Hypothyroidism and Type D Personality: Results From E-MPATHY, a Cross-sectional International Online Patient Survey

Context: Between 10% and 15% of people with hypothyroidism experience persistent symptoms, despite achieving biochemical euthyroidism. The underlying causes are unclear. Type D personality (a vulnerability factor for general psychological distress) is associated with poor health status and symptom burden but has not been studied in people with hypothyroidism. Objective: To investigate type D personality in hypothyroidism and explore associations with other characteristics and patient-reported outcomes. Design: Multinational, cross-sectional survey. Setting: Online. Participants: Individuals with self-reported, treated hypothyroidism. Intervention: Questionnaire. Main Outcome Measures: Type D personality and associations with baseline characteristics, control of the symptoms of hypothyroidism by medication, satisfaction with care and treatment of hypothyroidism, impact of hypothyroidism on everyday living. Results: A total of 3915 responses were received, 3523 of which were valid. The prevalence of type D personality was 54.2%. Statistically significant associations were found between type D personality and several respondent characteristics (age, marital status, ethnicity, household income, comorbidities, type of treatment for hypothyroidism, most recent TSH level), anxiety, depression, somatization, poor control of the symptoms of hypothyroidism by medication, dissatisfaction with care and treatment of hypothyroidism, and a negative impact of hypothyroidism on everyday living). Discussion: Our study found a high prevalence of type D personality among people with hypothyroidism who responded to the survey. Type D personality may be an important determinant of dissatisfaction with treatment and care among people with hypothyroidism. Our findings require independent confirmation. Close collaboration between the disciplines of thyroidology and psychology is likely to be key in progressing our understanding in this area.

The prevalence of overt and subclinical hypothyroidism globally is reported to be 0.2% to 5.3% and 10%, respectively (1).Despite achievement of a serum TSH within the reference range, persistent symptoms occur in 10% to 15% of people with hypothyroidism (2).Hypotheses for the cause of these symptoms include (1) the inability of levothyroxine (L-T4) to emulate normal physiology and restore T3 levels in tissues; (2) confounding effects of comorbidities; (3) inflammation due to autoimmunity; (4) L-T4 prescribed or taken by patients suboptimally; (5) a high prevalence of unexplained symptoms, known as "somatic symptom disorder" (SSD) (2,3); (6) being more likely to be investigated and diagnosed with minor, incidental perturbations of thyroid function (2); (7) the impact of the diagnostic label of having a chronic disease.
Type D personality has been linked to poor health and has been most often studied in patients with heart disease (4,5).It is associated with persistent symptoms (6,(7)(8)(9), impaired quality of life (10,11), mental health (11)(12)(13)(14), treatment outcomes (11,(13)(14)(15), and adherence with medication (11,13,14,16,17).Type D personality is characterized by a predisposition to pessimism, worry, stress, negative affectivity (NA), and social inhibition (SI) (18,19).The Type D Scale-14 (DS14) questionnaire, a self-administered and validated instrument, can be used to assess for type D personality (18).DS14 is subdivided into the subscales NA and SI.Individuals with high NA tend to experience negative emotions (such as feelings of dysphoria, anxiety, and irritability), have a negative view of themselves, and scan the world for signs of impending trouble.People with high SI tend to feel inhibited, tense, and insecure when with others (18).Individuals who display both high levels of NA and SI have a "distressed" or type D personality (18).Based on DS14 data, the prevalence of type D personality in the world population is estimated to be 21.0% to 38.5% (18,(20)(21)(22).It is found more frequently in primary care patients and in patients with a variety of morbidities (coronary heart disease, hypertension, chronic pain, asthma, tinnitus, sleep apnea, vulvovaginal candidiasis, mild traumatic brain injury, vertigo, melanoma, and diabetic foot syndrome) (18,23).Studies of type D personality in people with thyroid disease are limited to survivors of thyroid cancer, where it did not predict quality of life or adherence with medication (24,25).
In this study, type D personality in people with hypothyroidism was explored.The study questions were (1) what is the prevalence of Type D personality among people with hypothyroidism, and (2) what are the relationships between (2a) type D personality and respondent characteristics and (2b) type D personality and hypothyroidism-related patientreported outcomes?

