Considering Research Denominators

The pandemic of 2020 created a daily barrage of data released to the consumer for interpretation. Nations and states were compared on new cases, hospitalizations, recoveries, and deaths. National and international data were presented on interactive maps with circle sizes illustrating the spread of the COVID-19 virus. A concept key to understanding how to interpret data is sampling.

Every researcher has a sampling plan, the process of determining who or what will provide data for the study. The target population consists of all the individuals who meet the sampling criteria. For example, a researcher may use a target population of all perianesthesia nurses for a study on resilience during a pandemic. But a study of the target population may not be feasible or affordable. The accessible population is the portion of the target population that the researcher can access, such as membership in the American Society of PeriAnesthesia Nurses (ASPAN). Members who go on to provide data for the study are considered the sample. When reporting the number of subjects participating in a study, the researcher must be careful to provide a denominator.

The denominator in mathematics represents the whole and the number of pieces that the whole has been equally divided by. For example, your mother bakes a cake and cuts it into ten pieces. Two pieces are eaten by your sister and brother; eight of ten pieces remain. Mathematically, ten represents the whole, or the population of cake pieces, the denominator; eight represents the sample of pieces remaining. If the researcher only states that eight pieces of cake are available, the reader does not know how large or small the cake was before it was sampled by your sister and brother.

It is important for the perianesthesia reader to know the size of the target population or, at the least, information about the target population. Such information is frequently provided in the methodology section of a report under a heading of ‘Sampling.’ A target population of patients in the perianesthesia unit could be estimated by the number of patients cared for each year, or the number of surgeries per month, or even the number of admissions per week. In the last issue of the Journal, Andersson et al¹ collected anxiety data on 37 postoperative orthopedic patients. The reader can infer that the target population was 120 patients over a 4-week period, and thus, the sample represented 37 of 120 patients (31% of the population). While the authors provided a target population denominator, the small percentage of the population included in the sample limits the applicability of the findings. Repeated studies are needed.

Epidemiologists use the entire population for their studies. Two important calculations, prevalence and incidence, are made using large databases. Prevalence is the number of cases in a particular population at a given time. The denominator is the population. For example, on April 14, there were 1,020 positive cases of COVID-19 in New Hampshire and 752 cases in Vermont. It is tempting to say that the New Hampshire population was more infected. If the denominator is the number of tests performed, the infection rate is 8.7% and 7.1%, respectively. However, when the denominator is the total population, the rate is 0.075 and 0.12, respectively. Thus, from this data snapshot, on April 14, Vermont had a greater disease prevalence than New Hampshire.

Incidence is the number of new cases, with a denominator of the total population. On April 13, there were 56 new cases reported in New Hampshire and 21 new cases in Vermont. The incidence in the respective states was 0.0041 versus 0.0033; thus, there was a lower risk of contracting the disease in Vermont on April 13.

Sample data can be best interpreted if knowledge of the denominator is made available. The perianesthesia reader must carefully examine the research report to determine the target and accessible populations of a study. Without these data, study findings are difficult to translate into clinical practice.