Real-World Evidence of Treatment Patterns and Costs of Turner Syndrome and Noonan Syndrome in the USA

Abstract

Introduction

Short stature is a common clinical feature of Turner syndrome (TS) and Noonan syndrome (NS). Growth hormone (GH) treatment increases height in patients with TS and NS. We aimed to assess treatment patterns and costs in a real-world setting of GH-treated patients with TS and NS in the USA.

Methods

Patients with TS (aged 4 to <14 years) or NS (aged ≤18 years) were included in this retrospective analysis using the Komodo Health claims database. The study period from January 1, 2016, to September 30, 2022, included ≥6 months pre-index (baseline) and ≥1 year post-index (follow-up); the index date was the first (TS) or latest of two visits (NS) with diagnostic coding for the condition. Data were analyzed descriptively.

Results

Among 2,530 patients with TS (mean age [SD] 8.4 [3.0] years) and 1,119 patients with NS (mean age 7.0 [4.4] years), 36% and 16% initiated GH treatment during follow-up, respectively. Of these patients, 48%/41% were early initiators (started ≤6 months post-index), 66%/70% had high adherence (≥80% of GH treatment days covered), and 56%/52% had long persistence (treated ≥2 years). Mean GH treatment duration was 24.9 (15.9) and 17.6 (10.1) months for patients with TS and NS, respectively. Median (IQR) total all-cause costs per patient/year were USD 10,232 (USD 2,850–USD 36,590) and USD 17,937 (USD 5,616–USD 47,552), respectively.

Conclusion

Of the few patients on GH therapy, less than half were initiated on treatment early after their TS or NS diagnosis. Greater awareness regarding early diagnosis and treatment of TS and NS is needed.

Plain Language Summary

Turner syndrome (TS) and Noonan syndrome (NS) are common genetic conditions. Individuals with TS and NS tend to be shorter than other people. Treatment with growth hormone can help children with TS and NS reach a near-average adult height, especially if they receive the treatment early in their diagnosis and continue to take it. In this study, we looked at how children with TS and NS are treated for short height in the USA. We also assessed the costs associated with each condition. The patients included in this study were from a database of health insurance records between 2016 and 2022. We looked at 2,530 patients with TS aged 4 to 13 years and 1,119 with NS under age 18. We found that only 36% of patients with TS and 16% of those with NS had started growth hormone treatment. About half of the patients began treatment within 6 months of their diagnosis. Most patients took the growth hormone treatment as prescribed, but only about half stayed on the medication for at least 2 years. Yearly healthcare costs (including medication) were approximately USD 10,000 for patients with TS and USD 18,000 for patients with NS. Although growth hormone treatment can be effective, it is not widely used for these patients. The study suggests that healthcare providers and families need more awareness of treatment options to help children with TS and NS reach a more typical adult height.

Introduction

Turner syndrome (TS) and Noonan syndrome (NS) are common pediatric nongrowth hormone (GH) deficiency (GHD) conditions. Caused by a complete or partial loss of the second sex chromosome, TS affects approximately 1 in 2,500 female births [1]. Diagnosis of TS has a multimodal age distribution, and peak times of diagnosis are birth, adolescence, and later adulthood [2]. Caused by genetic mutations, NS has an estimated incidence of 1 in 1,000–2,500 live births [3]. Many features of NS are present at birth or emerge in early childhood, and others emerge throughout adulthood [4].

Short stature is one of the most common clinical features of these genetic conditions, occurring in as many as 95% of females with TS; prevalence estimates of short stature in individuals with NS are more variable, ranging from 50% to 80% [5–7]. Both syndromes also result in characteristic facial features, cardiovascular anomalies, and multiple comorbidities [5, 8]. Approximately 20% of girls with TS have a birth weight at or below −2 standard deviations (SDs) [9]. Although the birth weight and height of patients with NS are typically average, they experience growth deceleration in the first few years of life [10].

Treatment with daily recombinant human GH for short stature in TS and NS was approved by the US Food and Drug Administration in 1996 and 2007, respectively [11]. Early initiation and continued GH treatment substantially increase near-final adult height for both TS and NS [12–21]. Clinical practice guidelines include GH therapy for short stature in both syndromes [6, 22, 23]. In TS, it is recommended that GH treatment be started early because of rapid growth failure in the first few years of life; guidelines suggest beginning treatment at the first evidence of growth failure or short stature, as early as 2 years of age, to prevent further loss of potential height [22]. GH treatment has also been shown to be safe and effective in increasing height in patients with NS [24]. However, little literature has been published regarding real-world treatment patterns, adherence, and persistence of GH therapy in TS and NS [25–29]. Furthermore, healthcare costs and healthcare resource utilization (HCRU) for these conditions are unknown. We aimed to (1) understand real-world treatment patterns of incident GH-treated pediatric patients with TS or NS in the USA, including GH treatment initiation, time-to-treatment initiation, adherence, and persistence; (2) assess baseline characteristics associated with early treatment initiation, high adherence, and long persistence; and (3) explore all-cause HCRU and costs among patients with TS or NS by GH treatment initiation status.

Methods

Study Design

This was a descriptive, non-interventional, retrospective cohort study of medical and pharmacy claims data of patients with TS or NS using the Komodo Health claims database. The study period was from January 1, 2016, to September 30, 2022. We chose January 1, 2016, as the start date to represent recent advancements in care and therapeutic options for pediatric non-GHD conditions. Treatment guidelines and patterns, clinical management and diagnosis, newer therapies, HCRU, and costs may be less relevant with older data. The end date was selected to exclude a period of GH shortage in the USA starting October 1, 2022; the shortage may have affected treatment initiation, adherence, and persistence and caused changes in treatment patterns.

The patient identification period spanned July 1, 2016, through September 30, 2021 (Fig. 1). The index date (proxy for diagnosis date) was defined as the earliest date during the patient identification period corresponding to the date of the earliest qualifying claim, defined as the first visit or the latest of two visits with diagnostic coding indicating TS or NS, respectively. The baseline period was defined as the 6 months before the index date, allowing for the ascertainment of pre-diagnosis patient characteristics, HCRU, and costs. The follow-up period was defined as the minimum 1-year period after the index date to assess post-diagnosis treatment patterns and all-cause HCRU/costs. Follow-up for each patient began on the index date and ended at the end of the eligible insurance coverage period or on September 30, 2022, whichever came first; each patient had a variable follow-up time.

Fig. 1.

Patient identification schematic.

GH treatment was defined as a period of at least 6 months of GH therapy using pharmacy claims, which have a “days’ supply” value that was used to determine the time coverage for the pharmacy refill. Pharmacy claims for GH were identified by National Drug Code-11 (NDC) codes and medical encounters with GH administration by the Healthcare Common Procedure Coding System codes. GH treatment initiation was considered to be the first GH treatment period lasting at least 6 months post-index. Discontinuation of GH treatment was defined as 90 days or more from the end of the “days’ supply” of a GH refill claim without a new GH claim. Duration of GH treatment refers to the first GH treatment period lasting at least 6 months post-index and was defined as the last GH refill claim date of service, plus the “days’ supply,” minus the first GH fill claim date of service.

Data Source

Komodo Health is a data warehouse that acquires de-identified medical and prescription claims data on over 330 million patients from various sources across the USA, including hospital networks, physician networks, healthcare claim processing companies (i.e., claims clearinghouses), pharmacies, and health insurers (many private and some public, including Medicaid managed-care and Medicare Advantage plans). Available to researchers for generating real-world evidence, the administrative and claims dataset includes basic patient demographics, pharmacy claims, and visit claims; diagnoses and procedures from healthcare visits are also collected. Key components include information on insurance enrollment (plan, start, and end dates), medical claims data (date of service, admission, and discharge [for hospitalizations]; diagnosis code; procedure code; flag for emergency department [ED] visits; cost), and pharmacy claims data (11-digit NDC, fill date, days’ supply, cost). Allowed amounts (maximum amount health plans pay for a covered service) paid by insurers (cost calculated as the total of insurance and patient payments) are reported for each claim. Komodo Health imputes or adjusts missing or outlier values based on proprietary methods.

Patient Population

Two patient cohorts were included in this study. Patients with TS or NS were identified based on pre-specified definitions using International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) diagnosis codes. Patients were included in the study if they had continuous enrollment (uninterrupted record of healthcare encounters) and closed (complete/paid patient encounters) medical and pharmacy claims for 6 months pre- and at least 1 year post-index date, representing an approximate eligibility date for GH treatment initiation. Patients were excluded if they had both TS and NS, were missing the birth year, or had GH claims data before the index date.

