Methodological Framework for Meta–Analysis: the Case of Clinical Studies
Meta–analysis is a statistical analysis summing up data from several analogous studies. Meta–analysis may combine, mathematically, two or more studies aimed at testing the same hypothesis. Its advantages include the increased statistical capacity of a study, entailing a higher accuracy of the measur...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
State Statistics Service of Ukraine, the National Academy of Statistics, Accounting and Audit (NASAA), the National Academy for Public Administration (NAPA) under the President of Ukraine
2015-12-01
|
| Series: | Статистика України |
| Subjects: | |
| Online Access: | https://su-journal.com.ua/index.php/journal/article/view/46 |
| _version_ | 1818493316330684416 |
|---|---|
| author | L. O. Yaschenko M. V. Lyesnikova |
| author_facet | L. O. Yaschenko M. V. Lyesnikova |
| author_sort | L. O. Yaschenko |
| collection | DOAJ |
| description | Meta–analysis is a statistical analysis summing up data from several analogous studies. Meta–analysis may combine, mathematically, two or more studies aimed at testing the same hypothesis. Its advantages include the increased statistical capacity of a study, entailing a higher accuracy of the measured effect of an analyzed parameter, and its cost–effectiveness.
The purpose of the study is to outline the meta–analysis framework and demonstrate its implementation by use of data from clinical research: use of the medical drug Proteflazid® in medical treatment of papillomavirus infection. The meta–analysis covered 8 studies “occurrence – control”. In general, 539 patients with pathologies of cervix and urogenital pathologies caused by papillomavirus infection, and 498 patients of the control group were studied.
The meta–analysis is conducted on the basis of the three indicators:
Frequency of recurrences of condylomatous deceases.
Frequency of detection of human papilloma (F1P) using the method of polymerase chain reaction (PCR).
Frequency of disappearance (essential decrease) of F1P replication using PCR method.
Z–criterion of Fisher is used to measure statistical significance of the combined ratio of chances. The heterogeneity of the set of data from a separate study is measured using the test on heterogeneity The criterion .r–squared is used to evaluate the heterogeneity between the studies. P < 0.05 indicates on a statistically significant heterogeneity. Also, heterogeneity effects are measured by F–test, which is the ratio of the variability of the studies more favorable to the heterogeneity than to the chances. When a result is significant, .r–squared of the criterion (P <0.05) or the result of F–test is higher than 50%, which indicates on the existence of a heterogeneity between the studies, and for this case the model of random effects (DerSimonian–Lerd method) is chosen. Otherwise, the model of fixed effects (Mantel – Flaenszel model) is used. Also, the sensitivity analysis is made, to measure the quality and the coherence of results by skipping each study in the analysis. Funnel plot is used to measure the biasness of publications. The two software packages (RevMan and Comprehensive meta–analysis) are used to provide for reliability and accuracy of results; the results produced are the same.
The analysis shows that the probability of frequency of recurrence of condylomatous deceases in the main group is 5.5 times lower than in the control group; the probability of frequency of TIP detection by PCR method in the main group is 6.7 times lower than in the control group; the probability of frequency of disappearance (essential decrease) of TIP replication by PCR method in the main group is 4.7 times higher than in the control group. It can be concluded that the conducted meta–analysis shows high effectiveness of the medical drug Proteflazid® for medical treatment of patients with pathologies of cervix and urogenital pathologies caused by papilloma virus infection. |
| first_indexed | 2024-12-10T17:53:16Z |
| format | Article |
| id | doaj.art-42e2e687f070489b99d2c42fab912d16 |
| institution | Directory Open Access Journal |
| issn | 2519-1853 2519-1861 |
| language | English |
| last_indexed | 2024-12-10T17:53:16Z |
| publishDate | 2015-12-01 |
| publisher | State Statistics Service of Ukraine, the National Academy of Statistics, Accounting and Audit (NASAA), the National Academy for Public Administration (NAPA) under the President of Ukraine |
| record_format | Article |
| series | Статистика України |
