Advertisement
  • Loading metrics

Methodological Bias Can Lead the Cochrane Collaboration to Irrelevance in Public Health Decision-Making

  • Antonio Montresor ,

    montresora@who.int

    Affiliation: Control of Neglected Tropical Diseases, World Health Organization, Geneva, Switzerland

  • David Addiss,

    Affiliation: Children Without Worms, Task Force for Global Health, Decatur, Georgia, United States of America

  • Marco Albonico,

    Affiliation: WHO Collaborating Centre for Neglected Tropical Diseases, Public Health Laboratory Ivo de Carneri, Zanzibar, United Republic of Tanzania

  • Said Mohammed Ali,

    Affiliation: WHO Collaborating Centre for Neglected Tropical Diseases, Public Health Laboratory Ivo de Carneri, Zanzibar, United Republic of Tanzania

  • Steven K. Ault,

    Affiliation: Neglected Infectious Diseases, Pan American Health Organization, World Health Regional Office for the Americas, Washington, DC, United States of America

  • Albis-Francesco Gabrielli,

    Affiliation: Neglected Tropical Diseases, World Health Organization Regional Office for the Eastern Mediterranean, Cairo, Egypt

  • Amadou Garba,

    Affiliation: Control of Neglected Tropical Diseases, World Health Organization, Geneva, Switzerland

  • Elkhan Gasimov,

    Affiliation: Malaria, Other Vectorborne and Parasitic Diseases, World Health Organization Regional Office for Europe, Copenhagen, Denmark

  • Theresa Gyorkos,

    Affiliation: WHO Collaborating Centre for Research and Training in Parasite Epidemiology and Control, Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, Quebec, Canada

  • Mohamed Ahmed Jamsheed,

    Affiliation: Vector-Borne and Neglected Tropical Diseases Control, World Health Organization Regional Office for South-East Asia, New Delhi, India

  • Bruno Levecke,

    Affiliation: WHO Collaborating Centre for the Monitoring of Anthelminthic Drug Efficacy for Soil-Transmitted Helminthiasis, Department of Virology, Parasitology and Immunology, University of Ghent, Ghent, Belgium

  • Pamela Mbabazi,

    Affiliation: Control of Neglected Tropical Diseases, World Health Organization, Geneva, Switzerland

  • Denise Mupfasoni,

    Affiliation: Control of Neglected Tropical Diseases, World Health Organization, Geneva, Switzerland

  • Lorenzo Savioli,

    Affiliation: Global Schistosomiasis Alliance and Public Health Laboratory Ivo de Carneri, Zanzibar, United Republic of Tanzania

  • Jozef Vercruysse,

    Affiliation: WHO Collaborating Centre for the Monitoring of Anthelminthic Drug Efficacy for Soil-Transmitted Helminthiasis, Department of Virology, Parasitology and Immunology, University of Ghent, Ghent, Belgium

  •  [ ... ],
  • Aya Yajima

    Affiliation: Malaria, Other Vectorborne and Parasitic Diseases, World Health Organization, Regional Office for Western Pacific, Manila, Philippines

  • [ view all ]
  • [ view less ]

Methodological Bias Can Lead the Cochrane Collaboration to Irrelevance in Public Health Decision-Making

  • Antonio Montresor, 
  • David Addiss, 
  • Marco Albonico, 
  • Said Mohammed Ali, 
  • Steven K. Ault, 
  • Albis-Francesco Gabrielli, 
  • Amadou Garba, 
  • Elkhan Gasimov, 
  • Theresa Gyorkos, 
  • Mohamed Ahmed Jamsheed
PLOS
x

Aims and Limits of the Cochrane Collaboration

The Cochrane Collaboration (CC) was created in 1993 with the aim of systematically reviewing published research in order to facilitate the selection of appropriate interventions by health professionals and policy-makers [1]. CC systematic reviews focus on a wide range of health-care interventions and typically consider evidence only from randomized controlled trials (RCTs). Because they rely on chance to minimize the potential for epidemiological confounding, RCTs are commonly acknowledged as the strongest, least biased source of evidence on particular therapies or medical interventions for clinical practice. Similarly, it is well known that their utility in evaluating public health interventions is not always optimal [2] and, as a consequence, can result in distorted conclusions [3].

We specifically refer to a series of CC systematic reviews aimed at assessing the benefits of deworming for soil-transmitted helminthiases (STH) [46] in children, which we believe is affected by a significant methodological bias. Two essential characteristics of this intervention, and of the infections it targets, were not considered by the reviewers when they chose to restrict evidence to data derived only from RCTs.

Recovery from STH-Associated Morbidity Is a Long-Term Process

The World Health Organization (WHO) recommends a sustained program of mass deworming for preschool-age and school-age children in areas endemic for STH, corresponding to 10–12 years of treatment for each child [7]. The aim of such a programme is to keep STH infections to as low a level of intensity as possible in order to prevent and eliminate morbidity, thus protecting a child during his or her physical and cognitive development [7]. The “intervention” to evaluate is therefore not represented by one or two rounds of treatment but by the cumulative deworming experience extending throughout childhood. The fact that RCTs have considerably shorter follow-up times means they cannot capture the real effects of the deworming intervention, and conclusions drawn from this evidence risk being severely biased.

