Systematical reviews and meta-analyses

Размещено на http://www.allbest.ru/

Chair of clinical pharmacology and evidence medicine.

Discipline "Evidence basic medicine».

Course 3

Speciality 051301 «General medicine».

THE INFORMATION-DIDACTIC BLOCK

ON « Evidence basic medicine »

TO PRACTICAL EMPLOYMENT № 12.

Theme: «Systematical reviews and meta-analyses»

Systematical reviews

Personal clinical experience is important for formation of hypotheses, but because of the factor of partiality and high susceptibility to displacement of estimations it isn't so useful to quantitative estimations though before introduction of controllable clinical tests by a pharmacotherapy basis there were personal recommendations.

Comparison of proofs

The main lack of descriptive reviews -- their susceptibility to displacement of estimations. For displacement minimization it is necessary to approach the review to accessible proofs by means of the scientifically- systematical approach which has received the name of the metaanalysis and powerful development in works of Association Kohrane. The review executed thus is considered as systematical and includes six basic stages.

* definition of a main objective of the review;

* definition of ways of an estimation of results;

* regular information search;

* summation of the quantitative information;

* summation of proofs with use approaching statis-ticheskih methods;

* interpretation of results.

As well as in any scientific research, these stages are fixed in the report.

The systematical review -- a version of scientific research with in advance planned methods where as object of studying results of some original researches serve. The systematical review synthesizes results of these researches, using the approaches reducing possibility systematical and random errors. In the majority of systematical reviews use statistical methods of generalization of the data -- the metaanalysis. In process of receipt of the new data their conclusions can change.

The systematical review -- the structured process including:

correctly formulated question;

high-grade and professional information search;

not displaced (faultless) process of selection of publications and extraction of the fact sheet from them;

a critical estimation of the data;

synthesis of the data.

Correctly formulated clinical question -- a basis of the high-quality systematical review. At badly formulated question the review hardly will be informative enough. Besides, accurately formulated question allows the doctor to understand quickly, whether results of the review in its clinical practice are applicable. On the basis of such review corresponding conclusions become:

intervention, is doubtless, effective also it it is necessary to apply;

intervention is inefficient also it is not necessary to apply;

intervention harms also it is necessary to forbid;

the advantage or harm aren't proved, the further researches are required.

Systematical reviews consider results only well planned scientific researches which take place examination on reliability. It basically controlled prospective (participants divide into groups which will be or won't be subjected studied intervention before there have come outcomes) researches, with division of patients on skilled (new treatment is applied) and control groups (standard treatment is applied). The word "controlled" means that at the organization of tests action of the extraneous factors, capable to affect result of comparison of examinees of methods, is considered (is under control). That average indexes in both groups differed as less as possible, randomization (English random -- casual) is spent.

For this reason the scientifically-evidence medicine doesn't consider usual literary reviews as information reliable sources. Systematical reviews spend to find answers to certain, often narrow clinical questions in which accurate information following parameters should be reflected:

Certain population and clinical base (for example, children, outpatients);

* studied disease (for example, a bronchial asthma);

* a used method of research or treatment (for example, certain medicinal therapy);

* one or more certain clinical outcome (for example, frequency of attacks, quality of life).

The quantitative systematical review of the literature, or quantitative synthesis, the primary data for the purpose of reception of total statistics is called as the metaanalysis.

The systematic review (systematic overview) -- the special kind of the scientific research executed by specially developed technique which object are results of others, are original scientific researches. The analysis and generalization of results of the researches included in the review, it is possible to spend with application of statistical methods and without them. Systematic reviews allow to spend adequate generalization of results of scientific researches. The purpose of the systematic review is the weighed and impartial studying of results before the spent researches. Often in the systematic review the quantitative estimation of the total effect established on the basis of results of all studied researches (metaanalysis) is spent.

