Various techniques have been introduced in recent years to predict heart disease. Tools and techniques in data mining help to discover hidden patterns and valuable information to help make important decisions. Recent research conducted by a group of Japanese researchers concluded that continuing CPR (Cardiopulmonary Resuscitation) for around half an hour after cardiac arrest could save the patient’s life. They claim that providing CPR for around 38 minutes would increase the chances of survival with better brain function.
The Japanese researchers conducted systematically mined data relating to a group of 280,000 cardiac patients. They narrowed down the initial group to 32,000 people, whose hearts started beating on their own after resuscitation. Analysis of the data of this group of patients 30 days of their cardiac arrest revealed that more than 27 percent had good brain function! From this, they concluded that brain function was the best with a return to spontaneous circulation by 13 minutes and that neurological outcomes would probably be not as favorable beyond 38 minutes.
According to the American Heart Association, heart disease is the prime cause of death in U.S. Estimates show that around 1,100,000 people in the U.S. have a coronary attack each year. In such a scenario, data mining is making an important contribution for the evaluation of the most important criteria for predicting patient survival and profiling patients according to their survival chances. This is helping to develop appropriate techniques to improve health care outcomes.
Mining Software for Quality Data Mining
Studies show that extracting information using data mining software was far better than the performance of a trained medical nurse who manually extracted information from patient data. Efficient software often take 4.5 hours to compute the value of the 4 medication variables for 327 cardiac patients, while the nurse took 176 hours to complete the same task. Data mining with the help of software usually include the following stages.
- Sampling: Collecting a statistically representative sample of data
- Explore: Applying statistical and visualization techniques
- Modify: Selecting the most significant predictive variables
- Model: Modeling the variables to predict outcomes
- Access: Confirming a model’s accuracy
Providing accurate diagnosis and effective treatment have become a great challenge for healthcare providers. Today, hospitals use information systems to manage their healthcare or patient data. These systems typically generate huge amounts of data as numbers, text, charts, and images. Unfortunately, this data is rarely used to support clinical decision-making. A data mining company can help extract the wealth of hidden information in such data efficiently and at affordable rates. This information can be effectively used to make appropriate decisions for improving care.
