Abstract
Objective:
To compare healthcare utilization in voice patients with vs. without mental health (MH) diagnoses
Study Design:
Retrospective study using electronic medical records from large regional health care system
Methods:
We examined data on sociodemographic characteristics, comorbidities, voice-related diagnoses, and patterns of healthcare utilization (including medication use, tests and procedures, and outpatient visits). The study period spanned January 2005 through June 2017.
Results:
24,672 patients had at least 1 voice-related diagnosis. 11,483 (47%) also had at least 1 MH diagnosis, compared to 14% in the overall repository (p <0.0001). The most common voice-related diagnoses were nonspecific dysphonia (80%), acute laryngitis (30%), and vocal fold paresis/paralysis (7%). The 11,483 patients with both voice-related and MH diagnoses were more likely to have acute laryngitis and/or nonspecific dysphonia; less likely to have laryngeal cancer and/or paresis/paralysis; and more likely to have seen a primary care provider, to have received medications, and to have undergone radiology studies. In contrast, the 13,189 patients with only voice-related diagnoses had more overall voice-related visits, were more likely to have seen an otolaryngologist, and were more likely to have undergone a voice evaluation with a speech language pathologist.
Conclusion:
Voice patients with MH diagnoses were less likely to see otolaryngology and more likely to have radiology studies than voice patients without MH diagnoses. Further study is warranted to characterize temporal sequences of care in this group of patients and determine whether these differences are attributable to referral patterns from primary care.
Keywords: Laryngology, voice, dysphonia, mental health, health services
Introduction
The human voice is a fundamental means of expression. About 30% of people have a voice problem at some point in their lives1. When the voice is dysfunctional, a patient can experience significant emotional distress and disability2–5. Dysphonic patients report quality of life scores comparable to patients with chronic diseases such as congestive heart failure, angina, sciatica, and chronic obstructive pulmonary disease6. Previous cross-sectional studies of patient-reported symptoms in tertiary care voice clinics have shown a close relationship between mental health (MH) and voice-related diagnoses, with a high prevalence of anxiety, depression, and somatic concerns2–4,7. Similarly, a study using formal psychiatric evaluations showed that patients with functional dysphonia had a higher rate of anxiety and mood disorders,8 although the generalizability of those findings to larger groups of patents with broader voice-related diagnoses is unknown.
MH disorders affect about 17% of Americans every year9. Specific MH disorders, such as somatization, are associated with an increase of up to 50% in the utilization rate of non-MH care10. Despite the high prevalence of distress in patients with voice-related diagnoses, the impact of MH on their health care utilization has not yet been characterized. In a prior population-based study, we had observed that those who reported depression symptoms were less likely to report receiving voice care, and among those who did, were more likely to report non-laryngeal diagnoses such as reflux or allergies, compared to those who reported a voice problem but no depression symptoms.11 In that study we were unable to observe actual voice-related health care utilization. Understanding how patients with both voice-related and MH diagnoses utilize health care could facilitate the eventual development of more tailored treatment approaches for voice disorders in these patients.
In this study, our objective was to compare associations between voice-related diagnoses and health care utilization in voice patients with vs. without MH diagnoses. Using a large health care system’s electronic medical records, we aimed to describe sociodemographic characteristics, comorbidities, voice-related and mental health diagnoses, and patterns of voice-related health care utilization (including medication use, tests and procedures, and outpatient visits) in these groups. We hypothesized that patients seen for voice disorders would have a higher prevalence of MH diagnoses than the overall health system baseline prevalence of MH diagnoses. We also hypothesized that patients with both voice and MH diagnoses would be less likely to have specialty care, such as otolaryngology, compared to patients with voice diagnoses but no MH diagnoses.
Methods
The University of Minnesota Institutional Review Board approved this retrospective study (#1512M81630) using the University of Minnesota Clinical Data Repository (CDR). Beginning in 2005, the CDR includes data from more than 2.5 million patients seen at 8 hospitals and more than 40 clinics. We collected information on patients (living or deceased) who were at least 18 years old and who had not opted out from having their data used for research at the time we queried the CDR on July 1, 2017.
To identify CDR records of patients with voice-related and/or MH diagnoses, we used codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and 10th Revision, Clinical Modification (ICD-10-CM) (see Appendix, Table 1).
We identified all voice-related diagnoses and categorized them as follows12: acute laryngitis, chronic laryngitis, benign neoplasm of the larynx, laryngeal cancer, nonspecific dysphonia (i.e., combined with other unspecified diseases), vocal fold paresis/paralysis (unilateral or bilateral), laryngeal stenosis, and laryngeal spasm. In addition, we identified 4 relevant categories of MH diagnoses: depression, anxiety, somatoform disorders, and stress-related disorders (including adjustment disorders). We divided patients into 2 cohorts: those with only voice-related diagnoses and those with both voice-related and MH diagnoses.
Outcome measures included cohort differences in sociodemographic characteristics, comorbidities, voice-related diagnoses, and patterns of health care utilization (including medication use, tests and procedures, and outpatient visits). Specific sociodemographic characteristics noted at each patient’s first voice disorder visit included age, gender, race, zip code, and insurance coverage. We did not analyze substance use (including tobacco and alcohol use) because that information was incomplete for 38% of patients. To determine comorbidities, we used the Charlson Comorbidity Index (CCI), which facilitates prediction of 1-year survival through a weighted index that adjusts for 19 conditions13,14. The CCI does not include MH comorbidities or voice-related diagnoses. We determined the most common voice-related diagnoses in each of our 2 cohorts.
