Does synthetic data augmentation improve the performances of machine learning classifiers for identifying health problems in patient-nurse verbal communications in home healthcare settings?

BACKGROUND: Identifying health problems in audio-recorded patient-nurse communication is important to improve outcomes in home healthcare patients who have complex conditions with increased risks of hospital utilization. Training machine learning classifiers for identifying problems requires resource-intensive human annotation. OBJECTIVE: To generate synthetic patient-nurse communication and to automatically annotate for common health problems encountered in home healthcare settings using GPT-4. We also examined whether augmenting real-world patient-nurse communication with synthetic data can improve the performance of machine learning to identify health problems. DESIGN: Secondary data analysis of patient-nurse verbal communication data in home healthcare settings. METHODS: The data were collected from one of the largest home healthcare organizations in the United States. We used 23 audio recordings of patient-nurse communications from 15 patients. The audio recordings were transcribed verbatim and manually annotated for health problems (e.g., circulation, skin, pain) indicated in the Omaha System Classification scheme. Synthetic data of patient-nurse communication were generated using the in-context learning prompting method, enhanced by chain-of-thought prompting to improve the automatic annotation performance. Machine learning classifiers were applied to three training datasets: real-world communication, synthetic communication, and real-world communication augmented by synthetic communication. RESULTS: Average F1 scores improved from 0.62 to 0.63 after training data were augmented with synthetic communication. The largest increase was observed using the XGBoost classifier where F1 scores improved from 0.61 to 0.64 (about 5% improvement). When trained solely on either real-world communication or synthetic communication, the classifiers showed comparable F1 scores of 0.62-0.61, respectively. CONCLUSION: Integrating synthetic data improves machine learning classifiers' ability to identify health problems in home healthcare, with performance comparable to training on real-world data alone, highlighting the potential of synthetic data in healthcare analytics. CLINICAL RELEVANCE: This study demonstrates the clinical relevance of leveraging synthetic patient-nurse communication data to enhance machine learning classifier performances to identify health problems in home healthcare settings, which will contribute to more accurate and efficient problem identification and detection of home healthcare patients with complex health conditions.

Cyberincivility among Adolescents: A Systematic Review of Qualitative Evidence.

BACKGROUND: The ubiquitous use of social media places adolescents at high risk for cyberincivility (disrespectful, insensitive, or disruptive online behaviors). Adolescents who experience cyberincivility can have mental health issues including depression and suicidal ideation. However, no reviews synthesized findings from qualitative studies on cyberincivility among adolescents, which was the aim of this review. METHODS: Eleven articles were extracted from MEDLINE, Embase, PsycINFO, ERIC, and Education Full Text. Study qualities were assessed using the JBI Critical Appraisal Checklist. A meta-ethnographic framework guided the thematic analysis. RESULTS: Five themes were identified: forms of cyberincivility, platforms used for cyberincivility, perceptions and understanding of cyberincivility, adverse impact on health, and perceptions of seeking support. Social media was often used for cyberincivility, and common forms of cyberincivility included impersonation, aggressive verbal behaviors, and social exclusion. Adolescents perceived cyberincivility as persistent and relentless, with severe health impacts including depression, but they hesitate to seek support. CONCLUSIONS: Considering the negative impact of cyberincivility on adolescents' physical and mental well-being, it is important to develop anti-cyberincivility programs in schools. Schools have opportunities to efficiently reach both perpetrators and victims of cyberincivility with anti-cyberincivility programs rooted in trust building and engagement from family, community, and peers.

Addressing bias in preterm birth research: The role of advanced imputation techniques for missing race and ethnicity in perinatal health data.

OBJECTIVES: To evaluate the effectiveness of Bayesian Improved Surname Geocoding (BISG) and Bayesian Improved First Name Surname Geocoding (BIFSG) in estimating race and ethnicity, and how they influence odds ratios for preterm birth. METHODS: We analyzed hospital birth admission electronic health records (EHR) data (N = 9985). We created two simulation sets with 40 % of race and ethnicity data missing randomly or more likely for non-Hispanic black birthing people who had preterm birth. We calculated C-statistics to evaluate how accurately BISG and BIFSG estimate race and ethnicity. We examined the association between race and ethnicity and preterm birth using logistic regression and reported odds ratios (OR). RESULTS: BISG and BIFSG showed high accuracy for most racial and ethnic categories (C-statistics = 0.94-0.97, 95 % confidence intervals [CI] = 0.92-0.97). When race and ethnicity were not missing at random, BISG (OR = 1.25, CI = 0.97-1.62) and BIFSG (OR = 1.38, CI = 1.08-1.76) resulted in positive estimates mirroring the true association (OR = 1.68, CI = 1.34-2.09) for Non-Hispanic Black birthing people, while traditional methods showed contrasting estimates (Complete case OR = 0.62, CI = 0.41-0.94; multiple imputation OR = 0.63, CI = 0.40-0.98). CONCLUSIONS: BISG and BIFSG accurately estimate missing race and ethnicity in perinatal EHR data, decreasing bias in preterm birth research, and are recommended over traditional methods to reduce potential bias.

