Secondary Logo

Journal Logo

CLINICAL RESEARCH

Minimum Clinically Important Differences of the Hospital for Special Surgery Dysphagia and Dysphonia Inventory and Other Dysphagia Measurements in Patients Undergoing ACDF

Okano, Ichiro MD; Ortiz Miller, Courtney BA; Salzmann, Stephan N. MD; Hoshino, Yushi MD, PhD; Shue, Jennifer MS; Sama, Andrew A. MD; Cammisa, Frank P. MD; Girardi, Federico P. MD; Hughes, Alexander P. MD

Author Information
Clinical Orthopaedics and Related Research: October 2020 - Volume 478 - Issue 10 - p 2309-2320
doi: 10.1097/CORR.0000000000001236

Abstract

Introduction

Postoperative dysphagia is one of the most common complications of anterior cervical surgery. Since most dysphagia resolves over time after surgery, the observed frequency of dysphagia varies based on the timing of assessment. The reported frequencies are higher when the assessment is conducted earlier. One study showed that more than 80% of patients reported dysphagia on the first postoperative day [27]. At 1 month after surgery, the observed dysphagia rates were as high as 50% and dropped to 20% at 6 months follow-up [33].

For investigations of dysphagia after anterior cervical discectomy and fusion (ACDF), the four-grade Bazaz Dysphagia Score [2], which is based on the interpretation of dysphagia by clinicians, has been the most widely used [26]. The Bazaz score classifies postoperative dysphagia into four grades based on patient symptoms when swallowing liquids or solid foods, from none, or no symptoms while swallowing liquid or solid food, to severe, which is frequently symptomatic. The classification is quick, easy to apply, and requires no additional work for patients because the provider completes grading. However, the Bazaz score only assesses severity by the frequency of swallowing difficulty and does not include a qualitative assessment of symptom severity. Thus, the score may not accurately reflect actual symptom burden. Additionally, it is also well-known that the perspectives of patients and surgeons regarding clinical outcomes do not always match [13, 38]. To help address these issues, patient-reported outcome measures (PROMs) have been used to assess dysphagia after ACDF [12, 16, 17, 21, 29, 32, 39, 42, 43], including the SWAL-QOL [30] and the MD Anderson Dysphagia Inventory (MDADI) that assess dysphagia related to chronic illnesses [4]. Dysphagia after ACDF, however, is usually less severe and more self-limited than dysphagia associated with chronic illnesses. Therefore, the SWAL-QOL and MDADI may underestimate the frequency and severity of dysphagia in patients who have undergone ACDF because the instruments include questions that do not apply to this specific patient population. The Hospital for Special Surgery Dysphagia and Dysphonia Inventory (HSS-DDI) is a patient-derived, validated, and condition-specific PROM that consists of 31 patient-reported questions about swallowing and speaking functions, with a scale ranging from 0 (worst) to 100 (normal) [18]. The HSS-DDI was created to specifically to assess dysphagia and dysphonia after anterior cervical spine surgery [18].

Statistical differences in PROM scores are not always clinically meaningful, either to patients or providers [24, 25] because the likelihood of detecting statistical differences increases with a study’s sample size and studies with a large sample size can detect small changes that may be unimportant or even imperceptible to patients and providers. To address this issue, the minimum clinically important difference (MCID) was conceptualized more than 30 years ago [15]. The MCID is the smallest change in a PROM that is considered meaningful for a patient and/or provider. The concept of the MCID has become common in various medical fields, including general orthopaedic and spine surgery [6, 24, 28]. Because the HSS-DDI was recently introduced, the MCID of the HSS-DDI has not, to our knowledge, been established. Furthermore, although the SWAL-QOL and MDADI have been used in previous dysphagia studies, little research has addressed the MCIDs of these instruments for patients undergoing ACDF.

We therefore sought to determine (1) the minimum detectable changes (MDC) of the HSS-DDI, SWAL-QOL, and MDADI using a distribution-based approach and (2) the MCID of the HSS-DDI, SWAL-QOL, and MDADI, using an anchor-based approach.

Patients and Methods

This study was approved by our institutional review board. In this study, we used the database that was originally established for the HSS-DDI validation study and was longitudinally maintained after the study using the same inclusion criteria.

Enrollment and Study Attrition

We checked the eligibility for study enrollment for 323 patients who underwent elective ACDF in our clinic surgical booking record from 2015 to 2018. We excluded patients who were younger than 18 years (0% [0 of 323]), did not speak English (4% [12 of 323]), had a medical history of neurologic conditions (1% [3 of 323]) or head and neck cancers (< 1% [1 of 323]) that may affect swallowing and/or speech function, had been participating in another study that may impact study outcomes (2% [5 of 323]), and did not consent to participate in the study (9% [28 of 323]). Overall, 85% (274 of 323) of patients agreed to participate, and we obtained written informed consent from all patients at the time of enrollment. Two percent (6 of 323) of patients were excluded due to missing data in the Week 4 HSS-DDI survey or other demographic factors, and 83% (268 of 323) patients had complete Week 4 surveys.

