Sunday, April 19, 2015

Hockey stick effect, yo-yo effect, honeymoon effect, halo effect, Hawthorne effect, John Henry effect, and Pygmalion effect

In clinical trials, people need to pay attention to the various effects in designing the clinical trials. A while ago, I mentioned the hockey stick effect that described an initial rebound in a treatment measure followed by the true deterioration of the treatment measures, which occurred in diseases such as Alzheimer’s disease, alpha-1 antitrypsin deficiency, where the purpose of the treatment is to prevent the further deterioration of the disease progression instead of improving.

Other these kinds of effects were described below.

In a clinical trial design discussion, the KOL (key opinion leader) used the yo-yo effect to describe a outcome measure that could fluctuate during the screening period – before the treatment intervention started. 

According to Wikepedia, yo-yo effect, also known as weight cycling, is a term "yo-yo dieting" coined by Kelly D. Brownell at Yale University, in reference to the cyclical loss and gain of weight, resembling the up-down motion of a yo-yo. In this process, the dieter is initially successful in the pursuit of weight loss but is unsuccessful in maintaining the loss long-term and begins to gain the weight back. The dieter then seeks to lose the regained weight, and the cycle begins again.

In a FDA ADCOM meeting minutes, the yo-yo effect is used to describe the hemoglobin level “And hemoglobin levels often plummet to less than 10 before therapy is resumed. This causes a yo-yo effect and is not in the best interests of patients in treating their anemia. There is also no safety net to prevent patients' hemoglobin levels from falling so low that a blood transfusion is required. “


Honeymoon Effect

Honeymoon effect is used to describe the temporary rebound in insulin level right after the initial treatment. After starting treatment with insulin a person's own insulin levels may temporarily improve. This is believed to be due to altered immunity and is known as the "honeymoon phase". The clinical trial for type 1 diabetes must consider honeymoon effect in the study design.

The Honeymoon Phase (or Honeymoon Period) amongst people with type 1 diabetes refers to the period of time shortly following diabetes diagnosis when the pancreas is still able to produce a significant enough amount of insulin to reduce insulin needs and aid blood glucose control.

This does not, unfortunately, indicate that the diabetes is in remission or can be cured. 

There is no hard and fast rule for how long the honeymoon phase lasts amongst people with type 1 diabetes. The diabetes honeymoon phase can last for weeks, months or in some cases years.

In an ADCOM meeting, the honeymoon effect was also used to describe the transient initial effect in epilepsy trial “Beneficial effect of stimulation was lasting. We did not see the familiar and transient honeymoon effect of adding medication after medication to the regimens of these refractory patients. By three years of stimulation, seizures were at a median 68 percent improvement relative to baseline, that is to say a third of their initial level”


The halo effect
is a cognitive bias in which an observer's overall impression of a person, company, brand, or product influences the observer's feelings and thoughts about that entity's character or properties. It was named by psychologist Edward Thorndike in reference to a person being perceived as having a halo. Subsequent researchers have studied it in relation to attractiveness and its bearing on the judicial and educational systems. The halo effect is a specific type of confirmation bias, wherein positive feelings in one area cause ambiguous or neutral traits to be viewed positively. Edward Thorndike originally coined the term referring only to people; however, its use has been greatly expanded especially in the area of brand marketing.

The term "halo" is used in analogy with the religious concept: a glowing circle that can be seen floating above the heads of saints in countless medieval and Renaissance paintings. The saint's face seems bathed in heavenly light from his or her halo. Thus, by seeing that somebody was painted with a halo, the observer can tell that this must have been a good and worthy person. In other words, the observer is transferring their judgment from one easily observed characteristic of the person (painted with a halo) to a judgment of that person's character.

The halo effect works both in both positive and negative directions (the horns effect): If the observer likes one aspect of something, they will have a positive predisposition toward everything about it. If the observer dislikes one aspect of something, they will have a negative predisposition toward everything about it.

