Lesson finished
Unlock all 18 lessons of Medical InformaticsFull course ₹40,000 · paid
Healthcare Data Analytics, Clinical Decision Support and AI
A free lesson from Skill-Lync's Medical Informatics course, on Grow by Skill-Lync · Last updated 16 Nov 2026 · Session: “Module 1 Session 3”
What will you learn in this free Medical Informatics lesson?
The instructor explains descriptive, diagnostic, predictive and prescriptive analytics, with hospital dashboard uses such as readmissions, length of stay and bed utilization. The lesson then covers clinical decision support systems, their three-layer architecture and examples like drug allergy checks, and closes with applications of AI.
- Four types of healthcare analytics
- Hospital reports and BI tools such as Power BI
- CDSS architecture and drug interaction alerts
- AI applications such as image analysis
What comes after lesson 3?
Later in Skill-Lync's Medical Informatics course (paid): lesson 4, Module 2 Session 1; lesson 5, Module 2 Session 2; lesson 6, Module 3 Session 1. The full course has 18 lessons (6 h 42 min of video); see the course overview or its price and syllabus.
Lesson transcript
Machine transcript of the lesson audio (lightly cleaned). Technical terms may contain errors; the video is the reference.
In this section, we shall see data analytics and reporting. So let's look at the definition. So data analytics in healthcare is the process of analyzing any healthcare related data to derive meaningful information or reports to meet any clinical objective, improve patient care quality, predict trends, improve outreach, clinical data, diagnosis and business management. So data analytics helps us derive meaningful information. So all the data that have been collected and stored over the years would be put to analysis and the information that we derive from this analysis helps us take some decision that could be either to improve the patient care quality or it could improve the outreach or even help us with diagnosis or it could be business management related. For example, how many patients we have had this month compared to last year, similar month and such reports helps us keep on par with the repeat last statement.
So such reports helps us and basically the hospital management to take decisions that would benefit the organization. Now, typically the DAR reports are classified into four categories, descriptive analytics, diagnostic analytics, predictive analytics and prescriptive analytics. Descriptive analytics deals with uses historical data to draw comparisons, discover patterns to gain insight into the past what has already occurred. So coming to the diagnostic analytics, diagnostic analytics, diagnostic analytics, while it works on the data that was saved in the past, it also tries to analyze what was the reason behind.
So diagnostic analytics, while it takes into consideration the data that was used for descriptive analytics, it also tries to analyze what could have gone wrong or it basically tries to answer the what question, right? That is about diagnostic analytics. Now, coming to the predictive analytics with the data that we have in our repository. Now, analytics that are designed for predictive will actually give you futuristic outcomes from this data. Like, now example, if you have collected data that is pertinent to, let's say, kidney and you know the kidney fails under certain circumstances and when you have a new patient coming in with a similar clinical condition, your analytical tool could predict that potentially such patients are likely for kidney failure.
Now, these are all predictive analytics and when you try to build furthermore intelligence into the analytics, it is called prescriptive analytics. Now, prescriptive simply means that now that you with predictive, now that you know the outcome, prescriptive analytics will help you with the likely course of action that needs to be taken to treat, let's say, the kidney failure that would help save the patient now the data sources for such kind of analytics could be from different sources, it could be medical care, it could be genetics and biology, or it could be social circumstances or physical environment, or individual behavior, individual behavior could be like anything, the person, let's say, it could based on eating behavior, there will be different sets of data that would generated generated physical environment, for example, where do you live, which part of the world you live, what is the temperature or is it cold, hot, humid, based on such circumstances, the data would be collected and then you will have social circumstances like how is your lifestyle and related data, genetics and biology, again, which part of the world you come from, like how your ancestors has been, that also determines your health, how it would shape in coming years, medical care, again, what do you consume to live a healthy life and what care has been taken, such data sources from which we collect data helps us do different types of data analytics, which gives different types of meaningful information.
So applications of data analytics and reporting has been widely used and some of the prominent usage are for let's say hospital dashboards, for instance, in hospital, the top management would always like to know how many admissions do they have currently in their hospitals or you know how many readmissions has been there. Now readmissions is an indicator of how many patients were actually moved from let's say ICU to ward and the patients health deteriorated in the ward and have been admitted back to ICU for taking some medical care.
So such numbers also gives them the indicator how the hospital has been doing, we have reports on length of stay, basically, it helps understand how long the patient on an average is staying in an ICU. So there is always a number that the hospital management looks for, it should not be either too low or too high and such reports help them take necessary action at the earliest. There are reports on hospital bed utilization, such reports are necessary for the management to take decisions if it is underutilized, then they would need to take some necessary actions so that the bed occupancy is optimum or if it's overloaded, then again, they have to take a call that if they need to expand the beds.