TS was defined as evidence of the ICD-10-CM diagnosis code for TS (Q96.x) in at least two outpatient visits at least 30 days apart or one inpatient stay code at any position during the study period. The timing of the first diagnosis for TS during the study period determined the index date. If the first diagnosis for TS was after age 4 years, the index date was the date of diagnosis. If the first diagnosis for TS was before age 4, the index date was when the child turned 4 as many clinicians wait until age 4 to initiate GH therapy in these patients. The database only contains the birth year; therefore, we arbitrarily assigned the month and day during the patient’s birth year to the month and day of the diagnosis date. Because GH treatment may not be initiated for patients diagnosed after age 13, we only included patients with TS who were aged <14 years at the index date.

NS was defined as evidence of the ICD-10-CM diagnosis code for “other congenital malformation syndromes predominantly associated with short stature” (Q87.19) in at least two outpatient visits at least 30 days apart or one inpatient stay code at any position. We also required evidence of “short stature” (R62.52) at any time before or after diagnosis. The index date was the latest of the two diagnosis codes. Because the designated code is not specific to NS, we also required the presence of other characteristics hierarchically. First, at least one cardiac characteristic commonly associated with NS (e.g., pulmonary stenosis, hypertrophic cardiomyopathy, or congenital heart defect) was required at any time before the index date and for up to 2 years post-index date. For patients without NS-specific cardiac diagnoses, we required the presence of at least two characteristics listed as occurring very frequently in patients with NS on the Orphanet rare disease reference Website during the baseline and follow-up periods (online suppl. Table S1; for all online suppl. material, see https://doi.org/10.1159/000550211).

Outcomes

The primary outcome was initiation of GH treatment (yes/no for each patient). For patients who initiated GH treatment, time to initiation of GH treatment, duration of GH treatment, adherence to GH treatment, and medication persistence were assessed. Early treatment initiation was defined as initiation of GH treatment (lasting at least 6 months) within 6 months of the index date. Adherence to GH treatment was calculated based on the proportion of days covered (PDC) method, which is the proportion of days that a patient has access to their medication over a period of interest [30–32]. As commonly described in the literature using PDC, we considered high adherence to be a PDC ≥0.80 [33, 34]. Medication persistence was calculated based on the duration of time from GH treatment initiation to GH treatment discontinuation. Specifically, long persistence was defined as continuation of treatment for ≥2 years from the index date. Patients with follow-up duration between 1 and 2 years, with persistence during their shorter available follow-up (censored), were also defined as having long persistence.

All-cause HCRU per patient per year (PPPY), including inpatient hospitalizations (any admission, number of admissions, length of stay), outpatient visits (any visits, number of visits), ED visits (any visits, number of visits), and all-cause costs PPPY (total, inpatient, outpatient, ED, pharmacy), were assessed for each pediatric non-GHD condition in total and by GH treatment status. To better understand the study population, the number of documented cases of TS and NS as defined by the study criteria (removing any known deaths before January 1, 2023) was estimated from January 1, 2016, to December 31, 2022.

Statistical Analysis

Descriptive statistics at baseline and follow-up were conducted and included measures of central tendency (mean, median) and dispersion (range, interquartile range [IQR], and SD) for continuous variables. For categorical variables, frequencies and proportions were calculated. Descriptive statistics for rates were presented as the number of visits per year; costs were presented as the amount in US dollars (USD) PPPY. Data analysis was conducted using SAS 9.4 (SAS, Cary, NC, USA), RStudio 2022.07.0 Build 548 (Posit PBC, Boston, MA, USA), R version 4.2.1 (R Foundation, Vienna, Austria), Snowflake 8.16.10 (Snowflake Inc., Bozeman, MT, USA), and Visual Studio Code 1.88.1 (Microsoft, Redmond, WA, USA).

Annualized rates of HCRU and annualized costs were calculated, including inpatient, outpatient, ED visits, and medication pharmacy claims. Annualization of healthcare utilization and costs was performed based on 6 months of baseline claims and separately for the years of follow-up during which each patient had claims. Descriptive statistics for rates and costs were only summarized among patients with at least one claim of that type. Costs were inflation adjusted annually to 2023 based on the year each claim was submitted [35]. Extreme cost values for outpatient visits over 1 million dollars (after inflation adjustment) were winsorized to 1-million-dollar costs to mitigate the effect of potential cost errors. Only pharmacy claims indicated as “paid” were included in cost calculations.

Subgroup analyses were conducted for (1) early vs. late GH treatment initiation relative to initial diagnosis date (GH treatment started ≤6 vs. >6 months post-index); (2) high vs. low adherence (≥80% vs. <80% PDC); and (3) long vs. short persistence (≥2 years vs. <2 years) of continuous GH treatment. Statistical significance was determined using chi-square tests for categorical covariates and two-sided t tests for continuous covariates.

Results

Patient Population and Baseline Characteristics

Based on the study definitions of each pediatric non-GHD condition, as of January 1, 2023, the estimated prevalence was 36 per 100,000 females for TS and 5 per 100,000 individuals for NS. Of the 1,795,280 patients in the Komodo Health claims database with any GH-related diagnosis code or pharmacy claim, 10,345 (0.6%) had TS and 3,715 (0.2%) had NS (online suppl. Tables S2, S3). A total of 5,089 (49.2%) patients with TS and 2,317 (62.4%) with NS had continuous closed medical and pharmacy claims from a commercial insurer for 6 months pre- and at least 1 year post-index date. Among the patients with TS, 2,820 (27.3%) met all other inclusion and exclusion criteria; the study included 2,530 (24.5%) of these patients who were aged <14 years at the index date. Among those with NS, 1,605 (43.2%) had other conditions commonly associated with NS; the study included 1,119 (30.1%) patients meeting all other inclusion and exclusion criteria.

The mean age at index date was 8.4 and 7.0 years for patients with TS and NS, respectively (Table 1). Of the patients with TS, 372 (14.7%) were diagnosed before the age of 4, and their index date was adjusted to age 4. Among patients with TS, 37.5% were white, 10.9% were black, and 24.2% were Hispanic or Latino; race/ethnicity was similar for those with NS – 38.4% white, 6.7% black, and 19.7% Hispanic or Latino (Table 1). The majority of patients with each pediatric non-GHD condition had Medicaid insurance. The most common comorbid conditions were hearing loss (10.8%) and attention-deficit/hyperactivity disorder (7.4%) for patients with TS and congenital heart disease (25.4%), hearing loss (14.2%), asthma (11.4%), and attention-deficit/hyperactivity disorder (8.5%) for patients with NS.

Table 1.

Baseline (index date) characteristics and follow-up duration by initiation of GH treatment for patients with Turner and Noonan syndromes

