Data analysis was conducted in three stages. First, descriptive statistical analysis was used to summarize respondent characteristics and describe perceptions of each HOT-Fit dimension. The descriptive measures included frequency, percentage, mean, and standard deviation. Second, the quality of the measurement instrument was evaluated through validity and reliability testing. Validity was assessed using item-total correlation, while reliability was tested using Cronbach’s Alpha, as presented in Equation (2) [25].

Third, multiple linear regression analysis was employed to examine the influence of the Human, Organization, and Technology dimensions on Net Benefit. In this model, the three HOT-Fit dimensions functioned as independent variables, whereas Net Benefit was treated as the dependent variable. The analysis was intended to determine the magnitude, direction, and statistical significance of the relationships between variables. Before estimating the regression model, classical assumption tests were performed to verify model suitability. A normality test was used to assess whether the residuals followed a normal distribution. Multicollinearity was examined using the Variance

Inflation Factor (VIF) to ensure that correlations among independent variables did not compromise coefficient estimation. A linearity test was conducted to confirm the existence of linear relationships between predictors and the dependent variable. In addition, residual analysis was undertaken to evaluate homoscedasticity and to detect potential deviations from regression assumptions. All statistical tests were interpreted using a significance level of p < 0.05.

This study involved 67 respondents who were active users of Medxa SIMRS at Mohammad Hoesin Hospital, Palembang. Descriptive analysis was first conducted to examine respondents’ perceptions across the four principal constructs of the HOT-Fit model, namely Human, Organization, Technology, and Net Benefit. The mean scores and category assessments for each indicator are presented in Table 1.

As shown in Table 1, all HOT-Fit dimensions were evaluated positively, with mean scores ranging from 3.70 to 4.20, indicating that respondents generally perceived the implementation of Medxa SIMRS favorably. Among the three explanatory dimensions, Technology obtained the highest average score (Mean = 4.20; Very Good), followed by Human (Mean = 3.93; Good) and Organization (Mean = 3.70; Good). The Net Benefit dimension also showed a positive evaluation (Mean = 4.10; Good), suggesting that respondents perceived the system as beneficial for work efficiency, decision-making, and service delivery. These findings indicate that the technological aspects of Medxa SIMRS— particularly system quality and service quality—constitute the strongest perceived advantages of the system. By contrast, the relatively lower score for organizational structure (Mean = 3.6) suggests that institutional arrangements, coordination, or role clarity may not yet be fully optimized to support implementation. Before inferential analysis, descriptive statistics were also generated for the composite variables used in the regression model. These results are presented in Table 2.

As presented in Table 2, the average scores for all variables were relatively high, indicating generally positive perceptions across the study sample. The Human dimension had the highest composite mean (70.12; SD = 7.79), followed by Technology (37.81; SD = 4.29), Net Benefit (36.90; SD = 4.13), and Organization (33.30; SD = 4.94). The standard deviations indicate moderate dispersion, suggesting that although respondents tended to evaluate the system positively, some variability remained in their assessments. Overall, these descriptive findings provide an initial indication that Medxa SIMRS was well received by users from the perspectives of human readiness, organizational support, technological quality, and perceived benefits.

To examine the extent to which the HOT-Fit dimensions influenced implementation outcomes, multiple linear regression analysis was performed with Net Benefit (TOTAL_N) as the dependent variable and Human (TOTAL_H), Organization (TOTAL_O), and Technology (TOTAL_T) as the independent variables. The regression coefficients are summarized in Table 3.

The regression model was statistically significant overall (F(3, 63) = 4.288; p = 0.008), indicating that the three HOT-Fit dimensions jointly contributed to explaining variation in Net Benefit. The model produced an R² value of 0.170, which indicates that 17.0% of the variance in Net Benefit was explained by the Human, Organization, and Technology dimensions, while the remaining 83.0% was attributable to factors not captured in the model. Although this explanatory power is modest, it remains acceptable for organizational and behavioral research, where perceptions and contextual influences are often shaped by multiple interacting factors.

Among the three predictors, only the Technology dimension showed a statistically significant positive effect on Net Benefit (B = 0.299; t = 2.608; p = 0.011). This result suggests that higher perceptions of system quality, information quality, and service quality are associated with greater perceived benefits of Medxa SIMRS. By contrast, the Human dimension (p = 0.816) and the Organization dimension (p = 0.103) did not demonstrate statistically significant effects at the 5% significance level, despite their positive coefficient directions. These findings indicate that, within this sample, technological factors were the most salient determinant of perceived implementation success.

To further assess the predictive behavior of the regression model, the same specification was also implemented as a supervised machine learning model using the Linear Regression algorithm in scikit-learn. The dataset was divided into 80% training data and 20% testing data, and model performance was evaluated using R², RMSE, and MAE. The results are presented in Table 4.

