Prof. (Dr.) Niranjan Mohanty
Two centuries back, a German physician called Dr.C.F.S Hahnemann who amazed this medical world with his astounding discovery of this scientific technique, to establish the curative power of substances in infinitesimal dilutions. The most brilliant minds of this era flocked to him. During the course of his life time he introduced ideas and principles of standardization which were trail blazers of that time. The methods were brilliant considering the technological limitations of that time. However technology has advanced astronomically but today’s scientific research methodology can run smoothly with the principles of homoeopathy or not is the subject of my discussion.
Let us delineate the principles of homoeopathy around which it is revolving. From advent of the Hahnemann’s era to the present era many erudite, veterans& scholars of Homoeopathy have contributed for the growth & development of homoeopathy and have evolved many principles for the purpose of practices. The major axioms which are universally accepted by classical Homoeopathy are as follows:-
- Law of Similia
- Law of Simplex
- Law of Minimum Dose
- Doctrine of Drug Proving
- Theory of Chronic Disease
- Theory of Vital Force
- Doctrine of Suppression
- Doctrine of Individualization
- Obstacles to Cure
- Doctrine of Analogy
- Doctrine of Concomitants
- Doctrine of Generalization & Others
- Totality of symptoms & others.
The principles required for clinical practices and necessarily relates directly for adoption of Research Methodology are as follows:
- Law of similia
- Law of simplex
- Law of minimum
Indirectly for building of totality of symptoms, principle involved are as follows:
- Doctrine of analogy
- Doctrine of concomitant
- Doctrine of generalization
- Totality of symptoms (1)
The principles involved during treatment & follow up of the cases are as follows:
- 1) Theory of chronic disease
- 2) Doctrine of suppression
- 3) Obstacles to cure
Now let me briefly describe few lines on research and research methodology.
Research is a quest for knowledge through diligent search or investigation or experiment aimed at the discovery and interpretation of new knowledge. (2)
Research is an art of scientific Investigation.
While conducting research in homoeopathy above axioms are to be adhered to “A careful investigation or inquiry specially through search for new facts in any branch of nowledge.” (3)
A systematized effort to gain new knowledge. (4)
Research methodology is a systematic body of procedures and technique applied in carrying out our investigation or experimentation targeted at obtaining new knowledge (WHO).
Research techniques that are used for conduction of research are three types such as
- Library Research.
- Field Research.
- Laboratory Research.
It is a way to systematically solve the research problem.
It is a science of studying how research is done.
Mainly research approaches are two types which are as follows:
Looking to the categories of research we visualize following types such as:
A. Empirical and theoretical Research:
The philosophical approach to research is basically of two types
b) Theoretical / Conceptual.
a) Empirical: – Health research mainly follows the empirical approach.
- It is based on observation and experience more than upon theory and abstraction.
- For example, Epidemiological research depends upon the systematic collection of observations on the health related phenomenon of interest in defined population.
- Empirical research can be qualitative or quantitative in nature.
- Health science research deals with information of a quantitative nature.
For the most part, this involves
- The identification of the population of interest.
- The characteristics (variables) of the individuals (units) in the population.
- The study of variability of these characteristics among the individuals in the population.
- Quantification in empirical research is achieved by three numerical procedures
i) Measurement of variables.
ii) Estimation of population parameters.
iii) Statistical testing of hypothesis.
- Empirical research relies on experience or observation without due regard to system or theory.
- We can call it experimental type of research.
b) Theoretical / Conceptual Research
- It is related to some abstract, idea(s), theories.
- In abstraction with mathematical models.
- Advances in understanding of disease occurrence and causation cannot be made without a comparison of the theoretical constructs with that which we actually observe in populations.
- Empirical & theoretical research complements each other in developing an understanding of the phenomena, in predicting future events.
B. Basic and applied:
a) Basic / fundamental / pure:
- It means formation of theory or generalization.
