Assessment of the Selection Process for the 2004 National List for Essential Drugs (NLED)
of Thailand
Yoongthong W1, Kunaratanapruk S1, Sumpradit N1, Chongtrakul P2
1 Thai Food and Drug Administration2 Chulalongkorn University
Abstract
Assessment of 2004 NLED selection -Slide 1- 04.02.04
Problem Statement: Essential drugs (ED) satisfy the healthcare needs of the majority of the population. They must be effective, safe, and accessible. Careful selection of ED is important because it leads to a higher quality of care, better management of medicines/diseases, and more cost-effective use of health resources; and therefore ultimately improves public health as a whole. To achieve the goal, the ED selection process in Thailand dramatically changed from the use of an implicit system in 1999 (i.e., primarily relying on expert opinions) to an explicit system in 2003 (i.e., evidence-based medicines). In the explicit system, the evidence regarding characteristics of information (I), safety (S), administration restriction (a), frequency of use (f), and efficacy (E) of each medicine is calculated to form an ISafE score. The ISafE score divided by cost per patient per day generates a final score for that medicine. Although a majority of experts in NLED working groups appear to agree on the benefit of the scoring system in ED selection, it is important to evaluate this new method objectively. This study focuses on further understanding the system and how to improve its quality, as well as to promote the system’s sustainability.
Objectives: To assess the NLED working groups’ opinions regarding the NLED selection. Specifically, we examine how the working groups think and feel about structure, process, and outcome associated with the establishment of the NLED.
Design/Setting/Population: The study is based on a mailed survey with a self-administered questionnaire.Questionnaires were mailed to a total of 270 participants who are physicians and pharmacists officially appointed as experts and secretariats in in the 15 working groups for the NLED selection. Data collection has been conducted during Jan – Feb 2004.
Outcome Measures: The questionnaire includes three major domains. The first set of questions relates to the assessment of structure, i.e., informational components and databases used in the NLED selection. The second set of questions involves the evaluation of process, i.e., the transformation of evidence-based drug information into scoring. The third set of questions involves the appraisal of outcomes, i.e., the composition (amounts and items) of the NLED. Responses will be recorded using a five-point scale from 1 (very inappropriate/ dissatisfied) to 5 (very appropriate/satisfied). Personal background will be requested as possible factors affecting opinions. The questionnaire was pre-tested for its validity.
Results and Conclusions: The preliminary analyses (N = 67) showed that participants thought that publication quality, severity of drug’s risks, efficacy information, and cost is appropriate to use for the NLED selection process (mean scores equal 4.4, 4.2, 4.6, and 4.1, respectively). However, they were not sure about the appropriateness of the use of number of publication/drug’s risks, databases, and scoring transformation (mean scores of 3.7, 3.8, and 3.7, respectively). Overall they were moderately satisfied with the use of scoring system (3.7). Community’s need, NLED philosophy, physician’s experiences, and scoring were major factors that participants mostly used in the NLED selection, whereas influences from drug companies, family/friends, and medical professional association were least likely to be used.
Submitted: February 15, 2004
Introduction
Assessment of 2004 NLED selection -Slide 2- 04.02.04
• Importance of the National List of Essential Drugs (NLED)– Promote rational use of medicines
• The NLED selection is changed.– In 1999, it mostly relied on scoring system that is based on expert opinions. – In 2004, the scoring system was exclusively based on evidence-based materials.
However, consensus from the expert panels can modify scoring as appropriate.
• NLED Score is a tool to enhance objectivity and systematic approach to the ED selection process. NLED Score = ISafE score / Cost of treatment
• ISafE score is derived from an integration of data regarding Information, Safety, administration restriction, frequency of use, and Efficacy
Study frameworkStructure–Process–Outcome Model of Quality of Care• Introduced by Avedis Donabedian
Quality of care Quality of NLED (operationalization)
Structure Attributes of care setting e.g.,
- Adequacy of facilities
- Qualification of medical staff
Attributes of scoring
- Information
- Databases
Process What is actually done in giving and receiving care e.g.,
- Technical style
What is actually done in getting scores
- Transformation process from information to scores
Outcome Results of the care e.g.,
- Clinical end points
- Satisfaction with care
- Well-being, and etc.