The distribution of worms among human hosts is not uniform; only a minority of individuals in a community will have infection at a level sufficiently high to cause morbidity (i.e., at moderate or high intensities of worm burdens) [8]. Estimates indicate that, where the prevalence of infection with Ascaris lumbricoides is 50%, approximately 20% of the children in the community will have moderate- or high-intensity infections, and therefore exhibit morbidity [9]. However, deworming tablets are administered to the entire child population living in an endemic area (because of public health considerations such as the high cost and logistical burden of test-and-treat approaches, the low sensitivity of field-applicable diagnostic techniques, the relative safety of the medicines, the limited health infrastructure and poor access to treatment, and the low health-seeking behavior, among others). Consequently, the deworming intervention will directly benefit only a portion of the treated children, and will obviously provide no benefit to children who are not infected. It is therefore unreasonable to evaluate the benefits of deworming among all the children who are treated, instead of only among those who are infected.

Not surprisingly, the CC review concludes that the intervention may improve weight gain only in children “known to have worm infection” [6]. A systematic review is hardly necessary to point out that children without worms do not directly benefit from the administration of a deworming tablet.

In conclusion, the main challenges of using RCTs to evaluate the benefits of deworming include: (1) the short evaluation time periods relative to the longer time needed to observe accrued benefits; and (2) the need to assess the outcome of the intervention by pooling together infected and uninfected children alike, thus diluting the known benefits.

The use of RCTs or quasi-RCTs for the evaluation of deworming interventions has already received considerable criticism [7,1015]. However, the most recent CC review on this topic [6] does not address these criticisms, but rather perseveres in its biased methodological approach, thus highlighting its considerable limitations. Not only is such a review of little value in guiding global deworming policy, it could also generate confusion among public health planners and practitioners in endemic countries, thus contributing to possible withdrawal from treatment of millions of children suffering from STH.

We are convinced of the need to properly evaluate deworming interventions which, despite their simplicity, carry significant logistic and cost burdens. The cost-benefits of such interventions need to be compared with other health interventions (such as vaccination, sanitation, and maternal and child health interventions). A proper evaluation should be organized.

Some Examples of Studies Demonstrating the Importance of Maintaining a Low Worm Burden

  • The amount of blood lost to heavy-intensity hookworm infection has been precisely measured [16,17]. Severely infected children lose more than double the daily iron requirement [17]. It is especially difficult for those with limited iron dietary input to compensate such a daily loss of iron [17,18].
  • In areas of high endemicity, women given albendazole had a lower rate of severe anemia during pregnancy [19]. Birth weight of infants of women who had received albendazole significantly improved, and infant mortality at 6 months fell dramatically [20].
  • Several hospital-based studies have documented and quantified an elevated mortality due to:
    • intestinal obstruction caused by heavy intensity infections with A. lumbricoides [21] (morbidity 12 million cases; mortality 10,000 cases); and
    • dysentery syndrome caused by heavy-intensity infections with Trichuris trichiura [22,23].
  • Evidence from veterinary studies in experimentally infected pigs has demonstrated damage to the gut surface—flattening or atrophy of villi—and the consequent maldigestion and malabsorption in moderate- and heavy-intensity ascariasis [24].