Postpartum Hospital Discharge: Birthing Parent Perspectives on Supportive Practices and Areas for Improvement.

BACKGROUND: Postpartum hospital care and individualized discharge preparedness should be part of person-focused health services. Yet, there are limited descriptions of birthing parents' experiences to identify clinical practice strengths and opportunities to improve systems of care. OBJECTIVE: To explore birthing parents' perspectives on supportive healthcare practices and areas for improvement around postpartum hospital discharge. METHODS: In this mixed-methods study, participants completed an online questionnaire and a semistructured, telephone interview at 2 to 3 weeks postpartum. Researchers summarized responses to 2 quantitative questions and conducted a thematic content analysis of interview data. RESULTS: Forty birthing parents participated (90% non-White). According to quantitative responses, most birthing parents were prepared to be discharged (82.5%). Responses to the interview generated 6 broad factors related to postpartum hospital discharge preparedness: inpatient postpartum support, physical and emotional health, patient priorities and agency, clear and relevant information, holistic care, and scheduling and continuity of care. Researchers further identified themes around specific healthcare practices participants described to be supportive and opportunities for improvement. CONCLUSION: Birthing parents articulated multiple contributors to their preparation for postpartum hospital discharge. These perspectives offer insights for strengthening systems of perinatal care and inform measures of quality postpartum care.

Maternal Psychological Symptom Trajectories From 2 to 24 Months Postpartum: A Latent Class Growth Analysis.

BACKGROUND: Postpartum women experience an array of psychological symptoms that are associated with adverse health behaviors and outcomes including postpartum suicidal ideation and long-term depression. To provide early management of postpartum psychological symptoms, it is important to understand how the symptom experiences change over time. OBJECTIVE: The aim of this study was to examine maternal psychological symptom trajectories over 2, 6, 15, and 24 months postpartum using latent class growth analysis and to examine how each trajectory is associated with maternal depression outcome at 24 months. METHODS: We used secondary data from the Family Life Project ( N = 1,122) and performed multitrajectory latent class growth analysis based on four observed symptom variables (depression, anxiety, somatization, and hostility). After the final model was identified, bivariate analyses were conducted to examine the association between each trajectory and (a) individual characteristics and (b) outcome (Center for Epidemiologic Studies Depression Scale [CES-D]) variables. RESULTS: A four-class model was selected for the final model because it had better fit indices, entropy, and interpretability. The four symptom trajectories were (a) no symptoms over 24 months, (b) low symptoms over 24 months, (c) moderate symptoms increasing over 15 months, and (d) high symptoms increasing over 24 months. More postpartum women with the trajectory of high symptoms increasing over 24 months (Trajectory 4) were in low economic status (92.16%), unemployed (68.63%), or did not complete 4-year college education (98.04%). Most postpartum women (95.56%) in Trajectory 4 also had higher CES-D cutoff scores, indicating a possible clinical depression at 24 months postpartum. DISCUSSION: Postpartum women who experience increasing symptom trajectories over 15-24 months (Trajectories 3 and 4) could benefit from tailored symptom management interventions provided earlier in the postpartum period to prevent persistent and worsening symptom experiences.

Examination of Social Determinants of Health Characteristics Influencing Maternal Postpartum Symptom Experiences.

Postpartum women experience multiple, co-occurring postpartum symptoms. It is unknown if social determinants of health (SDOH) influence postpartum symptom typologies. This secondary analysis used the Community and Child Health Network study data. Participants included for analysis varied depending on the availability of the SDOH data (N = 851 to 1784). Bivariate and multiple regression analyses were conducted to examine the association between SDOH and previously identified postpartum symptom typologies. Area under the receiver operating characteristics curve (AUROC) was calculated to examine if adding SDOH variables contributes to predicting postpartum symptom typologies. The adjusted odds (aOR) of being in high symptom severity or occurrence typologies were greater for participants who had less than high school education (aOR = 2.29), experienced healthcare discrimination (aOR = 2.21), used governmental aid (aOR = 2.11), or were food insecure (aOR = 2.04). AUROC improved after adding SDOH. Considering experiences of different social-economic hardships influence postpartum symptom typologies, future practice and research should address SDOH to improve postpartum symptom experiences.