Study Outcomes

We set six study outcomes: distribution-based minimum detectable changes (MDC) for the (1) HSS-DDI, (2) SWAL-QOL, and (3) MDADI in addition to anchor-based MCIDs for the (4) HSS-DDI, (5) SWAL-QOL, and (6) MDADI. Two approaches are generally used to define MCID: (1) a distribution-based approach, in which an MCID value is defined as the smallest difference above the measurement error, derived statistically from distributional characteristics of the sample; and (2) an anchor-based approach, in which an MCID value is determined based on the minimal symptom change over time that is considered clinically important by patients and clinicians. The anchor-based approach uses an independent external criterion called anchor questions. The anchor questions must reflect the symptom change and be easier to understand than the PROM itself for patients who have no medical knowledge [45]. We used a common type of anchor question: a global rating of symptomatic change, graded as much better, little better, no change, little worse, or much worse. Values derived from one approach are not always equivalent to those from another approach [41] and the combined use of both approaches has been recommended [7, 37]. In previous studies, multiple terms and definitions of MCID were used [14, 22, 31, 44]. To avoid confusion between the two approaches, we defined values from the distribution-based method as the minimum detectable change (MDC) and values from the anchor-based method as the minimal clinically important difference (MCID).

Study Participants and Survey Administration

In this cohort of 268 patients, the mean ± SD age was 55 ± 12 years. Among the cohort, 60% (162 of 268) of the patients were males (Table 1). Surveys were administered in person with the exception of re-testing of the HSS-DDI survey, which was implemented via a telephone interview. The first 18% (49 of 268) of patients were included in the test-retest group. The test-retest group sample size was determined based on previous evidence on the minimum required sample size for reliability studies, which was 35 to 50 patients [11]. In the test-retest group, patients had a repeat assessment 3 to 11 days (mean 5) after Week 4. We assumed that there would be no substantial change during this period and used the test-retest data to assess measurement error. Of the 18% (49 of 268) of patients in this test-retest group, 90% (44 of 49 or 16% (44 of 268) of the total cohort) had no missing questionnaire responses and were included in the final analysis. Among the 268 patients who had a Week 4 HSS-DDI survey, 86% (230 of 268) had both a Week 8 HSS-DDI survey and an anchor question without missing answers. One patient had a speech anchor assessment only and was included in the assessment of the HSS-DDI speech domain.

Table 1. - Patient demographics in each PROM group
Factor HSS-DDI overall (n = 268) HSS-DDI test-retest (n = 44) SWAL-QOL (n = 169) MDADI (n = 168)
Age (years), mean ± SD 55 ± 12 57 ± 12 55 ± 12 55 ± 12
Sex (%)
 Male 60 (162) 77 (34) 54 (92) 55 (92)
 Female 40 (106) 23 (10) 46 (77) 45 (76)
BMI (kg/m2), mean ± SD 28 ± 5 30 ± 5 28 ± 5 28 ± 5
Race (%)
 White 93 (250) 93 (41) 95 (159) 94 (158)
 Black 1 (3) 0 (0) 1 (1) 1 (1)
 Asian 2 (6) 2 (1) 2 (3) 2 (3)
 Other 3 (9) 5 (2) 4 (6) 4 (6)
Current smoking (%) 6 (15) 2 (1) 7 (11) 7 (11)
Prior cervical surgery (%) 9 (24) 9 (4) 11 (19) 11 (19)
Operating level (%)
 C3/4 25 (66) 23 (10) 24 (40) 24 (40)
 C4/5 52 (139) 57 (25) 50 (84) 50 (84)
 C5/6 80 (214) 77 (34) 77 (130) 77 (129)
 C6/7 70 (188) 68 (30) 70 (119) 71 (119)
 C7/T1 5 (13) 5 (2) 5 (9) 5 (9)
Number of fused levels
 1 20 (54) 25 (11) 24 (40) 23 (39)
 2 40 (108) 30 (13) 39 (65) 39 (65)
 3 or more 40 (106) 45 (20) 38 (64) 38 (64)
Operation time (minutes), median (range) 137 (46-294) 161 (79-294) 142 (46-294) 143 (46-294)

In addition to the HSS-DDI, two other patient-reported dysphagia instruments, the SWAL-QOL and MDADI, were administered to the first 63% of patients (169 of 268 for SWAL-QOL and 168 of 268 for MDADI). The administration of these two PROMs were for an external validation study of the HSS-DDI and continued until June 2017. Fifty-six percent (150 of 268 for SWAL-QOL and 151 of 268 for MDADI) completed the Week 8 survey and anchor questions without missing data (Fig. 1).

