Over the last few years, everyone I speak to seems to know about AI. Ethical issues around AI are being actively discussed and debated, not just in the professional sense but also around coffee tables and at informal discussions by users.
This post is an attempt to provide an overview of the issues and approaches.
What we have now are essentially three interfaces of or approaches to ethics in AI.
Besides the differences in budgets, access to VC funding and who gets favorably written about in the NYT, 1 the main difference between the various factions (strange to see factions in ethics, but that is what we get for trying to pre-print our way into a science) is the temporal profile and the concreteness of the problems they are talking about.
Anyway my irritations aside, the ̷f̷a̷c̷t̷i̷o̷n̷s̷ approaches or interfaces are 2
The professional ethics people
The AI risk People
The AI alignment n̸u̸t̸s̸ people
Professional AI ethics
The professional Ethics people are dealing with the immediate and current harms. They focus on identifying such harms and developing frameworks and knowledge that can be used to improve things now and for the future. They are a lot like the bioethics people
One group of professional ethics people make guidelines. Another group fights big tech.Back then, when LLMs were not all that AI was, and people were using regression models to predict recidivism, deciding who to hire, and identify people from video surveillance, these people were studying the harms of such systems and talking about what to do about it.3
This paper provides a model for understanding the harms and risks that arise in different parts of the model training and deployment process. This is a good high-level overview.4
For many language models as well as image generation models King – Man + Woman = Queen and Man :: Computer Programmer as Woman :: Homemaker. These are looked into in the following papers
Caliskan, A., Bryson, J. J., & Narayanan, A. (2017). Semantics derived automatically from language corpora contain human-like biases. Science, 356(6334), 183-186. DOI:10.1126/science.aal4230 |Preprint on arxiv
Manzini, T., Lim, Y. C., Tsvetkov, Y., & Black, A. W. (2019). Black is to criminal as caucasian is to police: Detecting and removing multiclass bias in word embeddings. arXiv preprint arXiv:1904.04047. (pdf)
Facial recognition tech has a long history of being super duper racist, creepy, used for oppression, as well as not being very good. Tech companies, especially the superduperbig guys have been getting into this game and are releasing models that are seemingly better, but only if you’re a white male.
Emily M. Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. 2021. On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜 In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’21). Association for Computing Machinery, New York, NY, USA, 610–623.
Google the authors, and what they went through when they came out against LLMs. This will give you a lot of info about the stakeholders involved in this. The paper itself is highlighting the issues with Large Language models.
What about some practical stuff ?
A philosophical framework
While the debates go on, a zillion guidelines on how to deal with data and how to things ethically have come up. Chances are you will be overwhelmed if you start looking at them. I haven’t found any specific frameworks that speak to me. Personally I recommend and use the medial ethics framework along with the spider-man framework. which states that
That is as self explanatory as an ethical dictum gets. The medical framework has the following four principles
Respect for autonomy – Don’t make stuff that interferes with the autonomy of the user. When you’re in a position to make decisions for them, do this only after clearly explaining the harms to them and with their consent. Consent is king.
Beneficence – an AI programmer should act in the best interest of the end user and not of the employer
Justice – Is about the distribution of scarce resources, the decision of who gets what. Which means if your product, algorithm or system worsens disparities between people and creates inequalities, do better than that, think very hard about how you’re using your power.
Non-maleficence5– to not be the cause of harm. Also, to promote more good than harm, to the best of your ability. Also known as Above all, do no harm
Checklists and guidelines
If you’re looking for checklists and stuff that you can start using immediately to do responsible/ethical AI here is a github repo with the best links Awesome AI guidelines
If you want just *one* guideline, read this: Mitchell, Margaret, et al. “Model cards for model reporting.” Proceedings of the conference on fairness, accountability, and transparency. 2019. Model cards for model reporting.
This paper is great to get a conceptual understanding of the need for clear declarations about models. But it is too ambitious and a bit too bulky for wide use. Despite this, hugging face has implemented this on their website, although it isn’t always used. I recommend developing your own model card/checklist and having them attached to wherever you store your models.
This github repo with links to responsible AI resources is one more resource that focuses on practical advice and frameworks : Awesome Responsible AI
Explainable AI
For me one foundational source of ethical issues with deep learning models is that the algorithm is a black-box. The interpretability of the predictions or results of a neural network algorithm is poor. And so people who are working on model interpretability and explainability are also working on things that are critical for ethical decision making.
The risk people are talking about existential and military and other big risks from AI. There is definitely some overlap between the risk people and the professional ethics people but a lot of what they discuss is about future or possible harms. Still very concrete harms and stuff that we can definitely see happening.
Think autonomous weaponized drones, robot soldiers, autonomous robot doctors etc. The key here being to prevent things from getting out of hand when algorithms and robots are deployed to make autonomous decisions. I see this as robocop-dredd-dystopia prevention work
Healthcare is another area where a great deal of harm could come from automation (a great deal of good too) and it is important that we think hard and work towards systems that safe-keep the interests of the patient.
Here is a report about them and how people are trying to stop them. Stop killer robots
This debate captures a great deal of nuance on AI risk. Melanie Mitchell is a delight to listen to.
I don’t think there is enough work being done about real risk. We have a lot of thought leaders talking about it, but the engineering and the science of it is not getting a lot of traction.
AI Alignment
Imagine a world where an AI who is much smarter than us all has emerged and is currently demanding that everyone should call it lord and master and pray to it thrice a day.
The AI alignment people are
Figuring out how to prevent such an AI from emerging and
How best to align the interests of such an AI with our own.
I am not kidding you.
There are some actual cults involved and a lot of though experiments that are so bizarre I cannot even. 6
The problem is that these folks are completely ignoring addressing the current harms using the doom and gloom of this possible outcome.
This is a doomsday cult kinda ideology. And sadly some very big names in AI have signed up on this cult.
Worse, this approach is being used by some large players to superficially meet the demands of ethical AI, while completely sidestepping accountability for the issues that are relevant today. As you can imagine, there are many important people in governments all over the world for whom the biggest worry is an AI that will replace them. So those guys are also treating this like its a real and credible threat right now.
There is no doubt that we need to ask ourselves this question, about how do we deal with systems that have more power than us. But I think the answer to those questions lies in building in accountability, transparency, safety, informed consent and things that we already know how to do pretty well and don’t do because its bloody inconvenient and cost money. I definitely believe that we need better engineering research into this, threat assessments, all that. But this issue is not so novel that we need to come up with an entirely new discipline which ignores and laughs at the stuff other experts on harm-reduction are saying. That is stupid.
I am not going to link to any of the alignment cultists but here are two analytical articles about them which I think are great, and they link to plenty of stuff that you can explore.
I dont really fully agree with these authors, but they make some sense and what would be an ethical guide without stuff that one disagrees with.
Concluding remarks
I admire greatly the activists who are fighting the good fight against Big-LLM, I really do. But I do not like agendas for change and progress being exclusively drafted by activists on social media. Social media is like reality TV, what you get if you do all your intellectual debate there is some form of Donald Trump.
I think that at some point the IT/AI engineering profession is going to realize the same thing that the medical world did. That if you don’t start doing things ethically, you will lose your power and create harms far beyond you can imagine and this shit will haunt you. If the tech world looks at the medical, i guess it can see a lot of unethical stuff. That is our shame. But at the same time, I hope they will investigate the issue historically, and ask just how many checks and balances are there in healthcare to ensure the patient is not harmed. There is a lot to learn from the history of medical ethics.
