Critique an Epidemiology Research Paper – A nutritional Disease of Childhood Associated with a Maize Diet

Critique an Epidemiology Research Paper – A nutritional Disease of Childhood Associated with a Maize Diet

 

            Williams (1933) presented in her epidemiological study a disease that lacks in name but could be described by its well-marked manifestations and symptoms among the children aged between one and four in the Gold Coast Colony. From about twenty cases witnesses from previous year, the researcher chose five cases to consider for the study. Williams liaised with the Research Institute at Korle Bu and the Medical and Health Services for the pathological findings of the study. The study discovered a consistent pattern of disease attacks among the children that are found out to be purely dietetic deficiency in nature leading to fatality. It was also determined that the cause of the disease involves the preparation of maize (corn) as arkasa and kenki as well as fu fu and garri in some instances. Arkasa refers to a watery maize gruel while kenki refers to a thicker maize paste, cooked several times. In this report, I will discuss the critical analysis of an epidemiological research in terms of its strong points and limitations. Study base, exposure assessment, outcome assessment, relations assessment, data collection, descriptive results, synthesis with other knowledge and findings and conclusions are addressed.

 

            Initially, there are a number of features of the study that is worth pointing out before conducting a detailed critical appraisal of the epidemiological research. First, the author provided a detailed account on the concept of lactation of Gold Coast Colony, drawing from the closely adhered into family customs. Second, the researcher also offered a detailed account of which she believed is the root of the deficiency (arkasa and kenki). Williams also noted how children could typically acquire the syndrome as exacerbated by the fact that these children lack in accessory substances. Third, the study presented a comprehensive discussion of the history of patients who suffer from the disease including its symptoms. Fourth, the cases presented are likewise in depth centering discussions of diagnosis prior and after death and treatments prescribed. Finally, Williams offered a holistic approach to prevention encompassing the role of authority, agriculturists and educationalists.

 

            Although this qualitative study is presented in a spontaneous format, the first part, which will be regarded as the abstract, identifies the subject and the setting of the study. Speaking of qualitative, although it is claimed by Yarnell (2007) that epidemiological studies are more effective using quantitative method, qualitative studies contribute to epidemiological literature through identifying the most appropriate variables that should be measure, increasing understanding of quantitatively assessing variables and explaining results particularly unexpected findings (pp. 21-22; Yarbro, Goodman and Frogge, 2005, p. 41). Children of either sex and between one and four year olds are the subject of the study. These are the patients presented to Children’s Hospital in Accra. Immediately, one realises that, in the process, the study could have excluded a large number of otherwise extreme or critical cases that might provide in-depth clinical significance to the study because it mainly relies on hospital admission. In a context, where communities and local people are more concern in livelihood, presentation of sick family members to hospitals might be less prioritised as it can incur cost and would eat time to do so. One should also take note of the particular location the study was conducted – urban or rural – because fundamental differences would appear considering hospital admission. Note, however, that the identified potential difference does not result in the lack of clarity when it comes to the definition of the subject.

 

            As opposed to randomisation, Williams selected the cases to consider for the study through an information-oriented sampling. As insightful as it is, the study illustrates critical to extreme cases. Yin (2002) defines a critical case as having strategic importance relating to the general problem while extreme case as complementary significance with relation to the problem. Cases 2, 3 and 5 could be considered as critical significance while Cases 1 and 4 are extreme cases. Three females and two males are tracked from the time they are presented to the hospitals and of these, three post-mortem investigations were conducted, on the morbid anatomy section. From these cases, nutritional deficiencies that are connected with maize preparation were of concern, and by which treatments were based. The studied cases were observed in the Children’s Hospital in year 1932, taking advantage of the hospital records and the presence of medical practitioners. Based on this, one can consider that the design of the study is rather weak due to the fact that cases are haphazardly chosen. According to Gail and Benichou (2000) and Silman and Macfarlane (2002, p. 7), epidemiological study designs are usually stronger is subject selection is guided by the need to make a valid comparison. One of the strong points of the epidemiological study is the consistency in the treatments prescribed: cod liver oil and malt with quinine and iron and mouthwash. A confounder, however, is the manner by which these treatments could help in delaying the onset of the disease and its indications or the method of delivery of treatment undertaken outside the hospital.