Characteristics	Patients with TS				Patients with NS
	total (N = 2,530)	GH treatment initiated (n = 902)	GH treatment not initiated (n = 1,628)	p valuea	total (N = 1,119)	GH treatment initiated (n = 176)	GH treatment not initiated (n = 943)	p valuea
GH treatment duration, n (%)	​	​	​	NA	​	​	​	NA
Treated for ≥6 months	902 (35.7)	902 (100.0)	0 (0.0)	​	176 (15.7)	176 (100.0)	0 (0.0)	​
Treated for <6 months	169 (6.7)	0 (0.0)	169 (10.4)	​	57 (5.1)	0 (0.0)	57 (6.0)	​
No GH treatment	1,459 (57.7)	0 (0.0)	1,459 (89.6)	​	886 (79.2)	0 (0.0)	886 (94.0)	​
Age, years	​	​	​	<0.001	​	​	​	<0.001
Mean (SD)	8.39 (3.03)	8.85 (2.80)	8.13 (3.12)	​	7.05 (4.36)	6.05 (3.96)	7.23 (4.41)	​
Median (IQR)	8.54 (5.34–10.91)	9.07 (6.53–11.09)	7.85 (4.86–10.74)	​	6.18 (3.28–10.35)	4.82 (2.56–9.54)	6.30 (3.51–10.62)	​
Min, max	4.00, 14.00	4.01, 14.00	4.00, 13.98	​	1.00, 17.53	1.01, 15.95	1.00, 17.53	​
Sex, n (%)	​	​	​	NA	​	​	​	0.142
Male	0 (0.0)	0 (0.0)	0 (0.0)	​	601 (53.7)	83 (47.2)	518 (54.9)	​
Female	2,530 (100.0)	902 (100.0)	1,628 (100.0)	​	510 (45.6)	91 (51.7)	419 (44.4)	​
Unknown	0 (0.0)	0 (0.0)	0 (0.0)	​	8 (0.7)	2 (1.1)	6 (0.6)	​
Race/ethnicity, n (%)	​	​	​	<0.001	​	​	​	0.241
White	948 (37.5)	362 (40.1)	586 (36.0)	​	430 (38.4)	76 (43.2)	354 (37.5)	​
Black	276 (10.9)	72 (8.0)	204 (12.5)	​	75 (6.7)	8 (4.5)	67 (7.1)	​
Hispanic or Latino	611 (24.2)	243 (26.9)	368 (22.6)	​	221 (19.7)	26 (14.8)	195 (20.7)	​
Asian	80 (3.2)	31 (3.4)	49 (3.0)	​	33 (2.9)	4 (2.3)	29 (3.1)	​
Other	97 (3.8)	42 (4.7)	55 (3.4)	​	43 (3.8)	9 (5.1)	34 (3.6)	​
Unknown	518 (20.5)	152 (16.9)	366 (22.5)	​	317 (28.3)	53 (30.1)	264 (28.0)	​
Geographic region, n (%)	​	​	​	0.570	​	​	​	0.325
Midwest	512 (20.2)	178 (19.7)	334 (20.5)	​	244 (21.8)	41 (23.3)	203 (21.5)	​
Northeast	406 (16.0)	149 (16.5)	257 (15.8)	​	226 (20.2)	40 (22.7)	186 (19.7)	​
South	1,050 (41.5)	389 (43.1)	661 (40.6)	​	436 (39.0)	70 (39.8)	366 (38.8)	​
West	554 (21.9)	183 (20.3)	371 (22.8)	​	213 (19.0)	25 (14.2)	188 (19.9)	​
Unknown	8 (0.3)	3 (0.3)	5 (0.3)	​	0 (0.0)	0 (0.0)	0 (0.0)	​
Medical insurance type, n (%)	​	​	​	0.020	​	​	​	0.130
Commercial (private)	875 (34.6)	295 (32.7)	580 (35.6)	​	379 (33.9)	69 (39.2)	310 (32.9)	​
Medicare (public)	16 (0.6)	6 (0.7)	10 (0.6)	​	7 (0.6)	2 (1.1)	5 (0.5)	​
Medicaid (public)	1,621 (64.1)	600 (66.5)	1,021 (62.7)	​	731 (65.3)	104 (59.1)	627 (66.5)	​
Unknown	18 (0.7)	1 (0.1)	17 (1.0)	​	2 (0.2)	1 (0.6)	1 (0.1)	​
Comorbidities, n (%)	​	​	​	​	​	​	​	​
Congenital heart disease	120 (4.7)	30 (3.3)	90 (5.5)	0.013	284 (25.4)	47 (26.7)	237 (25.1)	0.660
ADHD	188 (7.4)	74 (8.2)	114 (7.0)	0.270	95 (8.5)	14 (8.0)	81 (8.6)	0.781
Hearing loss	272 (10.8)	100 (11.1)	172 (10.6)	0.685	159 (14.2)	15 (8.5)	144 (15.3)	0.019
Lymphedema	16 (0.6)	3 (0.3)	13 (0.8)	0.157	5 (0.4)	2 (1.1)	3 (0.3)	0.135
Hyperlipidemia	28 (1.1)	15 (1.7)	13 (0.8)	0.047	6 (0.5)	0 (0.0)	6 (0.6)	0.289
Asthma	139 (5.5)	39 (4.3)	100 (6.1)	0.054	128 (11.4)	15 (8.5)	113 (12.0)	0.185
Precocious puberty	19 (0.8)	8 (0.9)	11 (0.7)	0.556	11 (1.0)	4 (2.3)	7 (0.7)	0.059
Body mass index, n (%)	​	​	​	0.794	​	​	​	0.252
Normal (≤24.9 kg/m2)	23 (0.9)	7 (0.8)	16 (1.0)	​	7 (0.6)	0 (0.0)	7 (0.7)	​
Overweight (25.0–29.9 kg/m2)	4 (0.2)	1 (0.1)	3 (0.2)	​	0 (0.0)	0 (0.0)	0 (0.0)	​
Obesity (≥30 kg/m2)	1 (0.0)	0 (0.0)	1 (0.1)	​	0 (0.0)	0 (0.0)	0 (0.0)	​
Unknown	2,502 (98.9)	894 (99.1)	1,608 (98.8)	​	1,112 (99.4)	176 (100.0)	936 (99.3)	​
Systemic steroid treatment, n (%)	162 (6.4)	54 (6.0)	108 (6.6)	0.524	128 (11.4)	17 (9.7)	111 (11.8)	0.419
Estrogen treatment, n (%)	40 (1.6)	6 (0.7)	34 (2.1)	0.006	1 (0.1)	0 (0.0)	1 (0.1)	0.666
Oxandrolone treatment, n (%)	6 (0.2)	3 (0.3)	3 (0.2)	0.463	2 (0.2)	0 (0.0)	2 (0.2)	0.541
Follow-up duration, years	​	​	​	<0.001	​	​	​	<0.001
Mean (SD)	3.58 (1.65)	3.97 (1.60)	3.37 (1.64)	​	2.24 (0.91)	2.45 (1.02)	2.20 (0.89)	​
Median (IQR)	3.39 (2.11–5.09)	4.05 (2.57–5.49)	3.15 (1.87–4.79)	​	2.10 (1.55–2.71)	2.32 (1.72–2.91)	2.06 (1.54–2.67)	​
Min, max	1.00, 6.25	1.02, 6.25	1.00, 6.25	​	1.01, 6.22	1.01, 6.22	1.01, 6.09	​

GH treatment initiation is defined as at least 6 months of continuous GH treatment post-index.

GH, growth hormone; IQR, interquartile range; NA, not applicable; SD, standard deviation; ADHD, attention-deficit/hyperactivity disorder.

^aChi-square test for categorical and two-sided t test for continuous variables.

Treatment Patterns and Patient Characteristics

Of the 2,530 patients with TS, 902 (35.7%) initiated GH treatment during the follow-up period; 1,459 (57.7%) had no follow-up GH claims; and 169 (6.7%) had claims for <6 months’ duration. Among the GH-initiated patients, 823 (91.2%) had recorded ICD-10-CM codes for “short stature,” compared with 888 (54.5%) of the GH-non-initiated patients (n = 1,628). Mean (SD) duration of follow-up was 4.0 (1.6) years for GH-initiated patients and 3.4 (1.6) years for GH-non-initiated patients (Table 1). The mean age at index date was 8.9 years and 8.1 years for GH-initiated and GH-non-initiated patients, respectively. A larger proportion of GH-initiated than GH-non-initiated patients were white (40.1% vs. 36.0%) or were on Medicaid (66.5% vs. 62.7%); more GH-non-initiated patients had congenital heart disease (5.5% vs. 3.3%).

Of the 1,119 patients with NS, 176 (15.7%) initiated GH treatment during their follow-up after the index date; 886 (79.2%) had no follow-up GH claims; and 57 (5.1%) had claims for <6 months’ duration; mean (SD) for duration of follow-up was 2.5 (1.0) years for GH-initiated patients and 2.2 (0.9) years for GH-non-initiated patients (Table 1). GH-non-initiated patients were older than GH-initiated patients (mean age at index date, 7.2 years vs. 6.1 years). A greater percentage of GH-initiated patients than GH-non-initiated patients were female (51.7% vs. 44.4%) and white (43.2% vs. 37.5%); more GH-non-initiated patients were on Medicaid (66.5% vs. 59.1%) and had hearing loss (15.3% vs. 8.5%) or asthma (12.0% vs. 8.5%).

For patients with TS, the overall mean (SD) time to GH treatment initiation was 11.5 (13.0) months from the index date (median [IQR], 6.46 [1.97–16.71] months) (Table 2). Nearly half (48.0%) of patients with TS were early initiators. Time-to-treatment initiation for early and late initiators was a mean of 2.4 and 19.9 months; mean duration of GH treatment overall was 24.9 months, longer for the early than the late initiators (28.3 vs. 21.8 months). The mean (SD) time to GH treatment initiation for patients with NS was 9.3 (8.0) months from index (median [IQR], 7.24 [2.76–13.27] months) (Table 2); 41.5% of patients with NS were early initiators. Mean time to initiation was 2.6 and 14.0 months for early and late initiators, and duration of GH treatment was 17.6 months overall: 20.8 and 15.3 months for early and late initiators, respectively.

Table 2.