To account for comorbidities and patterns of voice-related healthcare utilization, we assessed the number of voice-related visits within 1 year after the first voice-related visit, the patient’s use of voice-related medications within 2 weeks after a voice-related visit, and the voice-related tests and procedures performed within 1 year after a voice-related visit. We also subdivided into whether a patient ever or never saw an otolaryngologist. We subcategorized voice-related visits by type of provider: primary care (family medicine, internal medicine, medicine/pediatrics), otolaryngology, emergency care, and all others. We examined the use of 6 voice-related medication categories previously used by Cohen et al.: proton pump inhibitors, oral antibiotics, oral steroids, inhaled steroids, antihistamines, and H2 receptor antagonists.15 For voice-related tests and procedures, we used 2017 Current Procedural Terminology (CPT) codes as previously described (see Appendix, Table 2).15
To summarize the data, we used descriptive statistics (for categorical variables, counts and percentages; for continuous variables, means and standard deviations). To assess the effect of sociodemographic (including age, gender, race, insurance status) and clinical (CCI, voice-related diagnosis) characteristics on MH diagnoses and health care utilization outcomes, we controlled for these factors by applying multiple logistic regression models and reported the adjusted odds ratio (OR), 95% confidence interval (CI), and P value. Mean utilization was compared between groups with linear regression models. For statistical analysis, we used SAS version 9.4 (SAS Institute Inc., Cary, NC). Sensitivity analyses adjusting for whether a patient saw an otolaryngologist were also performed, with no meaningful impact on our observations.
Results
Of the 2.5 million patients with records in the CDR, 24,672 had at least 1 voice-related diagnosis during the study period (January 2005 through June 2017). Of those patients, 11,483 (47%) also had at least 1 MH diagnosis. By contrast, 14% of patients in the overall database had at least 1 MH diagnosis (p<0.0001), and these differences remained when the prevalence of MH diagnoses in the overall database was adjusted for the age distribution of the cohort with voice-related diagnoses.
The overall population of patients with voice-related diagnoses (Table 1) was predominantly female and white, with a mean age of 53 years (SD 18). Most lived in urban areas, with a high rate of insurance coverage. The cohort with both voice-related and MH diagnoses had a mean CCI score that was nearly twice as high (M = 2.7, ± 3.4) than in the cohort of patients with only voice-related diagnoses (M = 1.5, ± 2.6).
Table 1:
Characteristics of patients with at least 1 voice-related diagnosis (N = 24,672), by cohort
| Both V + MH n = 11,483 (47%) | Only V n = 13,189 (53%) | Total N = 24,672 | |
|---|---|---|---|
| Gender | |||
| Male | 3,302 (29%) | 5,418 (41%) | 8,720 (35%) |
| Female | 8,181 (71%) | 7,771 (59%) | 15,952 (65%) |
| Race | |||
| White | 9,951 (87%) | 10,406 (79%) | 20,357 (82%) |
| Black | 441 (4%) | 631 (5%) | 1,072 (4%) |
| Other (Asian, American Indian, Pacific Islander) | 305 (3%) | 400 (3%) | 705 (3%) |
| Unknown | 786 (7%) | 1,752 (13%) | 2,538 (10%) |
| Age, years, mean (± SD) | 53 (± 17.4) | 52 (± 18.4) | 53 (± 17.9) |
| RUCA | |||
| Urban | 10,270 (90%) | 11,362 (86%) | 21,632 (88%) |
| Rural (large, small, or isolated city/town) | 1,197 (10%) | 1,801 (14%) | 2,998 (12%) |
| Unknown | 16 (< 1%) | 26 (< 1%) | 42 (< 1%) |
| Insurance | |||
| Any coverage | 11,109 (97%) | 12,676 (96%) | 23,785 (96%) |
| Private coverage | 10,672 (93%) | 12,290 (93%) | 22,962 (93%) |
| Medicare coverage | 3,378 (29%) | 3,292 (25%) | 6,670 (27%) |
| Medicaid coverage | 2,009 (18%) | 956 (7%) | 2,965 (12%) |
| CCI score, mean (± SD) | 2.7 (± 3.4) | 1.5 (±2.6) | 2.1 (±3.1) |
| 0 | 3,528 (31%) | 7,164 (54%) | 10,692 (43%) |
| 1 | 2,493 (22%) | 2,283 (17%) | 4,776 (19%) |
| 2+ | 5,462 (48%) | 3,742 (28%) | 9,204 (37%) |
CCI = Charlson Comorbidity Index, MH = mental health, RUCA = Rural-Urban Commuting Area, SD = standard deviation, V = voice-related
The most common voice-related diagnosis was nonspecific dysphonia (n = 16,750), followed by acute laryngitis (n = 7,320) and vocal fold paresis/paralysis (n = 1,262) (Table 2). The most common MH diagnosis was anxiety (n = 7,880), followed by depression (n = 7,697), stress-related disorders (n = 2,140), and somatoform disorders (n = 1,395). Patients with multiple diagnoses could have diagnoses in more than once subcategory, so totals could exceed 100%.
Table 2:
Categories of patients’ diagnoses,† by cohort
| Both V + MH n = 11,483 (47%) | Only V n = 13,189 (53%) | Total N = 24,672 | |
|---|---|---|---|
| Voice-related categories, n or N (%) | |||
| Acute laryngitis | 3,684 (32.0%) | 3,936 (29.9%) | 7,320 |
| Chronic laryngitis | 222 (1.9%) | 192 (1.5%) | 414 |
| Benign neoplasm of the larynx | 292 (2.5%) | 450 (3.4%) | 742 |
| Laryngeal cancer | 268 (2.3%) | 446 (3.4%) | 714 |
| Nonspecific dysphonia | 7,890 (68.7%) | 8,860 (67.2%) | 16,750 |
| Vocal fold paresis/paralysis | 511 (4.5%) | 751 (5.7%) | 1,262 |
| Laryngeal stenosis | 135 (1.2%) | 211 (1.6%) | 346 |
| Laryngeal spasm | 216 (1.9%) | 384 (2.9%) | 600 |
| Mental Health categories, n or N (%) | |||
| Anxiety | 7,880 (67.0%) | - | 7,880 |
| Depression | 7,697 (68.6%) | - | 7,697 |
| Stress-related disorders | 2,140 (12.2%) | - | 2,140 |
| Somatoform disorders | 1,395 (18.6%) | - | 1,395 |
Individual patients could have more than 1 diagnosis.