The history of one of the first systematic reviews is remarkable. Results of the first clinical research devoted to efficiency of an inexpensive short course of corticosteroids which appointed to pregnant women with high risk of premature birth, have been published in 1972 Has been established that corticosteroids reduce probability of death of newborns from the complications caused недоношенностью. By 1991 messages even about seven clinical tests devoted to the same theme, confirmed and specified results of the first clinical test have been published. It has been proved that the risk of death at such children can be lowered on 30--50 %. The schematic image of results of this research became an emblem of Cochrane Collaboration. As till 1989 it has not been published any systematic review of the given clinical tests, the majority of doctors considered such therapy inefficient. Thus, has passed about 20 years since the publication of the first results of research before their wide introduction in practice. By a rough estimate of ten thousand children have suffered and have died of absence of necessary therapy. It is an impressing illustration of necessity of introduction of principles of demonstrative medicine in practice (The Cochrane Collaboration, 2001).

Systematic reviews -- the main result of activity of Cochrane Collaboration -- are regularly published in electronic form under the name «The Cochrane Database of Systematic Reviews» (the Cochrane database of systematic reviews).

The basic characteristics of the systematic review are resulted in table 1.

Table 1. The basic characteristics of the systematic review

Aim	Characteristic
Clearness of the purpose of the review	The report containing research problems, the description of objects and research methods
To avoid not inclusion in the review of relevant researches	Exhaustive, sensitive and documentary strategy of research with use of bibliographic databases, keywords, probably manual search, probably attempt of inclusion of the unpublished researches which have been not limited to the country or language
To avoid a casual choice / of an exception of researches	The detailed and verified criteria of a choice and the exceptions developed for an estimation of results of research
Summarization of the given researches	Use of tables of selection of the data with check of their completeness
To estimate a validity of results of research	To develop and use criteria of quality for an estimation of a validity of researches by means of an estimation of design, carrying out and the analysis of researches with an estimation of the size of an error, displacement and chances
To estimate the size of associations and a source of a variety of the data of research	The review of the bases why results of research can use various corresponding quantitative models for an estimation of a role of such factors, as a condition of the patient, doses, duration and the intervention nature. There where it is possible, researches unite for cumulative effect reception
To estimate quality of results of the review	Check of sensitivity of results of elections and assumptions made in the review, such as criteria of inclusion and a validity which influence research on a method used for association of the data
Critical estimation or review repetition	The report on key aspects of creation of the review, methods, the analysis and results. It should include the report resume, search strategy, the table of basic elements of each included research. The report can be supplemented with the graphic description
To help the reader to estimate application of the review for the needs of organizers of public health services practising doctors, researchers	Discussion of methodological restrictions, as concerning primary researches, and the review. To provide use of proofs of research, including by working out of recommendations about application of results of research

Advantages of systematic reviews and the synthetic approach to the analysis of results of researches consist in the following.

* Help to democratize scientific researches and use of their results, promote большей to a knowledge base openness, including for the public. It can reduce value of "experts" as concentrations of knowledge that in the past led to increase in number of errors.

* Gives to scientific community an invaluable source infor-matsii which allows to generalize the material saved up in the past, to order new materials and to develop methodology of definition by that and key directions of scientific researches. It allows to provide application of results of researches and carrying out of an estimation of theories all over the world.

* Gives possibility for researchers to define the basic blanks in researches, to offer new decisions within the limits of the saved up knowledge and to avoid unnecessary duplication of researches.

* Gives the knowledge base for practising doctors and organizers of public health services who can use reviews for an estimation of efficiency of various forms and kinds of rendering of medical aid.

* Promote cumulative development of a science. Extremely seldom research is spent on absolutely new soil, we «stand on shoulders» the previous researches.

* Help to define more clearly what we don't know also ambiguity degree.

Use of systematic reviews not only allows to receive the best results, but also to raise their quality and trust degree to them.

In the field of evidence medicine systematic reviews of Cochrane library (the Russian branch -- http://www.Cochrane.ru/cochrane/rus_otd.htm) are considered as the most authoritative. Unfortunately, the overwhelming majority of articles and reports of area of applied medicine which are published at us in the country, to evidence medicine have very far relation. Often group of patients, subjected to a new method of treatment, compare to historical control group or in general there is no control group, randomization is incorrectly spent or not spent at all, statistical methods are inadequately applied, sometimes researchers (and lecturers) happen are economically interested in certain results.

