Unmasking Alternative Funding Programs Comparator Study

1 Study Overview

1.1 Background

A growing number of self-employed companies have started using Alternative Funding Programs (AFPs) to completely exclude specialty medications (used to treat complex chronic, expensive conditions like cancer, MS, and RA) from being covered by a patient’s pharmacy benefit. These AFPs work by forcing patients to sign up for manufacturer assistance (meant for uninsured patients) or getting the medication internationally.

There is currently a complete lack of data on the frequency of these programs and their impact on patients.

1.2 Outcomes

The primary outcome was medication access, defined as the referred medication, or an appropriate alternative, being covered (by insurance or alternative means) so the patient could access treatment. Secondary outcomes were how the medication was accessed, time to access (defined as time from referral to the estimated date patient received medication), the occurrence of a gap in therapy, rate of manufacturer patient assistance program (PAP) application, and pharmacist time spent on the access process.

1.3 Inclusions and Exclusions

  • All patients identified to be in an AFP in 2024 or 2025 (not required to be the first time they are identified).

  • Comparator group of non-AFP patients that were randomly selected at a 2:1 ratio with the AFP patients matched by disease state and patient status. Patients with carve out plans or maximizers were excluded from the selection process.

2 Descriptive Statistics

2.1 Demographics

Characteristic N
AFP Status
AFP
N = 2991		 Non-AFP
N = 5471
Year 846

    2024
227 (75.9%) 473 (86.5%)
    2025
72 (24.1%) 74 (13.5%)
Age 816

    Mean (SD)
45.6 (14.8) 51.1 (17.3)
    Median (Q1 - Q3)
47.0 (35.0 - 57.0) 53.0 (39.0 - 63.0)
    Min - Max
2.0 - 83.0 4.0 - 93.0
    Missing
11 19
Sex 846

    Male
109 (36.5%) 252 (46.1%)
    Female
189 (63.2%) 293 (53.6%)
    Other
1 (0.33%) 2 (0.37%)
Ethnicity 828

    Not Hispanic, Latino/a, or Spanish origin
257 (91.5%) 484 (88.5%)
    Hispanic or Latino
5 (1.78%) 14 (2.56%)
    Unknown
1 (0.36%) 16 (2.93%)
    Decline to Answer
4 (1.42%) 15 (2.74%)
    Mexican, Mexican American, or Chicano/a
1 (0.36%) 0 (0%)
    Puerto Rican
1 (0.36%) 1 (0.18%)
    Cuban
0 (0%) 1 (0.18%)
    Other Hispanic, Latino/a, or Spanish origin
2 (0.71%) 4 (0.73%)
    Not available
10 (3.56%) 12 (2.19%)
    Missing
18 0
Race 846

    Black or African American
21 (7.02%) 65 (11.9%)
    Other
10 (3.34%) 24 (4.39%)
    Unknown
10 (3.34%) 11 (2.01%)
    White
258 (86.3%) 447 (81.7%)
Clinic 846

    Cardiology
6 (2.01%) 12 (2.19%)
    Dermatology
30 (10.0%) 56 (10.2%)
    Endocrinology
5 (1.67%) 10 (1.83%)
    GI/IBD
59 (19.7%) 102 (18.6%)
    Hepatology (non-hepatitis)
2 (0.67%) 4 (0.73%)
    HIV
6 (2.01%) 12 (2.19%)
    HCV
2 (0.67%) 3 (0.55%)
    Other Infectious Diseases
1 (0.33%) 2 (0.37%)
    Oncology/hematology
55 (18.4%) 98 (17.9%)
    Non-MS Neurology
9 (3.01%) 15 (2.74%)
    MS
30 (10.0%) 54 (9.87%)
    Pediatric Inflammatory
2 (0.67%) 4 (0.73%)
    Pediatric, Other
1 (0.33%) 2 (0.37%)
    Pulmonology
9 (3.01%) 18 (3.29%)
    Rheumatology
67 (22.4%) 133 (24.3%)
    Transplant
1 (0.33%) 2 (0.37%)
    Other
14 (4.68%) 20 (3.66%)
Referral Status 846

    Continuing therapy
171 (57.2%) 312 (57.0%)
    New medication
128 (42.8%) 235 (43.0%)
1 n (%)

2.2 Outcomes

Characteristic N
AFP Status
AFP
N = 2991		 Non-AFP
N = 5471
Medication Accessed 846