MH = mental health, V = voice-related
In multivariable regression analysis, several sociodemographic factors were significantly associated with having both voice-related and MH diagnoses: female gender, white race, Medicaid insurance, and a higher number of comorbidities, i.e., a higher CCI score (Table 3). Voice-related diagnoses associated with also having MH diagnoses were acute laryngitis (OR = 1.22, P = 0.0002) and nonspecific dysphonia (OR = 1.28, P < 0.0001). Conversely, patients with a voice-related diagnosis of laryngeal cancer (OR = 0.52, P < 0.0001) or vocal fold paresis/paralysis (OR = 0.72, P < 0.0001) were significantly less likely to have both voice-related and MH diagnoses.
Table 3:
Factors associated with having both voice-related and mental health diagnoses
| Adjusted OR | 95% CI | P | |
|---|---|---|---|
| Increasing age (per year) | 0.99 | 0.99–0.99 | < 0.0001 |
| Gender | |||
| Male | Referent | - | - |
| Female | 1.78 | 1.67–1.89 | < 0.0001 |
| Race | |||
| White | Referent | - | - |
| Black | 0.47 | 0.41–0.60 | < 0.0001 |
| Other | 0.70 | 0.59–0.83 | < 0.0001 |
| Insurance† | |||
| Private coverage | 1.15 | 1.03–1.29 | 0.015 |
| Medicare coverage | 1.07 | 0.98–1.15 | 0.12 |
| Medicaid coverage | 2.65 | 2.40–2.94 | < 0.0001 |
| CCI score | |||
| 0 | Referent | - | - |
| 2+ | 3.67 | 3.41–3.94 | < 0.0001 |
| Voice-related diagnoses‡ | |||
| Acute laryngitis | 1.22 | 1.10–1.35 | 0.0002 |
| Chronic laryngitis | 1.29 | 1.03–1.61 | 0.025 |
| Benign neoplasm of the larynx | 0.80 | 0.68–0.95 | 0.011 |
| Laryngeal cancer | 0.52 | 0.43–0.62 | < 0.0001 |
| Nonspecific dysphonia | 1.28 | 1.16–1.41 | < 0.0001 |
| Vocal fold paresis/paralysis | 0.72 | 0.63–0.82 | < 0.0001 |
| Laryngeal stenosis | 0.72 | 0.57–0.92 | 0.0075 |
| Laryngeal spasm | 0.75 | 0.62–0.90 | 0.0027 |
Referent = not having the specific type of insurance
Referent = not having the specific voice-related diagnosis
Adjusted OR, CI, and P per multiple logistic regression model with mental health diagnosis as the outcome
CCI = Charlson Comorbidity Index, CI = confidence interval, OR = odds ratio
The total number of voice-related visits was higher in patients with only voice-related diagnoses (M = 2.01) than in patients with both voice-related and MH diagnoses (M = 1.88 visits, P < 0.0001). The number of voice-related primary care visits was higher in patients with both voice-related and MH diagnoses (M = 0.68) than in patients with only voice-related diagnoses (M = 0.49, P < 0.0001). However, the number of voice-related otolaryngology visits was higher in patients with only voice-related diagnoses (M = 0.95) than in patients with both voice-related and MH diagnoses (M = 0.65, P < 0.0001) (Table 4).
Table 4:
Voice-related health care utilization, by cohort
| BothV + MH n = 11,483 (47%) | Only V (n = 13,189 (53%) | p* | |
|---|---|---|---|
| Voice-related visits† per patient within 1 year after a voice-related diagnosis, mean (± SD) | |||
| Total | 1.88(3.28) | 2.01 (3.28) | < 0.0001 |
| Primary care | 0.68 (0.84) | 0.49 (0.68) | < 0.0001 |
| Emergency medicine | 0.04(0.22) | 0.04(0.21) | 0.79 |
| Otolaryngology | 0.65(1.61) | 0.95 (2.04) | < 0.0001 |
| Other (including psychiatric) | 0.56(2.58) | 0.59 (2.54) | < 0.0001 |
| Laryngeal medications† per patient within 2 weeks after a voice-related diagnosis, mean (± SD) | |||
| Total | 0.63(1.283) | 0.42 (0.925) | < 0.0001 |
| Number of CPT code† per patient within 1 year after a voice-related diagnosis, mean (± SD) | |||
| Radiology tests | 3.67(6.37) | 0.96(2.34) | < 0.0001 |
| Swallow studies | 0.44(1.50) | 0.16(0.79) | < 0.0001 |
| Swallow therapy | 0.22(0.96) | 0.08 (0.56) | < 0.0001 |
| Voice evaluation/therapy by speech pathology | 0.51(1.93) | 0.69 (2.10) | < 0.0001 |
| Laryngoscopy | 0.50(1.31) | 0.43 (1.26) | 0.0056 |
| Stroboscopy | 0.08(0.50) | 0.16(0.71) | < 0.0001 |
| Laryngeal surgery | 0.06(0.59) | 0.05 (0.46) | 0.60 |
| Chemodenervation | 0.06 (1.09) | 0.11 (1.35) | 0.014 |
P per a linear model adjusting for age, gender, zip code category, insurance status, and CCI score
Visits and procedures capped at 1 year; medications capped at 2 weeks
CCI = Charlson Comorbidity Index, CPT = Current Procedural Terminology, MH = mental health, SD = standard deviation, V = voice-related
The rate of voice-related medication use within 2 weeks after a voice-related visit was higher in patients with both voice-related and MH diagnoses (M = 0.63) than in patients with only voice-related diagnoses (M = 0.42, P < 0.0001). The number of radiology tests within 1 year after a voice-related diagnosis was higher in patients with both voice-related and MH diagnoses (M = 3.67) than in patients with only voice-related diagnoses (M = 0.96, P < 0.0001), including more swallow studies per individual (M = 0.44 vs. M = 0.16, P <0.0001) and more swallow therapy codes per individual (M = 0.22 vs. M = 0.08, P < 0.0001).