Ethics make sense because it improve systems. Great AI will be ethical just like the best healthcare is ethical. And just like a doctor ultimately works with the patients best interest at heart, at some point AI engineers too will adopt this dogma because it is the rational best choice and has a proven track record for reducing harm, and no one wants to build stuff that harms people. Also, the workers of the world have a lot more in common with each other than with the bossmen. 7
This little fella knows ?she looks FABULOUS
Image from: Welch S. C. & Metropolitan Museum of Art (New York N.Y.). (1985). India : art and culture 1300-1900. Metropolitan Museum of Art : Holt Rinehart and Winston.
If you wish to cite this post here is the citation
Philip, A. (2023, July 11). Ethical AI: Basic reading, schools of thought, an overview. Anand Philip’s blog. https://anandphilip.com/ethical-ai-basic-reading-schools-of-thought-an-overview/
Ok that is all I have for you now folks, please do comment and subscribe and like and share this on boobtube instamart dreads and feathers as you wish.
This work is then being carried forward by the actually-LLMs-are-not-that-great activists. ↩︎
It however lacks the liberal-progressive-activism priorities ↩︎
No relation with Angelina Jolie, having horns or dressing in black ↩︎
I love thought experiments, they teach us a lot of things including that one must not confuse a thought experiment about a distant and remote possibility with something real and applicable now. ↩︎
I am a Bourgeoisie malayali, can you blame me for bringing up Marx?. ↩︎
The #SurvivingViolence team interviewed police across ranks. The police are the most visible and accessible (present in every area) support service provider. https://t.co/SUSUi9feZV
Technical Details: This is a multi-method study which collected primary data from survivors of domestic violence and other key stakeholders, centering the experience of victim-survivors. It was conducted in 3 districts in Tamil Nadu; Ariyalur, Chennai and Vellore. The sampling method was Snowballing. 60 woman and 1 transman survivors were interviewed. Besides them, witnesses, members of the society and family were interviewed, as well as 62 support-providers like lawyers, medical workers, NGOs, and the Police.
There are seven sections in the report.
Section 1 SETTING THE SCENE
The report opens with background information about the state of Tamil Nadu from key studies based on NFHS, SDG, Census etc. highlighting the data on women’s health and experience of violence, and goes on to describe the questions in the interviews and the responses and then describes the methodology in detail.
Section 2 UNDERSTANDING DOMESTIC VIOLENCE
It then goes on to describe attitudes and beliefs of the interviewees about gender violence, laws about it, what constitutes it, and stances on punishment and resolution etc. are elicited.
The interviews were done with great sensitivity and did not shy away from asking very uncomfortable questions. There were both structured as well as non-structured bits in each interview. Some of the questions were
Why do you think women experience Domestic Violence?
If a friend or family member told you they were experiencing domestic violence, what advice would you give them?
Is there ever a situation/circumstance when a woman should stay in an abusive/violent relationship? If so, why?
Some of the findings from attitudes towards violence are
Survivors understood that beyond physical violence, sexual, emotional, verbal and economic abuse constituted domestic violence.
They agreed that women who experience violence should seek help.
Community members agreed with these views, and that domestic violence is committed because of cultural factors, interpersonal misunderstandings and alcohol abuse.
Community members also believed that women should and would put up with the violence up to a certain threshold of cruelty or affecting children, before leaving.
Support service providers largely agreed with an inclusive definition of domestic violence, but were less sure about whether it was a criminal office and the perpetrator should be punished
Summarized in the report as:
While society’s understanding of domestic violence has broadened, the a significant percentage believe that it is to be expected in relationships and are ambivalent about how best to respond. A survivor therefore steps out of an abusive situation into a support ecosystem that may or may not validate her experience.
Section 3 THE EXPERIENCE OF VIOLENCE looks into the details of the experience of violence (who, how, what etc.), why survivors think they are subjected to violence, ways of coping, seeking help etc. The community is also asked the same questions but keeping the experience of the survivor as the center.
Section 4 looks into the support available to survivors. This includes friends, family, the police, NGOs, lawyers etc. and their experiences in dealing with domestic violence and their understanding of the experience of it, the causes of it and the ways to solve it.
Section 5 looks into how well the Protection of Women from Domestic Violence Act, 2005 is implemented (I didn’t read this too closely)
Section 6 asks what patriarchy looks like, with stories from survivors and witnesses and others. eg
Section 7 deals with the people’s views on developing resilience and what it means. This section spoke to me deeply.
This project arose out of a broader quest to understand resilience in both intellectual and practical terms. Where do humans find their resilience? What are the external supports and impediments for it? How does any society (or state) build the resilience of its people in the face of adversity? We took this question to a section of people vulnerable because of their gender in a patriarchal society and because of their everyday experience of violence in the very contexts that are meant to keep us safe—intimate relationships, family and home
The report ends with suggestions to what both the civil society organizations and the government apparatus could do to improve things, informed by study.
At a cultural level, the understanding of domestic violence has changed a lot, and while everyone seems to agree that it is a bad thing and ought to stop, the solutions are not all that clear or trustworthy (to the society).
We need to support survivors of domestic violence in the communities we are part of. The support can be in many ways, but it should probably begin with speaking up about this in our families and providing support to family members who are survivors. I think this is key. We need to be clear and loud as individuals about what our stance is in the family-social units we are part of.
Individuals should attempt to provide structural support in some forms more widely.
Is there a model of heterosexual marriage that is not based on privileging the male and asking for submission from the female? Without widespread demonstration of the alternate model, for the family unit and the greater culture that we are part of, there really is no alternative that can be adopted.
Women who leave abusive spouses and/or seek help and justice need to be celebrated.
There is really no benefit to women “adjusting”. Adjusting is based towards benefits the norms that created this problem. I recognizing that that is the only recourse available in many situations at present.
Corruption masquerades as pragmatism.
I could view this problem as legacy code issues. While the number of bugs in the system make you want to burn it all down or lose hope, most of what we have came via iterative development. I mean the methods, techniques, heuristics, design-patterns and more that make up culture. that is reductive, i agree but useful to me.
The report makes me grateful to the feminists who have fought to get the laws and legal structures in place to deal with this issues.
It makes me hopeful to read the vast majority of the interviewees want education, financial stability, and social support and see these as being part of the cure.
This is reinforces my belief that individual financial empowerment is severely understated and underutilized as a source for reducing harms in the nation at large.
Level 1 the PITA : EMRs that double the work — systems where the EMR is used for documentation or compliance, but all the real work happens in written form, so this just increases the net amount of work done. [Admittedly, this is a problem of the administrative system, not so much the EMR itself, at times] (a large majority of EMR deployments in india are this kind)
Level 2 the copycat : EMRs that successfully replicate all the physical record keeping and so work isn’t double, but worsens things because replicating physical workflows in-toto invariably will make everything slower (This is what most new EMR implementations are)
Level 3 The trying-very-hard-to-be-cool: EMRs that have some automation and short cuts built in, like medication templates, discharge summary templates and suchlike which in some areas, significantly reduce the time taken to do a task and are overall nearly as fast as paper and pen ( I’ve seen a few of these)
Level 4 The we-do-design-thinking-and-listen-to-the-money : EMRs that have thought through the clinical processes and removed all parts that do not need to be there, and have largely click and pick workflows these are faster to use than other EMR/EHR and might even be faster than pen and paper. (I’ve seen maybe one of these of these)
Level 5 : EMRs that are faster than pen and paper, assist and improve decision making and make things easier for patients. (Haven’t met one of these so far)
How approval for new drugs, new fixed drug combinations and new generics are supposed to work and how they actually work
What are the quality standards and monitoring mechanisms in place for scientific medicine, as well as AYUSH.