    No
24 (8.03%) 5 (0.91%)
    Unknown
19 (6.35%) 9 (1.65%)
    Yes
256 (85.6%) 533 (97.4%)
Referral Outcome 846

    Filled through manufacturer PAP
84 (28.1%) 23 (4.20%)
    Filled by a non-US pharmacy
56 (18.7%) 0 (0%)
    Filled by a US pharmacy (original product or alternative)
102 (34.1%) 493 (90.1%)
    Patient lost to follow up during access process
19 (6.35%) 8 (1.46%)
    Medication changed to medical benefit medication
8 (2.68%) 11 (2.01%)
    Patient changed insurance to get medication approved
12 (4.01%) 0 (0%)
    Filled through Manufacturer Bridge Program
2 (0.67%) 1 (0.18%)
    Nonstart- never accessed treatment
12 (4.01%) 5 (0.91%)
    Filled using samples or internal supply
2 (0.67%) 5 (0.91%)
    Filled at HSSP covered by internal grant
2 (0.67%) 0 (0%)
    Transitioned care during access process
0 (0%) 1 (0.18%)
Time to Access 789

    Mean (SD)
42.7 (36.0) 14.6 (18.7)
    Median (Q1 - Q3)
33.0 (19.0 - 55.0) 9.0 (5.0 - 17.0)
    Min - Max
0.0 - 235.0 0.0 - 225.0
    Missing
43 14
1 n (%)

2.3 PAP and appeals

Characteristic N
AFP Status
AFP
N = 2991		 Non-AFP
N = 5471
Applied for PAP 845

    No
104 (34.8%) 499 (91.4%)
    Yes
195 (65.2%) 47 (8.61%)
    Missing
0 1
PAP approved 242

    No
107 (54.9%) 6 (12.8%)
    Yes
88 (45.1%) 41 (87.2%)
    Missing
104 500
PAP denial reason 113

    Overqualified financially
9 (8.41%) 4 (66.7%)
    Identified AFP enrollment
53 (49.5%) 0 (0%)
    Incomplete PAP enrollment paperwork
4 (3.74%) 2 (33.3%)
    Unknown
24 (22.4%) 0 (0%)
    Other
17 (15.9%) 0 (0%)
    Missing
192 541
Samples 846

    Yes
34 (11.4%) 11 (2.01%)
    No
258 (86.3%) 534 (97.6%)
    Unsure
7 (2.34%) 2 (0.37%)
Override fill provided 299

    No
258 (86.3%) 0 (NA%)
    Yes
41 (13.7%) 0 (NA%)
    Missing
0 547
PA submitted 846

    No
85 (28.4%) 207 (37.8%)
    Yes
214 (71.6%) 340 (62.2%)
Appeal submitted 846

    No
278 (93.0%) 525 (96.0%)
    Yes
21 (7.02%) 22 (4.02%)
1 n (%)

2.4 Gaps

Only for patients continuing therapy

Characteristic N
AFP Status
AFP
N = 1711		 Non-AFP
N = 3121
Treatment gap 483

    Yes
49 (28.7%) 6 (1.92%)
    No
104 (60.8%) 295 (94.6%)
    Unknown
18 (10.5%) 11 (3.53%)
Number of days not covered 55

    0-30 days
20 (40.8%) 4 (66.7%)
    31-90 days
12 (24.5%) 2 (33.3%)
    >90 days
6 (12.2%) 0 (0%)
    Unknown
11 (22.4%) 0 (0%)
    Missing
122 306
1 n (%)

2.5 PharmD time

Characteristic N
AFP Status
AFP
N = 299		 Non-AFP
N = 547
PharmD time spent on access 846

    Mean (SD)
3.4 (3.5) 1.2 (1.1)
    Median (Q1 - Q3)
2.0 (1.0 - 4.0) 1.0 (0.5 - 2.0)
    Min - Max
0.5 - 30.0 0.0 - 10.0

2.6 Poor clinical outcome

Characteristic N
AFP Status
AFP
N = 2991		 Non-AFP
N = 5471
Any poor clinical outcomes 846

    No poor clinical outcome
272 (91.0%) 540 (98.7%)
    Poor clinical outcome
27 (9.03%) 7 (1.28%)
Number of poor clinical outcomes 846

    0
272 (91.0%) 540 (98.7%)
    1
27 (9.03%) 7 (1.28%)
Disease progression/worsening/flare 846 23 (7.69%) 6 (1.10%)
Disease-related non-scheduled clinic visit 846 3 (1.00%) 0 (0%)
Disease-related urgent care/ED visit 846 1 (0.33%) 0 (0%)
Disease-related hospitalization 846 0 (0%) 1 (0.18%)
None of the above 846 168 (56.2%) 525 (96.0%)
Unsure 846 102 (34.1%) 14 (2.56%)
1 n (%)

