Breumelhof and Smout used the Symptom Sensitivity and Symptom Specificity Indices to evaluate patient undergoing 24 hour esophageal pH monitoring. This can help identify patients with symptoms related to acid reflux. The authors are from the University Hospital in Utrecht, The Netherlands.
Method:
(1) A patient undergoes 24 hour esophageal pH monitoring.
(2) Each time a symptom was experienced the patient records the time.
(3) Upon completion of the monitoring the symptom log is compared to the record of low esophageal pH readings (< 4). A symptom was considered to be associated with a low pH if the symptom occurs within 2 minutes of a reflux episode (4 minute window).
symptom sensitivity index =
= (number of episodes symptom noted AND pH was < 4) / (total number of episodes with pH < 4) * 100%
In addition, the authors calculated the Symptom Specificity Index, which is the same as the Symptom Index (previous section, above):
symptom specificity index = symptom index =
= (number of times symptom noted AND pH was < 4) / (total number of times symptom noted) * 100%
Interpretation:
• minimum symptom sensitivity index: 0%
• maximum symptom sensitivity index: 100%
• In the study population, the symptom sensitivity index ranged from 0 to 50%, indicating that many reflux episodes were asymptomatic.
|
Symptom Sensitivity Index
|
Symptom Specificity Index
|
Interpretation
|
|
<= 10%
|
< 75%
|
non GERD
|
|
<= 10%
|
>= 75%
|
probably GERD (if exclude chance occurrence)
|
|
> 10%
|
< 75%
|
uncertain
|
|
> 10%
|
>= 75%
|
GERD
|
One subset in patients with a low sensitivity and high specificity had a single symptom episode while multiple reflux episodes. These patients can represent a chance occurrence rather than causal relationship. Once this group is excluded then the combination of low sensitivity and high specificity probably represents GERD; patients in this group may be unaware of most reflux episodes.
I would think subset analysis could be done on the uncertain group in the table above. Wiener et al (previous section) found that the symptom (specificity) index > 50% was suspicious for GERD.
