Ole K. Hejlesen

Screening for diabetic retinopathy using computer based image analysis and statistical classification

Diabetic retinopathy is one of the most common causes of blindness in Europe. However, efficient therapies do exist. An accurate and early diagnosis and correct application of treatment can prevent blindness in more than 50% of all cases. Digital imaging is becoming available as a means of screening for diabetic retinopathy. As well as providing a high quality permanent record of the retinal appearance, which can be used for monitoring of progression or response to treatment, and which can be reviewed by an ophthalmologist, digital images have the potential to be processed by automatic analysis systems. We have described the preliminary development of a tool to provide automatic analysis of digital images taken as part of routine monitoring of diabetic retinopathy in our clinic. Various statistical classifiers, a Bayesian, a Mahalanobis, and a KNN classifier were tested. The system was tested on 134 retinal images. The Mahalanobis classifier had the best results: microaneurysms, haemorrhages, exudates, and cotton wool spots were detected with a sensitivity of 69, 83, 99, and 80%, respectively.

DiasNet—a diabetes advisory system for communication and education via the internet

Intensive diabetes treatment can lead to a substantial reduction of the rate of the complications associated with diabetes. However, a number of patients may have poor control despite specialist care, and this along with devolution of care to non-specialists suggests that alternative interventions should be developed. The present paper describes an Internet based system where more emphasis is put on patient empowerment, the keywords being education and communication. The DiasNet system is based on a well documented decision support system, Dias, designed for use by clinicians. The scope of DiasNet has been widened from being used by clinicians to give advice on insulin dose, to also being used by patients as a tool for education and communication. Patients can experiment with their own data, adjusting insulin doses or meal sizes. In this way different therapeutic and dietary alternatives can be tried out, allowing the patient to gain experience in achieving glycaemic control. DiasNet is implemented in JAVA according to the client/server principle, enabling a new way of communication between patient and clinician: in case of any problems, the patient simply phones the clinician, who immediately, using his or her office PC, can take a look at the data the patient has entered.

Using preventive home monitoring to reduce hospital admission rates and reduce costs: a case study of telehealth among chronic obstructive pulmonary disease patients

We studied whether preventive home monitoring of patients with chronic obstructive pulmonary disease (COPD) could reduce the frequency of hospital admissions and lower the cost of hospitalization. Patients were recruited from a health centre, general practitioner (GP) or the pulmonary hospital ward. They were randomized to usual care or tele-rehabilitation with a telehealth monitoring device installed in their home for four months. A total of 111 patients were suitable for inclusion and consented to be randomized: 60 patients were allocated to intervention and three were lost to follow-up. In the control group 51 patients were allocated to usual care and three patients were lost to follow-up. In the tele-rehabilitation group, the mean hospital admission rate was 0.49 per patient per 10 months compared to the control group rate of 1.17; this difference was significant (P = 0.041). The mean cost of admissions was €3461 per patient in the intervention group and €4576 in the control group; this difference was not significant. The Kaplan-Meier estimates for time to hospital admission were longer for the intervention group than the controls, but the difference was not significant. Future work requires large-scale studies of prolonged home monitoring with more extended follow-up.

DIAS—the diabetes advisory system: an outline of the system and the evaluation results obtained so far

The present paper gives a description of the Diabetes Advisory System (DIAS), and the evaluation results obtained so far. DIAS is a decision support system for the management of insulin dependent diabetes. The core of the system is a compartment model of the human carbohydrate metabolism implemented as a causal probabilistic network (CPN or Bayesian network), which gives it the ability to handle the uncertainty, for example, in blood glucose measurements or physiological variations in glucose metabolism. The evaluation results suggest that, at least in our hands, DIAS can generate advice that is safe and of a quality that is at least comparable to what is available from experienced clinicians.

Moving prediction of exacerbation in chronic obstructive pulmonary disease for patients in telecare

We investigated whether physiological data can be used for predicting chronic obstructive pulmonary disease (COPD) exacerbations. Home measurements from 57 patients were analysed, during which 10 exacerbations occurred in nine patients. A total of 273 different features were evaluated for their discrimination abilities between periods with and without exacerbations. The analysis showed that if a sensitivity level of 70% is considered to be acceptable, then the specificity was 95% and the AUC was 0.73, i.e. it is possible to discriminate between periods of exacerbation and periods without. A system capable of predicting risk could provide support to COPD patients in their tele-rehabilitation.

Preliminary experience of the DIAS computer model in providing insulin dose advice to patients with insulin dependent diabetes

The Diabetes Advisory System (DIAS) is a model of human glucose metabolism which predicts hourly blood glucose concentrations and provides advice on insulin dose. Its ability to provide appropriate advice was assessed in 20 well-controlled IDDM patients (mean (SD) age 38 (11), duration 17 (9) years; HbA1 8.8 (0.9)%, reference range 5.4-7.6%). Patients recorded blood glucose measurements, insulin dose and food intake for 4 days. These data were used to generate insulin dose advice by both DIAS and a diabetes specialist nurse. Patients were then allocated to follow either DIAS or nurse advice for a further 4 days. There was no significant difference in mean recorded blood glucose values or frequency of reported hypoglycaemia between the DIAS and nurse groups either before or after insulin dose adjustment. The DIAS model, however, generated significantly lower insulin dose advice than the nurse (median (range)% change in insulin dose: DIAS group -13.3% (-25.0 to +11.6) versus nurse group 0% (-8.7 to +2.5), P < 0.05). We conclude that, in the patients studied, DIAS provided insulin dose advice which maintained good short term control of diabetes, despite significant reductions in dose in some cases.