GH treatment characteristics for patients with Turner and Noonan syndromes by treatment initiation timing

Characteristics	Patients with TS and GH treatment initiation			Patients with NS and GH treatment initiation
	total (N = 902)	early GH treatment initiation (n = 433)	late GH treatment initiation (n = 469)	total (N = 176)	early GH treatment initiation (n = 73)	late GH treatment initiation (n = 103)
% of N	100.0	48.0	52.0	100.0	41.5	58.5
Time to GH treatment initiation, months
Mean (SD)	11.49 (12.96)	2.42 (1.78)	19.87 (13.19)	9.28 (8.03)	2.59 (1.64)	14.02 (7.35)
Median (IQR)	6.46 (1.97–16.71)	1.91 (0.89–3.85)	16.09 (9.38–24.93)	7.24 (2.76–13.27)	2.47 (1.25–3.68)	12.34 (8.22–16.25)
Min, max	0.00, 64.77	0.00, 6.02	6.05, 64.77	0.10, 45.59	0.10, 5.95	6.05, 45.59
GH initiation duration, months
Mean (SD)	24.95 (15.95)	28.35 (16.71)	21.81 (14.53)	17.61 (10.14)	20.84 (10.59)	15.33 (9.19)
Median (IQR)	20.64 (11.74–35.33)	23.68 (14.31–38.85)	17.20 (9.84–30.76)	15.10 (9.34–22.29)	19.05 (11.51–27.34)	12.70 (8.62–18.95)
Min, max	5.95, 72.14	5.99, 72.14	5.95, 68.26	6.09, 51.09	6.38, 51.09	6.09, 50.56
PDC adherence to GH treatment
Mean (SD)	0.83 (0.12)	0.84 (0.11)	0.82 (0.12)	0.85 (0.11)	0.87 (0.09)	0.83 (0.13)
Median (IQR)	0.86 (0.76–0.92)	0.86 (0.77–0.92)	0.85 (0.75–0.92)	0.88 (0.78–0.93)	0.89 (0.82–0.93)	0.86 (0.77–0.93)
Min, max	0.38, 1.00	0.39, 1.00	0.38, 1.00	0.46, 1.00	0.60, 1.00	0.46, 1.00

GH treatment initiation is defined as at least 6 months of continuous GH treatment post-index. Early initiation is the start of GH treatment within 6 months post-index.

GH, growth hormone; IQR, interquartile range; PDC, proportion of days covered; SD, standard deviation.

Baseline characteristics significantly associated with early treatment initiation for patients with TS were age at diagnosis and medical insurance type (Table 3); early initiators were older at diagnosis than late initiators (mean 9.5 vs. 8.3 years, p < 0.001), were more likely to have commercial insurance, and were less likely to have Medicaid (p = 0.047). Age at diagnosis was significantly associated with early initiation of GH among patients with NS; at diagnosis, early initiators were older than late initiators (mean 6.8 vs. 5.5 years, p = 0.032) (Table 3). Although precocious puberty was also significantly associated with early GH initiation, the sample size was too small to interpret (n = 4).

Table 3.

Baseline characteristics for patients with Turner and Noonan syndromes by treatment initiation timing

Characteristics	Patients with TS and GH treatment initiation			Patients with NS and GH treatment initiation
	early GH treatment initiation (n = 433)	late GH treatment initiation (n = 469)	p valuea	early GH treatment initiation (n = 73)	late GH treatment initiation (n = 103)	p valuea
% of N	48.0	52.0	​	41.5	58.5	​
Age, years	​	​	<0.001	​	​	0.032
Mean (SD)	9.51 (2.71)	8.25 (2.76)	​	6.81 (4.23)	5.51 (3.69)	​
Median (IQR)	9.67 (7.60–10.52)	8.53 (5.56–10.52)	​	6.52 (2.99–8.48)	4.44 (2.32–8.48)	​
Min, max	4.01, 14.00	4.01, 13.93	​	1.08, 15.64	1.01, 15.95	​
Sex, n (%)	​	​	NA	​	​	0.417
Female	433 (100.0)	469 (100.0)	​	40 (54.8)	51 (49.5)	​
Male	0 (0.0)	0 (0.0)	​	33 (45.2)	50 (48.5)	​
Unknown	0 (0.0)	0 (0.0)	​	0 (0.0)	2 (1.9)	​
Race/ethnicity, n (%)	​	​	0.170	​	​	0.709
White	180 (41.6)	182 (38.8)	​	30 (41.1)	46 (44.7)	​
Black	29 (6.7)	43 (9.2)	​	3 (4.1)	5 (4.9)	​
Hispanic or Latino	104 (24.0)	139 (29.6)	​	12 (16.4)	14 (13.6)	​
Asian	15 (3.5)	16 (3.4)	​	1 (1.4)	3 (2.9)	​
Other	22 (5.1)	20 (4.3)	​	2 (2.7)	7 (6.8)	​
Unknown	83 (19.2)	69 (14.7)	​	25 (34.2)	28 (27.2)	​
Geographic region, n (%)	​	​	0.519	​	​	0.329
Midwest	95 (21.9)	83 (17.7)	​	16 (21.9)	25 (24.3)	​
Northeast	69 (15.9)	80 (17.1)	​	18 (24.7)	22 (21.4)	​
South	186 (43.0)	203 (43.3)	​	25 (34.2)	45 (43.7)	​
West	82 (18.9)	101 (21.5)	​	14 (19.2)	11 (10.7)	​
Unknown	1 (0.2)	2 (0.4)	​	0 (0.0)	0 (0.0)	​
Medical insurance type, n (%)			0.047	​	​	0.133
Commercial (private)	157 (36.3)	138 (29.4)	​	35 (47.9)	34 (33.0)	​
Medicare (public)	1 (0.2)	5 (1.1)	​	0 (0.0)	2 (1.9)	​
Medicaid (public)	274 (63.3)	326 (69.5)	​	38 (52.1)	66 (64.1)	​
Unknown	1 (0.2)	0 (0.0)	​	0 (0.0)	1 (1.0)	​
Comorbidities, n (%)	​	​	​	​	​	​
Congenital heart disease	10 (2.3)	20 (4.3)	0.102	18 (24.7)	29 (28.2)	0.605
ADHD	32 (7.4)	42 (9.0)	0.392	8 (11.0)	6 (5.8)	0.215
Hearing loss	41 (9.5)	59 (12.6)	0.137	7 (9.6)	8 (7.8)	0.670
Lymphedema	0 (0.0)	3 (0.6)	0.096	1 (1.4)	1 (1.0)	0.806
Hyperlipidemia	9 (2.1)	6 (1.3)	0.348	NA	NA	NA
Asthma	15 (3.5)	24 (5.1)	0.223	6 (8.2)	9 (8.7)	0.903
Precocious puberty	5 (1.2)	3 (0.6)	0.410	4 (5.5)	0 (0.0)	0.016
Body mass index, n (%)	​	​	0.267	​	​	NA
Normal (≤24.9 kg/m2)	5 (1.2)	2 (0.4)	​	0 (0.0)	0 (0.0)	​
Overweight (25.0–29.9 kg/m2)	1 (0.2)	0 (0.0)	​	0 (0.0)	0 (0.0)	​
Obesity (≥30 kg/m2)	0 (0.0)	0 (0.0)	​	0 (0.0)	0 (0.0)	​
Unknown	427 (98.6)	467 (99.6)	​	73 (100.0)	103 (100.0)	​
Systemic steroid treatment, n (%)	22 (5.1)	32 (6.8)	0.271	8 (11.0)	9 (8.7)	0.623
Estrogen treatment, n (%)	1 (0.2)	5 (1.1)	0.123	0 (0.0)	0 (0.0)	NA
Oxandrolone treatment, n (%)	2 (0.5)	1 (0.2)	0.517	0 (0.0)	0 (0.0)	NA
Follow-up duration, years	​	​	<0.001	​	​	<0.001
Mean (SD)	3.64 (1.63)	4.28 (1.51)	​	2.11 (0.95)	2.70 (1.00)	​
Median (IQR)	3.40 (2.14–5.66)	4.63 (3.00–5.66)	​	1.91 (1.31–2.99)	2.50 (2.04–2.99)	​
Min, max	1.02, 6.22	1.06, 6.25	​	1.01, 5.72	1.11, 6.22	​

GH treatment initiation is defined as at least 6 months of continuous GH treatment post-index. Early initiation is the start of GH treatment within 6 months post-index.

GH, growth hormone; IQR, interquartile range; NA, not applicable; SD, standard deviation; ADHD, attention-deficit/hyperactivity disorder.

