Comparison of teaching machine with an observer in detection of angina pectoris by questionnaire.

Method The Geriatric Research Unit in Edinburgh uses a teaching machine (tm 1024 Educational Systems Ltd.) with an attached general purpose data collecting system (ts 512) for psychological testing. A comparison has been made, using Rose's (1962) questionnaire on angina pectoris, of questions asked by this machine with questions asked in the ordinary way. The questionnaire was put on coloured film. The first frame displayed 'the pictures you are going to see will ask you questions. You answer each question by pressing RED or BLUE. Each picture will explain which colour you press in your reply. Now press RED'. The capitals represent large red and blue buttons on the teaching machine by means of which dichotomous responses are made. The second frame said 'After this picture the questions begin. Now press BLUE'. These two frames served to familiarize the subject with the machine and to show whether the procedure had been under stood. Succeeding frames carried Rose's questions 1 to 5 exactly as published. Questions 6 and 7 were altered as described below. Question 8, concerning pain other than precordial, and question 9, concerning consultation with a doctor about the pain, were not asked. Each frame offered alternative replies in the form of 'If yes press RED, if no press BLUE'. The machine was programmed to pass steadily through the questionnaire unless the subject pressed a button for an answer which in the published questionnaire would cause the observer to ask no more questions and reject the diagnosis of angina; the machine then displayed a frame saying 'End of Questions'. Question 6 was altered to 'How soon is it relieved?' and was, of course, not asked unless question 5 was answered 'Relieved'. To obtain answers to question 7 concerning the site of the pain, 'Will you show me where it was?', three frames were needed, each of which on a diagram showed a possible pain site and asked 'Do you feel it here?' The film, therefore, as far as possible asks Rose's questions as published but explains on each frame what the alternative answers are. All persons to whom the questionnaire was administered in the ordinary way were questioned by one observer (J.S.M.) who used the training material supplied by Rose and tested reproducibility before beginning the study (Milne, Hope, and Williamson, 1970). The subjects examined were men and women aged 63 years and upwards forming a random sample of older people living in a defined area of Edinburgh and taking part in a longitudinal study. This sample has been fully described elsewhere (Milne, Maule, and Williamson, 1971). The investi gation described was performed in the second (annual) examination of the longitudinal study. Two hundred and nine men and 244 women took part. During the first examination all subjects were asked the questions in the ordinary way. One year later subjects were assigned at random to two groups. The first was asked Rose's questions in the o dinary way. The other was questioned by the teaching machine. From the replies can be given the prevalence of angina at the first examination, and the incidence between first and second examina tions can be calculated. Replies fall into three groups: (1) to the ordinary questions, (2) to the machine, and (3) to the ordinary questions when the machine was not in working order or could not be used. This machine was an early model, now obsolete, which easily went out of order and as there was no service engineer in Scotland repairs were inevitably delayed. The machine could not be used on home visits, necessi tated by illness, or in blind people. These reasons for not using the machine cannot certainly be said to be random, so although bias from this reason 105