Growth Hormone Time to Access and Outcomes: Part 2

Authors
Affiliation

Ryan Moore

VUMC Department of Biostatistics

Leena Choi

Published

Last Updated on 12 September 2024

Project Overview

Background

  • Growth hormone products for pediatric patients can be costly and difficult to obtain.

  • Patients face many barriers in obtaining these medications such as high out-of-pocket costs and delays in medication acquisition due to insurance authorization

  • Access to these medications may be limited due to the complex process from the physician prescribing to delivery of the medications.

  • Identifying predictors of overall medication access and time to medication initiation may improve the speed in which patients are able to initiate treatment and therefore improve growth and overall health outcomes.

Outcome

Primary outcome: Determine whether time to growth hormone approval is associated with the following predictors

  • Diagnosis

  • Type of payer

  • Medication

  • Additional testing required

  • Need for outside medical records

Secondary outcome:

Evaluate whether time to therapy initiation impacts patient growth

Evaluate whether time to therapy initiation from decision to treat (measured as number of days between treatment decision and hGh initiation) predicts patient growth

Methods

  • Single center, retrospective review of electronic medical records of patients starting treatment with growth hormone at the Vanderbilt Pediatric Endocrinology Clinic between January 1, 2018, and December 31, 2019

  • Data are summarized using descriptive statistics.

  • Patient growth is estimated using z scores of height by age (and sex) using the zscorer or anthro packages and WHO growth standards/references or CDC growth standards. For patients 0 to 5 years of age, the WHO growth standards are used (https://iris.who.int/handle/10665/43413). For 5 to 24 years, the WHO growth reference is used. (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2636412/pdf/07-043497.pdf/). For patients 2 to 20 years of age the CDC growth standards are used. The hfa z scores of the CDC and WHO standards are compared. A random effect model is fit to predict z score height-for-age at 1 year after decision to treat for each individual. The growth change is then calculated by subtracted the predicted z-score hfa at 1 year post-dtt from the baseline z score hfa.

Exclusions

  • Patients with missing exclusion critiera were recoded to “None of the above”
Characteristic N N = 6411
Exclusion 641
    Change in hGH due to insurance change
11 (1.7%)
    Decision to treat rescinded prior to prescription approval
4 (0.6%)
    hGH renewal
193 (30%)
    None of the above
374 (58%)
    Other
30 (4.7%)
    Patient chose to pay out of pocket initially (rather than obtaining insurance approval)
28 (4.4%)
    Patient with Adult Growth Hormone Deficiency
1 (0.2%)
1 n (%)

Descriptive Statistics

Patient Characteristics

Characteristic N N = 3741
Age 374 11.2 (8.1, 13.3)
Sex 374
    Female
126 (34%)
    Male
248 (66%)
Ethnicity 373
    Not Hispanic, Latino/a, or Spanish origin
158 (42%)
    Unknown
166 (45%)
    Decline to Answer
11 (2.9%)
    Mexican, Mexican American, or Chicano/a
5 (1.3%)
    Cuban
1 (0.3%)
    Other Hispanic, Latino/a, or Spanish origin
32 (8.6%)
    Missing
1
Race 374
    Black
20 (5.3%)
    Other
29 (7.8%)
    Unknown
24 (6.4%)
    White
301 (80%)
Language 374
    English
353 (94%)
    Non-English
21 (5.6%)
Baseline height Z-score CDC 359 -2.5 (-3.0, -2.0)
    Missing
15
Baseline height Z-score WHO 374 -2.5 (-3.0, -2.0)
Basesline height Z-score WHO and CDC 374 -2.5 (-3.0, -2.0)
Baseline Height 374 128.4 (113.6, 141.2)
Diagnosis 374
    Chronic kidney disease
5 (1.3%)
    Growth hormone deficiency
185 (49%)
    Idiopathic short stature
113 (30%)
    Noonan Syndrome
1 (0.3%)
    Panhypopituitarism
7 (1.9%)
    Prader-Willi syndrome
11 (2.9%)
    Small for Gestational age
35 (9.4%)
    Turner's syndrome
17 (4.5%)
Insurance Primary 374
    Medicaid
117 (31%)
    Tricare
12 (3.2%)
    Commercial
245 (66%)
Insurance Secondary 374
    Medicaid
11 (2.9%)
    Tricare
1 (0.3%)
    Commercial
10 (2.7%)
    None
346 (93%)
    Other
6 (1.6%)
Insurance Tertiary 374
    Medicaid
2 (0.5%)
    Commercial
8 (2.1%)
    None
359 (96%)
    Other
5 (1.3%)
Testing 374
    No test
335 (90%)
    Patient tested
39 (10%)
VSP Patient 374
    No
181 (48%)
    Yes
193 (52%)
1 Median (IQR); n (%)