Distinctions between systematic reviews and literature reviews

All reviews concern retrospective observant researches and consequently are subject regular and to random errors. Accordingly quality of the review and, hence, its value depend on use at its carrying out of the scientific methods reducing to a minimum these errors. This core that distinguishes traditional reviews of the literature from regular (tab. 2.). If the review is spent according to the principles stated in the right column of the table, it is more probable that it will be systematic and will provide the most objective conclusions. If methods of carrying out of the review correspond resulted in an average column, it is faster the literature review; probability of that its conclusions are based on faultless association of all data concerning this point in question, much more low.

Table 7-2. Distinctions between reviews of the literature and systematic reviews

The characteristic	The literature review	The systematic review
Taken up questions	The wide spectrum of questions is often considered	It is often devoted a certain clinical question
Sources of the data and search strategy	Sources aren't always specified, strategy can be erroneous	Sources usually all volume, and search strategy is precisely stated
Principle of selection of the data	It is not always specified, can be erroneous	Selection is based on certain criteria, applied by the identical in the image
Methods of an estimation of the data	The various	Strict, critical methods of an estimation
Generalization of the data	Often qualitative	Quantitative (metaanalysis)
Conclusions	Sometimes scientifically proved	As a rule, scientifically proved

Systematic reviews spend to find answers to certain, often enough narrow clinical questions. In accurate information of these questions four parameters should be reflected:

* certain population and clinical base (for example, elderly outpatients);

* studied disease (for example, an arterial hypertension);

* a used method of research or treatment (for example, certain medicinal therapy);

* one or more certain clinical outcome (for example, cardiovascular and цереброваскулярная disease and death rate).

The example of well formulated clinical question -- whether prevents medicinal treatment of an arterial hypertension at elderly patients strokes and heart attacks of a myocardium and whether it increases life expectancy? If the question, has to answer which review, isn't clear from the name, the resume and introductions or if the section with the description of methods of carrying out of the review isn't included, article, most likely, it will appear the literature review, instead of the systematic review.

The majority of reviews of the literature are is more often devoted a wide range of the problems connected by the handed over theme, instead of the deep analysis of a concrete question. For example, the literature review on a diabetes (reminding head from the textbook) can contain sections on physiology and a pathophysiology of an exchange of carbohydrates, fats and fibers, epidemiology of diabetes and to the forecast at it, diagnostic and screening methods, preventive, medical, rehabilitation and palliative interventions. Thus, literature reviews are most useful to reception of wide representation about a theme, they are much less useful by search of the concrete answer to a certain clinical question. Literature reviews are good for the description of history of development of a question and its decision, they can present is better advanced achievements if the conducted researches were not numerous, preliminary or not so authentic because of lacks of their organization or execution. They can be especially useful to discussion of the data in the light of the basic theory. In literature reviews analogies, speculative association of researches in two independent areas (for example, devoted to a cancer and Acquired Immune Deficiency Syndrom) are admissible. However in such reviews communication between clinical recommendations and the facts, often weak, incomplete, or that is even worse, is based on prejudiced citing of certain original researches. As a result of the recommendation in the reviews of the literature published in magazines or managements, often differ from recommendations of systematic reviews. For example, literature reviews can with delay for a decade recommend the treatment methods which efficiency is proved, or to put arguments in protection of methods of treatment about which for a long time already it is known that they are useless or harmful. Besides, the systematic reviews using quantitative methods, more likely, than literature reviews, will allow to find out small, but clinically significant effects of treatment.

Possibilities and restrictions of systematic reviews

Value of correctly spent systematic review for the practical doctor is difficult for overestimating. Many of us are overloaded by volume of the let out medical literature and as a result often prefer to look through survey messages, instead of publications of original researches. Reviews allow us to remain at level of modern knowledge. High-quality regular reviews help to define borders known and to learn about facts in evidence all scientifically. Separate researches seldom give exact answers to certain clinical questions, and systematic reviews can help to solve to practical doctors them. Critically analyzing primary researches, systematic reviews help to reveal also contradictions between results of various researches. Quantitatively uniting results of several small researches, the methaanalysis allows to do more exact, versatile and convincing conclusions. As example that is served by recently published review which has underlined beneficial effect of restriction of fibers in food on progressing of diseases of kidneys of diabetic and other etiology. Besides, the systematic review of several researches helps to define better, for what subgroups of patients the received results are applicable.

