M E M O R A N D U M

DEPARTMENT OF HEALTH AND HUMAN SERVICES PUBLIC HEALTH SERVICE FOOD AND DRUG ADMINISTRATION CENTER FOR DRUG EVALUATION AND RESEARCH

                                                                             

DATE:                        February 2, 2004

 

TO:                              Paul Seligman, M.D., M.P.H., Acting Director

                                    Anne Trontell, M.D., M.P.H., Deputy Director

Office of Drug Safety

Immediate Office, HFD-400

 

THROUGH:                Mark Avigan, M.D., C.M., Director

                                    Division of Drug Risk Evaluation, HFD-430

                                    Gerald DalPan, M.D., M.H.S., Director

                                    Division of Surveillance, Research, and Communication Support, HFD-410

                                   

FROM:                        Allen Brinker, M.D., M.S., Epidemiologist Team Leader

                                    Parivash Nourjah, Ph.D., Epidemiologist

                                    Karen Lechter, Ph.D., Survey Analyst

                                    Division of Drug Risk Evaluation, HFD-430

                                    Division of Surveillance, Research, and Communication Support, HFD-410

                                    Office of Drug Safety

 

SUBJECT:                  PID D030417

Drug: Isotretinoin

Topic: Review of Isotretinoin Patient Survey Materials and Data with Interest in Compliance During the First Year of the System to Manage Accutane Related Teratology (SMART) Program.

 

 

EXECUTIVE SUMMARY

 

The analyses presented herein are based upon data supplied by the primary drug sponsor, Hoffmann-La Roche^[1], and subcontractors Slone Epidemiology Unit, Boston University Medical School [Slone] and The Degge Group with S.I. International [Degge/SI].  Other data resources include the IMS HEALTH National Prescription Audit (NPA) and National Disease and Therapeutic Index (NDTI).

 

The main findings of the report based on the Divison of Drug Risk Evaluation (DDRE) analyses are as follows:

 

·       Absolute participation in the Accutane Patient Surveys [Survey(s)] increased from 16% - 19% in the year before SMART to 22%-26% in the first year of SMART.

 

·       In comparison to two observational cohorts of Accutane-brand isotretinoin and/or generic isotretinoin recipients, the youngest users of isotretinoin appear to under-represented among participants in the Degge/SI cohort and this difference appears clinically significant.  Further comparison to these observational data also suggests patients receiving isotretinoin from non-dermatologists appear to be underrepresented in the Survey. In general, results from the patient survey should be considered with caution due to possible error from a low enrollment rate, recall bias, social desirability bias, bias due to both unit non-response and item non-response, and poor questionnaire design.

 

·       In the Degge/SI cohort, 76% of Survey participants reported they signed two consent forms, 4% reported they signed only one form, 9% reported that they did not sign any consent forms, and 11% were uncertain about whether they signed a consent form or not.

 

·       In the Degge/SI cohort, 92% of Survey participants reported that they received a prescription with yellow Acutane Qualification sticker. The remaining women reported there was no sticker, did not answer the question, or reported someone else handled the prescription for them.  

 

·       Data reported from the Slone Epidemiology Group suggests that there has been some improvement in the rate of ANY pregnancy testing prior to initiating therapy with isotretinoin, from 77%-85% to 91% in the two year interval 2QTR2001 through 1QTR2003.  However, the improvement appears to have plateaued shortly after new labeling requirements (April 1, 2002).

 

·       In the Degge/SI cohort, approximately 68% of apparently fertile and sexually active women reported two pregnancy tests prior to initiation of isotretinoin, a labeled requirement for isotretinoin therapy.  Approximately 7% of apparently fertile and sexually active women reported no pregnancy testing prior to initiation of isotretinoin.  Of menstruating, apparently fertile sexually active women, 28% had a pregnancy test within the first 5 days of their menstrual period immediately before starting isotretinoin.

 

·       In Quarterly reports to the sponsor and FDA, the Slone Epidemiology Group has reported a low (<=1%) prevalence of sexual activity without birth control based on a denominator of ALL Survey respondents.  After restriction to a denominator of apparently fertile, 15-45 yr-old sexually active participants, this rate becomes ~3%.

 

·       In the Degge/SI cohort, 4.2% of apparently fertile and sexually active 15-45 year-old participants reported no form of birth control.  In total, 46.4% of apparently fertile and sexually active 15-45 year-old participants reported use of “appropriate” birth control consisting of two methods, at least one of which is a “primary” birth control method (e.g., oral contraception).

 

·       Regarding the presence of a prescription Qualification sticker at initiation of isotretinoin treatment, pregnancy testing was generally high both in the presence of a sticker (91%) and in the absence of a sticker (91%) for apparently fertile, 15 to 45 year old survey participants. No difference was seen after imputation for records with missing data or upon restriction to sexually active Survey participants. Overall 9% of Survey participants who reported a qualification sticker was present also indicated a pregnancy test was not done.

 

·       The presence of a prescription Qualification sticker did not correlate with use of birth control at initiation of isotretinoin treatment.  Among sexually active, apparently fertile 15 to 45 year old Survey participants, any birth control was noted in 97% of enrollees with a sticker and 96% of enrollees without a sticker. No difference was seen after imputation for records with missing data.

 

·       Review of data reported from the Slone Epidemiology Group suggests that there has been some improvement in the rate of ANY pregnancy testing during isotretinoin therapy, from ~70% to ~85% in the two year interval 2QTR2001 through 1QTR2003.  However, the increase appears to have plateaued soon after initiation of the SMART program and remains at ~85%.  Based on DDRE analyses of the Degge/SI DAT2 cohort, 81% of apparently fertile, 15 to 44 years old Survey participants report monthly pregnancy testing in two consecutive months during treatment with isotretinoin.

 

·       The presence of a prescription Qualification sticker did not appear to correlate with performance of the pregnancy test during isotretinoin therapy.  Pregnancy testing was generally high both in the presence (99%) and in the absence (98%) of a sticker for apparently fertile, 15 to 45 year old survey participants. Similarly, the presence of a prescription qualification sticker did not appear to effect compliance with birth control during isotretinoin therapy.  Among sexually active, apparently fertile 15 to 45 year old survey participants, any birth control was noted in 99% of enrollees with a sticker and 100% of enrollees without a sticker. No difference was seen after imputation for records with missing data.

 

·       DDRE has reviewed additional information to the sponsor’s “1 Year Report” and Degge/SI dataset in the form of Quarterly reports for 2QTR2003 and 3Qtr2003 from Degge/SI (on Accutane-brand isotretinoin) and Slone Epidemiology Group (for both Accutane-brand isotretinoin and generic isotretinoin). Due to differences in survey, survey methods, and operating procedures, it is problematic to compare across these disparate surveys.  However, a review of selected and generally comparable variables (e.g., pregnancy testing, birth control practices) suggests similar compliance with these selected attributes as for April 1, 2002 through March 30, 2003 as described immediately above.  Independent FDA validation of results in quarterly reports is not possible.

 

·       The Degge/SI dataset contains 15 reports of pregnancy among 4277 women on their first-course of isotretinoin therapy.  The observed pregnancy rate for first-course users within the Degge/SI cohort is thus 15/4277 = 3.5 / 1000. Since this rate is censored, it likely represents an underestimate of the rate realized when all these Survey participants complete follow-up. However, this rate is virtually identical to the rate as reported by researchers from the Slone Epidemiology Group for their Accutane Survey cohort of 2.9 per 1,000 women.

 

INTRODUCTION

 

Goals/Objectives of SMART

 

The System to Manage Accutane Related Teratogenicity (SMART) program was initiated on April 10, 2002 to address two main goals raised at a public meeting in September 2000.^[2]

1) No woman should begin Accutane therapy if she is pregnant;

2) No pregnancies should occur among women taking Accutane.

 

Tools of the SMART program; Required Steps in Process and Patient labeling^[3]

 

The S.M.A.R.T. program requires the following:

·       Prescribers read the S.M.A.R.T. "Guide to Best Practices" provided by Roche, and then sign and return the Letter of Understanding certifying their knowledge of the measures to minimize fetal exposures to Accutane. In addition, Accutane prescribers are strongly encouraged to participate in a half-day Continuing Medical Education (CME) course developed by the manufacturer that includes specific, practical information about pregnancy prevention.

·       Prescribers receive self-adhesive Accutane Qualification Stickers which are to be attached to the prescription form. These stickers indicate to the pharmacist that the patient is "qualified" to receive Accutane, i.e., the female patient has had negative pregnancy tests, as well as education and counseling about pregnancy prevention. The pregnancy test is to be repeated every month throughout the Accutane treatment course, and no prescriptions should be given for more than a one month supply of Accutane.

