Specialized Versus Generic Allied Health Therapy and the Risk of Parkinson's Disease Complications

BACKGROUND

Specialized versus generic physiotherapy (PT) reduces Parkinson's disease (PD)-related complications. It is unclear (1) whether other specialized allied heath disciplines, including occupational therapy (OT) and speech and language therapy (S&LT), also reduce complications; (2) whether there is a synergistic effect among multiple specialized disciplines; and (3) whether each allied health discipline prevents specific complications.

OBJECTIVES

To longitudinally assessed whether the level of expertise (specialized vs. generic training) of PT, OT, and S&LT was associated with the incidence rate of PD-related complications.

METHODS

We used claims data of all insured persons with PD in the Netherlands between January 1, 2010, and December 31, 2018. ParkinsonNet-trained therapists were classified as specialized, and other therapists as generic. We used mixed-effects Poisson regression models to estimate rate ratios adjusting for sociodemographic and clinical characteristics.

RESULTS

The population of 51,464 persons with PD (mean age, 72.4 years; standard deviation 9.8) sustained 10,525 PD-related complications during follow-up (median 3.3 years). Specialized PT was associated with fewer complications (incidence rate ratio [IRR] of specialized versus generic = 0.79; 95% confidence interval, [0.74-0.83]; P < 0.0001), as was specialized OT (IRR = 0.88 [0.77-0.99]; P = 0.03). We found a trend of an association between specialized S&LT and a lower rate of PD-related complications (IRR = 0.88 [0.74-1.04]; P = 0.18). The inverse association of specialized OT persisted in the stratum, which also received specialized PT (IRR = 0.62 [0.42-0.90]; P = 0.001). The strongest inverse association of PT was seen with orthopedic injuries (IRR = 0.78 [0.73-0.82]; P < 0.0001) and of S&LT with pneumonia (IRR = 0.70 [0.53-0.93]; P = 0.03).

CONCLUSIONS

These findings support a wider introduction of specialized allied health therapy expertise in PD care and conceivably for other medical conditions. © 2022 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Mortality Prediction of Patients With Cardiovascular Disease Using Medical Claims Data Under Artificial Intelligence Architectures: Validation Study

Background

Cardiovascular disease (CVD) is the greatest health problem in Australia, which kills more people than any other disease and incurs enormous costs for the health care system. In this study, we present a benchmark comparison of various artificial intelligence (AI) architectures for predicting the mortality rate of patients with CVD using structured medical claims data. Compared with other research in the clinical literature, our models are more efficient because we use a smaller number of features, and this study could help health professionals accurately choose AI models to predict mortality among patients with CVD using only claims data before a clinic visit.

Objective

This study aims to support health clinicians in accurately predicting mortality among patients with CVD using only claims data before a clinic visit.

Methods

The data set was obtained from the Medicare Benefits Scheme and Pharmaceutical Benefits Scheme service information in the period between 2004 and 2014, released by the Department of Health Australia in 2016. It included 346,201 records, corresponding to 346,201 patients. A total of five AI algorithms, including four classical machine learning algorithms (logistic regression [LR], random forest [RF], extra trees [ET], and gradient boosting trees [GBT]) and a deep learning algorithm, which is a densely connected neural network (DNN), were developed and compared in this study. In addition, because of the minority of deceased patients in the data set, a separate experiment using the Synthetic Minority Oversampling Technique (SMOTE) was conducted to enrich the data.

Results

Regarding model performance, in terms of discrimination, GBT and RF were the models with the highest area under the receiver operating characteristic curve (97.8% and 97.7%, respectively), followed by ET (96.8%) and LR (96.4%), whereas DNN was the least discriminative (95.3%). In terms of reliability, LR predictions were the least calibrated compared with the other four algorithms. In this study, despite increasing the training time, SMOTE was proven to further improve the model performance of LR, whereas other algorithms, especially GBT and DNN, worked well with class imbalanced data.

Conclusions

Compared with other research in the clinical literature involving AI models using claims data to predict patient health outcomes, our models are more efficient because we use a smaller number of features but still achieve high performance. This study could help health professionals accurately choose AI models to predict mortality among patients with CVD using only claims data before a clinic visit.

Population-based incidence of diabetic ketoacidosis in type 2 diabetes: medical claims data analysis in Japan

PURPOSE

Determination of the incidence rate of diabetic ketoacidosis (DKA) in type 2 diabetes mellitus (T2DM) is urgent, in response to the safety issue with use of the glucose lowering drugs, sodium-glucose cotransporter-2 inhibitors, and DKA.

