Predicting the health requirements of Coronavirus (COVID-19) in Kenya

Predicting the health requirements of Coronavirus (COVID-19) in Kenya

Last updated: 2 April 2020

Please note: This model is based on a widely used Discrete Time-Based SIR model that computes the theoretical number of people infected with a contagious illness in a closed population over time. This model is adaptive and we may change inputs and outcomes over the coming week. A model is only as strong as its methodology and the validity of it’s inputs, therefore we recommend using this only as indicative of overall patterns and outcomes in Kenya.

Model Documentation

We have received significant interest in this model, for answers to common questions click here.


Model summary

Peak active infections: 2 – 30 million people

Hospital beds needed: 9k – 199k

ICU beds needed: 3k – 77k

Ventilators needed: 2k – 55k

Ambulances needed: 57 – 1,496


COVID-19 in Kenya

The current global coronavirus (COVID-19) pandemic is a major public health challenge. Almost every country in the world is currently in some stage of quarantining in an attempt to limit the spread of the virus. The first case of COVID-19 was confirmed in Kenya on 13th March 2020.

Rescue.co is the largest network of first responders in Kenya and has been preparing the last month to respond to a very likely outbreak in the country. As part of this preparation, we modelled the disease outbreak in Kenya to understand the gap in resources. 

As well as helping us prepare for a spike in the usage of Rescue.co (# of increased ambulance trips), we are making this model publicly available to investors, others working within the healthcare space, hospitals and the government to assess expected challenges and galvanize efforts to respond.

Beyond modelling we are working in a number of ways to prepare: 

1.Training ambulance partners in best practice for infection control and case management of COVID19 patients.

2. Purchasing and equipping ambulance partners with best-in-class personal protective equipment (PPE).

3. Ramping up efforts to quickly add more ambulances to our platform.

4. Mapping ICU beds, hospital bed and ventilators, keeping a close tab on availability and advocating for additional funding.

Coronavirus (COVID-19) is a new challenge for Kenya; information about the speed and conditions of it’s spread are fast-changing. Our model provides our best estimates based on the current data, it should be used only as advisory information.

Our inputs

Choosing a model

There are several epidemiological models to choose from when mapping COVID-19 in Kenya. However, perhaps the most orthodox is the SIR model that computes the theoretical number of people infected with a contagious illness in a closed population over time. We have used this model with Kenya specific inputs and environmental constraints.

Doubling time or R factor of COVID-19

A crucial factor when determining the spread of a virus is the time it takes for the number of infected patients to double. Research World Health Organisation data shows that the disease is highly infectious but it varies significantly depending on the conditions of reproduction (population density, poor sanitation etc). Data from European Center for Disease Control and Prevention shows that during the height of infection it takes between 3 and 7 days for the disease to double. We believe that Kenya has significant challenges to slow the spread of the infection. 60% of Nairobi’s residents live in informal settlements, with population density often as much as 300,000 people per square kilometer. Under these conditions we believe the doubling time of COVID-19 in Kenya will be at the low end of the scale and have used 3 days in our model. So far this is trending correctly.

Other inputs

Hospitalisation: Verity et al. suggests that around 5% of infected people with Coronavirus (COVID-19) will need hospitalisation. This figure is lower than in other studies of hospitalisation which have noted rates as high as 15%. However, given Kenya’s low elderly population and risk profile we have used a low-level rate.

ICU usage: Similarly referenced by Verity et al. and assumed to be about 30% of hospitalisations or 1.5% of infections. 

Ventilator usage: Similarly referenced by Verity et al. and assumed to be about 30% of hospitalisation or 1.5% of infections. Other studies have found lower rates of total infections and we believe that due to the slightly lower aeging population than in Verity et al Kenya will have a 1% period.

Hospital length of stay: In a 2020 study of a single-center case series involving 138 patients with Novel Coronavirus (click here for more info), the average hospital stay was 10 days. THis pars with other research e.g. The value is based on observed full hospital length of stay for ~10,000 Respiratory Failure patients at four Penn Medicine facilities over a 5 year period.

ICU length of stay: 9 days is based on observed full hospital length of stay for ~4,000 Respiratory Failure patients at four Penn Medicine facilities over a 5 year period requiring ICU support. 

Ambulance usage: At the writing of this report there were no studies of how many COVID-19 hospital admissions and transfers required an ambulance. based on the WHO joint mission data we believe 10% of hospitalised cases were admitted by hospital and based on exceeding bed capacity we believe that 5% of hospitalised patients will need transferring per day.