The number of speech pathology codes per individual (for voice evaluation/therapy) was higher in patients with only voice-related diagnoses (M = 0.69) than in patients with both voice-related diagnosis and MH diagnoses (M = 0.51, P < 0.0001). Similarly, the number of stroboscopy codes per individual was also higher in patients with only voice-related diagnoses (M = 0.16 vs. M = 0.08, P < 0.0001).
We next needed to determine whether these findings could be attributed to differences in sociodemographic or clinical characteristics such as age, comorbidity, or voice-related diagnosis. In addition, because number of visits could be expected to be different for different voice-related diagnoses, for this analysis we examined whether patients ever saw an otolaryngologist (not total number of voice-related visits). Multivariable regression analysis, which adjusted for these various characteristics, showed that a higher number of comorbidities and a diagnosis of acute laryngitis were both associated with seeing a primary care provider for voice-related care (Table 5). Five diagnoses (chronic laryngitis, laryngeal cancer, nonspecific dysphonia, vocal cord paresis/paralysis, and laryngeal stenosis) were also associated with ever seeing an otolaryngologist for voice-related care. Importantly, after adjusting for sociodemographic factors, comorbidities, and the specific voice-related diagnosis, we found that MH diagnoses were more strongly associated with seeing a primary care provider for voice-related care (OR = 1.52, 95% CI = 1.43–1.61, P < 0.0001), while a MH diagnosis decreased the odds of ever seeing an otolaryngologist (OR = 0.88, 95% CI = 0.82–0.94, P = 0.0001).
Table 5:
Factors associated with health care utilization for voice-related visits, by provider type
| Primary care provider | Otolaryngologist (ever) | |||
|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | |
| Increasing age (per year) | 1.01(1.01–1.01) | < 0.0001 | 1.00 (1.00–1.00) | 0.44 |
| Gender | ||||
| Male | Referent | Referent | ||
| Female | 1.03(0.97–1.10) | 0.30 | 0.98 (0.92–1.05) | 0.55 |
| Race | ||||
| White | Referent | Referent | ||
| Black | 0.98(0.85–1.11) | 0.70 | 0.81 (0.69–0.95) | 0.011 |
| Other | 1.03 (0.88–1.20) | 0.73 | 0.84 (0.70–1.01) | 0.070 |
| Insurance† | ||||
| Private coverage | 0.90(0.81–1.01) | 0.076 | 1.19 (1.04–1.35) | 0.012 |
| Medicare coverage | 0.77 (0.71–0.83) | < 0.0001 | 0.94 (0.86–1.02) | 0.14 |
| Medicaid coverage | 0.80 (0.73–0.88) | < 0.0001 | 1.05 (0.94–1.17) | 0.38 |
| CCI score | ||||
| 0 | Referent | Referent | ||
| 2+ | 1.47(1.37–1.59) | < 0.0001 | 0.44 (0.41–0.48) | < 0.0001 |
| Voice-related diagnoses‡ | ||||
| Acute laryngitis | 2.85(2.55–3.17) | < 0.0001 | 0.48 (0.43–0.55) | < 0.0001 |
| Chronic laryngitis | 2.21(1.76–2.78) | < 0.0001 | 2.89 (2.20–3.80) | < 0.0001 |
| Benign neoplasm of the larynx | 1.38(1.17–1.62) | 0.0001 | 2.58 (2.14–3.12) | < 0.0001 |
| Laryngeal cancer | 0.45 (0.37–0.54) | < 0.0001 | 7.48 (6.01–9.32) | < 0.0001 |
| Nonspecific dysphonia | 1.18(1.06–1.31) | 0.0024 | 16.17 (14.08–18.58) | < 0.0001 |
| Vocal fold paresis/paralysis | 0.33 (0.29–0.39) | < 0.0001 | 6.60 (5.61–7.75) | < 0.0001 |
| Laryngeal stenosis | 0.24 (0.18–0.33) | < 0.0001 | 14.32 (10.46–19.62) | < 0.0001 |
| Laryngeal spasm | 0.55 (0.45–0.67) | < 0.0001 | 4.93 (3.93–6.20) | < 0.0001 |
| Mental health diagnoses | ||||
| Yes | 1.52(1.43–1.61) | < 0.0001 | 0.88 (0.82–0.94) | 0.0001 |
| No | Referent | Referent | ||
Referent = having the specific type of insurance
Referent = having the specific voice-related diagnosis
CCI = Charlson Comorbidity Index, CI = confidence interval, OR = odds ratio
Finally, we observed that age, insurance, comorbidities, and laryngeal diagnoses were associated with different patterns of use of speech pathology voice evaluations, laryngoscopy, radiology, and medication use. After we controlled for those factors, MH diagnoses were associated with lower odds of undergoing voice evaluation/therapy by speech pathology and with higher odds of undergoing laryngoscopy, undergoing radiology tests, and using voice-related medications (Table 6).
