Artificial neural network approaches to identify maternal and infant risk and asset factors using Peridata.Net: a WI-MIOS study.

Objective: To analyze PeriData.Net, a clinical registry with linked maternal-infant hospital data of Milwaukee County residents, to demonstrate a predictive analytic approach to perinatal infant risk assessment. Materials and Methods: Using unsupervised learning, we identified infant birth clusters with similar multivariate health indicator patterns, measured using perinatal variables from 2008 to 2019 from n = 43 969 clinical registry records in Milwaukee County, WI, followed by supervised learning risk-propagation modeling to identify key maternal factors. To understand the relationship between socioeconomic status (SES) and birth outcome cluster assignment, we recoded zip codes in Peridata.Net according to SES level. Results: Three self-organizing map clusters describe infant birth outcome patterns that are similar in the multivariate space. Birth outcome clusters showed higher hazard birth outcome patterns in cluster 3 than clusters 1 and 2. Cluster 3 was associated with lower Apgar scores at 1 and 5 min after birth, shorter infant length, and premature birth. Prediction profiles of birth clusters indicate the most sensitivity to pregnancy weight loss and prenatal visits. Majority of infants assigned to cluster 3 were in the 2 lowest SES levels. Discussion: Using an extensive perinatal clinical registry, we found that the strongest predictive performance, when considering cluster membership using supervised learning, was achieved by incorporating social and behavioral risk factors. There were inequalities in infant birth outcomes based on SES. Conclusion: Identifying infant risk hazard profiles can contribute to knowledge discovery and guide future research directions. Additionally, presenting the results to community members can build consensus for community-identified health and risk indicator prioritization for intervention development.

Comparison of SIREN social needs screening tools and Simplified Omaha System Terms: informing an informatics approach to social determinants of health assessments.

OBJECTIVE: Numerous studies indicate that the social determinants of health (SDOH), conditions in which people work, play, and learn, account for 30%-55% of health outcomes. Many healthcare and social service organizations seek ways to collect, integrate, and address the SDOH. Informatics solutions such as standardized nursing terminologies may facilitate such goals. In this study, we compared one standardized nursing terminology, the Omaha System, in its consumer-facing form, Simplified Omaha System Terms (SOST), to social needs screening tools identified by the Social Interventions Research and Evaluation Network (SIREN). MATERIALS AND METHODS: Using standard mapping techniques, we mapped 286 items from 15 SDOH screening tools to 335 SOST challenges. The SOST assessment includes 42 concepts across 4 domains. We analyzed the mapping using descriptive statistics and data visualization techniques. RESULTS: Of the 286 social needs screening tools items, 282 (98.7%) mapped 429 times to 102 (30.7%) of the 335 SOST challenges from 26 concepts in all domains, most frequently from Income, Home, and Abuse. No single SIREN tool assessed all SDOH items. The 4 items not mapped were related to financial abuse and perceived quality of life. DISCUSSION: SOST taxonomically and comprehensively collects SDOH data compared to SIREN tools. This demonstrates the importance of implementing standardized terminologies to reduce ambiguity and ensure the shared meaning of data. CONCLUSIONS: SOST could be used in clinical informatics solutions for interoperability and health information exchange, including SDOH. Further research is needed to examine consumer perspectives regarding SOST assessment compared to other social needs screening tools.

Real-world implementation evaluation of an electronic health record-integrated consumer informatics tool that collects patient-generated contextual data.

OBJECTIVE: Use the RE-AIM framework to examine the implementation of a patient contextual data (PCD) Tool designed to share patients' needs, values, and preferences with care teams ahead of clinical encounters. MATERIALS & METHODS: Observational study that follows initial PCD Tool scaling across primary care at a Midwestern academic health network. Program invitations, enrollment, patient submissions, and clinician views were tracked over a 1-year study period. Logistic regression modeled the likelihood of using the PCD Tool, accounting for patient covariates. RESULTS: Of 58,874 patients who could be contacted by email, 9,183 (15.6%) became PCD Tool users. Overall, 76% of primary care providers had patients who used the PCD Tool. Older age, female gender, non-minority race, patient portal activation, and Medicare coverage were significantly associated with increased likelihood of use. Number of office visits, medical issues, and behavioral health conditions also associated with use. Primary care staff viewed 18.7% of available PCD Tool summaries, 1.1% to 57.6% per clinic. DISCUSSION: The intervention mainly reached non-minority patients and patients who used more health services. Given the requirement for an email address on file, some patients may have been underrepresented. Overall, patient reach and adoption and clinician adoption, implementation, and maintenance of this Tool were modest but stable, consistent with a non-directive approach to fostering adoption by introducing the Tool in the absence of clear expectations for use. CONCLUSION: Healthcare organizations must implement effective methods to increase the reach, adoption, implementation, and maintenance of PCD tools across all patient populations. Assisting people, particularly racial minorities, with PCD Tool registration and actively supporting clinician use are critical steps in implementing technology that facilitates care.