Scenarios

Our model computes the different outcomes of COVID-19 in Kenya depending on the % decrease in social contact in Kenya as a result of quarantine and self-isolation measures put in place in the country.

These scenarios are:

Scenario 1: 10% reduction in social contact:There is little to no social quarantining in place. Most people continue with their daily lives unchanged. There is no financial aid for disruption and most Kenyans are forced to continue working.

Scenario 2: 30% reduction in social contact: There is limited social quarantining. The government imposes a curfew between certain hours, which has a limited effect, but does not enforce tough quarantining and does not provide financial aid for Kenyans, forcing most to continue working.

Scenario 3: 50% reduction in social contact: The government attempts to impose nation-wide curfews and fines for anyone breaking social isolation. The measures imposed by government are partially adhered to and 50% of people continue moving around as normal.

Scenario 4: 60% reduction in social contact: The government starts to put in place more comprehensive financial aid and salary replacements for a majority of Kenyans. A majority of Kenyans adhere to the social quarantining measures and a majority limit their movement outside the home to essential travel only.

Scenario 5: 70% reduction in social contact: The government comprehensively steps in and provides financial aid to most affected industries and demographics. The government builds pupose build mass accomodation for high-density citizens and the police ensure near 100% compliance for quarantining efforts.

Scenario 6: 80% reduction in social contact: The government stops all movement of people for at least 60 days and provides a comprehensive salary replacement scheme accessible by millions of Kenya. The government reduces all non-critical movement and travel with 80% effectiveness.


Outcomes

Peak active infections

Active infections in Kenya are predicted to peak at anywhere between 2m and 30m people, depending on the Government actions taken. Please note, this is the peak active infections, we are not currently able to compute the total cumulative number of people infected.

Duration and peak of COVID-19

Depending on the range of actions taken by the Government, COVID-19 will take between 1 and 10 months to peak and will last between 290 and 800 days.

Hospital beds needed

We have used the number of hospitalisations to infer the number of hospital beds that will be needed in Kenya, this may not in reality be a 1:1 relationship and some hospitalisations may not require a hospital bed.

Based on analysis of Level 4-6 facilities in Kenya we believe there are around 48,000 hospital beds in the country. In the US, 68% of hospital beds are used at one time. Extrapolating this to Kenya we believe there are 7,680 available hospital beds for COVID-19 patients, this number may, in reality, range anywhere from around 4,000 to 24,000.

Depending on the range of actions taken by the Government, Kenya will need between 9,000 and 199,000 hospital beds to cater for the range of outcomes during the infection.

ICU beds needed

Based on our mapping of Kenya’s ICU facilities we believe there are around 290 ICU beds in Kenya. We will assume that 90% are currently in use for normal critical patients. Therefore, we believe Kenya has 29 ICU beds for COVID-19 cases.

Depending on the range of actions taken by the Government, Kenya will need between 3,000 and 77,000 ICU beds to cater for the range of outcomes during the infection.

Ventilators needed

Based on contact with all major public and private healthcare facilities in Kenya we believe there are between 114 and 290 ventilators within the country.  It is assumed that 90% of these assets are used currently at any one time for other non-coronavirus related incidents. 

Depending on the range of actions taken by the Government, Kenya will need between 2,000 and 55,000 ventilators to cater for the range of outcomes during the infection.

Ambulances needed

Depending on the range of actions taken by the Government, Kenya will need between 57 and 1,496 ambulances to cater for the range of outcomes during the infection. Rescue.co currently has 80 ambulances on or offline.


Cost of purchasing assets

Volume needs and the cost of purchasing health assets

Reduction in social contactHospital beds to buyAmbulances to buyICU beds to buyVentilators to buyApproximate cost
10%191,4181,41677,25955,249$6bn
30%149,8111,07259,87943,391$4.7bn
50%94,51464738,42028,141$2.3bn
60%63,51441726,47819,512$2.03bn
70%6,60618638,50814,257$1.45bn
80%64803,0532,253$128m

Conclusion

It is clear from this exercise that Kenya needs to purchase significant numbers of ventilators and construct ICU facilities as an urgent priority. At even the best-case scenario modelled here ventilator usage will 100x current maximum capacity and ICU beds will be exceed by 200x the current capacity maximum. A less urgent, but nonetheless important requirement is the construction of mass, quarantined hospital bed facilities. Lessons should be learnt from other countries in how to construct these facilities at scale in limited time. Ambulances will need to be procured and we recommend a minimum order of 50 to be placed immediately.

