Quality Improvement

So here is a bit of an explanation about the project I'm trying to work on during this visit.

I won't spend much time here reiterating all of the difficulties one faces when working in Mulago hospital, as I've spent some time on this in previous posts (especially Incentives).  Suffice it to say that the problems of the hospital are almost never attributable to a particular individual, or the presence or lack of their innate intelligence, motivation, commitment or humanity (though undoubtedly there can be large interpersonal variation in these attributes, much as there are at Yale New Haven Hospital, and any large institution). Rather, the problems of the hospital are largely systems problems - aging physical infrastructure, understaffing, lack of resources, and, as I discussed before, the unspoken incentive structures that do little to reward good or to punish poor provision of patient care. And yet, while the current state of the hospital and its systems issues is inextricably linked to the historical (and ongoing) limitation of monetary, human, and technological resources, not every system improvement requires an increase in these resources.

As an example - as it currently stands in the hospital, communication between nurses and physicians is haphazard, at best. Resource limitation played a significant part in shaping the current system - there is a severe nursing shortage, making it all but impossible for nurses to join physicians on rounds. Longstanding cultural and societal norms also play a significant role, which tend to make interactions between (supposedly) higher and lower cadres rather more paternalistic. And yet, good interpersonal communication does not rely upon monetary resources.

Here is another, more tragic, example. I spent some time during this trip on the pediatric Renal/GI ward. The challenges of pediatrics at Mulago are very similar to those of medicine - especially on a ward where half of the patients have end-stage renal disease, without access to dialysis (which is pay-per-session), or transplant (for which they would need to travel to South Africa or India, and pay out of pocket). Those are problems that are very directly related to monetary resources, and therefore aren't ones that I have any power to fix.

But during my time there I was confronted with another type of problem.  A young child, perhaps 10 or 12 months old presented to the hospital after 2 weeks of watery diarrhea.  Diarrhea is a common illness amongst children all over the developing world, and is still a significant source of infant mortality (when it leads to severe dehydration). I met this child as I rounded with a Ugandan Senior House Officer (a pediatric resident). Oddly, the child did not look as severely dehydrated as some diarrheal patients - the mouth was not dry and pasty, and the child still had moist eyes and tears.  What was odd was that the child, though alert, was quite limp, and appeared to be in respiratory distress despite having relatively clear lungs on auscultation. As I reviewed the chart with the SHO, we found that a set of blood electrolytes had been measured in the Emergency Department (the day prior), and that the patient had a Potassium of about 1.3 (mEq/L). The normal range is from 3.5-5, which means that this patient had very severe Hypokalemia.

Most likely the patient had had ongoing potassium loss in the watery diarrhea, and had had only potassium-poor oral intake (like water or diluted porridge), thereby slowly depleting his body of potassium. Hypokalemia of this degree is life threatening, and because potassium shifts play a key role in nerve and muscle depolarization and firing, this degree of derangement confers a high risk for cardiac arrhythmias, and was also the likely cause of the patient's flaccid paralysis and respiratory distress (from respiratory muscle weakness).

The treatment, clearly, was potassium supplementation.  This degree of hypokalemia necessitates IV repletion, both for speed of treatment, and because GI absorption can be erratic in the setting of diarrhea. Undoubtedly, in the US, a child as sick as this one would be taken care of in the ICU, and their potassium would likely have been checked hourly as it was gradually repleted. Here, on the other hand, there are few high-dependency beds available, and if you send a test for serum potassium, you will not receive it back until the next day - but only if you get it there before the lab closes at 5pm. After that you may have to wait for the following day.

Our only choice was to replete potassium on the ward.  The problem is that overly rapid infusion of potassium is also life threatening (as it may also cause cardiac arrest). I sat with the resident, and we calculated exactly how many milliliters of the available concentrated potassium chloride solution to dilute in how many milliliters of normal saline.  We decided to infuse in aliquots of 100ml of the mixture, which we would give 3 times by the next morning.  Both the SHO and I spoke directly with the nurses on the ward, and we saw the initial solution mixed and hung above the patient.

