You have some false assumptions in your argument, mainly that it's just as easy to train and distribute humans across Africa (and elsewhere) to do malaria diagnosis as it is to develop and distribute copies of this software.
Moreover, as the software is improved, updating the software -- and raising the baseline of quality -- is a unique advantage over the "just use humans" solution.
The scenario is not "download an app and it detects malaria".
The scenario is "person trained to work with blood samples without spreading/contracting HIV deployed into field packing a small digital microscope and support software built on WP7." This person is likely already doing a variety of other community medicine and is the closest thing to a medical doctor many of their patients will ever see.
There's no argument against giving the practicioner a helpful tool, only that it may be automating an insignificant part of the overall problem (looking at the image).
Actually, watching the video now rather than just reading the text, the software does do image analysis to determine whether or not malaria is present.
The fact that it's a smartphone, or that it's running Windows 7, or that it was developed in Silverlight doesn't seem to have anything to do with the fact that we're looking at a glorified microscope.
Surely there's a cheaper way to get a microscopic image of a blood sample in Africa than flooding villages with Windows smartphones.
If this "glorified microscope" allows people to do their jobs tangibly better -- in this case, detect malaria -- then what exactly is the cost problem in your mind? We're not talking about gold-plated custom-designed malaria detection devices... we're talking about (commodity) smartphones with a small lens affixed and some software installed.
Do we know that it helps people do their jobs better? The article left me with this question: Is this different than other microscopes that you could use in the field, and if so, how? Does this make it easier/cheaper for the trained personnel to detect malaria? Maybe it does--I'm just not clear on that point after having read the article.
The real point is that now the mobile worker doesn't have to carry a phone + a large microscope. The mobile phone is continuing to replace more and more objects. Radio, tv, camerea, video, internet browser, and now microscope. Many more devices will be adapted to take advantage of the phones computation capability as well as things like GPS, internet etc. BTW, if that mobile health worker has to make a choice between purchasing a microscope and purchasing a phone, they are probably going to purchase the phone.
It's a Boston-centric publication. Why wouldn't they anchor it to one of Boston's best known institutions?
If it said "Harvard grad student" there would be far less vitriol here. The fact that it is someone at HBS getting some startup press really pisses you off - doesn't it?
Wealthy Africans get malaria too, and besides, even the poor will be getting smart phones soon.
It's going to be like cell phones which were pretty ubiquitous in the third world years ago when it wasn't even the case here. They bypassed landlines and went straight to cell, and the same will probably be true of computers and smart phones.
The current tests have very low accuracy (the article says 40%). These tests are likely being administered by trained professionals, and if the cost of each test is sufficiently high (for materials, and that it would take several tests to get a reasonably accurate answer), it might be cheaper in the long run (or even short periods of time) to buy these medical professionals smartphones and the lenses instead of the supplies for existing tests. If the cost of a smartphone and lens is greater than a full microscope with a webcam attached (to allow the computer or a dedicated device to do the analysis), then smartphones are still a useful proof of concept as well as being far more portable and being easier to power.
It doesn't seem logical on the face, but once you start adding up costs it's quite likely that it would be, especially if they can convince a smartphone manufacturer to donate a few hundred or thousand units (I'm sure that there are smartphone manufacturers donating money to the anti-malaria efforts already, as it's good publicity and can be used in advertising. Donating smartphones instead of money is a very small step, as well as possibly cheaper - the retail price for a smartphone is not the price of manufacture).
Yeah, you're missing a lot. I don't know where to start. A lot of people in poor tropical countries live in dilapidated shacks but have a TV and smartphones. And anyone, rich or poor, in these countries can get malaria. A friend of mine died of malaria that he contracted in Africa, and he was a millionaire who only stayed in the best hotels.
If it works effectively, you wouldn't really need that many smartphones. A lot many countries go by with shared resources for many areas to get things done.