Study Design
E-MPATHY (E-Mode Patient self-Assessment of THYroid therapy) was a multinational, large-scale, cross-sectional, online study performed among people with a diagnosis of hypothyroidism from 68 countries (findings on patient satisfaction and SSD have been published) (3,26).

Survey Questionnaire
Survey development and delivery are described elsewhere (26).Briefly, the survey was cognitively tested across 5 rounds and translated from English into French, German, Italian, and Spanish by native speakers with idioms replaced.A pilot study of the English questionnaire preceded a full release for selfcompletion online between April 11, 2020, and January 3, 2021.Assessment and definitions of comorbidities are shown in Supplementary Table S1 (27).Potential participants were informed that the survey would take approximately 30 minutes to complete and were encouraged to complete it within 1 session.More than 1 response from the same IP address were automatically blocked.

DS14
The DS14 contains 2 subscales: NA and SI.The NA subscale measures the tendency to experience negative emotions (eg, anxiety, irritability, depressed mood), so people with high scores on this scale are likely to experience distress and dysphoria.The SI subscale assesses the tendency to inhibit selfexpression in social interactions due to fear of disapproval.People with high scores may be reserved and introverted and avoid social interactions to prevent potential negative evaluations from others.The highest score for each subscale is 28 (18).We classified respondents using a 4-group categorization (24).Scores of ≥10 for each subscale are denoted by positive and <10 by negative signs; thus the 4 groups were (1) NA+ SI−, (2) NA− SI+, (3) NA+ SI+ (type D personality), and (4) NA− SI− (reference).A maximal attribution approach was taken: Where someone unequivocally scored more than 10 on a subscale, they were included as positive for that subscale, regardless of how many items they had answered.Similarly, if someone could not score more than 9 when accounting for any unanswered items in that scale, they were included as negative for that subscale.The English version of the DS14 was used, as well as translations into French, German, Italian, and Spanish performed by 2 certified native translators for each language for the purpose of our survey.Although validated French, German, Italian, and Spanish versions of the DS14 are available, these were not used in our study, as this was unknown to the coauthors responsible for designing the questionnaire at the time (P.P., L.H., E.V.N., E.P.).The data presented are for the full dataset (validated English and nonvalidated French, German, Italian, and Spanish translations).In addition to analyzing the full data (all languages, n = 3523), we examined the validated English-language responses (n = 2370) and the pooled nonvalidated translation French, German, Italian, and Spanish responses (n = 1153) and contrasted these with the full data.

SSD
SSD was assessed using the Patient Health Questionnaire-15 (PHQ-15), and a score of ≥10 on the PHQ-15 was considered as compatible with probable SSD (28).A maximal attribution approach was taken.All questions in the PHQ-15 specifically ask about symptoms experienced during the past 4 weeks.

Anxiety, low mood/depression
Participants were asked, "During the past 4 weeks, how much have you been bothered by anxiety?" and "During the past 4 weeks, how much have you been bothered by low mood/depression?" with the following response options: "bothered a little" or "bothered a lot" (considered as having anxiety or low mood/depression) and "not bothered at all" (considered as not having anxiety or low mood/depression).We used the method described previously (3) to create binarized variables for anxiety and low mood/depression.As for exploratory studies, the simplicity of interpretation with clear differentiation of categories offered by binarization provides a starting point to identify subsequent research questions.

Control of symptoms of hypothyroidism by medication
Participants responded to the statement "My hypothyroidism medication controls my symptoms well" with 6 response options on a Likert scale from "strongly disagree" to "strongly agree" and "uncertain." Satisfaction with treatment and care for hypothyroidism Participants responded to "How satisfied are you with the overall care and treatment you have received for your hypothyroidism?" with 6 response options on a Likert scale from "very satisfied" to "very dissatisfied" and "don't know."

