The objective of this type of study is to estimate the SENSITIVITY (proportion of true positives) and SPECIFICITY (proportion of true negatives) of the osteopathic test. These are the operative characteristics of a diagnostic test. They measure the value of a diagnostic test.

- SENSITIVITY (SYNONYM: TRUE POSITIVE RATE)

Referring to a diagnostic test, this is the proportion of truly ill people who have been catalogued as such by means of this test.

- SPECIFICITY (SYNONYM: TRUE NEGATIVE RATE)

The specificity is the proportion of true negatives. Its complementary values, that is, the proportion of false positives and false negatives, are also used as indices of error.

A DIAGNOSTIC TEST MUST BE SENSITIVE (THAT IS, BE ABLE TO DETECT) AND SPECIFIC (CORRECTLY DETECT WHAT ONE WANTS TO DETECT).

**CARRIE OUT A RANDOM CONTROLLED OBSERVATIONAL DOUBLE-BLIND STUDY:**

- Healthy control group of patients (30 patients) and study groups (30 patients) with pathology in keeping with the studied test: same number of people in each group.
- The subjects in the two groups can be classified: prior to the start of the study (a priori), by means of the evaluation and results of the Gold Standard technique (a posteriori), or by means of the evaluation and results of the osteopathic test that we are evaluating (a posteriori).
- There must be a benchmark (Gold Standard, reference test*) which identifies healthy and ill people.

The study methodology should include the Inclusion/Exclusion criteria for each group.

On many occasions there is no objective standard, for example, radiography, histology test, etc. In this case, “consensus among the judges” can be used to decide who is ill and who is healthy.

It is necessary to know his interexaminator degree of agreement measured with the Kappa coefficient or other tests (3 or 4 examinators carry out the same steps: the person who carries out the study does not know their decision).

- Interexaminator with Kappa coefficient, 3 or 4 examinators carry out the same steps: the person who carries out the study cannot take part as examinator).
- The examiner who does the osteopathic test should not know whether the subject is healthy or ill (independent and blind verification).
- The benchmark should have been established without knowing the result of the osteopathic test. The objective is to estimate the RELIABILITY, degree of agreement or interchangeability among observers.
- The Kappa coefficient can only be used for qualitative (test) variables (most osteopathic tests are dichotomous, so it is very valid). In the case of quantitative variables, the results should be categorized (classify in intervals) or use other statistical graphs different from Kappa.

**Measures of agreement:**

- Kappa coefficient for qualitative variables, YES/NO: Example X-rays; Kappa between 0 and 1.
- Statistical contrast with McNemar non-parametric test (2×2). (Impossible without computer statistics programme).
- Intraclass Correlation Coefficient (ICC) for quantitative variables. Example, algometry or VAS. ICC between 0 and 1.
- Statistical contrast (F test) (Impossible without computer statistics programme).

Double or triple-blinded, being aware of the difficulties facing osteopathy on this point, requiring special attention in the study design.

Several variables are studied: the sensitivity and specificity should be accompanied by their Confidence Interval.

Several variables are studied:

**Principal or dependent variable:**Pain, mobility, function, etc.**Independent variables:**age, sex, severity, etc.

We should remember on this point that these variables may also be:

**Quantitative variables**(algometry* or VAS): compares means with parametric Student’s t-tests, two groups, or ANOVA if more than two groups.**Qualitative variables**(X ray*, Yes/No, present/absent; healed/not healed): compares proportions, Chi-squared.

The normality of the population must be studied (Smirnov-Kolgorov).

If the distributions are not normal (Gaussian), use non-parametric tests such as the Mann-Whitney U test (2×2) or the Kruskall-Wallis test for more than two groups.

The protocol used for Algometry and X-rays has to be scientific (rigorous in its application, identical for each patient), all the examiners should be instructed that they must all do the same thing, in the same way, in each patient (the same for the osteopathic test).

**Algometry:**Know the error margin of the apparatus, protocol concerning location of the painful area (Trigger) to be studied, protocol of the use of the algometer (angle of inclination position).**Radiographs:**Strict positioning protocol of patient and articulation to be studied (possible construction of a frame to fix an identical articular position in each patient), orientation of X-rays always identical, articulation-X-ray tube distance always identical.**Photography:**Strict positioning protocol of patient and articulation to be studied (possible construction of a frame to fix an identical articular position in each patient), camera orientation always identical. With Photoshop the same type of image always (TIFF), same image size, measure in millimetres or count pixels.

This type of statistics should include, mean, mode, median, maximum, minimum, confidence interval range, variance, standard deviation of **all** the variables **always**.