Understanding and communicating risk: Measures of outcome and the magnitude of benefits and harms

Neil R. Bell; James A. Dickinson; Roland Grad; Harminder Singh; Danielle Kasperavicius; Brett D. Thombs

Article Figures & Data

Tables

View popup

Table 1.

Outcome measures encountered in preventive screening

MEASURE	HOW TO CALCULATE	ADVANTAGES AND DISADVANTAGES IN PATIENT RISK COMMUNICATION
Mortality (overall and disease specific)	The number of patients who died divided by the total number of patients in the study population. Mortality would be calculated separately for the control and intervention groups in randomized controlled trials	Provides the highest-quality estimate of the benefits of cancer screening Results unaffected by lead-time, length-time, or overdiagnosis bias
5- or 10-year survival rates (absolute rate)	The number of individuals who are alive at 5 or 10 years after the time of diagnosis of disease divided by the total number diagnosed with the disease	Provides exaggerated estimates of the benefits of preventive screening owing to lead-time, length-time, and overdiagnosis bias
Incidence (new cases)	The number of new events or cases that develop during a given time period in the total population at risk	Provides exaggerated estimates of benefits of preventive screening owing to overdiagnosis

View popup

Table 2.

Measures of magnitude or effect size encountered in preventive screening

MEASURE	ABBREVIATION	HOW TO CALCULATE	ADVANTAGES AND DISADVANTAGES IN PATIENT RISK COMMUNICATION	*EXAMPLE^{^} (REDUCTION IN LUNG CANCER MORTALITY)**
Natural frequency	NF	Number of persons with events in a population	Highest levels of patient understanding and satisfaction Denominator of 1000 people increases patient understanding of harms and benefits Understanding increased when baseline risk is included	13 of 1000 people died of lung cancer with screening; 16 of 1000 people died from lung cancer without screening. Thus, there were 3 of 1000 fewer deaths from lung cancer with screening
Absolute risk	AR	The number of events in the screened or control groups divided by the number of people in that group	Increases patient understanding of risk Understanding increased when baseline risk is included	AR in control group = 1.66% AR in screened group = 1.33%
Absolute risk reduction	ARR	Difference in the event rates between the screened and control arms of the study		ARR = 1.66% − 1.33% = 0.33%
Relative risk	RR	Ratio of the outcome measure (eg, overall mortality) in the screened group compared with the unscreened group	Can cause exaggerated perceived screening or treatment effects	RR = 0.80
Relative risk reduction	RRR	The difference in event rates between the screened and control groups divided by the event rate in the control group	Can exaggerate the perceived treatment effect for both physicians and patients. Often presented as percentage without baseline risk	RRR = 1.66 − 1.33/1.66 = 0.20 RRR = 20%
Number needed to screen	NNS	Reciprocal of the ARR	Decreased level of patient understanding compared with other measures of magnitude or effect size	NNS = 308^†

↵* All measures describe the same reduction in lung cancer mortality.
↵† Differs slightly from 1/ARR in this example owing to rounding.
All examples are taken from the National Lung Screening Trial.²⁴
Estimates were taken from the Canadian Task Force on Preventive Health Care systematic review and meta-analysis on screening for lung cancer.^25,26