 

            Although Williams did not offer any conceptual definition, it is clear that the determinant of interest is how the preparation of arkasa and kenki could lead to the acquisition of the disease among the children. The author disclosed basic details on maize preparation and hinted that arkasa and kenki do not likely to prove an adequate diet particularly for the young people As the causal agent, informations on maize preparation could have included details on bacterial, virus and germs obtained in the preparation for the purpose of establishing what particular elements infect the children. One could figure that even though the hospitals might not have the most sophisticated medical facilities at that time; it could have been more plausible if the study involves pathological informations. Nevertheless, an extraneous determinant was identified as breastfeeding, and this makes this more difficult for the researcher to establish the relations between the determinant of interest and the outcomes of the study (Grobbe and Hoes, 2007). Though the study was able to make generalisations that will move from time and place-specific observations, the medical significance of the study could be likewise established in the process (Rohman and Greenland, 1998). Anyhow, the researcher noted that it is impossible to conduct a scientific investigation on the cause or to make controlled experiments on the nature of the cure with reasons undisclosed.

 

            As already mentioned, the cases are chosen from the twenty cases occurred in 1932 hence there is no definitive time frame of exposure assessment. Although there is also no apparent operational definition of exposure, the research tendered symptoms-based descriptions. The author drew her conclusions from the history of the patient suffering from the disease and the extent to which the symptoms are evident. It is intuitive to think that it would be difficult to determine the consistency of the time frame with the expected induction time especially that the dietetic treatments proved to be beneficial to the patient, which implicates the delay in the appearance of the disease. The consistency of implementing the treatment is an evident confounder as well. One could think that the willingness of the parent or guardians to zealously implement treatment is beyond the control of the researcher. Informations about the exposure are also incoherent possibly because informants embrace cultural stance when it comes to disclosures, which is also beyond the control of the researcher. Such dilemmas were addressed through conducting pre and post-mortem analysis.

 

            Etiologically, there are a number of elements that contribute in the formation of the syndrome such as the deficient lactation process, diet that is mainly concentrated on maize, amino acid and protein deficiency and vitamin deficiency. Such outcomes results from the qualitative analysis of the cases probably via a comparative contextual analysis of analysing the similarity and differences among the cases considered (Grobbe and Hoes, 2007). There is no concrete measure of outcomes used but the disease frequency was measured on an incidence basis. At risk children to the disease was not determined as well. As such, the research is basically a causal study that builds on the strength of component causes on the prevalence of the disease, complementary component causes and the interplay between the two. In simpler terms, the focus of the study is the nutritional deficiency which is said to be associated with arkasa and kenki. Such research further considers the role of breastfeeding and [lack thereof] of accessory substances.

 

            One could only suppose that the measure of effect utilised is through distinguishing the exposed group and the case fatality. Other than that it is difficult to solidly identify the measure used because as already noted the research is purely qualitative. Because the researcher assessed that if the child infected will not be treated immediately, s/he will die and symptoms will be severe as s/he approaches death. Apart from the diet full of accessory substances, various drugs were used as part of the treatment. Such situation implies an unbending relationship of induction with exposure and hence outcomes. Effect modification of the syndrome was considered by means of defective lactation hence the question: does the effect of maize preparation on acquiring the syndrome depend whether or not to breastfeeding processes?

 

            There is no particular data collection method specified but because it involves case studies one can assume that the researcher made use of informant interview, structured observation and hospital document analysis. Informations are collated from the mothers, if possible, and grandmothers of the children. While some won’t disclose necessary informations about the feeding, it was established however that arkasa and kenki are consistent in the diet of the children. Details about the symptoms are collated through medical observations and examinations all taken within the hospital particular in terms of the degree of depigmentation, number and extent of patches and irritability. There are no informations tendered on the participation rate, response rate or follow-up rate, as this epidemiology is case-based (Grobbe and Hoes, 2007; Sills, 1986, p. 99).

 

As such, there was an obvious systematic bias in both information and selection. This is because there are errors that had been introduced during data collection such as inconsistency in details that had been collected. According to Ahrens and Pigeot (2005), errors on data collection cannot be corrected later on and since there are no individual measurement values introduced in the study, there is no way that the bias could be adjusted (Rothman, Greenland and Lash, 2008). We should take note that epidemiological research is basically about collecting data related to health with the description of health and welfare in populations and with the goal of improving health outcomes. Although the researcher observably concerned the latter, potentially identifiable data collection was existent (McNeil, 1996). Considering again that the study was conducted several decades ago, it is pointed here that ethical stance are considered particularly in stating what data are required and how the data will be collected, used and protected (Van Zilj, 2004).

 

The researcher offered no discussion on the measures of effect or association. Thomas and Weber (2001) made mention that measure of effect and association are very important in epidemiological studies as the goal of such study was to evaluate interventions and determine risks factors. Risk differences and risk ratios are formed from the measures and so an adequate measure should be established. A measure that could have been used would be the attributable risk as there is a risk of developing the disease even if the children are exposed or unexposed (Toft, Agger and Bruun, 2004; Porta, 2008, p. 216). Since arkasa and kenki are considered staple food, one could presume that every child could be exposed specifically those whose diet is limited to maize diet. Further, while Williams do not offer public health impact assessment, she presented a possible prevention which centers the encouragement of incorporating milk in the daily diet of the children of Gold Coast. Of concern is she should also consider the benefits of proper breastfeeding practices in order to make recommended more convincing especially that she consider improper lactation as a factor that worsens the condition of the children.