Laws governing AYUSH, scientific medicine and the production of drugs in India.
What I liked about the book:
Thorough research into how things are supposed to work in the area of drug regulations in India and an even more detailed research and experience into the million ways in which it is failing to work. Seriously, these two know their drug regulation problems.
Thorough understanding of the main players, organizations, lobbies etc. involved in this messy situation
Stupendous amounts of cred as far as I am concerned due to being a whistleblower as well as a RTI crusader/activist and all the litigation they have tried.
Focus on transparency, access to information, rights to quality
Documentation of various types of issues and the corruption in different bodies
Clear-sighted tracking of the bureaucratic games and political shenanigans around drug laws in the country.
What I learned from the book
We have a lot of outdated laws regulating drugs and other allied health products. These are exploited and used by both the governance people and the drug companies to push a lot of outright shitty medicines to the public
There is a LOT of corruption, incompetence and lack of transparency in the organizations that are supposed to be watchdogs and regulators
Successive governments and bureaucracies have generally found ways to pass the buck, do ass-saving paper pushing and pandering to sentiments rather than make any substantial and scientific changes.
We have a lot of under-equipped and underfunded drug regulators and inspectors
The quality of drugs produced in India is for all practical purposes at the mercy of the individual pharma company that does the production and there’s very little if any, real oversight or really punishment for even serious lapses.
Every month somewhere in this country, at least one pharma company is shown to have produced sub standard drugs and even then the judiciary seems to think harsh punishments are not needed. I mean, people have died and the dudes who made those meds never saw jail time.
Problems with the book
The whole book is narrated by someone who is shouting at YOU AND ALL THE IDIOTS WHO RUN THIS COUNTRY AND ALL THE CORRUPT DRUG MANUFACTURERS, THE INCOMPETENT JUDICIARY, THE CLUELESS LAW MAKERS OF NEWLY-INDEPENDENT INDIA WHO DID NOT HAVE 21ST CENTURY VALUE SYSTEMS OR KNOWLEDGE AND DRUG RIGHTS ACTIVISTS WHO ARE ALL CORRUPT TOO.
My blood pressure was measurably higher when I was reading this book. Seriously. There’s like four people the authors have anything nice to say about from 1923 -2023 (and two of them are dead). Oh, and the USFDA. They LOVE the FDA and the way things are done in the United States.
<sarcasm> I mean, I don’t blame them, they’ve been fighting the good fight all their lives when literally everyone else in India is asleep, incompetent and corrupt. </sarcasm>
I get it, though, one of the authors has been constantly targeted and harassed by various government agencies and big pharma. So I get where the tone is coming from, I know righteous indignation, I used to have a lot of it. Now I just have reflux and indigestion.
But the roid-rage-hulk-smash tone of the book made me want to put it down and stop reading about fifteen times. I stayed the course only because they did thorough research, and what is an activist without some righteous rage, really.
The other issues are that the policy and legal recommendations made in the book without fail made me wince, flinch and facepalm. The authors forget just how poor this country is, and just how many people there are. The authors also don’t know a damn thing about the practice of medicine in the real India. They are operating from an america-centric and america-learned model of how health should be dealt with.
The most important implied recommendation of this book seems to be BURN IT ALL DOWN.
Throughout the book they mention the dozens of parliamentary and other committees that have looked into these problems, and how so few of these reports see the light of day and fewer still get anything changed. Because the suggestions are not useful, or due to the unwillingness of the governance system to make any real changes.
This book has the same problem.
For eg.
They repeatedly suggest stronger laws, after describing just how poorly existing laws are implemented.
They suggest that online pharmacies will solve the supply chain problem (omg i cannot even)
They suggest that small pharmacists who are found to have bad drugs should be punished heavily.
They repeatedly recommend increasing the legal pressure on smaller players in the supply chain.
You cant improve state capacity by making tougher laws.
You can’t fix corruption in high places by punishing the small guy.
Some medication is most definitely in many many cases better than no medication, talk to a doctor, any doctor.
You cannot copy-paste quality standards designed in USA and Europe to solve for India.
That’s not how anything works.
Would I prefer my patients got 40mg of Telmisartan every time they buy a drug, YES.
Do we need to improve the capacity of drug inspectors and dis-entangle the regulatory nightmare, HELL YES.
But do we do it by shutting down everyone who only has 20 mg of telmisartan in their 40mg tablets, heavens no. The 20 mg the patient gets at 0.05 Paise a tablet is worlds better than the 0 mg he will get at 2 rupees a tablet.
In the words of the y00ths I interact with, the authors desperately need to touch grass.
Note: This is a guide for people without a math and code background, written by a medico for other people from a medical/biology background. If you are the mathy type, this is not for you, try this instead
Artificial Intelligence (AI) is an umbrella term for scientific fields whose aim is to mimic or replicate human-like skills with computers. A large part of this field is driven by computer science and mathematics.
One of AI computing’s main goals is to create self-learning or self-training systems or algorithms. The field of AI computing that tries to create, test and use such algorithms is called Machine Learning (ML). Artificial Neural Networks (ANNs) are one such algorithm which are very popular and have lead to a lot of breakthroughs in what computers can do.
Artificial Neural Networks (ANNs) are a type of software algorithm that is composed of bits of code that can do math and store information (neurons), that pass information (inputs) back and forth between each other, making slight changes till a particular result (output) is achieved.
This article describes a largely non-mathematical and no-code explanation for how it does this.
What are ANNs
Layers of the retina
Consider the Retina, it has layers and when light hits these layers, it triggers (gets converted into) various types of signals (chemical and electrical) and all these signals gets passed on to the Optic nerve and then to the Occipital lobe, which working in concert with the rest of the brain interprets what the signal means and produces an output- vision.[1].
This is not how an ANN works.
Like with the retina though, ANNs are code (neurons) arranged in layers sandwiched between an Input layer, which receives the signal, and an output layer which produces the results, and finally a verifier which determines if the right output was received or not.
A whisper (neural) network diagram
Why are they called neurons and artificial neural nets (ANNs)?
There’s a theory that neurons learn by passing on information that meets a certain criteria (activation threshold) to other neurons (spreading of activation) and getting repeated feedback about how to correct these activations from those neurons till it activates for the right signal or passes on the right information. [2].
For example, when you learn to do a physical task, like opening the fridge, the very first time you do it, your muscles and the nerves that control the muscles don’t really know how much force to use, or how much signal should pass between which neurons controlling which muscles to open the door smoothly. But, the system picks a starting amount of power, and tries, and the sense organs give feedback that says is this working or not, and based on that, it will adjust the amount of power needed. Over time the flow of information and signals between all the nerves and the muscles and the various organs involved in opening this door is so good and so optimized that it becomes effortless.
This is thought to be because of the back and forth “this is working, this is not working, too much pressure, too little pressure, wrong angle, right angle” messages that are rapidly passing between the hand and the brain. And it is this self correcting back and forth messaging system (back propagation of error) which eventually figures out the right way to solve the problem by finding the right amount and type of information (signal) passed between the neurons involved in this action.
