Search

An integrated approach, progressively implemented in the ARTEMIS system since 1975, is described for the computerized management of hypertensive patients. From a medical point of view, computerized programs can be used to memorize patients' individual records and profiles, to facilitate patient management and follow-up, to store medical knowledge about hypertension and to provide facilities for decision making at the level either of the individual patient or of the population followed up. From a technical point of view, the methodology used integrates data and knowledge management facilities into the same software. Five hypertension clinics are presently using the system in France and more than 22,000 records have been registered. Answer rates to 12 mandatory questions regarding past history and examination at first visit were superior to 95% in 19,601 records created between January 1976 and December 1987. Patient database interrogation can be used to evaluate the sensitivity and specificity of various signs and symptoms for the diagnosis of secondary hypertension, and to predict, for each patient, his/her cardiovascular risk, the risk of drop-out, the risk of insufficient blood pressure control and the probable blood pressure level. It also serves to test the content and validity of the associated expert system which is progressively built up. A prospective evaluation of the performance of the expert system on 80 cases of hypertension showed overall agreement between the specialists and the expert system ranging from 58 to 91% depending on the decision.(ABSTRACT TRUNCATED AT 250 WORDS)

Six thousand six hundred thirty two subjects, employed in 420 small and medium-sized companies in the Paris region were examined in a cross-sectional study. Their alcohol consumption, as obtained by interview was found to be higher among males than among females, among workers than among managers, executives, and clerks. Alcohol consumption was positively associated with age, body mass index, coffee and cigarette consumption, occupational exposure to noise and working nights or alternating shifts. A positive, continuous, relationship was observed, for men and women, between alcohol intake and both systolic and diastolic blood pressure. This association was highly significant in the multivariate analysis (multiple linear regression) where alcohol intake, following age and body mass index, was the third predictive factor of blood pressure level in the stepwise regression. The positive association between alcohol consumption and prevalence of arterial hypertension was aggravated by the poor control of hypertension which was found among drinkers. Awareness of hypertension, compliance with an antihypertensive treatment and its efficacy, were negatively associated with alcohol intake. The findings stress the importance of alcohol consumption which was found to be a major risk factor for arterial hypertension and noncompliance with antihypertensive treatment in this population.

A computerized system called ARTEMIS has been taking part in the management of a referral hypertension clinic since September 1975. From January 1, 1976 to January 1, 1979, 3,624 patients' records were set up and updated, all of them accessible on line. The mean rate of response to 17 questions with obligatory answers concerning patients' past history was 97.6 per cent. The 10 routine laboratory tests included in the initial in-hospital work-up were registered in over 92 per 100 of the 2,651 hospital records examined. Negative responses were registered with the same accuracy as positive ones, when the computerized questionnaires were appropriately drafted. Standardized collection of information insured homogeneity of the medical staff and has already made possible epidemiologic studies. ARTEMIS has detected strategical errors such as the lack of yearly programmed tests for cardiovascular risk-factors associated with hypertension. The computer has helped to achieve a high rate of patient compliance at one year (81.3 per cent), by providing patients up-dated editions of their personal summary reports as well as sending periodical letters of recall to them and to their physicians. At a total annual cost of 11 dollars per patient, the computer provides continuous information on process and outcome, for each patient and physician, and for the entire population followed up at the clinic. It facilitates physician adherence to predefined protocols and reinforces follow-up continuity by creating a communications network between the referral clinic, the patient and the general practitioner.

The long-term efficacy and tolerance of spironolactone in essential hypertension was evaluated among 20,812 patients referred to the Broussais and St. Joseph systemic hypertension clinics between 1976 and 1985 by using information prospectively collected in the computerized ARTEMIS data bank. In 182 patients (51 men, 131 women) treated with spironolactone alone during a mean follow-up period of 23 months, a mean dose of 96.5 mg decreased systolic and diastolic blood pressure (BP) by 18 and 10 mm Hg, respectively, below pretherapeutic levels. The BP decrease was greater with doses of 75 to 100 mg (12.4% and 12.2%) than with doses of 25 to 50 mg (5.3 and 6.5%, p less than 0.001), but no additional decrease was found with doses above 150 mg. Plasma creatinine level increased modestly (8.3 mumol/liters), as did plasma potassium level (0.6 mmol/liters) (both p less than 0.001); uric acid level increased, but not significantly (10.5 mumol/liter). Fasting blood glucose and total cholesterol levels did not change, triglyceride levels increased slightly (0.1 mmol/liter, p less than 0.05). These changes were similar in both sexes and were not influenced by length of follow-up. Among the 699 men prescribed spironolactone alone or in association with another antihypertensive treatment, 91 cases of gynecomastia developed (13%). Gynecomastia was reversible and dose-related; at doses of 50 mg or less the incidence was 6.9%, but 52.2% for doses of 150 mg or higher. Despite limitations inherent in the interpretation of data banks, it is concluded that spironolactone administered in daily practice reduced BP without inducing adverse metabolic adverse effects and that in patients with essential hypertension, doses should be kept below 100 mg.

