Personalized analysis of pain-weather associations: a pilot study

It is a well known belief that weather can influence human health, including pain sensation. However, the current data are controversial, which might be due to the wide range of interindividual differences. The present study aimed to characterize the individual pain-weather associations during chronic pain by utilizing several data analytical methods.

The study included 3-3 patients with (P1, P3, and P4) or without (P2, P5, P6) diabetes mellitus and signs of trigeminal neuralgia or low back pain. Subjective pain scores (0-10) and 12 weather parameters (terrestrial, geomagnetic, and solar) were recorded for one month repeated three times daily. Nonparametric Spearman's correlation (Sp), multiple regression (Mx), and principal component (PCA) analyses were performed to evaluate associations between pain and meteorological factors obtained at the day of recorded pain value, 2 days before and 2 days after the recorded pain, and the changes in these parameters (5 x12 parameters). Complex scores were calculated based on the results of these analyses.

While the temperature had the highest effects on the pain levels in most of the participants, huge interindividual differences in the degree and the direction of the associations between pain and weather parameters could be obtained. The analytic methods also revealed subject specific results, and the synthesis of different statistical methods as total scores provided a personalized map for each patient, which showed disparate patterns across the study participants. Thus, Participants 2 and 5 had higher scores for Mx compared to Sp; furthermore, certain factors showed opposite direction in their associations with the pain level depending on the type of analysis (Sp vs Mx). In contrast, P3 had a lower score for Mx compared to Sp, which might suggest a low level of weather sensitivity on the association between the different weather parameters in this subject. Furthermore, participants P4 and P6 had a very high level of weather sensitivity, while P1 had an opposite pattern. Regarding the time point-related effects on the pain level, most patients were sensitive to parameters obtained at the same day or two days before, except the P1 subject, who had the highest sensitivity to weather parameters detected two days after.

The present study highlights the importance of integrating different data analysis approaches to elucidate the individual connections between pain and most of the weather parameters. In conclusion, complex personalized profiling should be considered for the characterization of pain-weather associations by applying different data analytical approaches, which may provide feedback to physicians and patients.