Pain, 49 (1992) 199-204 0 1992 Elsevier Science
199 Publishers
B.V. All rights reserved
0304-3959/92/$05.00
PAIN 02018
Pain complaint and the weather: weather sensitivity and symptom complaints in chronic pain patients Michael 0 Department
S. Shutty,
Jr. a,b, Gary Cundiff
b and Douglas
E. DeGood
b
of Psychology, Clinical Studies Unit, Western State Hospital, Staunton, VA 24401 (USA) and ’ Pain Management Center, Department of Anesthesiology, University of Virginia Medical School, Charlottescille, VA 22901 (USA) (Received
19 June 1991, accepted
10 October
1991)
Chronic pain patients frequently report that weather conditions affect their pain; however, no Summary standardized measures of weather sensitivity have been developed. We describe the development and use of the Weather and Pain Questionnaire (WPQ) which assesses patient sensitivity to meteorologic variables defined by the National Weather Service (e.g., temperature, precipitation). Seventy chronic pain patients (59% females) with an average age of 43 years completed the WPQ. The instrument was revised using factor analysis to produce a Weather Sensitivity Index (WSI) (48% of variance) with high internal consistency (0.93) and test-retest reliability (r = 0.89). Reporting patterns suggested that patients could reliably identify which meteorologic variables influenced their pain but could not reliably determine which physical symptoms were consistently affected. The most frequently reported meteorologic variables which affect pain complaint were temperature (87%) and humidity (77%). The most frequently reported physical complaints associated with the weather were joint and muscle aches (82% and 79%, respectively). Patients labeled as being ‘weather sensitive’, defined by greater than median scores on the WPQ, reported significantly greater pain intensity, greater chronicity of pain problems, and more difficulties sleeping than patients with low scores on the WPQ. No differences in gender, education level, disability status, or global psychological distress were found. Results are discussed with respect to physiological and psychological mediating variables. Key words: Weather sensitivity; Pain complaint; Meteorology
Introduction Chronic pain patients frequently report that weather conditions affect their pain. A recent survey of 211 chronic pain patients revealed that 60% believed that meteorologic variables influenced their pain, with cold and damp conditions being rated as most problematic (Jamison et al. 1989). However, despite a vast literature addressing the topic of biometeorology (Tromp 19631, empirical studies of this phenomenon in chronic pain patients are scant.
Correspondence to: Michael S. Shutty, Jr., Ph.D., Western State Hospital, Clinical Studies Unit, P.O. Box 2500, Staunton, VA 24401, USA.
Observations that meteorologic variables such as temperature and humidity affect health status date to antiquity (Rosen 1979). For example, the Greek physician Hippocrates (about 400 BC) emphasized the seasonal dependence of medical disease (Landsberg 19861. Many of these observations have been reflected in folklore such as ‘aches and pains, coming rains’ (Toth 1979). Even slang language underscores the relationship between meteorologic variables and health in phrases such as ‘feeling under the weather’ used to describe when one is ill or depressed (Rosen 1979). Clinical case observations have detailed striking associations between meteorologic variables and pain complaint. An early example entitled ‘The relations of pain to weather, being a study of the natural history of a case of traumatic neuralgia’ published in the American Journal of Medical Sciences in 1877 describes a case
of phantom limb pain. The patient kept an exhaustive diary of his pain which was interpreted by the physician as resulting from coincident weather conditions. The author concluded that approaching storms, dropping barometric pressure and rain were associated with increased pain complaint (Landsberg 1986). Clinical studies have identified subgroups of patients, for example those diagnosed with arthritis (Hill lY72; Patberg et al. 1985; Sibley 1985; Laborde et al. 1986; Rasker et al. 1986; Patberg, 1987. 1989) and fibromyalgia (Yunus et al. 1981), who frequently report pain exacerbations associated with meteorologic variables. Hill (1972) has estimated that 80-900/o of pcrsons with arthritis report weather sensitivity, concluding that cold climate and falling barometric pressure combined with increased humidity were associated with reported pain exacerbation. Hollander (196 1) reported similar results for arthritis patients when barometric pressure and humidity were manipulated in a controlled laboratory environment. In a related vein, Yunus et al. (1981) reported that 92% of their series of 50 consecutive primary fibromyalgia patients reported ‘cold or humid weather’ as modulating their pain symptoms. Consequently, modulation of symptoms by meteorologic variables has been suggested as a diagnostic criterion for primary fibromyalgia. Unfortunately, there have been no attempts in the clinical literature to develop reliable and standardized instruments to measure self-report of weather sensitivity despite some effort in the lay literature (Rosen 1979). Hence, the results of survey studies are difficult to interpret, with respect to the consistency with which specific meteorologic variables are reported and for which specific groups of patients are involved. In addition, there are many diagnostic groups of patients who have been relatively neglected in the research literature despite frequently reported antidotal associations with meteorologic variables (e.g., migraine hcadachcs (Tromp 1963; Rosen 1979; Persinger l%O), sinus headaches (Rosen 19791, and scar pain (Rosen 197’); Landsberg 1986)). In our own clinical experience, weather sensitivity has been noted for a variety of musculoskeletal pain disorders including low back, neck and shoulder pain. It is not uncommon for a patient to report non-compliance with prescribed treatments (i.e.. exercise) due to increased pain attributed to changing weather conditions. This study describes the development of a brief internally consistent and reliable self-report weather sensitivity index for use with chronic pain patients suffering from musculoskeletal disorders. In addition, the relationship between self-reported weather sensitivity and frequently reported physical symptoms are examined. Finally, differences between ‘high weathersensitive’ and ‘low weather-sensitive’ patients are explored.
Methods
Subjects Seventy consecutive outpatienls lor chronic subjects
pain problems
for this study.
(41’4
males) receiving treatment
at a large university
Data
were
collected
hospital
during
served as
the
patient‘s
initial visit to our clinic. The average age was 43.2 years (SD. range: IV-XI)
with a mean education
The most frequent
presenting
complaint
was low hack pain (60%).
followed
by upper back (3% 1. neck. arm and shoulder
Patients
with
rheumatoid
disorders
(including
excluded from study and represented of our clinic
population.
pain (37%).
fihromyalgia)
were
only a minority (less than
Pain duration
months and more than
= 13.7;
level of 12.2 years (S.D. = 5.Y).
ranged
widely
5%)
between
IO years in several cases and averaged
years (S.D. = 9.6). Accidents were the chief precipitating reported,
and 41%
of patients
treatment
for pain-related
reported
having
3 6.7
event (58%)
received
surgical
problems.
lnstrumentx Lkwoquphic DMQ
urui Mcd~ul
is a self-report
Hisrory Questionnuiw
(DMQI.
booklet which provides descriptive
about pain problems,
past medical treatments
and current
ing. Patients rated how much their pain interfered daily activities (work,
household
The
information function-
with a number of
chores. socializing,
sexual relations.
cxerciae and sleep) using a S-point scale in which I is ‘not at all’ and 5 is ‘extremely’.