2.7 Human Resources Communication

Characteristic N
AFP Status
AFP
N = 2991		 Non-AFP
N = 5471
Communicated with HR department during access process 298

    Yes
18 (6.04%) 0 (NA%)
    No
110 (36.9%) 0 (NA%)
    Unknown
170 (57.0%) 0 (NA%)
Appeal to override AFP process 846 8 (2.68%) 0 (0%)
Voiced concern about the AFP process 846 5 (1.67%) 0 (0%)
Discussed explanation of benefits/AFP process 846 5 (1.67%) 0 (0%)
Other 846 1 (0.33%) 0 (0%)
Unknown 846 3 (1.00%) 0 (0%)
1 n (%)

2.8 Method of AFP Identification by HSSP

Characteristic N N = 2991
Method of AFP identification 299
    Access approved, but forced to sign up for AFP
68 (22.7%)
    Access approved, patient must apply for PAP
5 (1.67%)
    Access denied, patient must apply for PAP
15 (5.02%)
    Access denied, patient must contact AFP
53 (17.7%)
    AFP contacted clinic for therapy change
5 (1.67%)
    AFP contacted clinic to apply for PAP
14 (4.68%)
    AFP contacted patient directly
9 (3.01%)
    Identified by PAP
26 (8.70%)
    No option to submit PA or appeal, directed to PAP
45 (15.1%)
    On BI, ineligible for HSSP, contact AFP
31 (10.4%)
    Prescription required to be sent internationally
21 (7.02%)
    Other
5 (1.67%)
    Unknown
2 (0.67%)
1 n (%)

3 Analysis

Note

Patients with an outcome referral of “Therapy no longer needed” are removed from every section below

3.1 Outcome Descriptive Statistics

3.1.1 Table

Characteristic N
afp
Non-AFP
N = 5471		 AFP
N = 2991
Medication Accessed 846

    No
5 (0.91%) 24 (8.03%)
    Unknown
9 (1.65%) 19 (6.35%)
    Yes
533 (97.4%) 256 (85.6%)
Time to Access 789

    Mean (SD)
14.6 (18.7) 42.7 (36.0)
    Median (Q1 - Q3)
9.0 (5.0 - 17.0) 33.0 (19.0 - 55.0)
    Min - Max
0.0 - 225.0 0.0 - 235.0
    Missing
14 43
Any poor clinical outcomes 846

    No poor clinical outcome
540 (98.7%) 272 (91.0%)
    Poor clinical outcome
7 (1.28%) 27 (9.03%)
Referral Outcome 846

    Filled through manufacturer PAP
23 (4.20%) 84 (28.1%)
    Filled by a non-US pharmacy
0 (0%) 56 (18.7%)
    Filled by a US pharmacy (original product or alternative)
493 (90.1%) 102 (34.1%)
    Patient lost to follow up during access process
8 (1.46%) 19 (6.35%)
    Medication changed to medical benefit medication
11 (2.01%) 8 (2.68%)
    Patient changed insurance to get medication approved
0 (0%) 12 (4.01%)
    Filled through Manufacturer Bridge Program
1 (0.18%) 2 (0.67%)
    Nonstart- never accessed treatment
5 (0.91%) 12 (4.01%)
    Filled using samples or internal supply
5 (0.91%) 2 (0.67%)
    Filled at HSSP covered by internal grant
0 (0%) 2 (0.67%)
    Transitioned care during access process
1 (0.18%) 0 (0%)
1 n (%)

3.1.2 Figures

3.2 Primary Outcome - Medication Access

3.2.1 Descriptives by status and afp

Group Non-AFP / Continuing therapy
N = 3121		 AFP / Continuing therapy
N = 1711		 Non-AFP / New medication
N = 2351		 AFP / New medication
N = 1281
med_access



    No 0 (0.0%) 11 (6.4%) 5 (2.1%) 13 (10.2%)
    Unknown 2 (0.6%) 6 (3.5%) 7 (3.0%) 13 (10.2%)
    Yes 310 (99.4%) 154 (90.1%) 223 (94.9%) 102 (79.7%)
1 n (%)

3.2.2 Mixed effects logistic regression model

Medication access has 28 values of “unknown” for certain patients, but a logistic mixed model requires a strictly two-level outcome. To proceed, several approaches were taken.