Changes in maternal psychological symptom profiles from 2 to 6 months postpartum: an application of latent transition analysis.

To identify subgroups of postpartum women with different psychological symptom profiles at 2 and 6 months postpartum and to examine how they transition between symptom profiles over time using latent transition analysis (LTA). We used secondary data from the Family Life Project (N = 1,117) and performed LTA based on observed variables (depression, anxiety, somatization, and hostility). We examined transition probabilities and changes in latent status prevalence from 2 to 6 months postpartum. Considering the known influences of social determinants of health on psychological symptoms, bivariate analyses were conducted to describe the characteristics of different transition patterns. A 3-class model with better fit indices, entropy, and interpretability was selected. Based on symptom severity, the identified profiles were Profile 1: Low, Profile 2: Moderate, and Profile 3: High. From 2 to 6 months postpartum, the prevalence of low symptom profile decreased (82 to 78.2%) while the prevalence increased for moderate (15.8 to 17.5%) and high symptom profiles (2.2 to 4.4%). For all profiles, it was most likely for postpartum women to stay in the same profile from 2 to 6 months (low to low, moderate to moderate, and high to high). Those in persistent or worsening transition patterns were significantly younger or had less social support or education. Postpartum women in moderate or high symptom profiles at 2 months were most likely to stay in the same profile at 6 months postpartum, indicating persistent symptom burden. Clinicians should consider providing early, targeted support to prevent persistent symptom burden.

Examination of Maternal Allostatic Load Among Postpartum Women With Distinct Postpartum Symptom Typologies.

BACKGROUND: An increased allostatic load (cumulative physiologic wear and tear of the body) can lead to adverse health outcomes. Symptom experiences are known to influence allostatic load. Yet, the relationships between postpartum symptom typologies and maternal allostatic load remain unknown. METHODS: We used Community Child Health Network data and included participants with allostatic load data at 6, 12, or 24 months postpartum. Bivariate and multivariate analyses were conducted to examine associations between postpartum symptom typologies and (a) overall allostatic load, (b) allostatic load subscales for body systems (neuroendocrine, cardiovascular, metabolic, and inflammatory), and (c) individual biomarkers within the subscale. RESULTS: Overall allostatic load at 12 months postpartum was different by symptom typologies before (p = .042) and after adjusting for confounders (p = .029). Postpartum women in typology 5 (high overall) had the highest adjusted overall allostatic load (M = 4.18, SE = .27). At 12 months, adjusted allostatic load for the cardiovascular subscale was higher in typologies 3 (moderate-high sleep symptoms, M = 1.78, SE = .13) and 5 (high overall, M = 1.80, SE = .17). Within the cardiovascular subscale, those in typology 3 had higher adjusted odds for a clinically significant level of pulse rate (aOR = 2.01, CI = 1.22, 3.31). CONCLUSION: Postpartum women who experienced high symptom severity across all symptoms (typology 5) at 6 months had higher overall allostatic load at 12 months postpartum. Typologies 3 and 5 had the highest symptom severity in sleep-related symptoms and higher cardiovascular subscale scores. Postpartum symptom management should target symptom burden in an effort to reduce allostatic load thereby improving postpartum women's health outcomes.

Transitions in Social Networks From Young-Old to Old-Old Stage of Life Using Latent Transition Analysis.

Introduction: This study aimed to identify the different types of social networks among young-old adults, and to examine the transitions in social networks as they become old-old adults. Methods: This is a secondary data analysis using the longitudinal data (N = 1092) from the National Social Life, Health, and Aging Project. Latent class analysis was conducted to identify optimal number of classes and latent transition analysis was conducted to examine the transition probabilities. Results: Young-old adults in Class 1: family-oriented, social (close, external) transitioned into Class 2: family-oriented, non-social over time. In contrast, young-old adults in Class 2: family-oriented, non-social and Class 3: less family-oriented, social (close) were less likely to transition to another class. Conclusion: Older adults engaged in less social activities over time. Older adults should be encouraged to continue their social engagement with close social network of friends and relatives, and to maintain their relationship with family members.

Identification of postpartum symptom subgroups and associated long-term maternal depressive symptoms and well-being.

