The rampant spread of the COVID-19 virus has stretched the Indian healthcare system beyond its capacity. Patients and healthcare providers scramble for beds in hospitals and at makeshift facilities each time there is a new surge in cases. The ability to predict and evaluate hospital admissions and vacant bed capacity has therefore become critical for providing the necessary healthcare during the pandemic.
In this analysis, we have used data released by the Brihanmumbai Municipal Corporation (BMC) to map the trends in total COVID-19 bed capacity and bed occupancy (private facilities and public facilities included) in Mumbai from June 2020 to May 2021. This helped us identify periods of bed shortages in the city. We have studied these trends across different ward types including temporary wards at dedicated COVID-19 health facilities and critical care wards in hospitals.
We plotted these trends along with test positivity rates to study the response time and found some lag in the current bed management strategy. We have made initial observations on the initial delay in ramping up dedicated COVID-19 bed capacity during the second wave of the pandemic and insufficient availability of beds in critical care units during the peaks of both waves.
This is with the intent of initiating conversations on the subject and developing new research questions.
Bed Capacity versus demand
As would be expected, the data shows a very close correlation between test positivity rate and total bed occupancy.
Here, total bed occupancy includes Dedicated COVID-19 Health Centres (DCHC), Dedicated COVID-19 Hospitals (DCH) and COVID-19 Care Centres (CCCs). The dedicated facilities are earmarked areas in hospitals for the treatment of ‘moderate’ to ‘severe’ cases. The CCCs are makeshift facilities for ‘mild’, ‘very mild’ and ‘suspect cases’, set up in hospitals, hotels, schools, lodges and other such public or private spaces.
The rise in test positivity rate is closely followed by a rise in bed occupancy suggesting that test positivity rate is an important predictor of hospital admissions and therefore demand for beds.
Not all people who test positive for COVID-19 require institutional care but due to the shortage of beds, some doctors and patients may preemptively put in a request for a bed, which can also clog the system. This means that a high test positivity rate is likely to lead to more demand for hospital beds even if patients do not require immediate hospitalisation.
The data shows that the bed capacity was ramped up following a rise in test positivity rate, but with a notable lag.
The test positivity rate started rising in early February this year. The test positivity rate was about 4% on average, in mid-February. By the third week of March the average positivity rate had already increased to about 12%. Between this period, there was barely any increase in bed capacity in dedicated COVID-19 health wards. Bed capacity started being notably ramped up only towards early April, when the test positivity rate was at a peak of about 19.8%. Media reports show that the availability of beds reduced sharply during this period leading to long waiting times for patients in need of beds.
This analysis suggests that it is worth monitoring the test positivity rate more closely as another indicator that can inform bed capacity ramp-up in future (provided there is sufficient testing). A recent study conducted in Italy found that using TPR as a predictor can allow decision-makers to forecast the number of beds in hospitals and intensive care units needed 12 days ahead.
II. Critical care infrastructure
In April last year, the union government had approved a Rs 15,000 crore (i.e. 150 billion) ‘COVID-19 Emergency Response and Health System Preparedness Package’ to help states ramp up the number of oxygen-supported beds, intensive care unit (ICU) beds and ventilators. The BMC data shows that Mumbai had about 2,995 ICU beds at peak capacity this year, about 2.5 times that of June last year. This shows a notable increase in critical care capacity in the last one year. However, despite these investments, the bed capacity in critical care units proved inadequate when the second wave hit. The number of COVID-19 cases was far more than anticipated. Over 300,000 new cases were recorded in Mumbai between early February and the end of April this year. From just 34% in mid-February, the ICU bed occupancy rate went up to about 95% by mid-April. A similar increase is observed in ventilator bed occupancy.
III. Comparison between first and second wave
As observed above, there have been some improvements in overall infrastructure and COVID-management in Mumbai this year compared to last year. Even the total bed capacity at dedicated COVID-19 facilities and COVID-19 care centres was substantially increased this year. This is possibly why the COVID-19 Case Fatality Rate, i.e. deaths as a percentage of all positive cases (calculated with a 2-week lag), was less this year than it was during the first wave last year. But the huge spike in the number of cases this year and initial lags in ramping up capacity still led to periods of bed shortages and long waits for critical care treatment.
IV Policy Implications & Way Forward
COVID-19 bed capacity, especially at temporary facilities such as hotels and schools, cannot be maintained at all times. Even at hospitals, space needs to be freed up for non-COVID patients. The bed management at these facilities, therefore, has to respond to the changing needs of the pandemic.
But the challenge is how fast the system can respond to the surge in infections and ramp up capacity.
The above analysis shows that BMC is capable of significantly ramping up its capacities. But we need to monitor indicators that can be better predictors of bed requirement and can give local governments enough time to ramp up capacity when needed.
Further research should explore the ability of the test positivity rate to predict hospital bed capacity. Studies mapping the lag between the increase in test positivity rate and the increase in bed capacity, across different states and across different countries would also be useful.
The infrastructure and capacity at oxygen and ICU wards also need greater investment. According to the Union Ministry of Health and Family Welfare's reply to Lok Sabha on 12 March, 2021, one ventilator bed in Maharashtra caters to over 25,000 people.
Hospital budgets need to be boosted as an urgent priority so they can undertake these critical, high-cost investments. Failing to do so could prove devastating if we were to witness another major surge in COVID-19 infections.
Notes on the data
BMC collects most of its data on a daily basis but it occasionally revises the previously collected data which can create discrepancies in the dataset. For instance, cumulative positive cases should increase incrementally on a daily basis. However, on some days a lower cumulative value has been observed compared to the previous day. Since the number of fresh cases recorded on any given day cannot be negative, such a discrepancy is likely to be a result of adjustments to the data. Hence, we have aggregated the data at a weekly level to minimise such data gaps.
Numbers on reported deaths also show sharp variations due to periodic updates. Both waves see a sudden spike in numbers due to a possible reconciliation of deaths data.
The reconciled numbers are added in bulk at a later date and not on the actual day of death. This can limit real-time understanding of the trend in deaths due to COVID-19 and make it challenging to monitor and respond to the pandemic. Reconciled death data can prevent us from accurately calculating the Case Fatality Rate, for instance, as it does not reveal when people actually died.
However, it's worth noting that data reconciliation is common to many databases tracking disease on a daily basis. Among others, Johns Hopkins University, the World Health Organization (WHO) and the European Centre for Disease Prevention and Control (ECDC), compile regular updates of COVID-19 cases and deaths globally and by individual country.
About the Database
The Brihanmumbai Municipal Corporation (BMC) started releasing a detailed daily PDF containing information on a variety of indicators such as mortality, containment zones, bed capacity, vaccinations, etc. from June 2020. Such data is extremely useful not only for formulating short-term policies but also for assisting longer-term academic research. A long-standing issue with using the data, however, has been that the authorities replace the PDF each day with fresh information and remove the prior day's document. This practice has limited any kind of temporal analysis till now.
At IDFC Institute, we have been able to gather all such PDFs from June 2020 until the present. We are in the process of converting all relevant indicators to a machine-readable output and will put them out as a public-facing database in due course. In the meantime, we will put out a brief analysis of them along with the related datasets.