The cross-sectional study was conducted in the perinatology room of the regional hospital in the province of East Java, Indonesia in March-June 2021. The perinatology room is an inpatient room for newborns who have health problems or are at risk of having health problems. The perinatology room is divided into three rooms. The first observation room, the second isolation room, and the third stable baby room. The observation room for low birth weight, asphyxia, respiratory distress syndrome, hyaline membrane disease. The isolation room for babies born outside the hospital, and babies with congenital abnormalities. The stable baby room for babies who come from other hospitals, or babies in isolation and observation rooms who are starting to stabilize.
The population was parents who had a newborn baby in the perinatology room. The criteria for parents are mother and father, have good awareness, able to read and write, and were present in the perinatology room. The estimated population that meets the inclusion criteria based on previous research is 106 [1]. Calculation of the number of samples is based on the 95% confidence level, 5% margin of error, and 50% population proportion. The sample required is 84 participants. Estimated drop-out participants are 10% (8.4). The total required participants is 84 + 8.4 = 92.4 = 92 participants. The sampling method is simple random sampling. The population was 604. 146 potential participants met the criteria from 604 participants. The results of a simple random draw resulted in 102 being included in the sample and 44 not being the sample. A total of 10 participants filled out incomplete data and were excluded from the study sample. So, number of samples size was 92 (see Fig. 1).
Collecting data using a questionnaire. The questionnaire was prepared by Rohmah, N. Yusuf., and Hargono in 2019. The questionnaire has been used in similar research in three districts of the East Java region of Indonesia [1]. Test the content validity by calculating the content-validity coefficient. Aiken's V (1985) coefficient value of 0.70 is considered to have adequate content validity. The results of the content-validity coefficient test have an average value between 0.80–0.93, indicating that the content validity is adequate. The results of the validity of the questionnaire with the provisions of n = 24; p < = 0.05; value of r table = 0.3297; Cronbach's Alpha value > 0.7. The results of testing the validity of the questionnaire have a value of r > 0.3297, and the value of Cronbach's Alpha > 0.7, and the questionnaire has met the criteria of validity and reliability [1].
The latent variables studied included: knowledge (ten indicators), preferences (three indicat ors), parental trust in nurses (five indicators), nurse and parent interactions (five indicators), and shared decision-making (five indicators). Latent variables are also called unobserved variables. Variables latent are, cannot be measured directly. Variable latent measured by one or more manifest variables or also called indicators [15]. Data analysis using Partial Least Square (PLS). Analysis includes measurements of the outer model and inner model. The results of the outer model testing, on the PLS test, were carried out on all indicators in the study.
Knowledge variable indicators include: 1) knowledge about how to treat breasts; 2) knowledge of how to express breast milk; 3) knowledge of how to store breast milk; 4) knowledge of how to deliver breast milk to the nursery; 5) knowledge of how to breastfeed correctly; 6) knowledge of how to increase breast milk; 7) knowledge of how to care for the kangaroo method; 8) knowledge of how to touch a baby; 9) knowledge of how to provide a sound stimulus; 10) knowledge about how to give affection. Each indicator consists of one statement on a Likert scale of 1–4. Indicators of preference variables consist of 1) autonomy; 2) cooperation; and 3) submission. Each indicator consists of one statement on a Likert scale of 1–4. The indicators of the trust variable consist of five indicators. The first, Giving the best action. The second is being able to take action in bad conditions. The third is able to provide accurate information. The fourth always provides information before taking action, and the five always is present if needed. Each indicator consists of one statement on a Likert scale of 1–4.
Indicators of interaction variables between nurses and parents consist of 1) information seeking; 2) providing information; 3) information verification; 4) socio-emotional communication, and 5) problem identification. Each indicator consists of one statement on a Likert scale of 1–4. The indicators for joint decision making consist of 1) discussing goals with a single statement; 2) options talk with four statements; 3) choice talk with two statements; 4) decision talk with four statements; and 5) commitment with four statements.
All statements with a Likert scale of 1–4. The results of the first measurements of the outer model have six loading factor values <0.5 and a T-statistic value < 1.96 so that the indicator is declared invalid and not significant. Invalid indicators are not included in the second test. The value of outer loading on the second test, all indicators of latent variables > 0.5 so that it has met the validity test. The result of the next test is the value of composite reliability or Cronbachs alpha of all latent variable constructs. The results of testing the value of composite reliability or Cronbach's alpha of all constructs showed > 0.7, so it can be concluded that all latent variables have met the reliability test. The result of the next test is Average Variance Extracted (AVE). AVE values above 0.5 are highly recommended. AVE value for all latent variables > 0.5. Next, evaluate of discriminant validity is measured to see cross-loadings. The test results show that the indicators in each component of the construct have a higher correlation with each construct. This means that the results of the examination of convergent validity and discriminant validity show that they have been met. Next, evaluate the structural model (inner) model.