In a FDA memo, the halo effect was used to described as a bias in subjects’ assessment of the device effectiveness. “In general, subject data have not been employed as a primary endpoint in device trials because it is difficult to determine whether adequate training of patients has been achieved. Such subjects are particularly prone to the “halo effect” in which subjects judge a return to a “baseline” appearance (after resolution of the swelling, redness and tenderness associated with implantation) to be a significant improvement. “


Hawthorne Effect

When people are under study, observation or investigation, this very fact can have an effect on them and on the results of the study. This is known as the Hawthorne effect.

In a FDA presentation, the Hawthorne Effect is described as one of the main contributor to the placebo effect.

In clinical trial auditing/inspection, the observer effect needs to be minimized. “When you observe the inspection, you change the dynamics of the situation. When people are aware they are being observed, they act differently than they would if they were not being observed. The problem is not unique to this program. It has sometimes been referred to as the experimenter effect, Hawthorne effect, restaurant manners effect, and job interview effect.”
Reference: De Amici D, Klersy C, Ramajoli F, Brustia L and Politi P (2000) impact of the Hawthorne effect in a longitudinal clinical study. The case of anesthesia. Control Clin Trials. 21: 103-14

Hello-goodbye effect

This is part of the psychodynamics of some people, whereby they initially present themselves in the worst possible light in order to become eligible for treatment, and then after treatment present themselves in the best possible light in an attempt to signal substantial gain in their condition. If uncorrected, the health gain of the patient will be overestimated. The hello-goodbye effect is a serious threat to the validity of clinical studies, meta-analyses and policy making.


John Henry effect

The John Henry effect is an experimental bias introduced into social experiments by reactive behavior by the control group.

In a controlled social experiment if a control is aware of their status as members of the control group and is able to compare their performance with that of the treatment group, members of the control group may actively work harder to overcome the "disadvantage" of being in the control group.

For example, if in an educational trial where the school classes who are in the treatment receive an extra support teacher, students who are in the control group may be induced to work harder to overcome that disadvantage.

The term was first used by Gary Saretsky (1972) to describe the behavior of a legendary American steel driver in the 1870s who, when he heard his output was being compared with that of a steam drill, worked so hard to outperform the machine he died in the process.

Pygmalion effect

The Pygmalion effect, or Rosenthal effect, is the phenomenon whereby the greater the expectation placed upon people, the better they perform. The effect is named after the Greek myth of Pygmalion, a sculptor who fell in love with a statue he had carved.
 

Wednesday, April 01, 2015

Standardized MedDRA Queries (SMQs) and Standardized Drug Groupings (SDGs)

Clinical trial data collection usually includes the collection of adverse events, concomitant medications, and medical histories. The data is usually collected as the verbatim term in free text fields. The data collected with free text fields is not possible for meaningful summary and analysis. Therefore, the first step will be to perform the medical coding to group the same type or similar type of events or drugs together.

The industry standard for medical coding in clinical trials is to use MedDRA (Medical Dictionary for Regulatory Activities) for adverse events and medical histories and to use WHO-DD (World Health Organization – Drug Dictionary) for concomitant medications.

MedDRA hierarchy includes five layers from top to bottom:
  • System Organ Class (SOC)
  • High Level Group Term (HLGT)
  • High Level Term (HLT)
  • Preferred Term (PT)
  • Lowest Level Term (LLT)
During the coding, verbatim or reported term will be mapped to LLT ->  PT -> HLT -> HLGT -> SOC. The summary and statistical analysis will typically be based on Preferred Term and / or System Organ Class.

WHO-DD adopts Anatomical Therapeutic Chemical (ATC) Classification System and the hierarchy includes five ATC levels from top to bottom:

  • ATC Level 1: The anatomical main group
  • ATC Level 2: The therapeutic main group
  • ATC Level 3: The therapeutic/pharmacological subgroup
  • ATC Level 4: The chemical/therapeutic/pharmacological subgroup
  • ATC Level 5: The chemical substance
The ATC classifications are often used as starting points in the development of protocol violation lists and other medication of interest lists in clinical trials. The classes are also used to identify class effects in safety signal detection.