This is called the connectionist theory of cognition, and this system of learning is called a connectionist system because the way it works is thought to be due to the strength (weight) of its connections and finding the right connections.
Most neuroscientists believe this theory to be completely wrong about how the brain actually learns [3]
But, it’s either got a kernel of truth in it, or due to some mysterious reasons, this system of learning works when turned into computer code and so is called an artificial neural network.
So what does each neuron in these ANNs do?
Two things
When it receives a signal it transforms it and sends it forward to other neurons it is connected to. (Forward pass)
When it receives some information back from neurons about what was wrong with the earlier output, it adjusts the output and then passes the information along to the neurons behind it. (Backward pass)
You’ve got people standing in a row, and the game host says “life is meaningless” to the first person. Person one hears “wife is meaningless” and blurts it out to the next person, who hears “wife is weaning less”, the next person hears “life is winning less”, and so it goes. Now, in this version of the game, instead of revealing the answer, the verifier who is standing right at the end says, you were wrong by x percentage. So the last person passes this to the person before them, and this process is repeated a very large number of times till the message being received back is “life is meaningless”. Note that the last person isn’t told that the right answer is “life is meaningless”, only that it was wrong by a small amount.
Now consider that this is being done in parallel with multiple rows and each rows hears a different parts of the message, and what the verifier wants to hear is ” life is meaningless, cried the nihilist, while munching on stale peanuts he stole from the existentialist’s larder”
Sounds like a nightmare, and a mathematical nightmare it is for people like me who start getting hypoglycemic when letters become numbers and numbers get letters added to them, sometimes before, sometimes after, sometimes on top etc. Fortunately, the programs do all the math themselves and we just need to know two things. What matrix multiplication is what a linear equation is.
Matrix multiplication is a method of multiplying two sets of numbers arranged in a particular way. A matrix looks kinda like this
So it’s like a table of numbers but they encode some other information like which axis a number is on and stuff. There are specific rules about how two or more matrices can be multiplied. We won’t go into that because it is boring.
The reason we want to know about this is that in machine learning one thing we do a lot is turn all our data into matrices and then multiply them.
We represent numerical data as vectors and represent a table of such data as a matrix. The study of vectors and matrices is called linear algebra Source: Math for ML
Irrespective of what kind of data you need to provide as input, all of it gets converted into numbers. But you already know that that’s how computers work. We just use a specific method for representing these when doing ML, which is turning everything into vectors or matrices.
Linear equations (which is what linear algebra uses) all kinda look like $ax+b =c$
This is a mathematical formula which will always produce a straight line if the values are changed and plotted.
This means that, if you know the $a$ and the $b$ and the $c$, you can guess what the $x$ is and solve the equation [no duh].
But even if you don’t know what $c$ is, but can find out if the number you came up with (your $c$) is bigger than or smaller than $c$, you can adjust the $x$ to get the right answer.
The reason why we need this kind of a mental contortion is because remember that in our Chinese whispers example, the person at the verification end doesn’t say what the answer is, only that the given answer is right or wrong and some directional information. This behaviors is not just to create confusion, it’s designed to facilitate learning, because what we want is not a network that has memorized the answers (which it would if you give it the answer), but a network that’s learned the method to arrive at the answer (which you torture it into learning) .
So if you have a formula like $50x + 70 = 520$ and you don’t know the answer is $520$, but have this oracle who can tell you if the number you come up with is bigger or smaller than the real answer, how would you solve this?
You could begin by guessing $x=1$, and get the answer $c_1= 120$, ask the oracle is $c_1 >= c$ and you will hear $FALSE$.
Which means your $x$ needs to be bigger. OK, how about $x=10?$ This will give us $570$.
Oracle, is $c_{10} >= c ? -> TRUE$
Now we know we (probably) need to make the $x$ smaller to get the right answer.
As you can imagine, depending on how you change the $x$, you will within a few “steps” up and down, get the right answer. [4]
This painful iterative way is exactly how neural networks self correct.
Let’s say we don’t really know the right $a$ or $b$ either. Can we still solve this problem?
How can we come up with the $a$ and the $b$?
Not surprisingly, we can use the earlier approach of just picking some random numbers, putting them into the equation and then adjusting them up and down till we get the formula we’re looking for.
To keep the moving parts minimal, let’s make it so that $b$ is a constant, so we just pick a random number and leave it as a constant throughout all the iterations. $x$, we pick randomly as before but we will adjust it as we get more information. To begin with we can pick a random number for $a$ and after that we can use a different strategy, described below to update it.
OK, let’s use this to solve a pressing real world problem. Lets say we need to make a neural network that correctly identifies the various parts of a goat from a photograph.
So first we convert a photo into a matrix of numbers. [5]
Then we show the input neuron a part of the photo near the goat’s head. And we’re gonna ask it to guess what it is, but in linear-equation-math.
So to begin with, let $a =1 , b=1 \space and \space x=0$
(keep your imaginations alive)
$Step \space 1: a_1x_1 + b = c_1$ or $1*1+0 =1$
(In non math, this humble output $1$ means “this is a goat’s horn”)
Now, based on the information that it gets back, it will know if the x should “step” up or down. Let’s say the feedback says, “too high bro”.
So then, let’s step down the $x$ and let $x_2 =0.5$
Given that we already have a piece of data sitting in our neuron, which is the last output ($c_1$), instead of starting with a brand new $a$ (or answer), we can just modify this answer using the new $x$ and the same constant $b$, that way it’s nudged in the right direction. So we use $a_2 = c_1$ . This is made possible because $c_1$ isn’t a random number, it’s the result of the assumptions you made earlier. Or, it is “This is a goat’s horn”, which you can step down or modify using the new information.
$Step \space 2: a_2*0.5 + b = c_2$
What happened mathematically is that we produce a result something like “this is a goat’s neck”.
You can imagine, that if you repeat this enough times, based on how large your “step” for $x$ is, and based on what the other neurons in the system are saying, you will at some point get the right value of $c$, which could be “this is a goat’s right earlobe”, and this is now what it has “learned”.
So the next time it sees something like a earlobe, it will be able to identify it instantly and not confuse it with other dangling objects that goats possess.
And isn’t it cool that we could come up with a way to just make a wild guess, then with feedback adjust the wild guess into a coherent answer? (Wait, is that how humaa learning works?)
I think now you might be able to see that $x$ is in a sense the the importance or weight you give the input $a$ to produce an output $c$ and the $b$ acts as a a constant nudge in a particular direction or a bias.
So, in summary
A neuron gets “inputs $(a)$”, multiplies it by a “weight $(x)$” adds it to a “bias $(b)$” to produce an “output $(Y)$” (I know we called it a $c$ earlier, but confusion is our friend). How do we get to a final output? By summing these iterations. In math:
$\displaystyle\sum (weights*inputs) + bias = Y$
That fancy squiggle means SUM of the things to its right.
And the updating of the weights happens by passing back information about if our answer was right or wrong and in which direction. In computer science this is called back-propagation of error with gradient descent, that’s not important, but it is useful for showing off.
We’re nearly done, there’s just one more thing I mentioned earlier, a threshold of activation, which will wrap up the whole thing neatly.
While a neuron produces all this data, we don’t want it to fire all the time right? Suppose what it “heard” was too soft for a meaningful answer, or if it was so “loud” so it PASSES ALONG A SHOUT LIKE THIS!!!!!!. These kinda signals can increase the error, or confusion.
