 | Developing HIV/AIDS Treatment Guidelines (UNAIDS, 1999, 43 p.) |
 |  | Part 2: Developing valid HIV/AIDS guidelines |
| | Steps in developing HIV/AIDS treatment guidelines |
|
|
6. Identifying and applying scientific evidence
At present, expert committees develop guidelines without formal literature reviews. This approach relies heavily upon the group's knowledge of the literature and its members experience of routine clinical practice. However, clinicians knowledge of the literature is often incomplete and their clinical experience may be biased, causing them to overestimate or underestimate the effectiveness of treatments. A selection of articles dealing with the methods of identifying and synthesizing evidence can be found in the systematic review series of the British medical journal 12,13.
|
Developing a standard treatment guideline in Zimbabwe and Malawi In Zimbabwe The Essential drugs list for Zimbabwe (EDLIZ) has been evolving continuously since the first edition of the guideline document was released in 1981. The original idea was to have a standard list of essential drugs to promote rational drug prescribing, reduce drug costs and ensure an equitable quality of care in all health facilities. Its content has been expanded in the 1994 issue to include HIV-related disease. The section on HIV/AIDS promotes the use of consistent clinical methods and a balance between possible interventions and available resources. Graphics show cost comparisons between different options, promoting cost consciousness when prescribing. The success of the guideline is in part due to the change in focus, from expert clinicians, to users at all levels of the health care system. All had an input in constructing, field-testing and revising the guidelines. This rigorous process of preparation and review has been used over the years and refined to the needs of the country. A National Drug and Therapeutics Advisory Committee is responsible for regular review of the guidelines whenever necessary. Promotion of the guidelines and their production in sufficient quantities have encouraged wider usage. A national survey of prescribing habits in 1993 found that 94% of the drugs prescribed in the public sector were prescribed by generic name and all but 2.5% of prescribed drugs were EDLIZ drugs14. A training manual for the 1994 edition was published to attract wider usage. In Malawi The Malawi guidelines have developed along more or less the same principles as those applied in Zimbabwe. Additional lessons drawn from the Malawi experience included which technical problems to anticipate in the printing and production process15. · Provision of a detailed specification of the guideline ensured that the printer produced the document exactly as and when it was required. The Malawi guidelines were delayed by almost two months before they were available for use in training seminars. Part of the costs for printing were covered by the National AIDS Control Programme thereby reducing the overall cost of printing. The benefit to the AIDS programme was the inclusion of more treatment recommendations for HIV/ AIDS. · The layout and quality of the final guideline influences its acceptability. The cover design included an illustration in colour of a map of Malawi and some tablets, with special fonts for the wording. Durable paper was used to ensure it lasted longer. · During the distribution a careful plan was made to have a formal introduction process attended by a respected political leader and the press. The audience was also informed about the aims and benefits of guidelines. · The section on HIV/AIDS refers the user to the related section in the rest of the guideline. For example, the section dealing with chronic diarrhoea is detailed in the general section on diarrhoea. This promotes better integration of the guideline and reminds users that symptoms of HIV/ AIDS can be treated with commonly available drugs. References are also made to the Malawi HIV management guidelines based on the WHO HIV algorithms. Users requiring greater detail on conditions they are dealing with can refer to this document. |
Studies conducted on HIV/AIDS clinical care are available from several sources. Databases such as Medline and Popline are easily accessible and can provide information on specific aspects of HIV/AIDS clinical care and guideline development. This can then be modified as the information required is identified. Some useful search terms are "HIV", "AIDS", "randomized controlled trial", "reviews", "meta-analysis", "clinical care", "opportunistic infections", "decision-making" and "treatment".
Research reports on HIV/AIDS clinical treatment conducted in the country should also be collected and reviewed. Information from physical searches of journals and publications or other bibliographic sources will supplement the databases. The current information on the nationwide morbidity patterns and clinical presentation of HIV infection would also be useful in determining the burden of disease and the quantity of drugs to procure.
Once the relevant articles have been identified, the committee prepares abstracts for review. Methods for assessing the quality of these studies and reports include consideration of the extent to which the studies provide valid information about the condition, and whether the results can be applied to other settings.
There are several ways to apply the evidence. The committee should adopt whichever is the most appropriate for sifting through the large mass of clinical data and arriving at recommendations. They include:
· clinician knowledge,
· unsystematic literature review,
· systematic reviews,
· ungraded systematic reviews,
· formal meta-analysis.