Analysis of Clinician and Patient Factors and Completion of Telemedicine Appointments Using Video.

Importance: Telemedicine provides patients access to episodic and longitudinal care. Policy discussions surrounding future support for telemedicine require an understanding of factors associated with successful video visits. Objective: To assess patient and clinician factors associated with successful and with failed video visits. Design, Setting, and Participants: This was a quality improvement study of 137 846 scheduled video visits at a single academic health system in southeastern Wisconsin between March 1 and December 31, 2020, supplemented with patient experience survey data. Patient information was gathered using demographic information abstracted from the electronic health record and linked with block-level socioeconomic data from the US Census Bureau. Data on perceived clinician experience with technology was obtained using the survey. Main Outcomes and Measures: The primary outcome of interest was the successful completion of a scheduled video visit or the conversion of the video visit to a telephone-based service. Visit types and administrative data were used to categorize visits. Mixed-effects modeling with pseudo R2 values was performed to compare the relative associations of patient and clinician factors with video visit failures. Results: In total, 75 947 patients and 1155 clinicians participated in 137 846 scheduled video encounters, 17 190 patients (23%) were 65 years or older, and 61 223 (81%) patients were of White race and ethnicity. Of the scheduled video encounters, 123 473 (90%) were successful, and 14 373 (10%) were converted to telephone services. A total of 16 776 patients (22%) completed a patient experience survey. Lower clinician comfort with technology (odds ratio [OR], 0.15; 95% CI, 0.08-0.28), advanced patient age (66-80 years: OR, 0.28; 95% CI, 0.26-0.30), lower patient socioeconomic status (including low high-speed internet availability) (OR, 0.85; 95% CI, 0.77-0.92), and patient racial and ethnic minority group status (Black or African American: OR, 0.75; 95% CI, 0.69-0.81) were associated with conversion to telephone visits. Patient characteristics accounted for systematic components for success; marginal pseudo R2 values decreased from 23% (95% CI, 21.1%-26.1%) to 7.8% (95% CI, 6.3%-9.4%) with exclusion of patient factors. Conclusions and Relevance: As policy makers consider expanding telehealth coverage and hospital systems focus on investments, consideration of patient support, equity, and friction should guide decisions. In particular, this quality improvement study suggests that underserved patients may become disproportionately vulnerable by cuts in coverage for telephone-based services.

Impact of Pre-visit Contextual Data Collection on Patient-Physician Communication and Patient Activation: a Randomized Trial.

BACKGROUND: Patient contextual data (PCD) are often missing from electronic health records, limiting the opportunity to incorporate preferences and life circumstances into care. Engaging patients through tools that collect and summarize such data may improve communication and patient activation. However, differential tool adoption by race might widen health care disparities. OBJECTIVE: Determine if a digital tool designed to collect and present PCD improves communication and patient activation; secondarily, evaluate if use impacts outcomes by race. DESIGN, SETTING, AND PARTICIPANTS: A pragmatic, two-armed, non-blinded, randomized controlled trial conducted during 2019 in a primary care setting. INTERVENTION: The PCD tool (PatientWisdom) invited patients to identify preferences, values, goals, and barriers to care. Patients were randomized to a standard pre-visit email or facilitated enrollment with dedicated outreach to encourage use of the tool. MAIN OUTCOMES AND MEASURES: Outcomes of interest were post-visit patient communication and patient activation measured by the Communication Assessment Tool (CAT) and Patient Activation Measure (PAM), respectively. Outcomes were evaluated using treatment-on-the-treated (TOT) and intention-to-treat (ITT) principles. KEY RESULTS: A total of 301 patients were enrolled. Facilitated enrollment resulted in a five-fold increase in uptake of the PCD tool. TOT analysis indicated that the PCD tool was associated with notable increases in specific CAT items rated as excellent: "treated me with respect" (+ 13 percentage points; p = 0.04), "showed interest in my ideas" (+ 14 percentage points; p = 0.03), "showed care and concern" (+ 16 percentage points; p = 0.02), and "spent about the right amount of time with me" (+ 11 percentage points; p = 0.05). There were no significant pre/post-visit differences in PAM scores between arms (- 4.41 percentage points; p = 0.58). ITT results were similar. We saw no evidence of the treatment effect varying by race in ITT or TOT analyses. CONCLUSIONS AND RELEVANCE: The inclusion of PCD enhanced essential aspects of patient-provider communication but did not affect patient activation. Outcomes did not differ by race. TRIAL REGISTRATION: Clincaltrials.gov identifier: NCT03766841.

Racial Disparities in Patient Activation: The Role of Economic Diversity.