As shown in Table 4, the model demonstrated limited predictive performance. On the training dataset, the model achieved R² = 0.206, indicating that approximately 20.6% of the variance in Net Benefit could be explained by the three HOT-Fit dimensions. The corresponding error values (RMSE = 3.62; MAE = 2.80) suggest a moderate level of prediction error. However, performance decreased substantially on the testing dataset, where the model yielded a negative R² value (-0.151) and higher error values (RMSE = 4.22; MAE = 3.36). A negative test R² indicates that the model performed worse than a baseline prediction based solely on the mean, thereby reflecting weak generalizability to unseen data.

This limitation is also illustrated in Figure 2, which displays the relationship between actual and predicted values. In principle, data points that lie close to the identity line (y = x) indicate better predictive accuracy. In the present study, the dispersion of points away from the identity line suggests that the model was only moderately effective in fitting the observed data and less effective in predicting new observations. These findings imply that while the regression model is useful for identifying statistically relevant relationships within the study sample, its value as a predictive tool remains limited.

The dotted line represents the identity line (y = x); the closer the observations are to this line, the better the predictive accuracy of the model.

The present study demonstrates that the implementation of Medxa SIMRS at Mohammad Hoesin Hospital was generally perceived positively across all dimensions of the HOT-Fit model. Respondents reported good to very good evaluations of the Human, Organization, Technology, and Net Benefit dimensions, indicating that the system has been accepted reasonably well during its early implementation phase. The strongest performance was observed in the Technology dimension, whereas the Organization dimension received the lowest average score. This pattern suggests that the technical characteristics of the system were more readily recognized by users than the organizational conditions that support its implementation.

The prominence of the Technology dimension in both descriptive and inferential analyses constitutes the central finding of this study. The regression results indicate that Technology was the only HOT-Fit dimension that significantly influenced Net Benefit, implying that users’ perceptions of system quality, information quality, and service quality were directly associated with the perceived benefits of Medxa SIMRS. In practical terms, this means that when the system is considered reliable, responsive, and capable of providing accurate information, users are more likely to perceive improvements in efficiency, decision-making, and service quality. This finding reinforces the proposition that technological performance is a fundamental driver of information system success, particularly in healthcare settings where speed, accuracy, and continuity of information are critical to daily operations.

By contrast, the Human and Organization dimensions did not show statistically significant effects on Net Benefit, although both variables had positive coefficients. This result does not necessarily imply that human and organizational factors are unimportant. Rather, it may indicate that these dimensions were relatively stable across respondents and therefore contributed less to differentiating perceived system benefits. The high mean scores observed for these dimensions suggest that users generally viewed competence, satisfaction, management support, and work culture positively. Under such conditions, the absence of statistical significance may reflect limited variability in responses, which reduces the ability of the model to detect distinct effects. It is also possible that the contribution of these dimensions is more indirect, for example by influencing system use, adaptation, or acceptance rather than perceived benefit directly.

Another important finding is the relatively modest explanatory power of the regression model (R² = 0.170). This suggests that a substantial proportion of the variance in Net Benefit remains unexplained by the three HOT-Fit dimensions included in the model. In the context of a large public referral hospital, this is understandable, as perceptions of information system benefit are likely influenced by many additional factors, including workload, departmental variation, leadership style, policy enforcement, training intensity, interoperability with other systems, and the broader organizational climate. Accordingly, the present results should be interpreted as evidence that the HOT-Fit dimensions contribute meaningfully—but not exhaustively—to perceived implementation success.

The supplementary machine learning analysis further supports this interpretation. Although the model displayed modest fit on the training data, its performance deteriorated on the test data, as indicated by the negative R² and increased prediction error. This outcome suggests that the relationships captured by the model may be sample-specific and insufficiently robust for prediction in new cases. From a methodological standpoint, these results imply that the current model is better suited for explanatory analysis than for predictive application. Future studies seeking stronger predictive performance may need to incorporate a larger sample, additional explanatory variables, or alternative modeling techniques capable of capturing more complex relationships among the determinants of system success.

Taken together, the findings highlight an important implication for hospital management. While user-related and organizational conditions should continue to be maintained, priority attention should be directed toward strengthening the technological quality of Medxa SIMRS, particularly in relation to system reliability, information accuracy, and service responsiveness. At the same time, the relatively lower descriptive score for organizational structure indicates that implementation may be further improved through clearer coordination mechanisms, stronger cross-unit integration, and more explicit operational roles. In this regard, the success of SIMRS implementation appears to depend not only on adoption itself, but on sustained efforts to ensure that the system remains technically effective and institutionally supported within the complexity of hospital operations.