^aChi-square test for categorical and two-sided t test for continuous variables.

Adherence to and Persistence with GH Treatment

Of the 902 patients with TS who initiated GH treatment, 599 (66.4%) had high adherence to treatment, and 507 (56.2%) had long persistence (Table 4). Adherence to GH treatment was similar for the early and late initiators, with an overall PDC of 0.83 (Table 2). Patients with high adherence were more likely to have commercial insurance and less likely to have Medicaid than patients with low adherence (p = 0.001) (Table 4). Race, geographic region, and the absence of congenital heart disease were significantly associated with long GH treatment persistence. Precocious puberty, no estrogen treatment, and no oxandrolone treatment were also significantly associated with long persistence, but sample sizes for these conditions were too small to interpret.

Table 4.

Baseline characteristics for patients with Turner and Noonan syndromes by treatment adherence and persistence

Characteristics	Patients with TS and GH treatment initiation			Patients with NS and GH treatment initiation			Patients with TS and GH treatment initiation			Patients with NS and GH treatment initiation
	high treatment adherence (n = 599)	low treatment adherence (n = 303)	p valuea	high treatment adherence (n = 124)	low treatment adherence (n = 52)	p valuea	long persistence (n = 507)	short persistence (n = 395)	p valuea	long persistence (n = 92)	short persistence (n = 84)	p valuea
% of N	66.4	33.6	​	70.5	29.6	​	56.2	43.8	​	52.3	47.7	​
Age at index, years	​	​	0.557	​	​	0.367	​	​	0.237	​	​	0.606
Mean (SD)	8.82 (2.84)	8.93 (2.72)	​	6.23 (4.14)	5.63 (3.50)	​	8.76 (2.69)	8.98 (2.93)	​	6.20 (3.94)	5.89 (4.00)	​
Median (IQR)	8.96 (6.51–11.06)	9.52 (6.53–11.06)	​	4.92 (2.37–7.80)	4.77 (3.18–7.80)	​	8.93 (6.64–11.58)	9.48 (6.30–11.58)	​	5.59 (2.48–8.98)	4.76 (2.63–8.98)	​
Min, max	4.01, 14.00	4.01, 13.87	​	1.01, 15.64	1.08, 15.95	​	4.01, 14.00	4.01, 13.98	​	1.18, 13.96	1.01, 15.95	​
Sex, n (%)	​	​	NA	​	​	0.089	​	​	NA	​	​	0.312
Female	599 (100.0)	303 (100.0)	​	58 (46.8)	33 (63.5)	​	507 (100.0)	395 (100.0)	​	47 (51.1)	44 (52.4)	​
Male	0 (0.0)	0 (0.0)	​	65 (52.4)	18 (34.6)	​	0 (0.0)	0 (0.0)	​	45 (48.9)	38 (45.2)	​
Unknown	0 (0.0)	0 (0.0)	​	1 (0.8)	1 (1.9)	​	0 (0.0)	0 (0.0)	​	0 (0.0)	2 (2.4)	​
Race/ethnicity, n (%)	​	​	0.102	​	​	0.859	​	​	0.026	​	​	0.440
White	254 (42.4)	108 (35.6)	​	55 (44.4)	21 (40.4)	​	216 (42.6)	146 (37.0)	​	46 (50.0)	30 (35.7)	​
Black	38 (6.3)	34 (11.2)	​	5 (4.0)	3 (5.8)	​	27 (5.3)	45 (11.4)	​	4 (4.3)	4 (4.8)	​
Hispanic or Latino	157 (26.2)	86 (28.4)	​	20 (16.1)	6 (11.5)	​	133 (26.2)	110 (27.8)	​	14 (15.2)	12 (14.3)	​
Asian	22 (3.7)	9 (3.0)	​	2 (1.6)	2 (3.8)	​	17 (3.4)	14 (3.5)	​	2 (2.2)	2 (2.4)	​
Other	29 (4.8)	13 (4.3)	​	6 (4.8)	3 (5.8)	​	25 (4.9)	17 (4.3)	​	4 (4.3)	5 (6.0)	​
Unknown	99 (16.5)	53 (17.5)	​	36 (29.0)	17 (32.7)	​	89 (17.6)	63 (15.9)	​	22 (23.9)	31 (36.9)	​
Geographic region, n (%)	​	​	0.496	​	​	0.457	​	​	0.019	​	​	0.577
Midwest	121 (20.2)	57 (18.8)	​	29 (23.4)	12 (23.1)	​	105 (20.7)	73 (18.5)	​	19 (20.7)	22 (26.2)	​
Northeast	107 (17.9)	42 (13.9)	​	31 (25.0)	9 (17.3)	​	100 (19.7)	49 (12.4)	​	21 (22.8)	19 (22.6)	​
South	248 (41.4)	141 (46.5)	​	45 (36.3)	25 (48.1)	​	208 (41.0)	181 (45.8)	​	36 (39.1)	34 (40.5)	​
West	121 (20.2)	62 (20.5)	​	19 (15.3)	6 (11.5)	​	93 (18.3)	90 (22.8)	​	16 (17.4)	9 (10.7)	​
Unknown	2 (0.3)	1 (0.3)	​	0 (0.0)	0 (0.0)	​	1 (0.2)	2 (0.5)	​	0 (0.0)	0 (0.0)	​
Medical insurance type, n (%)	​	​	0.001	​	​	0.600	​	​	0.147	​	​	0.154
Commercial (private)	218 (36.4)	77 (25.4)	​	46 (37.1)	23 (44.2)	​	176 (34.7)	119 (30.1)	​	41 (44.6)	28 (33.3)	​
Medicare (public)	6 (1.0)	0 (0.0)	​	2 (1.6)	0 (0.0)	​	5 (1.0)	1 (0.3)	​	0 (0.0)	2 (2.4)	​
Medicaid (public)	374 (62.4)	226 (74.6)	​	75 (60.5)	29 (55.8)	​	326 (64.3)	274 (69.4)	​	50 (54.3)	54 (64.3)	​
Unknown	1 (0.2)	0 (0.0)	​	1 (0.8)	0 (0.0)	​	0 (0.0)	1 (0.3)	​	1 (1.1)	0 (0.0)	​
Comorbidities, n (%)	​	​	​	​	​	​	​	​	​	​	​	​
Congenital heart disease	19 (3.2)	11 (3.6)	0.717	34 (27.4)	13 (25.0)	0.741	11 (2.2)	19 (4.8)	0.028	21 (22.8)	26 (31.0)	0.224
ADHD	48 (8.0)	26 (8.6)	0.769	13 (10.5)	1 (1.9)	0.055	36 (7.1)	38 (9.6)	0.171	9 (9.8)	5 (6.0)	0.348
Hearing loss	65 (10.9)	35 (11.6)	0.752	10 (8.1)	5 (9.6)	0.737	56 (11.0)	44 (11.1)	0.964	10 (10.9)	5 (6.0)	0.243
Lymphedema	3 (0.5)	0 (0.0)	0.217	1 (0.8)	1 (1.9)	0.524	1 (0.2)	2 (0.5)	0.424	0 (0.0)	2 (2.4)	0.137
Hyperlipidemia	9 (1.5)	6 (2.0)	0.596	NA	NA	NA	9 (1.8)	6 (1.5)	0.765	NA	NA	NA
Asthma	22 (3.7)	17 (5.6)	0.177	9 (7.3)	6 (11.5)	0.353	21 (4.1)	18 (4.6)	0.761	8 (8.7)	7 (8.3)	0.931
Precocious puberty	6 (1.0)	2 (0.7)	0.605	4 (3.2)	0 (0.0)	0.190	8 (1.6)	0 (0.0)	0.012	2 (2.2)	2 (2.4)	0.927
Body mass index, n (%)	​	​	0.430	​	​	NA	​	​	0.676	​	​	NA
Normal (≤24.9 kg/m2)	6 (1.0)	1 (0.3)	​	0 (0.0)	0 (0.0)	​	4 (0.8)	3 (0.8)	​	0 (0.0)	0 (0.0)	​
Overweight (25.0–29.9 kg/m2)	1 (0.2)	0 (0.0)	​	0 (0.0)	0 (0.0)	​	1 (0.2)	0 (0.0)	​	0 (0.0)	0 (0.0)	​
Obesity (≥30 kg/m2)	0 (0.0)	0 (0.0)	​	0 (0.0)	0 (0.0)	​	0 (0.0)	0 (0.0)	​	0 (0.0)	0 (0.0)	​
Unknown	592 (98.8)	302 (99.7)	​	124 (100.0)	52 (100.0)	​	502 (99.0)	392 (99.2)	​	92 (100.0)	84 (100.0)	​
Systemic steroid treatment, n (%)	37 (6.2)	17 (5.6)	0.735	13 (10.5)	4 (7.7)	0.567	29 (5.7)	25 (6.3)	0.702	10 (10.9)	7 (8.3)	0.569
Estrogen treatment, n (%)	4 (0.7)	2 (0.7)	0.989	0 (0.0)	0 (0.0)	NA	0 (0.0)	6 (1.5)	0.005	0 (0.0)	0 (0.0)	NA
Oxandrolone treatment, n (%)	1 (0.2)	2 (0.7)	0.224	0 (0.0)	0 (0.0)	NA	0 (0.0)	3 (0.8)	0.049	0 (0.0)	0 (0.0)	NA
Follow-up duration, years	​	​	0.001	​	​	0.072	​	​	<0.001	​	​	<0.001
Mean (SD)	3.85 (1.60)	4.22 (1.58)	​	2.36 (1.01)	2.67 (1.01)	​	4.14 (1.56)	3.75 (1.64)	​	2.72 (1.01)	2.16 (0.96)	​
Median (IQR)	3.78 (2.46–5.67)	4.57 (2.78–5.67)	​	2.16 (1.61–3.09)	2.55 (2.03–3.09)	​	4.43 (2.74–5.24)	3.69 (2.22–5.24)	​	2.48 (2.10–2.78)	2.03 (1.32–2.78)	​
Min, max	1.02, 6.25	1.15, 6.25	​	1.06, 6.22	1.01, 4.92	​	1.15, 6.25	1.02, 6.24	​	1.06, 5.72	1.01, 6.22	​