Table 6:
Factors associated with voice-related health care utilization
| Voice evaluation/therapy by speech pathology | Laryngoscopy | Radiology tests | Laryngeal medications | |||||
|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | OR (95% CI) | P | OR (95% CI) | P | |
| Age | 0.99 (0.99–0.99) | < 0.0001 | 1.01(1.00–1.01) | <0.0001 | 1.01 (1.01–1.02) | < 0.0001 | 1.00 (1.00–1.00) | 0.59 |
| Gender | ||||||||
| Male | Referent | Referent | Referent | Referent | ||||
| Female | 1.05 (0.97–1.14) | 0.26 | 0.87 (0.80–0.95) | 0.0018 | 0.96 (0.90–1.03) | 0.21 | 0.96 (0.90–1.02) | 0.17 |
| Race | ||||||||
| White | Referent | Referent | Referent | Referent | ||||
| Black | 1.16 (0.96–1.41) | 0.13 | 0.86 (0.70–1.05) | 0.14 | 1.14 (0.98–1.32) | 0.090 | 1.18 (1.03–1.35) | 0.018 |
| Other | 0.90(0.71–1.15) | 0.40 | 1.12 (0.89–1.43) | 0.33 | 0.86 (0.72–1.03) | 0.097 | 1.12 (0.95–1.31) | 0.19 |
| Insurance† | ||||||||
| Private coverage | 1.27(1.08–1.51) | 0.0052 | 0.70 (0.60–0.83) | <0.0001 | 0.92 (0.81–1.04) | 0.19 | 1.04 (0.92–1.16) | 0.53 |
| Medicare coverage | 1.49 (1.34–1.66) | < 0.0001 | 0.91 (0.81–1.01) | 0.083 | 1.02 (0.94–1.12) | 0.60 | 0.78 (0.72–0.85) | < 0.0001 |
| Medicaid coverage | 0.92 (0.80–1.05) | 0.20 | 0.95 (0.83–1.08) | 0.43 | 1.58 (1.43–1.75) | < 0.0001 | 1.07 (0.98–1.17) | 0.16 |
| CCI score | ||||||||
| 0 | Referent | Referent | Referent | Referent | ||||
| 2+ | 0.93 (0.84–1.03) | 0.16 | 1.55(1.39–1.72) | < 0.0001 | 5.86 (5.40–6.35) | < 0.0001 | 1.20(1.11–1.29) | < 0.0001 |
| Voice-related diagnoses‡ | ||||||||
| Acute laryngitis | 0.82 (0.70–0.96) | 0.015 | 1.37(1.18–1.59) | <0.0001 | 1.75 (1.55–1.97) | < 0.0001 | 3.07 (2.76–3.42) | < 0.0001 |
| Chronic laryngitis | 1.62(1.19–2.21) | 0.0024 | 2.36(1.73–3.23) | < 0.0001 | 1.66(1.29–2.15) | 0.0001 | 2.88 (2.30–3.60) | < 0.0001 |
| Benign neoplasm of larynx | 2.16(1.79–2.61) | < 0.0001 | 1.66 (1.34–2.05) | < 0.0001 | 1.19 (0.98–1.44) | 0.076 | 1.43 (1.20–1.69) | < 0.0001 |
| Laryngeal cancer | 2.46 (1.98–3.05) | < 0.0001 | 2.30(1.80–2.93) | < 0.0001 | 1.45 (1.19–1.77) | 0.0003 | 2.21 (1.85–2.64) | < 0.0001 |
| Nonspecific dysphonia | 9.57 (7.82–11.70) | < 0.0001 | 2.85 (2.45–3.33) | < 0.0001 | 1.65 (1.47–1.86) | < 0.0001 | 2.33 (2.10–2.59) | < 0.0001 |
| Vocal fold paresis/paralysis | 3.79 (3.27–4.39) | < 0.0001 | 1.28 (1.09–1.50) | 0.0026 | 1.54(1.32–1.79) | < 0.0001 | 0.78 (0.67–0.90) | 0.0008 |
| Laryngeal stenosis | 1.70(1.28–2.27) | 0.0003 | 4.60 (3.36–6.31) | < 0.0001 | 1.69(1.29–2.21) | 0.0002 | 2.85 (2.26–3.60) | < 0.0001 |
| Laryngeal spasm | 2.43 (1.94–3.04) | < 0.0001 | 1.22 (0.97–1.53) | 0.10 | 1.36(1.10–1.69) | 0.0042 | 1.18 (0.96–1.44) | 0.12 |
| Mental health diagnoses | ||||||||
| Yes | 0.84 (0.77–0.91) | < 0.0001 | 1.72 (1.58–1.88) | < 0.0001 | 3.48 (3.26–3.71) | < 0.0001 | 1.37 (1.29–1.46) | < 0.0001 |
| No | Referent | Referent | Referent | Referent | ||||
Referent = having the specific type of insurance
Referent = having the specific voice-related diagnosis
CCI = Charlson Comorbidity Index, CI = confidence interval, OR = odds ratio
Discussion
Several retrospective studies have examined concurrent MH and voice problems2,3,7,8. Typically, those studies have been based in tertiary care clinics and referral centers, and most have also been limited by small population size. One national survey study showed that those reporting both voice concerns and symptoms of depression were less likely to get voice care and also less likely to report improvement with voice care.11 Our study is the first to query a large health care system database to investigate the relationship between MH and voice-related diagnoses based on health care utilization.
Of the nearly 25,000 patients in our study with voice-related diagnoses, nearly half (47%) also had MH diagnoses at some point during the study period, per their electronic medical records. This figure is dramatically higher than the baseline prevalence (14%) of MH diagnoses in the overall clinical data repository. After we adjusted for other factors, we found that patients were more likely to have both voice-related and MH diagnoses if they were female, white, and/or had additional comorbidities. Similarly, previous retrospective studies have consistently shown that female patients experience greater psychosocial distress.2–4 Moreover, studies of patients with multiple morbidities find that physical medical diagnoses occur with MH diagnoses at a rate of about 30%.16,17 Overall, the results of this study confirm a high co-occurrence of MH and physical health diagnoses among persons with voice-related disorders and underscore the importance of being aware of potential mental health concerns among patients who present with voice disorders.
