Managing the COVID 19 crisis: A supply chain approach
While it is more deadly than the common flu, Covid-19 is less lethal than many diseases known to man. Yet, the global march of the novel Coronavirus is raising serious concerns. The human immune system had not experienced this organism before November 2019, when this virus is believed to have made the leap from animals to humans for the very first time. In all likelihood, we have no natural immunity against it.
The bigger problem is, that the virus is much more contagious than common flu. This is because unlike the common flu, which takes only two days to show symptoms, nCOVID-19 infected people take anywhere up to 14 days to experience symptoms. This long period of 14 days, means that infected people are likely to spread the illness asymptomatically for a long time before they themselves become ill.
Given the above, the disease has the potential to quickly overwhelm a country’s healthcare system and cause a severe demand-capacity asymmetry. The medical infrastructure of countries can quickly be overrun with an influx of patients, leading to a situation of rationing of lifesaving services and infrastructure. Such a situation of rationing can create a vicious loop of more people waiting for care, which in turn can infect even more, leading to a snowball effect.
Countries scrambling to handle this unprecedented public health challenge have been forced to take what are being described as draconian restrictions on travel and daily life to combat the virus and flatten the curve to levels that are manageable (like other diseases) by their healthcare systems. To most, a complete lockdown seems to be the only solution. Unfortunately, it comes at a huge economic cost.
The Lockdown Dilemma
In the early stages of any epidemic, the number of infected is always a small percentage of the total population. In India, at the time of writing this article, the infected is around 400, in a population of 1.3 billion. So, a lockdown means that in addition to the infected (who are few right now), the healthy too, are prevented from engaging in their daily activities such as traveling with dire consequences for their livelihood, and for the economic parameters of the whole country.
Unlike many of the relatively wealthy or developed nations, this is a serious issue for a country like India where more than 86% of employment is in the unorganized sector and more than 62% workers are employed as daily wage earners who may not have the savings to survive a few days, let alone the recommended two weeks. Even in the organized sector, work-from-home is not an option in many situations. For instance, the manufacturing sector of India which accounts for 75% of the country’s industrial output, is very labour intensive. So, a severe lockdown does not seem to be a viable solution.
On the other hand, if the country doesn’t lockdown, and instead, depends on the citizens to self-regulate and practice social distancing or self-quarantine, it may not yield the desired results. As with any systemic level changes, left to individuals, it is very difficult to get proper subordination to the global objective. Individuals will act based on their individual perception of risk and their other conflicting needs like caring for their family, going to work etc., and will end up violating quarantine.
"System-wide behavior changes cannot be attained by leaving it to individual choices!"
In a populous country like India, the number of infected can grow exponentially. If there is only one infected person at the beginning of a month, considering that the number of infected people doubles every three days, there will be 1024 infected by month end, about a thousand times as many as it began with. At this rate, the infection can quickly go out of control and much beyond the country’s medical infrastructure, leading to numerous deaths. So, a lockdown seems inevitable.
Either solution – allowing citizens to self-regulate OR enforcing a lockdown – can lead to havoc. How do we deal with this conflict? Using the TOC thinking process, we can arrive at a breakthrough solution to meet the competing needs of economy and health. This is possible, if and only if, we try to find the wrong assumption that is forcing us to stay in the dilemma.
The assumption that can be challenged is
- There is no practical way to identify the infected person, as soon as he gets the infection, and isolate the person.
Direction Of Solution
If we build a system that can catch every infected person in the net of testing, we can isolate the person from general population and give him or her the necessary medical care till completely cured. This will allow bulk of the population, the healthy ones, to transact business as usual without any risk of infection. This is the only approach that can achieve the twin objective of controlling the pandemic while allowing the economy to run as usual.
Lockdown The Infected Rather Than A Geographical Area
To implement this solution, the entire situation has to be conceptualized or designed as an operational work flow with dependent resource centres. The infected part of the population, an unknown variable, will form the demand side of this work flow. The first resource centres are the testing facilities that can identify the infected. If testing results take more than a day to arrive, there is need for a quarantine work center for that duration (from start of test to getting confirmed results).