A
summary
score
provided
an index of self-rated
disability
status has demonstrated
number of studies (Tail
averaged
disability.
across all
This method
good reliability
VisudAnulog SC& IVES).
of the pain,
nearest millimeter.
in a
The VAS consists of a IO-cm line with
is asked to indicate
to the intensity
and validity
et al. 1987. 1990).
the ends marked with ‘none’ and ‘severe’ (Ohnhaus The patient
activities
of assessing
and Adler
1975).
the point on the line corresponding
and the distance
This measure
is evaluated
to the
has been shown to be a valid and
reliable tool for measuring both clinical and experimental
pain (Price
et al. 19x.3. ,(;\.~q~t~rtt C‘hecXlisr-Yt/ Kcrisd
CSC‘L-OUR). The
rogatis 1977) has heen advocated
SCL-YOR
(De-
as a useful measure of psychologi-
cal distress in chronic pain patients (Shutty et al. 1086). Test-retest reliability
coefficients
internal
1 week
for
range from 0.78 to O.YO, whereas
consistency of the subscales range from 0.77 to 0.90 (De-
rogatis
lY77).
The
General
global psychological additional
Symptom
Index was used to provide
distress index. The
sleep items comprising
were used to provide
Depression
a sleep factor (Shutty et al. 1986)
an index of mood and sleep disorder,
tively. The Somatization
a
auhscale and 3 respec-
subscale was not used due to overlap
with
the WPQ.
Wcuther umi Pum L)utwwwu~~re CWPQI. The WPQ questionnaire a Symptom
which provides a Weather Complaint
meteorologic record affected ‘flow
variables
daily weather
(ranging
Checklist.
from
‘not
Part
used by the fluctuations:
at all’
by these variables.
1 of the WPQ National
patients
to ‘extreme’) For example,
literature
consists of symptoms
to he frequently
reported
and
consists of 7
Weather
Service
rate on a S-point how much
their
an item from part
much is your pain affected by temperature
2 of the WPQ
is a I-part
Sensitivity Index (WSI)
to
scale
pain
is
1 asked.
(hot or cold)?‘. Part
revealed
after
as affected
review
of the
by the weather.
This
lrsting of symptoms was reduced to 15 based on a pilot study in which SO consecutive
patients
with musculoskeletal
what symptoms, if any. were affected 2 of the WPQ, how
much
example.
the
patients
disorders
by weather
were asked
conditions.
rate on a S-point scale (as indicated
symptoms
are
made
worse
an item from part 2 asked, ‘How
made w’orse by the weather’!‘.
hy the
On part ahovcj
weather.
For-
much is stiffness in joints
The item content for parts I and 2 of
201 TABLE
I
Results
PERCENTAGE OF PATIENTS REPORTING WEATHER VARIABLES INFLUENCING PAIN, MEAN WEATHER INFLUENCE SCORES, AND STABILITY OVER TIME COEFFICIENTS Biometeorologic variable
Percentage reporting
Influence scores
Stability over time
Temperature (hot or cold) Sudden weather changes Humidity (damp or dry) Precipit;tion (rain or snow) Thunderstorms Sunshine
87 76 73 72 52 52
3.36 2.98 2.91 2.95 2.34 2.03
0.81 0.75 0.88 0.87 0.77 0.72
the WPS are listed in Tables I and development of the WPQ is described
II, respectively. in Results.
Factor analysis
Psychometric
Procedure The DMQ, VAS, SCL-90R and WPQ were mailed to patients prior to their first clinic visit. Patients who failed to complete the assessment instruments were encouraged to do so upon arrival to the clinic; consequently, the return rate was nearly 100%. Since stability over time is conceptually important in determining weather sensitivity, a 2nd WPQ was sent to all patients 1 month following the initial assessment with 31 returned (44%). There were no differences found between those who returned the WPQ and those who did not on gender, age, pain intensity, disability, or initial WPQ scores. Statistical analyses were conducted in 4 phases. First, factor analysis was used to evaluate the dimensionality of weather sensitivity. Second, internal consistency and test-retest reliability were assessed for both parts of the WPQ. Third, descriptive analysis of the WPQ was performed to provide information about the frequency and intensity with which meteorologic variables affect pain symptom complaints. Finally, group comparisons were made between patients who reported a high degree of weather sensitivity versus patients who reported relatively low weather sensitivity.
TABLE
II
PERCENTAGE OF PATIENTS REPORTING SYMPTOM PLAINTS INFLUENCED BY WEATHER VARIABLES, WEATHER INFLUENCE SCORES, AND STABILITY TIME COEFFICIENTS
COMMEAN OVER
Symptom complaint
Percentage reporting
Influence scores
Stability over time
Joint stiffness Muscle aches/soreness Trouble sleeping Muscle weakness Poor circulation Feeling depressed Feeling anxious Hot or cold spells Feeling angry Headaches Poor appetite Upset stomach Trouble breathing Faintness Chest pains
82 79 75 74 72 67 64 59 59 58 49 46 42 41 34
3.94 3.85 3.42 3.74 3.15 2.91 2.76 2.41 2.47 2.67 2.12 1.91 2.08 2.06 1.82
0.57 0.70 0.63 0.47 0.56 0.33 0.46 0.20 0.53 0.45 0.31 0.29 0.60 0.34 0.44
\
The factor structure of the WSI was examined using least-squares factor analysis with oblique (direct oblimin) rotation. This analysis sought to determine whether the construct of weather sensitivity as defined by the 7 meteorologic variables could be represented as a single dimension based on patient reporting patterns. The Kaiser-Olkin-Mayer index was 0.83, indicating that these data were appropriate for factor analysis (Dziuban and Shirkey 1974). Both the root-l criterion and the scree plot of eigenvalues (Hakstian et al. 1982) suggested a 2-factor model, accounting for 48% of the common variance, as most appropriate. Factor 1 appeared to reflect a global WSI emphasizing moisture variables including humidity, rain, thunderstorm activity, temperature, and weather changes. In contrast, factor 2 included winds and sunshine. In addition, sunshine was highly correlated with both factors. Since past literature has shown sunshine to be correlated with pain report (Laborde et al. 1986; Jamison and Parris 1990) coupled with our findings that sunshine was associated with both factors, only wind was omitted from a revised WSI. Factor analysis was not deemed appropriate for the second part of the WPQ since symptom complaints were expected to reflect multiple etiologies in addition to meteorologic variables; hence, the likelihood of finding an interpretable underlying factor structure was low. In addition, many of the symptom complaints were not shared by all patients. Reliability
Alpha reliability was 0.94 for the revised WSI, and each item correlated highly with the total score (0.760.90) indicating very good internal consistency of the scale. Test-retest reliability coefficients were high for each meteorologic variable (range: 0.71-0.87) with the stability of the total WSI score estimated as 0.89. In contrast, reliability estimates for the symptom complaints affected by meteorologic variables was lower. Internal consistency of the second part of the WPQ was 0.87; however, correlations between specific symptom complaints and total score ranged widely between 0.21 and 0.80. In addition, test-test reliability coefficients were generally low, ranging from 0.20 to 0.70 with total score reliability being less than adequate at 0.61. The symptoms most reliably reported as influenced by the weather were muscle aches and soreness (r = 0.70), difficulty sleeping (r = 0.63), trouble getting breath (r = 0.60) and stiffness in joints (r = 0.57). On balance, these findings suggest that, although patients consistently identify the meteorologic variables thought to influence their pain, the actual pain symptoms reported to be influenced by the weather are not stable.
