兰德-A-Psychometric-Study-of-the-Modified-Bridges-for-Newborns-Screening-Tool_82页_746kb

Psychometric Study of the Modified Bridges for Newborns Screening Tool Summary

Core Content

The Modified Bridges for Newborns Screening Tool is a maternal risk assessment instrument used by First 5 Los Angeles (First 5 LA) to identify mothers in need of additional supportive services through the Welcome Baby program. This program provides free services to pregnant women and families with children under nine months in Los Angeles County (LAC), particularly in Best Start communities. The tool consists of 26 questions grouped into three subscales: Medical, Psychosocial, and Demographics/Basic Needs, with each question weighted based on the severity of the risk it assesses.

The RAND Corporation conducted a psychometric study to evaluate the reliability and validity of the Modified Bridges. The study addressed three key research questions:

1. Do hospital liaisons assign similar scores when using the Modified Bridges?
2. Does the Modified Bridges identify a similar level of risk as other validated measures?
3. Does the Modified Bridges accurately distinguish between low-/moderate-risk and high-risk mothers?

Main Findings

1. Inter-Rater and Test-Retest Reliability

• Hospital liaisons showed acceptable levels of inter-rater reliability and test-retest reliability, indicating consistency in scoring.
• The Demographics/Basic Needs subscale had the highest inter-rater reliability, while the Medical subscale had moderate reliability.
• Experience and training significantly improved liaison performance.
• A cutoff score of 50 is used to classify mothers as high risk (50 or above) or low/moderate risk (below 50).

2. Comparison with Other Risk Measures

• The Modified Bridges is most comparable to the Florida Healthy Start Prenatal and Infant Risk Screens, which assess similar content.
• The current scoring protocol may overidentify high-risk mothers, and using a higher risk threshold could improve accuracy.
• Sensitivity and specificity analyses suggest the tool may not effectively distinguish between high and low/moderate risk mothers.

3. Reliability and Validity of the Tool

• The current reliability of the Modified Bridges is 0.65, which is below the widely accepted threshold of 0.70.
• Removing redundant and non-discriminatory items (e.g., infant medical problems and mother's use of English) and adjusting item weights could improve reliability to 0.73 (standardized reliability = 0.75).
• Factor analysis supported the three subscale structure of the tool.
• The psychosocial subscale had the highest factor loading, indicating it captures the most significant aspects of maternal risk.

Key Recommendations

• Adjust item weights for seven items to increase their contribution to the total score.
• Decrease weights for three items that may not be as relevant.
• Remove two items:
  • Infant medical problems (redundant with other items).
  • Mother's use of English (does not contribute to risk discrimination).
• These changes would improve the reliability and validity of the Modified Bridges.

Limitations

• The study used mock patient interviews and video vignettes, which may not fully reflect real-world scenarios.
• The sample size was limited, and the risk threshold may not be optimal for all populations.
• The reliability of the tool remains moderate, suggesting further refinements are needed.

Conclusion

The Modified Bridges is a valuable tool for identifying maternal risk in the Welcome Baby program, but it has limitations in reliability and risk discrimination. While hospital liaisons generally score consistently, the current scoring protocol could be enhanced to improve accuracy. The study recommends item weight revisions and removal of non-discriminatory items to better align the tool with validated risk measures and improve its psychometric properties.

Key Information

• Modified Bridges consists of 26 questions with weights ranging from 3 to 18 points.
• High-risk mothers are eligible for Select Home Visitation (SHV) programs, which offer more intensive support.
• Reliability of the Modified Bridges is 0.65, and revisions could bring it up to 0.73.
• The tool's validity is supported by its similarity to Florida measures, but there is room for improvement in risk threshold accuracy.

Appendices and References

• Appendix A provides details on the Legacy Instrument Protocol.
• Appendix B includes frequencies and descriptive statistics from the validity study.
• The References section lists all sources and instruments used in the study.

Figures and Tables

• Figure 2.1: Overall Hospital Liaison Performance and Outlier Detection
• Figure 3.1: Receiver Operating Characteristic for Prenatal Risk
• Figure 3.2: Receiver Operating Characteristic for Postnatal Infant Risk
• Table 1.1: Modified Bridges Questions and Scoring by Weight
• Table 2.1: Participating Hospitals and Medical Centers
• Table 2.2: Modified Bridges Total Score and Subscale Means and Standard Deviations
• Table 2.3: Kappa and Kendall Coefficients for Inter-Rater Agreement
• Table 2.4: Percentage of Perfect Agreement Between Liaisons and Between Times
• Table 3.1: Comparison Measures to Validate the Modified Bridges
• Table 3.2: Modified Bridges and Florida Risk Frequency and Percentages
• Table 3.3: Rules of Thumb for Correlation Coefficients
• Table 3.4: Correlations Between Items and Total Scores
• Table 3.5: Pearson Correlations Between Modified Bridges and Florida Risk Scores
• Table 3.6: Dichotomous Risk Cutoff Frequencies
• Table 3.7: Sensitivity and Specificity of Modified Bridges
• Table 3.8: Comparison of Modified Bridges Total Scores Across Known Risk Groups
• Table 4.1: Factor Loadings from Confirmatory Factor Analysis
• Table 4.2: Cronbach's Alphas for Modified Bridges
• Table 4.3: Item-Total Correlations Based on Raw and Standardized Scores
• Table 4.4: Proposed Changes to Item Weights Based on Correlations and Effect Sizes
• Table 4.5: Cronbach's Alphas Using Raw and Standardized Scores
• Table B.1: Modified Bridges Frequencies and Descriptive Statistics from the Validity Study