Impact of hypothyroidism on everyday living
Participants responded to the statement "My hypothyroidism has affected everyday activities that people my age usually do (eg, exercise, household chores, etc.)" with the 6 response options on a Likert scale from "strongly disagree" to "strongly agree" and "uncertain."

Dissemination of Questionnaire
Advertisements and information sheets to explain the purpose of the survey were prepared in the aforementioned 5 languages and promoted through Thyroid Federation International, a global network of patient thyroid disorder organizations, (https:// thyroid-fed.org/)affiliates, and partners via social media and web pages.

Inclusion Criteria
Participants self-identified as being more than 18 years old and using medication for hypothyroidism.

Institutional Review Board Waiver Statement
The noninterventional nature of the survey and the fact that data were anonymous rendered the study exempt from institutional review board approval.The study was conducted in accordance with the Declaration of Helsinki as revised in 2013.All participants gave informed consent.

Statistical Analyses
This study is a secondary analysis of a study (26), the data set of which was determined to be of sufficient size (>1066) to detect small to medium differences (delta 0.1) in statistics, with good confidence (95% power, alpha .05).The programming software Python 3.11 was used for statistical analyses.
We tested for association of the 4 type D personality groups with demographic and other baseline variables (sex, age, marital status, employment status, ethnic background, years of education, household income, comorbid conditions, current treatment for hypothyroidism, most recent TSH level, cause of hypothyroidism, self-reported anxiety, self-reported depression, PHQ-15 score) using chi-square tests, with a Bonferroni correction for multiple testing.The 4 type D personality groups were compared against the 3 patient-reported outcomes related to hypothyroidism (control of symptoms of hypothyroidism by medication, satisfaction with treatment and care for hypothyroidism, impact of hypothyroidism on everyday living).To explore what drove the associations, partial chi-square values (a standardization of the difference between observed and expected frequencies) were examined, as their size provides insight into the contribution to the chisquare statistic overall based on differences from the statistically expected data distributions.
Data are presented as percentages derived from the observed figures as a fraction of the total number of valid respondents (n = 3523), whereas chi-square data only included those cases where both type D personality status and the variable in question were known.

Respondent Characteristics
A total of 3915 responses were received, 3523 of which contained valid DS14 data (90.0%) and comprised the study population (Table 1 and Fig

Associations with respondent characteristics
Statistically significant associations were found between type D personality and age, marital status, ethnicity, household income, comorbidities, type of treatment for hypothyroidism, most recent TSH levels, anxiety, depression, and SSD (Table 2).With a 4-category outcome and multinominal independent variables, describing the directionality of the associations has challenges.One way to consider how an association is detected is to examine the "partial chi-squares" (Supplementary Table S2) (27).Partial chi-squares represent the contribution of the particular combination of factors.The larger the partial, the greater the contribution to any detected association.The largest contributions to the distribution were most commonly with the reference group (NA−SI−, not having type D personality) and effects arising due to youngest (age 18-40 years) and oldest (age >61 years), being single/divorced, non-White ethnicity, above and below average income, low TSH, desiccated thyroid extract (DTE) and L-T4 + L-T3 treatment, comorbidity, no anxiety, and no depression (Supplementary Fig. S1) (27).
The association with SSD was probably driven by divergences in the expected numbers in the reference group for type D personality (NA−SI−), also suggested by the prevalence of SSD (PHQ-15 score ≥10) in the 4 groups of type D, which was highest in the NA+SI (type D personality) and lowest in the NA−SI− (reference group): NA+SI− 57.3% (421/735), NA−SI+ 36.1% (108/299), NA+SI+ (type D personality) 66.1% (1261/1908), and NA−SI− (reference) 33.0% (192/ 581).No associations were found between type D personality and sex, cause of hypothyroidism, years of education, or employment status.

