Abstract
Objective
Patients considering surgery face many uncertainties and concerns. This investigation sought to develop an objective assessment tool for characterizing the areas of greatest concern among those considering thyroidectomy.
Study Design
Prospective cohort design.
Setting
Tertiary-care otolaryngology practice.
Subjects
Participants included individuals presenting with a thyroid nodule found to be of intermediate risk (15%-20%) of cancer based on fine-needle aspiration (FNA) biopsy and amendable to hemithyroidectomy.
Methods
As part of validating a clinical measure concerning perioperative concerns, patients presenting with thyroid nodules amendable to hemithyroidectomy were voluntarily recruited. Thirty individuals (6 men, 24 women) completed a novel 18-item questionnaire during their initial clinical visit and again 3 days later. Outcome measures included descriptive statistics and test-retest reliability.
Results
Individual patients have both general and specific preoperative concerns. A wide range of responses was obtained across the entire spectrum of the survey. The questionnaire demonstrated fair-to-excellent test-retest reliability, with correlation values from 0.467 to 0.954. Between-rater reliability was consistent with intraclass correlation coefficient values of 0.52 to 0.86. No relationship between question order and patient response was identified. Women were found to be more concerned about scar appearance (P = .028), and men were conversely more concerned about resuming daily activities (P = .026).
Conclusions
Patients considering thyroidectomy have concerns that remain stable in the early preoperative period. This is the first study to evaluate preoperative patient concerns and preliminarily establishes the Western Surgical Concern Inventory–Thyroid as a means of ensuring adequate patient counseling and a method of evaluating preoperative patient education.
More than 36 million Americans undergo elective surgery each year. An estimated 22,000 of these procedures are hemi- or total thyroidectomies. 1 Although necessary and elective surgical procedures have been performed since 3500 bc, no systematic evaluations of patient preoperative concerns have been previously undertaken.
As a patient considers surgery, surgeons strive to educate them on the risks, benefits, and biopsychosocial issues associated with the proposed surgical procedure. The goal of such discussions seeks to develop a strong therapeutic alliance, ensuring that patient expectations are in line with therapeutic realities and that all significant issues are adequately addressed. This alliance has been shown to yield superior physiological and psychological patient outcomes as well as to circumvent the leading source of malpractice litigation: miscommunication.2-4
Although each surgeon endeavors to effectively communicate with his or her patient, they face a formidable challenge in appreciating the vast array of potential concerns that face each patient as an individual. Surprisingly, no previous investigations have attempted to identify and describe the concerns experienced by patients during the early preoperative period.
Using thyroidectomy as a model, the goal of this investigation was to generate a simple yet comprehensive clinical instrument that could allow patients to more accurately convey their individual preoperative concerns. The proposed questionnaire would provide for improved physician-patient communication, more informed and holistic patient decision making, and ultimately improved patient outcomes.
Materials and Methods
Ethics approval was obtained via the Western University Health Sciences Research Ethics Board (review 13772E).
Surgical Model Selection
A hemithyroidectomy procedure has several features that make it an effective surgical model to improve the understanding of preoperative patient concerns. Patients considering this procedure frequently present with a fine-needle aspiration (FNA) biopsy indicating a lesion with an intermediate risk of malignancy. Through quality control assessments by the Department of Pathology at Western University, a FNA biopsy indicating a follicular neoplasm, Hurthle cell neoplasm, or indeterminate biopsy has been correlated with a 15% to 20% risk of malignancy on final pathology, coinciding with the recently published Bethesda System for reporting thyroid cytopathology. 5 Thus, individuals identified as having an intermediate risk nodular lesion face uncertainty as to whether their nodule is cancerous. In addition to the general risks and concerns associated with undergoing any surgical procedure, these individuals face the possibility that they will be diagnosed with thyroid cancer. Based on this unique patient circumstance and the presence of well-established surgical risks, the hemithyroidectomy procedure was selected as an ideal surgical model for this investigation.
Questionnaire Design
The development of a valid questionnaire required the implementation of a rigorous research protocol. A random set of individuals with indeterminate FNA biopsy results considering hemithyroidectomy were voluntarily interviewed in a systematic fashion to determine their concerns regarding their proposed surgery. These interviews were conducted by the principal author (M.G.B.) and took place immediately after a thorough discussion of the risks and benefits of the surgery by a single experienced otolaryngologist (J.H.F.). Each interview consisted of a series of 2 questions: “What are your concerns?” and “What is your greatest concern?” Responses were recorded by the interviewer without further discussion relative to the procedure to avoid bias. Following these interviews, all recorded responses were collated and reviewed by the research team. Twenty randomly selected patients were interviewed over a 2-week period. Based on these responses, a broad range of questions were then carefully drafted through group consensus to reflect the issues identified to be of greatest concern by the patients.