In our study, the most common overall voice-related diagnoses were nonspecific dysphonia, acute laryngitis, and vocal fold paresis/paralysis, consistent with prior work.15 The relatively nonspecific diagnoses of acute laryngitis and nonspecific dysphonia were more common in patients with both voice-related and MH diagnoses than in patients with only voice-related diagnoses. Our findings are consistent with previous cross-sectional work suggesting a lack of specificity of laryngeal diagnoses in patients with (vs. without) depressive symptoms.11 A potential contributor to the lack of specificity of laryngeal diagnoses could be differences in provider type. Although the provider type distribution in our study was similar to the distribution in previous studies of health care utilization in dysphonic patients,15 patients in our study with both voice-related and MH diagnoses were more likely to see a primary care provider; in contrast, patients with only voice-related diagnoses were more likely to see an otolaryngologist. Differences in specialty expertise and instrumentation could lead to differences in diagnosis patterns.12 Although these differences in expertise may seem to suggest that these differences in diagnosis are not meaningful, the presumptive diagnosis selected by the primary care provider may directly influence the likelihood and timing of imaging, medication, and/or specialty referral. Thus these differences in voice-related diagnoses, whether the diagnoses themselves are accurate or not, are highly relevant for the real-life experiences of patients with voice disorders.
In addition to a higher number of overall voice-related visits, we observed that patients with both voice-related and MH diagnoses had a higher rate of voice-related medication use and were more likely to undergo diagnostic tests such as radiology, but less likely to undergo voice therapy. It is unknown whether patients were less likely to undergo voice therapy due to lack of referral to voice therapy or lack of follow through from the patient. Lack of follow through could potentially be related to concerns that may be exacerbated by MH comorbidities including unclear potential improvement, not understanding the purpose of therapy, or concerns it would be too hard.18 Another potential contributing factor is insurance, since patients with both MH and voice diagnoses were more likely to have Medicaid; although this was mitigated by inclusion of insurance status as a covariate in all regression analyses, it is possible there may be more subtle impacts that remain to be observed. A previous systematic review of health care utilization showed that patients with (vs. without) depression had higher medical costs.19 Another study observed that patients with somatizing (vs. nonsomatizing) disorders had twice the rate of outpatient and inpatient health care utilization and twice the annual medical costs.10 It is possible that the differences we observe in utilization of services such as imaging may partially be related to patient distress that prompts providers to order additional testing or perform procedures as a way to supplement their evaluations and provide reassurance to concerned patients.
Strengths of our study include the use of data from a large health care system (including more than 2.5 million patients). The potential generalizability of our study is enhanced by robust data representing small community clinics as well as a large academic center, encompassing real-life diversity in care models and settings.
Limitations of the study are inherent to research using electronic health record data, including the potential for misclassified or missing data and the fact that we can only assess association, not causality, in this retrospective observational study of data from one large health system. Care outside of the integrated health clinics in this system is not captured in this data set. We based decisions regarding codes for diagnoses, procedures, and medications on the existing literature,12,20 yet the literature contains limited data on the use of claims to study mental health.21 Our estimates could be conservative given potential underreporting of MH diagnoses.22 Finally, our data source does not allow us to assess treatment outcomes. The strengths of our study outweigh these limitations, and overall the findings provide a valuable window into how MH diagnoses may influence voice-related care utilization.
In summary, we observed that nearly half of patients with voice disorders have MH diagnoses, a much higher prevalence than patients in general within this data source. This suggests that voice care providers should have a high sensitivity to potential MH concerns in our patient population. We also identified significant differences in voice-related care for patients with vs. without MH diagnoses, particularly regarding the use of primary vs. specialty care. Patients with voice-related and MH diagnoses were less likely to see an otolaryngologist than patients without MH diagnoses, and more likely to see primary care providers for voice-related visits. Those with MH diagnoses were also more likely to undergo imaging. Those with both mental health and voice diagnoses tend to see non-specialists and are at risk for greater medication use and imaging (with associated radiation exposure) suggesting a need for more focused investigation of this population. These differences in utilization thus potentially have important ramifications for patients with voice disorders and MH diagnoses, including costs of care, time to diagnosis, specificity and change of diagnosis, and utilization of voice care.12,15,23 Voice care providers need to be vigilant about biases that may influence care patterns.
Future studies will characterize the temporal sequence of mental health and voice diagnoses to identify potential avenues for improving appropriate access to specialty care. In addition, the impact of MH diagnoses on referral patterns will need to be studied to determine whether our findings are attributable to lower rates of specialty referral or to lower rates of presenting for specialty evaluation when referred. It could also be informative to examine how MH diagnoses influence voice-related health care patterns in primary care compared to otolaryngology, which are the two most common provider types in voice care. Such studies could also be extended to other data sets to assess the generalizability of our current findings. The current study is a necessary first step to identify how MH diagnoses may influence voice-related health care utilization.
Conclusions
Using data from a large health care system, we observed a high prevalence of MH diagnoses among patients with voice disorders. We also identified important differences in voice-related diagnoses and health care utilization in voice patients with vs. without MH diagnoses, after accounting for important covariates such as sociodemographics, comorbidities, and voice-related diagnoses. One potential explanation is that the presence of MH diagnoses change how providers treat patients with voice disorders, but further studies are necessary to identify factors that may explain voice-related care differences between those with and without MH diagnoses.