Height

Table

Summary
Characteristic N N = 3741
Baseline Height 374 128.4 (113.6, 141.2)
Basesline height Z-score WHO and CDC 374 -2.5 (-3.0, -2.0)
Follow-up time for height measurements (yr) 374 2.9 (2.4, 3.6)
1 Median (IQR)

Visualizations

Height over time

Medication data

Characteristic N N = 3741
Medication 374
    Norditropin
111 (30%)
    Nutropin
30 (8.0%)
    Genotropin
123 (33%)
    Humatrope
77 (21%)
    Omnitrope
24 (6.4%)
    Zomacton
9 (2.4%)
Time to approval from decision to treat 371 10.0 (4.0, 37.5)
    Missing
3
Time to med access from decision to treat 253 3.0 (1.0, 6.0)
    Missing
121
Time to appeal from decision to treat 33 42.0 (15.0, 62.0)
    Missing
341
Interim meds started 367
    Yes
25 (6.8%)
    No
342 (93%)
    Missing
7
Delay not from insurance 372
    No
321 (86%)
    Yes
51 (14%)
    Missing
2
Testing 374
    No test
335 (90%)
    Patient tested
39 (10%)
Outside records 374
    No
364 (97%)
    Yes
10 (2.7%)
PA completed for multiple insurance 374
    No
352 (94%)
    Yes
22 (5.9%)
Change insurance 21
    No
17 (81%)
    Yes
4 (19%)
    Missing
353
Number of PA 374
    none
11 (2.9%)
    1
288 (77%)
    2
67 (18%)
    3
7 (1.9%)
    >3
1 (0.3%)
PA denial 111
    Off-label use
0 (0%)
    Did not meet criteria or required formulary alternative
50 (45%)
    Non preferred agent
2 (1.8%)
    Diagnosis not covered
53 (48%)
    Other
6 (5.4%)
    Missing
263
pa_denial_other 6
    diagnosis of ISS not considered medically necessary
1 (17%)
    Gh excluded from plan
2 (33%)
    GH excluded from plan
1 (17%)
    GH not covered under plan
1 (17%)
    TennCare only pays for drug if they've tried other drugs that cost less
1 (17%)
    Missing
368
Appeal number 90
    1
51 (57%)
    2
28 (31%)
    3
11 (12%)
    Missing
284
Assistance 320
    Copay card
71 (22%)
    MAP
3 (0.9%)
    Other
1 (0.3%)
    None
138 (43%)
    Uknown
107 (33%)
    Missing
54
1 n (%); Median (IQR)

Time to insurance approval

Time differences in days
  0%  25%  50%  75% 100% 
   1   46   94  155  215
Characteristic N = 3741
Prior Authorization Approved
    N 264
    Median (IQR) 3 (1, 6)
    Range 0, 421
    Unknown 110
Appeal Approved
    N 56
    Median (IQR) 34 (22, 68)
    Range 3, 1,121
    Unknown 318
PAP Approved
    N 48
    Median (IQR) 84 (59, 122)
    Range 7, 430
    Unknown 326
Cash Pay Decided
    1 1 (14%)
    13 1 (14%)
    79 1 (14%)
    94 1 (14%)
    118 1 (14%)
    192 1 (14%)
    215 1 (14%)
    Unknown 367
1 n (%)

Analysis

Modeling time to appeal or PAP approval

Outcome

Characteristic N N = 3741
Time to approval (days) 374 5.0 (1.0, 32.0)
Type of approval 374
    Appeal approval time
55 (15%)
    Bi initiation
13 (3.5%)
    Cash pay
7 (1.9%)
    PA approval time
251 (67%)
    PAP approval time
48 (13%)
1 Median (IQR); n (%)

Data exploration visualizations

Diagnosis

Testing

CoxPH Analysis

Call:
coxph(formula = S ~ age + diagnosis_cat + insurance_prim + vsp_patient + 
    testing, data = moddat)

  n= 374, number of events= 367 

                                            coef exp(coef)  se(coef)      z
age                                     0.007299  1.007326  0.014507  0.503
diagnosis_catGrowth hormone deficiency  0.647982  1.911680  0.131795  4.917
diagnosis_catTurner's syndrome          0.686005  1.985767  0.270598  2.535
diagnosis_catSmall for Gestational age  0.470778  1.601240  0.217674  2.163
diagnosis_catPrader-Willi syndrome      1.043579  2.839361  0.343887  3.035
diagnosis_catPanhypopituitarism         0.972065  2.643397  0.403622  2.408
diagnosis_catNoonan Syndrome            3.217320 24.961143  1.024771  3.140
diagnosis_catChronic kidney disease     0.314772  1.369947  0.468542  0.672
insurance_primMedicaid                  0.327585  1.387613  0.120217  2.725
insurance_primTricare                   0.358984  1.431873  0.302285  1.188
vsp_patientYes                          0.411643  1.509295  0.118114  3.485
testingPatient tested                  -0.913522  0.401109  0.184960 -4.939
                                       Pr(>|z|)    
age                                    0.614845    
diagnosis_catGrowth hormone deficiency 8.81e-07 ***
diagnosis_catTurner's syndrome         0.011240 *  
diagnosis_catSmall for Gestational age 0.030559 *  
diagnosis_catPrader-Willi syndrome     0.002408 ** 
diagnosis_catPanhypopituitarism        0.016025 *  
diagnosis_catNoonan Syndrome           0.001692 ** 
diagnosis_catChronic kidney disease    0.501704    
insurance_primMedicaid                 0.006431 ** 
insurance_primTricare                  0.235004    
vsp_patientYes                         0.000492 ***
testingPatient tested                  7.85e-07 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