I had other responsibilities in the hospital that afternoon, and was off of the ward. When I returned the next morning, I learned that the child had died at around 1am.  The SHO recounted that she had popped back by the ward in the late afternoon and found that only perhaps 10 milliliters of the first 100ml aliquot had actually been infused into the patient. Remember, that there are no infusion pumps - no machines to ensure steady flow of IV fluids at a pre-determined rate.  Rather, you just adjust a little plastic gadget that squeezes the IV tubing, thus restricting flow and then you eyeball how fast the fluid is dripping. The SHO spoke once again with the nursing staff about the planned infusions, and ultimately had to go home for the evening.

It's difficult to know what happened, but my guess is that the patient was simply never infused with sufficient potassium, and ultimately succumbed to respiratory failure. Alternatively, if someone had sped the infusion too much, the patient could have had cardiac arrest.  More likely, however, as had happened during the day, a nurse had hung and/or adjusted the IV drip rate, and overestimated the rate at which the patient would receive the fluids, thereafter moving on to her many other tasks, and never finding a moment to check back.

In the morning, I watched the SHO speak with the nurses.  She was appropriately upset. This was, she told me, the second child she had lost to hypokalemia recently.  Still, there were no raised voices, and it did not seem that she was inappropriate with the nursing staff (who, of course, were not the ones who had been on the ward overnight). Certainly, though, she voiced her frustration.

Some minutes later, I heard the nurses talking amongst themselves. It sounded as if they felt maligned. In Europe, one nurse noted to another, the nurses have no more than 5 or so patients each, and if they have critically ill patients, perhaps only 2. On that particular pediatric ward, which had about 30 patients, there were 2 or 3 nurses during the day and probably 1 at night. What could they have done, when there were no doubt other critically ill patients on the ward who also demanded their attention?

Well, I'm not sure exactly what they could have done. Certainly, they are in no position to hire more nursing staff, or to buy more infusion pumps.  But they could have done something differently. It's true that resource limitation has allowed these systems to grow into the dysfunctional tangle in which they now exist, but it is not true that there was no way to save that child. It is not true that the medicines and the expertise to infuse them were not present. It cannot be true that the overnight nurse did not have even the 5 seconds it would have taken to check on the infusion's progress, or the 2 minutes to hang the next one. What is true, is that there is no formal system in place to ensure that the importance of those tasks was passed on at shift change, and no system to make sure that those tasks were performed in a timely manner.

Rather than lament, or become resigned to the problems of the system, though, I would rather that the nurses realize that it is within their power to change the system in which they work in small but meaningful ways, that do not require additional resources, and may have saved that child's life.

That is what Quality Improvement (QI) work is about. Unlike traditional medical research, which mostly concerns itself with the discovery of new knowledge, QI projects aim to improve the performance of systems towards pre-defined goals. Although QI got its start in commercial industries (such as manufacturing), its techniques have been co-opted for use other areas. In medicine, QI projects strive to ensure that health systems provide care that is safe, effective, efficient, patient-centered, timely, and equitable (in line with the Institute of Medicine Guidelines).

Those goals sounds pretty grandiose from way out here in Kampala, but the message is simple and applicable - we need to improve systems to ensure that they provide proven care, and provide it well.

What might that mean on the Mulago National Referral Hospital Pediatric Renal/GI ward? Well, paper and pens are easily available. Each day after rounds, could the physician team write a bulleted list of only the most critical nursing tasks, to ensure that they receive special attention? Could a similar list be used to pass on critical tasks during nursing shift-change? What can we do to encourage staff to consistently complete these tasks? How can we reward good performance and remediate poor performance?

If you have an eye for systems issues, it doesn't take long to identify opportunities for improvement in a difficult work environment; there are many low-hanging fruit. The question is, how do you improve systems in a sustainable and resource-reasonable manner? And how do you incentivize people to spend their time and energy doing it? That is what my project here is about.

