic tertiary care center. More than half of the physicians (52%) reported being in practice for more than 20 years.

Table 1.

Respondent Characteristics (N = 359)

Physician characteristics	N (%)
Specialty
Primary care	128 (36)
Endocrinology	114 (32)
Geriatrics	113 (32)
Practice setting
Private practice	173 (51)
Community-based academic affiliate	106 (31)
Academic tertiary care center	63 (18)
Years in practice
0–10	67 (19)
11–20	102 (29)
>20	186 (52)
Percent patients on thyroid hormone therapy, %
0–25	245 (70)
26–100	107 (30)
Thyroid hormone preparationa
Synthetic T4 only	267 (76)
Any preparation (including at least one of the following: desiccated thyroid, T4/T3 combination, dietary supplements)	85 (24)
Read guidelines on hypothyroidism
Yes	131 (36)
No	228 (64)

Missing data not included.

^aPreparations reported as almost always, often, or sometimes used.

Figure 2 shows a list of factors reported by physicians as somewhat to very likely to being barriers to appropriate thyroid hormone management. Common barriers included patient nonadherence (70%), patient requests for tests and treatments (46%), multiple providers managing thyroid medications (45%), polypharmacy (40%), and lack of patient education (39%). Interestingly, almost one-third of physicians reported concern for patient dissatisfaction as a significant barrier to thyroid hormone management (32%). There was a significant relationship between patient requests for tests and treatments and concern for patient dissatisfaction being reported as barriers, suggesting that physicians who report patient requests as a barrier also report concern for patient dissatisfaction (p < 0.001).

FIG. 2.

Physician-reported barriers to appropriate thyroid hormone management. Factors shown were reported as somewhat, likely, or very likely to be barriers to thyroid hormone management.

Figure 3 demonstrates the results of the multivariable logistic regression analysis of physician characteristics associated with patient requests for tests and treatments being reported as a barrier to appropriate thyroid hormone management. Endocrinologists (odds ratio [OR] = 2.29 [CI 1.03–5.23], compared with primary care physicians) and physicians with more than 25% of their patients on thyroid hormone therapy (OR = 1.90 [CI 1.05–3.46], compared with those having <25% of patients on thyroid hormone therapy) were more likely to report patient requests for tests and treatments as a barrier for managing thyroid hormone therapy. On the contrary, physicians with more years in practice (11–20 years: OR = 0.44 [CI 0.21–0.89]; >20 years: OR = 0.24 [CI 0.12–0.46] compared with those with ≤10 years in practice) were less likely to do so.

FIG. 3.

Physician characteristics associated with reporting of patient requests for tests and treatments as somewhat to very likely to being a barrier to appropriate thyroid hormone management.

Multivariable logistic regression analysis did not reveal any statistically significant associations between physician characteristics and patient nonadherence as a reported barrier. However, physicians who reported patient requests for tests and treatments as a barrier to appropriate thyroid hormone management were also more likely to report patient nonadherence as a barrier (p = 0.001). Physicians practicing for >20 years were significantly less likely to report multiple providers managing thyroid hormone as a barrier in multivariable logistic regression analyses (OR = 0.53 [CI 0.28–0.98]).

Figure 4 shows patient requests as reported by surveyed physicians and physician-reported practice patterns. Patient requests included adjusting thyroid hormone dose based on symptoms when biochemically euthyroid (52%), requests for preparations other than synthetic thyroxine (52%), maintaining TSH level below the reference range (32%), and adjusting thyroid hormone dose according to serum T3 level (21%). In addition, it is shown that 21% of physicians would adjust thyroid hormone dose based on symptoms when biochemically euthyroid, 12% of physicians would prescribe preparations other than synthetic thyroxine, 15% would maintain TSH level below reference range, and 8% would adjust thyroid hormone dose according to serum T3 levels.

FIG. 4.

Influence of patient requests for tests and treatments on physician-reported practice patterns. Patient requests shown were sometimes to almost always received per physician report. Practice patterns shown as sometimes to almost always being executed per physician report.

Patient management requests were associated with physician-reported practice patterns. That is, if a physician reported that patients sometimes to almost always requested the aforementioned unconventional management practices, then in all but one scenario, the physician reported sometimes to almost always executing this practice (adjusting thyroid hormone dose based on symptoms when biochemically euthyroid, p = 0.014; use of preparations other than synthetic thyroxine, p < 0.001; maintaining TSH level below reference range, p < 0.001).