Effectiveness and cost-effectiveness of telehealthcare for chronic obstructive pulmonary disease: study protocol for a cluster randomized controlled trial

BackgroundSeveral feasibility studies show promising results of telehealthcare on health outcomes and health-related quality of life for patients suffering from chronic obstructive pulmonary disease, and some of these studies show that telehealthcare may even lower healthcare costs. However, the only large-scale trial we have so far - the Whole System Demonstrator Project in England - has raised doubts about these results since it conclude that telehealthcare as a supplement to usual care is not likely to be cost-effective compared with usual care alone.Methods/DesignThe present study is known as ‘TeleCare North’ in Denmark. It seeks to address these doubts by implementing a large-scale, pragmatic, cluster-randomized trial with nested economic evaluation. The purpose of the study is to assess the effectiveness and the cost-effectiveness of a telehealth solution for patients suffering from chronic obstructive pulmonary disease compared to usual practice. General practitioners will be responsible for recruiting eligible participants (1,200 participants are expected) for the trial in the geographical area of the North Denmark Region. Twenty-six municipality districts in the region define the randomization clusters. The primary outcomes are changes in health-related quality of life, and the incremental cost-effectiveness ratio measured from baseline to follow-up at 12 months. Secondary outcomes are changes in mortality and physiological indicators (diastolic and systolic blood pressure, pulse, oxygen saturation, and weight).DiscussionThere has been a call for large-scale clinical trials with rigorous cost-effectiveness assessments in telehealthcare research. This study is meant to improve the international evidence base for the effectiveness and cost-effectiveness of telehealthcare to patients suffering from chronic obstructive pulmonary disease by implementing a large-scale pragmatic cluster-randomized clinical trial.Trial registrationClinicaltrials.gov, http://NCT01984840, November 14, 2013.

Use of the DIAS model to predict unrecognised hypoglycaemia in patients with insulin-dependent diabetes

The Diabetes Advisory System (DIAS) is a model of human glucose metabolism implemented in a causal probabilistic network. It handles data on insulin dose, carbohydrate intake and blood glucose concentration to predict hourly blood glucose concentrations and thus provide an indication of blood glucose values between home blood tests. DIAS was used to predict blood glucose profiles in eight patients with well-controlled insulin-dependent diabetes, who are at increased risk of hypoglycaemia (abnormally low blood glucose levels). DIAS predicted nocturnal hypoglycaemia in six patients and daytime hypoglycaemia in one patient. The occurrence of nocturnal hypoglycaemia was not recognised by the patient or suspected by their doctor but was subsequently confirmed by blood testing in five patients. It is known that unrecognised nocturnal hypoglycaemia is common in patients with well-controlled diabetes. The ability of DIAS to identify such periods of hypoglycaemia with reasonable accuracy illustrates how the advanced technology it employs may provide reliable decision support to clinicians.

Cost-Utility Analysis of a Telerehabilitation Program: A Case Study of COPD Patients

OBJECTIVE The present study seeks to conduct cost-utility analysis (CUA) of the Danish TELEKAT (Telehomecare, Chronic Patients and the Integrated Healthcare System) project. The TELEKAT project seeks to test and develop a preventive home monitoring concept across sectors for chronic obstructive pulmonary disease (COPD) patients. The concept of the TELEKAT project is to reduce admissions by enabling the COPD patients to conduct self-monitoring and maintain rehabilitation activities in their own home. COPD patients with severe and very severe COPD were included in the study. SUBJECTS AND METHODS This economic evaluation follows international guidelines for the conduction of a CUA alongside a clinical randomized controlled trial. The analysis is based on a health sector perspective. RESULTS The mean incremental cost efficiency ratio, located in the southeast quadrant, shows that telerehabilitation is less costly and more effective than the rehabilitation given to the control group. The telerehabilitation program produces more value for money and generates savings on healthcare budgets. CONCLUSIONS The telerehabilitation program appears to be more cost-effective than the conventional rehabilitation program for COPD patients. Further studies of cost-effectiveness with a focus on large-scale studies are needed.

A Novel Algorithm for Prediction and Detection of Hypoglycemia Based on Continuous Glucose Monitoring and Heart Rate Variability in Patients With Type 1 Diabetes

Background: Hypoglycemia is a common and serious side effect of insulin therapy in patients with diabetes. Early detection and prediction of hypoglycemia may improve treatment and avoidance of serious complications. Continuous glucose monitoring (CGM) has previously been used for detection of hypoglycemia, but with a modest accuracy. Therefore, our aim was to investigate whether a novel algorithm that adds information of the complex dynamic/pattern of heart rate variability (HRV) could improve the accuracy of hypoglycemia as detected by a CGM device. Methods: Data from 10 patients with type 1 diabetes studied during insulin-induced hypoglycemia were obtained. Blood glucose samples were used as reference. HRV patterns and CGM data were combined in a mathematical prediction algorithm. Detection of hypoglycemic periods, performed by the algorithm, was treated as a pattern recognition problem and features/patterns derived from HRV and CGM prior to each blood glucose sample were used to decide if that particular point in time was below the hypoglycemic threshold of 3.9 mmol/L. Results: A total of 903 samples were analyzed by the novel algorithm, which yielded a sensitivity of 79% and a specificity of 99%. The algorithm was able to detect 16/16 hypoglycemic events with no false positives and had a lead time of 22 minutes as compared to the CGM device. Conclusions: Detection accuracy and lead time were significantly improved by the novel algorithm compared to that of CGM alone.