GH treatment initiation is defined as at least 6 months of continuous GH treatment post-index. Early initiation is the start of GH treatment within 6 months post-index. High adherence is defined as PDC of at least 80% (i.e., taking at least 80% of the expected doses). Long persistence is defined as at least 2 years of continuous GH treatment or at least 1 year of treatment up to the end of follow-up.

GH, growth hormone; IQR, interquartile range; NA, not applicable; PDC, proportion of days covered; SD, standard deviation; ADHD, attention-deficit/hyperactivity disorder.

^bChi-square test for categorical and two-sided t test for continuous variables.

Of the 176 patients with NS who started GH treatment, 124 (70.5%) had high adherence, and 92 (52.3%) had long persistence (Table 4). As measured by the PDC method, adherence to GH treatment was 0.85, similar for the early and late starters (Table 2). A greater proportion of patients with high adherence were male, but no baseline characteristics were significantly associated with high adherence (Table 4). Although a greater proportion of patients with long persistence were White and had commercial insurance, no baseline demographic or clinical characteristics were significantly associated with long persistence for patients with NS.

HCRU and All-Cause Costs

Among all patients with TS, 9.3%, 99.9%, 48.1%, and 93.3% had at least 1 inpatient, outpatient, ED, and pharmacy claim, respectively; rates were similar for GH-initiated patients and GH-non-initiated patients (Table 5). The overall median (IQR) number of inpatient visits was 0.36 (0.21–0.57) PPPY; among these patients, the median (IQR) total length of stay was 1.31 (0.67–3.00) days, slightly higher for GH-non-initiated patients than GH-initiated patients (1.48 [0.77–3.58] vs. 1.00 [0.59–2.44]). The median (IQR) number of outpatient and ED visits was 16.32 (9.79–28.31) and 0.53 (0.32–0.93) PPPY, respectively. The median (IQR) number of pharmacy claims was 8.23 (2.91–15.18), 13.66 (9.00–19.39) for GH-initiated patients vs. 4.09 (1.66–10.52) for GH-non-initiated patients. Total overall all-cause costs PPPY were higher for GH-initiated patients than GH-non-initiated patients (median [IQR] USD 39,816 [USD 24,781–USD 62,295] vs. USD 3,932 [USD 1,870–USD 9,830]), driven primarily by the higher pharmacy costs for the GH-initiated patients (USD 32,966 [USD 18,197–USD 53,826] vs. USD 199 [USD 41–USD 969]) than the GH-non-initiated patients.

Table 5.