Results of scoring
- Satisfaction with scoring and outcome
- Compatibility of selected items with NLED philosophy
- Amount of selected items Assessment of 2004 NLED selection -Slide 3- 04.02.04
Objectives
• Goal: • To explore how experts in the NLED working groups think and feel about the 2004 selection process of NLED
• Objectives: • Objective # 1: To examine overall opinions about scoring used in the NLED selection• Objective # 2: To explore opinions about approaches and factors used in NLED selection •Objective # 3: To investigate relationships between structure, process, and outcome components of the scoring approach used in the NLED selection
Assessment of 2004 NLED selection -Slide 4- 04.02.04
Methods: Classification
15 Working Groups
Branch one: NLED Scoring Branch two: Non-NLED scoring
Questionnaire A Questionnaire B
Assessment of 2004 NLED selection -Slide 5- 04.02.04
Consists of 4 working groups e.g.,1. Occupational and Toxicology2. Nutrition subgroup3. Dentistry4. Radiopharmaceuticals
Consists of 11 working groups e.g.,1. Neurology and Psychiatry2. Cardiology 3. Infectious diseases4. Dermatology
Design
– Exploratory study based on mailed survey of self-administered questionnaires
Data collection procedures
– Questionnaires were mailed to all staff members
– Follow-up telephone calls were made to non-respondents
– Overall response rate was 43%
Measures
– Structure: Appropriateness of information and databases
– Process: Appropriateness of transformation from information to scoring
– Outcome:
• Satisfaction: Satisfaction with scoring, Satisfaction with the selected items
• NLED Quality: Compatibility b/w the selected items with the NLED philosophy, and Amount of the selected items in NLED
Methods: Study design
Assessment of 2004 NLED selection -Slide 6- 04.02.04
Results: DemographicsFactor Classification Branch one: Scoring Branch two: non-scoring
Number % Number %
Gender Male 46 56.1 10 38.5
Female 36 43.9 17 61.5
Age range
30-39 17 20.0 2 7.1
40-49 27 31.8 11 39.3
50-59 35 41.2 15 53.6
60 or older 6 7.1 0 0
Career Physician 62 71.3 14 50.0
Non-physician 25 28.7 14 50.0
Work place
University 43 49.4 14 50.0
Non-university 44 50.6 14 50.0
Assessment of 2004 NLED selection -Slide 7- 04.02.04
Responses from branch one are displayed for this presentation
Objective # 1: Overall opinions about NLED
Assessment of 2004 NLED selection -Slide 8- 04.02.04
Component Items Mean (SD) Range
Structure Appropriateness of ISafE and cost information (12 items) 3.96 (0.43) 3.46– 4.55
Appropriateness of evidence-based databases (6 items) 3.82 (0.52) 3.63 – 3.96
Satisfaction with evidence-based databases 3.70 (0.65) 2.00 – 5.00
Process Appropriateness of score transformation (5 items) 3.73 (0.52) 3.56 – 3.89
Outcome Satisfaction with using panel’s judgment to adjust scores 4.07 (0.57) 2.00 – 5.00
Satisfaction with scoring as a selection tool 3.73 (0.73) 1.00 – 5.00
Satisfaction with the items selected in the NLED 3.84 (0.57) 2.00 – 5.00
Compatibility b/w selected items and NLED philosophy 3.51 (0.68) 2.00 – 5.00
Amount of drug items in the NLED (3 = appropriate)* 3.29 (0.74) 1.00 – 5.00
Other opinions
Compatibility b/w the selected items and one’s decision 3.30 (0.67) 2.00 – 5.00
Understanding about scoring system 3.70 (0.68) 2.00 – 5.00
-Appropriateness/Satisfactory is measured with a 5-point bipolar scale: 5 = very appropriate/satisfied to 1 = very inappropriate/dissatisfied-* Amount of selected items is measured with a 5-point bipolar scale: 5 = too many to 1 = too few. -Compatibility/Understanding is measured a 5-point unipolar scale: 5 = very compatible/well understood to 1 = not at all compatible/understand