Systematic reviews are necessary also to researchers as allow to summarize the existing data, to improve hypotheses, to define the necessary sizes of sample for research, help to develop the plan of the future researches. Without these reviews researchers can pass by promising directions or study already solved questions. Reviews and other generalizing publications are necessary for organizers of public health services for development of a policy of rendering of the medical services, providing optimum results at available resources.

Systematic reviews can help with the decision of clinical problems, but never will replace clinical experience. Doctors make recommendations the concrete patient, being based on analogies, experience, logic constructions and the theory, no less than on scientifically well-founded results of researches. The awareness on efficiency of any therapy won't replace knowledge of how to use it for treatment of concrete patients. Scientifically well-founded recommendations can lead to bad practical results if are applied not critically or not realizy. The decision-making structure in medicine is difficult; for this purpose it is required to unite knowledge, skills, morally-ethical values and scientifically well-founded facts at each meeting of the patient with the doctor.

The past, the present and the future of systematic reviews

Prevalence of reviews quickly changes; the number of annually published systematic reviews has increased last decade at least in 500 times. Now it is often possible to meet some systematic reviews on the same or close clinical questions. For example, on arterial pressure two metaanalyses published within one month have been devoted an estimation of influence of calcium recently. Though independent reviews duplicating each other in which similar results are received, allow to concern them with the big trust, it is necessary to pay off for it with narrowing of possibilities -- they can distract researchers from carrying out of necessary systematic reviews in other areas. The situation can become even more difficult if in several reviews results are similar, and clinical recommendations differ; The such happens with five metaanalyses on conducting patients in branches of the intensive therapy, published during 5 years (and two of them have been published in one release of the same magazine). To resolve disagreements between metaanalyses it is possible, having spent at high methodological level the new systematic review in which all existing data will be synthesized. For the past of 10 years the methodical recommendations helping critically to estimate and apply review have been developed. The scheme of interpretation of the given original researches and creation of systematic reviews with use of levels evidense`s and degrees of validity of recommendations with success has been used by several groups of researchers. At conciliatory conferences of the American board of lung specialists on prevention of thromboses medical recommendations usually classify according to their degree scientific evidence -- depending on, whether they are based on results of large, strict, randomized researches, metaanalyses, observant researches or on opinion of experts. systematic review metaanalysis

The international commonwealth of scientists named Cochrane Collaboration, is created to help with preparation, constant updating and distribution of results of systematic reviews on medical interventions. Library Cochrane -- the first large-scale multidisciplinary product of this cooperation. It is updated quarterly, and in January, 1997 contained 159 finished systematic reviews prepared within the limits of Cochrane Collaboration (199 more systematic reviews prepare for release now), more than 1600 resumes of other systematic reviews, the list more than 110 thousand controllable researches and the bibliography from 400 articles by a technique of preparation of systematic reviews.

What prospects of systematic reviews? Coverage of various directions of medicine by them though and quickly extends, but is still limited. Cochrane Collaboration works in this direction, but it needs more active participation and support to operate successfully. The researches devoted to a science of synthesis of the data, will allow to raise quality and value of the results received in systematic reviews. More effective methods of carrying out of these reviews will be developed. Access to reviews of patients and heads of public health services that will induce doctors more actively will increase to search, analyze and apply results of systematic reviews in the work. Development of information technology can provide availability of the information directly on a place of rendering of medical aid, but practical doctors should get skills of use of such technology. Data presentation formats become more convenient both for physicians, and for patients. Practical methods of inclusion of the information from systematic reviews in process of development of the clinical decision will be created and are everywhere widespread. In expectation of this progress doctors already now can and should enjoy advantages of systematic reviews.

The basic features of systematic reviews

* In systematic reviews collect, critically estimate and generalize results of primary researches on a certain theme or a problem.

* By preparation of systematic reviews use the approaches reducing possibility systematic and random errors.

* systematic reviews -- analitiko-synthetic articles, concern them also economic analyses, clinical recommendations, analyses of algorithms of acceptance of clinical decisions.