Associations with patient-reported outcomes
Type D personality was associated with the expression of the view that the thyroid medication taken did not control the symptoms of hypothyroidism well (P = 6.32E-24), with dissatisfaction with care and treatment of hypothyroidism (P = 1.45E-18) and with a negative impact on everyday living (P = 5.00E-29) (Table 3, Fig. 2).Partial chi-squares suggest that NA largely drove these associations (Supplementary Table S2) (27).

English and Non-English Translations of DS14
The results presented were derived from all respondents using the validated English version of the DS14, as well as translations into French, German, Italian, and Spanish performed by 2 certified native translators for each language.The analyses for type D personality were repeated for respondents using data for the English and non-English language translations (Supplementary Table S3) (27).The concordance in the P-values between the analyses from all respondents and English-language respondents was high with only 1 (ethnic background) of 17 variables discordant.The discordance between all respondents and non-English language respondents was slightly greater (4 of 17; age, marital status, ethnicity, and most TSH), with similar results for English vs non-English language respondents (3 of 17, age, marital status, and most recent TSH).

Discussion
Type D personality (NA+SI+) is common, and it is associated with persistent physical symptoms and anxiety and depressive symptoms (13).While impaired quality of life and dissatisfaction with care and treatment are well documented in hypothyroidism (29), data on type D personality are confined to 2 studies in thyroid cancer survivors, showing no association with quality of life or adherence with medication (24,25).In our study, we provide a starting point that highlights areas for exploration.
We used the DS14 (18) to assess type D personality and to test the hypothesis that type D personality may be associated with clinical characteristics and patient-reported outcomes related to hypothyroidism.Respondents' characteristics were similar to people with hypothyroidism reported in the literature (1,30,31).

Type D Personality (NA+SI+)
The proportion of respondents with type D personality (NA+SI+) was higher (54.2%) than the reported prevalence in the general population (21-38.5%)(18,(20)(21)(22) and that reported in patients with heart disease (21-35%) (32), skin diseases (39-43%) (33), and obstructive sleep apnea (32.5%) (34) but similar to that reported for diabetes (52%) (35).It was higher than the 20.8% prevalence reported among survivors of thyroid cancer (25).Type D personality is common among patients with chronic diseases (36); thus it is not surprising there is a high prevalence in people with hypothyroidism, though selection bias may have inflated this figure.Both behavioral and biological mechanisms including common genetic predispositions that implicate type D personality in disease causation have been proposed as potential explanations of the association between type D personality and disease (19).

Type D personality and respondent characteristics
In our study, type D personality traits were associated with several respondent characteristics (clear drivers as revealed by the partial chi-square values were youngest and oldest ages, ethnicity other than White, having above average household income, having mental illness, treatment with DTE and L-T4 + L-T3, having a low most recent TSH level, having no anxiety, and having no depression and not having SSD).The trend noted for clustering of treatment with DTE or L-T4 + L-T3 and non-type D personality is interesting.Given that type D personality traits are stable across time (37), it is unlikely that treatment with DTE or L-T4 + L-T3 would reverse those features.Alternatively, people with hypothyroidism and non-type D personality may be more capable of accessing DTE or L-T4 + L-T3 treatment because they are less socially inhibited.The association of type D personality with a low TSH is interesting and potentially related to the use of DTE or L-T4 + L-T3 often resulting in overtreatment (38).In another study, we showed that probable SSD is highly prevalent in the same cohort of patients (3).An association between SSD and type D personality has been reported in people without hypothyroidism in the past (13) and is now noted in people with hypothyroidism.