During the coding, verbatim or reported drug names will be mapped to ATC Level 5 -> ATC Level 4 -> ATC Level 3 -> ATC Level 2 -> ATC Level 1. Summary and statisitcal analyses will typically be based on ATC level 1 and ATC level 3, but can also be based on ATC level 2 and ATC level 4.

ATC Level is not as obvious as MedDRA hierarchy, however, both dictionaries use the tree structure. From top to bottom, the choice of terms expands.

In certain situations, the typical coding based on these hierarchy levels is not sufficient. Additional grouping may be needed. For example, if we are specifically interested in any pulmonary hypertension adverse event, we will need to group any of the following terms together:

Acute right ventricular failure
Chronic right ventricular failure
Cor pulmonale
Cor pulmonale acute
Cor pulmonale chronic
Portopulmonary hypertension
Pulmonary arterial hypertension
Pulmonary arterial pressure abnormal
……
Pulmonary artery wall hypertrophy
Pulmonary endarterectomy
Pulmonary hypertension
Pulmonary hypertensive crisis
Pulmonary vascular resistance abnormality
Pulmonary vein occlusion
Pulmonary vein stenosis
Right atrial dilatation
Right ventricular hypertrophy
Right ventricular systolic pressure increased
Vascular resistance pulmonary increased
Angiogram pulmonary abnormal
……
Portal hypertension
Pulmonary infarction
Pulmonary microemboli
Pulmonary vascular disorder

 
Similarly, we may be specifically interested in certain type of medications to define the inclusion/exclusion criteria or to identify the protocol deviations for subjects who take the prohibited medications.  In a diabetes study, we could utilize ‘Drugs used in diabetes’ to find patients potentially having diabetes, then prohibit the drugs affecting the insulin level. In a study in menopausal women, we may need to define a broad or narrow scope of Estrogens. The narrow scope will include substances classified as estrogens based on their therapeutic and chemical properties and the broad scope include substances having estrogen like effects.

 
Fortunately, both  MSSO (the organization for maintaining MedDRA) and UMC (the organization for maintaining WHO-DD) are working on providing these additional groupings. For MedDRA, the additional groupings are called SMQ which stands for Standardized MedDRA Queries; for WHO-DD, the additional groupings are called SDGs which stands for Standardized Drug Groupings.


Standardised MedDRA Queries (SMQs) are developed to facilitate retrieval of MedDRA-coded data as a first step in investigating drug safety issues in pharmacovigilance and clinical development. SMQs are validated, pre-determined sets of MedDRA terms grouped together after extensive review, testing, analysis, and expert discussion. SMQs are a unique feature of MedDRA and provide a strong tool to support safety analysis and reporting. The SMQ topics are intended to address the important pharmacovigilance topics needed by regulatory and industry users. SMQs have been developed with the CIOMS Working Group on Standardised MedDRA Queries that provides pharmacovigilance expertise and validation of SMQs. The SMQs are maintained with each release of MedDRA dictionary by the MSSO.

SMQs are included in each release of the MedDRA dictionary with no additional charge.


The Standardized Drug Queries (SDQs) are new additional classification tools in the WHO Drug Dictionary Enhanced. They have been developed together with a group of expert dictionary users and cover different types of classifications e.g. NSAID, cardiac drugs, Old form, CYP etc.

A SDG is a grouping of medicines having one or several properties in common. The individual grouping can be based on indication, chemical properties, pharmacodynamic properties or pharmacokinetic properties as well as any other property of interest. SDGs can be used in pre-qualification of patients, prohibited medicines, or safety- and subgroup analysis. Standardized Drug Groupings simplify creation of protocol violation lists,  medications of interest, or other relevant
medication groupings.

SDGs are available per request (not automatically provided with WHO-DD release) and are free of charge to all users of the WHO Drug Dictionary EnhancedTM (WHO DDE).  The list of currently available SDGs can be found at SDG product leaflet for drug safety.pdf

 
There are many presentations describing the use of SDGs.