Fig. 1
Fig. 1:
The Hospital for Special Surgery Dysphagia and Dysphonia Inventory (HSS-DDI) is shown (reprinted with permission from the Hospital for Special Surgery, New York, NY, USA).

Postoperative Dysphagia Assessments

The HSS-DDI scores calculated from the responses at 4 weeks and 8 weeks after surgery were used to estimate the MCIDs. The HSS-DDI consists of 31 questions about dysphagia and dysphonia experienced in the past 2 weeks and each answer is weighted as 0 (all the time) to 4 (none of the time) based on the frequency of symptoms (Fig. 2). The scores are summed and converted to an actual number on a scale of 0 (worst) to100 (no symptoms) (Fig. 3). The HSS-DDI was validated for test-retest reliability and responsiveness longitudinally in a previous study [18]. The internal validity was assessed by comparing the HSS-DDI score and surgical invasiveness index [34], which is based on numbers of decompressed, fused, and instrumented levels. A worse or lower HSS-DDI score was associated with an increased invasiveness index. Additionally, the score was externally validated by evaluating associations with other previously validated surveys such as the SWAL-QOL and MDADI. Along with the total HSS-DDI score, we calculated the individual dysphagia domain scores separately for comparison with other dysphagia PROM assessment scores.

Fig. 2
Fig. 2:
Scoring of the Hospital for Special Surgery Dysphagia and Dysphonia Inventory (HSS-DDI) is shown (reprinted with permission from the Hospital for Special Surgery, New York, NY, USA).
Fig. 3
Fig. 3:
A patient flow diagram is shown.

The original version of the SWAL-QOL consists of a 44-item survey that evaluates quality of life in 10 psychosocial areas. It was originally created for patients with neurologic or mechanical oropharyngeal dysphagia [30]. In this study, we used the focused SWAL-QOL score [17, 39], which was introduced by Siska et al. [39] and consists of 14 selected items from the original 44-item SWAL-QOL questionnaire. The scale range of this focused SWAL-QOL is 0 (worst) – 100 (normal). This focused version has been more commonly used for dysphagia studies after ACDF due to improved administration than the original 44-item version that was lengthy, time-consuming, and does not have an established 100-scale scoring system. The MDADI, which consists of a 20-item survey, was originally created for patients with head and neck cancer [4]. The MDADI has been also used in ACDF studies [32]. The total MDADI (scale range: 20 [worst] to 100 [normal]) scoring system, as described in the original MDADI article, was used.

Distribution-based MDC Analysis

For the distribution-based MDC calculation, the values of half a SD (0.5 SD), originally described as the medium effect size by Cohen [5], for all three measurements (HSS-DDI, SWAL-QOL, and MDADI) and the subdomains of the HSS-DDI (dysphagia and dysphonia) at Week 4 (baseline) were calculated. For the HSS-DDI, because the database included a test-retest group, we calculated the interclass correlation coefficient of agreement of the HSS-DDI and its subdomains in the test-retest group as a parameter of reliability, using a two-way random effects model and the standard error of measurement (SEM), as described by de Vet et al. [9,10].

Anchor-based MCID Analysis

For the anchor-based MCID calculation, the anchor scores were implemented at Week 8. For the total HSS-DDI score, we used the nine-grade combined score (range -4 to + 4) of both dysphagia and dysphonia anchors, which we calculated by summing both five-grade global self-assessment scores (much better [+ 2], a little better [+ 1], no change [0], a little worse [-1], and much worse [-2]) for the swallowing and speech subdomains. For example, if a patient answered that their dysphagia at Week 8 was a little better (+ 1) than at Week 4, and their dysphonia at Week 8 was much better (+ 2), then the summed nine-grade score would be 3 (1 + 2 = 3).

To assess the HSS-DDI subdomains, we used either the swallowing or speech five-grade anchor score [40]. We also used the dysphagia subdomain anchor score for the SWAL-QOL and MDADI, since these PROMs specifically assess swallowing problems. We defined the minimum mean change values as the mean PROM values of those who had a self-assessment score improvement of + 1 in the subdomain and combined scores. The correlations between anchor scores and PROMs were calculated to assess the responsiveness of PROM changes to the anchor question. Additionally, another anchor-based method for the calculation of MCIDs is the receiver operation characteristic (ROC) curve based method that was also used (see Appendix, Supplemental Digital Content 1, https://links.lww.com/CORR/A329; Fig.1A-C, Supplemental Digital Content 2, https://links.lww.com/CORR/A330; Fig.2, Supplemental Digital Content 3, https://links.lww.com/CORR/A331; and Table 1; Supplemental Digital Content 4, https://links.lww.com/CORR/A332)

Statistical Analysis

Patients with one or more missing answers in questionnaires, PROMs, and anchor questions were excluded from the analyses. Analyses of correlations between anchor scores and changes of PROMs were performed using Pearson’s correlation and Spearman’s rank correlation coefficients. All statistical analyses were conducted with R (Ver. 3.5.2; R Foundation for Statistical Computing, Vienna, Austria; “lme4” package for intraclass correlation coefficients and SEM calculations). The statistical significance was set at p < 0.05.