E.g. In villages in India or Africa etc., there might be one person with a computer and an internet connection who might offer his/her services as a co-operative for the residents of the village. The same way, one smartphone equipped with the relevant applications per district or a village can effectively tender assistance to all those in need from the surrounding areas.
The article claims that 90% of 8+ million childhood deaths in sub-Saharan Africa are caused by malaria, but the World Health Organization says it's responsible for only 20% of child deaths. Moreover, the latest estimates for annual deaths from malaria (all ages) is 780,000. In any case, it's a lot, and I'm glad this team has delivered a useful tool!
The sad part is: This app is basically useless. Malaria has very distinct symptoms. Malaria also has a 3-day pre-roll period where you know you are about to get malaria, but you are not yet in bed. Also, Malaria is the #1 drug every chemist stocks in malaria-infested areas. Furthermore, taking a malaria pill even if you don't have malaria has no side effect.
What this means is that if all the criteria of bad luck come together:
1. Person does not know the symptoms for malaria
2. There is no chemist around who stocks malaria pills
3. The illness for some reason instantly hits the person without the pre-roll period
...then the person for whom all these criteria comes into question is about 99.99% not likely to have an Android phone around. And 100% not likely to have a lifelens attached to the phone.
Hopefully this sort of technology is expanded upon, with two directions possible. The first is creating more self-diagnosis tools for smartphones (more relevant to those who are likely to own one). The second is creating tools for those 99.999% not likely to have an Android phone around (and perhaps to help doctors/nurses/other aides).
CellScope, based in Berkeley, came up with a way to attach a smartphone to a microscope a few years ago in order to test for TB in remote places and have doctors elsewhere analyze results [ BBC story: http://news.bbc.co.uk/2/hi/technology/8161775.stm ].
Their next device is an otoscope that attaches to the iPhone so you can take pictures of ear infections [ see the CellScope profile on LAUNCH: http://lnch.is/qTqQK0 ].
Whether the malaria app works or not, I have no doubt we'll see more health-related hardware and apps for mobile phones.
The mainstay of malaria diagnosis has been the microscopic examination of blood. Although blood is the sample most frequently used to make a diagnosis, both saliva and urine have been investigated as alternative, less invasive specimens.
Areas that cannot afford laboratory diagnostic tests often use only a history of subjective fever as the indication to treat for malaria. Using Giemsa-stained blood smears from children in Malawi, one study showed that when clinical predictors (rectal temperature, nailbed pallor, and splenomegaly) were used as treatment indications, rather than using only a history of subjective fevers, a correct diagnosis increased from 2% to 41% of cases, and unnecessary treatment for malaria was significantly decreased.
It certainly sounds as there is use for diagnostics tools.
You still don't get it. It's not neccessary to diagnose for malaria via blood. Malaria patients know they have malaria in almost all cases, because they are familiar with the disease. As soon as the symptoms show, they take the tablets. If the problem is not fixed, then it was not malaria. But if you live in a malaria area and you have the typical malaria fever, 99.9% of the time it will be malaria.
Blood diagnosis is unnecessary and expensive. Just take the pills.
I'm no expert. I don't claim to be. My point is that people actually appear to do testing. Now maybe you should relay your message to these testers, but it doesn't appear to be uncommon.
In fact, there's a whole Wikipedia page dedicate to Malaria testing:
Again, I'm not expert (or even novice), but looking at the literature and what people in the field are working on -- it looks like they make use of diagnostic tools for Malaria. I've certainly seen no papers saying that diagnostics isn't useful in 99.99% of all cases, just pop a pill. Not saying that that's not what should be done, but I find the evidence thus far lacking.
The people who need to diagnose it are in a hospital, where they can just use one of many methods. They don't need to to peer through an android lens. Laymen rarely need to diagnose malaria because the symptoms are very very clear.