* systematic reviews help doctors to be well informed about the modern information, despite a large quantity of medical publications.

* systematic reviews can help to prove clinical decisions results of researches though in itself they don't allow to make decisions and don't replace clinical experience.

The metaanalysis

The metaanalysis (meta-analysis) -- application of statistical methods at creation of the systematic review with a view of generalization of the results included in the review of researches. systematic reviews sometimes name the metaanalysis if this method was applied in the review. The metaanalysis spend to generalize available information and to extend it in a clear kind for readers. It includes definition of a main objective of the analysis, a choice of ways of an estimation of the results, the systematized information search, generalization of the quantitative information, its analysis by means of statistical methods, interpretation of results. The metaanalysis -- the statistical method, allowing to unite results of independent researches. More often it use for an estimation of clinical efficiency of therapeutic interventions; for this purpose unite results of two and more randomized controllable researches. Informationally of the metaanalysis depends on quality of the systematic review on which basis it is spent. The qualitative metaanalysis assumes studying of all researches devoted to corresponding problem, an estimation of heterogeneity and definition informationally the basic results by the sensitivity analysis. There are some versions of the metaanalysis.

* the Cumulative metaanalysis allows to construct a cumulative curve of accumulation of estimations at occurrence of the new data.

* Prospektive metaanalysis -- attempt of working out of the metaanalysis of planned tests. Such approach can appear comprehensible in medicine areas where already there is a developed network of information interchange and joint programs.

* In practice instead of prospektive the metaanalysis often apply the prospektive-retrospective metaanalysis, uniting new results with earlier published.

* the Metaanalysis of the individual data is based on studying of results of treatment of separate patients. In the near future the metaanalysis of the individual data will be limited, most likely, to studying of the basic diseases which treatment demands the large-scale centralized capital investments.

Advantages:

* reception of authentic results;

* elimination of possible errors;

* accuracy of estimations;

* a transparency.

Complexities:

* revealing and selection of researches;

* heterogeneity of the presented information;

* probability of loss of the important information;

* the inadequate analysis of compared subgroups;

* the inadequate analysis of sensitivity of a method.

The main requirement to the informative metaanalysis -- presence of the adequate systematic review. Results of the metaanalysis usually represent in the form of the schedule and the relation of chances (odds ratio), the total indicator reflecting expressiveness of effect.

Why the metaanalysis

systematic reviews give scientific proofs for introduction of results of researches in practice and decision-making, while the metaanalysis -- an analytical part of regular reviews. The basic lines, characteristic for the metaanalysis are more low resulted.

Specification in the report of research of the research problems, estimated hypotheses (both in the field of medicine, and in the field of biology), the review of a material and methods of systematic reviews before researches will be undertaken.

Association of all accessible primary researches, including information search, with the accurate description of strategy of search and information sources. The choice of researches should be based on the accurate criteria proved by the report of research.

the Estimation of methodological quality of the selected researches (application of the methods reducing an error). An estimation of reproducibility of researches.

Definition of required results of researches, an explanation of differences which are whenever possible spent on each of primary researches.

the Choice and a method of an estimation of results of researches, and also a subject of researches characterize in the standardized form under the primary documentation of research with check of an error of sample. Procedure should be clear, reproduced and with the minimum statistical error.

There where the review and characteristics of the data are executed, the metaanalysis (quantitative synthesis of results of primary researches) uses corresponding methods and models (clearly proved) to consider at calculations all possible reasons of variability of signs (for example, distinctions of quality of researches, participants, doses, duration and character of intervention, definition and measurements of results).

When the data has considerable disorder, too poor quality or high heterogeneity, metaanalysis carrying out represents the big complexity.

Maintenance of clearness of results of the systematic review concerning elections and assumptions is spent at all stages of the analysis. In particular, in the metaanalysis should find reflection:

¦ influence of quality of research and (or) criteria of inclusion;

¦ credibility and possible influence of statistical errors;

¦ influence of various models of strategy of a choice and maintenance of reconstruction of values of the passed data in researches with incomplete results.

the Fair idea of key aspects of all analysis stages in the research report, carrying out of a critical estimation and reproducibility maintenance. This data can be presented in the form of the special table including key elements of each primary research. Graphic representation of results also can help with interpretation and it needs to be included there where it is necessary.