| spelling | doaj.art-42e2e687f070489b99d2c42fab912d162022-12-22T01:39:01ZengState Statistics Service of Ukraine, the National Academy of Statistics, Accounting and Audit (NASAA), the National Academy for Public Administration (NAPA) under the President of UkraineСтатистика України2519-18532519-18612015-12-014(71)364546Methodological Framework for Meta–Analysis: the Case of Clinical StudiesL. O. Yaschenko0M. V. Lyesnikova1Research Institute of Statistical StudiesResearch Institute of Statistical StudiesMeta–analysis is a statistical analysis summing up data from several analogous studies. Meta–analysis may combine, mathematically, two or more studies aimed at testing the same hypothesis. Its advantages include the increased statistical capacity of a study, entailing a higher accuracy of the measured effect of an analyzed parameter, and its cost–effectiveness. The purpose of the study is to outline the meta–analysis framework and demonstrate its implementation by use of data from clinical research: use of the medical drug Proteflazid® in medical treatment of papillomavirus infection. The meta–analysis covered 8 studies “occurrence – control”. In general, 539 patients with pathologies of cervix and urogenital pathologies caused by papillomavirus infection, and 498 patients of the control group were studied. The meta–analysis is conducted on the basis of the three indicators: Frequency of recurrences of condylomatous deceases. Frequency of detection of human papilloma (F1P) using the method of polymerase chain reaction (PCR). Frequency of disappearance (essential decrease) of F1P replication using PCR method. Z–criterion of Fisher is used to measure statistical significance of the combined ratio of chances. The heterogeneity of the set of data from a separate study is measured using the test on heterogeneity The criterion .r–squared is used to evaluate the heterogeneity between the studies. P < 0.05 indicates on a statistically significant heterogeneity. Also, heterogeneity effects are measured by F–test, which is the ratio of the variability of the studies more favorable to the heterogeneity than to the chances. When a result is significant, .r–squared of the criterion (P <0.05) or the result of F–test is higher than 50%, which indicates on the existence of a heterogeneity between the studies, and for this case the model of random effects (DerSimonian–Lerd method) is chosen. Otherwise, the model of fixed effects (Mantel – Flaenszel model) is used. Also, the sensitivity analysis is made, to measure the quality and the coherence of results by skipping each study in the analysis. Funnel plot is used to measure the biasness of publications. The two software packages (RevMan and Comprehensive meta–analysis) are used to provide for reliability and accuracy of results; the results produced are the same. The analysis shows that the probability of frequency of recurrence of condylomatous deceases in the main group is 5.5 times lower than in the control group; the probability of frequency of TIP detection by PCR method in the main group is 6.7 times lower than in the control group; the probability of frequency of disappearance (essential decrease) of TIP replication by PCR method in the main group is 4.7 times higher than in the control group. It can be concluded that the conducted meta–analysis shows high effectiveness of the medical drug Proteflazid® for medical treatment of patients with pathologies of cervix and urogenital pathologies caused by papilloma virus infection.https://su-journal.com.ua/index.php/journal/article/view/46meta–analysis, ratio of chances, heterogeneity, model of fixed effects, model of random effects, forest plot, funnel plot |
| spellingShingle | L. O. Yaschenko M. V. Lyesnikova Methodological Framework for Meta–Analysis: the Case of Clinical Studies Статистика України meta–analysis, ratio of chances, heterogeneity, model of fixed effects, model of random effects, forest plot, funnel plot |
| title | Methodological Framework for Meta–Analysis: the Case of Clinical Studies |
| title_full | Methodological Framework for Meta–Analysis: the Case of Clinical Studies |
| title_fullStr | Methodological Framework for Meta–Analysis: the Case of Clinical Studies |
| title_full_unstemmed | Methodological Framework for Meta–Analysis: the Case of Clinical Studies |
| title_short | Methodological Framework for Meta–Analysis: the Case of Clinical Studies |
| title_sort | methodological framework for meta analysis the case of clinical studies |
| topic | meta–analysis, ratio of chances, heterogeneity, model of fixed effects, model of random effects, forest plot, funnel plot |
| url | https://su-journal.com.ua/index.php/journal/article/view/46 |
| work_keys_str_mv | AT loyaschenko methodologicalframeworkformetaanalysisthecaseofclinicalstudies AT mvlyesnikova methodologicalframeworkformetaanalysisthecaseofclinicalstudies |