HCRU and costs by treatment initiation status

​	Patients with TS			Patients with NS
	total (N = 2,530)	GH treatment initiated (n = 902)	GH treatment not initiated (n = 1,628)	total (N = 1,119)	GH treatment initiated (n = 176)	GH treatment not initiated (n = 943)
All-cause HCRU, PPPY a
Inpatient
n (%)	236 (9.3)	72 (8.0)	164 (10.1)	267 (23.9)	44 (25.0)	223 (23.6)
Mean (SD)	0.58 (0.77)	0.37 (0.30)	0.67 (0.88)	1.02 (1.32)	0.88 (0.77)	1.05 (1.40)
Median (IQR)	0.36 (0.21–0.57)	0.29 (0.18–0.39)	0.39 (0.23–0.67)	0.67 (0.41–1.02)	0.70 (0.34–0.98)	0.66 (0.42–1.07)
Min, max	0.16, 4.98	0.16, 1.89	0.16, 4.98	0.00, 14.11	0.20, 3.89	0.00, 14.11
Total length of stay, days
Mean (SD)	5.84 (25.42)	1.93 (2.49)	7.55 (30.32)	7.27 (12.60)	4.80 (5.92)	7.76 (13.49)
Median (IQR)	1.31 (0.67–3.00)	1.00 (0.59–2.44)	1.48 (0.77–3.58)	2.89 (1.34–6.77)	2.16 (1.13–6.24)	3.09 (1.41–7.63)
Min, max	0.16, 356.06	0.16, 13.54	0.17, 356.06	0.00, 105.94	0.48, 24.23	0.00, 105.94
Outpatient
n (%)	2,527 (99.9)	902 (100.0)	1,625 (99.8)	1,118 (99.9)	176 (100.0)	942 (99.9)
Mean (SD)	30.74 (54.34)	27.29 (32.49)	32.66 (63.22)	85.56 (120.07)	77.11 (104.44)	87.14 (122.75)
Median (IQR)	16.32 (9.79–28.31)	18.40 (12.32–28.53)	14.46 (8.67–28.17)	37.04 (17.84–100.49)	36.36 (18.93–94.07)	37.24 (17.28–103.33)
Min, max	0.25, 917.43	0.46, 331.72	0.25, 917.43	1.78, 999.47	6.50, 706.66	1.78, 999.47
Emergency department
n (%)	1,218 (48.1)	450 (49.9)	768 (47.2)	631 (56.4)	105 (59.7)	526 (55.8)
Mean (SD)	0.81 (0.91)	0.70 (0.66)	0.88 (1.02)	1.42 (1.58)	1.42 (1.76)	1.42 (1.55)
Median (IQR)	0.53 (0.32–0.93)	0.50 (0.27–0.86)	0.56 (0.33–0.97)	0.92 (0.51–1.72)	0.89 (0.49–1.52)	0.94 (0.51–1.76)
Min, max	0.16, 11.19	0.16, 4.67	0.16, 11.19	0.20, 17.92	0.21, 12.95	0.20, 17.92
Total pharmacy
n (%)	2,360 (93.3)	902 (100.0)	1,458 (89.6)	1,029 (92.0)	176 (100.0)	853 (90.5)
Mean (SD)	10.97 (11.05)	15.35 (8.98)	8.26 (11.33)	14.79 (17.13)	19.14 (12.51)	13.90 (17.81)
Median (IQR)	8.23 (2.91–15.18)	13.66 (9.00–19.39)	4.09 (1.66–10.52)	9.19 (3.14–20.28)	16.24 (10.56–24.51)	7.32 (2.46–18.52)
Min, max	0.16, 109.23	1.00, 66.55	0.16, 109.23	0.23, 138.24	2.84, 94.57	0.23, 138.24
All-cause costs (USD), PPPY a
Total overall costs (medical and pharmacy)
n (%)	2,530 (100.0)	902 (35.7)	1,628 (64.3)	1,119 (100.0)	176 (15.7)	943 (84.3)
Mean (SD)	26,511.83 (53,343.85)	47,288.43 (31,528.39)	15,000.48 (59,167.11)	52,594.33 (139,881.54)	62,023.66 (69,958.04)	50,834.46 (149,310.84)
Median (IQR)	10,231.95 (2,850.05–36,589.74)	39,816.06 (24,780.85–62,295.27)	3,932.42 (1,870.46–9,830.26)	17,936.96 (5,616.37–47,552.20)	43,488.28 (26,132.23–71,404.39)	13,355.69 (4,654.25–39,102.75)
Min, max	20.32, 1,963,911.62	2,828.05, 262,430.02	20.32, 1,963,911.62	364.41, 2,449,855.19	7,211.70, 578,136.09	364.41, 2,449,855.19
Total medical costs
n (%)	2,530 (100.0)	902 (35.7)	1,628 (64.3)	1,119 (100.0)	176 (15.7)	943 (84.3)
Mean (SD)	11,076.94 (46,719.36)	7,374.37 (12,855.70)	13,128.36 (57,353.65)	44,281.12 (134,180.43)	31,916.08 (56,275.43)	46,588.91 (144,034.83)
Median (IQR)	3,183.76 (1,720.90–7,714.81)	3,485.26 (1,960.16–7,370.64)	3,102.74 (1,550.81–7,905.94)	11,460.50 (4,221.14–35,590.89)	11,023.92 (4,522.33–37,379.90)	11,562.76 (4,084.16–35,509.02)
Min, max	19.71, 1,951,697.43	274.82, 172,248.08	19.71, 1,951,697.43	313.85, 2,420,463.70	864.22, 384,241.84	313.85, 2,420,463.70
Inpatient
n (%)	236 (9.3)	72 (30.5)	164 (69.5)	267 (23.9)	44 (16.5)	223 (83.5)
Mean (SD)	20,068.71 (36,101.15)	10,184.30 (12,438.56)	24,408.20 (41,826.85)	39,459.46 (71,658.23)	26,464.43 (30,616.27)	42,023.50 (77,012.97)
Median (IQR)	8,642.44 (3,509.64–21,374.98)	5,053.65 (2,556.08–13,565.17)	10,273.60 (4,355.23–25,013.62)	17,214.68 (8,455.85–43,122.23)	18,194.54 (9,251.83–31,627.87)	16,979.82 (8,369.96–44,632.99)
Min, max	484.08, 293,176.82	484.08, 78,213.18	707.06, 293,176.82	346.56, 831,360.18	1,353.57, 172,452.79	346.56, 831,360.18
Outpatient
n (%)	2,527 (99.9)	902 (35.7)	1,625 (64.3)	1,118 (99.9)	176 (15.7)	942 (84.3)
Mean (SD)	8,912.07 (41,399.87)	6,342.25 (10,788.26)	10,338.51 (50,947.35)	34,078.33 (125,257.06)	24,508.21 (48,941.24)	35,866.38 (134,751.64)
Median (IQR)	2,905.73 (1,557.95–6,469.12)	3,158.31 (1,839.66–6,320.02)	2,712.16 (1,398.69–6,678.64)	8,617.46 (3,479.36–24,135.12)	8,079.79 (3,876.52–21,170.94)	8,680.75 (3,381.59–24,295.26)
Min, max	19.71, 1,831,222.21	66.89, 160,333.74	19.71, 1,831,222.21	313.85, 2,407,648.95	864.22, 335,662.22	313.85, 2,407,648.95
Emergency department
n (%)	1,218 (48.1)	450 (36.9)	768 (63.1)	631 (56.4)	105 (16.6)	526 (83.4)
Mean (SD)	630.25 (1,670.83)	439.35 (599.92)	742.11 (2,045.72)	1,450.60 (2,974.34)	1,327.15 (2,265.27)	1,475.24 (3,097.72)
Median (IQR)	265.43 (120.65–628.26)	224.76 (110.56–504.18)	295.31 (129.50–744.74)	631.51 (266.95–1,558.46)	552.79 (292.14–1,476.04)	651.01 (266.56–1,593.64)
Min, max	11.88, 37,812.75	19.26, 6,001.00	11.88, 37,812.75	22.73, 44,996.73	33.05, 19,092.99	22.73, 44,996.73
Total pharmacy costs
n (%)	2,360 (93.3)	902 (38.2)	1,458 (61.8)	1,029 (92.0)	176 (17.1)	853 (82.9)
Mean (SD)	16,546.73 (26,532.64)	39,914.06 (28,911.70)	2,090.40 (8,702.20)	9,040.31 (34,964.63)	30,107.57 (42,513.40)	4,693.49 (31,515.54)
Median (IQR)	1,681.28 (115.33–25,977.46)	32,966.03 (18,197.37–53,825.54)	198.92 (40.59–969.27)	570.85 (85.08–6,659.05)	19,362.23 (12,102.89–39,658.87)	306.81 (57.87–1,710.12)
Min, max	0.18, 261,736.30	1,527.38, 261,736.30	0.18, 196,805.88	0.83, 513,564.57	3,669.09, 513,564.57	0.83, 505,364.16

GH treatment initiation is defined as at least 6 months of continuous GH treatment post-index. Costs are inflation adjusted to 2023 USD.

GH, growth hormone; IQR, interquartile range; PPPY, per patient per year; SD, standard deviation.

^aAmong those who had at least one claim.

Among the patients with NS, 23.9%, 99.9%, 56.4%, and 92.0% had at least 1 inpatient, outpatient, ED, and pharmacy claim, respectively, with similar rates by GH treatment initiation status (Table 5). The median (IQR) number of inpatient visits and length of stay was 0.67 (0.41–1.02) and 2.89 (1.34–6.77) days PPPY, similar for both GH treatment initiation groups. Patients with NS had a median (IQR) of 37.04 (17.84–100.49) and 0.92 (0.51–1.72) outpatient and ED visits PPPY, respectively. As expected, GH-initiated patients had more pharmacy claims than the GH-non-initiated patients (median [IQR] 16.24 [10.56–24.51] vs. 7.32 [2.46–18.52]). GH-initiated patients had higher total costs PPPY than the GH-non-initiated patients (median [IQR] USD 43,488 [USD 26,132–USD 71,404] vs. USD 13,356 [USD 4,654–USD 39,103]) due to the higher pharmacy costs for the GH-initiated vs. GH-non-initiated patients (USD 19,362 [USD 12,103–USD 39,659] vs. USD 307 [USD 58–USD 1,710]).

Discussion

In this retrospective US-based cohort study, we observed a very low rate of treatment initiation – only 36% of patients with TS aged 4–13 years and 16% of patients with NS started GH therapy during the follow-up period. The mean age at index date (proxy for diagnosis date) was 8.4 and 7.0 years, respectively. Despite their older age at diagnosis, less than half of the patients with TS or NS taking GH treatment started treatment early (within 6 months of diagnosis), deviating from international clinical guidelines. The average time to initiation was 2–3 months for the early initiators and more than 1 year (NS) to nearly 2 years (TS) for late initiators; for both conditions, early initiators were older at the time of diagnosis. The majority of patients with TS or NS who started GH therapy had high adherence to treatment, but only about half had long persistence. Unadjusted total, all-cause PPPY costs were higher for GH-initiated patients compared with GH-non-initiated patients for both conditions, driven by the expected higher pharmacy costs for the GH-initiated patients. Our research highlights the need for further awareness and education about the need for early treatment and diagnosis of TS and NS among healthcare professionals and caregivers of these patients.

The current recommendation for TS per the International Turner Syndrome Consensus Group is to start GH therapy as early as age 2 years for short stature, growth failure, or likelihood of short stature, to ensure adequate treatment duration before the start of puberty [22]. Because most patients with TS have short stature [5], and the age at index date in our study was limited to those aged 4–13 years, the majority of our patient population were candidates for guideline-based GH therapy; however, a low rate of GH therapy use was observed. Most patients with TS were started on GH therapy approximately 1 year after GH treatment was indicated, yielding an average age of treatment initiation of approximately 10 years, which is slightly higher than that seen in a real-world study in Argentina (8.8 years) and an outcomes study in 4 European countries (range, mean 6.2–8.8 years) [27, 28] but lower than that seen in a small (n = 16) hospital-based, retrospective study in India (mean of 12.7 years) [29].

Approximately 80% of individuals with NS have short stature associated with specific mutations that cause this condition [5, 6]. Studies have demonstrated that patients with NS who have short stature benefit from early initiation of GH therapy [14–16]. A survey of European clinicians treating NS found that most pediatric endocrinologists reported the optimal age for starting GH therapy in patients with short stature was early childhood (from ages 4 to <7 years) [36]. However, few patients with NS in our US study were initiated on GH treatment, and the mean age at diagnosis (index date) was 6 years. Patients with NS started GH treatment an average of 9 months after diagnosis (at approximately 7 years of age), slightly younger than that observed in the Kabi International Growth Study database of GH-treated patients with NS (mean age of 9.6 years) [26].