Description of weuther .sensitiritJ The overall mean WSI score was 2.67 (SD. = 1.16) Cor our sample. As seen in Table I, the meteorologic variables most frequently endorsed as having at least some influence on pain complaint were temperature (mean: 3.36, 87%) endorsed) followed by sudden weather changes (mean: 2.98, 76% endorsed), humidity (mean: 2.91, 73.1% endorsed) and precipitation (mean: 2.95, 72.5% endorsed). Only 3% (N = 2) of patients indicated that no meteorologic variables affected their pain. As seen in Table II, symptom complaints influenced by the weather included stiffness in joints (mean: 3.94, 82%~ endorsed) and muscle aches/soreness (mean: 3.85, 79% endorsed) were reported most frequently. In contrast, faintness (mean: 2.06, 41% endorsed) and chest pains (mean: 1.82, 34% endorsed) were reported as least influenced by the weather; however, both of these symptoms were not reported reliably over time (0.43 and 0.34, respectively). In order to examine the associations between the meteorologic variables and specific symptom complaints, Canonical Correlation Analysis (CCA) was performed between the 6 meteorologic variables and the 4 most frequently reported symptom complaints (cf., Tables I and II). CCA allows for the evaluation of key multivariatc associations between each set of variables while accounting for the intercorrelations within each variable set. A significant association (Wilk’s F (24, 182) = 2.20, P < 0.002) was found accounting for 45% of the variance. Using a conservative cutoff score of 0.30 for judging canonical coefficients, the canonical variate appeared to be primarily a function of sudden wcathcr changes which were positively associated with muscle aches/soreness and joint stiffness.
Patients were divided into 2 groups (high or low weather sensitivity) using the revised WSI median score (median: 2.60) as a cutoff point. This grouping method was based upon the distribution of the WSI scores and was used to avoid making unfounded assumptions regarding extent of weather sensitivity. A series of planned t tests were conducted comparing weather sensitivity status with pain intensity, pain duration, disability status, global psychological distress, depression and sleep complaints. These analyses revealed that high weather-sensitive patients reported greater pain (t (59) = 3.61. P < 0.002), longer duration (t (61) = 2.54, P < 0.02), and more difficulty sleeping (t (60) = -3.23. P < 0.002) than low weather-sensitive patients. High weather-sensitive patients reported mean pain intensity levels of 4.12 on the VAS as compared to 3.52 for low weather-sensitive patients. Similarly, high weather-sensitive patients reported average pain duration of 9.8 years as compared to 3.8 years for low
wcathcr-sensitive patients. Finally, high weather-sensitive patients reported mean sleep disturbance levels of 2.69 on the SCL-90R as compared to 1.53 for low weather-sensitive patients. No significant group differcnces were found for disability status and global psychological distress. Of interest, there was trend (P < 0.10) for high weather-sensitive patients to report higher depression t scores on the SCL-90R (mean: 65.30) versus low weather-sensitive patients (mean: 55.25). A set of chi-square analyses compared weather sensitivity groups on measures of educational status and gender; no significant differences were found.
Discussion In summary, our findings suggest that judgments about global weather sensitivity (WSI) can be reliably measured in chronic pain patients. Nearly three-fourths of our sample indicated that temperature, humidity, precipitation and sudden weather changes affect their pain to some degree. In fact, very few patients (3%) reported no association between the weather and their pain. These findings are consistent with those of Jamison and Parris (1990) and Jamison et al. (1989) who also examined weather sensitivity in a heterogeneous group of pain patients. In contrast, our patients were generally unable to reliably identify specific symptoms which arc consistently influenced by the weather over time, except for a minority of complaints such as muscle aches and soreness. Consequently, it appears that patients could only identify if weather influenced their pain but could not reliably determine which symptoms were consistently affected. This pattern of findings suggests that the effect of weather may be mediated by psychological factors. such as unreliability of observation, beliefs and expectancies regarding the influence of weather on pain, and mood variables. A straightforward explanation is that patient judgments are inconsistent and inaccurate, due to biased observations. Turk and Salovey (1985) have characterized several common inferential errors people make including selective attention and confirmation bias which may account for patient unreliability in reporting specific weather and pain complaint relationships. For example, laboratory studies have demonstrated that the most perceptually salient stimuli can often dominate causal attributions even when the salient stimuli are presented randomly (Nisbett and Ross 1980). Our finding that patients who attributed their pain to fluctuating weather conditions also reported greater pain intensity is consistent with this attributional style. Potential misattribution of this sort is more likely when the true causes arc unclear, as is often the case in many chronic musculoskeletal pain syndromes. In fact, a frequently reported source of
203
treatment dissatisfaction among low-pain patients is lack of an adequate explanation for their pain (Deyo and Diehl 1986). In addition, patient beliefs and expectancies may contribute to misattribution through confirmation biases (Turk and Salovey 1985). Confirmation bias is defined as the tendency to seek out, notice, and recall things that support, rather than disconfirm, a particular belief. Consequently, patients may notice and remember weather conditiqns only when their pain worsens while discounting other less salient variables such as increased stress level and physical deconditioning. Similarly, if patients believe that a specific weather condition is the cause of their pain, they may interpret their pain differently during that particular weather condition. On balance, these sources of error in observation are likely to contribute to poor reliability of specific weather-pain complaint relationships while strengthening the global belief that weather variables influence pain. Our finding that self-reported weather-sensitive patients have longer pain histories than low weathersensitive patients is not inconsistent with the notion that, over time, beliefs about the influence of weather condition on pain have strengthened. Another potential mediating variable influencing judgments about weather and pain is mood state. There is a burgeoning literature examining seasonality in depression and, more specifically, the effect of ultraviolet light on mood (Wehr and Rosenthal 1989). In addition, there is a large number of anecdotal reports describing the effect of weather conditions upon mood (Rosen 1979). Coupled with the literature examining the co-occurrence of pain and depression (Roman0 and Turner 1985), it is not unreasonable to speculate that weather conditions may affect pain in part via changes in mood. Furthermore, there is literature to suggest that current mood state biases retrospective memory for pain (Erskine et al. 1990) and improves memory for negative events in general (Turk and Salovey 1985). Our tentative findings that high weather-sensitive patients reported more symptoms of depression and greater sleep difficulties than low weather-sensitive patients is consistent with the notion that mood mediates a causal relationship between weather conditions and pain or that mood mediates retrospective judgements about these events. Aside from the many factors biasing patient judgments regarding the influence of weather upon pain, our findings suggest consistent associations between meteorologic variables and specific pain complaints which are likely to form a basis for many patient judgments. The associations revealed in the CCA between sudden weather changes and symptoms of muscle aches/soreness and joint stiffness is perhaps the most cited association between weather and pain (Hollander 1961; Tromp 1963; Hill 1972; Rosen 1979;
Persinger 1980; Patberg et al. 1985; Roman0 and Turner 1985; Sibley 1985; Laborde et al. 1986; Landsberg 1986; Rasker et al. 1986; Patberg 1987 1989; Jamison et al. 1989; Jamison and Parris, 1990) and is typically associated with arthritis pain as this population has received the most research attention. However, our findings and those of others (Jamison and Parris 1990) suggest that patients with musculoskeletal pain may also suffer from similar weather-related problems. In this regard, Landsberg (1986) has reviewed evidence which suggests that scar pain is greater during stormy, wet seasons and is associated with high humidity. This latter finding is relevant to our sample, as 58% of our sample reported pain onset following accident and 41% of our patients report having had surgery; hence, many of our patients can be expected to evidence scarring, although this was not assessed. In summary, our findings underscore the importance of assessing patient beliefs and expectations about the weather and their pain. Since it is likely that some .patient beliefs about the weather and their pain have an organic basis, further studies are needed to tease apart ‘true’ meteorologic effects from those only thought to exist. Finally, our findings extend the literature detailing patients’ beliefs about their pain which can potentially influence compliance with conservative treatment recommendations.