Next, we need quarantines with basic medical care facilities where everyone with infection is housed till either completely cured or till it is necessary to move them to the next work centre- the intensive care for the critically ill. If this flow line is managed effectively using flow principles of TOC, one can resolve the need for lockdown, i.e., all infected individuals can be cared for under quarantine while the healthy are free to move around.
The flow line will fail, if there is any emergent bottleneck, with significant queuing in front of it. For example, a long wait for testing means more unidentified cases in the populace infecting others. Similarly, if the infected are not quarantined in some identified location, because of long queuing, the chances of them infecting others is high ( assuming self-quarantine is not effective).
In the design of a flow-line, it is important to understand where the real constraint will be and subordinate all decisions to ensure effective use of this constraint without allowing it to become a significant bottleneck. This, however, is complicated due the unique nature of this flow line. Like any conventional work flow each resource centre serves as dependent resource in the end-to-end flow line (from testing to critical care) but unlike a conventional operational flow line which has one input and one output at end of line, here every work centre has its own intermediate output (healthy/cured people moving out of the system). So, while it is important to identify the constraint for overall flow line, it is also important to ensure that none of the other work centres can become a bottleneck. Therefore, in this entire flowline, the resource with minimal spare capacity with maximum difficulty to ramp up should be the designated constraint for the overall end-to-end flowline. All other resources should have adequate spare capacity.
Even a cursory examination of this flowline makes it evident that capacity for intensive care is very difficult to expand in the short term due to the high capital expenditure and specialized resources needed to man it. This is also the entity where we can least afford a long queue in front of it. A large queuing (constraint turning to bottleneck) can force an ethical dilemma -the need to cherry-pick which patients to treat!
Hence, to maintain the Intensive Care Units as desired constraint with the additional condition of minimal queue, we should look at subordinating all decisions to ensure that. This implies that all the upstream work centres have to be ramped up adequately with the objective of ensuring no infected goes undetected and is immediately quarantined. If one is able to do that, the demand side will taper off and the exponential growth will not get triggered. This in turn will drastically reduce the potential load on ICU. The disease will be like any other normal disease in an area
Using the focussing steps of TOC.
Goal : Maximize the healthy and recovered people in a population (and ensure rate of recovered patients is higher than rate of new infections)
- Identify the Constraint Resource - Critical Care
- Decide how to exploit the constraint - Ensure critical care capacity is available immediately for all patients needing it.
- Subordinate to above decision - Ensure capacity of detection, primary quarantine is elevated when required. This will ensure number of new infection stays at lowest level (like any other disease) and consequent load on critical care is minimal
- Elevate: Add protective capacity to critical care, if solution deployment is delayed beyond stage 1
Capacity Management On Non-Designated Bottlenecks
The key to subordination lies in effective capacity management decisions of all work centres, upstream of critical care.
Aim of testing is to identify every infected person as early as possible. Depending on the stage of the disease in a geography, the testing coverage will have to be adjusted. Initially, it should cover all people arriving from infected regions. And at later stages, anyone who is likely to have come in contact with the infection should be traced and tested. But whichever stage the disease may be in, it is important that testing is not rationed. A rationing of testing by introducing strict policy of who can be tested will result in testing becoming an undesired bottleneck.
Often, testing, especially when there is significant load on testing facilities, can give false positive or false negative results. Both situations are not good; but not offering infected people care, and instead letting them unknowingly infect others is worse. So, it is not enough to test once, confirmatory tests will have to be performed. The idea is to identify all the infected quickly and isolate them from the general populace.
• Test Duration Quarantine
With current level of technology, the results of the test takes only 4-8 hours. But if due to backlog and load on testing facilities, the time taken for a test increases, it is important to provide temporary quarantine facilities to those awaiting results. This will ensure that no potentially infected people are free to infect others. Patients who test positive for the infection can be cared for in the quarantine facilities and discharged when cured.
• Quarantine with primary healthcare for the infected
Quarantine facilities with medical care is required for the infected so that they are not left to their own devices. Rationing of quarantine facilities by letting only few selected to be quarantined with medical care while others are requested to self-quarantine may not be effective. It is best to avoid any possibility of quarantine violation and escalation of infections in the population.
"Rationing of upstream resource centres to manage the load on it is incorrect subordination."
