Functional Symptoms and Psychological States: An Electronic Diary Study by Burton, C., Weller, D. And Sharpe, M. This study reported the findings from a time series study that compared day-to-day variation in physical symptoms that are related to Somatoform disorders with levels of anxiety, mood, and the subject's level of concern for their symptoms. The authors acknowledge that according to psychosomatic models of the relationship between vague physical symptoms and psychological concerns stress individuals often recognize the physical symptoms they experience and ignore the psychiatric components of their presentation. Theoretical models often assume that a variety of factors including psychiatric factors influence physical illness. The authors also note that there are a few other studies documenting the co-occurrence of psychological distress and functional somatic symptoms over time; however, previous studies have focused on a single psychological symptom such as depression or anxiety and have used pen and paper retrospective recording methods to document their presence (subjects record their symptoms well after they have occurred). These prior studies are open to biases for a number of reasons but most specifically a recording bias due to the retrospective nature of these methods and the tendency of these studies to focus on only one psychological variable limits their generalizability.

The current study used an electronic recording method in real time and tracked the association between levels of anxiety, depression, and symptom-related concern and the perceived severity of physical symptoms in a group of 26 individuals with indeterminate physical symptoms. The physical symptoms were broken down into three groups: 1) External Pain consisting of muscle, joint, and back pain; 2) Internal Pain consisting of headache, abdominal, and pelvic pain; and 3) Autonomic Pain consisting of several different symptoms from bowel pain, nausea, numbness, dizziness, etc. The subjects were recruited through six general practitioners or by self-referral as paid volunteers for the research. Subjects' ages were between 21-65 years of age. The subjects had at least three physical symptoms in at least two bodily systems that were inadequately explained by physical pathologies. Exclusion criteria included the presence of a severe physical illness such as cancer or coronary artery disease, concurrent psychiatric diagnoses (other than a past diagnosis of depression, subjects with first time diagnoses of depression were not included), and an inability to learn to use the electronic palm recording device. Recording was performed twice daily (prompted by an alarm at set times) for 12 weeks on a palm device via a visual analogue scale (VAS) of the three physical symptom groups and the psychological variables of stress, mood, anxiety, and illness concern. Three baseline measures were obtained for level of depression and anxiety with the Hospital Anxiety and Depression Scale (HAD) the Somatic Symptom Inventory (SSI), and the Illness concern Questionnaire (ICQ). Other methodological controls included having the VAS screens presented randomly when recording, the alarm used to indicate data recording was synchronized between subjects, once recorded the data entered was not accessible by the subject, thorough training for subjects by the researchers in the recording methods and use of the device, the repeating of one VAS screen at each recording for validation purposes, and frequent follow-ups for data collection.

The mean age of the participants was 46 years, 22 subjects (about 87%) were female, nine subjects had college or professional degrees, 18 were married or in a steady relationship, and 15 were not working or were not seeking work (this means that perhaps some subjects were not working but were seeking work, but we are not told how many). The baseline screening indicated that subjects reported moderate levels of baseline depression and anxiety that were significantly correlated with the ICQ but not with the SSI. However, on a daily basis subjects reported lower levels of depression and anxiety on the VAS scales than was measured at baseline. Predominate daily physical symptoms were pain, fatigue, and gastrointestinal symptoms. There was a strong diurnal variation for fatigue and stress with lower levels reported in the morning and higher levels as the day progressed. This diurnal variation was also observed to a lesser extent for depression and for the subject's concern for their symptoms.

The correlations between the variables demonstrated strong relationships between stress and anxiety, moderate correlations for mood with fatigue and internal pain with autonomic symptoms. Weaker relationships (below .3) were found between all other variables. The daily reported stress levels of the subjects were not significantly related to any of their physical symptoms (all correlations below .1). As would be expected in a regression analysis anxiety and low mood were found to be...

In addition, the overall regression coefficients between the physical symptoms and the psychological variables were most robust for low mood and symptom concern and were weakest for anxiety level.
Interestingly there were some differences in the baseline measures of anxiety and depression and daily recorded values. Daily recoded levels were lower than baseline levels for both anxiety and depression. At baseline overall anxiety levels were higher than depression levels in the subjects, but the opposite trend was observed in daily data recording (depression levels were consistently higher than anxiety levels).

There are some limitations to this data that the authors acknowledge. First, the study suffers from poor generalziability due to the sampling method. The authors also speculated that there is a potential problem with the assumption that subjects were accurate in their assessments of their physical symptoms and psychological states. The authors suggest that the differences at baseline and daily recording in anxiety levels and depression levels may reflect an under perception or under recording of anxiety in daily life.

There are several other concerns here: First, the study lacks a matched control group so there really is no comparison between people with these somatic concerns (such as IBS) and their psychological presentation and people with everyday somatic concerns and their psychological stressors. Perhaps a control group of depressed individuals without the somatic problems would have highlighted any differences or similarities between pure depression, bodily sensations, and other variables of interests. The authors do acknowledge that this lack of a control group is limiting but then defend themselves by stating that their design required the presence of physical symptoms outside the normal range of experience; however, without a comparison group we are left wondering if the subjects in this study just displayed normal variations in mood associated with variations in bodily states (McKillip, & Voss, 1978). Without a control group we really do not know how well psychological states predict somatic states in IBS.

This brings up a couple of other questions with respect to the demographic variables in the study. Most of the subjects were unemployed, single, and not well-educated. I am just wondering how these variables affected the outcome measures. The hypothesis of the study is that psychological variables affect somatic states, but I am wondering if perhaps the demographic characteristics of the group affected both.

The authors acknowledge that their study is not causal; however, there are some concerns based on the above observation of demographic contributions and the hypothesis the researchers investigate. In the introduction the researchers state the proposition that psychological variables lead to vague or indeterminate somatic symptoms. Is it not just as feasible to conjecture that vague or indeterminate physical symptoms lead to these psychological states? This study looks at associations and not at directionality, but the early presumption in the introduction is that there is a direction between the variables in question. Path analysis is a statistical technique that can answer questions about the directionality of multiple relationships (not necessarily cause and effect) and could determine multiple directionalities (how depression affects IBS symptoms, how demographic variables affect both IBS symptoms and depression, etc.). For example, Rutter (2002) found significant associations between IBS, coping, and depression such that coping and weaker beliefs of personal control lead to more depression in subjects with IBS.

Finally, there is a real problem with using numerically-based scales to measure a construct at baseline and then resorting to shorter visual analogue scales to measure the same construct later. The researchers found that baseline measures of anxiety and depression were higher (HAD, ICQ, SIS are numerical scales) than the daily recordings (visual analogue scales with end indicators only). The variation in the numerical scales would be expected to be greater than the visual analogue scales and the finding that low levels of anxiety and depression occurred daily in subjects might just be an artifact of the measurement system. Wewers and Lowe (1990) provided an early discussion these potential problems when using visual analogues in clinical work. The researchers reported that the low levels of daily psychological distress observed in the subjects in this study indicates that Somatoform patients do not make the connection between their mood state and their physical illness an cite an early study by Blankenstein, Van der Horst, Schilte, de Vries, Zaat, Knottnerus, van Eijk, and Haan (2002) that also found…