All five approaches have been successfully used to develop guidelines. Combining these different approaches will give the best possible interpretation of the evidence collected, although this may prolong the process and increase the overall cost of developing the guidelines.
The above approaches were specifically applied to HIV/AIDS clinical care to identify evidence for the development of the United States Public Health Service and the Infectious Diseases Society of America (USPHS/IDSA) Guidelines16. The lack of evidence in a specific area is a common difficulty encountered in attempting to frame evidence-based recommendations. The USPHS committee found more evidence in support of treatment of opportunistic infections than for counselling strategies, as many more studies have been conducted on HIV clinical management.
The United States AHCPR has developed a useful model for grading the level of evidence (Table 2). In most countries the commonest form of evidence is based on results of expert committees (Table 2, level IV) and often does not involve extensive review and analysis of scientific data. Increasingly, recommendations developed in the United States of America are based on scientific data using the levels Ib and IIa shown in Table 2. Where a combination of levels of evidence is used to define recommendations, it is necessary to indicate their strength. An explanatory note of this should also be included in the final document. Another use of level IV evidence (Table 2) is in the formulation of prospective studies or trials that produce more convincing evidence.
|
Methods of synthesizing evidence Clinician knowledge This is commonly used when clinical information is lacking and only the clinician's experience of the condition under discussion is available. Various methods used to collect this information are discussed below in section 7. The information may however be incomplete owing to clinicians not having time to research, collate and interpret the data from a variety of sources. A bias may also be introduced by the patient's or physician's enthusiasm for a cure, leading to an overestimation of the effectiveness of the treatment. Guidelines developed in this way also tend to reinforce prevailing biases and practices rather than produce new approaches to care. Unsystematic literature review This method has no explicit search strategy, inclusion criteria or a formal method of synthesis. Information of variable quality is collected in a random manner and may be difficult to collate and interpret. This method may also result in a selection bias due to omission of some studies in the search. Systematic reviews These avoid the pitfalls encountered in the previous method by applying search criteria. Only studies deemed to be methodically correct are included in the search and utilized in the analysis, thereby increasing the validity and reliability of the outcome. Ungraded systematic reviews This method typically reviews only randomized trials and can consequently omit some very important studies with a different design methodology. Formal meta-analysis This pools all the information from multiple randomized trials and summarizes all the reviewed evidence by a single statistic. Pooled relative risk of mortality is a commonly used statistic which has the effect of increasing the power of studies that lacked a large enough sample size to produce significant results. Generally these studies are more reliable and yield more validevidence. However, the selecting of only randomized trials may also result in wrong interpretation and misleading results. |
Organizations like WHO and UNAIDS can facilitate the development of evidence-based guidelines, by supporting studies on opportunistic infections and the efficacy of preventive interventions. Multicentre studies can be conducted to allow for application of results in a number of countries.
WHO's Reproductive Health Library has a systematic review of clinical trials on priority reproductive health topics. This is aimed at promoting evidence-based guideline development in developing countries. It is accessible electronically and is free of charge to developing countries from WHO17 or through a paid subscription to Update Software Ltd., Summertown Pavillion, Middleway Oxford OX2 7LG, United Kingdom. A similar initiative for HIV/AIDS guideline development, if promoted by UNAIDS, would support countries that did not have the financial capability to conduct clinical trials.
A conference in 1997 on improving the use of medicines (ICIUM)18 identified progress in developing countries over the past decade. Over 120 papers on several themes were presented. Conference proceedings are available from the WHO/DAP web site (http: //www.who.int/dap-icium/index.html).
Table 2: Levels of evidence
|
Level |
Type of evidence (based on AHCPR 1992 19) |
|
Ia |
Evidence obtained from meta-analysis of randomized control trials |
|
Ib |
Evidence obtained from at least one randomized control trial |
|
IIa |
Evidence obtained from at least one well designed control study without randomization |
|
IIb |
Evidence from at least one other type of well designed quasi-experimental study |
|
III |
Evidence obtained from well designed non-experimental descriptive studies, such as comparative, correlation and case-control studies |
|
IV |
Evidence obtained from expert committee reports or opinions and/or clinical experience of respected authorities |
|
|