As I said - there are many low-hanging fruit, and by a few weeks into my first visit here I had plenty of ideas to work on.  But (long story short) I learned what many people working in development have learned before me - sometimes my ideas won't work because they are my ideas.  I don't live here. I don't work here. I don't spend most of my waking hours in Mulago Hospital. I don't know the system well enough to recognize what may be clear to others. I can't get done by email what needs to be done in person, and I can't get someone to do my work for me. (If only I could...)

Ultimately, I realized that instead of putting together a project myself, I needed to figure out a way to incentivize staff here to do QI work.  And so my current project was born.  I have been lucky enough to receive some monetary support from the Office of Global Health at Yale, and we will be using that money to provide a grant to a Ugandan Senior House Officer to perform a QI project on the medicine wards. After a competitive grant application process, the winning SHO will receive a grant which is intended both to reimburse their time spent on the project, and also to pay for project costs (such as data collection and materials).

As you might imagine, changing systems can be difficult, and as such, Quality Improvement work is a bit more complicated and rigorous than just trying to convince someone to pick up a pen and paper twice a day before sign out.  This is largely because, even if you do convince someone to do a task for a while, as soon as you stop bugging them, they'll probably stop doing it.  Even a simple task, such as getting a staff member to consistently use of a new form, requires a variety of changes to the context in which that staff member works. It's ultimately the changes to the work environment - the system - that will ensure that form use continues after you're no longer watching.

I'm far from a Quality Improvement expert, and I am learning as I go.  But I'm optimistic about our project here. I have been greeted with interest by the SHOs, who recognize, more than anyone, the difficulties of the system in which they work. So in addition to time on the wards, I've been spending time recruiting applicants, doing some teaching on basic QI concepts, and working to ensure approval and buy-in from the Mulago administration. So far, so good.

Application Deadline: December 21st.

Ebola, no Marburg, no Ebola

If you're into viral hemorraghic fevers, Uganda is kind of the place to be. As I write this, Uganda is at the tail end of it's third outbreak in the last 6 months.

There are a number of viral diseases that can cause hemorrhage, but the ones that keep popping up in Uganda are Ebola, and Marburg virus diseases.  Ebola and Marburg are related viruses, both of the same family (Filoviridae). They cause short-lived, intense illnesses, characterized by acute onset of high fever, chills, malaise, general prostration and high mortality. The symptoms largely reflect the storm of pro-inflammatory mediators released by the infected cells of the innate immune system (mostly macrophages) which spiral out of control, and can lead to the hemorrhagic manifestations which make the diseases both so scary, and so apt for use in hollywood films - basically, bleeding from every orifice.

The case fatality rates range from 21% -- if you contracted Marburg virus in Germany in a 1967 outbreak -- to 90% -- if you contracted Marburg virus in Angola in 2005. The case-fatality ranges for most of the relevant Ebola strains are similar - from 50-90% (with all outbreaks occurring within Africa). There is debate as to whether the wide mortality range documented for Marburg is attributable to strain characteristics, or the quality of available medical care at the outbreak sites - I suspect that it was at least in large part due to the latter.

While Ebola Virus takes its name from a river in what is now the Democratic Republic of the Congo (then Zaire), interestingly, Marburg Virus actually takes its name from Marburg, Germany. Marburg was the city where hemorrhagic fever made its Western debut when in 1967 an outbreak infected 25 individuals harvesting tissue to produce the poliovirus vaccine.  They were harvesting tissue from -- you guessed it -- monkeys sent from Uganda. And while Germany has been lucky enough to avoid further outbreaks (probably because the Ugandan monkey tissue industry isn't what it used to be), central and East Africa have not been as lucky.

Most often, outbreaks occur when an index case comes in contact with an infected monkey. Monkeys are not thought to be the natural reservoir of the viruses, as they also become ill and often die from the disease. When encountered, though, weakened or dead monkeys often end up as food. In an outbreak in Gabon in 1996, a dead chimpanzee was butchered and eaten by 19 people, all of whom quickly became ill. There is also some evidence that fruit bats may be a reservoir for Marburg virus, as several European tourists developed the disease after separately visiting a particular (bat-ridden) cave in eastern Uganda (one of whom died).  Most often, however, the index case does not live to tell the story of their exposure.