- Basic research is usually considered to involve a research for knowledge without a defined goal of utility or specific purpose.
b) Applied / Action:
- To find out the solution for immediate problems.
- Applied research is problem oriented.
C. Other categories of Research:
a) Longitudinal / one time research.
b) Field setting / laboratory / simulation research.
c) Clinical / diagnostic research.
d) Historical research.
e) Conclusion oriented / decision oriented research.
(5) Beliefs / attitudes / practice in society by man)
Several fundamental principles are used in scientific inquiry.
- Scientific method is not common sense.
- In arriving at conclusion “common sense” can’t be employed e.g. – Draft of air causes Allergic Rhinitis.
- To arrive at conclusion an organized observation of entities or events which are classified or ordered on the basis of common properties and behaviors are required.
- It is this commonality of properties and behaviors that allows predictions which carried to the ultimate, become laws e.g. A number of Allergic Rhinitis cases are studied and it is found a number cases are having a group of common causes from which prediction is made and there after etiology become conclusive as from ‘Allergens’.
B) Inference and chance
- Reasoning or inference is the force of advances in research
- In terms of logic, It means that a statement / or a conclusion ought to be accepted because one or more other statements / premises (evidence) are true.
- Inferential suppositions, presumptions or theories may be so developed, through careful constructions, as to pose testable hypothesis.
- The testing of hypothesis is the basic method of advancing knowledge in science.
- Two distinct approaches or arguments have evolved in the development of inferences. They are such as: deductive and inductive.
- In deduction, the conclusion necessarily follows from premises (evidence) / statements, as in syllogism e.g. [All ‘A’ is ‘B’, all ‘B’ is ‘C’ therefore all ‘A’ is ‘C’] or in algebraic equations.
- Deduction can be distinguished by the fact that it moves from the general to the specific.
- It dose not allow for the elements of chance or uncertainty.
- Deductive inference, therefore are suitable to theoretical research.
Induction: - Inductive reasoning is distinguished by the fact that it moves from the specific to the general (from sample to population). It builds
Health research being primarily empirical depends entirely upon inductive reasoning.
The conclusion dose not necessarily follows from the premises or evidence (facts). We can say only that the conclusion is more likely to be valid it the premises are true, i.e. there is a possibility that the premises may be true but the conclusion is false.
Chance must, therefore, be fully accounted for.
Mill’s canons of inductive reasoning are frequently utilized in the formation of hypothesis. These methods include:
a) Method of difference – When the frequency of a diseases is markedly dissimilar under two circumstances (For example, the difference in frequency of Lung Cancer in Smokers and Non-smokers.
b) Method of agreement – In a factor or its absence is common to a number of different circumstances that are found to be associated with the presence of disease, the factor or its absence may be casually associated with the disease (e.g. the occurrence of Hepatitis A is associated with patient contact, crowding & poor sanitation and hygiene, each conducive to the transmission of the Hepatitis virus).
c) Method of concomitant variation, or the dose response effect –
Example – – Increase expression of goiter with direct level of iodine in the diet.
Increasing frequency of leukemia with increasing radiation exposure.
Increase prevalence of elephantiasis in areas of increasing filarial endemicity.
d) Method of analogy – The distribution and frequency of a disease or effect may be similar enough to that of a some other disease to suggest commonality in cause (e.g. Hepatitis B virus infection and Cancer of the Liver). (6)
Designing and methodology of an experiment or a study consists of a series of guideposts to keep one going in right direction and sometime it may be tentative and not final.
The steps are as follows:
Definition of the problem: – Define the problem you intend to study such as Smoking and Lung Cancer, Cholesterol & C.A.D etc.
- Relevance of the problem with fields of application of proposed research result.
- Rationale of the study: – What necessitate to carry out the study.
2 .Review of literature: – Critically review the literature on the problem under study
- Any such work done by others in the past
- Want to confirm the findings
- Challenge the conclusion
- Extend the work further
- Bridge some gaps in the existing knowledge
3. Aim & Objectives: -
- Define the aims and objectives of the study.
