Running head: A Review of Epidemiology and Causation Concepts

The concept of causation in epidemiology has been a difficult enterprise to students and researchers seeking to master it both in terms of evidence, and its causal relationships to diseases, or their legitimate causes, taking in consideration epidemiology as a scientific field (Parascandola &Weed, 2001; Olsen, 2003;Karhausen, 2000; Dumas et al.,2013).

The idea that epidemiology is at the heart of observational, descriptive and scientific studies seems to add an important argument to the core issue that causation is a practical tool capable of enhancing the analysis of deterministic and probabilistic values or considerations (Dumas et al.,2013; Parascandola &Weed, 2001). However, the scope of definition of causation is very wide. The goal of this paper is to use the background review of several peer reviewed articles on the definitions of causation, then compare the strengths and weaknesses of each definitions.The student/researcher will attempt to offer a limited criticism of the causation concept based on the literature, and an exploratory research article on gynecological health applying ethnographic methods to public health inquiry. Such article seems to present the limits of causation in epidemiology.

Parascandola and Weed (2001) offer five categories of causations resulting from an

intensive analysis of the literature. The authors posit the followings: production, necessary and sufficient, sufficient-component, counterfactual, and probabilistic causations. Various theories of causation of disease consist the backbone for further analysis (Karhausen, 2000; Parascandola &Weed, 2001):

1) Supernatural theory. Posits that disease is of supernatural origin, or is due to a curse from God.

2) Contagion theory (miasma). According to this theory, diseases are transmitted by a miasma or cloud that clung low on the earth surface.

3) Germ theory (Koch, 1880; Pasteur, 1866). Koch postulates that an understanding of

the causes of disease is important in the health field not only for prevention but also in

diagnosis, and the application of treatment. According to the science enhanced scope and sequence biology course and the Virginia Department of Education 2012, the core of the postulates contends that:

a) a specific organism ought to always be observed in association with the disease.

b) the organism is to be isolated from a host causing infection and grown in pure culture

in a laboratory.

c) upon inoculation of the organism from a pure culture into a susceptible host

organism, it must cause the disease.

d) the infectious organism will be subjected to re-isolation from the diseased causing

organism and grown in pure culture.

4) Epidemiological triad theory. This theory deals with: agent  which consists of  biological, chemical, physical, nutritional, and social, host factor such as age, sex, heredity, nutrition, occupation, custom, habits, immunity power, biological-blood sugar, cholesterol, housing, marital status, and socio-economic status. The environmental factor considers the physical, biological, and psychosocial environments.

5) BEINGS model. This model considers that biological, environmental factors, immunological factors, nutritional factors, genetic factors, and social, spiritual, and service factors are vital ingredients in preventing disease occurrence.

6) Web of causation. It provides a lens of multiple factorial causation of disease; mostly non-communicable disease. An example is factors causing ischemic heart disease such as

hypercholesterolemia, hypertension, obesity, tobacco use, physical inactivity, genetic predisposition, age, sex, and diabetes (Le, Bhushan, and Sochat, 2015).

7) The theory of necessary and sufficient cause advances that the cause of a particular disease is a constellation of various factors; and all of them come into play in an optimum combination.

Definitions

An understanding of the causes of disease is important in the health field not only for prevention but also in diagnosis and the application of treatment, the researcher/writer asserts. As maintained by numerous researchers, the cause of a disease is an event, condition, characteristic, or combination of these factors which plays an important role in producing the disease (Gordis, 2000; MacMahon & Pugh, 1970).

Production

According to Parascandola and Weed (2001, p.906), production results from causes or conditions that play integral parts in disease occurrences.

Necessary and sufficient cause

A necessary- and sufficient cause is a sine qua non condition for an effect to occur. Parascandola and Weed (2001, p.906) retained that “four different types of causal relations can be derived from these two definitions: necessary and sufficient, necessary but not sufficient, sufficient but not necessary, and neither necessary nor sufficient”.

Sufficient-component causes

A sufficient-component cause guarantees the occurrence of an effect. Thus, when the cause is not absent, the effect must be evident. All, not a single components play a key role in determining the validity of causation. Component causes foster togetherness in order to produce effect (Olsen, 2003). Others such as Dumas et al. (2013) advance that component cause models are tied with multicausal phenomenon generated due to chronic multifactorial diseases.