So for hygiene, it’s better that we pass on only information that is of uniform volume all over the network and only if it passes some kind of a test of importance (quality check). To do this, we could pass the output $(Y \space or \space c)$ through a mathematical transformation that achieves this. This transformation is called the activation function.
Depending on the type of task, the activation function could be something like, take the mean, or convert into a range between 0 and 1 or something else
Those of you who remember your physiology lectures might notice how much this is like an activation potential for a neuron. That is not a coincidence. What ANNs do is inspired by real neurons.
An activation function is often represented with the Greek letter phi ($\phi$)
To update our earlier equation
$\phi(\Sigma(weights*inputs) + bias) = Y$
So now we have a way of making guesses that slowly move in the right direction, and we have an activation function that makes sure only the good stuff gets passed along in a standardized way.
Wait, now we know how a linear equation can be solved with guessing, but this technique isn’t limited to linear equations. Almost any kind of mathematical equation could be solved by initially random guessing and then passing the information about how wrong you are back, and updating the guess. In math this is called the universal approximation theorem and it is kind of a big deal.
That’s it. That’s how a neural network do.
Congratulations you now have a working understanding of how math equations can be fiddled with to get to the right answer. And that is the basis of all artificial neural networks, and a whole lot of AI and ML. [6]
Further Reading
Han, Su-Hyun, Ko Woon Kim, SangYun Kim, and Young Chul Youn. “Artificial Neural Network: Understanding the Basic Concepts without Mathematics.” Dementia and Neurocognitive Disorders 17, no. 3 (2018): 83. https://doi.org/10.12779/dnd.2018.17.3.83.
[4]:Some of you smarty pants would like to point out that a $>=$ wont ever give us the right answer but please sit down this is a loosely true mathematical explanation and other people get it, this isn’t for you anyway. ↩︎
[6]: Some of you might want to ask, say, what is artificial intelligence then? Well, AI, is just artificial neural networks making a whole lot of really really complex guesses and then self correcting them till they are kinda accurate. ↩︎
This is what a handwritten history sheet (which is where we write down the most important details that we extract from a patient or responder about the case). Source
And this is what the most popular EHR/EMR User interfaces that are meant to enter the same details as the history sheet look like
See a difference?
No, I don’t mean the ridiculous amount of clutter, the confusing information architecture or just the general pathetic state of these UIs.
I mean, the history sheet is not a system of transcribing what the patient said or transcribing what data was collected. It is a notation system, a language of its own, with data arranged spatially in a particularly way, symbols used to convey information, abbreviations, and a system to indicate what is important and a method for tracking the temporal profile of the case.
The EHRs are incoherent because they are talk-transcription interfaces. Not notational interfaces.
Imagine asking an artist who is composing music, to write down, in English, the notes in each chords, the tempo and what not. Would that looks coherent? Why would you do that when you have a musical notation system?
The clinical case file, handwritten is a few hundred year old technology that’s had iterative improvements in its quality and information architecture and it works! A case note written by an doctor in Bengaluru makes perfect sense when read by a doctor in Mumbai. It makes sense when read by a different specialist. It makes sense to nurses, to pharmacists to therapists of various kinds.
It doesn’t, however make sense to the administration. EHRs serve administrative needs. Not communicative.
Dr. Gowri Kulkarni addresses doctor attitudes and components of practicing telemedicine, Jasmine George of Hidden Pockets speaks of patient experiences, experiences with seeking sexual and reproductive health online and access to safe abortions using telemedicine.
5:40 — How telemedicine became so prominent due to Covid and how reproductive health issues were handled. Provider and patient perspective — (JG)
12:00 — How have doctors responded to telemedicine? Fears, expectations etc. (GK)
38:00 — How does one deal with the nuances of consent via telemedicine? (GK, JG, AP) This part is very important, and the discussion bring out a lot of the complexities involved in informed consent and technology, and the national health stack.
56:24 — Negotiating privacy and safety in crisis helplines and abortion and reproductive health (JG)
58:00 — Is this the future? What can Telemedicine do and not do? –(GK, JG)
1:07:00 — Unethical practices in clinical medicine now, why they exist and the future — (GK)
1:12:00 — The changing nature of the social contract with doctors, and the need for change in practices — (JG) A vital point here being made by Jasmine, she speaks of how instead of the tort approach, we need positive laws keeping stakeholders in consultation.
1:15:20 — Challenges in scaling telemedicine — (GK)
This post originally appeared on Karana’s blog , this iteration has a TOC, more references and has been edited to make things clearer.Many of these updates especially the footnotes were due to Dr. Verghese Thomas‘s comments.
Introduction
This post builds on the questions that were raised in the talk: Telemedicine Policies and Standards in India , adds more information, references, a detailed prescription and creates a reference – friendly structure.
The first part is descriptive. I will attempt to provide a clear understanding of the different aspects of telemedicine as it stands in India with regards to policy and infrastructure. In the second part I will focus on prescription, or describing what I think are the key issues and what I (and others) think should be done about them.
Part I – The description
Current State of regulations on telemedicine in India
Telemedicine is not new in India; for about 20 years, various governmental and non-governmental organizations have been involved in various kinds of telemedicine projects in India. The ISRO currently maintains a network of 130 hospitals in India that are connected for telemedicine [1].
Most of these are doctor to doctor consults, in which a doctor in a rural or secondary care setup discusses the care and treatment of a patient with a specialist based in one of the nodal centers, like AIIMS, or other central or state institutes. We do not have clear information about how many patients are treated overall, and what kind of outcomes are being measured and if there are any specific interventions being conducted in these centers. Mishra (2008, 2009) DeSouza (2014) and others have delved into this and more in detail.
Despite this history, there has been very little in terms of legal or official documentation about what services count as telemedicine, what kind of services can be provided via telemedicine (and what cannot), what the liability structure for these consultations are, and what clinical guidelines or standards apply. While anecdotal information about health worker adoption of these technologies abound there are not a lot of published reports from academia.
In the last 5 years or so, with the boom in the number of smartphone users [2] , a large number of mobile health providors have been doing telemdicine in India. The legal status of these is discussed frequently in media, Quora etc. But not a lot of clarity exists. As recently as 2019, the Karnataka medical counil notified all doctors to stop doing telemedicine, and opined that telemedicine consultations are illegal. [3]
During this time startups based in Bangalore sent details of the legal provisions and clinical audits and security practices to the KMC, which silently gave a go ahead, without really going public about it. I know this because I was involved in the making of these responses.
I have been working in health technology in India since 2010 and have been focusing on mobile-based telemedicine since 2016. Over the years I have collaborated with legal and other organizations to understand and frame the legal and ethical issues in telemedicine and have been involved some of the policy conversations around telemedicine in India. This post is a result of those experiences.[4]
The legislative or official backing of telemedicine providers is framed like this by most private providers:
The The Indian Medical Council Act, 1956 specifies who can practice medicine in India (registered medical practitioner), and what a legally valid prescription is.
This indicates that as long as any consultation is done by a registered medical practitioner and they provide a prescription following this standard it is a legal consultation.
Pharmacy Council of India which regulates training and registration of pharmacists and pharmacies in India, in its Pharmacy Practice Regulations, 2015 No. 14-148/ 2012- PCI defines that prescriptions can be physical or electronic.
