Predictive Modeling of 30-Day Emergency Hospital Transport of Patients Using a Personal Emergency Response System: Prognostic Retrospective Study

Precision emergency medicine

BACKGROUND

Precision health is a burgeoning scientific discipline that aims to incorporate individual variability in biological, behavioral, and social factors to develop personalized health solutions. To date, emergency medicine has not deeply engaged in the precision health movement. However, rapid advances in health technology, data science, and medical informatics offer new opportunities for emergency medicine to realize the promises of precision health.

METHODS

In this article, we conceptualize precision emergency medicine as an emerging paradigm and identify key drivers of its implementation into current and future clinical practice. We acknowledge important obstacles to the specialty-wide adoption of precision emergency medicine and offer solutions that conceive a successful path forward.

RESULTS

Precision emergency medicine is defined as the use of information and technology to deliver acute care effectively, efficiently, and authentically to individual patients and their communities. Key drivers and opportunities include leveraging human data, capitalizing on technology and digital tools, providing deliberate access to care, advancing population health, and reimagining provider education and roles. Overcoming challenges in equity, privacy, and cost is essential for success. We close with a call to action to proactively incorporate precision health into the clinical practice of emergency medicine, the training of future emergency physicians, and the research agenda of the specialty.

CONCLUSIONS

Precision emergency medicine leverages new technology and data-driven artificial intelligence to advance diagnostic testing, individualize patient care plans and therapeutics, and strategically refine the convergence of the health system and the community.

Health information technology to improve care for people with multiple chronic conditions

OBJECTIVE

To review evidence regarding the use of Health Information Technology (health IT) interventions aimed at improving care for people living with multiple chronic conditions (PLWMCC) in order to identify critical knowledge gaps.

DATA SOURCES

We searched MEDLINE, CINAHL, PsycINFO, EMBASE, Compendex, and IEEE Xplore databases for studies published in English between 2010 and 2020.

STUDY DESIGN

We identified studies of health IT interventions for PLWMCC across three domains as follows: self-management support, care coordination, and algorithms to support clinical decision making.

DATA COLLECTION/EXTRACTION METHODS

Structured search queries were created and validated. Abstracts were reviewed iteratively to refine inclusion and exclusion criteria. The search was supplemented by manually searching the bibliographic sections of the included studies. The search included a forward citation search of studies nested within a clinical trial to identify the clinical trial protocol and published clinical trial results. Data were extracted independently by two reviewers.

PRINCIPAL FINDINGS

The search yielded 1907 articles; 44 were included. Nine randomized controlled trials (RCTs) and 35 other studies including quasi-experimental, usability, feasibility, qualitative studies, or development/validation studies of analytic models were included. Five RCTs had positive results, and the remaining four RCTs showed that the interventions had no effect. The studies address individual patient engagement and assess patient-centered outcomes such as quality of life. Few RCTs assess outcomes such as disability and none assess mortality.

CONCLUSIONS

Despite a growing body of literature on health IT interventions or multicomponent interventions including a health IT component for chronic disease management, current evidence for applying health IT solutions to improve care for PLWMCC is limited. The body of literature included in this review provides critical information on the state of the science as well as the many gaps that need to be filled for digital health to fulfill its promise in supporting care delivery that meets the needs of PLWMCC.

Predictive analytics and tailored interventions improve clinical outcomes in older adults: a randomized controlled trial

This study explored the potential to improve clinical outcomes in patients at risk of moving to the top segment of the cost acuity pyramid. This randomized controlled trial evaluated the impact of a Stepped-Care approach (predictive analytics + tailored nurse-driven interventions) on healthcare utilization among 370 older adult patients enrolled in a homecare management program and using a Personal Emergency Response System. The Control group (CG) received care as usual, while the Intervention group (IG) received Stepped-Care during a 180-day intervention period. The primary outcome, decrease in emergency encounters, was not statistically significant (15%, p = 0.291). However, compared to the CG, the IG had significant reductions in total 90-day readmissions (68%, p = 0.007), patients with 90-day readmissions (76%, p = 0.011), total 180-day readmissions (53%, p = 0.020), and EMS encounters (49%, p = 0.006). Predictive analytics combined with tailored interventions could potentially improve clinical outcomes in older adults, supporting population health management in home or community settings.