Many postpartum women experience postpartum symptoms which often occur in clusters (i.e., three or more co-occurring symptoms that are related to each other). To date, research has focused on individual symptoms, which limits our understanding of how postpartum symptom clusters manifest and influence health. This secondary analysis used the Community and Child Health Network study data (N = 1784). No patient or public directly participated or contributed to the current analysis. Guided by the Symptom Management Theory, latent class analysis was performed to identify subgroups of postpartum women with different symptom experiences using observed variables at 6 months postpartum: anxiety (MINI-anxiety), general stress (PSS-10), posttraumatic stress (PCL-C), postpartum depression (EPDS), sleep disturbance (PSQI-sleep disturbance), and sleep duration (PSQI-sleep duration). Bivariate and multiple regression analyses were conducted to examine the association between subgroups and (a) individual characteristics and (b) long-term depressive symptoms (CES-D-9) and well-being at 18 and/or 24 months postpartum. Five subgroups were selected that had better-fit indices, entropy, and interpretability. Subgroups were labeled as (1) Minimum overall, (2) Mild-moderate overall, (3) Moderate-high sleep symptoms, (4) High psychological symptoms, and (5) High overall. After adjusting for covariates, postpartum women in Subgroups 4 and 5 had higher CES-D-9 scores at 18 and 24 months and lower well-being scores at 24 months. More postpartum women in Subgroups 4 and 5 experienced a history of depression or unemployment. Clinicians should provide targeted interventions for postpartum women in high-symptom subgroups.

Envisioning career trajectory post-PhD in nursing: Lessons learned from a professional development project.

BACKGROUND: Nursing Doctor of Philosophy (PhD) students can pursue diverse career opportunities within and outside of academia upon graduation. However, mentor-mentee models, competing demands, and limited resources can challenge students as they search for guidance in navigating career decisions. This article describes the development, implementation, and evaluation of a project to support PhD nursing career development. METHODS: A student-designed project was implemented over 4 weeks which aligned with four career trajectories that students identified. Descriptive statistics were used to analyze quantitative survey questions. Responses to open-ended questions and field notes were also examined. RESULTS: Post-implementation survey data suggested that all participants found the sessions helpful and recommended providing the workshop annually. Students' questions focused on three areas: job searches, job selection, and experiences once in a career trajectory. Workshop speakers' discussions focused on important tasks and strategies and wisdom and personal reflections offered to PhD students. DISCUSSION: Nursing PhD students are interested in diverse career trajectories beyond academia and valued an opportunity to explore these options outside of the traditional mentor-mentee relationship. Leveraging resources from schools of nursing and the broader collegiate environment is important in helping students to explore potential career trajectories.

Postpartum symptom clusters in birthing individuals: A systematic review.

BACKGROUND: Birthing individuals experience various postpartum symptoms which have been associated with adverse health outcomes. Previous systematic reviews have focused on the examination of individual symptoms and their impact on health, which have limited our understanding of postpartum symptom clusters. OBJECTIVE: To examine the compositions of symptom clusters, analytic methodologies and predictors of symptom clusters in birthing individuals up to 1 year postpartum. DESIGN: Systematic review. METHODS: This systematic review was reported following the PRISMA. Five databases were searched: PubMed, Embase, CINAHL Complete, PsycINFO and Scopus. Two reviewers performed title and abstract and full-text screening independently. Standard Quality Assessment Criteria were used to assess the articles' qualities. Key information was extracted into evidence table, which was checked for accuracy and completeness. A narrative synthesis was conducted. RESULTS: A total of 30 articles were included. Studies were conducted in 16 countries, mostly in Europe (n = 9) and North America (n = 7). The majority were quantitative (n = 27) and cross-sectional (n = 27). Factor analysis was the most frequently used analytic methodology (n = 21). All three qualitative studies used grounded theory. Taking into consideration the variations among the studies, stress (n = 15, infant or partner-related and from other sources), depression (n = 12), somatic (n = 12) and anxiety clusters (n = 10) were most frequently identified. Symptom cluster predictors were examined in less than half of the included studies (n = 13). Among these, most were focused on how individual postpartum symptoms influence symptom clusters. CONCLUSIONS: Stress, depression, somatic and anxiety clusters are the most frequently identified postpartum symptom clusters. Future studies should examine the consistency, stability and clinical meaningfulness of these four symptom clusters. RELEVANCE TO CLINICAL PRACTICE: The identification and management of the four symptom clusters should be of particular interest to clinicians and researchers. NO PATIENT OR PUBLIC CONTRIBUTION: This systematic review did not directly involve patient or public contribution to the manuscript.