Methodological restrictions both primary researches, and systematic reviews, should be estimated. Any clinical or organizational recommendations should be practical and settling, and to provide clearness of proofs on which basis they are made. Offers of necessary researches should include clinical and methodological requirements to these researches.

Metaanalysis stages

Search of the researches containing the data on the interesting problem.

Definition of criteria of selection for inclusion or an exception of researches of consideration.

Abstragirurovanie of the data from concrete researches.

the Joint analysis of the abstracted data.

Definition of the size of effects for homogeneity check.

Metaanalysis problems

The displaced estimations

There are some types of displacement of estimations in the metaanalysis. Displacement concern the first type because of partiality to the publication of positive results in a counterbalance the negative. The statistical method is offered, allowing to reveal displacement and to eliminate it from estimations. Moreover, at the analysis of sensitivity of a total estimation is standard to estimate number of tests with negative result (reliability index) which would be required for prevention of any observable positive result.

Other types of potential displacement are caused:

* incompleteness of information search;

* unsuccessful criteria of inclusion/exception of sources;

* negligences of a statement in original messages.

It is possible to show that in systematic reviews essentially it is less than these problems in comparison with traditional narrative analogs.

Association of diverse researches

The criticism of the metaanalysis recognizes that in it «apples mix up with oranges» that yields not interpreted results, however metaanalysis competently executed within the limits of the systematic review should overcome this criticality as strict criteria of inclusion and careful interpretation of observable heterogeneity become its essential stages. Really, mixing of "slightly differing grades of a citrus» can essentially improve uniformity of sample by means of various statistical methods.

Inclusion of the unpublished data

In the metaanalysis the basic efforts are directed on identification of all set of tests in certain area -- published or unpublished. And though the last can be methodologically weak, the careful estimation of quality of both groups before inclusion in the metaanalysis eliminates this lack. The similar approach can give the most valuable synthesis of the data.

«The gold standard»

As such standard usually consider well conducted clinical test of corresponding design and the size of the sample, organized for the purpose of an estimation of efficiency of a medical product. When results of several such tests of efficiency of the given medicine are accessible to the analysis, the following from their analysis the facts will be conclusive as under such circumstances the metaanalysis gives the most exact estimations. Problems arise at a divergence of results between one big and group of earlier and small researches. The divergence reasons are necessary for finding out in details, without giving in to aspiration to sum up results in favor of the big research.

Search of researches for the analysis

* Spend systematic and total search of all articles in an interesting theme. If any researches the systemstic error of results of the metaanalysis is passed, possible. Unpublished researches can remain not found out. publicationally systematic error often mention as restriction for the metaanalysis.

* Researches with positive results publish more often, than with the negative.

* Inclusion of unpublished researches without reviewing can't be welcomed.

* Statistical or quasi-statistical methods are offered for an estimation publicationally a regular error, but not all believe that these methods are comprehensible.

Definition of criteria of selection

* the Criteria of selection resulted more low, should be written down in the report before the research beginning:

¦ adequacy of the size of sample;

¦ presence of control group;

¦ completeness of the information on research in a source;

¦ years of the publication;

¦ research type;

¦ similarity of an exposition (a dose, cofactors);

¦ similarity of effects;

¦ control of regular errors;

¦ research restrictions.

* Select system of calculation of the points, exposed to researches.

* On the basis of the established criteria develop the abstraction form.

* Register researches excluded from the analysis, specifying the exception reasons.

Heterogeneous results

If estimations of effects in individual researches are diverse, use of methods fixed and random effects, most likely, will yield various results. Whether it is necessary to unite results which are non-uniform? Many epidemiologists believe that the metaanalysis -- is faster means of ordering of the information accessible on a certain problem, than means of reception of incorporated estimations. In all situations when separate researches are non-uniform by results, it is necessary to establish the reasons of these distinctions.

Problems which can arise at metaanalysis carrying out

* the Sizes of effects completely гетерогенны.

* Revealing during an estimation of several dropping out values.

* In some researches the control group can not be used.

* Can not be effect estimations, only r-value or the F-statistican.