Type D personality and patient-reported outcomes
Type D personality was associated with poor control of symptoms of hypothyroidism by medication, dissatisfaction with overall treatment and care for hypothyroidism, and a negative impact of hypothyroidism on everyday living.These associations are probably driven by NA.Our findings are consistent with other studies showing an association of type D personality with negative outcomes and that this effect is mainly driven by NA (39)(40)(41).This may be related to NA predisposing to chronic stress and physiological and immunological responses, unhealthy behaviors as coping mechanisms, and adverse effects on social relationships and support systems.The experience of somatic symptoms in somatization has been associated with harm avoidance and NA (42).Compared to people without somatization, people with somatization show more self-defeating, depressive, and passive-aggressive personality traits and neuroticism and less agreeableness and extraversion (43).

English and Non-English Translations of the DS14
There were only a few differences in the results of the analyses between English and non-English language users; thus the fact that the non-English translations were not independently validated did not appear to impact on our findings.Ethnic background was the only respondent characteristic that was significant in the full data but not statistically significant in the validated English translation data.

Significance of Main Findings
Hypothyroidism is known to be associated with psychological morbidity both before and after the diagnosis of hypothyroidism (44).Our study indicates that type D personality may cluster with hypothyroidism.Given that the majority of patients develop hypothyroidism in middle life, it may be assumed that type D personality precedes the onset of hypothyroidism in most cases.This reasoning suggests that the type D personality traits are not the result of hypothyroidism or its treatment and the poor experiences often described by people with hypothyroidism but that the preexistence of type D personality traits may color the perception of the patient experience.The significance of our findings rests with how clinicians approach the common scenario of people with hypothyroidism and persistent unexplained symptoms.Research in Europe conducted between 2019 and 2021 (45,46) and in Latin America in 2022 (47) shows that thyroid specialists usually offer pharmacological solutions to such patients in the form of combination therapy of L-T4 + L-T3, despite evidence from randomized controlled studies indicating no benefit from combination treatment compared to L-T4 alone (29).
A study of people without hypothyroidism with a high prevalence of type D personality, SSD, anxiety, and depression (13) showed that psychological interventions were associated with improvements in anxiety, depression, and physical symptoms (13).Thus, psychological therapies may be appropriate for some people with hypothyroidism, particularly those with type D personality and SSD, who can be identified by using the validated questionnaires DS14 and PHQ-15.

Limitations
The study has limitations.Some nations were overrepresented, there were very few men, and the demographic characteristics and medical background in our sample could limit generalizability of the results.The sample may not represent the population of people with hypothyroidism since respondents were invited via patient organizations and social media and there was some sample heterogeneity.The diagnosis of hypothyroidism was not validated independently, and it is possible that some patients may have been treated for hypothyroidism despite normal thyroid biochemistry.In addition, the assessments of satisfaction, control of symptoms of hypothyroidism by medication, impact on daily life, and anxiety and depression did not utilize validated instruments.Data on whether depression was treated or not were not available.The survey was conducted during the COVID pandemic, which may have influenced responses.The results of our study    Patient reported outcomes by type D personality groups.White columns show percentage of respondents who indicated that their thyroid medication controlled symptoms of hypothyroidism poorly (sum of "tend to disagree" and "strongly disagree" in response to the statement "my hypothyroidism medication controls my symptoms well"); grey columns show percentage of respondents who indicated that they were dissatisfied with their care and treatment for hypothyroidism (sum of "slightly dissatisfied" and "very dissatisfied" in response to question "how satisfied are you with the overall care and treatment you have received for your hypothyroidism?");black columns show percentage of respondents who indicated that hypothyroidism had a negative impact on everyday living (sum of "tend to agree" and "strongly agree" in response to the statement "my hypothyroidism has affected everyday activities that people my age usually do eg, exercise, household chores, etc.").The largest differences noted were between groups NA+SI+ (type D personality) and NA−SI− (reference).The percentages were calculated as follows: white columns "tend to disagree" + "strongly disagree"/"tend to disagree" + "strongly disagree" + "tend to agree" + "strongly agree" + "neither agree nor disagree"; grey columns: "slightly dissatisfied" + "very dissatisfied"/"slightly dissatisfied" + "very dissatisfied" + "slightly satisfied" + "very satisfied" + "neither agree nor disagree"; black columns "tend to agree" + "strongly agree"/"tend to agree" + "strongly agree" + "tend to disagree" + "strongly disagree" + "nether agree nor disagree".Missing data and "don't know" (for grey columns), "don't know/can't recall," and "this does not apply to me" (for black columns) and "missing data" (for all columns) were excluded from the calculations.
needs to be considered in the context of the fact that dissatisfied patients are more likely to respond to surveys (26,48,49).
We used univariate analyses.The significant findings reported here may not persist after multivariate analyses.However, our aims were exploratory, and our findings are of interest themselves, providing useful information not previously available.In mitigation of these factors, the sample size (which was well beyond that specified by power calculations), cognitive testing, piloting, and inclusion of a patient representative in the research team were strengths.