From the above: "Effective and affordable treatment is recommended for all cases of malaria within 24 h of the onset of illness. Most cases of "malaria" (ie, fever) are self-diagnosed and most treatments, and deaths, occur at home. The most ethical and cost-effective policy is to ensure that newer drug combinations are only used for true cases of malaria. Although it is cost effective to improve the accuracy of malaria diagnosis, simple, accurate, and inexpensive methods are not widely available, particularly in poor communities where they are most needed."
Additionally people apparently can build up a resistance to the Malaria treatment. See:
“When people are sick in Mali, the doctor will usually tell them they have malaria whether or not they test for it,” said Fatou Faye, an infectious diseases researcher and trainer at a privately funded medical laboratory, the Charles Merieux Centre in Bamako.
“The patients then buy anti-malarial drugs in the street and build up a resistance to treatment.”
As a result, according to research by Dr. Imelda Bates at the Malaria Knowledge Project (MKP), part of the Liverpool University School of Tropical Medicine, this means people miss other causes of feverish illness such as pneumonia and meningitis, which can cause further illness and even death.
Again, I'm not saying your wrong necessarily, but everything I read seems to point to the fact that low-cost accurate Malaria diagnosis would be useful.
Actually you're telling him that he's acting exactly like people commonly do in malarial areas, and are informing him that simply taking that course of action breeds worse varieties of malaria, and causes people to die of unrelated diseases that could have been treated.
This strongly suggests that he has practical experience with malaria. And is useful information for him about how the advice he has based on experience is suboptimal.
Medicine is a game of probability. If you have malaria symptoms in a malaria area, the chance of you having malaria are around 99%. Every other disease with similar symptoms (e.g Typhoid) is pretty rare.
When it comes to peoples lives, people very quickly reach the optimal solution. It's what humans do best. This android solution is a theoretical solution from people who have never experienced malaria and don't even live in an area where malaria exists.
Yes, I have practical experience with malaria. I also have practical experience with android phones. I am actually qualified to make experience-based judgments on the feasibility of that tool, and quoting theoretical papers may win you arguments on the internet, but it will have little practical usage.
When there is a disease in an area that has existed for so long that people have built genetic resistance to it, at the cost of having occasionally sickly children, then trust me, modern methods that work will quickly be selected.
Modern methods that don't work (bed mosquito nets), will be selected away.
Armchair solutions for problems you only partially understand are pointless.
From WHO/USAID report on Malaria detection (executive summary):
"In malaria patients, a prompt and accurate diagnosis is the key to effective disease management. The two diagnostic approaches currently used most often, however, do not allow a satisfactory diagnosis of malaria. Clinical diagnosis, the most widely used approach, is unreliable because the symptoms of malaria are very non-specific. Microscopic diagnosis, the established method for laboratory confirmation of malaria, presents technical and personnel requirements that often cannot be met, particularly in facilities at the periphery of the health care system. In addition, delays in the provision of the microscopy results to the clinician mean that decisions on treatment may be taken without the benefit of the results. Thus, the recent introduction of rapid diagnostic tests (RDTs) for malaria is of considerable interest."
The report is completely disconnected from reality. Malaria patients often have very clear symptoms, and at the point where a diagnosis is necessary, a blood test WILL have to be done. Malaria is a very common disease, and there are very clear and available treatments, and the symptoms are very well known. The blood tests are also very available.
I have had malaria twice. When you have it you definitely know you have the disease. Your legs shake from side to side. You can diagnose disease as well becuase your temperature goes in a zig zag patern over a 10 to 20 min cycle. none of these methods will tell you what kind of maleria you have, which is usefull to know what the best treatment is. Sometimes this is not necessary as there is only one kind in the region. It should not be under estimated how hard it is tell one type of maleria from the other. The last time I had it took 4 hours for a major hospital in London to identfy that I had 3 types of parasite in my blood.
Saying that Android penetration is small now means nothing regarding growth potential. Mobile phone usage in sub-Saharan Africa has exploded over the past decade and is responsible for greater mobile banking penetration than there is in the US (despite our incredible Android penetration).
