Prediction of Drug-Related Risks Using Clinical Context Information in Longitudinal Claims Data

Prediction of Medical Disputes Between Health Care Workers and Patients in Terms of Hospital Legal Construction Using Machine Learning Techniques: Externally Validated Cross-Sectional Study

BACKGROUND

Medical disputes are a global public health issue that is receiving increasing attention. However, studies investigating the relationship between hospital legal construction and medical disputes are scarce. The development of a multicenter model incorporating machine learning (ML) techniques for the individualized prediction of medical disputes would be beneficial for medical workers.

OBJECTIVE

This study aimed to identify predictors related to medical disputes from the perspective of hospital legal construction and the use of ML techniques to build models for predicting the risk of medical disputes.

METHODS

This study enrolled 38,053 medical workers from 130 tertiary hospitals in Hunan province, China. The participants were randomly divided into a training cohort (34,286/38,053, 90.1%) and an internal validation cohort (3767/38,053, 9.9%). Medical workers from 87 tertiary hospitals in Beijing were included in an external validation cohort (26,285/26,285, 100%). This study used logistic regression and 5 ML techniques: decision tree, random forest, support vector machine, gradient boosting decision tree (GBDT), and deep neural network. In total, 12 metrics, including discrimination and calibration, were used for performance evaluation. A scoring system was developed to select the optimal model. Shapley additive explanations was used to generate the importance coefficients for characteristics. To promote the clinical practice of our proposed optimal model, reclassification of patients was performed, and a web-based app for medical dispute prediction was created, which can be easily accessed by the public.

RESULTS

Medical disputes occurred among 46.06% (17,527/38,053) of the medical workers in Hunan province, China. Among the 26 clinical characteristics, multivariate analysis demonstrated that 18 characteristics were significantly associated with medical disputes, and these characteristics were used for ML model development. Among the ML techniques, GBDT was identified as the optimal model, demonstrating the lowest Brier score (0.205), highest area under the receiver operating characteristic curve (0.738, 95% CI 0.722-0.754), and the largest discrimination slope (0.172) and Youden index (1.355). In addition, it achieved the highest metrics score (63 points), followed by deep neural network (46 points) and random forest (45 points), in the internal validation set. In the external validation set, GBDT still performed comparably, achieving the second highest metrics score (52 points). The high-risk group had more than twice the odds of experiencing medical disputes compared with the low-risk group.

CONCLUSIONS

We established a prediction model to stratify medical workers into different risk groups for encountering medical disputes. Among the 5 ML models, GBDT demonstrated the optimal comprehensive performance and was used to construct the web-based app. Our proposed model can serve as a useful tool for identifying medical workers at high risk of medical disputes. We believe that preventive strategies should be implemented for the high-risk group.

Haefeli W, M. Seidling H. Mortality and Hospitalizations Among Patients Enrolled in an Interprofessional Medication Management Program

BACKGROUND

Measures for improving medication safety in outpatient care are often complex and involve medication reviews. Over the period 2016-2022 (with a preceeding one-year pilot phase), an interprofessional medication management program- the Medicines Initiative Saxony-Thuringia (Arzneimittelinitiative Sachsen-Thüringen, ARMIN)-was implemented in two German federal states. More than 5000 patients received a medication review by the end of 2019 by a team composed of physicians and pharmacists and were provided with joint, continuous care thereafter.

METHODS

In the framework of a retrospectively registered cohort study, the mortality and hospitalizations of this population (5033 patients) were studied using routine data from a statutory health insurer (observation period 2015-2019) and compared with those of a control group (10 039 patients) determined from the routine data by propensity score matching. Mortality was compared by survival analysis (Cox regression), and hospitalization rates were compared in terms of event probabilities within two years of enrollment in the medication management program. Robustness was tested in multiple sensitivity analyses.

RESULTS

Over the observation period, 9.3% of the ARMIN participants and 12.9% of persons in the control group died (hazard ratio of the adjusted Cox regression, 0.84; 95% confidence interval [0.76; 0.94], P = 0.001). In the first two years after inclusion, the ARMIN participants were hospitalized just as often as the persons in the control group (52.4% versus 53.4%; odds ratio from the adjusted model, 1.04 [0.96; 1.11], P = 0.347). The effects were consistent in sensitivity analyses.

CONCLUSION

In this retrospective cohort study, participation in the ARMIN program was associated with a lower risk of death. Exploratory analyses provide clues to the potential origin of this association.

