Hong Kong Department of Health SARS Bulletin 20 May Of the fatalities, "70.5% of the deceased patients had pre-existing medical illnesses, such as hypertension, heart diseases, diabetes mellitus and stroke. 62.2% of the deceased were aged 65 or above. Taking into account both of these factors, 78.1% had either pre-existing illnesses or were aged 65 or above, or both." See interpretation of previous figures/charts Hong Kong

Many of the deaths may have been due to other medical complications, or arising from the treatment itself. For instance, Taiwan adjusted its recorded death figures from 180 down to 37 due to a new study which indicated that many of the deaths recorded were not the result of SARS. Global mortality figures were duly adjusted by WHO to 774 down from 916. [Associated Press - New York Times 5October03]

There are many causes of over reporting as well as under reporting of deaths. For over reporting, see below Causes of inaccurate reporting of deaths.

Most of the cases investigated appear to have been due to direct contact with the person ill - in many cases, staff and workers in hospitals, or very close communal or family groupings reducing the possibility that airborne transmission is involved to any degree. Hong Kong

Worldwide and country by country cases by date of onset over time appear to be in small cycles and not increasingly larger numbers spreading out. Last major peak worldwide was in late March. See worldwide epidemic curves by date of onset and not date of reporting.

There was spontaneous recovery (in some cases not so spontaneous) for most individuals (90-95%)².

To date, there is no universally accepted medicine which is helpful against the virus, there is no vaccine, further identification tests have not been completed. And yet most of the cases recover, there does not appear to be relapses, spread of the virus is extremely slow, and accurate identification of the virus would probably reduce the fatality rate significantly by increasing the population of those who have the virus but show few signs and had spontaneously recovery

Causes of inaccurate reporting of deaths

What must be taken into account in assessing the true number of deaths

When the number of deaths is very small as is currently the case with SARS worldwide, it is imperative not to give an inaccurate picture. Such inaccuracies can lead to severe measures being taken either to stem the tide of the disease, sway world opinion for economic or political reasons, to convince political leaders that the proper measures are being taken, or due to genuine fear and concern when in fact the true picture would lead to a different set of containment and treatment policies. This can be done by taking the following into consideration in the reporting process. Some of these factors can not be determined accurately but others can and must be taken into account in order to provide an accurate base from which to calculate percentages.

Note: This is particularly important for the first four on the list as these may have a substantial effect on the numbers of deaths when overall numbers are not large.

Causes of over reporting of death from SARS - initial listing

Death due to the side effects of drugs, or other care should be weighed carefully before lumping together with other SARS deaths. SARS Watch links to an article published in The Standard (no longer available on line) which gives reports on the possible effect of the drug cocktail used in the treatment of the disease. "Secretary for Health, Welfare and Food Yeoh Eng-kiong said yesterday medical experts were now examining whether the drug cocktail of ribavirin and steroids was responsible for some of the coronavirus pneumonia deaths in Hong Kong."

The side effects can be most easily seen in the effects of treatment on different age groups. Side effects may be greater for older individuals causing death due to damage to organs or other side effects occasioned by the treatment modalities.

Time lapses or partial reporting can produce large increases or drops when finally reported thus creating the instituting or withdrawal of protective measures not commensurate with the severity of the situation.

When it is difficult to report new cases or deaths, many cases will go unrecorded.

No centralized or regional collection means that incomplete national figures cause wide discrepancies in the statistical view of the situation, or lead to inflated figures when finally reported.

The normal age graded mortality rate should also be a factor in assessing the mortality rate due to SARS particularly in older ages where the disease may not be the determining factor. For instance, the stress of separation from family, stress of any emergency medical treatment whatsoever, the wish to end the suffering, etc.

Incorrect registration as SARS but actually related to an entirely different disease, depending on the outcome, will either increase the death rate incorrectly or change the base of cases from which the death rate is determined (thus lowering the rate).

Death caused by other unrelated reasons and also found to have SARS must be placed in a separate category for statistical purposes.

If records are not kept on a regular basis, data is often lost, minimized, or changed.

Deaths registered in one jurisdiction or country due to transfer shortly before death or after an individual has been infected. This is particularly relevant for travelers from other regions but also for SARS cases of one jurisdiction hospitalized in another jurisdiction.

The strain of the virus causes very high death rates for some strains and very low for others and should be recorded separately.

Two jurisdictions or administrations may unwittingly claim the same area or individual for statistical reasons leading to a doubling of certain data. On the other hand, areas which are disputed or with unclear boundaries may not be counted at all leaving potentially large gaps in the data.

Errors if unchecked can create a ripple effect through the data.

Reliance on tests on the presence of a particular virus or reaction to a virus can lead to biased results - particularly if there are no other symptoms giving confirmation as might happen if it is found that symptomless or new symptom SARS is found in the population. In many cases in the past, single tests alone have been found to be unreliable. If statistics as well as medical treatment are based on tests with low discrimination, death rates as well as possible deaths due to other causes may cause an artificial rise in the data.

With inadequate reporting for any reason, deaths are more often reported than the probable cases or those that have recovered. This tends to distort the rates since the base on which the calculation is made is significantly smaller.

Deaths may be reported for a number of reasons including funding (for facilities and equipment), political (to fulfill expectations), social (to support severe countermeasures), or economic.

May 7 percentage tables from Hong Kong

Hong Kong tables of age groups of recorded SARS cases (about 2 persons per 10,000 people are diagnosed as probable SARS cases) and the distribution of deaths by age groups (SARS Bulletin May 7).

Figures
Total accumulated SARS recorded cases in Hong Kong is 1646 since the beginning with 958 already recovered (58.2%) and 495 still in hospital (includes 89 convalescent). There have been a total of 193 SARS related deaths.

Percentage tables
The number of SARS recorded cases in the total population of Hong Kong is extremely small. If the data is used to calculate risk, the chances that it will cause death is even smaller - to date, ages under 25 have had no fatalities, and a very small possibility (3 out of 100 SARS cases) to the age group 25-34. There is a slight increase (11 out of 100) up to age 64. For the age group 65-74, the rate increases to (23 out of 100 SARS cases) and for age 74+ (38 out of 100 cases).

Interpretation
One consideration for the 65+ age group is the rate of pre-existing medical conditions such as hypertension, heart diseases, diabetes mellitus and stroke which is probably higher. In fact, It is stated that 68% of all fatalities had pre-existing medical conditions. Another consideration to be considered is that in the general population, not all of the symptoms of SARS are present in every case and many individuals may not be hospitalized or counted as having SARS. Hospitalization increases as the individual gets older and particularly when there is a pre-existing medical condition which elevates this group to the largest group recorded as having SARS and the largest group hospitalized when any of SARS symptoms appears.

Within Fatalities Rate
It must be emphasized that these figures are the percentages within the fatality group and have nothing to do with the death rate from SARS itself. For instance, if there are 2 deaths from 100 SARS cases with the age distribution of 1 in the ages 0 to 44 and 1 in the age group 45+ then the death rate from SARS is 2% but a 50% rate within the fatalities for each age group.

See next below for a summary of current update by WHO

Major Points World Health Organization Updates* (cut and paste) As of June 19 Major Points will not be given on this site. Please contact full reports. For the full current WHO reports, access Original WHO SARS updates Previous WHO Summaries 22 May-19 June 03 Previous WHO Summaries 19 April-21 May 03

Advice From Different Viewpoints