So such decision can be taken by such reporting tools. There are qualitative and quantitative reports, which are used by the quality heads from the hospitals, they typically use it for monitoring the patient care, and that they also take necessary actions against the reports that considerated if it's, if it's deviating from their standard expectations. Dar is also used for decision support, as we saw, most of the reports helps us take different decisions. Reports also helps the hospital with supply chain management and its optimization. For example, if you are maintaining some stock of the medicine, and if it has to be maintained at optimal level, you should not stock either too many of drugs or too less.
So that kind of reports also gets generated and are far effective for any hospitals to take right decisions. Let's look at the tools used in the DAR module. Some of the commonly used tools as part of DAR are Microsoft Reporting, Qlik Sense, Microsoft Power BI, Tableau, or from AWS, we have QuickSight. Now, these are the reporting tools, which are far more user-friendly and needs very little programming. The image on the right shows sample report for period Jan to December. Tip it last statement. The image on the right shows sample report for the patient admissions from Jan to December.
And we can see that in the month of June, the sample report shows that hospital had the maximum admissions, but in the month of July, there was the least. And generally hospitals would try to look at this variation and why that variation was, was it because of seasonal change or was there less of publicity or what such decisions are taken based on the reports that come out of the data. Now, let's move on to the next section that is on CDSS and AI. CDSS stands for Clinical Decision Support Systems and AI, Artificial Intelligence.
Now, let's look at the definition of Clinical Decision Support Systems. Clinical Decision Support Systems, CDSS, are computer-based programs that analyze healthcare-related data to assist healthcare providers with prompts and reminders to enhance decision-making in the clinical workflow at the point of care. So, on the right, we can see the image, we have diagnosis and from diagnosis, we come to prognosis. So, once you diagnose what's the problem and then prognosis is you predict how it is going to shape up in coming time. Then you also look at the treatment that would be required against that diagnosis.
And then once you give treatment, you again monitor, you evolve. And the next step again is basically to avoid such events or diagnosis or clinical decision, you do some prevention methods, right? Now clinical makers or support systems actually help in different sections of these areas. So CDSS are used in every section of this workflow, right from diagnosis, prognosis, treatment, evolution, prevention. So systems are built to address challenges that are faced in every step of this decision making. Now, if you look at the benefits of CDSS, it increases the quality of care because the right decisions are being taken at right time.
It strengthens the likelihood of favorable outcomes, avoidance of error and adverse medical events. It also boosts employee efficiency. Knowing that there is a CDSS system in place, the employees always know that there is someone to assist them in decision making. It also improves cost-benefit ratio and the bottom line because of the IT systems, less time is spent, you know, guessing what could be the right decision. This system helps take decision faster and hence the patient outcomes are much better. Increased patient satisfactions. It also increases patients' experience and with such systems because decisions are taken faster and patients have good progress against the actions taken.
So let's look at the architecture of the CDSS. A typical CDSS contains three core elements. A base or data management layer, inference engine or the processing layer and user interface. The data management layer holds all the clinical and patient-related data, while processing layer contains the algorithm that work on data to produce results. So you have a data management layer and then on top of that you have processing layer and the third layer is basically the user interface which shows the result or the output received from the processing layer.
So that's how a typical CDSS system is layered. So let's look at the examples. In the paper world, even today some of the complex clinical workflows has been managed on paper resulting in delayed action and error nears as well. The first CDSS called Mycine was built in the late 1950s. So it gave suggestions of treatment for patients who appeared to have a bacterial infection. So Mycine had a computing model applying over 500 rules. So Mycine actually considered the patient weight and based on the patient weight that was entered, the system used to prompt the recommended dosage of the antibiotics for the patient.
And so such systems were actually put in use in since the late 1950s. The other example of CDSS is drug allergy checking. So the system matches an ordered medication against a patient's list of documented allergies, evaluates the probability of unwanted reactions and generates alerts. So basically what happens is once you document the allergies for a patient, let's say if the patient is allergic to some food ingredients, it could be peanuts, tomatoes, or the patient could be allergic to some of the medicines for simplest could be let's say paracetamol.
In this case, once you document these allergies at the time of admissions and tomorrow if any of the caregiver is giving anything that the patient is allergic to, the system would immediately give a pop-up saying the so-and-so patient is having allergy to the item that has been ordered. Hence, such systems have helped the caregivers take right decision. And basic guidance on dosage is the other application. So for example, the software component generates a patient-specific list of recommended dosing parameters for a particular medicine. Now we have systems in place clinically.