Acknowledgments/Funding
Research reported in this publication was supported by the Health Services Research Grant (to V.J.) from the Centralized Otolaryngology Research Efforts (CORE) of the American Academy of Otolaryngology Head and Neck Surgery Foundation. This research was also supported by the National Center for Advancing Translational Sciences (NIH award number UL1TR000114 to the University of Minnesota Clinical and Translational Science Institute) and the National Institute on Deafness and Other Communication Disorders (K23DC016335 to S.M.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the American Academy of Otolaryngology Head and Neck Surgery Foundation or the NIH.
Additional Funding
Seth Cohen is a consultant for ZSquare and Syneos.
Appendix
Appendix Table 1:
Voice-related and mental health diagnoses, by category and by codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and 10th Revision, Clinical Modification (ICD10-CM)
| voice-related diagnoses | Mental health diagnoses |
|---|---|
| Acute laryngitis | Depression |
| Acute laryngitis without mention of obstruction (464.0, J04.0) | Major depressive disorder, single episode (296.2* F32*) |
| Acute laryngitis with obstruction (464.01, J05.0B | Major depressive disorder, recurrent episode (296.3* F33*) |
| Acute laryngotracheitis without obstruction (464.2, J04.2) | Atypical depressive disorder (296.82, F32.8) |
| Acute laryngotracheitis with obstruction (464.21, J05.0) | Depressive type psychosis (298.0, F32.3/F33.3) |
| Chronic laryngitis | Dysthymic disorder (300.4, F34.1) |
| Chronic laryngitis (476.0, J37.0) | Prolonged depressive reaction (309.1, F43.21) |
| Chronic laryngotracheitis (476.1, J37.1) | Depressive disorder, not elsewhere classified (311* F32.9) |
| Benign neoplasm of the larynx | Anxiety |
| Benign neoplasm of the larynx (212.1, D141) | Anxiety disorder, not otherwise specified (300.0* F41*) |
| Meoplasm of uncertain behavior of the larynx (235.6, D38.0) | Obsessive-compulsive disorder (300.3, F42) |
| Polyp of vocal cord or larynx (478.4, J38.1) | Hypochondriasis (300.7, F45.21/F45.29) |
| Laryngeal cancer | Chronic motor or vocal tic disorder (307.21, F95.0) |
| Malignant neoplasm of the glottis (161.0, C32.0) | Somatoform |
| Malignant neoplasm of the supraglottis (161.1, C32.1) | Somatoform disorder (300.8*, F45.0/F45.1/F45.8/F45.9/F48.8) |
| Malignant neoplasm of the subglottis (161.2, C32.3) | Conversion disorder (300.11, F44.4) |
| Malignant neoplasm of the laryngeal cartilage (161.3, C32.8) | Stress-related disorders (including adjustment disorders) |
| Malignant neoplasm of other specific sites of the larynx (161.8, C32.8) | Adjustment disorder with depressed mood (309.0, F43.21) |
| Malignant neoplasm of the larynx, unspecified (161.9, C32.9) | Adjustment disorder with anxiety (309.24, F43.22) |
| Nonspecific dysphonia | Adjustment disorder with mixed anxiety and depressed mood (309.28, F43.23) |
| Voice and resonance disorder, unspecified (784.4, R49.9) | |
| Aphonia (784.41, R49.1) | Other adjustment reactions with predominant disturbance of other emotions (309.29, F43.29) |
| Dysphonia (784.42, R49.0) | |
| Other voice and resonance disorders (784.49, R49.8) | Unspecified adjustment reaction (309.9, F43.20) |
| Other diseases of vocal cords (478.5, J38.3) | Acute reaction to stress (308* F43.0/R45.7) |
| Edema of larynx (478.6, J38.4) | Reaction to severe stress, adjustment disorder (F43*) |
| Unspecified disease of larynx (478.7, J38.7) | Other unspecified adjustment disorder, including posttraumatic stress disorder (309.8* F43.1/F43.12/F43.8) |
| Cellulitis and perichondritis of larynx (478.71, J38.7) | |
| Other diseases of larynx, not elsewhere classified (478.79, J38.7) | |
| Disorders of the 10th nerve (352.3, G52.2) | |
| Vocal fold paresis/paralysis | |
| Bilateral paralysis of the vocal cords or larynx, complete (478.34, 138.02) | |
| Paralysis of vocal cords or larynx, unspecified (478.30, J38.00) | |
| Unilateral paralysis of the vocal cords or larynx, complete (478.32, 138.01) | |
| Unilateral paralysis of the vocal cords or larynx, partial (478.31, J38.01) | |
| Bilateral paralysis of the vocal cords or larynx, partial (478.33, J38.02) | |
| Larvngeal stenosis | |
| Stenosis of the larynx (478.74, J38.6) | |
| Larvngeal spasm | |
| Laryngeal spasm (478.75, J38.50) |
includes all subdivided categories
Appendix Table 2:
Current Procedural Terminology (CPT) codes for health care utilization
| Radiology: 71010, 71015, 71020, 71021, 71022, 71023, 71030, 71034, 71035, 70450, 70460, 70470, 70490, 70491, 70492, 71260, 71270, 70552, 70553, 70540, 70542, 70543 |
| Swallow study: 74210, 74220, 74230, 92610, 92611, 92612, 92614, 92615, 92616, 92617 |
| Swallow therapy: 92526 |
| Voice evaluation/therapy by speech pathology: 92503, 92506, 92507, 92508, 92520 |
| Laryngoscopy: 31505, 31575, |
| Stroboscopy: 31579 |
| Laryngeal surgery: 31300, 31320, 31360, 31365, 31367, 31368, 31370, 31375, 31380, 31382, 31390, 31395, 31400, 31420, 31500, 31502, 31510, 31511, 31512, 31513, 31515, 31520, 31525, 31526, 31527, 31528, 31529, 31530, 31531, 31536, 31540, 31541, 31545, 31546, 31560, 31561, 31570, 31571, 31576, 31577, 31578, 31580, 31582, 31584, 31587, 31588, 31590, 31595, 31599, 64716, 64886, 31535, 31600, 31603, 31605, 31610, 43030 |
| Chemodenervation: 64613 |
| Laryngeal electromyography: 95865 |
| Reflux testing/evaluation: 91010, 91011, 91012, 91020, 91030, 91034, 91035, 91037, 91038 |
| Radiation therapy: 77263, 77014, 77295, 77290, 77334, 77338, 77300, 77413, 77427, 77336, 77414, 77470, 77301, 77280, 77418 |
Footnotes
Meeting information:
Presented in part at the AAO-HNSF annual meeting, Atlanta, GA, USA, October 7–10, 2018.