References Dequeker, J. and Wuestenraed, L., The effect of biometeorological factors on Ritchie Articular Index and pain in rheumatoid arthritis, Stand. J. Rheumatol., 15 (1986) 280-284. Derogatis, L.R., SCL-90 Administration, scoring and procedures manual, John Hopkins, Baltimore, MD, 1977. Deyo, R.A. and Diehl, A.K., Patient satisfaction with medical care for low-back pain, Spine, 11 (1986) 28-30. Dziuban, D.D. and Shirkey, E.C., When is a correlation matrix appropriate for factor analysis?, Psychol. Bull., 81 (1974) 358-361. Erskine, A., Morley, S. and Pearce, S., Memory for pain: A Review, Pain 41 (1990) 255-265. Hakstian, A.R., Rogers, W.T. and Cattell, R.B., The behavior of number-of-factors rules with stimulated data, Multivar. Behav. Res., 17 (1982) 193-219. Hill, D.F., Climate and arthritis in arthritis and allied conditions. In: J.L. Hollander and D.C. McCarty (Eds.), A Textbook of Rheumatology, 8th edn., Lea and Feringer, Philadelphia, PA, 1972, pp. 256-263. Hollander, J.L., The controlled-climate chamber for study of the effects of meteorological changes on human diseases, Trans. NY Acad. Sci., 24 (1961) 167-172. Jamison, R.N. and Parris, W.C.V., Weather patterns affect chronic pain: myth or reality?, Pain, Suppl. 5 (1990). Jamison, R.N., Lindsey, K.A., Philpo, B. and Parris, W.C.V., How does change in weather influence chronic pain?, Paper presented at the 8th Annual Scientific Meeting of the American Pain Society, 1989, Phoenix, AZ. Laborde, J.M., Dando, W.A. and Powers, M.J., Influence of weather on osteoarthritics, Sot. Sci, Med., 23 (1986) 549-554.
204 Landsberg, H.E.. Weather. climate and you. weatherwise, 10 (19X6) 24x-253. Nisbett, R.E. and Ross, L.D., Human Inference: Strategies and Shortcomings of Informal Judgement, Prentice-Hall, Old Tappan. NJ, 1980. Ohnhaus. E.E. and Adler. R., Methodological problems in the measurement of pain: a comparison between the Verbal Rating Scale and the Visual Analogue Scale, Pain. 1 (1975) 379-384. Patberg, W.R., Effect of weather on daily pain score in rheumatoid arthritis, Lancet, ii (1987) 3X6-387. Patberg. W.R., Effect of temperature and humidity on daily pain score in a patient with rheumatoid arthritis, Arthr. Rheum.. 32 (1989) 1627-1629. Patberg, W.R., Nienhuis, R.L.F. and Veringa, F., Relation between meteorological factors and pain in rheumatoid arthritis in a marine climate, J. Rheumatol., 12 (19X5) 71 l-715. Persinger, MA.. The Weather Matrix and Human Behavior, Praeger. New York. 1980. Price. D.D.. McCrath. P.. Rafii, A. and Buckingham, B.. The validation of visual analogue scales as ratio scaled measures for chronic and experimental pain, Pain, 17 (1983) 45-56. Rasker. J.J., Peters, H.J.G. and Boon, D.L.. Influence of weather on stiffness and force in patients with rheumatoid arthritis. Stand. J. Rheumatol.. IS (1986) 27-36. Romano. J.M. and Turner, J.A., Chronic pain and depression: does the evidence support a relationship. “. Psychol. Bull.. 97 (19X.5) 1x-34.
Rosen, S., Weathering, M. Evans, New York, 1979. Shutty, M.S.. Degood, D.E. and Schwartz, D.P., Psychological dimensions or distress in chronic pain patients: a factor analytic study of Symptom Checklist-90 responses, J. Consult. Clin. Psychol.. 54 (1986) 836-X42. Sibley, J.T.. Weather and arthritis symptoms, J. Rheumatol., 12 (1985) 707-710. Tait, R.C., Pollard. A., Margolis, R.B., Duckro, P.N., and Krause, S.J., The pain disability index: psychometric and validity data. Arch. Phys. Med. Rehabil., 6X (19X7) 43X-441. Tait. R.C., Chibnall. J.T. and Krause S., The pain disability index: psychometric properties, Pain. 40 (1990) 171-182. Toth, J.. Old weather sayings are rational - and here’s their rationale, Sci. Digest. SO (1979) 70-73. Tromp, SW., Medical Biometeorology. Elsevier, New York, 1963. Turk. D.C. and Salovey, P.. Cognitive structures, cognitive processes and cognitive-behavior modification. I. Client Issues, Cogn. Ther. Res.. 9 (19x5) l-17. Wchr. T.A. and Rosenthal. N.E.. Seasonality and affective illness. Am. J. Psychiat., 146 (19X9) X29-X34. Yunus. M., Masi, A.T., Calabro, J.J.. Miller, K.A. and Feigenbaum. S.L., Primary fibromyalgia (fibrositis): clinical study of SO patients with matched normal controls. Sem. Arthr. Rheum., 2 (1981) ISl-171.