                                       exp(coef) exp(-coef) lower .95 upper .95
age                                       1.0073    0.99273    0.9791    1.0364
diagnosis_catGrowth hormone deficiency    1.9117    0.52310    1.4765    2.4751
diagnosis_catTurner's syndrome            1.9858    0.50358    1.1684    3.3749
diagnosis_catSmall for Gestational age    1.6012    0.62452    1.0451    2.4532
diagnosis_catPrader-Willi syndrome        2.8394    0.35219    1.4471    5.5711
diagnosis_catPanhypopituitarism           2.6434    0.37830    1.1984    5.8308
diagnosis_catNoonan Syndrome             24.9611    0.04006    3.3495  186.0163
diagnosis_catChronic kidney disease       1.3699    0.72996    0.5469    3.4318
insurance_primMedicaid                    1.3876    0.72066    1.0963    1.7563
insurance_primTricare                     1.4319    0.69839    0.7918    2.5895
vsp_patientYes                            1.5093    0.66256    1.1974    1.9024
testingPatient tested                     0.4011    2.49309    0.2791    0.5764

Concordance= 0.663  (se = 0.016 )
Likelihood ratio test= 80.21  on 12 df,   p=4e-12
Wald test            = 80.48  on 12 df,   p=3e-12
Score (logrank) test = 89.48  on 12 df,   p=6e-14
Analysis of Deviance Table
 Cox model: response is S
Terms added sequentially (first to last)

                loglik   Chisq Df Pr(>|Chi|)    
NULL           -1818.9                          
age            -1817.7  2.4363  1   0.118556    
diagnosis_cat  -1802.4 30.6016  7  7.362e-05 ***
insurance_prim -1797.3 10.2578  2   0.005923 ** 
vsp_patient    -1793.2  8.0468  1   0.004559 ** 
testing        -1778.8 28.8686  1  7.746e-08 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Modeling height

Linear regression Model

Linear Regression Model

ols(formula = Predicted_Height_1yr_DTT ~ Predicted_Height_at_DTT + 
    Race + I(Time_to_Approval/365.25) + Interim_Meds_Started + 
    Diagnosis + Testing + VSP_Patient, data = h_mod)
Frequencies of Missing Values Due to Each Variable
Predicted_Height_1yr_DTT  Predicted_Height_at_DTT                     Race 
                       0                        0                        0 
        Time_to_Approval     Interim_Meds_Started                Diagnosis 
                       0                        7                        0 
                 Testing              VSP_Patient 
                       0                        0
Model Likelihood
Ratio Test		 Discrimination
Indexes
Obs 363 LR χ2 872.19 R2 0.910
σ 0.2698 d.f. 14 R2adj 0.906
d.f. 348 Pr(>χ2) 0.0000 g 0.905

Residuals

     Min       1Q   Median       3Q      Max 
-0.70942 -0.16826  0.01147  0.16071  1.24445
β S.E. t Pr(>|t|)
Intercept   0.1982  0.0492 4.03 <0.0001
Predicted_Height_at_DTT   0.9058  0.0170 53.30 <0.0001
Race=Black  -0.1711  0.0634 -2.70 0.0073
Race=Other  -0.0435  0.0419 -1.04 0.2998
Time_to_Approval   0.0052  0.0604 0.09 0.9317
Interim_Meds_Started=Yes  -0.0278  0.0580 -0.48 0.6318
Diagnosis=Chronic kidney disease   0.1136  0.1249 0.91 0.3639
Diagnosis=Idiopathic short stature  -0.0971  0.0349 -2.78 0.0057
Diagnosis=Noonan Syndrome  -0.3741  0.2715 -1.38 0.1691
Diagnosis=Panhypopituitarism  -0.0209  0.1050 -0.20 0.8426
Diagnosis=Prader-Willi syndrome   0.2411  0.0884 2.73 0.0067
Diagnosis=Small for Gestational age  -0.1114  0.0512 -2.18 0.0302
Diagnosis=Turner's syndrome  -0.1043  0.0694 -1.50 0.1336
Testing=Patient tested  -0.0043  0.0484 -0.09 0.9287
VSP_Patient=Yes  -0.0523  0.0299 -1.75 0.0810