Good practices in normal childbirth: reliability analysis of an instrument by Cronbach’s Alpha

ABSTRACT Objectives: to analyze the internal consistency of the evaluation instrument of the adherence to the good practices of childbirth and birth care in the professionals, through Cronbach’s Alpha Coefficient for each of the dimensions and for the total instrument. Method: this is a descriptive and cross-sectional study performed in obstetric centers of eleven public hospitals in the Federal District, with a questionnaire applied to 261 professionals who worked in the delivery care. Results: The study was attended by 261 professionals, 42.5% (111) nurses and 57.5% (150) physicians. The reliability evaluation of the instrument by the Cronbach Alfa resulted in 0.53, 0.78 and 0.76 for dimensions 1, 2 and 3, after debugging that resulted in the exclusion of 11 items. Conclusions: the instrument obtained Cronbach’s alpha of 0.80. There is a need for improvement in the items of dimension 1 that refer to attitudes, knowledge, and practices of the organization of the network of care to gestation, childbirth, and birth. However, it can be applied in the way it is used to evaluate practices based on scientific evidence of childbirth care.

cognitive and intellectual level, beliefs, and social context of which the person is a part. The terms adherence and compliance have been used to designate the degree of coincidence between the behaviors of the individual (patient or client) and the therapeutic recommendations of the health professional (10) . In this study, adherence was adopted as the coincidence between the knowledge, attitudes, and practices of the professionals regarding the technical recommendations and scientific evidence on the childbirth care (11) .
The instrument of measures analyzed in this article is focused on the context of the health services, the multiple knowledge of the professionals, the values, beliefs, and ideas that intermediate the relationship established with the parturients (10)(11) . It was constructed based on the practices proposed by the "Rede Cegonha" guidelines, treated as constructs to be measured (latent variable) through indirect observation of their manifestations in the daily work process in health, in the network of childbirth and birth care (12) . These manifestations were transformed into items (operational concepts) that constitute attitudes, knowledge, practices, behaviors, and opinions about the health and physical and psychological well-being of women (11)(12)(13) .
The research data were used to test the validity (ability to measure what is proposed in a certain phenomenon) and the reliability (ability to present measures faithful to reality) of the instrument (14)(15) . items (14)(15)(16) . questions about the socioeconomic, demographic and professional profile (11) . The items of the instrument referring to the professional practices follow the scale of five Likert points, transformed into values from 0 to 100, so ordered and scored: disagree completely (1=0 point), disagree partially (2=25 points) , does not know/does not apply (3=50 points), partially agrees (4=75 points), totally agrees (5=100 points).
The profile data were analyzed using descriptive statistics. Cronbach's alpha was calculated to evaluate the internal consistency of the instrument. It is a coefficient that measures the correlation between the answers in a questionnaire through the analysis of the profile of the answers given by the respondents, whose values vary from 0 to 1 (14)(15)(16)(17) . The closer to 1, the greater the reliability of the indicators. A generally accepted lower limit is 0.7, although it drops to 0.6 in exploratory research. The Cronbach Alpha Reliability classification occurs as follows: Very low (α ≤ 0.30); Low (0.30 <α ≤ 0.60); Moderate (0.60 <α ≤ 0.75); High (0.75 <α ≤ 0.90) and Very high (α> 0.90) (14)(15)(16)(17) .

Reliability (intensity of correlation between items)
was further tested by eliminating items from the questionnaire in the debugging process. If with the elimination of an item, the Coefficient increased, it was assumed that this item was not highly correlated with the other items of the scale and it could be eliminated from the instrument. If the Coefficient decreased, it was assumed that this item was highly correlated with the other items of the instrument. Alfa Cronbach was calculated for the instrument as a whole and for each dimension (14)(15)(16)(17) .

Discussion
In the reliability analysis of the instrument by Cronbach's Alpha, it was possible to obtain an improved Rev. Latino-Am. Enfermagem 2018;26:e3000.
proposal of the questionnaire "adherence to good practices in the normal birth care" (11). Dimension 1 presented a coefficient of 0.53, with low correlation among items, therefore, they had a low reliability. The professionals (1)(2)11) . This may explain the moderate consistency of this dimension (0.60 <α ≤0.75) (14)(15) .
In Dimension 2, after clearance, Cronbach's alpha was high in the total dimension and in all items (high 0.75 <α ≤ 0.90) (14)(15) . The two excluded items refer to the performance of vaginal touches by more than one professional and to the use of early amniotomy in the services, both reversal items in the instrument and unnecessary obstetric interventions (2)(3)(4)(5)(6)(7) .
The evaluation of the adherence to practices based on scientific evidence demonstrated in the items of Dimension 2 showed that professionals both perform good practices and still intervene unnecessarily in childbirth.
The results are consistent with data from other national studies, which demonstrate that episiotomy is still performed in 56% of vaginal deliveries; that lithotomy delivery occurs in 92% of the women and that 37% of the mothers underwent Kristeller's maneuver, none of them supported by the best available evidence (4)(5) .
Positive practices such as feeding during labor (26%) and There is a need for important actions to favor the incorporation of simple recommendations, such as wandering or not maintaining routine venous access (6)(7)(8) .
Devices such as birth plans constructed from basic health units during prenatal care deserve to be taken up to help professionals and women to jointly rebuild the technicality of gestation, childbirth and birth care in favor of a preventive, contemplative and humanistic attention (1)(2)(3)(4)(5)(6)(7)(18)(19)(20)(21) . The incorporation of practices with scientific foundations and respecting childbirth may be very rewarding for health professionals, but implying a re-signification of asymmetric power relationships (18)(19)(20) .
Regarding It is important to highlight that the work process requires a permanent education in procedures, in clinical protocols, in the construction of information and decision-sharing spaces, and in technological support (1)(2)(3)(4)(5)(6)(7) . Although clearance has substantially altered the result by showing that the excluded items had low consistency, the researcher's decision about the relevance or otherwise of the exclusion is necessary (14)(15) .
It is observed that all are negative items that may have influenced the result and refer to extremely relevant aspects of the work process. It is important to emphasize that this coefficient assists the researcher on the pertinence or not of an item in a given questionnaire, but does not replace the decision on the relevance of the item within the general context of the construct under study (14)(15) . For this reason, although they did not obtain Alfa Cronbach high, the items of the first dimension were maintained. There organization of integrated systems for care to gestation, childbirth, and birth (16) .
Another result to be observed was the difference of Cronbach's Alpha between gender and between the professional categories of the study. By the results, women and nurses are more sincere in the answers. This data is relevant since the values of Alfa can be altered by other characteristics of the sample in new applications of the instructor (14,16) .
It is important to highlight that the robustness of the results of a study depends on the instrument used that must have internal consistency and quality (14)(15)(16) . The evaluation of health services still presents a challenge, considering the complexity of the object to be evaluated,

Conclusion
The value of the total Cronbach's Alpha coefficient of the instrument after the debugging was 0.80 with 39 items. Among the dimensions, this coefficient ranged from 0.53 to 0.76. The first dimension presented a low correlation with the other items, which requires a new application after revision to achieve a higher reliability index. However, as it stands, it can be applied to evaluate the knowledge, attitudes, and practices of professionals who work in childbirth care.