Results

The MDCs of the HSS-DDI, SWAL-QOL, and MDADI

The MDCs of the HSS-DDI, SWAL-QOL, MDADI were 11 of 100, 9 of 100, and 8 of 80 points, respectively, using the 0.5 SD method. Using the data of HSS-DDI test-retest group, the intraclass correlation coefficient was calculated as 0.86 and the SEM-based MDC of HSS-DDI total score was 9 out of 100 points (Table 2).

Table 2. - Results of the distribution-based MDC analysis
Measurement Domain Scale range Number Mean ± SD at Week 4 Number of patients with minimum score at Week 4 Number of patients with maximum score at Week 4 0.5 SD at baseline Intraclass correlation coefficient SEM
HSS-DDI Total 0-100 268 75 ± 22 < 1 (1) 9 (24) 11 0.86 9
Dysphagia 0-100 268 69 ± 24 1 (2) 9 (24) 12 0.84 10
Dysphonia 0-100 268 85 ± 22 < 1 (1) 38 (101) 11 0.88 8
SWAL-QOL 0-100 169 80 ± 18 0 (0) 12 (21) 9
MDADI 20-100 168 84 ± 17 0 (0) 23 (38) 8
SEM = standard error of measurement.

The MCIDs of the HSS-DDI, SWAL-QOL, and MDADI

Since the values calculated with the minimum mean change method were consistently higher than those of ROC based method (see Appendix, Supplemental Digital Content 1, https://links.lww.com/CORR/A329), we reported only the minimum mean change method values as the MCIDs.

The total score MCID of the HSS-DDI using the anchor-based approach was 10 out of 100 points (Table 3) and the MCIDs of the SWAL-QOL and MDADI were 8 out of 100 and 6 out of 80 points, respectively (Table 4).

Table 3. - Results of the anchor-based HSS-DDI MCID analysis with the nine-grade combined anchor
Nine-grade total anchor score Percent (n) Mean ± SD change
All 230 10 ± 14
+ 4 7 (15) 13 ± 12
+ 3 4 (10) 18 ± 12
+ 2 20 (45) 19 ± 14
+ 1a 21 (48) 10 ± 10
0 42 (96) 6 ± 11
-1 5 (11) 7 ± 20
-2 2 (4) 18 ± 9
-3 0 (0)
-4 < 1 (1) -15
aIndicates the minimum mean change defining the MCID.

Table 4. - Results of the anchor-based MCID analyses with the five-grade anchor
Measurement HSS-DDI
Dysphagia Dysphonia
Anchor type Swallowing  Score change Speech Score change
Five-grade anchor score Percent (number) Mean ± SD change Percent (number) Mean ± SD change
 Much better (+ 2) 15 (34) 22 ± 16 10 (23) 11 ± 17
 A little better (+ 1)a 34 (78) 16 ± 14 17 (40) 14 ± 18
 No change (0) 46 (105) 9 ± 13 70 (161) 4 ± 13
 A little worse (-1) 4 (10) 10 ± 19 2 (5) -2 ± 32
 Much worse (-2) 1 (3) 8 ± 34 1 (2) 5 ± 16
All 100 (230)b 13 ± 16 100 (231)b 6 ± 17
Measurement SWAL-QOL MDADI
Anchor type Swallowing Score change Swallowing Score change
Five-grade anchor score Percent (number) Mean ± SD change Percent (number) Mean ± SD change
 Much better (+ 2) 19 (28) 10 ± 14 19 (28) 6 ± 9
 A little better (+ 1)a 37 (55) 8 ± 11 37 (56) 6 ± 11
 No change (0) 41 (61) 3 ± 9 40 (61) 2 ± 7
 A little worse (-1) 3 (4) -1 ± 9 3 (4) 0 ± 7
 Much worse (-2) 1 (2) 21± 6 1 (2) 9 ± 4
All 100 (150)b 6 ± 11 100 (151)b 4 ± 9
aIndicates the minimum mean change defining the MCIDs.
bThe total number for the all categories is different by surveys or domains due to differences in patients with complete surveys and missing answers.

All MCIDs and MDCs were around 10% of their possible ranges (Table 5), except the HSS-DDI subdomain (dysphagia and dysphonia) scores, of which MCIDs were 16 out of 100 and 14 out of 100, respectively.

