[What can pain diaries achieve? Advantages and limitations of their use with reference to single case evaluation.].

Pain diaries are modern evaluation instruments for assessment of pain dynamics and treatment outcome. There are syndrome-specific and non-syndrome-specific diaries. They exist as paper-and-pencil versions, as electronic diaries or as uptime recorders. Pain diaries are especially suitable for considering pain course over time. There-fore, they have many fields of application, which extend to the topography, differential diagnosis, therapeutic evaluation, classification, indications and, obviously, to questions of research. A particular advantage of pain diaries is the possibility of individual design and realization. Despite numerous advantages, problems with the application and data analysis of diaries must be kept in mind. They are instruments with a large potential reactivity of measurements, e.g. answer tendencies, memory effects, simulation. In addition, items in pain diaries usually have different scale levels, which can lead to problems in adaequate analyses of data. The data analysis of pain diaries is demonstrated with the help of selected individual cases. Three different phases are distinguished, and useful suggestions for analysis of individual cases are submitted. The first evaluation step includes the visual inspection of primary data. This provides a survey of the process of pain and at the same time a reference level for assessment of the secondary data. The secondary data can be divided into agglomeration, variation and correlation data. Agglomeration data, such as frequencies, sums or means, are a first quantitative, but still superficial, process of analysis. Additionally, the findings should be complemented by variation data to characterize these results with regard to their variability. The significance of deviation measures is demonstrated by a selected case example. A small deviation in this case reveals a relatively constant course of pain parameters, whereas large deviation reflects an unstable course. Moreover, the variability of data can be analysed by trend coefficient and time series analysis. Time-related measures of deviation, developed in the 1930s by von Neumann et al., and a simple version of time-series analyses developed by Tryon, are described and compared with one another. Finally, the use of correlation statistics is useful to reveal important relationships between secondary data. For the case example such statistically significant relations exist between pain intensity, therapeutic interventions, and sleep duration. Due to the effort required by diary analysis the use of pain diaries seems to be indicated only if there is the intention to generate quantitative data by mathematical-statistical methods. Problems arise, in spite of the quantitative data level, from the limited comparability of results and the undeniable reactivity of measurements. At present open questions refer to the desirable or non-desirable stability of surveys by means of pain diaries and the possible generation of secondary effects as the result of long-term diary recording.