Probabilistic cause

The key factor is that an increase in probabilistic cause leads to the probability of the occurrence of an effect.

Counterfactual causes

The counterfactual approach describes causality by comparing observed events and the counterfactual events or the event that would have been observed. In both, the probabilistic cause and counterfactual causes models, statistical analysis are paramount in order to test the limits of traditional scientific inquiries, and causality of questions, the writer/researcher contends.

Strengths and Weaknesses

The literature surveyed on causation (Parascandola and Weed, 2001; Karhausen, 2000; Maclure and Schneeweiss, 2001) offers a very interesting lens of inquiry and tests the limits of causation. One key difference retained by the writer/researcher is that the definitions proposed by Parascandola and Weed (2001) lie in the deterministic and probabilistic continuum in assessing the concept as a scientific tool for epidemiology as a science. Whether the debate has been on the ontological or epistemological nature of the issues raised: causation and effect, one argument seems to the difficult nature of a sufficient-component cause definition, and the probabilistic causation offer different models of analysis, the writer /researcher posits. For instance, is it possible to have a sufficient model full of its effects? A probabilistic model showing statistical causation as a viable scientific tool, and an inclusive approach has been provided as a good option, while the necessary and sufficient definition, deterministic in nature, lacks a measurement effect (Parascandola and Weed, 2001; Karhausen, 2000; Maclure and Schneeweiss, 2001), Nonetheless, the weakness in the probabilistic model, based on the literature surveyed has not been fully assessed or presented.

In terms of offering a critical platform for understanding, Maclure and Schneeweiss (2001), through their episcope propose a very clever and critical lens of analysis of causation by looking at the causation of bias in epidemiology. In a nutshell, the model presents causal diagrams and graphs to apply observational data or phenomena through a network of information systems to analyze morbidity within a population. Through an episcope, evidence that is observed through the “lenses” and “filters” are classified as directions of causation, random confounding, correlated causes, diagnosis, encoding of exposure, data, grouping through the generalization of hypotheses, inference and  publication bias(journal article) in order to tract a risk difference(RD) in the “domain of agency” and  the “ domain of knowledge”(p.115).

It is important to note that the causal effects in analyzing the ways of causation have bias that appear to be inherent in the model when looking at factors such as agent leading to morbidity and background causes in a population under study, the writer/researcher adds. Whether one needs to look at the many cofounders laid out by this model or determinants of diagnosis, evidence of exposure, grouping criteria, it seems implausible to quantitatively theorize about the many associations presented through the diagrams and graphs; and there is a missing link for descriptive epidemiology, the writer/ researcher argues.

The case of the sociocultural context of gynecological health among Haitian immigrant women in Florida: applying ethnographic methods to public health inquiry.

The above definitions of causation ,and the role of epidemiology as a scientific domain linking to quantitative measures offers a limited view on determinants of diseases. The impact on disease prevention is to be fully addressed and assessed through various research methods: qualitative and mixed-methods designs, the writer/researcher claims. A case of reference is the exploratory research on a feminine practice among a culturally important group: “Twalèt déba” (Feminine hygiene relating to vagina cleanliness), which may have important significance for transmission of disease, namely human papillomavirus [HPV] (Menard et al., 2010), the main cause of cervical cancer (Le, Bhushan, and Sochat, 2015).

Foster (1976) elucidates two key domains in his classic system of etiologic classification in terms of understanding women’s beliefs about disease causation. They are: personal causes of illness: personalistic and naturalistic. This exploratory research consisted in observation, participant observation, and semi-structured in-depth-interviews of 35 Haitian immigrant women participants about a feminine hygiene practice in Little Haiti, Miami, Florida termed Twalet déba (Feminine hygiene relating to vagina cleanliness).The feminine hygiene of these Haitian women in Little Haiti, Miami, Florida, is culturally construed, and the basis for sexual health and virility has not been fully researched. The cause/causative agents based on the etiologic beliefs about cervical cancer, and the commonly named feminine hygiene agents for this population have been depicted in the tables 1and 2 below where causative agents of cervical cancer, and commonly

named feminine hygiene products are summarized.