Conclusions
This study demonstrates a high prevalence of type D personality among people with hypothyroidism who responded to the survey.There were associations between type D personality, several respondent characteristics, somatization, and negative patient-reported outcomes regarding hypothyroidism in an adequately powered sample.Type D personality may be an important determinant of dissatisfaction with treatment and care among some people with hypothyroidism.Our findings require independent confirmation from studies that focus on type D personality.Close collaboration between the disciplines of thyroidology and psychology is likely to be key in progressing our understanding in this area.

Figure 2 .
Figure 2.Patient reported outcomes by type D personality groups.White columns show percentage of respondents who indicated that their thyroid medication controlled symptoms of hypothyroidism poorly (sum of "tend to disagree" and "strongly disagree" in response to the statement "my hypothyroidism medication controls my symptoms well"); grey columns show percentage of respondents who indicated that they were dissatisfied with their care and treatment for hypothyroidism (sum of "slightly dissatisfied" and "very dissatisfied" in response to question "how satisfied are you with the overall care and treatment you have received for your hypothyroidism?");black columns show percentage of respondents who indicated that hypothyroidism had a negative impact on everyday living (sum of "tend to agree" and "strongly agree" in response to the statement "my hypothyroidism has affected everyday activities that people my age usually do eg, exercise, household chores, etc.").The largest differences noted were between groups NA+SI+ (type D personality) and NA−SI− (reference).The percentages were calculated as follows: white columns "tend to disagree" + "strongly disagree"/"tend to disagree" + "strongly disagree" + "tend to agree" + "strongly agree" + "neither agree nor disagree"; grey columns: "slightly dissatisfied" + "very dissatisfied"/"slightly dissatisfied" + "very dissatisfied" + "slightly satisfied" + "very satisfied" + "neither agree nor disagree"; black columns "tend to agree" + "strongly agree"/"tend to agree" + "strongly agree" + "tend to disagree" + "strongly disagree" + "nether agree nor disagree".Missing data and "don't know" (for grey columns), "don't know/can't recall," and "this does not apply to me" (for black columns) and "missing data" (for all columns) were excluded from the calculations.

Table 2 . Chi-square analysis for observed variables against the 4 type D personality groups as defined by the subscales NA and SI
Flow chart showing included and excluded data.The Journal of Clinical Endocrinology & Metabolism, 2024, Vol.00, No. 0 7 Downloaded from https://academic.oup.com/jcem/advance-article/doi/10.1210/clinem/dgae140/7640726 by guest on 13 April 2024

Table 3 . Associations between type D personality and patient-reported outcomes
The adjusted threshold by Bonferroni method was for the P-level of .002294.The table includes all data (both English and non-English translations of Type D Scale-14).The column labeled "adjusted significance" is derived after Bonferroni correction.
Abbreviations: E, expected values; NA, negative affectivity; O, observed values; SI, social inhibition.Chi-square analysis for hypothyroid patient reported outcomes against the 4 type D personality groups as defined by the subscales NA and SI (NA+ SI−, NA− SI+, NA+ SI+ (Type D personality) and NA− SI− (reference), where + denotes a score ≥10 and − a score of <10.