199 Publishers
B.V. All rights reserved
0304-3959/92/$05.00
PAIN 02018
Pain complaint and the weather: weather sensitivity and symptom complaints in chronic pain patients Michael 0 Department
S. Shutty,
Jr. a,b, Gary Cundiff
b and Douglas
E. DeGood
b
of Psychology, Clinical Studies Unit, Western State Hospital, Staunton, VA 24401 (USA) and ’ Pain Management Center, Department of Anesthesiology, University of Virginia Medical School, Charlottescille, VA 22901 (USA) (Received
19 June 1991, accepted
10 October
1991)
Chronic pain patients frequently report that weather conditions affect their pain; however, no Summary standardized measures of weather sensitivity have been developed. We describe the development and use of the Weather and Pain Questionnaire (WPQ) which assesses patient sensitivity to meteorologic variables defined by the National Weather Service (e.g., temperature, precipitation). Seventy chronic pain patients (59% females) with an average age of 43 years completed the WPQ. The instrument was revised using factor analysis to produce a Weather Sensitivity Index (WSI) (48% of variance) with high internal consistency (0.93) and test-retest reliability (r = 0.89). Reporting patterns suggested that patients could reliably identify which meteorologic variables influenced their pain but could not reliably determine which physical symptoms were consistently affected. The most frequently reported meteorologic variables which affect pain complaint were temperature (87%) and humidity (77%). The most frequently reported physical complaints associated with the weather were joint and muscle aches (82% and 79%, respectively). Patients labeled as being ‘weather sensitive’, defined by greater than median scores on the WPQ, reported significantly greater pain intensity, greater chronicity of pain problems, and more difficulties sleeping than patients with low scores on the WPQ. No differences in gender, education level, disability status, or global psychological distress were found. Results are discussed with respect to physiological and psychological mediating variables. Key words: Weather sensitivity; Pain complaint; Meteorology
Introduction Chronic pain patients frequently report that weather conditions affect their pain. A recent survey of 211 chronic pain patients revealed that 60% believed that meteorologic variables influenced their pain, with cold and damp conditions being rated as most problematic (Jamison et al. 1989). However, despite a vast literature addressing the topic of biometeorology (Tromp 19631, empirical studies of this phenomenon in chronic pain patients are scant.
Correspondence to: Michael S. Shutty, Jr., Ph.D., Western State Hospital, Clinical Studies Unit, P.O. Box 2500, Staunton, VA 24401, USA.
Observations that meteorologic variables such as temperature and humidity affect health status date to antiquity (Rosen 1979). For example, the Greek physician Hippocrates (about 400 BC) emphasized the seasonal dependence of medical disease (Landsberg 19861. Many of these observations have been reflected in folklore such as ‘aches and pains, coming rains’ (Toth 1979). Even slang language underscores the relationship between meteorologic variables and health in phrases such as ‘feeling under the weather’ used to describe when one is ill or depressed (Rosen 1979). Clinical case observations have detailed striking associations between meteorologic variables and pain complaint. An early example entitled ‘The relations of pain to weather, being a study of the natural history of a case of traumatic neuralgia’ published in the American Journal of Medical Sciences in 1877 describes a case
of phantom limb pain. The patient kept an exhaustive diary of his pain which was interpreted by the physician as resulting from coincident weather conditions. The author concluded that approaching storms, dropping barometric pressure and rain were associated with increased pain complaint (Landsberg 1986). Clinical studies have identified subgroups of patients, for example those diagnosed with arthritis (Hill lY72; Patberg et al. 1985; Sibley 1985; Laborde et al. 1986; Rasker et al. 1986; Patberg, 1987. 1989) and fibromyalgia (Yunus et al. 1981), who frequently report pain exacerbations associated with meteorologic variables. Hill (1972) has estimated that 80-900/o of pcrsons with arthritis report weather sensitivity, concluding that cold climate and falling barometric pressure combined with increased humidity were associated with reported pain exacerbation. Hollander (196 1) reported similar results for arthritis patients when barometric pressure and humidity were manipulated in a controlled laboratory environment. In a related vein, Yunus et al. (1981) reported that 92% of their series of 50 consecutive primary fibromyalgia patients reported ‘cold or humid weather’ as modulating their pain symptoms. Consequently, modulation of symptoms by meteorologic variables has been suggested as a diagnostic criterion for primary fibromyalgia. Unfortunately, there have been no attempts in the clinical literature to develop reliable and standardized instruments to measure self-report of weather sensitivity despite some effort in the lay literature (Rosen 1979). Hence, the results of survey studies are difficult to interpret, with respect to the consistency with which specific meteorologic variables are reported and for which specific groups of patients are involved. In addition, there are many diagnostic groups of patients who have been relatively neglected in the research literature despite frequently reported antidotal associations with meteorologic variables (e.g., migraine hcadachcs (Tromp 1963; Rosen 1979; Persinger l%O), sinus headaches (Rosen 19791, and scar pain (Rosen 197’); Landsberg 1986)). In our own clinical experience, weather sensitivity has been noted for a variety of musculoskeletal pain disorders including low back, neck and shoulder pain. It is not uncommon for a patient to report non-compliance with prescribed treatments (i.e.. exercise) due to increased pain attributed to changing weather conditions. This study describes the development of a brief internally consistent and reliable self-report weather sensitivity index for use with chronic pain patients suffering from musculoskeletal disorders. In addition, the relationship between self-reported weather sensitivity and frequently reported physical symptoms are examined. Finally, differences between ‘high weathersensitive’ and ‘low weather-sensitive’ patients are explored.
Methods
Subjects Seventy consecutive outpatienls lor chronic subjects
pain problems
for this study.
(41’4
males) receiving treatment
at a large university
Data
were
collected
hospital
during
served as
the
patient‘s
initial visit to our clinic. The average age was 43.2 years (SD. range: IV-XI)
with a mean education
The most frequent
presenting
complaint
was low hack pain (60%).
followed
by upper back (3% 1. neck. arm and shoulder
Patients
with
rheumatoid
disorders
(including
excluded from study and represented of our clinic
population.
pain (37%).
fihromyalgia)
were
only a minority (less than
Pain duration
months and more than
= 13.7;
level of 12.2 years (S.D. = 5.Y).
ranged
widely
5%)
between
IO years in several cases and averaged
years (S.D. = 9.6). Accidents were the chief precipitating reported,
and 41%
of patients
treatment
for pain-related
reported
having
3 6.7
event (58%)
received
surgical
problems.
lnstrumentx Lkwoquphic DMQ
urui Mcd~ul
is a self-report
Hisrory Questionnuiw
(DMQI.
booklet which provides descriptive
about pain problems,
past medical treatments
and current
ing. Patients rated how much their pain interfered daily activities (work,
household
The
information function-
with a number of
chores. socializing,
sexual relations.
cxerciae and sleep) using a S-point scale in which I is ‘not at all’ and 5 is ‘extremely’.