Conclusion

The knowledge of causation is an integral part of epidemiology as it enables the researcher to make the proper diagnosis, formulate the correct treatment plan and take necessary measures in the prevention of a certain disease. But such knowledge can be limited due to the type of causal relationships one seeks to analyze, research designs pursued and implemented. The various definitions and the role of the episcope offering a constructive approach are good, debatable and difficult aggregates in the web of epidemiological research as a viable scientific field.

This final paper is about synthesizing three interesting research articles dealing with important health issues whose tentacles are deeply rooted in ethnicity and health practices. The researcher/writer saw in the process of reviewing them, an avenue for the concept of ethnic health. The ramifications are beyond the scope of the required assignment. The proposed synthesis starts with the manner in which the four concepts of internal validity, external validity, conclusion validity and construct validity validity are assessed.

The requirements for sound analysis in the considered researches are serious benchmarks that the researcher/writer hopes to follow throughout his graduate experience. It is a noble endeavor to dwell in the research inquiry web, then expect to find a way out through perfection on the dissertation continuum. There are important considerations found in the researched articles surveyed during this semester that will stick to the mind.

Three articles, by Rahill & Rice (2010), Malow et al. (2013), and Huffman et al. (2013) consist the bulk of the intended analysis. The researcher started an inquiry at the onset of this semester about health beliefs and cultural factors affecting the promotion of prevention of diseases, and was unfortunate to briefly find the same concepts were indirectly considered in two of the research articles. The followings serve as a platform for analysis and further inquiries into reliability.

1- Correlates of picuriste use in a sample of health-seeking Haitian immigrants and adult children of immigrants in Miami-Dade County, Florida. Rahill & Rice (2010).

Rahill and Rice’s ( 2010) article, “ Correlates of picuriste  use in a sample of health-seeking Haitian immigrants and adult children of immigrants in Miami- Dade county, Florida”, presents a  considerable look at a health practice from Haitians in Miami-Dade County, Florida, whose origin is deeply rooted in old health practices in Haiti, namely picuriste injections. A picuriste is a traditional health worker/person with no formal medical training providing intramuscular, subcutaneous, and intravenous injections. The impact of such practices on transmission of diseases such as hepatitis B, hepatitis C, tuberculosis, and malaria, has not been  fully studied, except by few researchers such as (Rahill, 2007; Pape, 1986; Clerisme, 2005).

The goal of this mixed-methods research was to explore covariates of the use of picuristes (traditional health workers with no formal medical training providing intramuscular, subcutaneous, and intravenous injections) in the Haitian community of Miami-Dade County, Florida. Rahill and Rice (2010) contend that although the picuriste practices have been researched in other parts of the world: Salvador, Uganda and Pakistan (Flaskerud & Nyamathi, 1996), the experience of Haitians in Miami-Dade, Florida ought to be understood in terms of cultural contexts, and the comprehension of the correlates attached to the practices can help prevent infections. Using the sample size and a survey element for the qualitative aspect of the research, the authors were able to bridge quantitative measures and analysis to present the followings after logistic measures were calculated and statistical interpretations made through:

1)                  The variables: gender, age, education, length of time in the U.S., recommendations by acquaintances, or friend, picuriste injections use is deep in a belief that benefits outweigh problems.

2)                  The following research questions were created by the researcher/writer:

RQ1: Is there a relationship between picuriste users and non-users? The proposed hypothesis was:

H0: There is a relationship between picuriste users and non-users.  The results of bivariate analyses showed that a significant difference between picuriste users (clients) and non-users (non clients) for education (X² =11.7; DF =2 and p < 0.01).

H1: Participants with a post-secondary diploma patronized a picuriste. The proposed research question by the researcher/writer was:

RQ2. Can education play a role in choosing the service of a picuriste? Only 7% of the population responded affirmatively.

RQ3: Is there a relationship between length of time in the U.S. and picuriste use? The proposed hypothesis by the authors was:

H3: The length of time participants were in the United States was inversely associated with a positive decision to use picuriste services. The chi-square was (X² =6.9; DF=2 and p = 0.03. The manner in which validity was analyzed is depicted below:

Internal validity: Bivariate analyses although show significant difference between picuriste users and non-users for education. However, it was not met because no significant differences between picuristes and non picuristes for any demographic variables.