They are so intelligent that if you have a patient weighing X kgs, so many drugs are which are weight dependent, so against that weight, you will get a recommended dose or even against a specific gender or, you know, if it's a hormonal therapy. So such guidance on dosage are also given by the system. Drug interactions checking is another example. So based on clinical documentation at hand, a CDSS considers interactions of a newly prescribed medication with other drugs administered to the patient. So basically, sometimes you have drugs-to-drug interaction.
So when two types of drugs are given to same patient, they interact inside the body and could be adverse event could happen. And hence, such advices are given to the caregiver to take the necessary precautions. Checking of duplicate therapy. So that's another application. So within clinically, CDSS features addressing overdose problem compares a newly added drug with active ingredients of drugs in a patient's profile. And if a match is detected, the system generates an alert. For example, if you are giving crocine to a patient which has paracetamol and for example, if a new doctor comes and orders for dolo as well, in that case, the system would be smart enough to give us the reminder or an alert to the caregiver saying you have already administered crocine which has paracetamol and the new order or the drug that has been ordered is also having paracetamol as its ingredient and it could be a duplicate therapy or it could be the overdose for the patient.
Hence, such systems, I mean, the logics are being incorporated under 3DSS systems. The other example is for diagnostic support wherein the CDSS for disease identification compare information on a patient's condition with a knowledge base and generate a list of possible diagnoses. The other example is clinical management. Hospital rules can be encoded into a knowledge-based CDSS. Such solutions perform various tasks from prompting nurses to take specific measurements according to a protocol to informing doctors about patients who don't follow their treatment plans. So, clinical management is one of the application areas of CDSS.
The other areas where CDSS is used are basically cardiac risk assessment, anemia assessment, sepsis, myocardial ischemia, acute kidney injury, behavioral health criteria, medication management are some of the CDSS that are used in practice. Let's move on to new section that's on AI. So, what's artificial intelligence? Let's look at the definition. Artificial intelligence in healthcare is an overarching term used to describe the use of machine learning algorithms and softwares or artificial intelligence. to mimic human cognition in the analysis, presentation and comprehension of complex medical and healthcare data, right?
So, basically, artificial, as it says, it's an overarching term. Basically, it's a broad term, right? And it's basically is used to describe the machine learning algorithms and softwares that mimic human brain. And that's why it is called our artificial intelligence. Now, we have a complete module in the end on artificial intelligence. Hence, we shall cover the brief applications of AI in healthcare here. Now, AI has helped in automation, vastly automation of clinical processes and workflows, smart decision makings with CDSS, AI has also with the machine learning has, you know, given excellent results of giving clinical outcomes, improved accessibility with AI, accessibility of data has improved early diagnosis.
Many times, as we can see here on the last point, image analysis like diagnosis, AI has been able to detect cancers or tumor markers from the images that are acquired from the imaging devices. Sometimes humans may... Missed to find such findings but the AI based systems have been of great help for the doctors to capture such information. It has increased the speed and reduced cost. It has enhanced human abilities and mental health support. As we discussed, such systems have assisted doctors and the caregivers and that has put them under huge relief from the stress work environment of doing all the things and taking decisions on their own. Such systems have assisted them in taking decisions smartly and effectively.
So workflow assistance is always there with AI and efficient and unique assistance in surgery as well. These days we have seen with the assistance of AI. AI. So to picturize the role of AI in healthcare, it has early detection of ailments, improved decision makings, it helps in treatment, expanded access to medical services, it has associated with care, it has given superior experience, right? And then checking health through wearable devices. We have seen smart watches which picks up the heart rate and other parameters and then it also helps you know how your sleep pattern has been. These are all the AI tools that work on the data that gets collected and obviously at end of life care also has been part of the AI journey.
With this we end module 1. With this we end module 1. Let's look at the summary of this module. So in this module we covered overview of medical informatics and its importance. We saw in detail EHR, HIS, its advantage and challenges. We also looked at the healthcare IT standards like HL7, FHIR, IHE, SNOMED CT, etc. Later we covered essential concepts of biomedical computing, understanding computer architecture and its component. We also saw different coding systems for biomedical data and it needs for example ICD, CPT, SNOMED CT, etc.
We also saw how data analytics and reporting helps and the different tools that are used for it. We also saw in brief clinical decision support systems and artificial intelligence and its applications. These are some of the image credits from which the image were used. With this we end our module 1 on introduction to medical informatics. Thank you. So that we way learn our and You