Critical Evaluation of Ayurveda-Based Articles - Essay Example

Critical Review of Three Ayurveda-based Articles The following three articles were critically evaluated. Each article is reviewed separately and all quotes and citations in a given article review refer only to the paper under review and as such are not individually cited. Article 1 - Park, J. & Ernst, E. 2005, "Ayurvedic medicine for rheumatoid arthritis: A systematic review", Seminars in Arthritis and Rheumatism, vol. 34, pp. 705-713. Article 2 - Thabrew, M., Senaratna, L., Samarawickrema, N., & Munasinghe, 2001, "Antioxidant potential of two polyherbal preparations used in Ayurvda for the treatment of rheumatoid arthritis", Journal of Ethnopharmacology, vol. 76, pp. 285-291 Article 3 - Govindarajan, R., Vijayakumar, P., & Pushpangadan, P., 2005, "Antioxidant approach to disease management and the role of 'Rasayana' herbs of Ayurveda", Journal of Ethonopharmacology, vol. 99, pp. 165-178. Article 1 Park and Ernst's (2005) "Ayurvedic Medicine for Rheumatoid Arthritis: A Systematic Review," though a qualitative study by design, has many of the scientific rigors associated with quantitative studies. These two research methodologies, quantitative and qualitative, generally polarize along the objective (quantitative) and subjective (qualitative) continuum, yet Park and Ernst's study has aspects of both. They summarize their study as "a systematic review of all randomized controlled trials (RCTs) on the effectiveness of Ayurvedic medicine for RA." Thus, it is qualitative in nature. However, their research design and meticulous literature search are consistent with quantitative studies - structured, repeatable, well documented and theory based. The integrity of a literature study depends on the integrity of trial selection. Biased trial selection can bias the study. The authors took noteworthy precautions to ensure the study's integrity. They found only seven RCTs meeting their inclusion criteria: Randomized and controlled trials, testing any form of Ayurvedic medicine, using no other therapies other than herbal medicines (no single-herb studies) Trials performed on RA patients (not osteoarthritis alone) After their comprehensive search, the authors indicate they identified 33 trials. This is a satisfactory sample size, but their review revealed trials with methodological concerns, particularly non-randomization, observational structures, or lacking controls. Once these trials were excluded, they had a sample size of seven. The seven selected, meeting the strict inclusion criteria, were methodologically sound. They incorporated important tenets of scientific research, being systematic, controlled, and empirical. The authors detailed their team's search strategy including: Electronically searching major databases (Medline, Embase, AMED, Cochrane Controlled Trial Register, and the abstract service of Central Council for Research in Ayurveda and Siddha [CCRAS]) Hand searching 4 relevant Sri Lankan and Indian journals not in the electronic databases Electronically and hand searching the authors' personal files Further, they detailed their keyword selection for the searches and indicated imposing no language restrictions. Documenting that trial selection was predicated upon documented keywords reduced the potential for both overt and covert bias, enhancing objectivity while minimizing subjectivity. Documenting that trial selection was performed scientifically and not according to the authors' personal choices allows the reader to better interpret the information as they see fit, rather than reading a study that reflects the authors' opinions. To further educate readers, Park and Ernst provide a comprehensive paragraph on how they scored and reviewed the articles. In addition to ensuring that each article was reviewed by a person speaking the language of the trial article, they noted that they followed the QUOROM guideline for systematic reviews and used a standardized score (Jadad) for assessing methodological quality. The authors document scientifically why they believed the study of value and provide a concise definition of Ayurvedic medicine. They do not identify themselves, their institutions, background, or other "personal" information in the study. As with most scholarly publications, the authors' university affiliations were disclosed on the study's front page, and the Acknowledgements section documents persons providing technical and editorial support. Including personal or demographic information in the body of the article could undermine article neutrality. The authors wrote in the first-person ("we") in the Discussion. They did so when commenting on their research in terms of methodology and limitations, not in terms of outcome or content analysis. Their closing line, "In our opinion, the study of Ayurvedic medicine is sufficiently important to merit more rigorous research" was personal. Given the lengths to which they went to perform empirical, theory-based research, closing with a personal opinion statement weakens study integrity. With the authors stating their bias, the neutrality of the research is in question. A statement such as, "Given the long tradition of Ayurvedic medicine and successful treatments to date, Ayurvedic medicine warrants continued rigorous research," could