Medialization thyroplasty versus injection laryngoplasty: a cost minimization analysis

Abstract

Background

Medialization thyroplasty and injection laryngoplasty are widely accepted treatment options for unilateral vocal fold paralysis. Although both procedures result in similar clinical outcomes, little is known about the corresponding medical care costs. Medialization thyroplasty requires expensive operating room resources while injection laryngoplasty utilizes outpatient resources but may require repeated procedures. The purpose of this study, therefore, is to quantify the cost differences in adult patients with unilateral vocal fold paralysis undergoing medialization thyroplasty versus injection laryngoplasty.

Study design

Cost minimization analysis conducted using a decision tree model.

Methods

A decision tree model was constructed to capture clinical scenarios for medialization thyroplasty and injection laryngoplasty. Probabilities for various events were obtained from a retrospective cohort from the London Health Sciences Centre, Canada. Costs were derived from the published literature and the London Health Science Centre. All costs were reported in 2014 Canadian dollars. Time horizon was 5 years. The study was conducted from an academic hospital perspective in Canada. Various sensitivity analyses were conducted to assess differences in procedure-specific costs and probabilities of key events.

Results

Sixty-three patients underwent medialization thyroplasty and 41 underwent injection laryngoplasty. Cost of medialization thyroplasty was C$2499.10 per patient whereas those treated with injection laryngoplasty cost C$943.19. Results showed that cost savings with IL were C$1555.91. Deterministic and probabilistic sensitivity analyses suggested cost savings ranged from C$596 to C$3626.

Conclusions

Treatment with injection laryngoplasty results in cost savings of C$1555.91 per patient. Our extensive sensitivity analyses suggest that switching from medialization thyroplasty to injection laryngoplasty will lead to a minimum cost savings of C$596 per patient. Considering the significant cost savings and similar effectiveness, injection laryngoplasty should be strongly considered as a preferred treatment option for patients diagnosed with unilateral vocal fold paralysis.

Background

The aim of treatment for unilateral vocal fold paralysis (UVFP) is, firstly, to decrease aspiration, and secondly, to improve voice quality. One treatment paradigm is medialization of the paralysed vocal fold to allow for contact with the mobile vocal fold. Two options for medialization include Type 1 medialization thyroplasty (MT) and injection laryngoplasty (IL). MT, as described by Isshiki et al., is considered the gold standard treatment and involves permanent medialization of the vocal fold with an alloplastic stent in the paraglottic space [1]. However, with the development of reliable injectable soft tissue fillers and distal chip flexible endoscopes, office-based IL has become a new alternative [2]. An injectable filler, such as methylcellulose, collagen, or calcium hydroxylapatite, is used to medialize the vocal fold via percutaneous or transoral injection.

Since MT and IL utilize the same medialization treatment paradigm, these techniques are often applied in similar clinical scenarios. Outcomes between treatment options have been compared and have been found to yield similar clinical outcomes [3-6]. MT offers a permanent solution but requires operating room time and sedation. In contrast, IL is performed in the outpatient clinic setting. However, the soft tissue fillers are resorbed over time and treatment may require multiple injections over one's lifetime [7].

Given the current fiscal constraints in our health care system, the cost of both procedures should be considered when deciding on the ideal intervention for UVFP. The purpose of this study is therefore to quantify the cost differences between MT and office-based IL in adults with UVFP.

Methods

This economic analysis was conducted from the London Health Sciences Centre (LHSC) perspective-a tertiary academic hospital in Canada. All costs were reported in 2014 Canadian dollars. A 5-year time horizon was used. The discount rate was set at 5% to account for inflation and interest over time [8]. Time to relapse of IL was set to 1 year as the expected lifespan of calcium hydroxylapatite is between 1 to 2 years [4].