A randomized trial examining the effect of predictive analytics and tailored interventions on the cost of care

This two-arm randomized controlled trial evaluated the impact of a Stepped-Care intervention (predictive analytics combined with tailored interventions) on the healthcare costs of older adults using a Personal Emergency Response System (PERS). A total of 370 patients aged 65 and over with healthcare costs in the middle segment of the cost pyramid for the fiscal year prior to their enrollment were enrolled for the study. During a 180-day intervention period, control group (CG) received standard care, while intervention group (IG) received the Stepped-Care intervention. The IG had 31% lower annualized inpatient cost per patient compared with the CG (3.7 K, $8.1 K vs. $11.8 K, p = 0.02). Both groups had similar annualized outpatient costs per patient ($6.1 K vs. $5.8 K, p = 0.10). The annualized total cost reduction per patient in the IG vs. CG was 20% (3.5 K, $17.7 K vs. $14.2 K, p = 0.04). Predictive analytics coupled with tailored interventions has great potential to reduce healthcare costs in older adults, thereby supporting population health management in home or community settings.

Predictive Modeling of 30-Day Emergency Hospital Transport of German Patients Using a Personal Emergency Response: Retrospective Study and Comparison with the United States

BACKGROUND

Predictive analytics based on data from remote monitoring of elderly via a personal emergency response system (PERS) in the United States can identify subscribers at high risk for emergency hospital transport. These risk predictions can subsequently be used to proactively target interventions and prevent avoidable, costly health care use. It is, however, unknown if PERS-based risk prediction with targeted interventions could also be applied in the German health care setting.

OBJECTIVE

The objectives were to develop and validate a predictive model of 30-day emergency hospital transport based on data from a German PERS provider and compare the model with our previously published predictive model developed on data from a US PERS provider.

METHODS

Retrospective data of 5805 subscribers to a German PERS service were used to develop and validate an extreme gradient boosting predictive model of 30-day hospital transport, including predictors derived from subscriber demographics, self-reported medical conditions, and a 2-year history of case data. Models were trained on 80% (4644/5805) of the data, and performance was evaluated on an independent test set of 20% (1161/5805). Results were compared with our previously published prediction model developed on a data set of PERS users in the United States.

RESULTS

German PERS subscribers were on average aged 83.6 years, with 64.0% (743/1161) females, with 65.4% (759/1161) reported 3 or more chronic conditions. A total of 1.4% (350/24,847) of subscribers had one or more emergency transports in 30 days in the test set, which was significantly lower compared with the US data set (2455/109,966, 2.2%). Performance of the predictive model of emergency hospital transport, as evaluated by area under the receiver operator characteristic curve (AUC), was 0.749 (95% CI 0.721-0.777), which was similar to the US prediction model (AUC=0.778 [95% CI 0.769-0.788]). The top 1% (12/1161) of predicted high-risk patients were 10.7 times more likely to experience an emergency hospital transport in 30 days than the overall German PERS population. This lift was comparable to a model lift of 11.9 obtained by the US predictive model.

CONCLUSIONS

Despite differences in emergency care use, PERS-based collected subscriber data can be used to predict use outcomes in different international settings. These predictive analytic tools can be used by health care organizations to extend population health management into the home by identifying and delivering timelier targeted interventions to high-risk patients. This could lead to overall improved patient experience, higher quality of care, and more efficient resource use.

Secure training of decision trees with continuous attributes

We apply multiparty computation (MPC) techniques to show, given a database that is secret-shared among multiple mutually distrustful parties, how the parties may obliviously construct a decision tree based on the secret data. We consider data with continuous attributes (i.e., coming from a large domain), and develop a secure version of a learning algorithm similar to the C4.5 or CART algorithms. Previous MPC-based work only focused on decision tree learning with discrete attributes (De Hoogh et al. 2014). Our starting point is to apply an existing generic MPC protocol to a standard decision tree learning algorithm, which we then optimize in several ways. We exploit the fact that even if we allow the data to have continuous values, which a priori might require fixed or floating point representations, the output of the tree learning algorithm only depends on the relative ordering of the data. By obliviously sorting the data we reduce the number of comparisons needed per node to O(N log2 N) from the naive O(N 2), where N is the number of training records in the dataset, thus making the algorithm feasible for larger datasets. This does however introduce a problem when duplicate values occur in the dataset, but we manage to overcome this problem with a relatively cheap subprotocol. We show a procedure to convert a sorting network into a permutation network of smaller complexity, resulting in a round complexity of O(log N) per layer in the tree. We implement our algorithm in the MP-SPDZ framework and benchmark our implementation for both passive and active three-party computation using arithmetic modulo 264. We apply our implementation to a large scale medical dataset of ≈ 290 000 rows using random forests, and thus demonstrate practical feasibility of using MPC for privacy-preserving machine learning based on decision trees for large datasets.