Table 2 demonstrates physician perspectives on potential interventions that may help to optimize thyroid hormone therapy. A total of 84% of respondents stated that a physician-directed intervention would be somewhat to very likely to reduce overtreatment, 85% that it would reduce harm to patients, 89% that it would increase overall adherence to guidelines, and 82% that it would reduce undertreatment. Overall, 81% of respondents stated that they would be somewhat to very likely interested or willing to use a clinical decision support tool to facilitate the shared decision-making process. A total of 79% of respondents stated that a patient-directed intervention would be somewhat to very likely to reduce overtreatment, 79% that it would reduce harm to patients, 84% that it would increase overall adherence to guidelines, and 81% that it would reduce undertreatment. Overall, 80% of respondents stated that their patients would be somewhat to very likely interested or willing to use a decision aid to facilitate the shared decision-making process.

Table 2.

Physician Perspectives Regarding Potential Interventions to Optimize Management of Thyroid Hormone Therapy

Potential interventiona	Reduce overtreatment, N (%)	Reduce harm to patients, N (%)	Increase overall adherence to evidence-based guidelines, N (%)	Reduce undertreatment, N (%)
Physician-directed intervention, such as a clinical decision support tool	301 (84)	306 (85)	320 (89)	295 (82)
Patient-directed intervention, such as a decision aid	284 (79)	285 (79)	300 (84)	290 (81)

Missing data not included.

^aReported as somewhat to very likely to have an effect.

Discussion

Findings from this large nationwide survey of a diverse group of physicians treating patients on thyroid hormone therapy showed that patient requests for tests and treatments are commonly reported as a barrier to appropriate management of thyroid hormone, in addition to patient nonadherence and multiple providers managing thyroid hormone therapy. Endocrinologists, high case volume physicians, and physicians with fewer years in practice were more likely to report patient requests as a barrier in the management of hypothyroidism.

Interestingly, patient requests for tests and treatments were cited as a significant barrier to the management of hypothyroidism by almost half the treating physicians. It has been previously shown that patient requests for diagnostic testing or treatment occur in approximately half of primary care visits (13). In addition, it has been demonstrated that most patient requests for treatments are honored, even though they may not necessarily represent the physicians' first choice of management, and may be influenced by direct consumer advertising (14–17). Similarly, in our study, physicians who felt pressure from patient requests were more likely to honor them. A survey of 544 patients and their 15 primary care physicians concluded that the physicians' perceptions of patients' expectations were the strongest predictors of the decision to prescribe a particular treatment, even if they considered it inappropriate (18).

In addition, in an observational study of 824 patient visits within 135 primary care physician practices in Northern California, physician-perceived visit difficulty was associated with patient requests for diagnostic tests, especially for those in a multidisciplinary specialty practice (19). This suggests that similar to our study, patient requests constitute a barrier to optimal patient care in other clinical settings.

In our study, endocrinologists were more likely to report patient requests as a barrier compared with primary care physicians and geriatricians. Patients on thyroid hormone replacement who are referred to endocrinologists may be more likely to have a complicated treatment course, such as patients in whom it is difficult to render and maintain a euthyroid state. It is possible that these patients may be more likely to request alternative thyroid hormone treatments and/or dose adjustments, in an effort to achieve euthyroidism and alleviate symptoms believed to be related to their thyroid hormone management.

More than half of the physicians in our study reported that their patients request thyroid hormone preparations other than synthetic thyroxine and adjustment of thyroid hormone dose based on symptoms even when biochemically euthyroid. Guidelines on the management of hypothyroidism have long recommended against both these clinical practices due to insufficient evidence to support their implementation (4,5). However, despite the notion that levothyroxine monotherapy maintains an adequate level of serum T4, and that the iodothyronine deiodinases provide physiologic regulation of T3 availability (20), the use of combination T4 and T3 therapy is still employed in select patients (21–23). This may stem from dissatisfaction in a subset of hypothyroid patients with their current therapy or their physicians, as shown in a prior study (11), patient preference or external influences, such as media publicity, or influence from pharmaceutical companies.

A nationwide survey of 389 members of the American Thyroid Association showed a marked increase in the willingness of physicians to prescribe combination therapy or thyroid extract in specific circumstances (21). In this other study, the presence of patient symptoms (p < 0.001), TSH levels (2.2 vs. 3.9 mIU/L, p = 0.03), T3 levels (75 vs. 120 ng/dL, p < 0.001), and request for T3 therapy (p < 0.001) significantly increased the likelihood that alternative therapies would be prescribed in multivariable analyses (21). Interestingly, a qualitative interview study of health professionals in general practice from the United Kingdom (N = 16) showed that most physicians and nurses followed guidelines and relied on blood tests over clinical symptoms to adjust levothyroxine dose (24). However, a qualitative interview study of the attitudes and perceptions of patients to thyroid hormone replacement therapy conducted by the same investigators revealed that patients who remained unwell despite a normal serum TSH level felt that their normal results presented a barrier to further evaluation of their symptoms by their physicians (25).