The Patient Activation Measure (PAM) assesses a person's level of knowledge, skills, and confidence to self-manage their day-to-day health. We conducted a mediation analysis to examine potential direct effects of race on significantly lower baseline PAM scores in Black than in White participants (p<0.001) who were a subset of 184 adults who participated in a randomized controlled trial. In the mediation analysis, using natural indirect effects, the continuous outcome was the PAM score. The mediators were income, education, ability to pay bills, and health literacy; race (Black or White) was the "exposure." The results indicate that income (p=0.025) and difficulty paying monthly bills (p=0.04) mediated the relationship between race and baseline PAM score, whereas health literacy (p=0.301) and education (p=0.436) did not. Researchers must further investigate the role of economic diversity as an underlying mechanism of patient activation and differences in outcomes.Clinical Trial Registration: Avoiding Health Disparities When Collecting Patient Contextual Data for Clinical Care and Pragmatic Research: NCT03766841https://clinicaltrials.gov/ct2/show/NCT03766841?term=crotty&draw=2&rank=1.

The Impact of Previsit Contextual Data Collection on Patient-Provider Communication and Patient Activation: Study Protocol for a Randomized Controlled Trial.

BACKGROUND: Patient-centered care is respectful of and responsive to individual patient preferences, needs, and values. To provide patient-centered care, clinicians need to know and incorporate patients' context into their communication and care with patients. Patient contextual data (PCD) encompass social determinants of health and patients' needs, values, goals, and preferences relevant to their care. PCD can be challenging to collect as a routine component of the time-limited primary care visit. OBJECTIVE: This study aims to determine if patient-provider communication and patient activation are different for patient users and patient nonusers of an electronic health record (EHR)-integrated PCD tool and assess if the impact of using PCD on patient-provider communication and patient activation differs for Black and White patients. METHODS: We describe a randomized controlled trial of a prospective cohort of non-Hispanic White and Black patients who receive primary care services at a midwestern academic health care system in the United States. We will evaluate whether providing PCD through a consumer informatics tool enhances patient-provider communication, as measured by the Communication Assessment Tool, and we will evaluate patient activation, as measured by the Patient Activation Measure for PCD tool users and nonusers. Furthermore, owing to racial disparities in care and communication, we seek to determine if the adoption and use of the tool might narrow the differences between patient groups. RESULTS: The trial was funded in November 2017 and received local ethics review approval in February 2019. The study began recruitment in April 2019 and enrollment concluded in October 2019 with 301 participants. The analysis was completed in May 2020, and trial results are expected to be published in winter 2020. CONCLUSIONS: Recently, there has been increased attention to the role of health information technology tools to enable patients to collaborate with providers through the sharing of PCD. The adoption of such tools may overcome the barriers of current EHRs by directly engaging patients to submit their contextual data. Effectively, these tools would support the EHR in providing a more holistic understanding of the patient. Research further supports that individuals who have robust digital engagement using consumer informatics tools have higher participation in treatment follow-up and self-care across populations. Therefore, it is critical to investigate interventions that elicit and share patients' social risks and care preferences with the health care team as a mechanism to improve individualized care and reduce the gap in health outcomes. TRIAL REGISTRATION: ClinicalTrials.gov NCT03766841; https://clinicaltrials.gov/ct2/show/NCT03766841. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): RR1-10.2196/20309.

The impact of patient-generated contextual data on communication in clinical practice: A qualitative assessment of patient and clinician perspectives.

BACKGROUND: Effective communication is integral to patient-centered care, yet external pressures can impede the ability to discuss important topics. One strategy to facilitate communication is pre-visit collection and sharing of patient contextual data (PCD), including life circumstances such as their beliefs, needs, and concerns. OBJECTIVE: To understand how patients and care team members perceive the electronic collection of PCD and its impact on communication in the context of a large academic health system that implemented PatientWisdom, a new technology that elicits PCD from patients and integrates within the electronic health record (EHR). METHODS: We conducted focus groups with patients (n = 26) and semi-structured interviews with primary care team members (n = 20). Qualitative analysis of focus group/interviews included an iterative and reflexive inductive technique to uncover emergent themes. RESULTS: Four themes were reflected among both patient and care team: (1) the technology enhances the patient's voice; (2) the technology creates a safe space for patients to share sensitive topics; (3) PCD facilitates rapport not only between patient and provider but the entire care team; (4) PCD aligns patient and clinician goals. Two unique themes emerged among patients: (1) PCD provides opportunity for reflection; (2) PCD humanizes patients in the clinical context. One theme was evident in provider comments: collecting PCD may potentially undermine trust if not reviewed by clinical teams. CONCLUSION: PCD collected directly from patients and available within the EHR was seen by patients and care team members as beneficial to communication. PCD collection supports a paradigm shift towards coproduction of health information and a shared responsibility for information gathering but requires investment from patients and care team to ensure the data are effectively utilized. PRACTICE VALUE: PCD may be useful for team-based care, enabling physicians and non-physician staff to more quickly and responsively connect with patients.