A model of problem solving
Clinical decision-making processes are a specialized form of problem solving and have been studied within the context of problem solving and theories of information pro- cessing. It is often assumed that clinical decisions are made by the process of inductive reasoning, which involves collecting evidence and data and using this data to develop a conclusion and a hypothesis. For example, within this framework, a general practitioner would start a consultation with a patient without any prior model of their problem.
The GP would then ask the appropriate questions regarding the patient’s history and symptoms and develop a hypothesis about the presenting problem. However, doctors’ decision-making processes are generally considered within the framework of the hypothetico-deductive model of decision making. This perspective emphasizes the development of hypotheses early on in the consultation and is illustrated by Newell and Simon’s (1972) model of problem solving, which emphasizes hypothesis testing. Newell and Simon suggested that problem solving involves a number of stages that result in a solution to any given problem. This model has been applied to many different forms of problem solving and is a useful framework for examining clinical decisions (see Figure 4.2).
The stages involved are as follows:
1 Understand the nature of the problem and develop an internal representation.
At this stage, the individual needs to formulate an internal representation of the problem. This process involves understanding the goal of the problem, evaluating any given conditions and assessing the nature of the available data.
2 Develop a plan of action for solving the problem. Newell and Simon differentiated between two types of plans: heuristics and algorithms. An algorithm is a set of rules that will provide a correct solution if applied correctly (e.g. addition, multiplication, etc. involve algorithms). However, most human problem solving involves heuristics, which are rules of thumb. Heuristics are less definite and specific but provide
guidance and direction for the problem his solver. Heuristics may involve developing parallels between the present problem and previous similar ones.
3 Apply heuristics. Once developed, the plans are then applied to the given situation.
4 Determine whether heuristics have been fruitful. The individual then decides whether the heuristics have been successful in the attempt to solve the given problem.
If they are considered unsuccessful, the individual may need to develop a new approach to the problem.
5 Determine whether an acceptable solution has been obtained.
6 Finish and verify the solution. The end-point of the problem-solving process involves the individual deciding that an acceptable solution to the problem has been reached and that this solution provides a suitable outcome.
According to Newell and Simon’s model of problem solving, hypotheses about the causes and solutions to the problem are developed very early on in the process. They regarded this process as dynamic and ever-changing and suggested that at each stage of the process the individual applies a ‘means end analysis’, whereby they assess the value of the hypothesis, which is either accepted or rejected according to the evidence. This type of model involves information processing whereby the individual develops hypotheses to convert an open problem, which may be unmanageable with no obvious end-point, to one which can be closed and tested by a series of hypotheses.
Fig. 4-2 A simplified model of problem solving
Clinical decisions as problem solving
Clinical decisions can be conceptualized as a form of problem solving and involve the development of hypotheses early on in the consultation process. These hypotheses are subsequently tested by the doctor’s selection of questions. Models of problem solving have been applied to clinical decision making by several authors (e.g. MacWhinney 1973; Weinman 1987), who have argued that the process of formulating a clinical decision involves the following stages (see Figure 4.3):
1 Accessing information about the patient’s symptoms. The initial questions in any consultation from health professional to the patient will enable the health professional to understand the nature of the problem and to form an internal repre- sentation of the type of problem.
2 Developing hypotheses. Early on in the problem-solving process, the health professional develops hypotheses about the possible causes and solutions to the problem.
3 Search for attributes. The health professional then proceeds to test the hypotheses by searching for factors either to confirm or to refute their hypotheses. Research into the hypothesis testing process has indicated that although doctors aim to either confirm or refute their hypothesis by asking balanced questions, most of their questioning is biased towards confirmation of their original hypothesis. Therefore, an initial hypothesis that a patient has a psychological problem may cause the doctor to focus on the patient’s psychological state and ignore the patient’s attempt to talk about their physical symptoms. Studies have shown that doctors’ clinical information
Fig. 4-3 Diagnosis as a form of problem solving
collected subsequent to the development of a hypothesis may be systematically distorted to support the original hypothesis (Wallsten 1978). Furthermore, the type of hypothesis has been shown to bias the collection and interpretation of any information received during the consultation (Wason 1974).
4 Making a management decision. The outcome of the clinical decision-making process involves the health professional deciding on the way forward. Weinman (1987) suggested that it is important to realize that the outcome of a consultation and a diagnosis is not an absolute entity, but is itself a hypothesis and an informed guess that will be either confirmed or refuted by future events.