External validity: No generalization was possible due to the limited sample size and other culturally specific health beliefs for this population. A factor analysis could have been used.

Conclusion validity: Correlates in the data: patronizing the practice of picuriste injections, health beliefs, efficacy of injections, and differentiation between picuriste users and non-users did not clearly establish a causal relationship between the demographics: age, education, gender, residence in the US considered in the multiple logistic regressions used.

Construct validity: The inferences generated are in conformity with the variables in the mixed-methods study where survey data used in the qualitative part, yielded fair statistical analyses.

2-Effect of medical advice for diet on diabetes self-management and glycemic control for Haitian and African Americans with type 2 diabetes. Huffman et al. (2013).

This research article posits an interesting claim about the effect of medical advice for type 2 Haitian-Americans and African-Americans diabetic participants from an ethnic standpoint, which is a solid contribution to the healthcare and research fields. The key point is that medical advice needs to be clearly communicated across ethnic spectrums to be effective. The goal of the research was to uncover the enablers and barriers of diet management specific to Black ethnicities with type 2 diabetes.

The research variables considered were: medical advice, level of following a meal plan or diet, and diabetes self-management (DSM. Using a cross-sectional design, the authors looked at a sample size: 254 Blacks with type 2 diabetes: Haitian-American =129 & African-American =125. The 3 Hypotheses testing the independent and dependent variables used were:

1) Has your health care provider or nurse ever told you to follow a meal plan or diet?

2) How often do you use the exchange lists or food group lists to plan your meal?

3) How often do you have trouble getting exercise?

These questions shed light on the main study variables and allowed for the consideration of a diabetic self-management composite (DSM) score. The DSM composite scores considered were DSM adherence, dietary patterns, exercise barrier scale, health beliefs, and heart disease. In terms of establishing some correlations between the many variables, covariates: socio demographics, psychosocial variables, depressive symptoms, anthropometrics and physical activity were used. Logical regression models were used to test hypotheses: participants were advised to follow diet, and participants use exchange lists or food groups to plan meals.

The research construct and results showed that elements of internal validity, external validity, conclusion validity and construct validity were analyzed and espoused. The salient findings were:

Internal validity: The effect of medical advice for diet has been measured via the variables according to the proposed study’s goal through adherence and various composite scores. Statistical analysis allowed for the original hypotheses to be accepted.

External validity: No generalization was made, although the 254 sample size was sound for statistical purposes. The rationale is due to differences in medical advice for diet for the 2 ethnic groups.

Conclusion validity: This cross-sectional study provided sound conclusions from the findings of the data analysis that cultural factors, socio demographics, education, and coping abilities must be taken in consideration when providing advices and education to patients managing their diabetes.

Construct validity: Statistical sound analysis and inferences from: Chi-Square test for categorical variables; past studies, and scales: behavioral risk factor surveillance, (BRFSS), the Michigan Diabetes Questionnaire, and the Joint Commission on Accreditation of Healthcare Organization, and the National Medical Association reports, confirmed the test of construct validity.

3. Depression, substance abuse and other contextual predictors of adherence to antiretroviral therapy (ART) among Haitians. (Malow et al., 2013).

The bulk of this cross-sectional research is fresh in considering its contribution to the Haitian society and the world of academic research. There is a need for a better understanding of depression in poor countries, and the urgency mandated by a transmissible disease such as HIV, is pronounced. From a sample of a previous research baseline of 256 HIV-positive Haitian adults in Haiti, the authors rightly looked at the characteristics of the sample and decided on the following variables: age, gender, depression, education level, higher alcohol impact, partner conflict, negative attitudes, maladaptive coping. Key outcome measures: adherence % & sum of endorsement barriers to adherence, and descriptive statistics of measured variables: mean, standard deviation, confirmatory factor analysis towards adherence to ART regimens, led to the testing of the 5 hypotheses.

1. Low health literacy, lack of education, and strong beliefs in alternative medicine practices may contribute to non-adherence to treatment.
2. Psychosocial and behavioral constructs of alcohol use, partner conflict, and maladaptive coping attitudes would have a depressive effect on medication adherence.
3. Negative attitudes about HIV medications predict poorer adherence
4. Alcohol problem predicts both partner conflict and maladaptive coping.
5. Being female and younger predicted more maladaptive coping.