have expressed the same sentiment, though grounded in science, not opinion. Certain issues generally associated with research studies were not relevant in this study. As a literature review with no human subjects ethical issues were reduced. Still, the authors had to monitor the trials to ensure that no patient or identifying data was published or re-published. In the paper under review, no ethical issues were identified. The authors discussed generalizability, noting, "Generalization based only on the currently available material is problematic" devoting two paragraphs to study limitations: Definitions - interpretation and definition of words such as "random" and "controlled" by other researchers Inclusion - despite the rigorous search, inevitably articles were missed Publication bias - as negative studies rarely get published, there could be a "false-positive overall impression" of Ayurvedic medicine's effectiveness Quality of the trials studied - some trials were methodologically weak but met the inclusion criteria These issues are expected in scientific research. Responsible researchers recognize limitations in their study, noting them in their write-up for mitigation by future researchers. Park and Ernst' study is a sound example of a qualitative, exploratory study. Though reviewing only seven studies, the authors document their selection criteria, so the seven articles provide a scientific and methodologically sound base. Relevant assessment criteria were presented and the content was presented in a logical and readable manner. Unfortunately, concluding this well-written, well-researched, semi-quantitative yet still-qualitative paper on a note of personal opinion, not evidence-based research, diminished its potency. Article 2 Thabrew et al.'s "Antioxidant potential of two polyherbal preparations used in Ayurveda for the treatment of rheumatoid arthritis" is a quantitative study, objective in design, but design flaws render it weak as a true explanatory study. Integral to explanatory studies is the presence of a hypothesis. Thabrew et al. write, "an investigation has been carried out to determine whether the beneficial effects of WC [Weldehi choornaya] and MRQ [Maharasnadi quathar] in RA [Rheumatoid Arthritis] are mediated at least in part, through alterations in the activities of their antioxidant effects." However, they state no formal hypothesis, or proposal to be tested, such as "MRQ is superior to WC in altering the antioxidant effects resulting in an improved quality of life for patients with RA" or "MRQ and WC, as herbal preparations that mediate alterations in antioxidant effects, will provide a greater increase in the activities of SOD, GPX, and catalase than patients taking neither preparation." The study has many elements of an exploratory process, undertaking a study to "see the results" rather than stating at the outset the expected results to be upheld or disproved. Given their comment, "Reasons for the MRQ mediated increase in plasma antioxidant enzyme activities is not clear" it is not surprising the authors did not enter the study with a working hypothesis of the superiority of one particular polyherbal. Providing a testable hypothesis, however, could help structure the study; the authors could have hypothesized that the two polyherbals would have increased antioxidant enzyme concentrations as compared to patients with no treatment, and formulated a secondary hypothesis based on the composition of the polyherbals or other factors about which would be more efficacious. Experimental and control group selection was based on convenience sampling, not random sampling, as they selected only patients with RA. Sample size was a significant problem with the study. The authors noted beginning with 50 patients in each group (WC and MRQ); however, they only collected blood samples from 35 healthy patients, the control group. At the conclusion of the study, 5 MRQ patients had dropped out, but 30 WC patients had done so. The authors comment that more WC patients may have dropped out because of the lesser efficacy of WC, as demonstrated in the results, as compared to MRQ but they do not publish data to inform readers of same. The authors do not state the critical data on trends or adverse events during the treatment phase for the WC patients leading to the high attrition rate. Such events need to be monitored, with an attrition rate greater than 50% one questions whether the study should have been aborted and the WC patients offered the MRQ, apparently a superior polyherbal. The authors did not indicate whether the WC patients were offered the MRQ at study conclusion - a standard practice when comparing medications or treatments when one presents as clearly superior to the other. The high attrition rate makes outcome comparison difficult. Twice as many MRQ patients (45) in the final study as compared to the WC patients (20) renders statistical analyses difficult. The authors' choice of measures (mean/standard deviation and a one-way ANOVA) are inadequate to control for sample size differentials or weighting issues. Minimally, they needed to use a geometric mean, not arithmetic mean, to control for outliers and sampling issues. They indicate that with ANOVA, their confidence levels - representing statistical power -- were p Read More