- State whether nature of the problem has to be studied or solution has to be found by different methods.
- State your hypothesis.
- After the problem and purpose are clear and literature is reviewed.
- You have to start precisely with an assumption positive or negative, e.g. constitutional medicine is more effective for ‘Lymphangitis’ than pathological prescription with Hydrocotyle ‘Q.’
5. Plan of action: - “Prepare an over all plan or design of the investigation for studying the problem and meeting the objectives.”
A) Definition of the population under study
i) It may be country / state / districts / town / village / families / specific groups.
ii) Age group
iii) Income group
vi) Define clearly who are to be included and who are not to be included, i.e. (Inclusion and exclusion criteria)
B) Selection of the sample
a) It should be unbiased.
b) Sufficiently large in size to represent population under study.
a) Sample size
The size of sample is very vital in an scientific study. Ordinarily should not be less than 30. A sample small in size, is a biased one & should never be depended upon for drawing any conclusions, therefore however a large sample is considered as large enough. Normally cut off is taken as 30. A sample of size greater than 30 is considered large enough for statistical purpose.
For Qualitative Data
In such data we deal with proportions such as morbidity rates and cure rates. For finding the suitable size of the sample, the assumption usually made is that the allowable error does not exceed 10% or 20% of the positive character. The size can be calculated by the following formula with a desired allowable error (L) at 5% risk that the true estimate will not exceed allowable error by 10% or 20% of ‘p’ n=4pq/L2
Where ‘p’ is the positive character, q =1-p and L= allowable error, 10% or 20% of ‘p’
For Quantitative Data
In such data we deal with the means of a sample and of the universe. If the SD (s) in a population is known from the past experience, the size of sample can be determined by the following formulae with the desired allowable error (L). At 5% risk the true estimate will lie beyond the allowable error (variation).
Hence, the first step is to decide how large an error due to sampling defects can be tolerated or allowed in the estimates. Such allowable error has to be stated by the investigator.
The second step is to express the allowable error in terms of confidence limits. Suppose L is the allowable error in the sample mean and we are willing to take a 5% chance that error will exceed L. so we may put:
L=2s/Ön or Ön=2s/L or n= 4s2/L2
Sample size for analytical studies:
a. Testing equality of two proportions: p1 = p2
The sample measures used are the sample proportions, and the sampling distribution used in testing this null hypothesis is either the standard normal distribution (z), or equivalently the chi-square (c2).
- Set type I error:a;
- Determine ‘minimum clinically significant difference’:d;
- Make a guess as to the ‘proportion’ in one group (usually ‘control’): p1;
- Determine the power required to detect this difference: (1-b).
The sample size required is:
For example, suppose we are interested in determining the sample size required in a clinical trial of a new drug that is expected to improve survival. Suppose the traditional survival rate is 40%, i.e., p1 = 0.4. We are interested in detecting whether the new drug improves survival by at least 10%, i.e., d = 0.10, therefore p2 = 0.50. Suppose we want a type I error of 5%, i.e.,a = 0.05, therefore Z1-a = 1.96; we also want the type II error (b) to be 5%, or we want to detect a difference of 10% or more with a probability of 95%: therefore Z b = -1.645.
Substituting these values in the above equation given n = 640. Thus the study would require 640 subjects in each of the two groups to assure a probability of detecting an increase in the survival rate of 10% or more with 95%certainty, if the statistical test used 5% as the level of significance.
b. Sample size for a case – control study
Suppose that long term use of oral contraceptives (OC) increased the risk for coronary heart disease (CHD) and that one wished to detect an increase in relative risk of at least 30% (equivalently, OR>1.3) by means of a case control study, What would be the proper sample size?