An equal-appearing-interval response-style questionnaire was subsequently generated. This questionnaire style was selected to obtain the greatest amount of information while remaining time-efficient and comprehensive and of low burden to the patient. In addition, the questionnaire was designed to represent a consistent structure in that the response categories were maintained throughout all questions: the questionnaire was constructed with all potential responses ranging from 1 representing “no concerns” to a 9, which indicated the presence of “major concerns.”
Once the preliminary questionnaire was developed, it was then subjected to pilot testing to confirm content validity. This process involved proctoring questionnaires to 20 randomly selected patients considering hemithyroidectomy. All were instructed to independently complete each question included in the survey. After the completion of each survey question, participants were then asked to explain their interpretation of what each question was asking and why they had chosen their particular rated response. Participants were also asked to comment on any potential concerns that were not adequately addressed by the questionnaire. As no existing questionnaire that specifically addressed perioperative concerns nor the concerns of patients considering hemi- or total thyroidectomy had previously been developed, we were unable to reference the set of questions we generated to standardized norms or existing criteria.
Based on the data generated through pilot testing and the determination of content validity, the questionnaire was further revised, and a final questionnaire was formulated (appendix, available at otojournal.org). This new questionnaire contained 2 distinct groups of questions: those specific to thyroid surgery and those addressing general concerns about undergoing an operative procedure. To determine if question order influenced patient response, 3 versions of the questionnaire were generated. Version A ordered specific surgical concerns prior to general concerns, version B ordered general surgical concerns prior to specific surgical concerns, and version C inquired into general and specific surgical concerns in a randomized order.
Sample Size Calculation
As part of this initial validation, we calculated a minimum sample size using Horatio Version 3.0. 6 Based on our calculations, a minimum sample size (N) of 21 individuals was determined to be sufficient to detect the hypothesized effect (r2 = 0.30) of a single predictor variable 81.2% of the time using a .05 α level. Thus, at least 21 patients would be required to generate adequate statistical power to meet significance.
Participant Recruitment, Patient Population, and Research Protocol
The patient population included any individual presenting to a tertiary-care otolaryngology practice with a thyroid nodule found to be of intermediate risk (15%-20%) of cancer based on FNA biopsy. 5 Those individuals with biopsies consistent with a thyroid malignancy were excluded from study participation. Following a thorough evaluation of the patient and review of thyroid ultrasonography, those patients deemed to be appropriate surgical candidates were advised of the potential risks and benefits of total versus hemithyroidectomy. Those individuals choosing to undergo hemithyroidectomy were then voluntarily recruited for study participation by a clinic nurse familiar with the details of the study. English-fluent and -literate individuals wishing to participate were asked to complete a randomly selected version (A, B, or C) of the study questionnaire. Because this sample population was limited to patients undergoing hemithyroidectomy, patients were instructed to ignore 1 survey question, which addressed concern about calcium problems—included in the survey for future administration to patients undergoing total thyroidectomy (appendix, question 5). Participants completed the questionnaire independently, with no input from any member of the medical team. The questionnaire was then collected by the clinic nurse and stored in a secure location during the data collection phase of the study. Participants were provided with a second questionnaire (identical to that completed initially (ie, version A, B, or C) to be completed at home 3 days following their initial clinic appointment and mailed back to the research team via a preaddressed, stamped return envelope.
Once all data had been collected and collated, data entry by the principal author (M.G.B.) took place. This was then submitted for independent statistical analysis.
Main outcome measures focused on individual patient variability, demographic variability (relationship between age and gender on overall degree of concern), and test-retest reliability. As a secondary outcome measure, the influence of question order on response rating was also assessed.
Statistical Analysis
All statistical analyses were performed using SPSS version 18 (SPSS, Inc, an IBM Company, Chicago, Illinois). This included comparisons for participant age and gender specific to questionnaire version, as well as all reliability analyses. Descriptive statistics were obtained for the entire data set. To evaluate the level of reliability, Cronbach α measures were determined for each list randomization across the 3 days of assessment. This followed an initial evaluation of correlation facilitated by the development of a Pearson correlation matrix. Furthermore, the range of intraclass correlation coefficients (ICCs) for average measures were calculated for each list randomization to assess the level of response consistency.