Decision analytic model

A decision tree model, consistent with the usual treatment pathways for MT and IL for patients presenting with UVFP, was developed to perform our cost-minimization analysis (Fig. 1). Analysis was conducted using TreeAge Pro 2009 software (TreeAge Software, Inc., Williamstown, MA). [ Table Omitted - see PDF ]

Fig. 1

Decision Tree of Cost Minimization comparing Medialization Thyroplasty and Injection Laryngoplasty. The square represents the initial decision to undergo MT or IL after the identification of UVFP. Circles represent chance events, and triangles represent terminal nodes beyond which no further interventions and costs occurred. One month following the procedure (MT or IL), patients were stratified into 3 groups based on the post-procedural outcomes: voice symptoms (V), voice and swallowing symptoms (V & S), and asymptomatic (Asymp). There were no patients who complained of swallowing symptoms without voice symptoms. Patients with symptoms (V or V & S) after the implementation of initial IL have three possible paths: immediate revision IL (Revision IL), MT (Switch to MT), or observation if the patient was satisfied despite their symptomology (Satisfactory). Due to the temporary nature of the fillers used for IL, patients who were initially satisfied with their treatment despite symptomology could have three possible paths: repeat IL (2nd IL), MT (Switch to MT), or observation if the patient remained satisfied despite their symptoms (Remain Satisfactory). Similarly, patients who were asymptomatic after the initial IL could have three possible paths: relapse after the fillers are resorbed over time and have a repeat IL (2nd IL), undergo MT (Switch to MT), or remain asymptomatic (Remain Asymp). MT Subtree: For patients with symptoms (V or V & S) after MT, there were two possible paths: immediate revision MT (Revision MT), or observation if the patient was satisfied despite their symptomology (Satisfactory). For patients who were asymptomatic after MT, there was no further intervention as of MT is considered permanent

Medialization procedures

MT was performed in the operating room under light sedation and local anesthesia. A nasopharyngoscope was used for visualization of the vocal folds preoperatively and left in place intra-operatively to confirm medialization. The standard thyroplasty approach using the Montgomery® Thyroplasty Implant System (Boston Medical Products, Westborough, MA) was employed, as described by Montgomery and Montgomery [9]. Patients were brought to the post-operative recovery unit following the procedure. Patients were stratified by risk for post-operative complications as described by Zhao et al. and admitted or discharged home accordingly [10]. Reasons for admission include previous neck radiation or surgery, and other comorbidities such as myocardial infarction and stroke.

IL was performed in an outpatient, hospital-based Otolaryngology-Head and Neck Surgery clinic. The patient was positioned in the examination chair in the semi-recumbent position. The cricothyroid approach was most preferred, but transthyrohyoid membrane or transthyroid approaches were also used as indicated. Local anesthesia was infiltrated into the anterior neck soft tissues overlying the cricothyroid membrane and the airway. Under direct vision with the nasopharyngoscope, the needle was advanced to the thyroarytenoid muscle and an injection of calcium hydroxylapatite (RadiesseTM Voice, Merz Aesthetics Inc., San Mateo CA) was completed until the paralyzed vocal fold was medialized. Patient vocalization confirmed medialization. The patient was monitored in the clinic waiting area for at least 30 min, then discharged home.

Probabilities

Probabilities for model parameters were derived from a retrospective cohort of patients from LHSC. Patients were accrued after approval from the institution ethics review board (Institutional Research Ethics Board #105711). Consecutive patients treated by four Otolaryngology-Head and Neck surgeons with MT or IL for UVFP from April 2008 to April 2014 were eligible for inclusion. Other inclusion criteria were: 1) adult over the age of 18, 2) patients were medically eligible for both treatment modalities, and 3) at least one post-procedural follow-up. Patients were excluded if there was any possibility of resolution of UVFP as recovery may be mistaken for success of the procedure and skew probability of resolution of symptoms in the model. These included patients with idiopathic UVFP with a duration of less than 1 year at the time of presentation. Selection of MT versus IL was a joint decision by the surgeon and patient, considering patient preference and surgeon comfort. Date of decision to treat was set as the date of obtaining consent for the procedure.

Patient diagnosis, age, gender, and intervention were recorded. Post-procedural swallowing and voice symptoms were recorded at 1 month. Need for repeat or revision procedure and complications were assessed. Gastrostomy tube dependence rate was recorded where available.