Objective # 2: Comparison of approaches and factors used in the NLED selection
Assessment of 2004 NLED selection -Slide 9- 04.02.04
Comparison of approaches Descriptive statistics t-statistics
Mean (SD) df t p
Expert opinion vs. NLED Scoring 2.33 (0.92) vs. 2.40 (0.82) 74 0.55 0.587
Expert opinion vs. Combined approach 2.33 (0.92) vs. 4.32 (0.68) 74 14.88 0.000
NLED Scoring vs. Combined approach 2.40 (0.82) vs. 4.32 (0.68) 74 16.91 0.000
Rank Factors Mean (SD) Rank Factors Mean (SD)
1 Community need 4.01 (0.99) 2 Ind. patient’s need 3.55 (0.97)
NLED philosophy 3.80 (0.89) 3 Reimbursement 2.99 (0.98)
2 Physician’s experience 3.68 (0.77) 4 Professional association 2.35 (1.11)
NLED Scoring 3.59 (0.74) 5 Family/friends’ need 1.55 (0.91)
Government budget 3.58 (0.85) Drug company 1.52 (0.78)
Factors used by the panel in NLED selection (1 = never used, 5 = used a lot)
Approaches used in the NLED selection (1=very inappropriate, 5=very appropriate)
Objective # 3: Structure-Process-Outcome relationship for the scoring approach
= 0.28**
= 0.15**
= 0.003
= 0.56*
= 0.45*
= 0.78* = 0.59*Information
characteristicsR2 = 0.08
Evidence-based databases
Process of scoring
transformationR2 = 0.25
Satisfaction with the use of panel’s
judgment in score adjustment
Satisfaction with scoring system
R2 = 0.32Quality of
NLED R2 = 0.36
Structure Process Outcome
Satisfaction with items in NLED
R2 = 0.29
Intermediate Outcomes
Assessment of 2004 NLED selection -Slide 10- 04.02.04
The panel thought that NLED scoring approach could increase the NLED quality.- Appropriate information increases the appropriateness of the scoring transformation.- When the scoring transformation is appropriate, satisfaction with the scoring system increases. - Satisfaction with scoring system and panel’s judgment in score adjustment leads to the satisfaction with items selected in the NLED- Satisfaction with the items selected into the NLED increases participants’ evaluation of NLED quality.
* p < 0.001 ** p < 0.05
Discussions
Assessment of 2004 NLED selection -Slide 11- 04.02.04
• Objective 1: • Participants had favorable opinions about the 2004 NLED scoring approach (i.e., majority of responses had mean scores >= 3.5). This is probably because:
• NLED scoring approach may simulate panel’s thinking process of drug evaluation• It is based on well-established clinical evidence, and therefore it is accepted by the panel
• The magnitude of favorability may be improved if: • the panels understand NLED scoring approach better (mean score = 3.7)• revisions are made to address the panel’s concerns about NLED scoring approach
• Objective 2: NLED Scoring approach with expert opinion is viewed as the most appropriate way. This is probably because:
• The combined approach may minimize limitations of each other.
• Objective 3: Structure-Process-Outcome relationship • NLED Scoring approach partially addresses participants’ views of the NLED quality (R2 = 0.36).
Conclusions, Limitations & Recommendations
• Conclusions– The 2004 NLED scoring approach is useful for the establishment of the
national list of essential medicines.
• Limitations– Data collection was conducted during different stages of selection process
among working groups
– Adoption level of scoring approach is different among working groups
• Recommendations– NLED scoring system needs further revisions.
– Intensive training for the panel is necessary.
Motivation & DTC advertising -Slide 12- 04.01.04Assessment of 2004 NLED selection -Slide 12- 04.02.04