·       All female patients must have two negative urine or serum pregnancy tests before the initial Accutane prescription is written.  Additionally, patients must have a negative pregnancy test prior to receiving subsequent prescriptions, regardless of whether or not they are sexually active. Patients who are, or might become, sexually active with a male partner must also select and use two forms of effective contraception simultaneously for at least one month prior to initiation of Accutane therapy, during therapy, and for one month following discontinuation of therapy.  Patients are required to sign a Patient Information/Consent form about Accutane and birth defects, in addition to the General Consent Form that all patients receive about other potentially serious risks (e.g., suicidality, etc.). Finally, female patients must be given the opportunity to enroll in the Accutane Survey. This confidential survey collects data on the utility of S.M.A.R.T. in reducing pregnancies among female Accutane users.  This survey will help to identify aspects of S.M.A.R.T. that could be improved upon.

·       Pharmacists dispense Accutane only upon presentation of a prescription with the special Accutane Qualification Sticker.  Moreover, pharmacists dispense a maximum one-month supply of Accutane, fill prescriptions within seven days from the date of "qualification," (defined above) and provide a Medication Guide for patients with each Accutane prescription. Requests for refills (i.e. more Accutane without a new prescription) and phoned-in prescriptions are not to be filled.

 

METHODS

 

Background on Patient Survey

 

Accutane (isotretinoin, Hoffmann-La Roche) was initially marketed in the US in 1982.  Isotretinoin is an effective medication for the treatment of severe recalicitrant acne, although it has potent teratogenic effects.  To minimize the risk of patients becoming pregnant while taking isotretinoin and avoiding exposure during pregnancy, the sponsor designed and implemented the Accutane Pregnancy Prevention Program (PPP) in 1988.  In 1989, the Slone Epidemiology Unit of Boston University designed and conducted the Survey of Accutane Use in Women (Slone Survey / Accutane [Patient] Survey) to evaluate the PPP program.  The general format^[4] of the survey remains largely unchanged.   In the original survey, women who chose to enroll in the program were randomized (at initiation of therapy) to either a During and After Treatment [DAT] arm (consisting of 2 or more questionnaires, DAT-1, DAT-2, and DAT-3) or an After Treatment [AT] arm.  The DAT arm was restricted to a total enrollment of 5,000 women per year.  Between 1989 and 1994, women in the DAT arm were queried by phone while women assigned to the AT arm were queried by mail.  Women who enrolled but failed to respond to mailed questionnaires were contacted by air courier or, if needed, telephone.  Since 1995, mailed questionnaires have been used as the primary contact instrument for both the DAT and AT arms with follow-up contact by air courier or telephone for non-responders.  The Slone Epidemiology Unit was the sole administrator of the Accutane patient survey from inception through September 2002.  During the fourth quarter of 2002, Hoffmann-La Roche re-awarded the contract for the survey of enrollees using Accutane-brand isotretinoin to Degge/SI.  Sponsors of generic isotretinoin introduced into the market during the fourth quarter of 2002, elected to use the Slone Epidemiology Group to conduct the survey, a regulatory requirement, on their behalf.  The Slone Epidemiology Unit has continued to survey women who enrolled before the SMART program, those who enroll with cards addressed to Slone Epidemiology Unit, and those who declined transfer to Degge/SI.  The DAT questionnaire was also modified to include new questions specifically designed to measure compliance with SMART.^[5]  The new questionnaire was introduced in 4QTR2002 by Degge/SI and in 2QTR2003 by Slone for generic isotretinoin.  Chart 2 provides a timeline for Accutane/generic isotretinoin patient surveys^[6].

 

Data Sources

 

The analyses presented herein are based upon data in a report supplied by the primary drug sponsor, Hoffmann-La Roche^[7] and primary dataset provided by subcontractors Slone Epidemiology Unit, Boston University Medical School [Slone] and The Degge Group with S.I. International (Degge/SI).  Other data resources include the IMS HEALTH National Prescription Audit (NPA) and National Disease and Therapeutic Index (NDTI).

RESULTS

 

1.         SURVEY PARTICIPATION, REPRESENTATIVENESS, AND LIMITATIONS

 

1.1       Absolute Participation in Patient Surveys

 

Absolute participation in the Patient Survey by quarter, as provided to FDA by Hoffmann-La Roche is shown in table 1.  These data appear to show that, although participation has generally been poor, participation is increasing. 

 

Table 1.  Trends in participation in Accutane Patient Survey per Sponsor
era	Last four quarters before SMART				First four quarters of SMART
Qtr	2Qtr01	3Qtr01	4Qtr01	1Qtr02	2Qtr02	3Qtr02	4Qtr02	1Qtr03
Survey Enrollees	7,903	7,897	8,211	10,044	8,533	7,560	10,249	10,289
New female starts	49,795	47,402	54,348	31,262	42,642	34,424	38,528	19,761
Enrollment Rate	17.3%	16.8%	15.1%	32.1%	20.0%	22.0%	26.6%	52.1%

 

To this end, the sponsor states,

 

“Accutane Survey enrollment rate of 28.2% [for the first year of SMART], based on a denominator of new female patient starts, represents an approximate 10 percentage point increase in Survey enrollment rate from the 4 quarters prior to SMART and reverses a downward trend in Survey enrollment rates observed in a retrospective review.”

 

In order to provide for an independent assessment of quarterly enrollment since initiation of the SMART program (April 1, 2001) though March 31, 2002, DDRE produced an independent estimate of unique females on Accutane therapy for the 8 quarters of interest.  These data, along with quarterly Survey enrollees (numerators) shown in Table 1, permit calculation of enrollment rates.  The approach used to calculate quarterly estimates of unique females on Accutane is based on the following assumptions: 1) there is a steady-state condition of drug usage for well-established drugs such as Accutane; and 2) the average total number of Rxs per unique individual until the end of therapy is 3.7.^[8] This estimate of average number of Rxs per unique patient is provided in a large, observational study of Accutane recipients followed to the apparent end of an Accutane treatment course.  The assumptions listed previously, and lack of proven applicability of this metric to the population of Accutane recipients are limitations of this analytic approach.  Steps taken to calculate enrollment rates were as follows: First, retail female prescription volume for the 8 quarters of interest was obtained by dividing the observed quarterly prescription volume by two and then summed.  [Sum Rx volume to females over the eight quarters of interest = 1,334,000 Rx.]  Second, the total number of unique females on Accutane for the 8 quarters of interest was estimated by dividing the total female Rx volume by 3.7.  This is outlined in the following equation:

Unique females (total for 8 quarters) = (1,334,000 Rx) (unique patient / 3.7 Rx^[9])

= 360,500

 

Finally, total female utilization (360,500) over the 8 quarters was multiplied by the fraction of total utilization for any specific quarter to result in an estimate of unique females for each quarter.

 

Table 2. Trends in participation in Accutane Patient Survey per DDRE based on estimate of unique females on therapy.
era	Last four quarters before SMART				First four quarters of SMART
Qtr	2Qtr01	3Qtr01	4Qtr01	1Qtr02	2Qtr02	3Qtr02	4Qtr02	1Qtr03
Survey Enrollees	7,903	7,897	8,211	10,044	8,533	7,560	10,249	10,289
Rx volume for quartera	377,000	338,000	394,000	399,000	295,000	248,000	292,000	325,000
Rx volume to femaleb	188,500	169,000	197,000	199,500	147,500	124,000	146,000	162,500
% of total female utilizationc	14.1%	12.7%	14.8%	15.0%	11.0%	9.3%	10.9%	12.2%
Estimate of total unique femalesd	50,900	45,700	53,200	53,900	39,900	33,500	39,500	43,900
Enrollment Rate	15.5%	17.3%	15.4%	18.6%	21.4%	22.6%	26.0%	23.4%
a  US Data Source: IMS Health, IMS National Prescription Audit PlusTM, Accutane and generic isotretinoin for 3/01 to 6/03; accessed 7/28/03. b  Based on assessment of 1:1 male: female Rx distribution; FDA custom calculations were based on IMS Health data. c  Based on division by female Rx volume of quarter by total Rx volume to females (1,334,000); FDA custom calculations were based on IMS Health data. d  Based of multiplication by percent of total female utilization of quarter by 360,500; unique female count of 360,500 based on an average of 3.7 Rxs per patient (see text); FDA custom calculations were based on IMS Health data.

 

The analyses provided by the sponsor and by DDRE are highly similar in all eight quarters before and after initiation of SMART except for 1QTR 2002 and 1QTR 2003.  In these quarters, the DDRE analysis of enrollment is approximately half of the enrollment estimate by the sponsor.  Since the two analyses use the same numerator data, the difference for the two quarters of interest must be due to differences in the denominators for 1QTR 2002 and 1QTR 2003.  This difference is captured in Tables 1 and 2.  In these tables, the respective denominators for 1QTR 2002 and 1QTR 2003 are approximately half of each other. In order to investigate trends due to seasonality, average duration of treatment in relation to season were explored within the AdvancePCS claims database.  No differences were observed based on standard calendar quarters (data not shown) for 2002.  Thus, the reason the two models result in almost identical counts for six quarters yet diverge strikingly for 1Qtr02 and 1Qtr03 is not readily apparent.  Based on DDRE analyses, absolute participation in the patient surveys increased from 16% - 19% in the year before SMART to 22%-26% in the first year after implementation of SMART.