Patient demographics

A total of 228 patients were screened for inclusion via retrospective review. One hundred four patients met the inclusion and exclusion criteria. Sixty-three patients underwent MT whereas 41 underwent IL. Patient demographics and clinical characteristics are reported in Table 1. Patient demographic characteristics were compared using the paired t-test and chi-squared test. [ Table Omitted - see PDF ]

Table 1

Baseline Patient Demographics

Medialization Thyroplasty

Injection Laryngoplasty

Mean Age

61.5

70.4

p?=?0.003a

Males (%)

37 (57.8%)

27 (63.9%)

p?=?0.466

Voice Complaints (%)

63 (100%)

41 (100%)

p?=?1.000

Swallowing Complaints (%)

28 (44.4%)

21 (51.2%)

p?=?0.499

Etiology (N)

Idiopathic (15)

Iatrogenic (23)

Neoplastic (20)

Traumatic (4)

Stroke (1)

Idiopathic (3)

Iatrogenic (13)

Neoplastic (25)

astatistically significant with ? <0.05

Probabilities for medialization thyroplasty

Of the 63 patients initially treated with MT, 53 (84.1%) were asymptomatic at 1 month. Four patients continued to have voice complaints. Six patients reported both voice and swallowing symptoms. Two patients, one with voice complaints and one with voice and swallowing complaints, underwent revision MT. The patient with initial voice complaints was treated successfully with revision MT. However, the patient with initial voice and swallowing complaints remained symptomatic, underwent a second revision surgery, and was subsequently asymptomatic. As a second revision surgery is rare in the clinical setting, this was not included in the base case analysis, but the impact of its inclusion is discussed in the sensitivity analysis. Thirty-eight patients (60.3%) were admitted for overnight stay based on criteria of Zhao et al. [10] All patients undergoing revision MT were admitted due to prior neck surgery.

Probabilities for injection laryngoplasty

Forty-one patients underwent IL. Eighteen patients (43.9%) were asymptomatic at 1 month. Eight patients continued to have voice complaints and 15 patients continued to have voice and swallowing complaints at follow-up. Of these patients, three underwent repeat IL. Two of these patients were subsequently asymptomatic, but one continued to have voice complaints. This patient was consented for MT but passed away prior to the procedure due to a prior lung neoplasm.

Six patients had satisfactory initial results from IL but became symptomatic and underwent subsequent MT. Time from IL to MT was a mean of 18 months. All patients were asymptomatic following MT. No patients were admitted following IL.

Costs

Where available, LHSC costs were utilized. Only direct medical care costs (costs for personnel, equipment, and materials) were considered in the base case analysis to reflect LHSC's perspective. Indirect costs were included in the scenario analysis to provide a societal perspective.

Costs for medialization thyroplasty

Cost for MT consisted of four components: surgeon, anesthesiologist, operating room and recovery personnel, and equipment costs (Table 2). Surgeon and anesthesiologist costs were based on the schedule of benefits published by the Ontario Health Insurance Program (OHIP) [11]. Ontario operates as a single payer system and this government published list consists of uniform fees paid for physician services throughout the province. Operating room and recovery unit personnel costs were calculated by the duration of each procedure at LHSC's rate of C$5.23/min. Material costs were based on the last 12 cases performed at the LHSC as prior data were not available. Total cost of MT was C$1440.11. Revision MT was assumed to cost the same as initial MT. [ Table Omitted - see PDF ]

Table 2

Parameter Table - Base Case

Parameter

Base case

Reference

Probabilities a (%)

Having a voice issue right after the initial MT

6.4

LHSC Datad

Undergoing revision MT for patients with voice issue after the initial MT

25.0

LHSC Datad

Having a voice and swallowing issue right after the initial MT

9.5

LHSC Datad

Undergoing revision MT for patients with voice and swallowing issue after the initial MT

16.7

LHSC Datad

Having a voice issue right after the initial IL

19.5

LHSC Data4

Having a voice and swallowing issue right after the initial IL

36.6

LHSC Datad

Undergoing revision IL for patients with voice issue after the initial IL

37.5

LHSC Datad

Having a voice issue after the revision IL for patients with voice issue after the initial IL

33.3

LHSC Datad

Switching to MT if asymptomatic right after the 1st IL but relapse over time

33.3

LHSC Datad

Admission after MT

60.3

LHSC Datad

Direct costs ($):

?Costs components for MT b

Surgeon

632.85

[11]

Nursing and OR aides

325.38

LHSC Datad

Supplies

316.42

LHSC Datad

Anesthesia

165.46

[11]