References
- 1.Roy N, Merrill RM, Gray SD, Smith EM. Voice disorders in the general population: prevalence, risk factors, and occupational impact. Laryngoscope. 2005;115(11):1988–1995. [DOI] [PubMed] [Google Scholar]
- 2.Dietrich M, Verdolini Abbott K, Gartner-Schmidt J, Rosen CA. The frequency of perceived stress, anxiety, and depression in patients with common pathologies affecting voice. J Voice. 2008;22(4):472–488. [DOI] [PubMed] [Google Scholar]
- 3.Siupsinskiene N, Razbadauskas A, Dubosas L. Psychological distress in patients with benign voice disorders. Folia Phoniatr Logop. 2011;63(6):281–288. [DOI] [PubMed] [Google Scholar]
- 4.Misono S, Peterson CB, Meredith L, et al. Psychosocial distress in patients presenting with voice concerns. J Voice. 2014;28(6):753–761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ruben RJ. Redefining the survival of the fittest: communication disorders in the 21st century. Laryngoscope. 2000;110(2 Pt 1):241–245. [DOI] [PubMed] [Google Scholar]
- 6.Cohen SM, Dupont WD, Courey MS. Quality-of-life impact of non-neoplastic voice disorders: a meta-analysis. Ann Otol Rhinol Laryngol. 2006;115(2):128–134. [DOI] [PubMed] [Google Scholar]
- 7.Mirza N, Ruiz C, Baum ED, Staab JP. The prevalence of major psychiatric pathologies in patients with voice disorders. Ear Nose Throat J. 2003;82(10):808–810, 812, 814. [PubMed] [Google Scholar]
- 8.Willinger U, Völkl-Kernstock S, Aschauer HN. Marked depression and anxiety in patients with functional dysphonia. Psychiatry Res. 2005;134(1):85–91. [DOI] [PubMed] [Google Scholar]
- 9.National Institute of Mental Health: Mental Illness. https://www.nimh.nih.gov/health/statistics/mental-illness.shtml. Accessed March 13, 2018.
- 10.Barsky AJ, Orav EJ, Bates DW. Somatization increases medical utilization and costs independent of psychiatric and medical comorbidity. Arch Gen Psychiatry. 2005;62(8):903–910. [DOI] [PubMed] [Google Scholar]
- 11.Marmor S, Horvath KJ, Lim KO, Misono S. Voice problems and depression among adults in the United States. Laryngoscope. December 2015. doi: 10.1002/lary.25819 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Cohen SM, Kim J, Roy N, Wilk A, Thomas S, Courey M. Change in diagnosis and treatment following specialty voice evaluation: A national database analysis. Laryngoscope. 2015;125(7):1660–1666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–383. [DOI] [PubMed] [Google Scholar]
- 14.Stavem K, Hoel H, Skjaker SA, Haagensen R. Charlson comorbidity index derived from chart review or administrative data: agreement and prediction of mortality in intensive care patients. Clin Epidemiol. 2017;9:311–320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cohen SM, Kim J, Roy N, Asche C, Courey M. Direct health care costs of laryngeal diseases and disorders. Laryngoscope. 2012;122(7):1582–1588. [DOI] [PubMed] [Google Scholar]
- 16.Walker ER, Druss BG. A Public Health Perspective on Mental and Medical Comorbidity. doi: 10.1001/jamapsychiatry.2015.2688 [DOI] [PMC free article] [PubMed]
- 17.Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet. 2012;380(9836):37–43. [DOI] [PubMed] [Google Scholar]
- 18.Misono S, Marmor S, Roy N, Mau T, Cohen SM. Factors Influencing Likelihood of Voice Therapy Attendance. Otolaryngol Head Neck Surg. 2017;156(3):518–524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Jansen L, van Schijndel M, van Waarde J, van Busschbach J. Health-economic outcomes in hospital patients with medical-psychiatric comorbidity: A systematic review and meta-analysis. PLoS One. 2018;13(3):e0194029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Cohen SM, Kim J, Roy N, Courey M. Factors influencing referral of patients with voice disorders from primary care to otolaryngology. Laryngoscope. 2014;124(1):214–220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Townsend L, Walkup JT, Crystal S, Olfson M. A systematic review of validated methods for identifying depression using administrative data. Pharmacoepidemiol Drug Saf. 2012;21 Suppl 1:163–173. [DOI] [PubMed] [Google Scholar]
- 22.Kessler RC, Berglund P, Demler O, et al. The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R). JAMA. 2003;289(23):3095–3105. [DOI] [PubMed] [Google Scholar]
- 23.Cohen SM, Dinan MA, Roy N, Kim J, Courey M. Diagnosis Change in Voice-Disordered Patients Evaluated by Primary Care and/or Otolaryngology. Otolaryngol Head Neck Surg. 2013;150(1):95–102. [DOI] [PubMed] [Google Scholar]