* Absence of the information on a standard deviation or a variation.

* Set of measurements in each research.

Definition of restrictions of research

It is important to know restrictions of research and to define the areas approaching for the further researches.

The most general restrictions

* the Inadequate information on potential stirring factors.

* Absence of the exact information on an exposition.

* Availability of suitable population to comparison.

* Absence of the information on interaction of an exposition and illness at individual level in ecological researches, absence of exact knowledge of the biological mechanism of illness, possibility of a regular error at formation of groups.

* Expositions without casual selection, various percent of a response of respondents in researches "case-control", incorrect classification. Restrictions on studying of time interrelations in cross-section researches. Losses at procaking in longitudinal researches.

Task 1

The metaanalysis of researches of communication between passive smoking of the house and a lung cancer has been spent. In 9 found researches various methods were used. For example, one research was cohort, and others -- a case -- control. One researches used the information on the revealed cases in hospital, others included the healthy people living the houses. Two researches were blind so that the interviewer didn't know with whom deals, with a case of illness or with control. The information on smoking fluctuated from zero (in research among healthy faces) to almost 70 % at patients. Definitions "not smoker" and smoking degree also differed: in one research of the person which informed on smoking off and on, have been classified as not smokers, and in other research comparison spent not between exposed to influence of smoking and not exposed, and between persons who were exposed to influence of tobacco more and less than 4 hours per day.

Whether it is possible to unite relations of chances in these nine researches?

Task 2

Search in MEDLARS together with viewing Current Contents on articles and references has allowed to identify 12 controllable clinical tests of effect of an additive of vitamin A for children's death rate. Four tests spent in hospitals on children sick of measles. Eight were the researches spent in communities, included in children lived in various villages, areas, and their inhabitants have been defined in treatment or control group.

Whether

* It is possible to include all 12 tests in the metaanalysis?

Total control of knowledge

Answer questions

1. Make definition and the basic characteristic of systematic reviews.

2. List the basic differences of systematic reviews from literature reviews.

3. Specify possibilities and restrictions of systematic reviews.

4. Result the basic stages of drawing up of the systematic review.

5. Name principles of selection of researches for systematic reviews.

6. Make definition of concept the metaanalysis.

7. Name the basic types of carrying out of the metaanalysis.

8. Make definition of indicators of results of the researches used in practice of the metaanalysis.

9. List ways of representation of results of the metaanalysis.

10. Give the analysis of the abstract of the spent metaanalysis of interrelation between seropositive on cagA and a stomach cancer.

The metaanalysis of interrelation between it is seropositivity on cagA

And a stomach cancer

Substantiation and the purpose

Messages of the literature concerning mutual relations of an infection with cagA-positive types Helicobacter pylori and a stomach cancer in comparison with only one infecty H. pylori are inconsistent. The purpose of this research consisted in estimating size of risk of a cancer of the stomach connected with cagA-seropozitive, and to identify some reasons of heterogeneity between researches.

Methods

The metaanalysis of researches by a principle "case-control" in a combination to the control coordinated on age and a floor which has provided the preliminary data about frequency of infection Н pylori and cagA Н pylori, revealed serology with a method or by means of PCR been undertaken.

Results

All-round literary search identified 16 qualified researches with 2284 basic cases and 2770 -- control. Н pylori and cagA-seropozitivy authentically increased risk of a cancer of a stomach in 2,28 and in 2,87 times accordingly. Among Н pylori - the infected population the infection with cagA-positive shtamms increased further risk of a cancer of a stomach as a whole in 1,64 times (a confidential interval of 95 % ДИ=1,21--2,24) and in 2,01 times (95 % ДИ=1,21-- 3,32) not cardially a stomach cancer. The kardially cancer of a stomach hasn't been connected with infection Н pylori or cagA-positive types Н pylori. The age of patients and localization of a cancer of a stomach bring the contribution to heterogeneity between researches.

Conclusions

The infection with cagA-positive types Н pylori increases risk of a cancer of a stomach in relation to risk which is caused only by one infection Н pylori. Definition of the cagA-status of infection Н pylori can give additional effect in revealing of populations with high risk of a cancer of a stomach.

Размещено на Allbest.ru