 

1.2       Representativeness

 

The sponsor includes the following summary with regard to the representativeness of Accutane Survey participants:

 

“The representativeness of Accutane Survey respondents to all Accutane female patients in the United States was demonstrated for age, prescriber, payer and geographic region of the country. While this does not ensure absolute comparability of all Survey respondents to all Survey non-respondents, it suggests that the convenience sample is reasonably representative of the population on these selected parameters.”

 

DDRE would agree with the conclusion of the sponsor that the Survey represents a “convenience sample.”  With reference to the sponsor’s data (not shown) for clinical characteristics of Slone and Degge/SI cohorts versus data on the isotretinoin-user population at large, we would note the following:

 

• Age distribution - differences are mild but noticeable

·       Young (< 15 years) and 45+ age groups are underrepresented in the Accutane Survey populations when compared to the female Accutane population/reference population^[10] (For < 15 year olds: SEC 13.2%, Degge/SI 9.6%, and reference population 17.3%; For 45+ age group SEC 3.6%, Degge/SI 5%, and reference population 10.9%). 

·       The 20-29 year age-group is over-represented (SEC 35.7%, Degge/SI 37.2%, and reference population 24.9%). 

• Geographic Region - Generally similar, east somewhat under-represented in the survey population when compared to reference population^[11] (SEC 15.6%, Degge/SI 16.5%, and reference population 21.2%)
• Payor Type - Based solely on Degge/SI cohort; unable to compare given the large fraction (12%) of unknown responses.
• Prescriber - Data suggest that up to 20% of Accutane prescriptions are generated by non-dermatologists.  However, non-dermatologists account for only 5% of prescribers in the Slone cohort and 10% of the Degge/SI cohort.^[12]  This is also noted by the sponsor.

 

The Degge/SI primary dataset as provided to FDA by the sponsor does not include data on geographic region nor payer type.  DDRE is unable to generate tables comparable to the sponsor’s.  Data on age and prescriber was provided by the sponsor in their dataset. DDRE analyses of these data are outlined as follows. The age distribution of the Degge/SI cohort is compared to the distribution of recipients of isotretinoin (both Accutane-brand and generic) within the IMS Health NDTI^TM and with recipients of Accutane-brand isotretinoin with coverage managed by AdvancePCS (described presented previously within the Isotretinoin Utilization section of this document^[13].) In brief, the IMS Health National Disease and Therapeutic Index™ (NDTI™) is an ongoing survey designed to provide descriptive information on the patterns and treatment of disease encountered in office-based practice in the continental United States by collecting data on drug products mentioned during visits to roughly 2,000 - 3,000 office-based physicians.  The data are projected nationally to reflect national prescribing patterns.  These data are displayed in Table 3 below.  In comparison to the IMS Health NDTI™, it appears the youngest isotretinoin users are underrepresented in the Survey.  While the NDTI™ is designed to be representative, it is limited by size.  Thus, the age distribution of Degge/SI participants was also compared to isotretinoin recipients with drug benefits managed by AdvancePCS, which is not designed to be representative but is quite large.  These data also suggest that the youngest isotretinoin users appear to be under-represented among Survey participants.  Furthermore, while it is inappropriate to subject these data to statistical testing, these comparisons provide evidence that there is a large and clinically meaningful difference in Survey participation for the youngest recipients of isotretinoin.

 

 

Second, in DDRE analyses of the Degge/SI dataset, 94% of DAT-2 participants indicated their Accutane prescriber was a dermatologist.  In comparison, approximately 80% of recent isotretinoin prescriptions in the IMS Health NPA^TM were associated with dermatologists.^[14]  Thus, as shown in analyses by DDRE and the sponsor, patients receiving care from non-dermatologists are underrepresented in the Survey.  It is not known if there is differential encouragement for the patient survey between dermatologists and non-dermatologists or if compliance with isotretinoin labeling and pregnancy varies between dermatologists and non-dermatologists.

 

1.3       Overview of Study Limitations due to Internal Variability

 

In consideration of the generalizability of the Survey, the sponsor notes,

 

“Examination of invalid and absent responses to the Accutane Survey indicated relatively high proportions of invalid and absent responses to several key questions that measure S.M.A.R.T. compliance.”

 

The Slone Survey, implemented in 1989, was designed to, “assess the compliance of physicians and patients with the Accutane Pregnancy Prevention Program and to identify the rate of pregnancy during treatment with isotretinoin and during the month after treatment.” Appendix 6^[15] provides a comprehensive discussion on both the internal variability and external generalizability of the Survey. DDRE/DSRCS analyses also indicated a high percentage of missing responses on key variables. For instance, patient survey information on contraception may be incomplete or unreliable due to the complexity of the questionnaires.  In addition, some questions may be skipped inappropriately due to misinterpretation of survey instructions.  While the sponsor does not address what inference should be gleaned from “invalid and absent responses,” we take the position that these issues introduce biases, degrade interval validity, and preclude statistical testing of Survey results. Furthermore, it is our position that any changes between calendar quarters represents too short a time interval to detect meaningful changes; larger changes over time may be considered more robust, but not definitive evidence of change.

 

2.         DAT-1 SURVEY (INITIATION OF TREATMENT)

 

2.1       DAT-1 Univariate Analyses

 

2.1.1    Timing of Enrollment relative to the start of Accutane therapy

 

Timing of enrollment relative to the start of therapy with isotretinoin is important to assure accuracy of responses. The sponsor reports 73% of enrollees enroll within 1 month of starting treatment with isotretinoin.  Approximately 27% of participants enroll at a time point of > 1 month.

 

Based on a dataset provided by the sponsor, 5489 women responded to a DAT-1 questionnaire sent by Degge/SI between October 2002 and April 1, 2003.  Twenty (20) of the respondents reported no plans to start Accutane therapy and have been excluded from all further DDRE analyses.  Of the remaining 5469 respondents, 5334 women started treatment with Accutane, and 135 women reported they will start Accutane therapy later or did not provide an answer.  Of this baseline dataset of 5469 women, 80.9% completed the questionnaire within 30 days of enrollment.  Twelve percent (12%) responded later than 30 days, and 7.1% had an invalid or missing date.Thus, we concur with the sponsor in assessment that most enrollees in the Survey do so early in therapy. Among the baseline dataset of 5469 respondents, 1416 (26%) reported they had undergone a previous course of Accutane therapy.

 

2.1.2    Consent forms

 

The SMART program requires women to sign two consent forms, one required of all patients on isotretinoin and the other required only of female patients. The sponsor does not include an analysis of the Degge/SI cohort for consent.

 

Based of DDRE analyses of the Degge/SI cohort, 76.3% of women reported that they signed two consent forms, 3.7% reported they signed only one form, 9.3% reported they had not signed a consent form, and 10.7% were uncertain about whether they signed a consent form or not.

 

2.1.3    Medication Guide

 

A Medication Guide should be distributed with each prescription. The sponsor does not include an analysis of the Degge/SI cohort for receipt of a Medication Guide within the Degge/SI cohort in their “1 Year Report.”

 

Based of DDRE analyses of the Degge/SI cohort, 81% of respondents reported receipt of a Medication Guide.  Four (4) percent reported that a Medication Guide was missing and 15% reported they were unsure or did not answer the question.

 

2.1.4    Patient Recall of Qualification Sticker

 

The sponsor’s analysis of patient recall of a qualification sticker is consistent with the analysis of the Pharmacy Compliance Survey^[16] .  Based on data from the 3^rd and 4^th quarters of SMART, the sponsor notes 97% of survey enrollees recalled the presence of a qualification sticker.

 

In DDRE analysis of the Degge/SI cohort, a yellow Accutane Qualification sticker on the prescription was reported by 92.1% of the 5469 participants. Another 2.3% reported the Accutane qualification sticker to be missing and 5.6% of respondents were unsure or did not answer the question.

 

2.1.5    Compliance with 7-day Limit of Qualified Prescriptions

 

According to the new label for isotretinoin, the qualification sticker signifies that a patient was “qualified” for therapy upon receipt of the prescription and the prescription is to be filled within 7 days.  The sponsor conducted no analyses within the Degge/SI Survey on distribution of receipt and fill dates for Survey participants.

 

In DDRE analyses of Degge/SI survey data for women who started Accutane and further reported the presence of a Qualification sticker (n=4403), 89.9% initiated treatment within 7 days of receiving their Accutane prescription. Among this same group of participants, 93.2% filled their prescriptions within 7 days of receipt.  These data are compared to an analysis of the Pharmacy Compliance Survey^[17], where, after DDRE review, 94% to 99% of prescriptions with qualification stickers were filled within 7 days.

2.1.6    Pregnancy Testing

 

2.1.6.1 Any Pregnancy testing: The sponsor provided the trends in any pregnancy testing prior to starting isotretinoin therapy based on the DAT-1 survey data collected by SEC and Degge/SI. As shown in table 4 below, SEC data show that the percentage of women reporting ANY pregnancy testing during the initial four quarters of SMART appears constant at about 92%.  It follows that 7% to 8% of women reported NO pregnancy testing.