The test of hypothesis in the study will be equivalent to testing if the proportion of women using (OC) is the same among those with CHD and those without CHD. We need to determine what proportion of women without CHD (controls) use OC; let us say 20%. Then we decide what will be the minimum difference that should be detected by the statistical test. Since we need to detect an OR >1.3, this translates to an increased use (24.5%) among the CHD patients to give a difference of 4.5% to be detected. Choosing a and b to be 5% each, the sample size, using the above formula, would be 2220, i.e., we need to study 2220 cases and 2220 controls for the disease.
Sometimes the ratio of cases and controls may not be one one, e.g., when the disease is rare then the number of cases available for study may be limited, and we may have to increase the number of controls ( I-2,1-3 etc.) to compensate. In such cases, the calculation of the sample size will incorporate these differences. Computer programmes such as EPIINFO allow for these variations.
c. Comparison of two population Means.
When the study involves comparing the means of two samples, the sample measure that is used as the difference of the sample means. This has an approximately normal distribution. The standard error of difference depends on the standard deviations of the measurements in each of the population, & depending on whether these are the same or the different, different formulae have to be used. In the simplest (and most commonly used) scenario, the two standard deviations are considered to be the same. We will illustrate the procedure.
We need to determine, as in case a, the minimum difference (d) in the means that we are interested in detecting by statistical tests: The two types of statistical errors (a and b) and the standard deviations (s). Then the sample size required is calculated using the following formula: n = [(Z1-a - Zb) s / d ]2
For example, suppose we want to test a drug that reduces blood pressure. We want to say the drug is effective if the reduction in blood pressure is 5mm of Hg or more, compared with the ‘ placebo’. Suppose we know that systolic blood pressure in a population is distributed normally with a standard deviation of 8 mm of Hg. If we choose a = 0.05 and
b = 0.05, the sample size required in this study will be: n=[(1.96 + 1.645) 8/3 ]2 = 34 subjects in each group.
If the design is such that the two groups are not independent (e.g., matched studies or paired experiments) or if the standard deviations are different for the two groups, the formulae should be adjusted accordingly.
d. Comparison of more than two groups and methods
When considering sample size calculations for studies involving comparisons of more than two groups, either comparing proportions or means several other issues (e.g. which comparison is more important than others; whether errors of paired comparison, or for the study as a whole are more important, etc.) have to be taken in to account. Accordingly the formulae for each of the situations will be much more complicated.
In multivariate analysis, such as those using multiple linear regression, logistic regression, or comparison of survival curves, simple formulae for the calculation of sample sizes are not available. Some attempts at estimating sample sizes using nomograms, or by simulating experiments and calculating sample sizes based on these simulated experiments, have recently appeared in the statistical literature. We will not discuss these here. When planning experiments, one of the crucial steps is in deciding how large the study should be, and appropriate guidance should be sought from experts
b) Sampling methods:
1) Simple random sampling – Choose random number from the table.
ii) Stratified sampling - (Selecting 50% male & 50% female)
iii) Systemic sampling – Systematically it is chosen.
iv) Cluster sampling - Cluster may be identified (households) and random samples of cluster.
v) Multistage sampling – In several stages.
c) Specifying the nature of study:
i) Longitudinal studies
- Prospective study
- Retrospective study
ii) Cohort studies: – A group of persons exposed to some sort of environment e.g. new born & mother exposed to radiation.
iii) Interventional studies: – In these there are three phases.
- Diagnostic / Identification
- Intervention by treatment
- Assessment phase for result
iv) Experimental studies
- Experimental or trial are made
- A drug is given and results are wanted
v) Cross sectional studies (Non experimental):- Such studies are one time or at a point of time study of all persons in a representative sample. It is conducted in field and not in laboratories.
Example: – Examination of children 2- 12 yrs and classify their nutritional grade.
Prevalence of pregnancy in age group of 20-25 yrs.
vi) Control studies: – Most of the experimental studies need a control as a yard stick of evidence.
Example: – Growth of child with constitutional medicine & control group with no medicine. Control group must be identical.