Table 5. - Summary of the MDC and MCID values of the HSS-DDI, SWAL-QOL, and MDADI
Measurement Domain Distribution-based MDC Anchor-based MCID
0.5 SD SEM Minimum mean change
HSS-DDI Total 11 9 10
Dysphagia 12 10 16
Dysphonia 11 8 14
SWAL-QOL 9 8
MDADI 8 6
HSS-DDI and SWAL-QOL scores ranged from 0-100; MDADI scores ranged from 20-100; SEM = standard error of measurement.

Discussion

Several previous studies have investigated dysphagia after ACDF including its prevalence, risk factors, effect of preventive interventions, and PROMs used[12, 39, 42]. However, few of these studies considered MCIDs, although most studies reported statistically significant results. The risk of maintaining an exclusive focus on statistical significance in clinical research is that patients (and providers) do not perceive p values, they perceive effect sizes. For an intervention as large as ACDF, to justify the risk to the patient or the cost to the healthcare system, the effect size must be large enough to care about. The MCID addresses this concern by defining the smallest increment of improvement that a patient is likely to perceive as clinically important. Our study bridges the gap between clinical and statistical viewpoints in this area. In our study, we found that MCDs of HSS-DDI, SWAL-QOL, and MDADI were 9-11, 9, and 8, respectively. The MCIDs of the HSS-DDI, SWAL-QOL, and MDADI were 10, 8, and 6, respectively. Improvements smaller than those should not be considered clinically important and certainly should not be used to justify a surgical intervention or invasive treatment.

This study has several inherent limitations. First, only a limited number patients from the study cohort was included in the test-retest group (16%, 44 of 268). Also, 63% of patients were included in the SWAL-QOL (169 of 268) and MDADI (168 of 268) survey groups. Although patients were included in these groups in an unbiased manner, the results, especially the SEM MDCs based on the test-retest analysis, may not be highly robust due to the relatively small sample sizes. Additionally, the SEMs could only be calculated for the HSS-DDI and its sub-domain scores because it was the only PROM administered to the test-retest group. Of the consented participants, 16% (44 of 274) of patients did not complete Week 8 surveys and/or anchor questions. Thus, the results might only reflect patients who were highly cooperative and willing to complete lengthy surveys over multiple time points. Our data did not include the provider-reported Bazaz classification and other more objective measurements of dysphagia such as a laryngoscopy or a fluoroscopic swallowing examination. Thus, direct comparisons of the previously reported prevalence or improvement rates measured with these methods were not possible. Another limitation was that the PROMs were administered during the subacute postoperative phase in our study, and our results may not apply to dysphagia assessments conducted immediately after surgery. Additionally, dysphonia was assessed using only the HSS-DDI, while three different PROMs were used for dysphagia. Among dysphagia PROMs, we could not include other PROMs used in previous ACDF studies such as the 16-item abridged form of SWAL-QOL [29] or the 10-item Eating Assessment Tool (EAT-10) that is a short and validated survey for general dysphagia [3, 21, 23]. Also, our study was conducted at a single academic institution in an urban area where 93% (250 of 268) of patients were white. Since MCID can be affected by patients’ characteristics and comorbidities [28], the generalizability of these study results should be confirmed in other patient groups.

When the MCID is higher than the MDC, then all clinically important changes can be detected because they are above the margin for measurement error. However, the MCIDs for all PROMs in our study were not higher than the MDCs, which means that clinically important differences could not necessarily be differentiated from measurement errors. To avoid this issue, some researchers recommend adding mathematical weights to MCID values and making them higher than MDCs [1, 8]. However, there are multiple recommended weights, and the selection of weighting methods is another concern. A consensus-based approach should be used in the future to address this issue.