A
summary
score
provided
an index of self-rated
disability
status has demonstrated
number of studies (Tail
averaged
disability.
across all
This method
good reliability
VisudAnulog SC& IVES).
of the pain,
nearest millimeter.
in a
The VAS consists of a IO-cm line with
is asked to indicate
to the intensity
and validity
et al. 1987. 1990).
the ends marked with ‘none’ and ‘severe’ (Ohnhaus The patient
activities
of assessing
and Adler
1975).
the point on the line corresponding
and the distance
This measure
is evaluated
to the
has been shown to be a valid and
reliable tool for measuring both clinical and experimental
pain (Price
et al. 19x.3. ,(;\.~q~t~rtt C‘hecXlisr-Yt/ Kcrisd
CSC‘L-OUR). The
rogatis 1977) has heen advocated
SCL-YOR
(De-
as a useful measure of psychologi-
cal distress in chronic pain patients (Shutty et al. 1086). Test-retest reliability
coefficients
internal
1 week
for
range from 0.78 to O.YO, whereas
consistency of the subscales range from 0.77 to 0.90 (De-
rogatis
lY77).
The
General
global psychological additional
Symptom
Index was used to provide
distress index. The
sleep items comprising
were used to provide
Depression
a sleep factor (Shutty et al. 1986)
an index of mood and sleep disorder,
tively. The Somatization
a
auhscale and 3 respec-
subscale was not used due to overlap
with
the WPQ.
Wcuther umi Pum L)utwwwu~~re CWPQI. The WPQ questionnaire a Symptom
which provides a Weather Complaint
meteorologic record affected ‘flow
variables
daily weather
(ranging
Checklist.
from
‘not
Part
used by the fluctuations:
at all’
by these variables.
1 of the WPQ National
patients
to ‘extreme’) For example,
literature
consists of symptoms
to he frequently
reported
and
consists of 7
Weather
Service
rate on a S-point how much
their
an item from part
much is your pain affected by temperature
2 of the WPQ
is a I-part
Sensitivity Index (WSI)
to
scale
pain
is
1 asked.
(hot or cold)?‘. Part
revealed
after
as affected
review
of the
by the weather.
This
lrsting of symptoms was reduced to 15 based on a pilot study in which SO consecutive
patients
with musculoskeletal
what symptoms, if any. were affected 2 of the WPQ, how
much
example.
the
patients
disorders
by weather
were asked
conditions.
rate on a S-point scale (as indicated
symptoms
are
made
worse
an item from part 2 asked, ‘How
made w’orse by the weather’!‘.
hy the
On part ahovcj
weather.
For-
much is stiffness in joints
The item content for parts I and 2 of
201 TABLE
I
Results
PERCENTAGE OF PATIENTS REPORTING WEATHER VARIABLES INFLUENCING PAIN, MEAN WEATHER INFLUENCE SCORES, AND STABILITY OVER TIME COEFFICIENTS Biometeorologic variable
Percentage reporting
Influence scores
Stability over time
Temperature (hot or cold) Sudden weather changes Humidity (damp or dry) Precipit;tion (rain or snow) Thunderstorms Sunshine
87 76 73 72 52 52
3.36 2.98 2.91 2.95 2.34 2.03
0.81 0.75 0.88 0.87 0.77 0.72
the WPS are listed in Tables I and development of the WPQ is described
II, respectively. in Results.
Factor analysis
Psychometric
Procedure The DMQ, VAS, SCL-90R and WPQ were mailed to patients prior to their first clinic visit. Patients who failed to complete the assessment instruments were encouraged to do so upon arrival to the clinic; consequently, the return rate was nearly 100%. Since stability over time is conceptually important in determining weather sensitivity, a 2nd WPQ was sent to all patients 1 month following the initial assessment with 31 returned (44%). There were no differences found between those who returned the WPQ and those who did not on gender, age, pain intensity, disability, or initial WPQ scores. Statistical analyses were conducted in 4 phases. First, factor analysis was used to evaluate the dimensionality of weather sensitivity. Second, internal consistency and test-retest reliability were assessed for both parts of the WPQ. Third, descriptive analysis of the WPQ was performed to provide information about the frequency and intensity with which meteorologic variables affect pain symptom complaints. Finally, group comparisons were made between patients who reported a high degree of weather sensitivity versus patients who reported relatively low weather sensitivity.
TABLE
II
PERCENTAGE OF PATIENTS REPORTING SYMPTOM PLAINTS INFLUENCED BY WEATHER VARIABLES, WEATHER INFLUENCE SCORES, AND STABILITY TIME COEFFICIENTS
COMMEAN OVER
Symptom complaint
Percentage reporting
Influence scores
Stability over time
Joint stiffness Muscle aches/soreness Trouble sleeping Muscle weakness Poor circulation Feeling depressed Feeling anxious Hot or cold spells Feeling angry Headaches Poor appetite Upset stomach Trouble breathing Faintness Chest pains
82 79 75 74 72 67 64 59 59 58 49 46 42 41 34
3.94 3.85 3.42 3.74 3.15 2.91 2.76 2.41 2.47 2.67 2.12 1.91 2.08 2.06 1.82
0.57 0.70 0.63 0.47 0.56 0.33 0.46 0.20 0.53 0.45 0.31 0.29 0.60 0.34 0.44
\
The factor structure of the WSI was examined using least-squares factor analysis with oblique (direct oblimin) rotation. This analysis sought to determine whether the construct of weather sensitivity as defined by the 7 meteorologic variables could be represented as a single dimension based on patient reporting patterns. The Kaiser-Olkin-Mayer index was 0.83, indicating that these data were appropriate for factor analysis (Dziuban and Shirkey 1974). Both the root-l criterion and the scree plot of eigenvalues (Hakstian et al. 1982) suggested a 2-factor model, accounting for 48% of the common variance, as most appropriate. Factor 1 appeared to reflect a global WSI emphasizing moisture variables including humidity, rain, thunderstorm activity, temperature, and weather changes. In contrast, factor 2 included winds and sunshine. In addition, sunshine was highly correlated with both factors. Since past literature has shown sunshine to be correlated with pain report (Laborde et al. 1986; Jamison and Parris 1990) coupled with our findings that sunshine was associated with both factors, only wind was omitted from a revised WSI. Factor analysis was not deemed appropriate for the second part of the WPQ since symptom complaints were expected to reflect multiple etiologies in addition to meteorologic variables; hence, the likelihood of finding an interpretable underlying factor structure was low. In addition, many of the symptom complaints were not shared by all patients. Reliability
Alpha reliability was 0.94 for the revised WSI, and each item correlated highly with the total score (0.760.90) indicating very good internal consistency of the scale. Test-retest reliability coefficients were high for each meteorologic variable (range: 0.71-0.87) with the stability of the total WSI score estimated as 0.89. In contrast, reliability estimates for the symptom complaints affected by meteorologic variables was lower. Internal consistency of the second part of the WPQ was 0.87; however, correlations between specific symptom complaints and total score ranged widely between 0.21 and 0.80. In addition, test-test reliability coefficients were generally low, ranging from 0.20 to 0.70 with total score reliability being less than adequate at 0.61. The symptoms most reliably reported as influenced by the weather were muscle aches and soreness (r = 0.70), difficulty sleeping (r = 0.63), trouble getting breath (r = 0.60) and stiffness in joints (r = 0.57). On balance, these findings suggest that, although patients consistently identify the meteorologic variables thought to influence their pain, the actual pain symptoms reported to be influenced by the weather are not stable.