The IT act of 2000 and its amendments(chapters 2, 5 and 7) describe how to digitally sign an electronic health record, and make it clear that this makes it a legally valid record.
Put these together and we have the broad framework under which you can have an online consultation.
In March 2020, The Medical Council of India published a guideline for telemedicine in India, which provides a broad practice framework. The board of governers has since accepted this guideline and has decided to provide statutory basis for them under Professional conduct, Etiquette and Ethics regulations. MCI-211(2)/2019 (Ethics)/201858 [PDF]
While this is far from comprehensive, it provides some way in which a doctor can be held liable for online malpractice, and provides clearer legitimacy to online consultations.
EHRs, regulations and their relationship with telemedicine
Since telemedicine creates and stores medical information about patients, all telemedicine providers who store data digitally ned to adhere to EHR standards. The definition of what and EMR is and what EHRs are are detailed in the MoHFW – released EHR standard 2016, and it is clear that any agency that stores patient information must comply with these standards.
Besides this, in 2018, the ministry has also set up a National Resource Centre for EHR Standard (NRCeS) to ” augment facilitation for adoption of the notified EHR Standards in technical association with Centre for Development of Advanced Computing (C-DAC), Pune for providing assistance in developing, implementing and using EHR standards effectively in healthcare Information Technology (IT) applications”. This organization has been working with vendors and creators of EHRs in providing training etc.
The EHR standard 2016, lays down the best practices for storage, retrieval, and communication of health information. It follows international standards in EHR design and explains the complexity of EHRs very well. The enforcebility of these standards is still unclear, as are penalties, if any that exist for not complying. There doesn’t seem to be any national certifying methodology or agency for EHRs.
Privacy and security of health data in India
Health data In India is owned by the patient, at least in the broad sense.
At present, the IT act of 2000 and its amendments are what form the legal basis of the right to privacy and security of personal information. This bill covers health information but is not very exhaustive.
The (PDP) bill provides for protection of personal data of individuals, and establishes a Data Protection Authority for the same. It defines what personal health information is and lays out penalties for breaches etc. But it also makes it clear that the government has ultimate power in making decisions about health data and lays out a large set of non-exhaustive circumstances or reasons for breaching consent. It also states that the government may ask “data fiduciaries to provide it with any: (i) non-personal data and (ii) anonymized personal data (where it is not possible to identify data principal) for better targeting of services.”
The law doesn’t speak of the right of a patient to be forgotten, and the entire system assumes the national health stack and which in turn is built on top of aadhar, and so anonymity doesn’t seem to be an option, and it very much wants every patient to be identified.
Consent
The current law [IT act] does not address the matter of consent very well. As a result of this, consent for using reusing, researching and doing what ever needs to be done is taken by most health apps upfront as part of the EULA . Chances are, if you clicked on one of those I Agree buttons, you’ve provided a blanket agreement for the use of your data. There is some distinction made about anonymizing and de-identifying data.
Anonymized data is data that has been stripped of all information that could be considered as personally identifiable. De-identified removes identifying information in a reversible manner, eg. replacing names with a unique code or number shown to some people, but separately maintaining a way to look up the name given the number or code.
The current legal framework gives software providers and other health providers almost unfettered access to data as long as it is anonymized, and doesn’t specify how often and in what situations consent must be taken.
In the PDP, consent is deliberated on in some detail and an XML standard for logging consent has been proposed.
India’s digital health infrastructure
Before going further into what the government is doing for the creation of digital health infrastructure, let me state that
The public health system in India is extremely good in some places and extremely bad in some places. And the difference between these places is not technology, it’s the way they solved the people problems. You cannot solve people problems with technology. Some examples of the people problems are caste, favoritism, and the informational, financial and power asymmetry between people who deliver healthcare and the people who receive it.
If we don’t address these structural issues first, and this is not something you can do in parallel, and you throw tech at it, there is plenty of evidence from high quality studies around the world, that this worsens the problem.
The current health system is extremely bloated in administrative and “overhead” areas and lacks resources in the “delivery” areas. This administrative bloat needs to be addressed and pruned. It cannot efficiently deliver what needs to be delivered, and I have felt many times that maybe we need to delete it all and restart.
An insurance scheme – which covers ~10 Cr families at 5 lakh per family.
In order to deliver this (and only this), the govt set up an independent body called the National health authority. The PMJAY website is very clear about its scope.
The Niti Ayog was entrusted with figuring out the tech for this – how to get health insurance to those who need it, and came up with the national health stack.[PDF] in July 2018.
In this proposal the NITI ayog starts off by saying,
In this document, we present the idea of a national health Stack (nhS)—a digital infrastructure built with a deep understanding of the incentive structures prevalent in the Indian healthcare ecosystem. The NHS, a set of building blocks which are essential in implementing digital health initiatives, would be “built as a common public good” to avoid duplication of efforts and successfully achieve convergence. Also, the NHS will be “built for nhpS but designed beyond nhpS” as an enabler for rapid development of diverse solutions in health and their adoption by states
Its Components:
A. National health electronic registries: to create a single source of truth for and manage master health data of the nation; (Suddenly we are not talking about 50 crore people.)
B. A coverage and claims platform with fraud detection;
C A Federated personal health records (PHR) Framework:
D. A national health analytics platform:
E. Other things including, Digital Health ID, Health Data Dictionaries and Supply Chain Management for Drugs, payment gateways etc shared across all health programs.
With all of this built on top of the aadhaar
In about 40 pages of the proposal it goes from being an insurance providing system for 50 crore people, to the central and unified system for accessing health for everyone in the country.
Since then, an organization called iSPIRT Foundation, which is a volunteer run non-profit, funded by some of the biggest names in the tech industry are currently going ahead and building the national health stack. In fact parts of it are already ready and the code’s on github.
From what I could gather from their website and materials provided by them, the organization takes a systems thinking approach and the volunteers clearly have experience in building tech infrastructure. They are very open about their work. It seems like the foundation has brought together smart minds and industry to work on creating an ecosystem for building digital products and business in India, including the NHS.
For the last few weeks they have been holding an open house discussing the NHS, parts of which have already been made! and a certification of some kind is in the works.
They are working with private industry very well, and I have reached out to health startups who mention that they are informed about the work being done and are generally happy about the quality of the discussions, although how far recommendations from policy, disability and patient rights organizations etc. are considered is unclear.
The Community
Before I jump into the prescriptive part of this post, I think it’s important to discuss some of the initiatives, communities and organizations that are involved in the discussion around telemedicine and digital health infrastructure of the country. This list is in no particular order and is not exhaustive. If you are an organization or community interested in this, please comment.
Jan Swasthya Abhiyan (JSA) The JSA forms the Indian regional circle of the global People’s Health Movement (PHM). They do a lot of advocacy around universal health coverage and gender and patient rights and commonly comment on health related legislation in India.
Digital health india, an NGO. No policy briefs so far, but a corona CDSS was made publicly available by them. They, in collaboration with the NRCeS have created and maintain a Telemedicine provider registry, which is a great project. It also conducts evaluations of telemedicine providers and the research is available on their website. Run by health and social work professionals.
Digital Health Providers Association, consisting of a few healthcare startups, has recently come up with a policy brief for telemedicine. Run by Health technology professionals.