We identified the MCIDs of three PROMs for dysphagia as 10 in the HSS-DDI, 8 in the SWAL-QOL, and 6 in the MDADI. The basic concept of MCID is that the differences below these levels cannot be considered clinically meaningful, even if the results showed “statistically significant” p values. Using the MCIDs, providers can determine if the study results are actually meaningful for their patients. For instance, there have been seven studies using the same PROMs we investigated as the outcome measures or one of the explanatory variables [12, 16, 17, 32, 39, 42, 43], and five studies reported “statistically significant” differences in some analyses related to PROMs [12, 32, 39, 42, 43]. Among these five, three studies reported the differences over our MCIDs [12, 32, 39] and these results can be considered clinically meaningful. However, two other studies reported differences below our MCID [42, 43]. The differences of score in these studies were too small to be recognized as meaningful change by patients, and care must be taken to interpret their study results. Although all three of the PROMs we studied had different MCIDs, the net values of these PROMs were small and less than 10% of the possible score. Other PROMs used in other general orthopaedics and spine surgery fields have higher values than these ranging from 10% to 30% of their possible scores [19, 28]. One possible reason is that the range of postoperative dysphagia severity was relatively small in the possible dysphagia spectrum. A small MCID means that relatively small changes in these PROMs after surgery are discernible to patients as clinical improvements. Providers should be aware of the characteristics of the MCIDs of these three PROMs. Surprisingly, there have been few studies identifying MCIDs for dysphagia PROMs including post-stroke dysphagia, head and neck cancer, or other etiologies [36]. Since the HSS-DDI is a relatively new survey instrument, no other study that we know of has determined its MCID. However, in terms of other PROMs, one post-ACDF dysphagia study using the SWAL-QOL reported “the minimum needed for clinical relevance” of the instrument [16]. The authors, however, only mentioned that they used the average 6-week postoperative SWAL-QOL and did not specify how they calculated the value in their article. Their value was 8 points, which is very close to our SWAL-QOL MDC (9 points) and the same as our SWAL-QOL MCID (8 points). Regarding the MDADI, no ACDF-related studies mention an MCID. One study about dysphagia in patients with head and neck cancer reported their MDCs and MCIDs using multiple distribution-based and anchor-based methods [20]. The authors reported higher anchor-based MCID values (around 10 points) than we did (6 points). A possible reason for this discrepancy was that this study used different anchors that were specific for head and neck cancer patients, such as feeding tube and/or aspirator requirements. A direct comparison between their study anchor-based MCIDs and ours is not possible. Although the patient populations were different, the distribution-based MDC calculated by the same method as ours (0.5 SD) was around 8 points, which is the same as our MDADI MDC. This means that the possible ranges of measurement error are not much different between head and neck cancer patients in their study and post-ACDF patients in our study.

In this study, we adopted the values calculated with minimum mean change method as anchor-based MCIDs. We also used the ROC based method, in which the MCIDs were consistently lower than those of the minimum mean change method. This discrepancy in MCID values between different assessment methods has been reported in studies of other PROMs [35, 41]. One small study compared MCIDs among six PROMs that used different methods, and reported that the MCIDs of the ROC-based method often differed from the MCIDs of the mean-change method [41]. Application of multiple methods to determine the MCID of a PROM will almost always result in a range of potential MCID values [37]. Choosing a single MCID value is a subjective process and some methodologists have recommended a systematic review of expert consensus based on Delphi methods to converge the range into a single value or at least a narrower one [37]. However, it may take time to establish expert consensus in this area because the clinical experience and use of PROMs to evaluate dysphagia after ACDF has only recently become popular. Currently, we recommend that users choose the higher values as shown in this article because the higher values are less likely to overestimate the clinical meaning of potentially harmful interventions for the prevention or treatment of dysphagia, such as topical steroid use, or novel but expensive less invasive implants.

In conclusion, the MDCs of the HSS-DDI, SWAL-QOL, and MDADI were similar, all within 8 points to 11 points. The MCIDs of these PROMs were 10 points or less on 100-point scales, which is considered small. The differences of less than 10 points on the HSS-DDI, 8 points on SWAL-QOL, and 6 points on MDADI were unlikely to be perceived by patients to be clinically important. Future studies on ACDF should report between-group differences in light of this, rather than focusing on statistical significance based on p values. Further studies in different patient populations are warranted to reach the generalizable consensus for MCIDs of dysphagia PROMs after ACDF.