Results
Thirty participants were included in this investigation: 6 men and 24 women with average age of 54 ± 2 years (mean ± SD). Biopsy results identified 11 follicular lesions, 6 Hurthle cell lesions, 4 intermediate-risk lesions, and 9 benign nodules. Mean concerns appear in Figure 1 with summary data appearing in Table 1 . To determine if age and gender influenced overall patient concern, an overall summary score—termed the concern index—was generated. This score was a sum of all rated responses with a possible range of 17 to 153 (minimum score of 1 for 17 questions to a maximum score of 9 for 17 questions). The mean (min; max) overall concern scores were 53.93 (47.61; 60.25, 95% confidence interval [CI]). This score did not vary based on the version of the survey, nor did it vary based on the age or gender of the participant.
Responses for Each of the Questionnaire Domains a
Range of responses for each of the 17 questionnaire domains inquiring into individual patient concerns. Reported are the items of concern and their respective mean, minimum confidence interval, maximum confidence interval (mean [min CI; max CI]), and overall minimum (Min) and maximum response (Max) values. Correlations from initial test administration to subsequent test administration (day 1 to 3) are reported as α values and their respective CIs (α [min CI; max CI]). Correlations from initial test administration to subsequent test administration (day 1 to 3) are reported as Pearson r values (Pearson r), with statistically significant values identified with an asterisk* (P < .01, 2-tailed).
Items using the full spectrum of the scale.
Mean concern ratings (±standard error) across the 17 questionnaire domains for patients considering hemithyroidectomy.
To further address the influence of age and gender on participant responses, individual analysis of each survey item took place. No relationship was found between age and each survey item. All survey items were then stratified across gender, and individual 2-tailed t tests were performed for each item. Only the items referencing the appearance of a scar and returning to daily activities demonstrated a statistically significant difference (P = .028 and P = .026, respectively), with women being more concerned than men about their scar appearance, with scores of 3.17 ± 0.38 and 1.83 ± 0.40, respectively, and men being more concerned about resuming daily activities than women, with scores of 5.33 ± 0.80 and 2.88 ± 0.33, respectively (mean ± standard error).
Qualitative evaluation of the survey data revealed a broad range of responses and use of the full range of the scaled score (min-max; Table 1 ). Administering the questionnaire initially and 3 days following the participants’ initial appointment allowed us to assess the stability of questionnaire responses during the early preoperative period ( Table 1 ). Test-retest reliability for each of the 3 versions of the study were excellent, with α values of .894 (.779; .964) for version A, .910 (.795; .976) for version B, and 0.940 (.862; .964) for version C. No relationships were found between questionnaire version (question order) and participant responses. Based on this finding, we then determined the intrarater reliability for the measures. First, a correlation (Pearson r) was performed independently on each of the 17 questions addressed in the measure. Correlations across the group of questions ranged from 0.467 to 0.954 ( Table 1 ). All questions demonstrated significant correlation (P < .01), except for the question referencing surgical wait times (r = 0.467). When assessing correlations for all questions, 10 of them exhibited r values of ≥0.70, with 8 of those exceeding 0.817. Next, ICCs were found to range from 0.863 to 0.956, which represents high levels of consistency for the responses obtained. Between-rater reliability was then calculated using an interitem assessment of responses provided by the participants. Results of this statistical assessment indicated very good consistency of responses, with ICC consistency ranging from 0.52 to 0.86. Finally, Cronbach α values were obtained for the 17 questions and ranged from .695 to .975 ( Table 1 ). While some variability existed across the correlations in each of the 17 questions evaluated for test-retest reliability, relatively tight correlation CIs were noted for most questions. The narrowest range of lower and upper bounds of the correlation CIs was noted for the question relative to being judged (0.948-0.988), with the widest range of correlations observed for the question relative to voice changes (0.370-0.861). These data suggest very good test-retest consistency across the 30 participants who were included in the study.
Discussion
This survey was designed to provide a preliminary approach to the evaluation and quantification of the preoperative concerns of patients undergoing hemithyroidectomy. The results have the potential to exert a significant effect on physician-patient communication, ensuring that the most significant concerns of patients are being addressed. This questionnaire provides a means of determining individual patients’ concerns, which may be discordant from those commonly reviewed by their surgeons. At the very least, these data allow for improved understanding of issues that are important to patients who are considering hemithyroidectomy. Physician knowledge of individual patient concerns allows for a more complete and pertinent risk-benefit analysis, which allows for more informed decision making by the patient. Finally, this questionnaire allows for future analyses of thyroidectomy patients who may make decisions (ie, hemi- vs total thyroidectomy for intermediate FNA biopsy results) based on real or assumed concerns. Because this is the first instrument of its type that focuses on patient concerns specific to thyroid surgery, it provides a foundation from which further evaluation and documentation can occur.
