5
$\begingroup$

Thanks for all the answer for the question Calculation of incidence rate for epidemiological study in hospital. And here come's the second part of the question:

What about the prevalence rate then? I have read The new public health as suggested by Chi, the book says that prevalence is usually not available when using ordinary incidence rate, but I saw another formula here:

total case count in that period of time/total patient bed days during that period of time

It puzzled me again, what is it? I have never heard of prevalence calculated using denominator as patient-bed days.

Thanks!

$\endgroup$
1
  • $\begingroup$ There is no such thing as a "prevalence rate" despite its widespread usage. "Prevalence" is a more epidemiologically sound term. $\endgroup$
    – AdamO
    Jan 3 '18 at 19:21
4
$\begingroup$

It's a somewhat unusual way to calculate prevalence, but it makes some sense to use the patient-bed days as the denominator. Consider two scenarios:

A hospital has a single patient, who stay for 1,000 days, and in that time, has a single infection.

B hospital has 1000 patients, who stay for 1 day each, and in that time, they have 50 infections.

Using "N" as the denominator:

A Prevalence = 1.00 B Prevalence = 0.05

Using Patient-Days as the denominator:

A Prevalence = 0.001 B Prevalence = 0.05

The latter accurately reflects the higher burden of disease. Generally speaking, when you calculate a prevalence just using N, it's under the assumption that all persons at risk are at risk for the same amount of time. In the example above, that isn't true.

$\endgroup$
1
  • $\begingroup$ One must underscore the "population" in this instance is people who are hospitalized. If one aims to make inference about the general population or even the catchment for the hospital, it is another example of Berkson's bias. $\endgroup$
    – AdamO
    Jan 3 '18 at 19:23
0
$\begingroup$

Person time is mostly used when calculating incidence. That is, when you are taking from a specific point in time to another. It is also called the incidence density. When using the total population as the denominator you will be calculating the cumulative incidence.

$\endgroup$
1
  • 1
    $\begingroup$ Welcome to the site, @Femi. I edited your answer to make the English smoother. I made some educated guesses regarding what you were trying to say. Please ensure it still says what you want. $\endgroup$ Dec 26 '12 at 23:25
0
$\begingroup$

Similar to my comment on the earlier thread, there is a great deal of heterogeneity regarding the terminology that epidemiologists use to measure disease. Rates, ratios, proportions, prevlance, incidence, risk - all of these words, and others, get combined and re-combined into a messy nomenclature that is more confusing than helpful.

Perhaps it would be more helpful to simply think of the fundamental difference between a prevalence and an incidence. The latter refers only to how fast new cases appear in a population, thus the denominator should be some measure of time.

The former should simply reflect the number of cases in a population, relative to the size of that population. It is a measure of total disease burden, and should reflect both the incidence of disease (how fast new cases appear) AND the duration of disease.

Which is more important will depend on your question.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.