Poor compliance with appointments and drug treatment is one of the recognised factors preventing effective management of hypertension. Factors predictive of poor attendance and inadequate blood pressure control in patients attending a hypertension clinic were therefore determined using univariate analyses and a multivariate logistic model. Out of 1346 patients with blood pressure exceeding 160/95 mm Hg followed up for three years, 209 (15.5%) dropped out during the first year. Variables that were significantly related to increased drop out rates were male sex, young age, obesity at entry, cigarette smoking, direct referral to the clinic as a result of screening instead of referral by a general practitioner, absence of pre-existing antihypertensive treatment at the first visit, moderate hypertension, and low socioeconomic category. Variables at entry that were significantly related to poor blood pressure control at one year were old age, evidence of coronary heart disease, severe hypertension, and raised blood glucose concentrations. Early detection of patients at high risk of drop out or poor blood pressure control might improve treatment of hypertension and allow management to be more individually adapted to each patient.

Alcohol consumption and glycosuria were found to be associated (p less than 0.001) in a population of 6571 salaried employees who underwent a systematic examination. The prevalence of glycosuria was found to range from 1.3% among 2609 non-drinkers to 5% among 816 heavy drinkers (six glasses or more of alcoholic beverage daily). This association was still significant after adjustment for age, sex and body mass index. Similarly, a positive association was observed between fasting glycemia and alcoholic intake in a subgroup of 998 subjects when such a result was available (p less than 0.05).

The influence of blood pressure levels on the prevalence of symptoms was studied in a group of 1771 untreated hypertensive patients referred to the Saint-Joseph Hypertension Clinic in Paris. Information on symptoms was obtained from a standardized physician-conducted interview during the patient's first visit at the outpatient clinic. The most frequent symptoms were headache (40.5%), palpitation (28.5%), nocturia (20.4%), dizzines (20.8%) and tinnitus (13.8%). Except for nocturia, symptom prevalence was higher in females than in males. In males as well as in females, no correlation was found between blood pressure levels and the presence of headaches, dizziness, palpitation and tinnitus when results were adjusted for age. In contrast, the relationship between two behavioural characteristics, anxiety and lack of regular physical activity, and symptom prevalence was more pronounced than the relationship with the blood pressure level itself. Moreover, the symptom prevalence reported differed significantly between the five permanent physicians of the clinic; comparison of results obtained by physician conducted interview and self-administered questionnaire indicated that differences between physicians were not due to differences in patient's characteristics but to differences in physician behaviour.

An expert system has been integrated to the data management system of the ARTEMIS programme for hypertensive patients. The patient database, which has been used since 1975, contains the medical records of about 20,000 patients. Information is interactively entered by physicians, nurses and secretaries on video display units. The computerised medical record has replaced the traditional handwritten medical record. The database management system is used to produce different summary reports (inpatient and outpatient care) and personalized recall letters which are mailed to the patients before their appointments. Suggestions provided by the expert system include additional information to be obtained (complementary patient interrogation, biological or radiological investigations, etc.), possible causes of hypertension, and medical prescriptions. The information base allows the description of both static knowledge (in the form of a semantic network) and dynamic knowledge (in the form of production rules). The inference system sequentially uses a combination of forward and backward chaining and performs both exact and approximate reasoning. The diagnostic performance of the expert system was evaluated in 100 cases of hypertension (50 of essential hypertension and 50 of secondary hypertension. Concordance between the diagnosis proposed by the expert system and the one proposed by the specialist was achieved in 92% of secondary hypertension cases and 88% of essential hypertension cases. It is suggested that the integration of data and knowledge management might enhance the overall acceptance by medical staff of a computerised system, and facilitate the validation of a knowledge base.