Many efforts were taken to ensure construct and content validity. These included the initial structured interviews, proctored questionnaires, and expert opinion, which allowed for the generation of a questionnaire that expressed what we believe was a comprehensive spectrum of concerns affecting most individuals undergoing hemithyroidectomy. Further efforts to improve validity included a single surgeon reviewing surgical risks to standardize preoperative information and variation in question order to negate the potential influence of preresponse bias.
Patients undergoing hemithyroidectomy face a broad array of concerns. While we fully anticipated that different patterns of response would be demonstrated by each individual participant, our evaluation of the data demonstrated remarkable consistency in what was judged as more or less important. In addition, statistical analyses of the data revealed that responses were highly consistent within each participant over all the assessments obtained; that is, those participants identified as having a high level of overall concern ranked individual items higher than those individuals having a lower overall level of concern. While some questions revealed very strong correlations during the test-retest phase of the validation, others were less consistent. This may suggest that there is a critical core set of questions that remain of consistent concern. In contrast, other questions may be initially viewed in a manner that upon reflection of even several days may change (ie, the practical implication of surgical waiting times may become more of a concern). This finding provides further support for the notion that a somewhat common series of concerns are ranked in a consistent manner, although they may be judged with relative levels of perceived concern. Individual patient concerns appeared to remain stable throughout the early preoperative period. Intrarater reliability was very high, with α values ranging from .894 to .940 across the 3 questionnaires. Overall, the results were consistent and reliable across the early preoperative period.
Overall, high levels of concern did not appear to correlate with either age or gender. Individual items also did not appear to correlate with age, but the items relative to “appearance of a scar” and “resuming daily activities” did demonstrate statistical significance, with women being more concerned about scarring and men being more concerned about resuming daily activities. Although these results appear to relate to social norms, the influence of a small sample size (6 men, 24 women) must be considered. Furthermore, future evaluations using this tool should include larger samples and consider identifying those individuals placing high value on appearance and remaining active to further extrapolate on the identified differences and avoid gender bias.
Despite the stringent methodology employed, there are limitations that require acknowledgment. Intrarater reliability was based on retesting 3 days following the initial session. Although the results suggest excellent test-retest reliability, this may potentially be a function of participants simply recalling their response during their initial appointment and not accurately indicating their present degree of concern. Further evaluation of intrarater reliability may benefit from testing at wider time intervals. This may, however, introduce the influence of recall decay with respect to the discussion of risks and benefits.7,8 Furthermore, specific concerns and the degree to which they are concerning may also change with time, and therefore, intrarater reliability may be less meaningful over time. A subsequent investigation will provide greater insight into this particular question and may help shed light on the influences of other sources of information (ie, the Internet) on patient preoperative concerns.
A further limitation involves survey content. The concerns addressed in the survey were based on participants in 20 initial structured interviews, expert opinion, and subsequent pilot testing at a tertiary-care otolaryngology clinic. It is possible that there are issues that are of concern to other patients that were not addressed by this limited group. To overcome this limitation, an area was included on the questionnaire in which participants could write in additional comments or areas of concern. Although none of the 30 individuals involved in this initial validation phase opted to make use of this, the final version of the survey continues to include this opportunity for further comment. Furthermore, given that there were no previous questionnaires investigating perioperative concern nor the concerns of patients considering hemi- or total thyroidectomy, we were unable to verify that our set of questions conform to standardized norms or an accepted set of reference criteria.
Finally, the questionnaire requires a degree of patient literacy and insight and thus excludes a particular subset of patients. Future investigation will be required to provide greater insight into the potential effects of education on patient concerns. Nonetheless, this survey may have the capacity to serve as a springboard to initiate physician-patient dialogue and permit the concerns of these patients to be elicited directly.
Through rigorous methodology, this investigation has provided preliminary validation for the Western Surgical Concern Inventory–Thyroid (WSCI-T). This tool provides a means of improving preoperative physician-patient communication, discovering individual patient concerns, ensuring the provision of information that is pertinent and complete for the individual patient, and allowing for more informed patient decision making. Thus, we believe the WSCI-T provides a valuable means of investigating patient decision making and bridging patient expectations with therapeutic goals, thereby ensuring a strong therapeutic alliance between physicians and patients.
Author Contributions
Disclosures
Footnotes
Acknowledgements
We would like to express our gratitude to Karen Canton, RN, and all the nursing staff at the London Health Sciences Centre–Victoria Hospital, Otolaryngology–Head and Neck Surgery clinic for their efforts in administering the survey. We further wish to acknowledge Wendy Reason, RN, for her insights during the development of the survey.
No sponsorships or competing interests have been disclosed for this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