??Inpatient stay if admitted after a MT c

1,595.78

LHSC Data4

Costs components for IL

Equipment

CaHa)

305.71

LHSC Datad

Physician

256.11

[11]

Nursing staff

5.42

[12]

Discount rate

5%

[1]

Time to relapse

1 year

[2]

a Only non-zero probabilities are listed. All probabilities for the remaining branches in the model are zero

b Revision MT was assumed to cost the same as initial MT

c Assumed rate of inpatient stay after MT is the same for patients with or without post-surgery symptoms

d Based on London Health Sciences Centre (LHSC) retrospectively collected patient cohort

Inpatient stay following MT was C$1477.58/day based on the mean cost of postoperative stay at LHSC. Mean length of stay was 1.08 days. As 60.3% of all patients undergoing initial MT were admitted, an average inpatient stay cost of C$962.26 was included in the cost of MT. All revision MT patients were inpatients, and an average inpatient stay cost of C$1595.78 was added to the cost of revision MT.

Costs for injection laryngoplasty

The cost for IL consisted of three components: surgeon, personnel, and equipment costs (Table 2). Surgeon costs were determined by the OHIP Schedule of Benefits [11]. Personnel costs included clinic nursing costs. Average hourly wage as per the Ontario Nursing Association agreement was multiplied by the average length of time for injection (4.86 min) [12]. This was added to an average 5 min preparation time for room set up and clean up for a total duration of procedure of 9.86 min. All patients, regardless of undergoing MT or IL, would require initial consultation and consent in the clinic. Therefore, this cost was considered equal between both groups and not included in the cost calculations. Equipment costs were based on the institutional pharmacy and the healthcare materials management services department. Cost for the nasopharyngoscope and processing were obtained from the hospital and included processing time which was timed at an average of 6.5 min, multiplied by an average technician wage of C$22/h [13]. Following the procedure, all patients recovered in the clinic waiting area. This was during an active clinic and did not require any special personnel or equipment. Therefore, no cost was associated with patient recovery time. Total cost for an IL was C$567.24.

Complications

One patient suffered from a postoperative hematoma following MT. This required a return visit to the operating room. While in the hospital, this patient experienced urinary retention requiring a longer hospital stay of 4 days. There were no complications following IL. Complication probabilities were set to zeros in the model in order to avoid undue influence of one special case in the MT arm.

Sensitivity analysis

Deterministic and probabilistic sensitivity analyses were conducted to account for variability in the input parameters presented in Tables 3 and 4. One-way sensitivity analyses were conducted for all costs with a known range, for a discount rate of 3 and 10%, an IL effect duration of 2 years and the percent of patients undergoing repeat IL (Table 4). [ Table Omitted - see PDF ]

Table 3

Probabilities for Sensitivity Analyses a

Probabilities b

Base case value (%)

Parameters for Beta distributions for PSAc

? d

? e

Having a voice issue right after the initial MT

6.4

4

59

Undergoing revision MT for patients with voice issue after the initial MT

25.0

1

3

Having a voice and swallowing issue right after the initial MT

9.5

6

57

Undergoing revision MT for patients with voice and swallowing issue after the initial MT

16.7

1

5

Having a voice issue right after the initial IL

19.5

8

33

Having a voice and swallowing issue right after the initial IL

36.6

15

26

Undergoing revision IL for patients with voice issue after the initial IL

37.5

3

5

Having a voice issue after the revision IL for patients with voice issue after the initial IL

33.3

1

2

Switching to MT if asymptomatic right after the 1st IL but relapse over time

33.3

6

12

Admission after MT

60.3

38

25

Having a voice issue right after the initial MT

4.3 [15]

a All probabilities are derived from the LHSC data set except the probability of having voice issue right after the initial MT

b Only non-zero probabilities are listed. All probabilities for the remaining branches in the model are zero

c PSA: probabilistic sensitivity analysis

d The parameter ? for beta distribution equals the number of occurrence

e The parameter ? for beta distribution equals the difference of sample size and number of occurrence

[ Table Omitted - see PDF ]

Table 4

Parameters used for Sensitivity and Scenario Analysis

Other Parameters

Base Case

Low Value

High Value

References

Isshiki N. Progress in laryngeal framework surgery. Acta Otolaryngol (Stockh). 2000;120(2):120-7.Google Scholar