 

The data for the Degge/SI cohort are generally similar for the 4QTR02 and 1QTR03. The percentage of women reporting ANY pregnancy testing is reported as 88% and those reporting no pregnancy testing as 9%. The difference between the two analyses may be in the larger fraction of women classified as “unknown” in the Degge/SI survey (3%) versus Slone (<1%).

 

 

The sponsor also conducted analyses to examine the relation between pregnancy testing and age (data not shown).  These data suggest that pregnancy testing is related to age to some degree, with slightly higher compliance among women aged 16-39 (86 to 92% receiving any pregnancy testing).  Poorer compliance with testing at the extremes of the age distribution was among relatively small numbers of survey participants at those extremes (78 to 86% for any pregnancy testing).

 

While the sponsor provides a reference for ANY pregnancy testing (as shown in Table 4 above) based on the interval 1995-March 2002, this large time interval averages results during a time of public heightened attention and regulatory attention to reducing isotretinoin-exposed pregnancies.  Therefore, trends in pregnancy testing before start of Accutane therapy based on DAT-1 surveys have been compiled by DDRE for the four quarters prior to April 1, 2002 and the following quarters of SMART based on Slone Survey quarterly reports (Table 5). 

 

 

 

 

 

 

Table 5. Trends in Pregnancy Testing before start of Accutane therapy based on DAT-1 surveys (aggregated from Slone Survey quarterly reports) for the four quarters before SMART and the first four quarters of SMART.
Era	Last four quarters before SMART				First four quarters of SMART
Qtr	2Qtr01	3Qtr01	4Qtr01	1Qtr02		2Qtr02	3Qtr02	4Qtr02	1Qtr03
N	1101	1050	1102	1141		1116	1178	1190	966
	(%)	(%)	(%)	(%)		(%)	(%)	(%)	(%)
Any test	80	77	81	85		92	92	92	91
No Test	19	21	18	15		8	8	8	8
Unknown	1	1	1	<1		<1	<1	<1	1

 

As in the analysis supplied by the sponsor, some 8% of enrollees in the Slone Survey report no pregnancy testing prior to initiating therapy with isotretionoin following April 1, 2002.  These data also suggest that there has been some improvement in the rate of pregnancy testing prior to initiating therapy with isotretinoin, from 77%-85% to 92%.  However, the improvement is not as great when the comparison is to the recent past and appears to have stabilized / plateaued shortly after requirements per new labeling (April 2002).

 

In DDRE analyses of the Degge/SI cohort (N=5469), there were 5334 women who started treatment with Accutane. All analyses of pregnancy testing or birth control further exclude 249 women who reported a reproductive status of menopause or who had undergone hysterectomy and another 341women of age >=45 years or age <15 years. Of the remaining 4744 women, the number of available patients for any specific analysis is further decreased by exclusion of records with missing information. Of 4719 apparently fertile women aged 15 to 45 years who provided information on sexual activity, 1846 (39.1%) reported no sexual intercourse during or before treatment, 1855 (39.3%) women reported sexual activity before and during Accutane treatment, and 1018 (21.6%) women reported sexual intercourse before starting Accutane treatment but not during treatment.

 

DDRE analyses of pregnancy testing are outlined in Table 6.  Of 4596 apparently fertile women aged 15 to 45 years who provided information on pregnancy testing, 1806 (90.8%) reported pregnancy testing prior to initiating treatment with Accutane.  This is generally unaffected upon stratification by sexual activity status.  Any pregnancy testing was reported by 92.1% among women reporting sexual activity before and during Accutane treatment, 92.7% among women reporting sexual intercourse before starting Accutane treatment but not during treatment, and 88.5% among women reporting no sexual intercourse before or during Accutane treatment.

 

Table 6. Patient report of pregnancy testing, all and stratified by sexual activity status, for apparently fertile women 15-45 years of age who provided data on pregnancy testing.
	N	At least one pregnancy test	Two or more pregnancy tests
Total*	4596	90.8%	66.0%
Sexually active since starting Accutane therapy	1806	92.1%	68.3%
Sexually active before but not during Accutane therapy	993	92.7%	66.8%
Not sexually active before or during Accutane therapy	1797	88.5%	63.0%
*Excludes women who report: (1) age younger than 15 years old or older than 45 years, (2) women who report they have undergone hysterectomy or menopause, (3) women who have not started Accutane treatment, and (4) records with missing values.

 

2.1.6.2 Two pregnancy tests: According to the Accutane label, women should not get the first prescription for Accutane before assessment by 2 pregnancy tests.  While this criterion is not new, the number of pregnancy tests was not included as a question in the old Survey.  The number of pregnancy tests is now included as a question in the new Survey.  Per the sponsor’s analysis, 63% of apparently fertile participants report two pregnancy tests prior to starting Accutane (data not shown).

 

Among the subset of 1806 apparently fertile women 15-45 years of age women who reported sexual activity since starting Accutane therapy, 68.3% reported two pregnancy tests in the month prior to starting their treatment (Table 6, shaded area).

 

2.1.6.3 Pregnancy testing in relation to menses: According to new labeling for isotretinoin, the first test must be done when the prescriber decides to prescribe Accutane, and the second pregnancy test must be done during the first 5 days of the menstrual period preceding initiation of Accutane therapy, assuming the woman menstruates.  Compliance with testing around menses is generally poor at 27% (data not shown), suggesting that clinicians cannot adopt visit / treatment cycles in sequence with recommendations for testing in relation to menses.

 

To address whether the test was taken at appropriate time per approved labeling, the number of days between start of menses and start of Accutane treatment was examined on the subset of women who received pregnancy testing between these two events (Table 7). 

 

Table 7. Distribution of days between starting menses and starting Accutane treatment for apparently fertile women aged 15-45 years who had a pregnancy test sometime between these two events.
	N	0-5 days**	6-10 days	11-20 days	21-30 days	>30 days
Total*	1851	27.1%	23.9%	26.0%	15.6%	7.2%
Sexually active before and during Accutane treatment	767	28.1%	23.7%	24.6%	16.3%	7.2%
Sexually active before Accutane but not during	429	26.8%	24.5%	25.9%	15.9%	7.0%
Not sexually active before or during treatment w/Accutane	655	26.3%	23.8%	27.8%	14.7%	7.5%
*Excludes women who report: (1) age younger than 15 years old or older than 45 years, (2) women who report they have undergone hysterectomy or menopause, and (3) and women who have not started Accutane treatment. Furthermore, 1524 women were not included in the analysis because of missing values for pregnancy testing, date of starting Accutane treatment, or date(s) of menses. ** “correct” per labeling

Based on DDRE analyses, 27.1% of apparently fertile women aged 15-45 years old on whom data were available received pregnancy testing during the first 5 days of menses. As is further shown in Table 7, compliance with pregnancy testing during menses (for menstruating participants) appeared to be unaffected by sexual activity status.  [It is noteworthy that the data for about 45% of apparently fertile women aged 15-45 is unusable due to missing values for either pregnancy testing, dates of menses, or isotretinoin therapy start date.]

 

2.1.7    Sexual Activity / Birth Control

 

2.1.7.1 Sexual activity: The isotretinoin label states that women should use two separate forms of contraception for at least 1 month before and during therapy, at least one of which should be a “primary^{Error! Bookmark not defined.}” method. The sponsor investigated the reproductive status and contraceptive practices for women enrolled in the Degge/SI survey which suggest a large fraction of women report they are not sexually active (50%).

 

2.1.7.2 Birth Control: Among the participants in the Degge/SI survey, approximately 26% note they are sexually active and use 2 forms of birth control, one of which is primary.  An additional large fraction of women (18%) report they are sexually active and do not use 2 forms of birth control.  The sponsor further stratified the data by age, where non-compliance with use of “two forms” of contraception is shown to generally increase in direct relation to age through age 40 to 44 years where it peaks at 39%.  Non-compliance appears to fall thereafter, although there is only one remaining age band beyond age 40-44 years.  It should be noted that within the age band with the most members (20-29 years, n=1577)), non-compliance with 2 forms of birth control (one primary) among sexually active enrollees is reported at 20%.

 

DDRE further compared trends in fertility/contraceptive status of enrollees in the Slone survey in the four quarters before April 1, 2002 and the following four quarters (Table 8).  These data are derived from “Pregnancy Risk Category” as used in Slone Survey quarterly reports and show status (per Slone categorization) at initiation of Accutane therapy based on DAT-1 surveys.