References

1. Ailliet L, Rubinstein SM, de Vet HC, van Tulder MW, Terwee CB. Reliability, responsiveness and interpretability of the neck disability index-Dutch version in primary care. Eur Spine J. 2015;24:88-93.
2. Bazaz R, Lee MJ, Yoo JU. Incidence of dysphagia after anterior cervical spine surgery: a prospective study. Spine (Phila Pa 1976). 2002;27:2453-2458.
3. Belafsky PC, Mouadeb DA, Rees CJ, Pryor JC, Postma GN, Allen J, Leonard RJ. Validity and reliability of the Eating Assessment Tool (EAT-10). Ann Otol Rhinol Laryngol. 2008;117:919-924.
4. Chen AY, Frankowski R, Bishop-Leone J, Hebert T, Leyk S, Lewin J, Goepfert H. The development and validation of a dysphagia-specific quality-of-life questionnaire for patients with head and neck cancer: the M. D. Anderson dysphagia inventory. Arch Otolaryngol Head Neck Surg. 2001;127:870-876.
5. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Academic Press; 1988.
6. Copay AG, Subach BR, Glassman SD, Polly DW Jr., Schuler TC. Understanding the minimum clinically important difference: a review of concepts and methods. Spine J. 2007;7:541-546.
7. Crosby RD, Kolotkin RL, Williams GR. Defining clinically meaningful change in health-related quality of life. J Clin Epidemiol. 2003;56:395-407.
8. de Vet HC, Ostelo RW, Terwee CB, van der Roer N, Knol DL, Beckerman H, Boers M, Bouter LM. Minimally important change determined by a visual method integrating an anchor-based and a distribution-based approach. Qual Life Res. 2007;16:131-142.
9. de Vet HC, Terwee CB, Knol DL, Bouter LM. When to use agreement versus reliability measures. J Clin Epidemiol. 2006;59:1033-1039.
10. de Vet HCW, Terwee CB, Mokkink LB, Knol DL. Measurement in Medicine: A Practical Guide. Cambridge: Cambridge University Press; 2011.
11. Donner A, Eliasziw M. Sample size requirements for reliability studies. Stat Med.1987;6:441-448.
12. Fengbin Y, Xinwei W, Haisong Y, Yu C, Xiaowei L, Deyu C. Dysphagia after anterior cervical discectomy and fusion: a prospective study comparing two anterior surgical approaches. Eur Spine J. 2013;22:1147-1151.
13. Finkelstein JA, Schwartz CE. Patient-reported outcomes in spine surgery: past, current, and future directions. J Neurosurg Spine. 2019;31:155-164.
14. Frans FA, Nieuwkerk PT, Met R, Bipat S, Legemate DA, Reekers JA, Koelemay MJ. Statistical or clinical improvement? Determining the minimally important difference for the vascular quality of life questionnaire in patients with critical limb ischemia. Eur J Vasc Endovasc Surg. 2014;47:180-186.
15. Guyatt G, Walter S, Norman G. Measuring change over time: assessing the usefulness of evaluative instruments. J Chronic Dis. 1987;40:171-178.
16. Haws BE, Khechen B, Narain AS, Hijji FY, Bohl DD, Massel DH, Mayo BC, Ahn J, Singh K. Impact of local steroid application on dysphagia following an anterior cervical discectomy and fusion: results of a prospective, randomized single-blind trial. J Neurosurg Spine. 2018;29:10-17.
17. Haws BE, Khechen B, Patel DV, Yoo JS, Guntin JA, Cardinal KL, Singh K. Swallowing function following anterior cervical discectomy and fusion with and without anterior plating: A SWAL-QOL (Swallowing-Quality of Life) and radiographic assessment. Neurospine. 2019;16:601-607.
18. Hughes AP, Salzmann SN, Aguwa OK, Miller CO, Duculan R, Shue J, Cammisa FP, Sama AA, Girardi FP, Kacker A, Mancuso CA. HSS Dysphagia and Dysphonia Inventory (HSS-DDI) following anterior cervical fusion: patient-derived, validated, condition-specific patient-reported outcome measure outperforms existing indices. J Bone Joint Surg Am. 2018;100:e66.
19. Hung M, Saltzman CL, Kendall R, Bounsanga J, Voss MW, Lawrence B, Spiker R, Brodke D. What are the MCIDs for PROMIS, NDI, and ODI instruments among patients with spinal conditions? Clin Orthop Relat Res. 2018;476:2027-2036.
20. Hutcheson KA, Barrow MP, Lisec A, Barringer DA, Gries K, Lewin JS. What is a clinically relevant difference in MDADI scores between groups of head and neck cancer patients? Laryngoscope. 2016;126:1108-1113.
21. Jenkins TJ, Nair R, Bhatt S, Rosenthal BD, Savage JW, Hsu WK, Patel AA. The effect of local versus intravenous corticosteroids on the likelihood of dysphagia and dysphonia following anterior cervical discectomy and fusion: a single-blinded, prospective, randomized controlled trial. J Bone Joint Surg Am. 2018;100:1461-1472.
22. King MT. A point of minimal important difference (MID): a critique of terminology and methods. Expert Rev Pharmacoecon Outcomes Res. 2011;11:171-184.
23. Leonard R, Belafsky P. Dysphagia following cervical spine surgery with anterior instrumentation: evidence from fluoroscopic swallow studies. Spine (Phila Pa 1976). 2011;36:2217-2223.
24. Leopold SS, Porcher R. Editorial: The Minimum Clinically Important Difference-The Least We Can Do. Clin Orthop Relat Res. 2017;475:929-932.
