The concept of a reference variable
The most important stage in evaluating the results of laboratory studies is the establishment of a difference in the norm from pathology. This is not difficult to do with an obvious deviation from the norm. However, most of the results of laboratory analyzes are not easily divided into "norm" and "pathology", because they are by nature not dichotomous and do not have distinct breaks or two different peaks, one of which would correspond to the normal result and the other would be pathological. This is explained by several reasons.
First, the division of the biological population of people in many laboratory indicators into patients and healthy is impossible even from a theoretical point of view. The disease can develop imperceptibly, manifesting itself as a gradual transition from small deviations of the indicator to high as dysfunction increases.
Secondly, healthy and sick actually belong to two different populations, but when these two populations are mixed, it is almost impossible to recognize each of them in the total mass, since in different patients the same index can take different values, overlapping the values of this indicatorin healthy individuals;in addition, the number of patients in the general population is small.
In order to interpret laboratory data, it is necessary to compare them with normal values, so it is important to determine what a normal indicator is. Normal indicators are indicators found in healthy people, but in groups of the latter they can have different numerical values. This is due to individual features of metabolism, hemopoiesis, the functioning of certain organs. Normal laboratory indicators are determined by sampling a healthy population of people, for example, specially selected conscripts or students grouped by age and sex. During the research, some factors should be standardized. For example, in the study of blood, it should be taken away on an empty stomach, the sampling method for all subjects should be the same,
, as well as the method of determining the indicators to be studied. A mathematical analysis of the results obtained in such studies has made it possible to distinguish two main classes of parameters of healthy human biomaterials. Some of them obey the law of Gaussian( normal) distribution, others - binomial distribution.
For example, all subjects are assessed the concentration of glucose in the blood and build a distribution curve. The average value is calculated by dividing the sum of all the results by their number.
where: Хср - the average value;n is the number of results, X is the value of a single result.
The average variance in the Gaussian distribution can be expressed by the standard deviation( SD), which is calculated using the following formula.
As a rule, the distribution of biological objects in terms of the severity of one of the characteristics describes the Gauss curve, this implies that in an interval where the magnitude of the tag varies within M + 2SD( with respect to the concentration of glucose in the blood - 3,96,38 mmol / L)more than 95% of biological objects enter;nevertheless, in almost 5% of individuals in a healthy population, the glucose concentration is not included in the M + 2SD interval. That's why the criterion for diagnosing diabetes is the blood glucose concentration of 7 mmol / L and above, and patients with results within the range of 6.38-6.9 mmol / l are considered to be at risk for this disease.
Thus, if the distribution of the characteristic is Gaussian law, the normal laboratory parameters are defined as the mean value of the indicator for a healthy population of +2 standard deviations( + 2SD).
n
Fig. Gaussian distribution( for example, the concentration of glucose in the blood in healthy individuals).
Fig. Gaussian distribution( for example, the concentration of glucose in the blood in healthy individuals).
However, in 5% of healthy people the value of the indicator is beyond the specified range. The given mathematical regularity obeys the distribution of a significant part of the laboratory indices of the chemical and cellular composition of the blood.
The second group of laboratory indicators include results for which the calculation of the mean value and the standard deviation is not possible. Therefore, for such indicators, instead of the most frequent normal value, the limits of normal vibrations are determined and indicated. You can simply indicate the range of results from the smallest to the largest, but more often cut 3% of the first values (from below) and 3% of the last( from above).
At the same time, normal laboratory indicators of various substances that are often used in laboratory diagnostics include only general biological variation without taking into account individual factors, which reduces the diagnostic value of laboratory tests. Therefore, the term "normal laboratory indicators" is replaced by the concept of reference values. The reference values give an idea of the range in which the normal quantities are located. The meaning of this introduction is that the results of a laboratory study are compared with reference values obtained under clearly defined conditions, taking into account individual factors that affect biological variation. The reference values are currently set for a limited number of indicators( approximately 150).Establishment of reference vibration intervals for each laboratory parameter is essential for the whole problem of the reliability of laboratory information, since comparison with them is the basis for making diagnostic and therapeutic decisions.
When assessing the results of laboratory studies, it must be remembered that the reference values are statistical data of 95% of the population, and deviations from the range do not necessarily indicate a pathology.
Typically, a standard set of biochemical studies used in conventional institutions includes at least 10-12 tests. The probability that the results of all 12 tests will turn out to be normal is not high. In the statistical analysis, it was found that in determining 8 indicators, the result of one of them would be "pathological" in approximately 25% of healthy individuals, and in 20 tests one or more abnormalities would be detected in 55% [Statland B. E. et al., 1977].The data cited confirm the idea that each laboratory test should be given deliberately, according to strict indications, and the list of screening tests should be limited.
Thus, approximately 5% of healthy people reveal "abnormal" laboratory indicators, so not all values that go beyond the norm should be regarded as pathological. Conversely, the indicator lying in the interval M + 2SD is not always normal, since the range of many parameters is quite wide. For example, normal hematocrit( Ht) in men varies from 42 to 52%.Massive blood loss can lead to a drop in Ht from 52 to 42%.The 42% index will not cause anxiety in doctors, since it refers to the range of normal values, although for a specific patient this decrease can be