References

  1. 1. Hill GB. Archie Cochrane and his legacy. An internal challenge to physicians’ autonomy? J Clin Epidemiol. 2000;53(12):1189–92. pmid:11146263
  2. 2. Shelton JD. Evidence-based public health: not only whether it works, but how it can be made to work practicably at scale. Glob Health Sci Pract. 2014;3:253–8. doi: 10.9745/ghsp-d-14-00066
  3. 3. Smith GCS, Pell JP. Parachute use to prevent death and major trauma related to gravitational challenges: systematic review of randomized controlled trials. BMJ. 2003;327:1460–1. doi: 10.1136/bmj.327.7429.1459
  4. 4. Dickson R, Awasthi S, Demellweek C, Williamson P. WITHDRAWN: Anthelmintic drugs for treating worms in children: effects on growth and cognitive performance (Cochrane Review). Cochrane Database Syst Rev. 2007 Jul 18;(2):CD000371. doi: 10.1002/14651858.cd000371
  5. 5. Taylor-Robinson DC, Maayan N, Soares-Weiser K, Donegan S, Garner P. Deworming drugs for soil-transmitted intestinal worms in children: effects on nutritional indicators, haemoglobin and school performance. Cochrane Database Syst Rev. 2012; Cochrane Database Syst Rev. 2012;11:CD000371. doi: 10.1002/14651858.CD000371.pub5. pmid:23152203
  6. 6. Taylor-Robinson DC, Maayan N, Soares-Weiser K, Donegan S, Garner P (2015) Deworming drugs for soil-transmitted intestinal worms in children: effects on nutritional indicators, haemoglobin and school performance. Cochrane Database Syst Rev. 2015 Jul 23;7:CD000371. doi: 10.1002/14651858.CD000371.pub6. pmid:26202783
  7. 7. Eliminating soil-transmitted helminthiases as a public health problem in children: progress report 2001–2010 and strategic plan 2011–2020. Geneva: World Health Organization; 2012.
  8. 8. Gabrielli AF, Montresor A, Engels D, Savioli L. Preventive Chemotherapy in Human Helminthiasis: Theoretical and operational aspects. Transactions of the Royal Society of Tropical Medicine and Hygiene 2011:105: 683–693. doi: 10.1016/j.trstmh.2011.08.013. pmid:22040463
  9. 9. Montresor A, à Porta N, Albonico M, Gabrielli AF, Jankovic D, Fitzpatrick C et al. Soil-transmitted helminthiasis. Relationship between prevalence and classes of intensity of infection. Trans R Soc Trop Med Hyg. 2015;109:262–7. doi: 10.1093/trstmh/tru180. pmid:25404186
  10. 10. Bhargava A. Treatment for intestinal helminth infection. Conclusions should have been based on broader considerations. BMJ. 2000;321:1225.
  11. 11. Bundy D, Peto R. Treatment for intestinal helminth infection. Studies of short term treatment cannot assess long term benefits of regular treatment. BMJ. 2000;321:1225. pmid:11185586
  12. 12. Cooper E. Treatment for intestinal helminth infection. Message does not follow from systematic review’s findings. BMJ. 2000;321:1225–6.
  13. 13. Michael E. Treatment for intestinal helminth infection. Contrary to authors’ comments, meta-analysis supports global helminth control initiatives. BMJ. 2000;321:1224–5. pmid:11073529
  14. 14. Savioli L, Neira M, Albonico M, Beach MJ, Chwaya HM, Crompton DW et al. Treatment for intestinal helminth infection. Review needed to take account of all relevant evidence, not only effects on growth and cognitive performance. BMJ. 2000;321:1226–7. pmid:11185587
  15. 15. Montresor A, Gabrielli AF, Engels D, Daumerie D, Savioli L. Has the NTD community neglected evidence-based policy? [Expert commentary of the viewpoint by Nagpal S, Sinclair D, Garner P]. PLoS Negl Trop Dis. 2013;7:e2299. doi: 10.1371/journal.pntd.0002299. pmid:23875037
  16. 16. Hall A, Hewitt G, Tuffrey V, de Silva N. A review and meta-analysis of the impact of deworming of intestinal worms on child growth and nutrition. Matern Child Nutr. 2008;4:118–236. doi: 10.1111/j.1740-8709.2007.00127.x. pmid:18289159
  17. 17. Stoltzfus RJ, Albonico M, Chwaya HM, Savioli L, Tielsch J, Schulze K, Yip R. Hemoquant determination of hookworm-related blood loss and its role in iron deficiency in African children. Am J Trop Med Hyg. 1996;55:399–404. pmid:8916795
  18. 18. Jonker FAM, Calis JCJ, Phiri K, Brienen EAT, Khoffi H et al. Real-time PCR Demonstrates Ancylostoma duodenale Is a Key Factor in the Etiology of Severe Anemia and Iron Deficiency in Malawian Pre-school Children. PLoS Negl Trop Dis. 2012;6(3):e1555. doi: 10.1371/journal.pntd.0001555. pmid:22514750
  19. 19. Christian P, Khatry SK, West KP Jr. Antenatal anthelminthic treatment, birthweight, and infant survival in rural Nepal. Lancet. 2004;364(9438):981–3. pmid:15364190 doi: 10.1016/s0140-6736(04)17023-2
  20. 20. Joseph SA, Casapía M, Blouin B, Maheu-Giroux M, Rahme E, Gyorkos TW. Risk factors associated with malnutrition in one-year-old children living in the Peruvian Amazon. PLoS Negl Trop Dis. 2014;8(12):e3369. doi: 10.1371/journal.pntd.0003369. pmid:25503381
  21. 21. Silva De et al. Morbidity and mortality in intestinal obstruction caused by Ascaris lumbricoides. Trop Med Int Health. 1997;2(2):189–90. doi: 10.1046/j.1365-3156.1997.d01-241.x
  22. 22. Stephenson LS, Holland CV, Cooper ES. The public health significance of Trichuris trichiura. Parasitology. 2000;121 Suppl:S73–95. pmid:11386693 doi: 10.1017/s0031182000006867
  23. 23. Kaminsky RG, Castillo RV, Flores CA. Growth retardation and severe anemia in children with Trichuris dysenteric syndrome. Asian Pac J Trop Biomed. 2015;5:581–6. doi: 10.1016/j.apjtb.2015.05.005
  24. 24. Martin J, Crompton DW, Carrera E, Nesheim MC. Mucosal surface lesions in young protein-deficient pigs infected with Ascaris suum (Nematoda). Parasitology. 1984;88:333–340. pmid:6718059