Description of weuther .sensitiritJ The overall mean WSI score was 2.67 (SD. = 1.16) Cor our sample. As seen in Table I, the meteorologic variables most frequently endorsed as having at least some influence on pain complaint were temperature (mean: 3.36, 87%) endorsed) followed by sudden weather changes (mean: 2.98, 76% endorsed), humidity (mean: 2.91, 73.1% endorsed) and precipitation (mean: 2.95, 72.5% endorsed). Only 3% (N = 2) of patients indicated that no meteorologic variables affected their pain. As seen in Table II, symptom complaints influenced by the weather included stiffness in joints (mean: 3.94, 82%~ endorsed) and muscle aches/soreness (mean: 3.85, 79% endorsed) were reported most frequently. In contrast, faintness (mean: 2.06, 41% endorsed) and chest pains (mean: 1.82, 34% endorsed) were reported as least influenced by the weather; however, both of these symptoms were not reported reliably over time (0.43 and 0.34, respectively). In order to examine the associations between the meteorologic variables and specific symptom complaints, Canonical Correlation Analysis (CCA) was performed between the 6 meteorologic variables and the 4 most frequently reported symptom complaints (cf., Tables I and II). CCA allows for the evaluation of key multivariatc associations between each set of variables while accounting for the intercorrelations within each variable set. A significant association (Wilk’s F (24, 182) = 2.20, P < 0.002) was found accounting for 45% of the variance. Using a conservative cutoff score of 0.30 for judging canonical coefficients, the canonical variate appeared to be primarily a function of sudden wcathcr changes which were positively associated with muscle aches/soreness and joint stiffness.
Patients were divided into 2 groups (high or low weather sensitivity) using the revised WSI median score (median: 2.60) as a cutoff point. This grouping method was based upon the distribution of the WSI scores and was used to avoid making unfounded assumptions regarding extent of weather sensitivity. A series of planned t tests were conducted comparing weather sensitivity status with pain intensity, pain duration, disability status, global psychological distress, depression and sleep complaints. These analyses revealed that high weather-sensitive patients reported greater pain (t (59) = 3.61. P < 0.002), longer duration (t (61) = 2.54, P < 0.02), and more difficulty sleeping (t (60) = -3.23. P < 0.002) than low weather-sensitive patients. High weather-sensitive patients reported mean pain intensity levels of 4.12 on the VAS as compared to 3.52 for low weather-sensitive patients. Similarly, high weather-sensitive patients reported average pain duration of 9.8 years as compared to 3.8 years for low
wcathcr-sensitive patients. Finally, high weather-sensitive patients reported mean sleep disturbance levels of 2.69 on the SCL-90R as compared to 1.53 for low weather-sensitive patients. No significant group differcnces were found for disability status and global psychological distress. Of interest, there was trend (P < 0.10) for high weather-sensitive patients to report higher depression t scores on the SCL-90R (mean: 65.30) versus low weather-sensitive patients (mean: 55.25). A set of chi-square analyses compared weather sensitivity groups on measures of educational status and gender; no significant differences were found.
Discussion In summary, our findings suggest that judgments about global weather sensitivity (WSI) can be reliably measured in chronic pain patients. Nearly three-fourths of our sample indicated that temperature, humidity, precipitation and sudden weather changes affect their pain to some degree. In fact, very few patients (3%) reported no association between the weather and their pain. These findings are consistent with those of Jamison and Parris (1990) and Jamison et al. (1989) who also examined weather sensitivity in a heterogeneous group of pain patients. In contrast, our patients were generally unable to reliably identify specific symptoms which arc consistently influenced by the weather over time, except for a minority of complaints such as muscle aches and soreness. Consequently, it appears that patients could only identify if weather influenced their pain but could not reliably determine which symptoms were consistently affected. This pattern of findings suggests that the effect of weather may be mediated by psychological factors. such as unreliability of observation, beliefs and expectancies regarding the influence of weather on pain, and mood variables. A straightforward explanation is that patient judgments are inconsistent and inaccurate, due to biased observations. Turk and Salovey (1985) have characterized several common inferential errors people make including selective attention and confirmation bias which may account for patient unreliability in reporting specific weather and pain complaint relationships. For example, laboratory studies have demonstrated that the most perceptually salient stimuli can often dominate causal attributions even when the salient stimuli are presented randomly (Nisbett and Ross 1980). Our finding that patients who attributed their pain to fluctuating weather conditions also reported greater pain intensity is consistent with this attributional style. Potential misattribution of this sort is more likely when the true causes arc unclear, as is often the case in many chronic musculoskeletal pain syndromes. In fact, a frequently reported source of
203
treatment dissatisfaction among low-pain patients is lack of an adequate explanation for their pain (Deyo and Diehl 1986). In addition, patient beliefs and expectancies may contribute to misattribution through confirmation biases (Turk and Salovey 1985). Confirmation bias is defined as the tendency to seek out, notice, and recall things that support, rather than disconfirm, a particular belief. Consequently, patients may notice and remember weather conditiqns only when their pain worsens while discounting other less salient variables such as increased stress level and physical deconditioning. Similarly, if patients believe that a specific weather condition is the cause of their pain, they may interpret their pain differently during that particular weather condition. On balance, these sources of error in observation are likely to contribute to poor reliability of specific weather-pain complaint relationships while strengthening the global belief that weather variables influence pain. Our finding that self-reported weather-sensitive patients have longer pain histories than low weathersensitive patients is not inconsistent with the notion that, over time, beliefs about the influence of weather condition on pain have strengthened. Another potential mediating variable influencing judgments about weather and pain is mood state. There is a burgeoning literature examining seasonality in depression and, more specifically, the effect of ultraviolet light on mood (Wehr and Rosenthal 1989). In addition, there is a large number of anecdotal reports describing the effect of weather conditions upon mood (Rosen 1979). Coupled with the literature examining the co-occurrence of pain and depression (Roman0 and Turner 1985), it is not unreasonable to speculate that weather conditions may affect pain in part via changes in mood. Furthermore, there is literature to suggest that current mood state biases retrospective memory for pain (Erskine et al. 1990) and improves memory for negative events in general (Turk and Salovey 1985). Our tentative findings that high weather-sensitive patients reported more symptoms of depression and greater sleep difficulties than low weather-sensitive patients is consistent with the notion that mood mediates a causal relationship between weather conditions and pain or that mood mediates retrospective judgements about these events. Aside from the many factors biasing patient judgments regarding the influence of weather upon pain, our findings suggest consistent associations between meteorologic variables and specific pain complaints which are likely to form a basis for many patient judgments. The associations revealed in the CCA between sudden weather changes and symptoms of muscle aches/soreness and joint stiffness is perhaps the most cited association between weather and pain (Hollander 1961; Tromp 1963; Hill 1972; Rosen 1979;
Persinger 1980; Patberg et al. 1985; Roman0 and Turner 1985; Sibley 1985; Laborde et al. 1986; Landsberg 1986; Rasker et al. 1986; Patberg 1987 1989; Jamison et al. 1989; Jamison and Parris, 1990) and is typically associated with arthritis pain as this population has received the most research attention. However, our findings and those of others (Jamison and Parris 1990) suggest that patients with musculoskeletal pain may also suffer from similar weather-related problems. In this regard, Landsberg (1986) has reviewed evidence which suggests that scar pain is greater during stormy, wet seasons and is associated with high humidity. This latter finding is relevant to our sample, as 58% of our sample reported pain onset following accident and 41% of our patients report having had surgery; hence, many of our patients can be expected to evidence scarring, although this was not assessed. In summary, our findings underscore the importance of assessing patient beliefs and expectations about the weather and their pain. Since it is likely that some .patient beliefs about the weather and their pain have an organic basis, further studies are needed to tease apart ‘true’ meteorologic effects from those only thought to exist. Finally, our findings extend the literature detailing patients’ beliefs about their pain which can potentially influence compliance with conservative treatment recommendations.