Misono S, Merati AL. Evidence-based practice: evaluation and management of unilateral vocal fold paralysis. Otolaryngol Clin North Am. 2012;45(5):1083-108.Google Scholar

Siu J, Tam S, Fung K. A comparison of outcomes in interventions for unilateral vocal fold paralysis: A systematic review. Laryngoscope. 2016;126(7):1616-24.Google Scholar

Shen T, Damrose EJ, Morzaria S. A meta-analysis of voice outcome comparing calcium hydroxylapatite injection laryngoplasty to silicone thyroplasty. Otolaryngol-Head Neck Surg. 2013;148(2):197-208.Google Scholar

Morgan JE, Zraick RI, Griffin AW, Bowen TL, Johnson FL. Injection versus medialization laryngoplasty for the treatment of unilateral vocal fold paralysis. Laryngoscope. 2007;117(11):2068-74.Google Scholar

Vinson KN, Zraick RI, Ragland FJ. Injection versus medialization laryngoplasty for the treatment of unilateral vocal fold paralysis: follow-up at six months. Laryngoscope. 2010;120(9):1802-7.Google Scholar

Upton DC, Johnson M, Zelazny SK, Dailey SH. Prospective evaluation of office-based injection laryngoplasty with hyaluronic acid gel. Ann Otol Rhinol Laryngol. 2013;122(9):541-6.Google Scholar

Guidelines for the Economic Evaluation of Health Technologies: Canada. 3rd Ed. Ottawa: Canadian Agency for Drugs and Technologies in Health (CADTH); 2006.

Montgomery WW, Montgomery SK. Montgomery thyroplasty implant system. Ann Otol Rhinol Laryngol Suppl. 1997;170:1-16.Google Scholar

Zhao X, Roth K, Fung K. Type I thyroplasty: risk stratification approach to inpatient versus outpatient postoperative management. J Otolaryngol-Head Neck Surg. 2010;39(6):757-61.Google Scholar

Ministry of Health Long Term Care. Schedule of benefits for physician services under the Health Insurance Act. 2007. Accessed 28 August 2015.Google Scholar

Ontario Nurses Association. Hospital Central Collective Agreement (English) | Expiry 2016-03-31. 2016. https://www.ona.org/wpcontent/uploads/ona_hospitalcentralagreement_20180331.pdf?x72008. Accessed 17 Feb 2017.

Chin CJ, Sowerby LJ, John-Baptiste A, Rotenberg BW. Reducing otolaryngology surgical inefficiency via assessment of tray redundancy. J Otolaryngol-Head Neck Surg. 2014;43:46.Google Scholar

Statistics Canada. Average hourly wages of employees by selected characteristics and occupation, unadjusted data, by province (monthly) (Ontario). http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/labr69g-eng.htm. Accessed 17 Feb 2017.

Roy N, Merrill RM, Gray SD, Smith EM. Voice disorders in the general population: prevalence, risk factors, and occupational impact. Laryngoscope. 2005;115(11):1988-95.Google Scholar

Hillel AT, Ochsner MC, Johns MM, Klein AM. A cost and time analysis of laryngology procedures in the endoscopy suite versus the operating room. Laryngoscope. 2016;126(6):1385-9.Google Scholar

Rosen CA, Gartner-Schmidt J, Casiano R, et al. Vocal fold augmentation with calcium hydroxylapatite: twelve-month report. Laryngoscope. 2009;119(5):1033-41.Google Scholar

Prendes BL, Yung KC, Likhterov I, Schneider SL, Al-Jurf SA, Courey MS. Long-term effects of injection laryngoplasty with a temporary agent on voice quality and vocal fold position. Laryngoscope. 2012;122(10):2227-33.Google Scholar

Abraham MT, Gonen M, Kraus DH. Complications of type I thyroplasty and arytenoid adduction. Laryngoscope. 2001;111(8):1322-9.Google Scholar

Mathison CC, Villari CR, Klein AM, Johns MM. Comparison of outcomes and complications between awake and asleep injection laryngoplasty: a case-control study. Laryngoscope. 2009;119(7):1417-23.Google Scholar