The fact that this technology is being built means that mobile clinic operators can take advantage of faster, easier, cheaper, more accurate diagnostics in the field. When will an everyday consumer enjoy the ability to self-diagnose a blood-borne disease? When someone gets AIDs in the US, can they self-diagnose?
If the scan can be expanded to other blood-borne diseases, it could potentially be useful. But it's pointless for malaria.
Malaria is a bit unique among diseases in that it's a parasite and not a virus or bacteria. I'm not sure this method extends to non-parasitic diseases.
However, I see a few practical problems when these high end phone get distributed in Africa by some aid organization. Because an expensive high end phone is also a good gift. I hope they will maintain statistics of how many of these high end phones actually reach the medical centers. And how many of them remain in the med center after a year. Sometimes its better to just send a microscope that can not be used for anything else but blood samples.
Okay, I just spent the better part of an hour working on a comment on this (hey, I just passed a hematology course, sooo... ;), and then, as I finally got around to checking out the website, I catch this line in the video (emphasis mine):
"All that's really needed from an individual is a drop of blood from the finger to create a thin smear on a slide for analysis with standard protocols."
That's a lot different from what the linked article claims (again, emphasis mine):
"A single drop of blood from the patient is required then smeared on the microscopy lens."
Smearing the lens directly seemed an interesting idea, although it poses a few challenges I'll touch on shortly. However, a close listen to the video makes it clear that's not what the project purports to do. The Lifelens "hardware" -- the lens -- is just meant to give a phone camera microscopic capabilities. How well that works isn't too clear, but it's definitely not the same thing.
It would actually be quite difficult to get an acceptable specimen on a tiny lens. The standard laboratory method is to place a drop of blood on a slide, then use another slide to draw it out into a thin film. Here's a quick video that shows a pretty good example:
Note how the tech drags the slide back at an angle over the drop, then pushes it forward almost immediately. It takes good timing and a lot of practice to get it right, and then you have to find just the right part of the specimen for a proper analysis. What you get in a typical thin smear is a range of population densities, from a clumpy mess of cells crowding and overlapping each other, to rarefied lines and clusters of cells spaced widely apart. The nice, evenly spaced fields shown in the article are just a small region of the full sample, usally a couple fields behind the feathered edge. Getting that is a challenge just in the laboratory. On a tiny lens in the field? I'll believe that when I see it, and this is just the second step! (Getting the sample is the first) The specimen still has to be stained.
While RBCs with their hemoglobin do en masse make blood red to reddish-purple, individual red cells don't actually show a lot of color without staining. Nor do white cells; they just look a bit larger and grainier than reds. To better examine cell structures, a stained specimen is needed. The standard analytical sample is a Wright stained thin film smear, which requires three different stains be applied and washed in sequence. This stuff doesn't just wash off, either, so don't expect to use the same lens twice. Again, however, that's not what the project actually claims to do. What it does claim, on the other hand, strains credulity in other directions.
Sad to say, but a half-sphere lens on a phone camera does not a microscope make. Without going too deep into details (I hope), proper microscopy requires close attention to the properties of light and lenses. You can't just point a powerful lens at a specimen and expect to get a good image. You also need to ensure sufficient lighting, proper contrast, and control your depth of field!!!
I cannot emphasize that last point enough! Even squashed nearly flat under a cover slip, any specimen you examine is still three dimensional, with details you can only see by shifting the microscopic stage up and down by very small amounts. The depth of field is very narrow. A professional scope's fine adjustment is fine enough to shift focus on a cell from "front" to "back" and anywhere in between. As you might imagine, that requires keeping the stage very, very still. That's why scopes have specimen stages. You won't get that out of a phone camera without clamping it down.
That's what made the article's statement so interesting. If an acceptable specimen could be obtained by smearing the lens directly, the fixed position might allow for proper analysis, but ...
