2014 10-22-can computerized alerts help in medical practice
•Download as PPTX, PDF•
1 like•158 views
This document summarizes a presentation on using computerized alerts to help detect malnutrition in hospital patients. An alert system was developed using .NET and C# that integrated data from different systems using standards like LOINC, ICD-10, and HL7. The study found the alert system helped significantly, increasing detection of malnutrition risk by 14% compared to the control group without alerts. While alerts can help, it is important to avoid alert fatigue from too many notifications.
Recommended
More Related Content
Viewers also liked
Viewers also liked (20)
Similar to 2014 10-22-can computerized alerts help in medical practice
Similar to 2014 10-22-can computerized alerts help in medical practice (20)
Recently uploaded
Recently uploaded (20)
2014 10-22-can computerized alerts help in medical practice
- 1. APAMI 2014 8th Asia Pacific Association for Medical Informatics Conference New Delhi, India Can computerized Alerts Help in Medical Practice? Dr. Humberto F. Mandirola Brieux Diego Kaminker HL7 ARGENTINA, HOSPITAL ITALIANO, HOSPITAL BELGRANO www.hl7latam.org 1
- 2. AGENDA Introduction. Malnutrition CPOE Research. www.hl7latam.org 2
- 3. 3 Acute Hospital of medium level 200 Beds 8 Critical care unit beds 250 Physicians latitud 34°33'0.54"S longitud 58°31'32.33"O
- 4. INTRODUCTION Focus on nutritional risk detection. An alert software was used. Malnutrition often goes unnoticed. www.hl7latam.org 4
- 5. MALNUTRITION PREVALENCE Malnutrition is a serious concern. The prevalence of malnutrition problems is higher than believed. It’s a serious underestimated problem. The prevalence is 30%, out of which 70% goes undiagnosed.. www.hl7latam.org 5
- 6. MALNUTRITION FACTORS Factors which have negative impact on nutrition: Difficulties in monitoring weight. Patients do not like hospital food. They may lose appetite. The disease itself consumes energy. Doctors may neglect the nutritional aspect. They focus mainly on the disease and studies, and not on nutritional aspects. www.hl7latam.org 6
- 7. History Project Our aim to diminish the malnourishment rate. Nutritional aspects assessed by doctors. Multidisciplinary approach. Creation of an automatic alarm. www.hl7tatam.org 7
- 8. CPOE Positive effects delay in order completion and processing duplicate orders, overdoses, allergic reactions and drug interactions. Negative effects Alert fatigue www.hl7latam.org 8
- 9. THE ALERT ENGINE The alert engine in .NET and in C#. We used different standards. LOINC to laboratory data ICD10 to medical record data. HL7 2.4 to transfer data. www.hl7latam.org 9
- 10. Alert group www.hl7latam.org 10 Control group SAMPLES Both groups Total 200 200 400 Age max 96 90 96 Age min 34 22 22 Age average 65,25 56,00 60,63
- 11. www.hl7latam.org 11 WORK FLOW
- 12. HOW THE ALERT ENGINE WORKS LIS EMR CLINCIAL ALTERT ENGINE www.hl7latam.org 12 HL7 INTERFACE
- 13. www.hl7latam.org 13 RESULTS Alert Control Total 198 197 Undetected nutritional risk 130 157 Detected nutritional risk 68 40 (p<0.001) 34,3 % 20,3 %
- 14. CONCLUSION The nutritional risk alert we built was of significant help. Detection of malnutrition in patients increased by 14%. It is very important to use alerts. Physicians using alerts could diagnose more problems. 1. . www.hl7latam.org 14
- 15. DISCUSION It is important to avoid overabundance of alerts. Early detection of malnutrition plays a fundamental role in reducing morbidity in patients. www.hl7latm.org 15
- 16. Thank you Email me at hmandirola@biocom.com www.hl7latam.org 16
Editor's Notes
- First of all, let me thank you for being here. It’s a pleasure to be here with you. Let me introduce myself. My name is Humberto Mandirola, I am a doctor and I am a member of the board of H L 7 Argentina, I work in the Health Informatics department of the Italian Hospital and in the critical care unit of the Belgrano Hospital, in Buenos Aires, Argentina. The goal of my presentation is to show you the results of a research connected with a software-based alarm system to detect malnutrition in hospitalized patients. We carried out this study at Belgrano Hospital from 2012 to 2013. Before I start, I would like to mention some people who were also involved in this research: Diego Kaminker of H L 7 Argentina, who helped in the research planning, Sebastian Guillén and Javier Alejandris, of Belgrano hospital, who helped in data collection. Finally, Analia Baum, Daniel Luna and Fernán Quirós, of the Italian Hospital, who were in charge of the revision of this research. To all of them I am very thankful.
- I have divided this presentation in several parts: I will start by giving an introduction of the research. Then I will have a look at the background of malnourishment in hospitalized patients and C P O E. Finally, I will talk about the research and the results.
- Belgrano Hospital is located in the surrounding area of Buenos Aires city, Argentina, South America. It is a medium level acute care hospital. It has 200 beds, 8 critical care unit beds and 250 physicians.
- Our research was focused on detecting malnutrition in hospitalized patients by using a software-based alarm system. Later, we assessed the efficacy of this malnutrition-detecting system. The idea of creating this alarm system to detect malnutrition was developed when we understood how difficult it is to detect it in hospitalized patients. The nutritional aspects of patients often go unnoticed because of several reasons.
