Study Characteristics
PEM
M-PEM
SCEM
Setting
Primary care
Secondary care
Prescribers
General Practitioner
Specialist
Design
Non-interventional observational cohort
Data Source
Secondary use of data from existing medical records
Ethics
Waiver under Section 251 of NHS Act 2006
Yes
Sample size
10,000+
Bespoke to targeted events
Period of Observation
6–12 months
3–12+ months
≤3 months
Questionnaire
Standard (single A5) across all studies
Bespoke design (multiple A4), by study
Drug Utilization
All populations
Targeted sub-populations considered at risk
Surveillance
General surveillance
General and targeted events (identified, potential and missing risks)a
Risk factors
Event (case) specific on follow-up
For targeted identified and potential risks for all patients
Analysis
Crude and age/sex adjusted measures of occurrence and association
Multiple covariate adjusted measures of occurrence and association; survival methods for signal generation; intra- and inter- cohort comparisons possible
PASS RMP
No
Yes
ENCEPP registration
No
Yes
Prescriber remuneration
No
Yes
For each approach, events of interest (e.g. identified risks, such as suicide events) may be followed up for purposes of further evaluation. For each patient, trained coding staff prepared a computerised, longitudinal, chronological record of demographic, exposure and outcome data (including additional follow-up). Quantitative and qualitative data analyses are conducted. Quantitative analyses include descriptive statistics to summarise patient and prescriber (where provided) characteristics such that drug utilisation and compliance with recommended prescribing regimens may be described. Calculations of event risk and rates (incidence densities for a fixed period (t) – IDt– usually expressed in units of first event reports per 1000 patient-months) can give estimates of real-world frequency. Calculation ID differences between periods of observation are effective methods by which disproportionality in risk or ID may be observed, suggestive of signals of treatment effects. Inter-cohort IDs and ID difference estimates may be calculated, as well as intra-cohort estimates for special populations of interest. Other complementary analyses include describing frequency of events given as reasons for stopping and also application of survival methods to provide further detail on pattern and time to onset of an event after starting treatment with a study drug. Qualitative analyses are undertaken for exploration of signals or events of interest. Through medical evaluation, individual or clusters of event reports are examined and important attributes identified. Examples of these approaches will be presented in the remainder of this chapter.
4.3 Completed Studies
A wide range of drugs used to manage mental health conditions have been studied using standard and modified approaches. These include agents to manage depressive disorders, anxiety disorders, attention deficit hyperactivity disorder (ADHD), schizophrenia and dementia (Table 4.2). Treatments used to manage conditions associated with addictive behaviours such as smoking habit, obesity and chronic non-cancer pain are also presented for purposes of demonstrating particular methodological considerations. All these medicines were all intended for wide-spread, long-term use in primary care. Other medications have also been studied (Layton and Shakir 2011), but these are not the focus of this chapter and so will not be described further.
Table 4.2
List of completed studies (standard (PEM)/modified (M-PEM)) of psychotropic medicines, by ATC code (World Health Organisation 2015)
Therapeutic class and ATC code | Name: generic [UK proprietary] | Design/cohort size | Collection period |
|---|---|---|---|
Antiobesity preparations (excl. diet products) | |||
A08AA10 | Sibutramine [REDUCTIL™] | Standard (N = 12,336) | Oct 2001–Jun 2002 |
A08AX01 | Rimonabant [ACOMPLIA™] | Modified (N = 10,008) | Jun 2006–Oct 2008 |