In addition to some of the aforementioned “unconventional practices” not being concordant with guidelines (e.g., using preparations other than thyroxine alone), others could potentially cause patient harm (e.g., maintaining TSH below reference range and adjusting thyroid hormone dose to symptoms or to serum T3 levels). Older patients are particularly vulnerable to cardiovascular and skeletal risks resulting from prolonged iatrogenic hyperthyroidism (26–30). Specifically, adverse cardiac events of overtreatment with thyroid hormone may include increased risk of atrial fibrillation, especially among those 60 years and older, as well as cardiovascular and all-cause mortality (26–28). In addition, there may be an increased risk of osteoporosis, particularly among postmenopausal women (29,30).

Congruent with patient requests for tests and treatments lies the issue of patient satisfaction. Our study indicates that physicians who reported patient requests as a barrier to thyroid hormone management also reported concern for patient dissatisfaction as a barrier. Physicians are increasingly facing pressure to enhance patient satisfaction in the context of time-constrained visits (31). However, it remains unclear about how large an impact patient requests have on physician-prescribing decisions and how they may be modified by perceived and actual patient satisfaction. Even though some studies have shown that denial of patient requests may lead to lower patient satisfaction (13,32–35), others have not (35,36).

Unfulfilled requests may have consequences for the physician–patient relationship, and physicians must learn to manage these requests in a respectful and clinically sensible fashion. In a study conducted by Paterniti et al. using audio-recorded visits and postvisit questionnaires of standardized patient visits to primary care offices, patients reported significantly higher visit satisfaction when approaches relying on the patient perspective were used to deny the request (37).

Not surprisingly, patient nonadherence and multiple physicians managing thyroid hormone were commonly reported barriers. In an era of increasingly complex health care, multiple physicians are often involved in patients' care, especially for older adults, which often leads to fragmentation of care (38). For example, Medicare beneficiaries see a median of seven providers, including two primary care providers and five specialists, in four practices each year (39). Prior studies have shown that gaps in information and poor communication across providers caring for the same patient can lead to medication errors and patient harm (40,41). Similarly, even though primary care physicians are typically the ones to initiate thyroid hormone, adjustments in dose and/or change in type of preparation may be undertaken by both general practitioners and specialists. These practices can lead to inappropriate treatment, over- or undertreatment. Conflicting views on how to manage thyroid hormone may stem from differential awareness and adherence to guidelines, disparate views on benefit–harm balance, and the role of the patient in the decision-making process.

This study provides important data on physician-perceived barriers to thyroid hormone management from a national, diverse, and multidisciplinary cohort of physicians. In addition, it provides information on the interest and willingness of physicians to use both physician and patient-directed interventions to optimize thyroid hormone management by promoting the shared decision-making process.

However, there are some potential study limitations that merit consideration. First, similar to other survey studies, there is a risk for nonresponse selection bias. However, this was somewhat mitigated by our high response rate. Second, several of the independent variables are based on physician report. Third, we acknowledge that physician report to the survey questions may differ from actual clinical practice. Finally, even though this study focuses on physician perceptions alone, we recognize that patient perceptions and behavior, as well as physician-patient communication, also play a critical role in the decision-making process.

In an era emphasizing the significance of shared decision making, patient-centered and patient satisfaction-driven care, there is a need to train clinicians to deal effectively with patient requests regarding thyroid hormone management to enhance both patient and clinician experiences, while avoiding overtreatment and patient harm and improving patient quality of life. In addition, efforts should be undertaken to improve guideline dissemination and enhance both physician–patient communication and communication among providers to overcome existing barriers.

The willingness of surveyed physicians in our study to utilize shared decision-making instruments to optimize thyroid hormone management is encouraging, and it highlights a high comfort level among physicians with this process. Implementation of physician and patient-directed interventions, such as clinical decision support tools or decision aids, may aid in improving congruency between treatment decisions, patient preferences, and patient satisfaction. In conclusion, besides addressing over- and undertreatment and patient dissatisfaction with thyroid hormone, key to improving care and avoiding patient harm is also further understanding physician-reported barriers to managing thyroid hormone replacement, and factors associated with physician perception that patient requests are a barrier.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work is supported by K08 AG049684 from the National Institute on Aging to Dr. Papaleontiou. Dr. Haymart is funded by R01 CA201198 from the National Cancer Institute and R01 HS024512 from the Agency for Healthcare Research and Quality (AHRQ).