53 comments
[ 4.3 ms ] story [ 119 ms ] threadSo...it's a microscope lens for a smartphone? Isn't that something already developed? What about this app "detects malaria"?
You're going to need a trained human in the loop just to handle the blood and apparatus in a sanitary way.
Humans are inherently better than any app at the image task, so why not just let the human do it?
This looks like a non-solution to me.
Moreover, as the software is improved, updating the software -- and raising the baseline of quality -- is a unique advantage over the "just use humans" solution.
The scenario is "person trained to work with blood samples without spreading/contracting HIV deployed into field packing a small digital microscope and support software built on WP7." This person is likely already doing a variety of other community medicine and is the closest thing to a medical doctor many of their patients will ever see.
There's no argument against giving the practicioner a helpful tool, only that it may be automating an insignificant part of the overall problem (looking at the image).
Surely there's a cheaper way to get a microscopic image of a blood sample in Africa than flooding villages with Windows smartphones.
I believe you mean "HBS Student Takes Credit for People who Created an App to Detect Malaria".
If it said "Harvard grad student" there would be far less vitriol here. The fact that it is someone at HBS getting some startup press really pisses you off - doesn't it?
</sorry-yes-cynical>
If these people could afford smart phones, they probably wouldn't have a huge malaria problem.
Am I missing something?
It's going to be like cell phones which were pretty ubiquitous in the third world years ago when it wasn't even the case here. They bypassed landlines and went straight to cell, and the same will probably be true of computers and smart phones.
It doesn't seem logical on the face, but once you start adding up costs it's quite likely that it would be, especially if they can convince a smartphone manufacturer to donate a few hundred or thousand units (I'm sure that there are smartphone manufacturers donating money to the anti-malaria efforts already, as it's good publicity and can be used in advertising. Donating smartphones instead of money is a very small step, as well as possibly cheaper - the retail price for a smartphone is not the price of manufacture).
E.g. In villages in India or Africa etc., there might be one person with a computer and an internet connection who might offer his/her services as a co-operative for the residents of the village. The same way, one smartphone equipped with the relevant applications per district or a village can effectively tender assistance to all those in need from the surrounding areas.
http://www.who.int/features/factfiles/malaria/en/index.html
http://healthland.time.com/2011/04/26/malaria-deaths-are-dow...
What this means is that if all the criteria of bad luck come together:
1. Person does not know the symptoms for malaria
2. There is no chemist around who stocks malaria pills
3. The illness for some reason instantly hits the person without the pre-roll period
...then the person for whom all these criteria comes into question is about 99.99% not likely to have an Android phone around. And 100% not likely to have a lifelens attached to the phone.
Their next device is an otoscope that attaches to the iPhone so you can take pictures of ear infections [ see the CellScope profile on LAUNCH: http://lnch.is/qTqQK0 ].
Whether the malaria app works or not, I have no doubt we'll see more health-related hardware and apps for mobile phones.
The mainstay of malaria diagnosis has been the microscopic examination of blood. Although blood is the sample most frequently used to make a diagnosis, both saliva and urine have been investigated as alternative, less invasive specimens.
Areas that cannot afford laboratory diagnostic tests often use only a history of subjective fever as the indication to treat for malaria. Using Giemsa-stained blood smears from children in Malawi, one study showed that when clinical predictors (rectal temperature, nailbed pallor, and splenomegaly) were used as treatment indications, rather than using only a history of subjective fevers, a correct diagnosis increased from 2% to 41% of cases, and unnecessary treatment for malaria was significantly decreased.
It certainly sounds as there is use for diagnostics tools.
Blood diagnosis is unnecessary and expensive. Just take the pills.
In fact, there's a whole Wikipedia page dedicate to Malaria testing:
http://en.wikipedia.org/wiki/Malaria_antigen_detection_tests
And they note more than 20 commercial products for sale that test for Malaria.