Evaluating the Impact of a Risk Assessment System With Tailored Interventions in Germany: Protocol for a Prospective Study With Matched Controls

BACKGROUND

With a worldwide increase in the elderly population, and an associated increase in health care utilization and costs, preventing avoidable emergency department visits and hospitalizations is becoming a global priority. A personal emergency response system (PERS), consisting of an alarm button and a means to establish a live connection to a response center, can help the elderly live at home longer independently. Individual risk assessment through predictive modeling can help indicate what PERS subscribers are at elevated risk of hospital transport so that early intervention becomes possible.

OBJECTIVE

The aim is to evaluate whether the combination of risk scores determined through predictive modeling and targeted interventions offered by a case manager can result in a reduction of hospital admissions and health care costs for a population of German PERS subscribers. The primary outcome of the study is the difference between the number of hospitalizations in the intervention and matched control groups.

METHODS

As part of the Sicher Zuhause program, an intervention group of 500 PERS subscribers will be tracked for 8 months. During this period, risk scores will be determined daily by a predictive model of hospital transport, and at-risk participants may receive phone calls from a case manager who assesses the health status of the participant and recommends interventions. The health care utilization of the intervention group will be compared to a group of matched controls, retrospectively drawn from a population of PERS subscribers who receive no interventions.

RESULTS

Differences in health care utilization and costs between the intervention group and the matched controls will be determined based on reimbursement records. In addition, qualitative data will be collected on the participants' satisfaction with the Sicher Zuhause program and utilization of the interventions offered as part of the program.

CONCLUSIONS

The study evaluation will offer insight into whether a combination of predictive analytics and case manager-driven interventions can help in avoiding hospital admissions and health care costs for PERS subscribers in Germany living at home independently. In the future, this may lead to improved quality of life and reduced medical costs for the population of the study.

TRIAL REGISTRATION

Deutsches Register Klinischer Studien (DRKS), DRKS00017328; https://www.drks.de/drks_web/navigate.do?navigationId=trial.HTML&TRIAL_ID=DRKS00017328.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

DERR1-10.2196/17584.

Home-based monitoring of falls using wearable sensors in Parkinson's disease

INTRODUCTION

Falling is among the most serious clinical problems in Parkinson's disease (PD). We used body-worn sensors (falls detector worn as a necklace) to quantify the hazard ratio of falls in PD patients in real life.

METHODS

We matched all 2063 elderly individuals with self-reported PD to 2063 elderly individuals without PD based on age, gender, comorbidity, and living conditions. We analyzed fall events collected at home via a wearable sensor. Fall events were collected either automatically using the wearable falls detector or were registered by a button push on the same device. We extracted fall events from a 2.5-year window, with an average follow-up of 1.1 years. All falls included were confirmed immediately by a subsequent telephone call. The outcomes evaluated were (1) incidence rate of any fall, (2) incidence rate of a new fall after enrollment (ie, hazard ratio), and (3) 1-year cumulative incidence of falling.

RESULTS

The incidence rate of any fall was higher among self-reported PD patients than controls (2.1 vs. 0.7 falls/person, respectively; P < .0001). The incidence rate of a new fall after enrollment (ie, hazard ratio) was 1.8 times higher for self-reported PD patients than controls (95% confidence interval, 1.6-2.0).

CONCLUSION

Having PD nearly doubles the incidence of falling in real life. These findings highlight PD as a prime "falling disease." The results also point to the feasibility of using body-worn sensors to monitor falls in daily life. © 2019 The Authors. Movement Disorders published by Wiley Periodicals, Inc. on behalf of International Parkinson and Movement Disorder Society.