Diagnosestrategien in der Allgemeinmedizin: Der Beitrag von deduktivem Hypothesentesten und getriggerten Routinefragen

Background: Determining a diagnosis is one of the most demanding cognitive tasks for a physician. In a short time a large number of diagnostic possibilities have to be evaluated. Especially in general medical practice the history taking, which requires no technical means, is of central importance. Previous research on diagnostic procedures where performed under unrealistic conditions. It postulated a primarily hypothetico-deductive approach (Elstein et al. 1978). According to this model early hypotheses are formed, which structure an active hypothesis-guided information search. This approach is considered to be cognitively highly demanding. It can be assumed that it cannot be used effectively in all cases. According to the adaptive toolbox (Gigerenzer and Todd 1999) there are probably more heuristic questioning strategies used by physicians. The role of routine questions as a cognitive strategy was largely ignored in previous research and usually negatively formulated. In this study triggered routine questions are formulated and examined for the first time as a positive diagnostic instrument. These routine questions are triggered by a symptom and evaluate an organ system without being guided by a hypothesis. In this study, the role of deductive hypothesis testing and triggered routine questioning is being investigated for the first time using real empirical data in general practice. Methods: A cross-sectional study with 12 general practitioners was conducted. 282 consultations were video recorded, regardless of patients’ symptoms or concerns. 134 of those consultations contained at least one diagnostic episode, a total of 163 diagnostic episodes were analysed. After each consultation the doctors reflected upon their diagnostic process in a semi-structured interview. The consultations as well as the interviews were transcribed and evaluated by qualitative and quantitative methods. The code-tree was derived from previously published research and iteratively adapted to the data material. Results: Deductive hypothesis testing could only be identified in 39% of all diagnostic episodes. The developed diagnostic hypotheses vary, depending on the situation, between very specific and abstract. In the hypothesis development and the search for information, the peculiarities of the general medical settings are strongly reflected. Especially the early development of hypothesis on basis of complex and multi-layered context knowledge was identified as a main source of information. Each participating general practitioner used triggered routine questions as diagnostic strategy. They were used in 38% of the diagnostic episodes. There were respiratory, gastrointestinal, urogenital and global triggered routine questions. They were content associated with the presenting complaint of the patient. Triggered routine questions were used to win a systematic overview of the patient's symptoms, for reorientation in the diagnostic space and for the reduction of diagnostic uncertainty. Discussion: As a result of the present study, the hypothetico-deductive model is clearly put into perspective. Deductive hypothesis testing is only one part of the used cognitive strategies. Triggered routine questions have been identified also as an important component of the adaptive Toolbox. Triggered routine questions are the link between open strategies, such as the inductive foraging and the very specific deductive hypothesis testing. They help to explore interesting areas of the diagnostic space, when a search for information by means of specific hypotheses would unduly restrict it and would be cognitively too costly. Deductive hypothesis testing and triggered routine questions are part of a highly adaptive process of information search in the uncertain decision environment of general practice. They can be integrated in a comprehensive model of bounded rationality for complex decision situations.