To rule out subjective bias in subjects under study single or double blind trial should be made. (7)
C) Research strategies & design:
The selection of a research strategy is the case of research design and is probably the single most important decision the investigator has to make:
Research strategy must include the followings:
1) Use of controls
2) Blinding – double or single
3) Study of instruments
4) Case recording format
5) Categorization – a) test group and b) control group
6) Parameter to assess the improvement – positive and negative response
7) Observations / Results
8) Presentation of data
9) Result analysis / Statistical evaluation
Statistical tools or tests of significance:-
For testing of hypothesis there is a large no. of tests available in the statistics. The most commonly tests for clinical study are follows:
Z – test, t – test and c2 – test. The other tests are also being used e. g. Variance ratio test and Analysis of Variance test.
Z – test:
It has two applications:
a) To test the significance of difference between a sample mean (X) and a known value of population (m).
Z = X – m / SE (X)
Where X = Sample mean
m = Population mean
SE = Standard error
b) To test the significance of difference of two sample means or between experiment sample mean and a control sample mean.
Z = observed difference between two sample means / Standard error of difference between two sample means.
Requirements to apply Z – test:
1. The sample or samples must be randomly selected.
2. The data must be quantitative.
3. The variable is assumed to follow normal distribution in the population.
4. The sample size must be larger than 30.
t – test:
Requirements to apply t – test:
1. The sample or samples must be randomly selected.
2. The data must be quantitative.
3. The variable is assumed to follow normal distribution in the population.
4. The sample size must be less than 30.
t = X – m / SE (X)
Where X = Sample mean
m = Population mean
SE = Standard error
Chi – square test (c2 – test):
It is a non-parametric test not based on any assumption or distribution of any variable. It is very useful in research. It is most commonly use when data are in frequencies such as in the no. of responses in two or more categories.
It has got the following three very important applications in medical statistics as tests of:
- Proportion – To find significance in same type of data.
- Association between two events.
- Goodness of fit – To test fitness of an observed frequency distribution of qualitative data to a theoretical distribution.
The test determines whether an observed frequency distribution differs from the theoretical distribution by chance or if the sample is drawn from a different population.
To apply c2 – test three essential requirements are needed such as:
- A random sample
- Qualitative data
- Lowest expected frequency (value) not less than 5
c2 = å (O – E)2 / E
Where O = Observed value
E = Expected value
Variance ratio test: (F – test):
This means comparison of sample variance. It is applied to test the homogeneity of variances.
F = S12 / S22 S12 = Variance of first sample
S22 = Variance of second sample
(S12 > S22)
ANOVA test (Analysis of Variance test):
This test is not confined to comparing two sample means but more than two samples drawn from corresponding normal population. (8)
A model case of scientific paper presentation on the caption “ Psoriasis in Homoeopathic practice” of the author will be projected to justify that homoeopathic principles can run hand in hand in conducting research work on homoeopathy as per the modern Research Methodology.
1, Roberts, A.Herbert, The Principles and art of cure of Homoeopathy, reprint edition 1997.
2,5,7. Health Research Methodology, A guide for training in research Methods,2 nd edition.W.H.O.regional office for the Western Pacific, Manila 2001,page-1.
3. Collin’s cobuild English Dictionary for advanced learners major new edition, Harper Collins publishers, 1995.
4. English Dictionary, Read man.
6. Klinbanm D.G., Kupper L.L, Morgenstern H.,, Epidemiological Research Principles and auantitative methods London, Life time Learning Publication 1982.
7,8. Mahajan B. W., Methods in Biostatistics,Jaypee brothers Medical Pub., New Delhi,6 th edition,reprint-1994.
Prof. (Dr.) Niranjan Mohanty. M. D. (Hom.)
Dean of the Homoeopathic Faculty, Utkal University, Orissa.
Principal-cum-Superintendent, H.O.D, P.G Department of Repertory
Dr. A. C. H. M. C. & H, Bhubaneswar.
National President, I. I. H. P.
Member C.C.H New Delhi.