Generally, the disease spreads by direct contact with the bodily fluid of actively ill patients, or by iatrogenia. In one terrible outbreak in 1976, the index case presented to a small missionary health outpost in (then) Zaire. Because he presented with high fevers, he was treated with injectable quinine for severe malaria.  The needle was then washed in a pan of water with the other needles with which they routinely injected all febrile patients with anti-malarials. Almost 100 patients were infected and died. It is also possible that the 2006 Marburg outbreak in Angola was caused by inappropriately reused blood bank equipment.

Generally, however, infection only occurs with direct contact with bodily fluids, either during illness, or during the common (sometimes religious) custom of washing the body of the deceased before burial. There is no evidence of infection being acquired from asymptomatic patients during the incubation period. There has also never been clear evidence of airborne transmission in previous outbreaks (despite what Dustin Hoffman says). There was, however, one episode in 1996 in which a patient presented to a hospital in Zaire with abdominal pain, and underwent exploratory laparotomy. There wasn't much to find in the abdomen, but they did manage to aerosolize some blood, thereby infecting the entire OR staff and subsequently their families and caretakers when they became ill. There was also a recent article purportedly documenting airborne transmission in laboratory conditions between pigs (another potential reservoir) and monkeys.

Perhaps the most unexpected tidbit I found about transmissibility was that, in fact, Ebola and Marburg are also potentially sexually transmitted diseases.  There has been well documented viral persistence in (amongst other places) the semen of surviving men for up to 3 months after infection, and following the 1967 outbreak in Marburg, there was one documented sexual transmission from a survivor.  Oops!  So I guess the lesson is this: pity sex with a cancer patient - OK; pity sex with a hemorrhagic fever survivor - best wait on that.

So what's happening here in Uganda?

The first outbreak this year was an Ebola outbreak, occurring in Kabaale (a southwestern area of Uganda). It was announced on July 29th, and the last confirmed case was discharged from the hospital on August 24th, though the outbreak was not officially declared over until October 4th (observing a pre-determined observation period that is about twice the maximum incubation period). There were 24 confirmed and suspected cases (only a portion of which had confirmatory viral testing performed), including 16 deaths. The majority of deaths, as in most outbreaks, were amongst the family of the index case.

Shortly thereafter, Uganda also suffered an outbreak of Marburg Virus this year.  The outbreak began in October, and was declared over in late November (days before the most recent outbreak), and ultimately included 20 probable or confirmed cases, and 9 deaths which were spread across 4 districts (Kabaale, Ibanda, Mbarara, and Kampala). The index case was a school teacher, and it is unclear what his exposure might have been.  Both he and later his infected caretakers travelled by bus around central and southwestern Uganda, including through the busy capital here, mostly to assist in the care of sick family members (who initially, at least, were sick with something other than Marburg).

The most recent, and not yet officially concluded outbreak of Ebola began in Luweero, a small city about 40 miles north Kampala. The index case was a Boda Boda driver, with an unknown exposure, who died on October 25th, having been taken care of in the village by his family and never having visited a health center. The next two deaths occurred on November 10 and 12th, were both family members and caretakers of the index case, and died at the local health center, thereby increasing the number of exposed individuals.  As of now, there have been 7 confirmed and suspected cases, with 5 deaths.

If nothing else, the string of recent outbreaks ensured that the infrastructure necessary to respond to the most recent outbreak was in place.  There were pre-existing isolation facilities at several hospitals (including Mulago Hospital, here in Kampala), and international expertise (from the likes of the CDC, MSF, etc.) to help coordinate a response.  The most recent outbreak was quickly contained, and really limited to family members of the index case and their immediate caretakers. Outbreak control is done primarily by 1) isolation of known or suspected patients, and 2) contact tracing for known contacts of patients.  Essentially, if you had contact with a case but are asymptomatic, someone checks up on you every day. If you develop any symptoms, you're off to isolation.