Here's a paper from the NIH on the importance and state of rapid Malaria diagnosis:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2223842/
Here's another NIH paper talking about an Malaria eradication, and a big part of their plan is diagnosis:
http://www.ncbi.nlm.nih.gov/pubmed/21311583
Again, I'm not expert (or even novice), but looking at the literature and what people in the field are working on -- it looks like they make use of diagnostic tools for Malaria. I've certainly seen no papers saying that diagnostics isn't useful in 99.99% of all cases, just pop a pill. Not saying that that's not what should be done, but I find the evidence thus far lacking.
"Clinical diagnosis, the most widely used approach, is unreliable because the symptoms of malaria are very non-specific."
From the last NIH paper I referenced, "Many of malaria's signs and symptoms are indistinguishable from those of other febrile diseases."
Here's a paper from the Lancet about Malaria misdiagnosis:
http://www.ncbi.nlm.nih.gov/pubmed/15555670
From the above: "Effective and affordable treatment is recommended for all cases of malaria within 24 h of the onset of illness. Most cases of "malaria" (ie, fever) are self-diagnosed and most treatments, and deaths, occur at home. The most ethical and cost-effective policy is to ensure that newer drug combinations are only used for true cases of malaria. Although it is cost effective to improve the accuracy of malaria diagnosis, simple, accurate, and inexpensive methods are not widely available, particularly in poor communities where they are most needed."
Additionally people apparently can build up a resistance to the Malaria treatment. See:
http://www.irinnews.org/Report.aspx?ReportId=78014
“When people are sick in Mali, the doctor will usually tell them they have malaria whether or not they test for it,” said Fatou Faye, an infectious diseases researcher and trainer at a privately funded medical laboratory, the Charles Merieux Centre in Bamako.
“The patients then buy anti-malarial drugs in the street and build up a resistance to treatment.”
As a result, according to research by Dr. Imelda Bates at the Malaria Knowledge Project (MKP), part of the Liverpool University School of Tropical Medicine, this means people miss other causes of feverish illness such as pneumonia and meningitis, which can cause further illness and even death.
Again, I'm not saying your wrong necessarily, but everything I read seems to point to the fact that low-cost accurate Malaria diagnosis would be useful.
Actually you're telling him that he's acting exactly like people commonly do in malarial areas, and are informing him that simply taking that course of action breeds worse varieties of malaria, and causes people to die of unrelated diseases that could have been treated.
This strongly suggests that he has practical experience with malaria. And is useful information for him about how the advice he has based on experience is suboptimal.
When it comes to peoples lives, people very quickly reach the optimal solution. It's what humans do best. This android solution is a theoretical solution from people who have never experienced malaria and don't even live in an area where malaria exists.
Yes, I have practical experience with malaria. I also have practical experience with android phones. I am actually qualified to make experience-based judgments on the feasibility of that tool, and quoting theoretical papers may win you arguments on the internet, but it will have little practical usage.
When there is a disease in an area that has existed for so long that people have built genetic resistance to it, at the cost of having occasionally sickly children, then trust me, modern methods that work will quickly be selected.
Modern methods that don't work (bed mosquito nets), will be selected away.
Armchair solutions for problems you only partially understand are pointless.
From WHO/USAID report on Malaria detection (executive summary):
"In malaria patients, a prompt and accurate diagnosis is the key to effective disease management. The two diagnostic approaches currently used most often, however, do not allow a satisfactory diagnosis of malaria. Clinical diagnosis, the most widely used approach, is unreliable because the symptoms of malaria are very non-specific. Microscopic diagnosis, the established method for laboratory confirmation of malaria, presents technical and personnel requirements that often cannot be met, particularly in facilities at the periphery of the health care system. In addition, delays in the provision of the microscopy results to the clinician mean that decisions on treatment may be taken without the benefit of the results. Thus, the recent introduction of rapid diagnostic tests (RDTs) for malaria is of considerable interest."