 

Table 8. Fertility/contraceptive status of enrollees at initiation of Accutane therapy in the Slone survey in the four quarters before SMART and the first four quarters of SMART.
era	Last four quarters before SMART				First four quarters of SMART
Qtr	2Qtr01	3Qtr01	4Qtr01	1Qtr02	2Qtr02	3Qtr02	4Qtr02	1Qtr03
N	1148	1103	1153	1177	1165	1233	1234	1024
	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)
Hysterectomy /Postmenopausal	4	5	4	3	4	4	4	3
Not Sexually Active, Birth Control*	28	29	28	33	32	34	35	38
Not Sexually Active, No Birth Control	26	28	26	25	26	23	25	23
Sexually Active, Birth Control	39	36	38	37	36	36	35	25
Sexually active, No Birth Control	<1	<1	<1	<1	<1	<1	<1	<1
Unknown	1	1	1	1	1	2	2	1
*Refers to at least one method of birth control including either primary methods (e.g., oral contraception, tubal ligation, injectable/implantable/insertable/patch contraceptives, vasectomy, IUD) or secondary methods (e.g., diaphragms, latex condoms, cervical cap)

Each survey enrollee cohort contains a small group (up to 1%) classified as “sexually active, not on birth control.” In aggregated analyses of these Quarterly Reports, this 1% “rate” does not change over the SMART era and thus appears unaffected by new labeling requirements. It should be noted that this 1% “rate” is based on a denominator of ALL Survey respondents.  After restriction to a denominator of apparently fertile, 14-45 yr-old sexually active participants, this rate becomes ~3%.

 

Trends on contraceptive type (e.g. birth control pills) stratified by age bands and are included in Appendix 4^[18]. Although the data for 15 –24 year old enrollees may suggest an increase in use of a “primary” birth control method (51%-60%), variability precludes assessment of a definitive trend since initiation of SMART. These data further appear to indicate that the fraction of enrollees classified as “Not sexually active but on birth control” increased after initiation of SMART (28% to 38%).

 

Of 4719 apparently fertile women aged 15 to 45 years who provided information on sexual activity, 1846 (39.1%) reported no sexual intercourse during or before treatment, 1855 (39.3%) women reported sexual activity before and during Accutane treatment, and 1018 (21.6%) women reported sexual intercourse before starting Accutane treatment but not during treatment.  In DDRE analyses of the Degge/SI cohort, 95.8% of apparently fertile and sexually active 15-45 year-old participants reported that they used some form of birth control early in the course of treatment with Accutane (Table 9); 4.2% of such women reported no form of birth control.  This is consistent with the recalculated rate of ~3% from Slone with restriction to sexually active women.  In total, 46.4% of apparently fertile and sexually active women reported use of “appropriate” birth control consisting of two methods, at least one of which is a “primary” birth control method (e.g., oral contraception).

 

Table 9.  Distribution of birth control method used among apparently fertile women 15-45 years of age within the DAT-1.
		Any form of birth control	Any primary method of birth control	Appropriate per Accutane label**
Total	4688*	70.2%	56.8%	23.7%
Sexually active since starting Accutane therapy	1841	95.8%	83.0%	46.4%
Sexually active before but not during Accutane therapy	1008	76.0%	57.7%	24.2%
Not sexually active before or during Accutane therapy	1838	42.0%	30.0%	0.7%
*Excludes women who report: (1) age younger than 15 years old or older than 45 years, (2) women who report they have undergone hysterectomy or menopause, (3) women who have not started Accutane treatment, and (4) records with missing information on birth control. **Appropriate per Accutane label requires two forms of which one must be a “primary” method (e.g., oral contraception, tubal ligation, injectable/implantable/insertable/patch contraceptives, vasectomy, IUD). As per the Accutane label, women who claim absolute abstinence may elect not to use birth control. Thus, the latter two rows of the table may include women who chose absolute abstinence and this may result in lowering for percent appropriate. There is no question in the patient survey for abstinence: sexual activity used as a surrogate.

 

Further analyses (data not shown) were conducted to examine changes in pregnancy testing or birth control over time among apparently fertile and sexually active women aged 15-45. Between 4QTR2002 and 1QTR2003, we observed a 3% absolute increase in reporting of two pregnancy tests and a 4% absolute increase in reporting of using birth control prior to starting Accutane. Furthermore, the use of a “primary” method of birth control increased 2% and “appropriate” use of birth control declined by 2%.

 

2.2       DAT-1 Bivariate Analyses

 

The sponsor does not include bivariate analyses as outlined in the following section.  Thus, this section includes only DDRE analyses. Data (as tables) are included in this section.  The overall effect of the sticker at initiation of therapy (thus through the DAT1 instrument) is shown in DDRE analyses of sticker and pre-therapy pregnancy test (Tables 10 through 13) and sticker and birth control (Tables 14 & 15).  Due to the potential bias of imputation of missing data, these tables are shown initially with exclusion of missing records, and then with imputation of missing data as a negative result [“missing = NO”].

 

2.2.1    Relationship between Qualification Sticker and Pregnancy Testing

 

The overall effect of the sticker at initiation of therapy is shown in DDRE analyses of sticker and pre-therapy pregnancy test in Table 10.  As shown, the sticker does not appear to relate to performance of the pregnancy test, as pregnancy testing was very high both in the presence of a sticker (91%) and in the absence of a sticker (90%) for all apparently fertile women 15 to 45 year old enrollees. Nine (9) percent of Survey participants who report a Qualification sticker report no pregnancy test.

 

Table 10. Relationship between Qualification sticker and pregnancy testing for all, apparently fertile*, 15-45 year-old enrollees. [Missing=excluded]
	Qualification sticker Yes	Qualification sticker No	Row total
Pregnancy test Yes	3908 (91%)	90 (90%)	3998
Pregnancy test No	392 (9%)	10 (10%)	402
	4300	100	4400**
*excludes women reporting hysterectomy or postmenopausal **excludes 319 records with missing data for either qualification sticker or pregnancy test; pregnancy testing includes any pregnancy testing performed at physician’ office or performed at home but reported to the physician; restricted to women who reported starting Accutane therapy.

 

As shown in the following table (Table 11), imputation of “NO” for missing data does not appear to change our assessment.  Although compliance with pregnancy testing falls slightly, it does so

both in the presence of a sticker (89%) and in the absence of a sticker (85%) for all apparently fertile women 15 to 45 years old.

 

 

Table 11. Relationship between Qualification sticker and pregnancy testing for all, apparently fertile*, 15-45 year-old enrollees. [Missing=NO]
	Qualification sticker Yes	Qualification sticker No	Row Total
Pregnancy test Yes	3908 (89%)	268 (85%)	4176
Pregnancy test No	495 (11%)	48 (15%)	543
	4403	316	4719**
*excludes women reporting hysterectomy or postmenopausal **pregnancy testing includes any pregnancy testing performed at physician’s office or performed at home but reported to the physician; restricted to women who reported starting Accutane therapy.

 

No effect of the Qualification sticker is seen after restriction to sexually active, apparently fertile women 15 to 45 year old enrollees, where the delta between groups is 1% (Table 12).

 

Table 12. Relationship between Qualification sticker and pregnancy testing for Sexually active, apparently fertile*, 15-45 year-old enrollees. [Missing=excluded]
	Qualification sticker Yes	Qualification sticker No	Row Total
Pregnancy test Yes	1575 (92%)	41 (91%)	1616
Pregnancy test No	136 (8%)	4 (9%)	140
	1711	45	1756**
*excludes women reporting hysterectomy or postmenopausal **excludes 90 records with missing data for either qualification sticker or pregnancy test; pregnancy testing includes any pregnancy testing performed at physician’ office or performed at home but reported to the physician; restricted to women who reported starting Accutane therapy.

 

As was seen above, little change is noted with imputation of NO for missing responses (Table 13, below). Since the vast majority of these enrollees noted the presence of the qualification sticker, inference about the “effect” of the sticker is limited. 

 

Table 13. Relationship between qualification sticker and pregnancy testing for Sexually active, apparently fertile*, 15-45 year-old enrollees. [Missing=NO]
	Qualification sticker Yes	Qualification sticker No	Row Total
Pregnancy test Yes	1575 (90%)	90 (87%)	1665
Pregnancy test No	168 (10%)	13 (13%)	181
	1743	103	1846**
*excludes women reporting hysterectomy or postmenopausal **pregnancy testing includes any pregnancy testing performed at physician’ office or performed at home but reported to the physician; restricted to women who reported starting Accutane therapy.

 

2.2.2    Relationship between Qualification Sticker and ANY Birth Control

 

Per the new isotretinoin label, the Qualification sticker is intended to document that the patient has received education and counseling on pregnancy prevention.  DDRE conducted analyses of Qualification sticker and birth control (as ANY birth control) among sexually active Survey participants (Tables 14 and 15) in order to study the potential impact of the Qualification sticker.  As above, the first table (Table 14) excludes records with missing data while the second table (Table 15) is based on imputation of NO for missing data.  As was seen above for pregnancy testing, there does not appear to be a strong relationship of the Qualification sticker to compliance with birth control as birth control was noted in 97% of enrollees with a sticker and 96% of enrollees without a sticker (Table 14). Three (3) percent of sexually active Survey participants who report a Qualification sticker further report no use of ANY birth control.