METHODS

We extracted data of adult patients with T2DM from a medical claims database in Japan, which included 1 million individuals. The study period was 2005-2013, before the era of sodium-glucose cotransporter-2 inhibitors. The inclusion criteria were patients with a diagnosis of T2DM who had at least 1 prescription for a glucose-lowering drug. We further examined the number of DKA-related admissions and medication use prior to admission. Calculation of DKA incidence was on a patient-years basis, with 95% confidence interval using a Poisson distribution.

RESULTS

Of 36,674 adult patients with T2DM, we identified 74 cases admitted for DKA, with an estimated DKA incidence of 0.48/1000 (95% confidence interval, 0.38-0.60/1000) patient-years. Of these 74 cases, approximately two-thirds of patients had no past or recent history of glucose-lowering drug use before admission for DKA.

CONCLUSIONS

This study estimated that the incidence of DKA was 0.48/1000 patient-years in Japan, similar to previous studies. We also found that underuse of glucose-lowering medication was common among patients hospitalized for DKA. KEY POINTS The population-based incidence of DKA in T2DM is largely unknown. In Japan, the estimated DKA incidence was 0.48/1000 (95% CI, 0.38-0.60/1000) patient-years. Drug-induced DKA is an emerging research area, and our results can be applied to evaluating the risk of DKA.

Surveillance rates for hepatocellular carcinoma among patients with cirrhosis, chronic hepatitis B, and chronic hepatitis C based on Japanese claims database

AIM

Surveillance is recommended for patients with risk of hepatocellular carcinoma (HCC) such as viral hepatitis and liver cirrhosis (LC). Although populations are at higher risk in Asia, surveillance rates have not been evaluated in those countries. Therefore, we aimed to examine surveillance rates for HCC and to compare predictors for surveillance among each type of liver disease in Japan.

METHODS

We carried out an observational study using the medical claims data of Japan. A total of 4713 patients who were diagnosed with LC, hepatitis B virus (HBV), or hepatitis C virus (HCV) infection were followed for 12 months. We described surveillance rates in patients with HBV or HCV infection or non-viral LC and evaluated potential predictors for follow-up surveillance among them by using multivariable logistic regression analysis, in which we set explanatory variables as sex, age, medical facility, HBV or HCV infection, and LC.

RESULTS

The HCC surveillance rate during follow-up was 42.3% among patients with chronic viral hepatitis, 56.5% with viral cirrhosis and 26.0%, with non-viral LC. Significant predictors for follow-up surveillance were presence of LC and hospital visits. Hepatitis B virus-infected patients aged <50 years and HCV-infected patients aged ≥50 years were more likely to undergo exams, that is, different tendencies between HBV and HCV patients were shown.

CONCLUSIONS

Follow-up surveillance for HCC was not adequately carried out among patients at risk for HCC in Japan, despite of recommendation in guidelines. In future, methodological research might be needed to increase surveillance rates for HCC in patients with chronic hepatitis or those who received treatment at a clinic.

Use of statins offsets insulin-related cancer risk

AIM

There is firm evidence of a relation between type 2 diabetes (T2DM) and increased risks of cancer at various sites, but it is still unclear how different antihyperglycaemic therapies modify site-specific cancer risks. The aim of this study was to provide a complete characterization of all possible associations between individual T2DM therapies, statin use and site-specific cancers in the Austrian population.

METHODS

Medical claims data of 1 847 051 patients with hospital stays during 2006-2007 were used to estimate age- and sex-dependent co-occurrences of site-specific cancer diagnoses and treatment with specific glucose-lowering drugs and statins.

RESULTS

Patients treated with insulin or insulin secretagogues showed up to ninefold increased risks for cancers of the colon [males only (m)], liver (m), pancreas, lung (m) and brain (m), as well as a strongly decreased risk for prostate cancer (m). In patients taking statins, the risks were generally decreased, with a greater risk reduction in patients not receiving antihyperglycaemic therapies. The strongest effects were observed for use of insulin and pancreatic cancer [m: OR 4.5, 95% CI: 3.1-6.6; females (f): OR 4.2, 95% CI: 2.5-7.1], sulfonylureas (m: OR 2.8, 95% CI: 1.7-4.6; f: OR 3.0, 95% CI: 2.1-4.2) or glitazones and skin cancer (f: OR 0.54, 95% CI: 0.36-0.80), as well as metformin and cancer of the prostate (m: OR 0.82, 95% CI: 0.75-0.91) and corpus uteri (f: OR 1.7, 95% CI: 1.4-2.0) and non-Hodgkin's lymphoma (f: OR 0.76, 95% CI: 0.64-0.91).

CONCLUSIONS

The use of statins offsets insulin-related cancer risks in patients with diabetes independently of sex and age. Overall, our data support the hyperglycaemia-cancer hypothesis. A reduction in endogenous or exogenous hyperinsulinaemia may be beneficial for cancer prevention. Therefore, insulin-sparing and insulin-sensitizing drugs should be the preferred treatment choices.