Telemedicine society of india – The oldest Indian organization. Involved in conferences and research. Run by School of Telemedicine and Bio-Medical Informatics at Sanjay Gandhi Post Graduate Institute of Medical Sciences. No policy briefs, but many papers have been produced by them and they were early advocates of telemedicine. Run by doctors and health informatics professionals.
Software Freedom Law Center (SFLC) does a lot of advocacy, PILs, and other policy related work in the privacy, information security and related areas. Their responses to NDBP and others are well researched. Run by FOSS (Free/Libre and open source software) practitioners and has some academic background.
Center for internet and society (CIS) a non-profit organization that undertakes interdisciplinary research on internet and digital technologies from policy and academic perspectives. has produced some of the finest policy work when it comes to technology and digital living in india. Run by policy specialists and a research team.
Kaarana – organized the aforementioned talk and is involved with discussions around privacy, aadhar etc.
You should note that I have been unable to find academic departments or chairs in Indian universities who have responded to or been involved from the public’s side in all these discussions. I think there is a grave paucity of policy makers engaging in health technology in India.
Part II: The prescription
In this section I will list problems and my recommendations referencing all the elements discussed in the description section.
MCI Telemedicine Practice Guidelines
Overall, while the guideline was much needed and came at the right time, the guidelines seem hurried.
Issues
They fail to take in account the telemedicine that’s already happening in India. So it’s more a guide for someone new to this.
There are no background papers or surveys of existing practices in telemedicine in India as the foundation of this document.
It also tries to do too many things, and offers different levels of detail in different areas. For example, it mentions a list of medication that may or may not be used online. And it states that violating this directive can be construed as malpractice. This sort of an approach, where you dictate what medications are OK and what are not are not in line with research from other countries or with the dynamic nature of medicine. The guideline and the MCI should instead discuss safe and unsafe prescription habits. It already had to amend the list of drugs, because the first version made it illegal to prescribe psychiatric medication in India. Keep in mind that lack of access to psychiatry and mental healthcare are among the top five reasons people use telemedicine!
The MCI is also geared to come up with practice guidelines on how to manage different issues online, which I think is not a good idea, because the various medical academic societies need to think this through and come up with guidelines, and for this a fair bit of background research is needed.
There is a great need for a collaborative approach. There needs to be at least a few studies into what kind of things are already being treated online, what kind of people are accessing health this way, and understand the system before trying to govern it.
What direction is needed from MCI:
Create a collaboration with industry and academia in understanding how telemedicine can be delivered safely and efficaciously.
Identify research lacunae in clinical practice and policy and ethics of online consultations
Propose and study safe online prescription habits
Delineate what kind of training someone who practices telemedicine needs
Guide on how EHR and telemedicine providers can get ethical oversight from medical institutions.
In summary, instead of focusing on getting lost in the details , it should focus on creating a framework that is non restrictive and safe, and allows doctors to practice fearlessly, and promotes collaboration and research that leads to better guidelines and practice.
EHR standards
It has too many standards that apply, some 20 different standards of ISO referenced, all of which are paid. This sets the entry cost too high for smaller organizations. [5]
The recommendations for clinical terms- The snomed CT does not have Indian language versions of the or any localization. The point of a clinical terminology dictionary is to understand and communicate local health problems with the greater community. Use of SNOMED CT will causes loss of information, and preserving local languages of health is very important.
Also, while the government has bought access to SNOMED CT, this only applies to Government agencies, private players would have to pay thousands of dollars yearly to get access.
While it’s a very comprehensive document, it makes sets the bar too high for people making EHRs. I’m not saying we should be lax, I’m saying we need to be practicaal.
The other standards it recommends like the LOINC, have been mentioned in a way that hospitals with older machines and smaller labs would not be able to comply.
Recommendations:
Recommend free and open source standards where ever possible.
Recommend using standards that have some benefit. So far, using SNOMED terminology is beneficial to a very small subset of EHR providers.
Create or open the creation of India specific clinical terminologies, personal health information standards etc.
Understand that with the advent of modern algorithmic computing, the need for each entity to follow strict standards for terminology is going away. As long as locally accepted standard terminology is being used, interoperability can be established using other means.
Privacy and health data.
The blanket permission given to the government to use patient data without consent is in direct opposition to many learned commissions and committees constituted in this area, the supreme court rules on related issues and research into the importance of privacy in healthcare. [6]
There exist apps out there that do not mention clinical research as part of the EULAs but go head and do it anyway.
In part, this is because of the lack of ethical literacy among technologists. To be clear, I am not saying us technologists are an unethical lot, but it seems like ethics is not part of the CS curriculum, and tech till recently maintained that they were just tool builders and didn’t have to worry about the effects.
Over the years, my experience in bringing up ethics in software circles has not be wonderful, mainly because there just isn’t enough literacy about the issue and because ethics are often confused with moral policing.
We need to keep in mind that beyond the lack of literacy, here is plenty of current data and research into the harms that are being caused by unethical practices in technology or 7ignoring of ethics in technology. [7]
While there has been some work in the area of teaching software professionals in making and using EHRs, there has been no talk of ethics in this policy space.
Neither the agencies dealing with EHRs nor any documents from Niti Ayog, which leads the policy making, have any mention of the need for ethical literacy for software makers or mention ethical oversight of digital health providers.
With the advent of AI, there is now a lot of evidence that just removing someone’s name and such details doesn’t actually anonymize data. Further steps need to be taken and this is an area that will keep needing to catch up with various misuses of data and so needs a flexible framework.
Recommendations:
Data ethics literacy for health technologists. This needs to be part of the computer science curriculum, and periodically discussed and dealt with in organizations.
Ethical oversight of health data providers. Academia and ethical experts in the country need to make it easy for digital health providers to access their expertise and financially viable to receive ethical oversight for research and development.
We clearly need a LOT more focus on individual rights and the evidence for this in the healthcare context, and our laws need to be informed of the advances in this area. The current law and the proposed Data protection law fall short in reassuring people that their interests are being taken care of.
The national health stack and the NDHB
One of the foundational assumptions of this stack is that the identity of the individual MUST be verified via aadhaar, or other methods.
The issues with the national digital health blueprint whose problems have been explored in detail in a talk at Kaarana and there are comments and reports on it available.
Comments by JSA, SFLI and CIS in particular stand out, and not with any coordination, they all point out the problems of consent, inclusion, and privacy.
It could work- but more often than not, as global experience shows it does not- though in the process it could provide many lucrative contracts to India’s IT majors. In a worst case scenario it could disrupt not only an ongoing incremental process of IT development that is ongoing, but also the organization of healthcare services at the district and sub-district levels- especially when new systems are being proposed as replacing all others. An approach where the biggest and newest software seeks to undermine or stop all others, even if they may be working well in their local settings is one reason- why some of these bold new ‘disruptive” innovations- can be literally disruptive of progress being made, without offering any alternative.
We therefore would call for an incremental approach that builds on the current situation and processes, with center providing technical support and guidance to multiple decentralized efforts. We set out some of the main features of such an alternative below
The main purpose of IT systems in the states and districts should be for decentralized management at that level.The center should limit itself to data that is actionable for the center-it need not be able to “see” every facility, let alone every individual
A central repository is neither required nor manageable nor desirable.Though these repositories are justified in the name of universal coverage and reaching the poor, it will like most such systems provide little in the way of entitlements to the poor. However in the hands of a powerful state, it can be used to encroach on privacy harms elect individuals who are perceived as hospital by the government of the day. Such large data banks have also commercial value and there is much data mercantilism-on which the entire document is silent. This silence is of great concern.There needs to be safeguards and guarantees against this.