• Intensive care
Out of primary health care, few will move to critical care. If upstream resources are well managed early during the epidemic, the demand will flatten out and the numbers will remain low. This will in turn prevent the intensive care from turning into severe bottleneck.
Interplay Of Capacity And Demand
It is important to understand the unique interplay of demand and capacity in this environment to be able to manage the situation effectively. As the graph above shows, if capacity is not put in place at the right time, then later additions don’t help, Catching up with demand becomes difficult.
In this environment, level of demand affects capacity requirement, and in turn available capacity impacts the level of demand. Higher demand needs higher capacity, however, if capacity is lower than demand, queue outside the quarantine will lead to more infected people in society which in turn further increases demand.
"Demand impacts capacity and capacity impacts demand!"
Hence typical supply chain strategies of rationing to manage scarcity of capacity will not work. If capacity is available ahead of demand, then demand will taper off (rate of recovery will be higher than rate of fresh infected due to quarantine of infected). Hence timing of capacity addition is crucial. If the timing is correct, one can manage with lower capacity. However, if the timing is delayed, more capacity will also not help prevent the exponential curve effect. As shown in the above curve, if capacity is added at C3 and time T3, then one can control the epidemic without the need for further addition in capacity. While if you miss the bus here, then adding more and more capacity will also not help, till the time, you have added in excess of the demand curve.
Capacity Management Cell
No major capacity enhancement will be required, if a country has managed the first wave (stage 1) of epidemic and is able to prevent the infection from getting a foothold in the country. However, at later stages, capacity enhancement becomes crucial for successfully handling the epidemic. The biggest challenge then, to ensure timely ramp-up of capacity in testing and quarantine, is the speed with which decisions are taken. When different agencies are involved, decision making slows down. So, separate cross-expert administrative teams, with complete authority to decide or modify existing policies, have to handle the preparatory activities for a quick ramp up. For instance, the cell looking at ramping up testing facilities should look at best testing practices around the world, identify private sector partners, arrange for personal protection equipment, sanitation equipment, etc. The capacity cell of quarantine care can identify a ready bank of possible locations (like exhibition centres, hotels, schools etc.) that can be converted into quarantine units. They should also be empowered to take policy decisions. For example, over 70,000 doctors who have been trained in medicine abroad in countries like Eastern Europe, Central Asia but do not have license to practise in India , can be temporarily allowed to help in primary quarantine care facility. These dedicated capacity teams for each of the entities in the flow line can thus keep a bank of deployable capacity for each upstream resource centre and enable rapid deployment of these facilities when the need arises. This will prevent testing and quarantine from becoming bottlenecks in the flow line.
If the solution is implemented in stage 1, capacity addition in intensive care may not be required but if not and the infection is in stage 2 or 3 when the solution is implemented, a dedicated team will have to be put together to ensure timely elevation of capacity for intensive care. However, better the identification and isolation, less is the overall demand flowing in, lower is the capacity enhancement needed for critical care.
It is very difficult to forecast capacity in this environment. The only way to manage is by keeping a hawk’s eye on the load and queues developing on a daily basis at these facilities. Facilities approaching full load should be quickly expanded.
Expanding capacity of the resource centres across the country will require 50 to 500 items – from consumables to expensive medical equipment – to be made available together. Even after these are set up initially, depending on the inflow, much of these will have to be replenished again, without delay. Different locations will have different levels of inflow of patients and therefore different levels of consumption of items. Stock out of these items can be detrimental to capacity utilization of the resource centres, e.g., lack of testing kits, can put a halt to testing; without enough ventilators critical care units cannot function at full capacity.
Each of these items could be supplied by at least half a dozen companies. And these suppliers have no visibility to what the other suppliers are making. Each supplier manufactures as per their understanding of how much is required (forecasts). This tends to be very inaccurate. So, with this way of working, some items may get produced in excess while others continue to be in shortage. The solution is a collaborative body of all suppliers with the objective of an efficient use of the capacity or rationing the capacity of all the suppliers in such a way as to ensure that all items are available when required. This can be spearheaded by or under the auspices of a central governmental body.