But I expect that the question on everyone's mind is - what is it like being here during an outbreak? Well, I'm sorry to disappoint, but the answer is - it's the same as before.  Although these are scary diseases, the rapidity of the initial response ensured that Ebola patients were not showing up on the wards at Mulago hospital. I think there were 2 suspicious cases seen in the Emergency Department, who were quickly rushed to isolation, though they both ultimately they received alternative diagnoses.  As such, during this outbreak, the only Ebola patients at Mulago hospital were known contacts who were taken directly to the separate isolation area (located in a cluster of tents set up in a fenced off area behind the hospital).  So I cannot feign bravery or altruism in deciding to stay here - it just didn't come up.  If there had been Ebola cases popping up on the wards, that would have been a different story.

As one of the senior physicians here put it, Ebola is not a "smart disease."  If you want to be successful, be indolent - like HIV - infect millions before we know what happened.  But if you just go and kill your host in a few fever-ridden days!? You won't make it very far.

Luckily, this outbreak hasn't made it very far.  Here's hoping it stays that way.

180 degree panorama of Murchison Falls - of The African Queen fame, Northwest Uganda

Perspective on Assault from Kla

Hello and Greetings from Kampala (it is often abbreviated kla)!

I've been back in Uganda working at Mulago hospital for almost 4 weeks now.  As usual it has taken me a while to find a chance to write a post, but I'm hoping to put up a few in the coming days to catch everyone up.

At some point I'll put together a post on the Ebola outbreak here (spoiler alert: I don't have Ebola), and another on the Quality Improvement project I'm working on putting together.  First, though, I wanted to write about some events that happened just before I came here to Kampala.

About a week and a half before I came back to Kampala, I was the victim of an assault in New Haven. It was midnight, and I was walking from a friend's apartment to my car in one of New Haven's nicer residential neighborhoods. I was walking with a friend of mine (also a resident). Although our cars were only about 2 blocks away from our friend's apartment, as we walked we crossed paths with a large group of adolescents.  There were a lot of them - probably 20 or more - all looked to be teenagers (from younger to older), and the group consisted of boys and girls, some walking, some on bikes, and all talking, chattering and interacting amongst themselves as the walked down the street.  Although it was an odd scene, they were not overtly threatening.  We had actually walked through the group without incident (they were heading one way down the street, and we in the other), but moments after we had passed them a smaller group of about 5 boys, mostly wearing ski masks, returned and assaulted us.

I didn't really see my attackers because they ran up from behind me and began hitting me in the head. I didn't realize it at the time, though in retrospect, and given the degree of the injuries they caused, we ultimately think they were hitting me with rocks. While three of the boys attacked me, another two snatched my friend's purse, shoving her to the ground and hitting her once (without causing significant injuries). They made no effort to rob me (my wallet and phone were in my pockets).

I was dazed as everything occurred, but as I was receiving repeated blows from behind, I decided to try to turn around to face my attacker, in the hope that a face to face confrontation might dissuade them.  Frankly, at the time, I didn't realize the gravity of what was happening, and I half thought that I would turn around and see one punk kid who had punched me a few times and gone running. Instead I saw several adolescents in ski masks standing behind me. I think the fact that I did turn around and face them was unexpected, and everyone stood still for a moment.

I saw that my friend was okay, and was not being pursued by anyone, so I kind turned around and sort of jogged and stumbled off into the adjacent city park (a few blocks of lawn and trees in the middle of this generally pleasant neighborhood). I could see that my friend was moving off in the same direction, and seemed okay. I could also see that one of the assailants followed me into the park, and I thought I saw him stoop to pick up a rock (or perhaps he already had one in his hand). I was not moving very fast (as you can imagine), so he was quite close behind me. I thought he was going to throw the rock at me (not realizing that they had already been hitting me with them), so I tried to bob and weave a bit, though lost my balance and fell to my knees. I expected another blow, as I tried to stand up, though I never received one. My friend later told me that the final attacker left me in part due to the shouted warning of one of the other assailants - something along the lines of 'what are you thinking? let's go!'