Predicting Hospital Readmissions from Health Insurance Claims Data: A Modeling Study Targeting Potentially Inappropriate Prescribing

BACKGROUND

 Numerous prediction models for readmissions are developed from hospital data whose predictor variables are based on specific data fields that are often not transferable to other settings. In contrast, routine data from statutory health insurances (in Germany) are highly standardized, ubiquitously available, and would thus allow for automatic identification of readmission risks.

OBJECTIVES

 To develop and internally validate prediction models for readmissions based on potentially inappropriate prescribing (PIP) in six diseases from routine data.

METHODS

 In a large database of German statutory health insurance claims, we detected disease-specific readmissions after index admissions for acute myocardial infarction (AMI), heart failure (HF), a composite of stroke, transient ischemic attack or atrial fibrillation (S/AF), chronic obstructive pulmonary disease (COPD), type-2 diabetes mellitus (DM), and osteoporosis (OS). PIP at the index admission was determined by the STOPP/START criteria (Screening Tool of Older Persons' Prescriptions/Screening Tool to Alert doctors to the Right Treatment) which were candidate variables in regularized prediction models for specific readmission within 90 days. The risks from disease-specific models were combined ("stacked") to predict all-cause readmission within 90 days. Validation performance was measured by the c-statistics.

RESULTS

 While the prevalence of START criteria was higher than for STOPP criteria, more single STOPP criteria were selected into models for specific readmissions. Performance in validation samples was the highest for DM (c-statistics: 0.68 [95% confidence interval (CI): 0.66-0.70]), followed by COPD (c-statistics: 0.65 [95% CI: 0.64-0.67]), S/AF (c-statistics: 0.65 [95% CI: 0.63-0.66]), HF (c-statistics: 0.61 [95% CI: 0.60-0.62]), AMI (c-statistics: 0.58 [95% CI: 0.56-0.60]), and OS (c-statistics: 0.51 [95% CI: 0.47-0.56]). Integrating risks from disease-specific models to a combined model for all-cause readmission yielded a c-statistics of 0.63 [95% CI: 0.63-0.64].

CONCLUSION

 PIP successfully predicted readmissions for most diseases, opening the possibility for interventions to improve these modifiable risk factors. Machine-learning methods appear promising for future modeling of PIP predictors in complex older patients with many underlying diseases.

Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validity

OBJECTIVE

To explore factors that potentially impact external validation performance while developing and validating a prognostic model for hospital admissions (HAs) in complex older general practice patients.

STUDY DESIGN AND SETTING

Using individual participant data from four cluster-randomised trials conducted in the Netherlands and Germany, we used logistic regression to develop a prognostic model to predict all-cause HAs within a 6-month follow-up period. A stratified intercept was used to account for heterogeneity in baseline risk between the studies. The model was validated both internally and by using internal-external cross-validation (IECV).

RESULTS

Prior HAs, physical components of the health-related quality of life comorbidity index, and medication-related variables were used in the final model. While achieving moderate discriminatory performance, internal bootstrap validation revealed a pronounced risk of overfitting. The results of the IECV, in which calibration was highly variable even after accounting for between-study heterogeneity, agreed with this finding. Heterogeneity was equally reflected in differing baseline risk, predictor effects and absolute risk predictions.

CONCLUSIONS

Predictor effect heterogeneity and differing baseline risk can explain the limited external performance of HA prediction models. With such drivers known, model adjustments in external validation settings (eg, intercept recalibration, complete updating) can be applied more purposefully.

TRIAL REGISTRATION NUMBER

PROSPERO id: CRD42018088129.

Disease-dependent variations in the timing and causes of readmissions in Germany: A claims data analysis for six different conditions

BACKGROUND

Hospital readmissions place a major burden on patients and health care systems worldwide, but little is known about patterns and timing of readmissions in Germany.

METHODS

We used German health insurance claims (AOK, 2011-2016) of patients ≥ 65 years hospitalized for acute myocardial infarction (AMI), heart failure (HF), a composite of stroke, transient ischemic attack, or atrial fibrillation (S/AF), chronic obstructive pulmonary disease (COPD), type 2 diabetes mellitus, or osteoporosis to identify hospital readmissions within 30 or 90 days. Readmissions were classified into all-cause, specific, and non-specific and their characteristics were analyzed.

RESULTS

Within 30 and 90 days, about 14-22% and 27-41% index admissions were readmitted for any reason, respectively. HF and S/AF contributed most index cases, and HF and COPD accounted for most all-cause readmissions. Distributions and ratios of specific to non-specific readmissions were disease-specific with highest specific readmissions rates among COPD and AMI.