The findings of this research corroborated the tests of validity expressed below:

Internal Validity: This research measured what it sought to measure by establishing levels of significance among the variables via statistical analysis. Poorer adherence was associated with female gender -18, p <0.05.

Depression & negative attitudes about HIV medication, p <0.001, and depression and maladaptive coping, partner conflict, negative attitudes about medication, and alcohol use were also significant.

External Validity: This research may not be generalizable to other more rural regions in Haiti.

Conclusion Validity: Depression and other psychosocial issues, and trust in medication regimen have a significant impact on adherence to ART.

Construct Validity: Inferences generated from the statistical analysis showed that depression had a significant relationship with partner conflict, maladaptive coping, and negative attitude about HIV medication.

Conclusion

The three researches considered elucidate a particular concept of health prevention through participation and better understanding of cultural factors capable of affecting health issues. Notions of compliance between two ethnic groups with type 2 diabetes offer a powerful insight due to proper medical advices, which need to always be clearly spelled out to minimize miscommunication. There are profound ethnic resonances that seem to offer new worldviews on health practices such as the use of picurstes, and Haitian health beliefs in Haiti where assessments of attitudes, participants’ personal and cultural beliefs of efficacy of drug treatment such as ART seem to call for constant understanding and review. Perhaps, poorer adherence may not be totally related to depression and negative attitudes.

References

• Clérismé, C. (2005). “Haitian health seeking behaviors and implications for policy formulation and program development in the United States.” November 18, 2005 Presentation at Center for Research on U.S. Latinos, HIV/AIDS and Drug Abuse (CRUSADA), Florida International University.
• Dumas, A., et al. (2014). Approches d’analyse causale en épidémiologie. Revue d’Épidemilogie et de Santé Publique, 62(1), 53-63.
• Flaskerud, J. H., & Nyamathi, A. M. (1996). Home medication injection among Latina women in Los Angeles: implications for health education and prevention. AIDS Care, 8(1), 95-102.
• Foster, O. (1976).Disease etiologies in non-western medical systems. American Anthropologist, 78,773-782.
• Gordis, L. (2000). Epidemiology, (2^nd ed.). Philadelphia: W.B. Saunders.
• Huffman. G., et al. (2013). Effect of medical advice for diet on diabetes self-management and glycemic control for Haitian and African Americans with type 2 diabetes. Food and Nutrition Sciences, 4, 1094-1101.
• Karhausen, L.R. (2000). The elusive grail of epidemiology. Medicine, Health Care and Philosophy, 3:59-67.
• Koch, R. (1880). Investigations into the etiology of traumatic infective diseases. The New Sydenham Society.
• Le, T., Bhushan, V., & Sochat, M. (2015). First Aid for the USMLE Step 1. New York: McGraw Hill.
• MacMahon, B., Pugh, T. (1970). Epidemiology, principles, and methods. Boston: Little, Brown.
• Malow, R., et al. (2013). Depression, substance abuse and other contextual predictors of adherence to antiretroviral therapy (ART) among Haitians. AIDS and Behavior, 17, 1221-1230.
• Ménard, J. et al. (2010). The sociocultural context of gynecological health among Haitian immigrant women in Florida; applying ethnographic methods to public health inquiry. Ethnicity & Health, 15(3), 253-267.
• Olsen, J. (2003). What characterizes a useful concept of causation in epidemiology? Journal of Epidemiology and Community Health, 57(2), 86-88.
• Pape, J. W., et al. (1986). Risks factors associated with AIDS in Haiti. Am J Med Sci., 291(1), 4-7.
• Parascandola, M., Weed D.L. (2001). Causation in epidemiology. Journal of Epidemiology and Community Health, 55(12), 905-911.
• Pasteur, L. (1866). Études sur le vin: ses maladies, causes qui les provoquent, procédés nouveaux pour le conserver et pour le vieillir. Paris: Imprimerie Impériale.
• Rahill, G. J., & Rice, C. (2010). Correlates of Picuriste Use in a Sample of health-seeking Haitian immigrants and adult children of immigrants in Miami-Dade County, Florida. American Journal of Public Research, Supplement 1 (100), 140-145.
• Virginia Department of Education. (2012).Science enhanced scope and sequence biology course.