A pull-based distribution collaborative model (TOC) can ensure proper distribution without shortage and excess in any geography. This model calls for a central warehouse which will source and supply all identified items needed by the resource centres. Having this central warehouse is critical as it allows for near 100 per cent availability for any fluctuating requirement from the resource centres. This is possible as consumption from all units will not peak at the same time. Additionally, by the principle of aggregation, the inventory required at the central warehouse can be nearly half of the total required at all the consumption points put together.
The suppliers will have to produce to ensure availability in the central warehouse. To replenish the stocks of the central warehouse, the inventory levels for each item will have to be decided. At the end of the day, the quantity of each item consumed by all the resource centres from the central warehouse can be triggered to all the suppliers. Simple colour signals based on relative chance of stockout of each item can be used for indicating priority and need for expediting. Red is top priority (as it has highest risk of stock out), followed by Yellow and Green. Every company can have a defined share of business which will be dynamically adjusted based on their ability to supply and on emergent requirements. Thus, now, the availability of items will continuously match actual demand pattern even when demand is varying. Further if the resource centres provide visibility to consumption from them, then the total replenishment times to these centres can be reduced to just the batching and transport time. A few personnel at the suppliers, central warehouse and resource centres will have to be trained to work on this system of daily consumption data transfers to the previous node and suppliers taught to react to colour priorities. Required IT infrastructure and facilities will also need to be put up to enable these daily data transfers.
The Timing Challenge
Even after doing everything possible to quicken ramp up abilities, it will be naïve to assume that one can always get the timing right for capacity enhancement. Hence, it is important to understand what stage of infection the country is in and manage using the right demand suppression (when one is late) levers and rapid capacity enhancement tactics to get over the crisis. However, it is important to know that, any attempt at demand suppression will impact economy negatively and hence these measures should be used sparingly – only when badly needed to mitigate damage created by delay in capacity enhancement.
Stage 1 – Beginning Of Pandemic (Infection Not Yet Within Country)
At this stage, it is far easy to manage because capacity required (for test and quarantine) is very low as compared to subsequent stages. However, all incoming people from the affected countries should be tested. This could cause testing (and quarantine work centre) load to go up. Hence, demand side throttling will be required to get some breathing space to add capacity. For this, the frequency of flights (and other ways of transportation) from the affected countries should be reduced. Once, testing and quarantine capacity is ramped up, frequency of these flights can be increased. The frequency of flights, etc, can be continually adjusted up and down based on load on capacity at any point of time. Or, only passengers with valid ‘negative’ certificate can be allowed in. This could help manage the load on testing and quarantine. If this is not reliable then solution of throttling frequency will be required.
But the key direction at this stage is that only tested and virus free people are allowed in. The affected are put through the flowline till cured. This will prevent the epidemic from taking a foothold in the country.
The economic activity of imports and exports with affected countries will be affected but significant portion of unorganized sector of the country will be unaffected and can continue with their regular activities. Thus, the competing needs of overall economic health and health of citizens will be met. There may be a stage where the country may have to stay isolated from many other countries as infection spreads in other countries. But large part of local economy will continue to thrive.
Stage 2 – Rising Infections But Traceability Remains
There are chances that one has missed implementation at stage 1 and stage 2 has kicked in. At stage 2, though many infected people may have potentially mingled with the population (safe to assume many do not strictly adhere to self-quarantine guidelines), contact tracing is still feasible. But the capacity to do contact tracing and put all the affected people into flow line of test and quarantine, may fall short of demand. This may lead to more infections and a rapid movement to stage 3.
So, a demand side throttling may be imperative. Since at this time, most of the places affected are areas with incoming overseas traffic, it is important to ring-fence these areas totally (states/districts/cities). No one should be allowed to enter or go out of these designated areas; there should be lockdown of economic activities and a curfew on movement of people. This is not an easy step to implement considering how porous state or district boundaries are in a populous country like India. We assume that, with the help of the armed forces and citizen volunteers, these ring fences can be put in place (while allowing flow of essential services). Once the ring fence is in place, it will prevent the network of infected from growing (demand side management). This will provide the authorities time to test and move the affected to designated quarantine centres. Simultaneously, build up in testing, quarantine will have to be taken up along with a further ramp up of supply chain capacities of personal protection equipment, disinfectants, medicines etc. At this stage, it will also be best to activate a cell dedicated to building some protective capacity even for critical care facilities (as these take some time to be up and running) for the patients entering the flowline; and in readiness - just in case the region moves to stage 3. This is important in order to ensure that patients won’t have to wait for intensive care at any point in time. At the same time, all the other geographies should test and isolate people who had come into those areas in the recent past (before the area was ring-fenced) from affected areas.