CONCLUSION

German claims are well-suited to investigate readmission causes if longer periods than 30 days are evaluated. Conditions closely related with the primary disease are the most frequent readmission causes, but multiple comorbidities among readmitted cases suggest that a multidisciplinary care approach should be implemented vigorously addressing comorbidities already during the index hospitalization.

AgeWell.de - study protocol of a pragmatic multi-center cluster-randomized controlled prevention trial against cognitive decline in older primary care patients

BACKGROUND

In the absence of treatment options, the WHO emphasizes the identification of effective prevention strategies as a key element to counteract the dementia epidemic. Regarding the complex nature of dementia, trials simultaneously targeting multiple risk factors should be particularly effective for prevention. So far, however, only few such multi-component trials have been launched, but yielding promising results. In Germany, comparable initiatives are lacking, and translation of these complex interventions into routine care was not yet done. Therefore, AgeWell.de will be conducted as the first multi-component prevention trial in Germany which is closely linked to the primary care setting.

METHODS

AgeWell.de will be designed as a multi-centric, cluster-randomized controlled multi-component prevention trial. Participants will be older community-dwelling general practitioner (GP) patients (60-77 years; n = 1,152) with increased dementia risk according to CAIDE (Cardiovascular Risk Factors, Aging, and Incidence of Dementia) Dementia Risk Score. Recruitment will take place at 5 study sites across Germany. GP practices will be randomized to either intervention A (advanced) or B (basic). GPs will be blinded to their respective group assignment, as will be the statistician conducting the randomization. The multi-component intervention (A) includes nutritional counseling, physical activity, cognitive training, optimization of medication, management of vascular risk factors, social activity, and, if necessary, further specific interventions targeting grief and depression. Intervention B includes general health advice on the intervention components and GP treatment as usual. We hypothesize that over the 2-year follow-up period the intervention group A will benefit significantly from the intervention program in terms of preserved cognitive function/delayed cognitive decline (primary outcome), and other relevant (secondary) outcomes (e.g. quality of life, social activities, depressive symptomatology, cost-effectiveness).

DISCUSSION

AgeWell.de will be the first multi-component trial targeting risk of cognitive decline in older adults in Germany. Compared to previous trials, AgeWell.de covers an even broader set of interventions suggested to be beneficial for the intended outcomes. The findings will add substantial knowledge on modifiable lifestyle factors to prevent or delay cognitive decline.

TRIAL REGISTRATION

German Clinical Trials Register (reference number: DRKS00013555 ).

A Systematic Review of Studies of the STOPP/START 2015 and American Geriatric Society Beers 2015 Criteria in Patients ≥ 65 Years

BACKGROUND

Polypharmacy remains problematic for individuals ≥65.

OBJECTIVE

To summarise the percentages of patients meeting 2015 STOPP criteria for Potentially Inappropriate Prescriptions (PIPs), 2015 Beers criteria for Potentially Inappropriate Medications (PIMs), and START criteria Potential Prescribing Omissions (PPOs).

METHODS

Searches conducted on 2 January 2019 in Medline, Embase, and PubMed identified 562 studies and 62 studies were retained for review. Data were abstracted independently.

RESULTS

62 studies (n=1,854,698) included two RCTs and 60 non-randomised studies. For thirty STOPP/START studies (n=1,245,974) average percentages for ≥1 PIP weighted by study size were 42.8% for 1,242,010 community patients and 51.8% for 3,964 hospitalised patients. For nineteen Beers studies (n = 595,811) the average percentages for ≥1 PIM were 58% for 593,389 community patients and 55.5% for 2,422 hospitalised patients. For thirteen studies (n=12,913) assessing both STOPP/START and Beers criteria the average percentages for ≥1 STOPP PIP were 33.9% and Beers PIMs 46.8% for 8,238 community patients, and for ≥ 1 STOPP PIP were 42.4% and for ≥1 Beers PIM 60.5% for 4,675 hospitalised patients. Only ten studies assessed changes over time and eight found positive changes.

CONCLUSION

PIP/PIM/PPO rates are high in community and hospitalised patients in many countries. RCTs are needed for interventions to: reduce new/existing PIPs/PIMs/PPO prescriptions, reduce prescriptions causing adverse effects, and enable regulatory authorities to monitor and reduce inappropriate prescriptions in real time. Substantial differences between Beers and STOPP/START assessments need to be investigated whether they are due to the criteria, differential medication availability between countries, or data availability to assess the criteria.