 

Table 14. Relationship between qualification sticker and ANY birth control for Sexually active, apparently fertile*, 15-45 year-old enrollees. [Missing=excluded]
	Qualification sticker Yes	Qualification sticker No	Row total
Any birth control Yes	1671 (97%)	70 (96%)	1741
Any birth control No	44 (3%)	3 (4%)	47
	1715	73	1788**
*excludes women reporting hysterectomy or postmenopausal **excludes 54 records with missing data for qualification sticker; birth control is defined as those who report currently use of ANY birth control method as well as any report of tubal ligation or vasectomy in partner.  ANY birth control use as shown in this analysis is probably an overestimate as women who reported vasectomy in their partner were included as ANY birth control regardless of whether they indicated another option as required per approved labeling.  This analysis approach was necessary because of the poor structure of the questionnaire with regard to birth control using tubal ligation and vasectomy leading to potential for confusion.  In addition, vasectomy may not pertain to all potential partners; restricted to women who reported starting Accutane therapy.

 

With imputation of a “NO” for missing information, compliance with any birth control fell slightly for both arms, to 95% among those reporting a Qualification sticker and 92% for those indicating no Qualification sticker (Table 15), but was otherwise high in both arms.

 

Table 15. Relationship between qualification sticker and ANY birth control for sexually active, apparently fertile*, 15-45 year-old enrollees. [Missing=NO]
	Qualification sticker Yes	Qualification sticker No	Row Total
Any birth control Yes	1671 (95%)	70 (92%)	1741
Any birth control No	95 (5%)	6 (8%)	101
	1766	76	1842**
*excludes women reporting hysterectomy or postmenopausal **Birth control is defined as those who report currently use of ANY birth control method as well as any report of tubal ligation or vasectomy in partner.  ANY birth control use as shown in this analysis is probably an overestimate as women who reported vasectomy in their partner were included as ANY birth control regardless of whether they indicated another option as required per approved labeling.  This analysis approach was necessary because of the poor structure of the questionnaire with regard to birth control using tubal ligation and vasectomy leading to potential for confusion.  In addition, vasectomy may not pertain to all potential partners; restricted to women who reported starting Accutane therapy.

 

3.         DAT-2 SURVEY (DURING TREATMENT)

 

NOTE: The sponsor’s submission (“1 Year Report on the SMART Program”) was conducted before accrual of enough DAT-2 questionnaires to be informative.  Thus, the following section describing DDRE analyses of DAT-2 data from the SI/Degge cohort does not have a direct corollary within the “1 Year Report on the SMART Program.”  These data permit inference on SMART attributes/compliance during the middle and later windows of Accutane therapy.  It should be anticipated that the sponsor will also conduct similar analyses to present at the Advisory Committee.

 

A dataset sent to FDA in October 10, 2003 included 3569 women who enrolled in the Degge/SI Accutane Survey from November, 2002 through March 31, 2003 and who returned a DAT-2 questionnaire. One hundred twenty-two (122) of these women apparently never completed a DAT-1 questionnaire and have only a DAT-2. About 47% (N= 1670) of these women reported that they were still taking Accutane.  Thirty-five (35) percent (N=1246) reported that they completed treatment.  The status of remaining 18% (N=652) could not be determined.

 

3.1       DAT-2 Univariate Analyses

 

3.1.1    Qualification Sticker

 

As in the DAT1 questionnaire, DAT-2 participants are asked about the presence of a qualification sticker on prescriptions.  Monthly prescriptions are required for Accutane therapy lasting over 30 days since no refills are permitted under the SMART program.  As shown in Table 16, 94.7% of DAT-2 respondents who continued on therapy with Accutane indicated the presence of a qualification sticker on their prescriptions. 

 

Table 16.  Percent distribution of Qualification sticker from DAT2 for participants who continued on therapy with Accutane.
Total number of sample	1670
Yes	94.7%
I saw the prescription but not I did not see a Qualification sticker	0.8%
Other*	4.5%
*Other includes missing values, someone else filled the prescription, participant did not see the prescription.

 

3.1.2    Pregnancy Testing

 

3.1.2.1 Per Slone Survey: Pregnancy testing during Accutane therapy based on DAT-2 surveys for the interval April 1, 2001 through March 30, 2003 is show in Table 17.  As with compliance with pregnancy testing at initiation of therapy, the proportion of women reporting any testing appears to increase in close association with the SMART program, from ~70% to ~85%.  As seen before, the increase appears to plateau soon after initiation of the SMART program and remain at ~85%.

 

Table 17. Trends in pregnancy testing during Accutane therapy based on DAT-2 surveys (based on quarter of enrollment, based on Slone Survey Quarterly Reports)
era	Last four quarters before SMART				First four quarters of SMART
Qtr	2Qtr01	3Qtr01	4Qtr01	1Qtr02	2Qtr02	3Qtr02	4Qtr02	1Qtr03*
N	1281	1352	1212	1324	1308	1419	1224	1206
	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)
Any test**	66	66	71	81	87	86	85	87
No Test	33	33	29	18	12	14	14	12
UK	1	1	<1	1	<1	<1	1	1
*Based on receipt of 10-week DAT-2: these counts reflect only the fraction of women enrolled in 1Qtr03 that answered DAT-2 through June 30, 2003. **Serum or urine pregnancy test

 

3.1.2.2 Per DDRE analyses: In DDRE analyses of the Degge/SI DAT-2 responder cohort for number of pregnancy tests taken during the past two months among apparently fertile women aged 15 to 44 years women still on treatment, 81% of those reporting sexual activity reported apparently monthly pregnancy testing (Table 18).

 

Table 18. Compliance with pregnancy testing during Accutane therapy among apparently fertile* Survey participants by sexual activity status.
	Sexually active (N=622)	Not sexually active (N=832)
Missing	3.2%	3.7%
Mentioned “0” test	10.9%	13.5%
One test	4.8%	7.3%
Two tests	53.4%	51.0%
More than 2 tests	27.7%	24.5%
Appropriate**	81.1%	75.5%
*excludes women reporting hysterectomy or postmenopausal **represents the sum of the preceding 2 rows as the DAT2 pregnancy testing question applies to a >=2-month window.  Thus, patients should report >=2 pregnancy tests to indicate monthly pregnancy testing

 

3.1.3    Contraception Practices

 

Birth control practices by sexual activity status /treatment status for apparently fertile women aged 15 to 44 years is shown in Table 19. As was seen among respondents to DAT-1, most apparently fertile, sexually active women who remain on therapy with Accutane (97.6%) or completed treatment within one month (94.4%) reported use of some birth control.  Compliance with two forms, one of which is primary (“appropriate” per label) was similar to levels reported in DAT-1 (~55%).  Thus, even during therapy with Accutane, compliance with two forms of contraception as outlined in approved labeling is problematic, with compliance among the highest risk group (young, fertile women who report sexual activity) to be just above 50%.

 

 

 

Table 19. Distribution of birth control use by sexual activity status/treatment status among apparently fertile*, 15-45 year-old participants in DAT2.
	Total**	Any form of birth control	Any primary form of birth control	Appropriate*** per Accutane label
Sexually active, on treatment w/Accutane	608	97.5%	91.0%	54.9%
Sexually active, completed treatment for > 1 month	143	94.4%	87.4%	51.1%
Not sexually active, on treatment w/Accutane	813	54.4%	44.2%	8.1%
Not sexually active, completed treatment for > 1 month	145	52.4%	47.0%	9.0%
*excludes women reporting hysterectomy or postmenopausal **excludes 37 women with missing information on birth control ***appropriate per Accutane label requires two forms of which one must be a “primary” method (e.g., oral contraception, tubal ligation, injectable/implantable/insertable/patch contraceptives, vasectomy, IUD)

 

3.2       DAT-2 Bivariate Analyses

 

3.2.1    Relationship between Qualification Sticker and Pregnancy Testing

 

The relationship between Qualification sticker and pregnancy testing among all, apparently fertile 15-45 year Survey participants continuing on Accutane therapy is show in Tables 20 and 21.  [As in the DAT-1 Bivariate Analyses section above, these tables are included in this section and are stratified by missing=excluded (Table 20) and missing=NO (Table 21).]

 

Table 20. Relationship between pregnancy testing and Qualification sticker for all, apparently fertile*, 15-45 year-old DAT-2 participants.  [Missing=excluded]
	Qualification sticker Yes	Qualification sticker No	Row total
Pregnancy test Yes	1074 (99%)	254 (98%)	1328
Pregnancy test No	5 (1%)	5 (2%)	10
	1079	259	1338**
*excludes women reporting hysterectomy or postmenopausal **restricted to those participants continuing therapy with Accutane; excludes 116 records with missing data; pregnancy testing includes any pregnancy testing performed at physician’s office or performed at home but reported to the physician; as the DAT2 pregnancy testing question applies to a >=2-month window, patients were categorized as “Yes” only if they reported >=2 pregnancy tests.

 

As was seen in analyses of DAT-1 data, any “effect” of the qualification sticker is limited as pregnancy testing is high among both groups.