Opioids | |||
N02AB03 | Fentanyl buccal [EFFENTORA™] | Modified (N = 556) | Mar 2009–Jun 2011 |
N02AB03 | Fentanyl nasal [PECFENT™] | Modified (N = 63) | Dec 2010–Sep 2012 |
Antipsychotics | |||
N05AE05 | Sertindole [SERDOLECT™] | Standard (N = 436) | Dec 1996–Dec 1997 |
N05AH03 | Olanzapine [ZYPREXA™] | Standard (N = 8858) | Dec 1996–May 1998 |
N05AH04 | Quetiapine [SEROQUEL™] | Standard (N = 1728) | Jun 1998–Jan 2000 |
N05AH04 | Quetiapine [SEROQUEL XL™] | Modified (N = 13,276) | Sep 2008–Feb 2013 |
N05AX08 | Risperidone [RISPERDAL™] | Standard (N = 7684) | Jul 1993–Dec 1996 |
Anxiolytics | |||
N05BE01 | Buspirone [BUSPAR™] | Standard (N = 11,113) | Mar 1988–Feb 1989 |
Hypnotics and sedatives | |||
N05CF01 | Zopiclone [ZIMOVANE™] | Standard (N = 11,543) | Mar 1991–Jul 1991 |
N05CF02 | Zolpidem [STILNOCT™] | Standard (N = 13,460) | Jul 1994–Jan 1996 |
Antidepressants | |||
N06AB03 | Fluoxetine [PROZAC™] | Standard (N = 12,692) | Mar 1989–Mar1990 |
N06AB05 | Paroxetine [SEROXAT™] | Standard (N = 13,741) | Mar 1991–Mar1992 |
N06AB06 | Sertraline [LUSTRAL™] | Standard (N = 12,734) | Jan 1991–Sep1992 |
N06AB08 | Fluvoxamine [FAVERIN™] | Standard (N = 10,983) | Feb 1987–Feb 1988 |
N06AG02 | Moclobemide [MANERIX™] | Standard (N = 10,835) | Jun 1993–Aug 1995 |
N06AX06 | Nefazodone [DUONIN™] | Standard (N = 11,834) | Jan 1996–Feb 1997 |
N06AX11 | Mirtazapine [ZISPIN™] | Standard (N = 13,554) | Sep 1997–Feb 1999 |
N06AX12 | Bupropion [ZYBAN™] | Standard (N = 11,735) | Jul 2000–Aug 2000 |
N06AX16 | Venlafaxine [EFEXOR™] | Standard (N = 12,642) | May 1995–Jun 1996 |
N06AX21 | Duloxetine [CYMBALTA/YENTREVE™] | Standard (N = 19,485) | Sep 2004–Apr 2005 |
Psychostimulants (agents used for ADHD and nootropics) | |||
N06BA07 | Modafinil [PROVIGIL™] | Modified (N = 2092) | Jul 2004–Aug 2005 |
N06BA09 | Atomoxetine [STRATTERA™] | Modified (N = 5079) | May 2004–Oct 2004 |
Antidementia drugs | |||
N06DA02 | Donepezil [ARICEPT™] | Standard (N = 1762) | Apr 1997–Feb 1999 |
Drugs used in addictive disorders | |||
N07BA03 | Varenicline [CHAMPIX™] | Modified (N = 12,135) | Dec 2006–March 2007 |
The sample sizes of standard, M-PEM and SCEM studies of psychotropics are remarkably different. The median cohort size of the 19 standard PEM studies of psychotropic medicines presented is 11735 (IQR 9847, 12713), whilst that of the 7 M-PEM studies is 5079 (IQR 1324,11072). One SCEM study has been completed to date (N = 869). The explanation is related to the difference in principle study objective: standard PEM studies were intended for general surveillance with a target sample size of at least 10,000 patients to allow for the detection of rare events occurring with a frequency of at least 1 in 2000 patients (assuming the background rate is zero) with 85 % power (Machin et al. 1997a, b). The lower sample size for M-PEM and SCEM reflects a bespoke study specific need, which generally requires fewer numbers than the general surveillance studies.
4.4 Exploring Drug Utilisation Factors Within Pharmacovigilance
Drug utilisation research is an essential part of pharmacoepidemiology, as it describes the extent, nature and determinants of drug exposure at the patient level. Collectively standard, M-PEM and SCEM approaches permit the examination of the characteristics of prescriber and new drug user populations and contribute to the accumulation of safety data, particularly with regard to vulnerable populations for whom off-label prescribing has occurred.
There are several ways in which this may occur: by prescribing a dose in excess of that specified by the MA, prescribing for an unlicenced indication, prescribing for a special group outside of the MA specification and altering the dosage form. Off-label prescribing is permitted in circumstances where a physician concludes that for medical reasons, treatment with the product is necessary to meet the specific needs of the patient (General Medical Council 2015).