The Government of India has formed multiple committees and held multiple rounds of consultations to decide upon the issue of Privacy and Data Protection. Justice A.P. Shah Committee formed by the Planning Commission released a report on privacy in 2012.[1] In its report, nine National Privacy Principles were recommended.[2] In 2017, a nine-judge bench of the Supreme Court of India unanimously recognized the existence of a fundamental right to privacy under Article 21 of the Constitution of India
The pressing concern with the National Digital Health Blueprint (NDHB) report is that it suggests a framework that severely infringes upon the fundamental right to privacy. These concerns are heightened in the absence of a comprehensive data protection law. The report also ignores a series of advancements on privacy and data protection that have taken place over the years. It does not adhere to the privacy principles recommended by Group of Experts on Privacy (Justice A.P. Shah Committee) and the more recent, Justice B.N. Srikrishna Committee report whose recommendations on data protection form the core foundation for the draft Personal Data Protection Bill, 2018.
A detailed analysis of the National health stack has been done by Smriti Mudgal Sharma
In conclusion it may be said that NHS is a great move towards monitoring and evaluation of the implementation of ABY. However, technology can at best streamline processes and help create a digital backbone for execution of public health programmes; it alone cannot solve the greater public health challenges. This endeavour needs to be complemented by strengthening the implementation capacity of states. The real need of the hour is to fix accountability of the medical professionals, improve standards of care, ensure transparency, and procure high-quality data without compromising privacy and choice of beneficiaries.
We also note that the nature of data which would be subject to processing in the proposed digital framework pre-supposes a robust data protection regime in India, one which is currently absent. Accordingly, we also urge ceasing the implementation of the framework until the Personal Data Protection Bill is passed by the parliament. The NDHB also assumes that access and delivery of the services promised under the ecosystem would be facilitated by the prospect of ‘near universal coverage’ of smart phones across India. However, this ‘mobile first’ premise rests on an assumption of widespread digital literacy, which is simply absent when one considers the social realities of the country.
Section 3.5 of the NDHB states the standards that will be in place for privacy and security, which includes provisions that are to be included in the operational aspects. This includes a provision on immutability, which states that a record cannot be deleted without following due process. We recommend that such due process takes into consideration the right of the data principal to delete specific entries or the entire set of records containing their personal information. We had also made this recommendation for the Digital Information Security in Healthcare Act 2018 49 , and reiterate it for the NDHB.
We need digital infrastructure but this (NHS and NDHB) do not address systemic inequalities which are the root cause of the problems this system is trying to solve. This is foolhardy and suspiciously represents and solves the problems of the industry and not the patient.
We need good data protection provisions in our laws, and without that, there is great deal of misuse that can happen due to this stack and the blueprint.
This is creating a system that might perpetuate the exclusion that pervades health and industry in India.
The Ispirt foundation.
Pretty much all learned groups so far have opined that before embarking on this glorious project we need to
Address systemic inequalities, and don’t ignore the fact that the lack of tech is not the core issue with health delivery in India.
Improve the data protection standards in India, pass strict data protection laws and then start this project
However, as I already pointed out, a non-profit with no official links to the Niti Ayog or the MoHFW is currently holding consultations with the industry for building the NHS and has already built parts of it.
Some questions that I am unable to find any answers for in official documentation or RTIs filed on this issue by others are
Who appointed them? What was the process? Who entrusted them with this highly complex job of creating a digital infrastructure for India before we finished discussing what infrastructure we really need?
Who do they answer to?
Do I and other civil society organizations have a right to be heard by them?
Is it even legal to start building the NHS using an informal agreement when neither the NDHB is finalized, nor are the laws around data protection passed?
Neither their website, nor the publications from Niti Ayog or MoHFW have any clues to give us.
What we see here is an organization that uses public resources and is creating public goods, but has no accountability to the public.
We do not know if the government designed and operates this or it has been subcontracted to them
You could say that has been designed to “get things done” and avoid the red tape.
Which is great if you’re building one app, but when you’re building national infrastructure, and if you are outside the purview of the RTI act, or any parliamentary oversight, and you are funded by a small group of tech billionaires, there is a problem.
Overall Recommendations
Transparency about who is building the NHS and who they are accountable to and if this is even legal
Create Systems that make consultative progress easier – I would love to have signed up for a newsletter that tells me that comments are elicited on a health policy related issue from the govt. or its organizations.
For the industry and the folks at NITI etc. to understand that consultative building, doesn’t mean slow, it means deliberate and harm reducing and exclusion free. The voices of the most vulnerable people in this nation are not being represented or consulted with while designing a system for them
Civil society, policy specialists, activists, FOSS proponents – Participate – join ispirt consultations, and listen and comment. Get involved.
For all these groups working in isolation to start talking to each other. Like the people’s health movement, we need a coalition of health technologists, policy specialists and health advocates.
References
Mishra SK, Kapoor L, Singh IP. Telemedicine in India: current scenario and the future. Telemedicine and e-Health. 2009 Jul 1;15(6):568-75.
Mishra SK. Current status of E-health in India. Retrieved from openmed. nic. in/1265/01/skm12. pdf on. 2008;30(06).
DeSouza SI, Rashmi MR, Vasanthi AP, Joseph SM, Rodrigues R. Mobile phones: The next step towards healthcare delivery in rural India?. PloS one. 2014 Aug 18;9(8):e104895.
[2]: I maintain an updated scientific bibliography called the “case for telemedicine” , you’re welcome to comment. ↩︎
[3]: Objective and good quality data on this is sparse as it’s largely large consulting firms that have provided numbers, here’s one figure to explain this from the McKinsey Digital India report of 2019 ↩︎
The longer I work in clinical medicine, the firmer my belief that a truly patient-centrist health system can be built only if we move away from hospital-centric medicine and let patients take charge of their health.
We need to consciously/purposefully move towards a clinical model where parts of the decision making process are augmented and even replaced by the machine, and happen between the patient and health-care tech. Making humans do the things we are so bad at makes no sense when we can have machines do it better.
People’s health in people’s hands means reducing the intervention and power doctors and health professionals have in the care-giving.
I see this increasingly in primary care where so much of the issues do not need a medical intervention. So many of these encounters are for answering “is this a serious problem?” and “make me well right now”.
Meeting a health professional if you have a non-serious issue is bad for both patient and doctor. Going to a doctor with a Upper respiratory infection increases your likelihood of getting an (unnecessary) antibiotic manyfold (stop justifying this, please).
On the medical side — we know that there are a lot of things GPs should be doing that they don’t have the time for, how about we welcome those things that truly do give us time — automation?
I long for a day when my patient can ask his mobile phone app if the sore throat he has is a simple viral thing or if he needs to see a human clinician (which he rarely does) and make an informed decision.
I don’t think that day is far.
PS: We have over the counter drugs, but still need medicos to tell people when/how to use them.
I heard the term “outercourse’ for the first time in college. We were in a consultation room and in walks a resident to discuss a case with the prof. He says “sir patient has done outercourse and wants to take ipill”.
The prof gave him a lecture about how you don’t have to discuss every case and anyone who asks for contraception should just be given contraception.
Then when he looked at us and our puzzled faces, he explained what it meant, and then a comment on many things. “All these kids claims to have only done outercourse, but so many become pregnant, must be indian fertility”.