 

 

 

 

 

Table 21. Relationship between pregnancy testing and qualification sticker for all, apparently fertile*, 15-45 year-old DAT-2 participants. [Missing=No]
	Qualification sticker Yes	Qualification sticker No	Row Total
Pregnancy test Yes	1074 (95%)	305 (95%)	1379
Pregnancy test No	58 (5%)	17 (5%)	75
	1132	322	1454**
*excludes women reporting hysterectomy or postmenopausal **restricted to those participants continuing therapy with Accutane; pregnancy testing includes any pregnancy testing performed at physician’s office or performed at home but reported to the physician; as the DAT2 pregnancy testing question applies to a >=2-month window, patients were categorized as “Yes” only if they reported >=2 pregnancy tests.

 

3.2.2    Relationship of Qualification Sticker and Birth Control

 

The impact of the Qualification Sticker on birth control among sexually active, apparently fertile 15-45 year Survey participants continuing on Accutane therapy is shown in Tables 22 and 23. These data provide little potential to examine the relationship of the Qualification sticker to birth control as these tables are dominated both by Qualification sticker use and birth control.  As noted for bivariate DAT-1 analyses of Qualification sticker and ANY birth control, it is possible that these analyses overestimate any birth control as the analysis approach accepted affirmative responses to questions which should have been skipped.

 

Table 22. Relationship between qualification sticker and ANY birth control for sexually active, apparently fertile*, 15-45 year-old DAT-2 respondents. [Missing=excluded]
	Qualification sticker Yes	Qualification sticker No	Row Total
Any birth control Yes	560 (99%)	15 (100%)	575
Any birth control No	5 (1%)	0 (0%)	5
	565	15	580**
* excludes women reporting hysterectomy or postmenopausal **excludes 28 records with missing data for qualification sticker; restricted to women who are currently on treatment or had stopped the treatment within the past month.  Per approved labeling, sexually active women are to remain on birth control / refrain from conception for one month following the end of Accutane treatment; questions pertaining to birth control are different in the DAT-2 instrument in comparison to the DAT-1 instrument. However, it is still possible for women to become confused in answering the question. Thus, as in the analysis of ANY birth control use in DAT-1, ANY birth control was defined as those who reported that currently using any birth control method including tubal ligation and vasectomy.  As before, vasectomy was considered a birth control method regardless of selection of an alternative method, as required per labeling;

 

 

 

 

 

Table 23. Relationship between qualification sticker and ANY birth control for sexually active, apparently fertile*, 15-45 year-old DAT-2 respondents. [Missing=NO]
	Qualification sticker Yes	Qualification sticker No	Row total
Any birth control Yes	560 (94%)	15 (100%)	575
Any birth control No	33 (6%)	0 (0%)	33
	593	15	608**
* excludes women reporting hysterectomy or postmenopausal **Restricted to women who are currently on treatment or had stopped the treatment within the past month as, per approved labeling, sexually active women are to remain on birth control / refrain from conception for one month following the end of Accutane treatment; questions pertaining to birth control are different in the DAT-2 instrument in comparison to the DAT-1 instrument. However, it is still possible for women to become confused in answering the question. Thus, as in the analysis of ANY birth control use in DAT-1, ANY birth control was defined as those who reported that currently using any birth control method including tubal ligation and vasectomy.  As before, vasectomy was considered a birth control method regardless of selection of an alternative method, as required per labeling.

 

4.         PREGNANCIES, POSSIBLE RISK FACTORS, AND PREGNANCY RATES WITHIN THE DEGGE/SI COHORT

 

In other analyses conducted by DDRE but not the Accutane sponsor, the Degge/SI cohort was examined for reports of pregnancy for calculation of an Accutane-exposed pregnancy rate.  Based on DDRE analyses, the Degge/SI dataset contains 28 reports of pregnancy among a total cohort of 5469 women as of the lock date on the dataset.  This is based entirely on information received on/within DAT-1 question 37 and DAT-2 question 17 (“Have you been pregnant at any time since you first started taking Accutane?”).  However, application of this numerator and denominator to calculate an Accutane-exposed pregnancy rate is problematic as women undergoing a repeat course of therapy (~25% of the total cohort) may correctly answer this question as “Yes” based on a un-exposed pregnancy in-between treatment courses.  Thus, it is not possible to interpret the answers reported for this segment of the cohort.  This precludes examination of pregnancy risk among women on repeat therapy.

 

After restriction to the 4277 women on their first-course of therapy, 15 participants reported the occurrence of a pregnancy since starting Accutane therapy. At the time of survey, the median age of these 15 women was 22 years (mean 24.5 years) with range 16 to 39 years.  For comparison, the median and mean of the total cohort of first-time users are 22 years and 24.4 years, respectively.  Eight women reported pregnancy on DAT1; 7 on DAT-2. So as to highlight the problems encountered in analysis of these data for risk factors, 4 out of these 15 women reported no sexual intercourse in both DAT1 and DAT2 questionnaires.  Thus, it is difficult to extract putative risk factors for an Accutane-exposed pregnancy based on the limited information captured by the Patient Survey.

 

In presentation before the FDA Dermatologic Drugs Advisory Committee Meeting (Gaithersburg, MD) on September 18, 2000, Mitchell et al reported 992 pregnancies among 339,994 women completing follow-up in the Survey (through 1999) for a rate of 2.9 per 1,000 women. Based on report of 15 pregnancies, the crude observed pregnancy rate for first-course users within the Degge/SI cohort is 15/4277 = 3.5 / 1000.  It should also be noted that the denominator used in this calculation (4277) includes some women who have not yet completed treatment and thus potentially able to report an Accutane-exposed pregnancy with further follow-up. Thus, since this rate is censored, it likely represents an underestimate of the rate to be realized when all these participants complete follow-up, as was described by Mitchell.

 

5.         SURVEY ADDENDUM (RECENT DATA)

 

FDA has received additional data, in addition to use data, since receipt of the Degge/SI dataset addendum (Oct 10, 2003) in the form of quarterly reports.  As shown in Chart 2^[19] there are currently three isotretinoin patient surveys.  All three of these groups currently submit “Quarterly Reports” summarizing data accrued to the end of the quarter.  Quarterly reports from the Slone Epidemiology Group on the Accutane-brand of isotretinoin have been cited previously in this document and Slone continues to submit data (using the old survey) for patients who enrolled to Slone via Accutane materials and/or deferred transfer to Degge/SI.  Degge/SI submitted a Quarterly report for 1QTR2003 based on very limited enrollment as of March 31, 2003 and two additional (and larger) Quarterly reports for 2QTR2003 and 3Qtr2003.  The Slone Epidemiology Group has also submitted two quarterly reports for 2QTR2003 and 3Qtr2003 for generic isotretinoin.  Due to differences in the survey, survey methods, and operating procedures, it is problematic to compare across these disparate surveys.  However, for completeness, selected and generally comparable variables have compiled from the Quarterly reports for 2QTR2003 and 3Qtr2003.^[20] In general, these data support results reported for the interval of April 1, 2001 through March 30, 2003. Specifically:

 

·       Report of any pregnancy testing prior to initiation of isotretinoin therapy continues at ~90%.

·       Report of two (or more) pregnancy tests prior to initiation of isotretinoin therapy continues at ~65%

·       Report of NO birth control among apparently fertile, sexually active respondents continues at ~3%.

·       Report of any pregnancy testing during therapy (DAT-2) continues at ~82%.

 

Thus, these data DO NOT show any remarkable improvements in compliance with these selected attributes.

 

DISCUSSION

 

On many points, analyses conducted and described herein by DDRE are in agreement with analyses as reported by the sponsor in their “1 Year Report.”  Some important items of agreement include:

 

·       absolute participation (22%-26% per DDRE; 28% per the sponsor)

·       > 90% report of use of the Qualification sticker

·       > 90% report of one pregnancy test prior to start of therapy

·       ~65% report of two pregnancy tests prior to start of therapy

·       use of two forms of contraception (57% per sponsor based on two forms; 46% per DDRE based on “appropriate” as per label)

 

However, bivariate analyses of the Qualification sticker have only been developed by DDRE. These analyses suggest pregnancy testing and birth control were generally similar between those reporting Qualification stickers and those not reporting Qualification stickers.  In addition, results for the respondents noting a Qualification sticker in these stratified analyses were very similar to those observed in univariate (un-stratified) analyses.