Situations leading to a warning or precaution use include conditions to be fulfilled before or during use, special populations at increased risk, risks associated with starting or stopping the product and possible medication errors. These situations represent vulnerable populations that require more careful management in order to avoid harm. They also represent areas of potential risk for pharmacovigilance because the populations being exposed have not been studied, or there is limited information.
4.4.1 Off-Label Use
As an example, the summary demographic information on the study populations identified for the four antipsychotic studies conducted using standard methods is presented in Table 4.3 (Mackay et al. 1999). It is notable that all drugs listed first gained MA for the treatment of schizophrenia. This is consistent with the narrowly defined patient population studied premarketing. Also notable in Table 4.3 is the frequency of apparent ‘off-label’ prescribing. In terms of indication, quetiapine IR was the antipsychotic for which the proportion of use was highest (47.4 %) followed by risperidone (44.9 %) for indications other than within the MA at launch. The observation that quetiapine IR appeared to have the highest frequency is not unexpected, given it was the last of the four drugs listed to be marketed. Nevertheless, off-label prescribing in terms of indication was very common. This observation is supported by similar findings elsewhere (Hodgson and Belgamwar 2006).
Drug (n) | Study population | |||||
|---|---|---|---|---|---|---|
Indication [year] | MA Min age years | Sex | Age Mean (SD) years [Range] | MA Indication n (%a) | Comments on use in elderly and children | |
Male n (%a) | Female n (%a) | |||||
Risperidone (n = 7684) | ||||||
Acute and chronic schizophrenic psychoses, other psychoses and affective symptoms [1993] | 15 | 4124/7624 (54.1) | 3500/7624 (45.9) | 44.1 (19.4) | 3376/6129 (55.1) | 98/6678 (1.5 %) <15 years; 1271/6678 (19.0 %) ≥65 years Indication: 1120/6129 (18.3 %) had psychosis; 87/6129 (1.4 %) dementiab |
[4–99] | ||||||
N = 6678 | ||||||
Sertindole (n = 462) | ||||||
Schizophrenia [1996] | 18 | 230/459 (50.1) | 229/459 (49.9) | 40.8 16.4) | 210/311 (67.5) | 2/402 (0.5 %) <18 years; 48/402 (11.9 %) ≥65 years Indication: 54/311 (17.4 %) had psychosis; 3/311 (1.0 %) dementiab |
[16–95] | ||||||
N = 402 | ||||||
Olanzapine (n = 8858) | ||||||
Schizophrenia [1996] | 18 | 4972/8810 (56.4) | 3838/8810 (43.6) | 42.2 (17.1) | 3470/5902 (58.8) | 81/7455 (1.1 %) <18 years; 995/7455 (13.3 %) ≥65 years Indication: 1104/5902 (18.7 %) had psychosis; 47/5902 (0.8 %) dementiab |
[14–97] | ||||||
N = 7455 | ||||||
Quetiapine immediate release (n = 1728) | ||||||
Schizophrenia [1997] | 18 | 807/1725 (46.8) | 918/1725 (53.2) | 44.6 (19.3) | 596/1133 (52.6) | 20/1425 (1.4 %) <18 years; 266/1425 (18.6 %) ≥65 years Indication: 215/1133 (18.9 %) had psychosis; 45/1133 (4.0 %) dementiab |
[15–98] | ||||||
N = 1425 | ||||||
None of these products studied were indicated for the use in children <15 years. However, the frequency of prescribing appeared to be consistently of the order of 1 %. The literature suggests that in the UK, antipsychotics are prescribed generally for aggressive behaviour in this population (Doerry and Kent 2003). Indeed the study for risperidone reported that of the 98 children aged <15 years, 49 (50 %) had been prescribed risperidone for hyperactivity. For all the antipsychotics studied, the use in the elderly was considered a special warning for use, and the proportions of each cohort aged >65 years was consistently between 10 % and 20 %. Whilst this in itself does not constitute off-label prescribing, the frequency of use in elderly patients for the treatment of behavioural and psychological manifestations of dementia is. In the majority of these studies, such prescribing was common (>1 % of population studied) although the frequency was the highest for quetiapine (4 %). Further discussion on the contribution of use and safety antipsychotics in patients with dementia is given later in this chapter.