Overall this model presents in stark terms the important of social quarantining and tough government led measures that ensure a 80%+ reduction in social contact for at least 3-5 months. With these measures in place Kenya’s heath system, with approximately $150m of investment, will be able to cope with predicted cases.


Get in touch

For those tackling the outbreak

Rescue.co is capturing important proprietary information during the coronavirus (COVID-19) outbreak in Kenya. We are keen to support public health organisations and companies tackling this outbreak. If you are interested in learning more about this model and/or our coronavirus cases and data, please contact info@rescue.co

Cover you and your family now

We are expecting and projecting a major outbreak of coronavirus (COVID-19) in Kenya over the coming weeks. To ensure you and your family are covered for free quality ambulance response (24/7) please visit Rescue.co to sign-up now.

Donate to help us fight COVID-19

Help us make sure that all EMTs in Kenya have proper PPE (Personal Protective Equipment) and training. 100% of your gift will go directly toward training and equipping first responders or toward a rescue or safe transfer for somebody who couldn’t otherwise afford it. Click here.

3 COMMENTS
  • paul moseley
    Reply

    This model appears to take into account the insights generated from highly contagious countries in a specific latitude range. Looking at the John’s Hopkins dashboard at https://gisanddata.maps.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6, geographies close to the equator have a significantly lower infection rate.

    Some may be due to lower reporting, but it is clear that tropical climates are not experiencing the same acceleration of infection. In fact, the majority of Kenya cases until the last couple of days were attributed to travellers coming in from highly affected areas.

    Take the US and Brazil as an example for comparison. Population 327M vs 209M. Reported cases: 245K vs 8K. Infection rate per population .075% vs .004%. Reported cases on 28 Feb: 60 vs 1.

    While it is good for planning to estimate high for the purpose of resource utilization, these numbers will immediately be seized upon by the population as fact…as I know from experience living in Nairobi and spending much of the day helping locals and small businesses understand the published numbers.

    Finally, I encourage you to look at the recent op-ed in the Washington Post by Mohammad Sajadi and Anthony Amoroso, infectious disease researcher with much experience in sub saharan Africa, and to consider the hypothesis that COVID 19 is acting like a seasonal flu that is much more severe in latitudes with a specific temperature and humidity range that encourages spread…’…major affected areas in January through early March were in the Northern Hemisphere between 30 and 50 degrees north latitude. And the areas had nearly identical temperatures of 41 to 52 degrees Fahrenheit and low humidity.’

    https://www.washingtonpost.com/opinions/2020/03/27/covid-19-is-probably-seasonal-thats-no-reason-relax/

    You are clearly experts in modeling these events and your model is impressive….and I am simply looking at the potential impact from a broad review of the global trend without your detailed insights. Thanks for considering these alternative variables as you communicate your model and results. A comparison of similar countries in similar latitudes may provide additional valuable insight into the potential impact of the virus.

    1. Hugo Winn
      Reply

      Hi Paul,

      Thanks for this reply, great points and I think you are picking up on some trends in opinion we have seen since publishing our model. I think you are absolutely right that there is significant discrepancies depending on the latitude of infected countries and it is hard to say whether low testing levels, under-reporting or significant external factors might be at play.

      As you write, the difference in observed viral reproduction at different latitudes may be due to temperature, or more likely, humidity. We believe there is currently not enough data to conclusively support this and studies have shown a wide discrepancy on the effect of environmental factors in the R factors of COVID-19. We have written a fuller response to your question and some others shared with us which might help in giving you a sense of our assumptions and hypotheses since publishing the model: https://rescue.co/a-response-to-some-questions-about-our-covid-19-model/

      We really appreciate your support and interest in our epi-model and please do email us at info@rescue.co if you would like to discuss in more depth.

  • Brian
    Reply

    How accurate are your projections especially on the peak active infections? On April 23rd, the Kenyan government reported that 67% of the current cases were asymptomatic while only 33% were symptomatic.This means that its obvious that most mild and and asymptomatic cases are not tested.Also given that we are conducting Antibody Tests which can give a fairly better number of who has been infected and developed antibodies to protect future infections.

What do you think?

0

Your Cart