Clinical studies have demonstrated that longer GH treatment duration is associated with improvement in adult height for both conditions [16, 18, 19, 21]. We observed an average treatment duration of 25 and 18 months among patients with TS and NS, respectively, lower than that seen in other research studies (ranging from approximately 4–6 years) [26–28], likely due to the short follow-up period in our analysis. We did not assess the reasons for the low initiation of GH treatment in our study, but possible factors could be the mode of administration (currently, daily injections), lack of insurance coverage/cost, or reluctance to use GH therapy in patients with NS because of safety concerns related to cardiomyopathy and/or potential inherent oncologic risk. However, evidence shows that the use of GH therapy is safe in patients with NS [24, 37, 38]. Another factor for low initiation of GH treatment may be the lack of healthcare professionals’ adherence to clinical practice guidelines, specifically for TS [39, 40].

Our study provides real-world insights into adherence and persistence to GH therapy in patients with TS and NS, which is lacking in the literature. We observed considerably lower rates of adherence and persistence in our study population for both pediatric non-GHD conditions compared with research reporting a GH-treatment adherence rate of 89% among patients with each of 5 different conditions, including TS [27]. In that study, researchers noted a significant difference in adherence among patients with private health coverage compared with public sector coverage [27], a finding also observed among patients with TS in our study. Additional GH therapy options, including long-acting GH treatment devices, could potentially improve adherence and persistence, as seen in pediatric GH deficiency [41, 42].

Additionally, our study offers new information about HCRU and costs in patients with TS or NS. An analysis of hospitalizations in a US sample of patients with TS from 2017 to 2019 [43] found that the total mean hospital costs for patients with TS were about USD 18,000, higher than the average US 2016 annual hospitalization cost of USD 13,400 [44]. Our study also found differences in comorbidities typically associated with patients with TS and NS in the clinical literature. Specifically, we found lower rates of hearing loss in both conditions (11% for TS and 14% for NS) compared with 60%–80% seen in the literature for TS and 25% for NS [5, 8]; however, the difference in rates for TS could be a result of patients not being diagnosed until later in life. We also observed a considerably lower proportion of patients with NS who had congenital heart disease (25%) compared with the 80% commonly reported in this patient population [5]. These differences could be related to the reporting window for the claims analysis or our study population’s age, which was beyond the typical timing (age) for cardiac evaluations in these patients.

The estimated prevalence of NS is 1 in 1,000–1 in 2,500 live births, or 4–10 per 10,000 live births [6]. Our estimate of 5 per 100,000 in the US population is likely a result of our conservative inclusion criteria. Because there is no specific ICD-10-CM disease code for NS, our disease classification method required further clinical conditions commonly associated with NS to increase confidence in cohort identification; this method may have underestimated the NS population in the claims database and focused on a more severely ill subset of patients with NS. Because NS is often mild [45], it is possible that our study captured only more complex cases. In contrast, our TS-estimated prevalence of 36 per 100,000 is similar to the reported prevalence of 1 per 2,500 [1], indicating that our study represents the majority of the patients with TS in the USA.

A key strength of our study is the use of a large-scale administrative claims database spanning multiple years, thereby providing a robust patient population for analysis. Limitations to our research include incomplete claims (i.e., if patients disenroll from health plans, switch plans, or pay out of pocket), missing information, and a lack of specific billing codes for some conditions, impeding the identification of specific causes for medical encounters and the determination of appropriate care. Data within a claims database are also subject to coding limitations and data entry errors, leading to inaccuracies in disease classification, patterns of medication use, HCRU, and associated costs. Classification of medical encounters and expenses related to a specific condition is limited when using claims data because claims are coded for billing purposes and are not designed to provide a complete clinical rationale for visits and procedures. Insurance claims data may overestimate drug costs because they lack information on rebates paid by drug manufacturers to health plans. Costs were unadjusted for covariates.

This is a non-interventional study; therefore, potential confounding factors cannot be ruled out. Data collection reflects routine clinical practice rather than mandatory assessments at pre-specified time points, which may impact the amount of data available and its interpretation. Because the average follow-up periods in our study were 2–4 years, HCRU and costs may not apply to later periods following diagnosis, which will vary considerably over the disease trajectory. The claims database only included patients’ birth year, limiting precision in calculating age and duration of GH therapy. We defined GH treatment initiation as being treated for at least 6 months to consider patients who received adequate GH treatment. Furthermore, because TS and NS diagnoses were based on available ICD-10-CM disease codes, misdiagnosis is a possibility. We approximated the date of diagnosis based on pertinent coding of medical claims. There may be underlying clinical differences between early and late GH treatment initiators. We did not have access to patients’ height data and therefore cannot verify if patients had short stature requiring GH treatment. However, all patients with NS were required to have a diagnosis code for short stature, and more than two-thirds of patients with TS (n = 1,711/n = 2,530) had documented short stature. Although coding of short stature may not be consistent among healthcare professionals, the majority of patients in our study met the criteria for treatment with GH therapy. Additionally, reasons why GH treatment was or was not initiated, GH dose, and GH treatment side effects were unknown. Height percentile for age threshold limits for GH therapy insurance coverage, and a lower prevalence of congenital heart disease in patients with NS may limit the generalizability of our research. Because claims databases rely on self-reported information from patients, data on race and ethnicity were missing for approximately 20–30% of our patient population. Lastly, our study findings may not generalize to other countries with different healthcare systems. Future real-world studies of TS and NS are needed with longer follow-up periods; further exploration of HCRU and costs in these patient populations is warranted to better understand the financial impact of these conditions and their treatment with GH therapy.

Conclusions

Our US-based study identified that approximately 36% patients with TS and 16% of patients with NS initiated GH treatment after diagnosis and that only up to half of these patients started GH treatment within 6 months of diagnosis. These findings highlight the need for better implementation of guideline-recommended clinical management of these patients and more awareness and education about the rationale for earlier diagnosis and treatment.

Acknowledgments

The authors thank Rebecca Hahn, MPH, of KJT Group, Inc., Rochester, NY, USA, for providing medical writing support, which was funded by Novo Nordisk, per Good Publication Practice (GPP 2022) guidelines.

Statement of Ethics

This was a non-interventional, retrospective cohort study using the Komodo Health claims database. Ethical approval and patient consent were not required for this study in accordance with local or national guidelines. All methods were carried out following relevant guidelines and regulations. The clinical data of the patients were collected from the database, and all data were anonymized before being used them in this study. No administrative permissions were required to access the raw data used in our research.

Conflict of Interest Statement

Andrew Dauber has served as a consultant for Novo Nordisk, Pfizer, BioMarin, and QED; has received research funding from Pfizer and BioMarin; and was a member of the journal’s Editorial Board at the time of submission. M Jennifer Abuzzahab has received research funding from Ascendis, Lumos, Medtronic, Novo Nordisk, Rhythm, and Soleno and has served as a consultant for Ascendis, Endo, Neurocrine, and Pfizer. Alicia Romano has served as a consultant for Novo Nordisk and BioMarin and is a speaker for Novo Nordisk. Vaneeta Bamba receives funding administered directly to Children’s Hospital of Philadelphia for industry-sponsored clinical trials administered by Lumos Pharma and Amryt Pharma for treatment of conditions unrelated to those described in this abstract. Jing Voon Chen is an employee of Novo Nordisk Inc. Nicky Kelepouris is an employee and stockholder of Novo Nordisk Inc. Moshe Fridman and Morgan Shy are employees of AMF Consulting, which was contracted to conduct the study analyses. Scott Bunner is an employee of Genesis Research Group, which was contracted to conduct the study analyses.

Funding Sources

Novo Nordisk Inc., Plainsboro, NJ, USA, funded this research and medical writing support.

Author Contributions

Jing Voon Chen and Nicky Kelepouris were responsible for the study concept, design, and data collection. Moshe Fridman, Morgan Shy, and Scott Bunner were responsible for the design and data analysis. All authors (Andrew Dauber, M Jennifer Abuzzahab, Alicia Romano, Vaneeta Bamba, Jing Voon Chen, Nicky Kelepouris, Moshe Fridman, Morgan Shy, and Scott Bunner) were responsible for data interpretation and analysis, drafting/revising the manuscript, and reviewing/approving the final version for submission.

Funding Statement

Novo Nordisk Inc., Plainsboro, NJ, USA, funded this research and medical writing support.

Data Availability Statement

The data supporting this study’s findings are not publicly available due to privacy reasons but are available from the corresponding author upon reasonable request.