One of the limitations of standard PEM methodology was that information on baseline characteristics, such as prior medical history, concurrent morbidities, as well as treatment patterns was limited. M-PEM retains all the strengths of the original method, with the same underlying process, but also tried to overcome these limitations through bespoke targeted surveillance, enhanced data collection and application of new analytical methods. This permits more detailed exploration of the heterogeneity of the new user population, as well as appropriate or inappropriate use.
There are a number of examples in the published literature where off-label prescribing has been reported within M-PEM studies conducted on psychotropic medicines. Davies et al. reported on the prescribing of modafinil (Provigil™),which was marketed in the UK in 1998 to promote wakefulness in the treatment of narcolepsy (Davies et al. 2013). Its licence was extended in 2004 to include chronic pathological conditions. Following a review of the safety of modafinil (European Medicines Agency 2011), risk minimisation measures were introduced. These included updates to the SPC to reflect nature of adverse reactions, additions to types of patients with conditions contraindicated for use and restriction of indication to patients with shift work sleep disorder, narcolepsy and obstructive sleep apnoea/hypopnoea syndrome. The M-PEM study looked specifically at use post the 2004 extension in 1096 patients prescribed modafinil in primary care. Study results reported that the prevalence of use in multiple sclerosis was very common (n = 372, 33.9 %) and use in children aged 16 years or under was uncommon (n = 9, 0.8 %) – both regarded as off-label (Davies et al. 2013).
4.4.2 Contraindications and Special Warnings and Precautions for Use
M-PEM studies were recently conducted to support a regulatory requirement to examine the use in general medical practice in England of two novel formulations of an opioid analgesic. Fentanyl citrate buccal tablets (Effentora™; Cephalon), were approved in the EU on April 2008 for the management of breakthrough pain in patients with cancer who are already receiving and who are tolerant to opioid therapy for their underlying persistent cancer pain. Effentora™ was launched in the UK in January 2009 (Cephalon 2010). Fentanyl pectin nasal spray (PecFent™; Archimedes Development Ltd.), licenced for the same indication, gained marketing approval in the EU in August 2010 and was launched in the UK in October 2010 (Archimedes Pharma UK Ltd. 2012). An objective of both studies study was to examine the frequency of inappropriate use, i.e. use without long-term opioid therapy and off-label use. Prescribing indicators were developed based on the SPC Sects. 4.3 and 4.4. The results from both studies regarding the number of reports of contraindications are summarised in Table 4.4. The majority of patients within either Effentora™ or PecFent™ cohorts had no contraindications for use (n = 482, 87.5 % vs n = 56, 88.9 %). However, the remainder were contraindicated because they were reported to be opioid naïve or not receiving maintenance opioids and had pre-existing COPD or respiratory depression. The prevalence was similar for both products (n = 69, 12.5 % vs n = 7, 11.1 %) albeit the study cohort size was small for PecFent™.
Table 4.4
Characteristics of Effentora™ or PecFent™ M-PEM cohorts
Characteristic n (% cohort) | Effentora™ (N = 551) | PecFent™(N = 63) |
|---|---|---|
Age at start of treatment (years): median (IQR) | 62 (50–72) | 62 (49–73) |
Sex: | ||
Male | 248 (45.1) | 22 (34.9) |
Female | 302 (54.6) | 41 (65.1) |
Not specified | 1 | – |
Indication: | ||
Break through pain | 341 (61.9) | 41 (65.1) |
Break through pain and other | 2 (0.4) | 2 (3.2) |
Other indications | 133 (24.1) | 16 (25.4) |
Not specified | 75 (13.6) | 4 (6.3) |
Contraindications: | ||
Opioid naïve/non-tolerant | 31 (5.6) | 3 (4.8) |
Age <18 years | 1 (0.2) | 0 (0) |
COPD | 35 (6.4) | 3 (4.8) |
Respiratory depression | 7 (1.3) | 1 (1.6) |
Respiratory failure | 3 (0.5) | 0 (0) |
Obliterative bronchiolitis | 1 (0.2) | 0 (0) |
Breastfeeding | 0 (0) | 0 (0) |
MAOI use <15 days of starting (or concomitant) | 1 (0.2) | 0 (0) |
Indicators of special warning/precautions for use: | ||
65+ years | 235 (42.6) | 28 (44.4) |
CKD stage >=3
 Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
Get Clinical Tree app for offline access
| ||