Although we were left alone at that point, moments later the same group attacked another resident just down the block who was heading to the same apartment we had just come from.

I only realized the degrees of my injuries as we waited for the ambulance. I had large lacerations on my scalp, my right forehead, and my right eyebrow.  My nose was broken and somewhat crooked. My most significant injury was to my left ear. A decently sized piece (about 1-1.5cm wide and about 2-3cm long) of my ear had been avulsed, and remained attached only by a sliver of skin near the very top of my ear. I have some great pictures from the ER, though even for a blog where I photo-documented cutting a parasite out of my own toe, they are a little much.  If you want to see them, email me.

As I write this, 5 or 6 weeks later, my injuries are all but healed.  I will have bit of a jagged Harry Potter-like scar on my right forehead, though the one in my eyebrow is largely hidden.  I went to the operating room about 5 days after the event to have an ENT doc straighten my nose, without complications.  And my ear, the survival of which was initially in question, ultimately regained its blood supply and has healed well (though the avulsed portion remains numb and there is a large scar).

I don't have too much to say about the higher meaning of the events at this point, in part because I'm tired of talking about them.  But I think I summed it up well, a day after the event, when I said: Random Violence is Bad.  I don't mean to say that all of this doesn't tell us something about the underlying problems of our society, but for those of us who are confronted with the results of poverty and hopelessness on a regular basis in the hospital, I'm not sure it tells us anything we didn't already know. Sure - I won't give large groups of adolescents in New Haven the benefit of the doubt on a dark street anymore, but I don't think that is a widely applicable lesson.

And what does this have to do with my trip to Kampala?  Well, in a way, coming back to Kampala rescued me from the aftermath of this event, in both a practical, and a philosophical sense.

After the assault, I spent about a week milling about at home, supposedly resting my brain after it was concussed -- which I am quite bad at -- and feeling generally restless. Even after the immediate medical issues were taken care of, it was difficult to overcome the inertia of the event - it had stopped my life. Every chance meeting with an acquaintance required another retelling of the episode, and my email inbox was backlogged with correspondance related to the assault.  Kampala, in that respect, was an escape. And while a quick trip to Uganda probably doesn't seem normal to many of you, I have done my best to make international work a regular part of my life, and so making it here was part of making it back to my normal.

And certainly, I did have to tell the story to a few Ugandans when they inquired about the wound on my forehead, though here the incident was largely forgotten.  It goes without saying, of course, that Ugandans are uniformly shocked that something like this could happen in the US. Interestingly, I found that Ugandans often had more difficulty wrapping their heads around it than Americans. While saying something like 'teenagers looking to wreak havoc' was generally enough of an explanation in the US, it is not sufficient here. Were they on drugs? Were they psychologically disturbed? A robbery would be one thing, or even a rebel army kidnapping children and raping and killing civilians - those are somehow more familiar here. That is not the case for truly random violence; in that category the US seems to stand out.

But Kampala rescued me from the implications of my assault in another way as well -- by putting them in perspective. I began this period at Mulago hospital working in the Intensive Care Unit. The difficulties of this work would fill a blog entry on its own, but suffice it to say that the ICU is full of unlucky people - victims of motor vehicle accidents (mostly from the motorcycle taxis), victims of assault (though rarely random), and victims of disease - people whose injuries are far worse than mine, who receive far less support and care, and many of whom don't survive. Walking past the emergency department, at any given time, you are liable to see a few patients far more disfigured than I was on my worst day, and who are unlikely to be helped to heal as much as I was.

It is easy to say to yourself "I'm lucky it wasn't worse" and then to go on feeling sad at just how bad it was. Living and working in a place where you are regularly confronted by "worse,"however, illuminates that type of sorrow in a harsh light.

Thanks to EgyptAir for having a pre-arranged tour to Giza for those of us with a few hours to kill in Cairo en route.