This android device is solving a solved problem.
The fact that this technology is being built means that mobile clinic operators can take advantage of faster, easier, cheaper, more accurate diagnostics in the field. When will an everyday consumer enjoy the ability to self-diagnose a blood-borne disease? When someone gets AIDs in the US, can they self-diagnose?
The need exists and this is a great step forward.
Malaria is a bit unique among diseases in that it's a parasite and not a virus or bacteria. I'm not sure this method extends to non-parasitic diseases.
However, I see a few practical problems when these high end phone get distributed in Africa by some aid organization. Because an expensive high end phone is also a good gift. I hope they will maintain statistics of how many of these high end phones actually reach the medical centers. And how many of them remain in the med center after a year. Sometimes its better to just send a microscope that can not be used for anything else but blood samples.
Okay, I just spent the better part of an hour working on a comment on this (hey, I just passed a hematology course, sooo... ;), and then, as I finally got around to checking out the website, I catch this line in the video (emphasis mine):
"All that's really needed from an individual is a drop of blood from the finger to create a thin smear on a slide for analysis with standard protocols."
That's a lot different from what the linked article claims (again, emphasis mine):
"A single drop of blood from the patient is required then smeared on the microscopy lens."
Smearing the lens directly seemed an interesting idea, although it poses a few challenges I'll touch on shortly. However, a close listen to the video makes it clear that's not what the project purports to do. The Lifelens "hardware" -- the lens -- is just meant to give a phone camera microscopic capabilities. How well that works isn't too clear, but it's definitely not the same thing.
It would actually be quite difficult to get an acceptable specimen on a tiny lens. The standard laboratory method is to place a drop of blood on a slide, then use another slide to draw it out into a thin film. Here's a quick video that shows a pretty good example:
http://www.youtube.com/watch?v=iA6ce-3sYgk
Note how the tech drags the slide back at an angle over the drop, then pushes it forward almost immediately. It takes good timing and a lot of practice to get it right, and then you have to find just the right part of the specimen for a proper analysis. What you get in a typical thin smear is a range of population densities, from a clumpy mess of cells crowding and overlapping each other, to rarefied lines and clusters of cells spaced widely apart. The nice, evenly spaced fields shown in the article are just a small region of the full sample, usally a couple fields behind the feathered edge. Getting that is a challenge just in the laboratory. On a tiny lens in the field? I'll believe that when I see it, and this is just the second step! (Getting the sample is the first) The specimen still has to be stained.
While RBCs with their hemoglobin do en masse make blood red to reddish-purple, individual red cells don't actually show a lot of color without staining. Nor do white cells; they just look a bit larger and grainier than reds. To better examine cell structures, a stained specimen is needed. The standard analytical sample is a Wright stained thin film smear, which requires three different stains be applied and washed in sequence. This stuff doesn't just wash off, either, so don't expect to use the same lens twice. Again, however, that's not what the project actually claims to do. What it does claim, on the other hand, strains credulity in other directions.
Sad to say, but a half-sphere lens on a phone camera does not a microscope make. Without going too deep into details (I hope), proper microscopy requires close attention to the properties of light and lenses. You can't just point a powerful lens at a specimen and expect to get a good image. You also need to ensure sufficient lighting, proper contrast, and control your depth of field!!!
I cannot emphasize that last point enough! Even squashed nearly flat under a cover slip, any specimen you examine is still three dimensional, with details you can only see by shifting the microscopic stage up and down by very small amounts. The depth of field is very narrow. A professional scope's fine adjustment is fine enough to shift focus on a cell from "front" to "back" and anywhere in between. As you might imagine, that requires keeping the stage very, very still. That's why scopes have specimen stages. You won't get that out of a phone camera without clamping it down.
That's what made the article's statement so interesting. If an acceptable specimen could be obtained by smearing the lens directly, the fixed position might allow for proper analysis, but ...