25. Leopold SS, Porcher R. Editorial: Threshold p values in orthopaedic research-We know the problem. What is the solution? Clin Orthop Relat Res. 2018;476:1689-1691.
26. Liu FY, Yang DL, Huang WZ, Huo LS, Ma L, Wang H, Yang SD, Ding WY. Risk factors for dysphagia after anterior cervical spine surgery: A meta-analysis. Medicine (Baltimore). 2017;96:e6267.
27. Liu J-M, Tong W-L, Chen X-Y, Zhou Y, Chen W-Z, Huang S-H, Liu Z-L. The incidences and risk factors related to early dysphagia after anterior cervical spine surgery: A prospective study. PloS one. 2017;12:e0173364-e0173364.
28. Maltenfort M, Díaz-Ledezma C. Statistics In Brief: Minimum clinically important difference-availability of reliable estimates. Clin Orthop Relat Res. 2017;475:933-946.
29. Mayo BC, Massel DH, Bohl DD, Patel DV, Khechen B, Haws BE, Narain AS, Hijji FY, Singh K. Dysphagia following anterior cervical spine surgery: assessment using an abridged SWAL-QOL. Int J Spine Surg. 2019;13:102-109.
30. McHorney CA, Bricker DE, Robbins J, Kramer AE, Rosenbek JC, Chignell KA. The SWAL-QOL outcomes tool for oropharyngeal dysphagiain adults: II. Item reduction and preliminary scaling. Dysphagia. 2000;15:122-133.
31. McHorney CA, Tarlov AR. Individual-patient monitoring in clinical practice: are available health status surveys adequate? Qual Life Res. 1995;4:293-307.
32. Mendoza-Lattes S, Clifford K, Bartelt R, Stewart J, Clark CR, Boezaart AP. Dysphagia following anterior cervical arthrodesis is associated with continuous, strong retraction of the esophagus. J Bone Joint Surg Am. 2008;90:256-263.
33. Min Y, Kim WS, Kang SS, Choi JM, Yeom JS, Paik NJ. Incidence of dysphagia and serial videofluoroscopic swallow study findings after anterior cervical discectomy and fusion: a prospective study. Clin Spine Surg. 2016;29:E177-181.
34. Mirza SK, Deyo RA, Heagerty PJ, Konodi MA, Lee LA, Turner JA, Goodkin R. Development of an index to characterize the "invasiveness" of spine surgery: validation by comparison to blood loss and operative time. Spine.(Phila Pa 1976). 2008;33:2651-2662.
35. Ogura K, Uehara K, Akiyama T, Shinoda Y, Iwata S, Tsukushi S, Kobayashi E, Hirose T, Yonemoto T, Endo M, Tanzawa Y, Nakatani F, Kawano H, Tanaka S, Kawai A. Minimal clinically important differences in Toronto Extremity Salvage Score for patients with lower extremity sarcoma. J Orthop Sci. 2020;25:315-318.
36. Oremus M, Santaguida P, Walker K, Wishart L. AHRQ Technology Assessments. Methodological Issues in Evaluation of Innovative Training Approaches to Stroke Rehabilitation. Rockville (MD): Agency for Healthcare Research and Quality (US); 2008.
37. Revicki D, Hays RD, Cella D, Sloan J. Recommended methods for determining responsiveness and minimally important differences for patient-reported outcomes. J Clin Epidemiol. 2008;61:102-109.
38. Schwartz CE, Finkelstein JA, Rapkin BD. Appraisal assessment in patient-reported outcome research: methods for uncovering the personal context and meaning of quality of life. Qual Life Res. 2017;26:545-554.
39. Siska PA, Ponnappan RK, Hohl JB, Lee JY, Kang JD, Donaldson WF 3rd. Dysphagia after anterior cervical spine surgery: a prospective study using the swallowing-quality of life questionnaire and analysis of patient comorbidities. Spine (Phila Pa 1976). 2011;36:1387-1391.
40. Turner D, Schunemann HJ, Griffith LE, Beaton DE, Griffiths AM, Critch JN, Guyatt GH. Using the entire cohort in the receiver operating characteristic analysis maximizes precision of the minimal important difference. J Clin Epidemiol. 2009;62:374-379.
41. Turner D, Schünemann HJ, Griffith LE, Beaton DE, Griffiths AM, Critch JN, Guyatt GH. The minimal detectable change cannot reliably replace the minimal important difference. J Clin Epidemiol. 2010;63:28-36.
42. Vaishnav AS, Saville P, McAnany S, Haws B, Singh K, Iyer S, Albert T, Gang CH, Qureshi SA. Is the likelihood of dysphagia different in patients undergoing one-level versus two-level anterior cervical discectomy and fusion? Spine J. [Published online ahead of print Feburuary 21, 2020]. DOI: 10.1016/j.spinee.2020.01.011.
43. Vaishnav AS, Saville P, McAnany S, Patel D, Haws B, Khechen B, Singh K, Gang CH, Qureshi SA. Predictive factors of postoperative dysphagia in single-level anterior cervical discectomy and fusion. Spine (Phila Pa 1976). 2019;44:E400-E407.
44. van Kampen DA, Willems WJ, van Beers LWAH, Castelein RM, Scholtes VAB, Terwee CB. Determination and comparison of the smallest detectable change (SDC) and the minimal important change (MIC) of four-shoulder patient-reported outcome measures (PROMs). J Orthop Surg Res. 2013;8:40-40.
45. Wyrwich KW, Norquist JM, Lenderking WR, Acaster S, the Industry Advisory Committee of International Society for Quality of Life Research. Methods for interpreting change over time in patient-reported outcome measures. Qual Life Res. 2013;22:475-483.

Supplemental Digital Content

© 2020 by the Association of Bone and Joint Surgeons