References Dequeker, J. and Wuestenraed, L., The effect of biometeorological factors on Ritchie Articular Index and pain in rheumatoid arthritis, Stand. J. Rheumatol., 15 (1986) 280-284. Derogatis, L.R., SCL-90 Administration, scoring and procedures manual, John Hopkins, Baltimore, MD, 1977. Deyo, R.A. and Diehl, A.K., Patient satisfaction with medical care for low-back pain, Spine, 11 (1986) 28-30. Dziuban, D.D. and Shirkey, E.C., When is a correlation matrix appropriate for factor analysis?, Psychol. Bull., 81 (1974) 358-361. Erskine, A., Morley, S. and Pearce, S., Memory for pain: A Review, Pain 41 (1990) 255-265. Hakstian, A.R., Rogers, W.T. and Cattell, R.B., The behavior of number-of-factors rules with stimulated data, Multivar. Behav. Res., 17 (1982) 193-219. Hill, D.F., Climate and arthritis in arthritis and allied conditions. In: J.L. Hollander and D.C. McCarty (Eds.), A Textbook of Rheumatology, 8th edn., Lea and Feringer, Philadelphia, PA, 1972, pp. 256-263. Hollander, J.L., The controlled-climate chamber for study of the effects of meteorological changes on human diseases, Trans. NY Acad. Sci., 24 (1961) 167-172. Jamison, R.N. and Parris, W.C.V., Weather patterns affect chronic pain: myth or reality?, Pain, Suppl. 5 (1990). Jamison, R.N., Lindsey, K.A., Philpo, B. and Parris, W.C.V., How does change in weather influence chronic pain?, Paper presented at the 8th Annual Scientific Meeting of the American Pain Society, 1989, Phoenix, AZ. Laborde, J.M., Dando, W.A. and Powers, M.J., Influence of weather on osteoarthritics, Sot. Sci, Med., 23 (1986) 549-554.
204 Landsberg, H.E.. Weather. climate and you. weatherwise, 10 (19X6) 24x-253. Nisbett, R.E. and Ross, L.D., Human Inference: Strategies and Shortcomings of Informal Judgement, Prentice-Hall, Old Tappan. NJ, 1980. Ohnhaus. E.E. and Adler. R., Methodological problems in the measurement of pain: a comparison between the Verbal Rating Scale and the Visual Analogue Scale, Pain. 1 (1975) 379-384. Patberg, W.R., Effect of weather on daily pain score in rheumatoid arthritis, Lancet, ii (1987) 3X6-387. Patberg. W.R., Effect of temperature and humidity on daily pain score in a patient with rheumatoid arthritis, Arthr. Rheum.. 32 (1989) 1627-1629. Patberg, W.R., Nienhuis, R.L.F. and Veringa, F., Relation between meteorological factors and pain in rheumatoid arthritis in a marine climate, J. Rheumatol., 12 (19X5) 71 l-715. Persinger, MA.. The Weather Matrix and Human Behavior, Praeger. New York. 1980. Price. D.D.. McCrath. P.. Rafii, A. and Buckingham, B.. The validation of visual analogue scales as ratio scaled measures for chronic and experimental pain, Pain, 17 (1983) 45-56. Rasker. J.J., Peters, H.J.G. and Boon, D.L.. Influence of weather on stiffness and force in patients with rheumatoid arthritis. Stand. J. Rheumatol.. IS (1986) 27-36. Romano. J.M. and Turner, J.A., Chronic pain and depression: does the evidence support a relationship. “. Psychol. Bull.. 97 (19X.5) 1x-34.
Rosen, S., Weathering, M. Evans, New York, 1979. Shutty, M.S.. Degood, D.E. and Schwartz, D.P., Psychological dimensions or distress in chronic pain patients: a factor analytic study of Symptom Checklist-90 responses, J. Consult. Clin. Psychol.. 54 (1986) 836-X42. Sibley, J.T.. Weather and arthritis symptoms, J. Rheumatol., 12 (1985) 707-710. Tait, R.C., Pollard. A., Margolis, R.B., Duckro, P.N., and Krause, S.J., The pain disability index: psychometric and validity data. Arch. Phys. Med. Rehabil., 6X (19X7) 43X-441. Tait. R.C., Chibnall. J.T. and Krause S., The pain disability index: psychometric properties, Pain. 40 (1990) 171-182. Toth, J.. Old weather sayings are rational - and here’s their rationale, Sci. Digest. SO (1979) 70-73. Tromp, SW., Medical Biometeorology. Elsevier, New York, 1963. Turk. D.C. and Salovey, P.. Cognitive structures, cognitive processes and cognitive-behavior modification. I. Client Issues, Cogn. Ther. Res.. 9 (19x5) l-17. Wchr. T.A. and Rosenthal. N.E.. Seasonality and affective illness. Am. J. Psychiat., 146 (19X9) X29-X34. Yunus. M., Masi, A.T., Calabro, J.J.. Miller, K.A. and Feigenbaum. S.L., Primary fibromyalgia (fibrositis): clinical study of SO patients with matched normal controls. Sem. Arthr. Rheum., 2 (1981) ISl-171.
