Dean's Speech: Coming Full Circle by Professor Gabriel Leung at 2017 David Todd Oration, Hong Kong Academy of Medicine

15 December 2017

Secretary Chan, President Lau, honoured colleagues of Council, Presidents and Officers of the Academy Colleges, fellows, distinguished guests, ladies and gentlemen,

Before I start, may I ask you to rise and observe a moment of silence in respectful memory of Professor Sir David Todd, who left us four months ago exactly.

The fact that this is the first time Professor Todd’s eponymous oration is delivered posthumously weighed heavily on me as I was preparing my talk. His life inspired the title – coming full circle. However the present context cannot be more different from the treacherous provenance of Shakespeare’s coinage in King Lear: Edmund, the illegitimate bastard son of the Duke of Gloucester, had turned his father against Edgar, the legitimate heir and his half-brother. The villainy returned when Edmund was mortally wounded by Edgar.

Professor Todd most certainly personified the perfect completion of an absolutely enormous medical circuit, casting a shadow so large that virtually every doctor in this town and far beyond, past or present, has always found shelter. Your President, Professor CS Lau in his letter to fellows of the Academy eloquently wrote that his first predecessor “departed from Queen Mary Hospital where his medical education matured and exceptional career achievement endured, taken care by those he had taught and surrounded by his family, beloved students and mentees.” In fact it was exactly 70 years ago that Sir David began his association with the University of Hong Kong and Queen Mary Hospital. Here you see a handsome 19-year-old who transferred from Lingnan University in Guangzhou and entered HKU as a second year medical student on September 9, 1947. He first lodged in Room 38 at the then St John’s Hall and now College, although he had always been identified as a Morrisonian. The University gave birth to Sir David’s legacy whereas he in turn delivered the Academy. Now it is time for the Dean of Medicine at the school where he started to deliver the oration named in the very establishment he built as its inaugural President.

The wheel is come full circle.

Turning now to the Faculty for which I am currently temporary custodian, we are celebrating our 130th anniversary. I feel honoured, as a public health doctor to be given the opportunity to lead our school during this important milestone. Every time I recall that my own first predecessor, Professor Sir Patrick Manson, was also a public health physician, I feel the burden of history. Not only was he first amongst equals in the Aberdonian triumvirate who founded the Hong Kong College of Medicine for Chinese in 1887; on his return home 12 years later, also a Year of the Boar, he established the London School of Hygiene and Tropical Medicine that continues to be a world leader in public health research and education.

This shows my home department of public health having come full circle, recently rehoused in the Patrick Manson Building, a long disused facility that once served as storage after the Faculty office and medical library were removed to their present location a few dozen yards further down Sassoon Road.

Manson pioneered HKU’s enduring strength in infectious disease research and practice. Before he came to Hong Kong, he spent a considerable amount of time experimenting with Filaria sanguinis hominis in Amoy, or present day Xiamen. Through this empirical work, he proposed the idea of mosquito-borne diseases and encouraged the young Ronald Ross who was working in India to test whether his theory would apply to malaria. Ross of course went on to win a Nobel for experimentally proving his mentor’s insight. To commemorate the work of its founder Patrick Manson, and that of Ronald Ross whose eponymous Institute subsequently became incorporated into the London School, its main doorway features the Anopheles and Aedes mosquitoes as you can see in this photograph.

Even though mosquito-borne infections had largely been relegated to the archives locally, notwithstanding the occasional scare of imported dengue and Japanese encephalitis, we still contribute to containing this global spectre of vector-borne conditions through infectious disease epidemiology and modelling research. For example, Joe Wu and I recently worked with two Harvard collaborators to evaluate a dose-sparing immunisation strategy, that is giving only one-fifth of the full antigen dose in order to vaccinate five times more people in the WHO Kinshasa vaccination campaign during the yellow fever outbreak. The green line indicates pre-campaign coverage. The black solid and dashed lines indicate post-campaign coverage if vaccines are administered in standard dose only and according to the WHO dose-sparing strategy with alternative age cutoffs for fractional dose vaccines. These various inset graphs show the infection attack rates, or the proportion of population who would become infected although not necessarily fall ill under the standard vs fractional dosing strategies for different scenarios of transmissibility, as indicated by the basic reproductive number R0. Our work provided direct evidence to inform WHO’s decision to recommend and administer a one-fifth reduced dose to extend supply that ultimately halted the outbreak in Kinshasa. Incidentally much aetiological research remains to be done, in case anybody in the audience is interested. The most vexed question is why yellow fever has not established endemicity in Asia or China, despite the abundant presence of the Aedes mosquito vector and almost certainly multiple introductions through travel and trade for centuries with Africa and South America where the disease is mostly found.