 

However interesting and potentially problematic as these results are, the generation of a pregnancy rate following new labeling could be argued as the final arbiter for success of the labeling, which was not limited to the Qualification sticker.  To this end, it is noteworthy to report the pregnancy rate of 3.5 /1000 for the initial Degge/SI cohort to be identical to the rate reported by the Slone Epidemiology Group for their Accutane Survey.  It appears, therefore, that the efforts implemented in the new labeling did little to affect the absolute pregnancy rate within the population included in the Patient Surveys. [As noted above, the denominator used in the calculation of a pregnancy rate for the Degge/SI cohort (4277) includes some women who have not yet to completed treatment and thus potentially able to report an Accutane-exposed pregnancy with further follow-up. Thus, since this rate is censored, it likely represents an underestimate of the rate to be realized when all these participants complete follow-up, as was described by Mitchell.]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

APPENDIX 1—GEOGRAPHIC REGIONS

 

East:               Connecticut, Delaware, Maine, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont

 

Midwest:        Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota, Wisconsin

 

South:             Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Virginia, Wet Virginia, District of Columbia

 

West:              Arizona, California, Colorado, Idaho, Montana, Nevada, New Mexico, Oregon, Utah, Washington, Wyoming

 

APPENDIX 2—DESCRIPTION OF ADVANCEPCS

 

AdvancePCS is one of the largest pharmacy benefit management companies in the U.S., currently covering an estimated 50 million patient lives and processing 300 million prescription claims annually.  FDA has on-line access to AdvancePCS’ database of paid claims for prescriptions filled in 55,000 pharmacies across the country.  Patients whose claims are processed by AdvancePCS include those covered under various types of insurance plans that cover prescription drugs, including some employer’s self-insured planes, selected managed care plans, Blue Cross/Blue Shield plans, and some other traditional insurers.  Demographically, these patients appear to represent all 50 states and include substantial numbers of the elder, children, and women of childbearing age.  Their representativeness of all patients receiving dispensed prescriptions in the U.S. is not known.

 

The data in AdvancePCS include date dispensed, drug name, drug strength, dosage form, quantity, days’ supply, and prescriber specialty.  All patients are assigned a unique identifier, and so can be followed through time for as long as they are in the same prescription plan.  Data are entered into the system within days of dispensing, so that real-time analysis is possible.  At any time, FDA is able to access the most recent 25 months of data online.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

APPENDIX 3—RECOMMENDATIONS FOR PCS AND PRESCRIPTION AUDIT

 

The following recommendations cover the PCS, and audit, and also suggestions for making the results of future survey waves easier to analyze and interpret.  The survey recommendations focus on increasing the number of prescriptions captured for analysis, and consistently obtaining an adequate number of prescriptions for all of the survey waves.  The operational audit procedures need to be clearly described in detail, and may need to be altered in order to achieve the goal of data validation.

 

Survey

·       Re-examine sample size and recruiting plan to ensure adequate capture of prescriptions for analysis

·       Consider unifying the survey (sponsor and generic companies) to increase the response rate and decrease the number of inquiries to individual stores for each survey

·       Consider extending the PCS indefinitely (beyond the current 2 year limit) as an indirect measure of physician compliance, assuming that an adequate number of prescriptions can be captured for analysis.  If an indefinite extension is not possible, consider extending the survey at least until a sufficient number of prescriptions has been obtained for a long enough period of time (at least 1 year) to permit a valid analysis of compliance.

·       Consider approaching the December 2002 dropouts to discuss their possible participation, since they account for a very large percentage of the isotretinoin prescriptions.  This approach might be successful if the separate surveys are unified into a single entity

·       If the December 2002 pharmacies will not reconsider participation, consider approaching PBM's or switch companies (firms that route the prescriptions from the pharmacy to the insurance provider) as a way to get to an adequate number of prescriptions.  Again, this approach might be successful if the separate surveys are unified into a single entity

·       Re-evaluate the recruiting process every 1 to 2 surveys until the desired number of prescriptions has been captured for 4 surveys (1 year of data)

 

Audit

·       Present a detailed summary of how the current audit is being implemented, focusing on the recruiting procedures are rules.

·       Re-evaluate the current audit implementation, and make the changes necessary to ensure that the audited pharmacies are randomly chosen from the participants, and balanced across the pharmacy analysis strata

 

Analysis

·       Provide an electronic copy of the cleaned data used for the analysis and production of the tables in the PCS reports (in addition to the paper report).  The data should include information on recruited pharmacies, as well as those that responded and were audited.

·       Include a table detailing the recruited vs. responded vs. audited pharmacies across pharmacy strata in all future reports.

 

 

 

 

APPENDIX 4—TABLES AND CHARTS

 

Table A.  Trends in Slone Contraceptive Status for age group 15-24 years
era	Last four quarters before SMART				First four quarters of SMART
Qtr	2Qtr01	3Qtr01	4Qtr01	1Qtr02	2Qtr02	3Qtr02	4Qtr02	1Qtr03
N	516	531	556	590	602	610	625	534
	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)
sterile*	2	1	1	1	2	1	2	1
abstinent	39	38	35	35	33	30	30	31
BCP, BC injection, IUD	51	52	54	56	58	60	55	60
Barrier, other, UK, none**	8	9	10	8	7	9	13	8
*self or partner **includes 0% to 1% of patients per quarter who report no contraceptive use and sexually active without a history of sterility (self or partner). BCP = birth control pills; BC injection = includes all injectable or implantable contraceptives; IUD=intrauterine device

 

Table B.  Trends in Slone Contraceptive Status for age group 25-34.
era	Last four quarters before SMART				First four quarters of SMART
Qtr	2Qtr01	3Qtr01	4Qtr01	1Qtr02	2Qtr02	3Qtr02	4Qtr02	1Qtr03
N	337	290	322	346	251	326	308	277
	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)
sterile*	24	22	20	20	20	28	23	23
abstinent	7	11	10	8	10	7	7	5
BCP, BC injection, IUD	58	59	59	62	61	56	60	60
Barrier, other, UK, none**	11	8	11	10	9	9	10	12
*self or partner **includes 0% to 1% of patients per quarter who report no contraceptive use and sexual activity without a history of sterility (self or partner). BCP = birth control pills; BC injection = includes all injectable or implantable contraceptives, IUD=intrauterine device

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table C.  Trends in Slone Contraceptive Status for age group 35-44 years.
era	Last four quarters before SMART				First four quarters of SMART
Qtr	2Qtr01	3Qtr01	4Qtr01	1Qtr02	2Qtr02	3Qtr02	4Qtr02	1Qtr03
N	176	158	155	139	176	158	155	106
	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)
sterile*	59	58	59	63	56	62	60	60
abstinent	6	7	8	6	6	4	4	6
BCP, BC injection, IUD	27	26	30	23	32	29	25	28
Barrier, other, UK, none**	8	9	3	8	6	5	11	6
*self or partner **includes 0% to 1% of patients per quarter who report no contraceptive use and sexual activity without a history of sterility (self or partner). BCP = birth control pills; BC injection = includes all injectable or implantable contraceptives, IUD=intrauterine device

 

 

Table D. Selected and generally comparable variables as reported for two calendar quarters by three disparate Isotretinoin Patient Surveys.
	1 April 2002 – 30 June 2003 5th Quarter after SMART			1 July 2003 – 30 Sept 2003 6th Quarter after SMART
	Slone Accutane	Degge/SI Accutane	Slone generics	Slone Accutane	Degge/SI Accutane	Slone generics
Any pregnancy test before Rx	92%	89%	92%	----**	91%	92%
Two + pregnancy tests	----*	64%	65%	----**	69%	62%
Sexually active but deny any birth control	1.1%	2.8%	3.2%	----**	2.6%	2.4%
§Any pregnancy test during Rx	87%	80%	----*	----**	82%	----*
* Data unavailable due to time interval or instrument **Data unavailable due to shift to 6-month frame of reference § pertains to cohort establishing in the preceding quarter

 

 

 

 

 

 

 

 

 

Chart 1 Accutane / Isotretinoin Patient Survey Operating Schema.

 

 

10w

18w

26w

34w

D A T

-1       0             1            2            3            4            5            6            7            8             9           10          11          12  13     14          15          16          17          18

DAT1

DAT2-10

DAT2-18

DAT2-26

DAT3

DAT3

DAT3

Random assignment of 1/3 of cohort to receive either DAT2-10, -18, or -26

Additional DAT3’s if original DAT3 was completed <6 months after stop date

Enrollment

34w

A T

-1       0                                                                      5            6            7            8             9

AT1

AT2

AT2

AT2

 

 

 

 

 

 

 

 

 

 

 

Chart 2. Overlap of 3 currently active, isotretinoin patient surveys.

 

 

S U R V E Y S

                                                +(old) SLONE Accutane Enrollment       ________________________________________________________________________________________________________            

 

                                                                                     ++(new) Degge/SI Accutane

  Initial Treatment Survey                                During Treatment Survey

                                                                           DAT 1 ONLY                                                                       DAT 2

                                                                                                      

                                                        ç_____________________________ ç____________________________________________________

                                                          Nov 20, 2002 – March 31, 2003                    April 1, 2003---------ongoing

     

 

                                                                                                                  ++ (new)  SLONE generics enrollment

                                                                                                                                                                            DAT 1                                                  

           ç___________________________________________________                                                 

 April 1, 2003-------------ongoing                                

                                        

                        

 

 

                                             SMART 1 yEAr  Evaluation  

                   ê___________________________________________ç

                        April 1, 2002---March 31, 2003                         

 

 

                                                                                ç                                                 ê                                                                  Ý       

               April 1^st Oct           Jan                _{1st quarter}          March        _{2nd quarter}          June                       Aug 15                                                                                                                                                                 31                                   30                      cutoff

                                                            

 

 

 

+Survey does not include questions to determine compliance with certain elements of SMART including issuance of a qualification sticker.

++Surveys contain questions that are specifically designed to determine compliance with SMART including issuance of a qualification sticker.