Mixed-Effects Logistic Regression Results: Medication Access (Unknown is dropped)
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) 4.68 107.52 43.32 266.87 <0.001
afpAFP -2.31 0.10 0.04 0.26 <0.001
3.2.2.0.0.1 Predicted probabilities and risk ratio

Predicted probabilities

95% CI
Predicted Probability Lower Bound Upper Bound
Non-AFP 0.9908626 0.9826317 0.9990936
AFP 0.9151272 0.8778527 0.9524017

Risk ratio

95% CI
Risk Ratio Lower Bound Upper Bound
AFP / Non-AFP 0.9235662 0.8861072 0.9610252
Mixed-Effects Logistic Regression Results: Medication Access (Unknown is No)
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) 3.65 38.39 19.81 74.41 <0.001
afpAFP -1.91 0.15 0.08 0.28 <0.001
3.2.2.0.0.2 Predicted probabilities and risk ratio

Predicted probabilities

95% CI
Predicted Probability Lower Bound Upper Bound
Non-AFP 0.9819209 0.9701729 0.9936689
AFP 0.8892030 0.8384155 0.9399905

Risk ratio

95% CI
Risk Ratio Lower Bound Upper Bound
AFP / Non-AFP 0.9055749 0.8578751 0.9532748
Mixed-Effects Logistic Regression Results: Medication Access (Unknown is Yes)
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) 4.69 109.37 44.22 270.50 <0.001
afpAFP -2.25 0.11 0.04 0.28 <0.001
3.2.2.0.0.3 Predicted probabilities and risk ratio

Predicted probabilities

95% CI
Predicted Probability Lower Bound Upper Bound
Non-AFP 0.9908923 0.9827199 0.9990646
AFP 0.9198320 0.8853187 0.9543452

Risk ratio

95% CI
Risk Ratio Lower Bound Upper Bound
AFP / Non-AFP 0.9282866 0.893432 0.9631412

Medication access has 28 values of “unknown” for certain patients, but a logistic mixed model requires a strictly two-level outcome. To proceed, several approaches were taken.

Mixed-Effects Logistic Regression Results: Medication Access (Unknown is dropped)
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) 5.24 188.16 62.99 562.12 <0.001
afpAFP -2.34 0.10 0.04 0.25 <0.001
referralNew medication -0.98 0.38 0.17 0.84 0.017
3.2.2.0.0.4 Predicted probabilities and risk ratio

Predicted probabilities

95% CI
Predicted Probability Lower Bound Upper Bound
Non-AFP Continuing therapy 0.9943195 0.9881381 1.0005010
Non-AFP New medication 0.9850452 0.9706927 0.9993978
AFP Continuing therapy 0.9438998 0.9040783 0.9837213
AFP New medication 0.8635993 0.7904286 0.9367700

Risk ratio

95% CI
Risk Ratio Lower Bound Upper Bound
AFP / Non-AFP Continuing therapy 0.9492922 0.9118097 0.9867747
AFP / Non-AFP New medication 0.8767103 0.8054664 0.9479542
Mixed-Effects Logistic Regression Results: Medication Access (Unknown is No)
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) 4.21 67.15 31.76 141.99 <0.001
afpAFP -1.93 0.14 0.08 0.27 <0.001
referralNew medication -1.01 0.37 0.20 0.67 0.001
3.2.2.0.0.5 Predicted probabilities and risk ratio

Predicted probabilities

95% CI
Predicted Probability Lower Bound Upper Bound
Non-AFP Continuing therapy 0.9882299 0.9795205 0.9969392
Non-AFP New medication 0.9684434 0.9482179 0.9886690
AFP Continuing therapy 0.9240837 0.8816615 0.9665058
AFP New medication 0.8164875 0.7373469 0.8956281

Risk ratio

95% CI
Risk Ratio Lower Bound Upper Bound
AFP / Non-AFP Continuing therapy 0.9350898 0.8964055 0.9737741
AFP / Non-AFP New medication 0.8430926 0.7678088 0.9183764
Mixed-Effects Logistic Regression Results: Medication Access (Unknown is Yes)
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) 5.22 185.19 62.25 550.91 <0.001
afpAFP -2.27 0.10 0.04 0.27 <0.001
referralNew medication -0.92 0.40 0.18 0.90 0.026
3.2.2.0.0.6 Predicted probabilities and risk ratio

Predicted probabilities

95% CI
Predicted Probability Lower Bound Upper Bound
Non-AFP Continuing therapy 0.9940790 0.9876623 1.0004958
Non-AFP New medication 0.9853126 0.9711798 0.9994455
AFP Continuing therapy 0.9456558 0.9067631 0.9845485
AFP New medication 0.8742639 0.8062781 0.9422498

Risk ratio

95% CI
Risk Ratio Lower Bound Upper Bound
AFP / Non-AFP Continuing therapy 0.9512884 0.9148134 0.9877633
AFP / Non-AFP New medication 0.8872960 0.8213227 0.9532693

3.2.3 Figure

This figures uses the results from counting unknown as no access.