Once all the affected are all quarantined, the lockdown, ringfence and curfew can be removed. Since, the demand side throttling along with capacity enhancement strategies help in rapid quarantine, economic activity can be restored in a short time in the affected area. Meanwhile, other unaffected areas continue normally.
Stage 3 – Widespread Community Level Spread Of Infection
If solution is not implemented in stage 1 and stage 2, and a geography has entered stage 3 wherein it is difficult to trace the source of infection, a more aggressive demand and capacity side management is required. At this stage, it becomes important to identify which state is in which stage of infection and take actions accordingly. In the states that have reached stage 3, impose curfew and lockdown and to ring fence them. Since the solution calls for identifying and isolating the affected, and who could be infected is not known, it is important to test all in those affected states. Based on the level of infected in a district, and starting with the district with the highest incidence, every-one in the state should be tested. Testing can be done district by district if capacity available is limited; but if there is more capacity, multiple districts can also be taken up at the same time. Simultaneously, there should be ramping up of capacity for testing and quarantine. At this stage, the intensive care capacity has to be further elevated to take care of the high influx of patients to the flowline. The cell activated for this resource centre in the flowline will have enabled this along with coordinating for smooth supply of the consumables, equipment and infrastructure needed.
Once all the affected are moved into quarantine, each area can be considered cleared to transact regularly within itself. These areas will be ringfenced from other areas but can allow people having negative infection certificates to enter and ensure that the area is supplied with essential goods.
This way the crisis can be overcome even at stage 3 with lesser damage to economy. In the meanwhile, the states in the country which may be at stage 2 or stage 1 can implement solutions as applicable to them.
Handling System Leakages/Relapse
One cannot assume that the solution will be perfectly implemented in the real world. So even with the above processes in place, there will be leakages from testing (false negatives) or from quarantine. Since there is no lockdown in place, the multiplier effect of infections remains the same. So, these patients who go unquarantined can kickstart the exponential progression of the virus - the pandemic might again start from stage 1 then move to 2 followed by 3. Persisting with the solution and continuously attempting to close the net so that most of the infected are isolated, can ensure that the base is continuously reset. For example, if this solution is implemented in India at the current levels of infection (~400 infected people) and we assume that only 80% of the infected are identified and isolated weekly into quarantine (leaving 20% to continue infecting others), the numbers of infected in the populace who can potentially infect others will still go into single digits in as few as 13 weeks.
Further, even after very rigorous implementation of the processes, it is quite possible that the disease will keep rearing its ugly head and trigger off a spate of infections intermittently in some areas. It is important keep a close watch, perpetually. Infection detection can happen only after symptoms show and the patient goes to a hospital or doctor. Hence the spread to contacts would have happened in interim. But as soon as an incident is known, it has to be handled, by contact tracing, testing and isolating. By this time, adequate capacity and capability should be available to manage the second spurt; so reaction will very likely be faster. However, if contact tracing and testing capacity falls short and testing is delayed, then stage 2 solution of ringfence, lockdown of affected areas, till the completion of testing and isolation of the infected is completed, has to be implemented. All other areas have to follow the rule of testing anyone coming from affected area, just before the ringfence.
The key insight which can help manage a pandemic is that, using the lever of interdependency between demand and capacity, the infected can be identified and isolated from population. Demand side management actions has a risk of affecting the economy negatively, but if these are judiciously used with along capacity ramp up tactics to prevent an emergent bottleneck in the flow line, a nation can ride over the crisis while managing the humanitarian costs of such a crisis.
Note by authors: This article has been written in March 2020, while the corona virus has already affected more than 192 countries. There are a lot of discussions in the media; much data has been gathered by now. Offering these suggestions, in hindsight, for handling the pandemic is far simpler now than it was at the onset of the virus’s rampage. Nevertheless, the authors feel that TOC way of thinking helps offer robust solutions, and can be used to solve systemic problems, whether these problems are faced by organizations or countries . This article is an attempt to showcase the value TOC principles can add.