- Malnutrition is a serious concern in hospitalized patients and it is a very important aspect to consider. Malnutrition prevalence upon hospital admission is higher than generally believed. This is a serious problem that is frequently underestimated, so it is therefore important to take steps to correct malnutrition in hospitalized patients. The prevalence rate of malnutrition problems in intensive care area is approximately of 30%, out of which 70% goes undiagnosed.
- Besides, there are also factors that may contribute to undetected malnourishment in patients, for example: difficulties in monitoring weight of bedridden patients. They cannot easily stand on a conventional weighing scale. Patients may not eat properly because they don’t like hospital food. They may lose appetite as a result of hospital confinement and hospitalism depression, among others. The disease itself consumes energy of the patient, causing malnutrition and lower level of body energy. Doctors may neglect the nutritional aspect of patients. They focus mainly on the disease and studies and not on nutritional aspects. All this makes undetected hospital malnutrition a matter of concern.
- It was our aim to diminish the malnourishment rate. We started our research by analyzing different ways to help doctors know the nutritional aspect of the patients. We considered some aspects related to training by using some clinical guidelines, but it didn’t work. We also tried to approach this subject in a multidisciplinary way but it also didn’t work, because all decisions were taken by acting physicians in the intensive care area. We thought that an automatic alarm could work, and we decided to create a software-based alarm system. We based our idea for a software base alarm on C P O E systems.
- I will talk now about the computerized order entry system. C P O E systems frequently include integrated decision support components. These decision support components improve patients’ safety. Studies on C P O E have shown positive effects of decision support components on patient outcomes, including the reduction of: delay in order completion, errors related to handwriting or transcription, duplicate orders, overdoses, allergic reactions and drug interactions. For these reasons, which are connected with the patient’s safety, many healthcare organizations and health insurance require the use of C P O E. Yet, the overabundance of reminders and alerts (also called alert fatigue) should be avoided as this may cause clinicians to neglect both important and unimportant alerts, in a manner that compromises the desired safety effect of integrating decision support with C P O E. After trying, without success, several measures to decrease hospital malnutrition, we decided to create a system with an alert engine.
- In the technical aspect, we decided to program the alert engine in dot NET technology and in C sharp language. We used different standards to codify controlled data, we used LOINC standard code to encode laboratory data and I C D 10 to encode medical record data. We used H L 7 standard version 2.4 to transfer data through messages. The laboratory and medical record data were sent from the laboratory system and from the medical records system to the alert system by an H L 7 messaging. If two or more nutritional rates are below the nutrition normal standard, an alarm goes off. Each time there is a data to be controlled by the alert engine either from the laboratory system or from the medical record, an H L 7 message with the information of the result is triggered to the engine.
- The next step was to measure the efficacy of the alarm system we created. We generated two sample groups of 200 patients each, 400 in total. One group was called “the alert group”, and used the alert system with them to detect malnourishment. The other group was called “control group” and this group did not have an alarm system to detect malnourishment. The average age of the alert group was sixty five years old. The average age of the control group was fifty six years old. There were no children in these samples. Additionally, both samples included patients of both sexes.
- Exclusion Criteria. This study is a randomized controlled trial, which provides the most effective way to exclude external influences between samples. We had to exclude five patients in total, (2 from the Alert group and 3 from the control group). We excluded these patients because some of these patient’s data were missing, like laboratory results, clinical data and anthropometric data. After the exclusion of these patients, the total number of patients of both groups was 395.
- Now let’s have a look at the way the alarm system works. The engine is fed with data from the laboratory and from the electronic medical records through an H L 7 interface, such as levels of albumin, triglyceride, cholesterol, leucocytes, other laboratory parameters, body mass index, among others. Each time there is a data to be controlled by the alert engine either from the laboratory system or from the medical record, an H L 7 message with the information of the result is triggered to the engine. The engine processes these data. If the engine detects something wrong, the alarm will be triggered.
- The results were better than expected. Nutritional risk was more easily detected in patients who were controlled with an alarm system than in those patients who were not. The difference between the two groups was significant. The system detected several cases of malnutrition. Doctors later confirmed that 80% of the cases were actual cases of malnutrition.
- Results have shown that the nutritional risk alert we built was of significant help. Detection of malnutrition increased by fourteen % in patients controlled with an alarm system. Detection of malnutrition does not mean the patient is malnourished, it simply means that patient has to be examined by the physician so that he may decide whether he or she is malnourished or not. The detection of malnutrition by the system make doctors pay more attention to these cases. We think it is very important to use alerts, and in our study we managed to show their usefulness. Physicians using alerts can diagnose more problems than those without them.
- When using alarms, we highly recommend paying attention to the systems design: it is important to avoid overabundance of alerts in electronic clinical record because the excess of information generates frustration in doctors and as result they tend to skip reading some useful information. Lastly, I would like to summarize that the control of malnutrition in hospitalized patients with alarm systems results in better attention and saves time. Early detection of malnutrition plays a fundamental role in reducing morbidity in patients. Indeed, alarm systems do have good results. And don’t settle for nothing less
- Finally I would like to share with you a quote of the father of the Nation "Be the change you want to see in the world" Thank you! Here I feel at home. If you have any question I will very be glad to answer.