3.3 Secondary Outcomes

3.3.1 Poor clinical outcome

  • Outcome events were limited: 27 cases among 299 AFP participants and 7 cases among 547 non-AFP participants.

  • Mixed-effects logistic regression may be unstable due to the small number of outcome events within sites.

  • A standard logistic regression model without random effects was fit as a sensitivity analysis; results were similar, indicating that sparse site-level may not affect conclusions.

3.3.1.1 Descriptives by status and afp

Group Non-AFP / Continuing therapy
N = 3121		 AFP / Continuing therapy
N = 1711		 Non-AFP / New medication
N = 2351		 AFP / New medication
N = 1281
Any poor clinical outcomes



    No poor clinical outcome 311 (99.7%) 158 (92.4%) 229 (97.4%) 114 (89.1%)
    Poor clinical outcome 1 (0.3%) 13 (7.6%) 6 (2.6%) 14 (10.9%)
1 n (%)
Mixed-Effects Logistic Regression Results: Poor clinical outcome
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) -4.85 0.01 0.00 0.02 <0.001
afpAFP 2.11 8.22 3.55 19.03 <0.001
referralNew medication 0.72 2.05 0.98 4.29 0.058 
Logistic Regression Model

lrm(formula = poor_clinc_any ~ afp + referral, data = moddat, 
    x = TRUE, y = TRUE)

                                  Model Likelihood     Discrimination    Rank Discrim.    
                                        Ratio Test            Indexes          Indexes    
     Obs                 846    LR chi2      32.70     R2       0.132    C       0.772    
 No poor clinical outcome812    d.f.             2     R2(2,846)0.036    Dxy     0.545    
     Poor clinical outcome34    Pr(> chi2) <0.0001    R2(2,97.9)0.269    gamma   0.675    
     Cluster on  moddat$site                           Brier    0.037    tau-a   0.042    
     Clusters             11                                                              
     max |deriv|       1e-09                                                              

                        Coef    S.E.   Wald Z Pr(>|Z|)
Intercept               -4.7043 0.5477 -8.59  <0.0001 
afp=AFP                  2.0463 0.6094  3.36  0.0008  
referral=New medication  0.6978 0.3745  1.86  0.0624

3.3.2 Time to access

3.3.2.1 Descriptives by status and afp

Group Non-AFP / Continuing therapy
N = 312		 AFP / Continuing therapy
N = 171		 Non-AFP / New medication
N = 235		 AFP / New medication
N = 128
Time to Access



    Mean (SD) 12.8 (15.3) 44.5 (37.7) 17.2 (22.2) 39.9 (33.4)
    Median (Q1 - Q3) 8.0 (4.3 - 14.0) 35.0 (19.0 - 61.3) 11.0 (6.0 - 21.0) 29.5 (20.0 - 47.5)
    Min - Max 0.0 - 95.0 0.0 - 235.0 0.0 - 225.0 3.0 - 162.0
    Missing 2 17 12 26
Characteristic N
outcome_referral
 Nonstart- never accessed treatment
N = 6		 Other
N = 0		 Filled using samples or internal supply
N = 0		 Filled at HSSP covered by internal grant
N = 0		 Transitioned care during access process
N = 0
Filled through manufacturer PAP
N = 29		 Filled by a non-US pharmacy
N = 28		 Filled by a US pharmacy (original product or alternative)
N = 222		 Patient lost to follow up during access process
N = 4		 Medication changed to medical benefit medication
N = 3		 Patient changed insurance to get medication approved
N = 5		 Filled through Manufacturer Bridge Program
N = 3		 Therapy no longer needed/PURSUED FOR CLINICAL REASONS DURING ACCESS PROCESS
N = 0
Time to Access 285