 

One could not established if there had been selection bias because although the cases are selected on the basis of non-confusion by any accompanying disease the cases considered are selected from a group exhibiting the same symptoms. There are specific confounders already identified which could have impacted the effects. Information bias is evident (Grobbee and Hoes, 2007; Abramson and Abramson, 1999) as well as the research generally rely on the cooperation of the mothers, grandmothers and aunts of the children. Despite this, the validity of the study was vindicated by the future studies. The results of the findings of Williams are consistent with the findings of Brock and Autret (1952). The only difference is that the syndrome/disease that Williams identified was given the name of kwarshiorkor. The findings of the latter pointed that the disease was caused by the lack of protein intake. After the weaning period, the study also noted that gruels of various kinds are given to the children resulting to the development of kwashiorkor a short time after the breastfeeding had ceased.    

           

            Further, the conclusion presented by the researcher with reference to the presented evidence is justified. This is because the researcher scrutinised the cases and associated it to maize preparation, breastfeeding and the development of the disease. Greenberg et al (2005) state that the concept of validity concerns the degree to which a measurement or study reaches a correct conclusion. Evidently, she was also able to provide the etiological factors that contributes to the formation of the disease. Because there is no specific aim or objectives stipulated, it would be therefore difficult is these conclusions responded to the objectives of the research, however. The findings are presented clearly in such a way that it enables a reader to judge them for him/herself. Aside from the mild inconsistency on the details of exposure, the findings are objectively presented. Savitz (2003) said that the implications of findings that is separate from the scientific merits of the study create incentive to reach the conclusion hence Williams’ study did not left readers in an ambiguous state. Though statistical analysis would have made the study more clinically significant, the data collected are worthy and the analysis is adequately performed. As such, Williams successfully answers whether maize preparation is responsible for the development of the disease.     

 

 

 

References

 

Abramson, J H and Abramson, Z H 1999, Survey Methods in Community Medicine: Epidemiological Research, Programme Evaluation, Clinical Trials, Elsevier Health Sciences.

 

Ahrens, W and Pigeot, I 2005, Handbook of Epidemiology, Birkhauser.

 

Brock, J F and Autret, M 1952, ‘Kwashiorkor in Africa,’ Bulletin of the World Health Organization (WHO), vol. 5, pp. 1-71.

 

Gail, M H and Benichou, J 2000, Encyclopedia of Epidemiologic Methods, John Wiley & Sons, New York.

 

Greenberg, R S, Daniels, S R, Flanders, D W, Eley, J W and Boring, J R 2004, Medical Epidemiology, Lange Medical Books/McGraw Hill Professional New York.

 

Grobbee, D E and Hoes, A W 2007, Clinical Epidemiology: Principles, Methods, and Applications for Clinical Research, Jones & Bartlett Publishers.

 

McNeil, D 1996, Epidemiological Research Methods, John Wiley and Sons, New York.

 

Porta, M 2008, Dictionary of Epidemiology, Oxford University Press.

 

Rohman, K J and Greenland, S 1998, Modern Epidemiology, 2nd edn, Lippincott, Philadelphia.

 

Rothman, K J 2005, ‘Causation and Causal Interference in Epidemiology,’ American Journal of Public Health, vol. 95, no. S1, pp. 1-7.

 

Rothman, K J, Greenland, S and Lash, T L 2008, Modern Epidemiology, Lippincott Williams and Wilkins.

 

Savitz, D A 2003, Interpreting Epidemiological Evidence: Strategies for Study Design and Analysis, Oxford University Press, US.

 

Sills, D L 1986, International encyclopedia of the social sciences, Macmillan.

 

Siman, A J and Macfarlane, G J 2002, Epidemiological studies: a practical guide, Cambridge University Press.

 

Thomas, J C and Weber, D J 2001, Epidemiologic methods for the study of infectious diseases, Oxford University Press, US.

 

Toft, N, Agger, J F and Bruun, J 2004, ‘Measures of association and effect,’ Biofolia.

 

Van Zilj, S 2004, Ethics in health Research: Principles, Structures and Processes.

 

Yarbro, C H, Goodman, M and Frogge, M H 2005, Nursing: Principles and Practice, Jones & Bartlett Publishers.

 

Yarnell, J 2007, Epidemiology and prevention: a systems-based approach, Oxford University Press, London.

 

Yin, R K 2002, Case Study Research Design and Methods, 3rd edn., Applied social research method series, vol. 5, Sage Publications, California.

 

Credit:ivythesis.typepad.com