The wheel is come full circle.

Let me tell you a little more about my research in infectious disease epidemiology. I must first confess that I did not receive formal training in the area, and learned every thing I know from doing. The story began with a panic telephone call from the then Director of Health and recently retired Director-General of the World Health Organisation Dr Margaret Chan about two-and-a-half weeks after the first SARS patient fell ill in March 2003. The late Tony Hedley, Lam Tai-hing and I volunteered our services very early on to the Department of Health that had been met with a deafening silence until that fateful call. We were then asked to help DH coordinate the case-contact questionnaires and to try and make sense of the deluge of data pouring in so as to inform public health interventions. Amongst other critical parameters, the case fatality ratio or CFR, that is the proportion of infected patients who would eventually succumb, had been a conundrum and in fact had been wrongly pronounced to be around 5% by clinical and public health “experts” locally and echoed in Geneva and Atlanta. Intuitively and based on back-of-the-envelope calculations, we “knew” the CFR could not have been that low and should have been considerably higher but we did not have the tools to prove it let alone robustly estimate it. While we were busily trying to figure out the complicated algebra of survival analysis, I received daily, and in fact also nightly, phone calls from Dr Klaus Stohr who was one of WHO’s point men during the outbreak, yelling down the line this vexed question ad nauseam, which is not unlike the remixed background sound track you hear. Running out of ideas, one of our Imperial College collaborators, the then young Dr Christl Donnelly (and now Professor Donnelly FRS), who was 36 weeks pregnant at the time ended up making a special trip from London to Oxford to consult Professor Sir David Cox of the Cox proportional hazards model that remains the gold-standard survival analytic method in clinical research. Sir David wrote out the long-form algebra for jointly estimating the admission-to-death and admission-to-discharge distributions that finally cracked our puzzle. So we were able to explain that because of the almost identical distributions and the well known lag between falling ill and either recovering or dying from the disease, the CFR was actually around 15%, three times the original estimate. Needless to say, the implications are enormous clinically and from a public health perspective.

Fast forward ten years to 2013. Our programme in infectious disease epidemiology that was baptised by fire in the 2003 SARS outbreak has now grown to become one of the largest and best funded groups in the School of Public Health. I distinctly remember a sense of déjà vu in March 2013 when another novel virus jumped the species barrier to infect and kill humans on a large scale – this time in the Yangtze River Delta. The causative bug was influenza A(H7N9). Unlike the 2003 SARS outbreak, Hong Kong was not at the epicentre and has in fact never recorded a local case to date. Our team of epidemiologists, statisticians and modellers had matured in the preceding decade to have become one of the most competitive in the world. Like during SARS, outside help was called in except this time around it was us that had become the mentors as opposed to having been mentees ten years prior. I too offered assistance to the China CDC, just as we did to the Hong Kong Department of Health in 2003. Sure enough we were invited to Beijing in April and May, squirreled away in a hotel close to the sixth ring road near the CDC campus. I brought half a dozen researchers to work side by side mainland colleagues in situ, who were backed up by a team of more than ten in Hong Kong working around the clock to define the epidemiologic characteristics of the new outbreak. After two frantic weeks, we worked out the key parameters, including the incubation period, case fatality ratio, basic reproductive number and modelled the potential effectiveness of various public health measures. These provided the key to controlling the epidemic. As a happy corollary, we also reaped the academic rewards of three papers in the Lancet and one in the BMJ.

Most impressively however, I learned another lesson – the same lesson about CFR during SARS but this time in reverse for H7N9. As you can see from the graph on the right, reprinted from one of the Lancet publications, the observed CFR dropped throughout the course of the epidemic as opposed to the simple and naïve estimate increasing during SARS. It was not as highly lethal as first feared, although it would still kill about 4 in 10 of symptomatic patients who were sick enough to have been admitted to hospital. What was different between H7N9 and SARS to have given rise to the divergent trends? We go back to the distributional characteristics of how long it would take those who were destined to recover vs die. This time we did not need to run to Oxford or Imperial to have the methods worked out. We did it ourselves. The mentee has become the mentor, the novice now the expert.