    Mean (SD)
50.9 (36.7) 52.7 (38.6) 18.3 (22.5) NA (NA) 58.0 (55.7) NA (NA) 21.7 (13.5) NA (NA) NA (NA) NA (NA) NA (NA) NA (NA) NA (NA)
    Median (Q1 - Q3)
43.0 (29.0 - 67.0) 45.5 (18.5 - 86.5) 10.5 (6.0 - 21.0) NA (NA - NA) 32.0 (20.0 - 122.0) NA (NA - NA) 22.0 (8.0 - 35.0) NA (NA - NA) NA (NA - NA) NA (NA - NA) NA (NA - NA) NA (NA - NA) NA (NA - NA)
    Min - Max
4.0 - 169.0 6.0 - 128.0 0.0 - 149.0 Inf - -Inf 20.0 - 122.0 Inf - -Inf 8.0 - 35.0 Inf - -Inf Inf - -Inf Inf - -Inf Inf - -Inf Inf - -Inf Inf - -Inf
    Missing
0 0 0 4 0 5 0 0 6 0 0 0 0

3.3.2.2 Modeling

  • Time to access is censored if patients are lost to followup or transitioned care

  • Patients for whom medication is not able to be accessed are excluded from the analysis as they are not applicable according to the table.

    • Nonstart

    • Therapy no longer needed

    • Patient changed insurance to get medication approved

  • 817 patients are included in the analysis

         
          censor dont censor
  No           0          29
  Unknown     28           0
  Yes          0         789
3.3.2.2.0.1 Kaplan-Meier curves

3.3.2.2.0.2 Model
HR 95% CI p-value
afp


    Non-AFP — —
    AFP 0.30 0.25 - 0.35 <0.001
referral


    Continuing therapy — —
    New medication 0.88 0.75 - 1.03 0.121
frailty(site)

<0.001
Abbreviations: CI = Confidence Interval, HR = Hazard Ratio

3.3.3 Pharmacist Time

3.3.3.1 Outcome

Characteristic N
afp
Non-AFP
N = 547		 AFP
N = 299
PharmD time spent on access 846

    Mean (SD)
1.2 (1.1) 3.4 (3.5)
    Median (Q1 - Q3)
1.0 (0.5 - 2.0) 2.0 (1.0 - 4.0)
    Min - Max
0.0 - 10.0 0.5 - 30.0

3.3.4 Treatment gap

3.3.4.1 Outcome

Characteristic N
afp
Non-AFP
N = 5471		 AFP
N = 2991
Treatment Gap 489

    Yes
6 (1.89%) 49 (28.7%)
    No
301 (94.7%) 104 (60.8%)
    Unknown
11 (3.46%) 18 (10.5%)
    Missing
229 128
1 n (%)
Characteristic N
Referral Status
Continuing therapy
N = 4831		 New medication
N = 3631
Treatment Gap 489

    Yes
55 (11.4%) 0 (0%)
    No
399 (82.6%) 6 (100.0%)
    Unknown
29 (6.00%) 0 (0%)
    Missing
0 357
AFP 846

    Non-AFP
312 (64.6%) 235 (64.7%)
    AFP
171 (35.4%) 128 (35.3%)
1 n (%)

3.3.4.2 Mixed effects logistic regression models

Note

  • Only including patients continuing therapy

  • Low event rate, but model appears fairly stable despite large coefs

  • Some patients were marked as “Unknown” - Treated them similiarly as the Unknowns for the primary outcome where a model is created for three cases.

Mixed-Effects Logistic Regression Results: Treatment gap outcome (Drop unknowns)
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) -3.694 0.025 0.010 0.064 <0.001
afpAFP 3.353 28.582 11.377 71.807 <0.001
Mixed-Effects Logistic Regression Results: Treatment gap outcome (Drop unknowns)
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) -2.740 0.065 0.035 0.119 <0.001
afpAFP 2.531 12.567 6.902 22.879 <0.001
Mixed-Effects Logistic Regression Results: Treatment gap outcome (Drop unknowns)
95% CI for OR
Term Log(OR) Odds Ratio Lower Upper P-value
(Intercept) -3.721 0.024 0.01 0.061 <0.001
afpAFP 3.187 24.205 9.85 59.480 <0.001

3.3.5 PAP approval

AFP Approval Status N Percent
Non-AFP Approved 41 0.0749543
Non-AFP Denied 6 0.0109689
Non-AFP Did not apply 499 0.9122486
Non-AFP NA 1 0.0018282
AFP Approved 88 0.2943144
AFP Denied 107 0.3578595
AFP Did not apply 104 0.3478261