十年人事幾番新。The wheel is come full circle.

Remember I said earlier that our team models public health policy interventions before they are actually implemented, which is akin to carrying out a clinical trial to test for efficacy before approving a new drug. Hong Kong met its first massive infectious disease challenge in living memory during the 1997 H5N1 outbreak when 18 people became infected and 6 of them died. Since then we and others had posited that wholesale and retail poultry markets provide an ideal environment for influenza viruses to reassort, amplify and infect people. Therefore if they were to be closed down, theory would predict that the outbreak would dissipate. The trouble was of course the overwhelming local preference for choice, in that residents would rather retain the option of live chickens as opposed to exclusively consuming frozen, chilled or other pre-slaughtered alternatives. Public health authorities were therefore loath to impose an outright ban. Various half measures were progressively and cumulatively implemented, where each new policy diktat followed yet another outbreak of a highly pathogenic ’flu variant. The urgent challenge at the time was to find a way to prospectively monitor the potential for an outbreak, in order to steer away from the reactive mode. Virologists like Malik Peiris suggested to use a highly prevalent but relatively benign (that is of low pathogenicity) influenza strain, such as H9N2, as an indicator virus to track the ecological viral load in the wet markets. Assuming similar replication behaviour between H9N2 and its highly pathogenic relatives, coupled with advanced statistical methods, we now had a reliable routine sentinel surveillance system to forewarn us of any impending outbreak. This figure shows a 12-year time series of H9N2 isolation rates and the effects of the introduction of 1 rest day per month, 2 rest days per month, and banning the keeping of live poultry overnight.

In fact, key to the control of the H7N9 outbreak that eventually spread nationwide was the closing down of all live retail markets. This national directive was very much grounded in evidence we provided based on initial data from Nanjing, Shanghai, Hangzhou and Huzhou during the first wave of the outbreak in the spring of 2013.

Some of you have no doubt fallen asleep during the last few minutes and those who managed to stay awake must be wondering: surely this is all common sense, is that all public health researchers can offer? Do we really need them? Aren’t they merely formalising the obvious? Sadly for medicine and public health, but fortunately for the survival of the public health species, sometimes things that may appear to be motherhood-and-apple-pie and statements of the obvious can be counterintuitive with massive consequences. For instance, continuing on in the vein of closing down places of intense contact to prevent infectious disease spread, this next slide presents such a counterpoint. On the left was a post hoc evaluation of the government mandate to close schools in 2008, following the high-profile reporting of several paediatric deaths from influenza pneumonia. The effective reproductive number graph demonstrates that the epidemic had already been on the wane and would have faded out even before schools were shut. In other words, it was a case of “too little, too late”. In contrast, during the 2009 H1N1 pandemic, schools were again closed and all children had an earlier summer break. This time, our team documented a 12% reduction in transmissibility during the 4-week school closure, a 25% reduction in transmissibility during summer vacation, and that the H1N1 peak was delayed until September when schools reopened. Thus it worked on this occasion. The take-home lesson is: don’t just assume in medicine because it makes an ass of u and me. EBM stands for “evidence-based medicine”, not “eminence-based medicine”.

Ladies and gentlemen,

You have indulged me for far too long. Let me end by coming back to the titan whose name graces this Oration. He once said:

“Being a doctor is fine but the improvement in health in the past decades has really not so much to do with doctors or nurses but with public health measures – clean water, sanitation, better food, better education, better housing. All these have done much more to improve the general health of the community than doctors and nurses, though of course if you’re really sick, you still need a doctor.”

Such deeply felt and simply expressed humility of one of the most eminent, if not the most eminent medical practitioner of all time in Hong Kong, Sir David Todd will forever live on in our hearts, and with his blessing from the heavens, in our actions too.

May the circles of wisdom, compassion and commitment live on, just like those “Big Shots” captured by the camera lens in this newly published twin album set.

In Sir David’s memory, who became a medical student 70 years ago, I will let his juniors, our current students and trainees, have the last word, in song!