Introduction

Medical students (henceforth named trainees) are often present when patients visit a doctor or a clinic. Although most patients accept that doctors in training must develop their clinical skills, the presence of a student might add to the arousal that the medical environment already imposes on a patient. This could lead to a substantial increase in blood pressure, as suggested by casual observations in our own general practice.

Numerous publications already addressed the 'white-coat' effect, that is, the transient rise in a patient's blood pressure when measured by an observer in a medical environment.^{1, 2, 3, 4, 5, 6, 7} However, to the best of our knowledge, no published report investigated whether a doctor-in-training when present is a neutral bystander or an obtrusive observer provoking a reactive increase in a patient's blood pressure. The relevance of this research question is underscored by the fact that treatment of hypertension is a frequent reason to consult primary care physicians.² We therefore undertook a randomised clinical trial to evaluate to what extent the presence of a medical student might influence a patient's blood pressure when it is measured by an experienced practitioner.

Methods

Doctors and patients

A total of 22 general practitioners affiliated with the University of Ghent (Belgium) as trainers of medical students, agreed to enrol at least 10 patients visiting for any reason between March 2001 and June 2001 and to re-examine these patients at a subsequent occasion.⁸ Patients were eligible if they were 25 years or older, and if they were registered at the practice for at least 1 year. We excluded patients with an acute medical condition, recent surgery, severe chronic disorders, secondary hypertension, or pregnancy. We also did not consider patients with recent psychological or social problems.

Study protocol

In all, 281 consecutive patients complied with the entry criteria and were randomly assigned to a 'trainee' or 'no trainee' group. Randomisation was balanced within practices, with five patients allocated to each of the 'trainee' and the 'no trainee' groups. To randomise a patient, the practitioner drew a lot from an opaque envelope. Sample size calculation assuming a standard deviation of systolic pressure of 15 mmHg demonstrated that 131 patients per group might be needed to reveal a between-group difference of 6 mmHg on a two-sided test with the -level and power set at 5 and 90% respectively.⁹ The blood pressure of all patients was measured by the experienced doctor-trainer at baseline and at the follow-up visit approximately 2 months later. In the 'trainee' group, the first blood pressure measurement took place in the presence of the trainee.

Blood pressure measurement

Before the study, all investigators were trained to measure blood pressure according to the WHO-NHG guidelines.^{10, 11, 12, 13, 14, 15, 16, 17, 18, 19} After the patients had rested for 5 min in the sitting position, doctors obtained one reading of systolic and diastolic (phase V) blood pressure. The patients refrained from smoking and caffeinated beverages for at least 20 min before the blood pressure was recorded. Within practices, doctors used the same mercury sphygmomanometer for all patients. The sphygmomanometers were used as calibrated for usual practice. No special calibration was carried out for this specific study. Hypertension was defined as a blood pressure of at least 140 mmHg systolic or 90 mmHg diastolic. Hypertensive patients whose treatment changed during follow-up were excluded from the analysis.

Statistical analysis

For database management and statistical analysis, we used the SPSS Inc., version 10. We compared means of normally distributed data by Student's t-test for paired or unpaired observations, as appropriate. Medians were compared by Wilcoxon's signed rank test and proportions by a ² statistic.

We searched for correlates of the blood pressure changes from the first to the second visit. Using stepwise linear regression, we considered as potential explanatory variables: trainee vs no trainee group; gender of patient, doctor and trainee (separate analysis for 'trainee' group); patient characteristics including age, initial blood pressure, smoking status, history of diabetes and cardiovascular diseases, overweight (body mass index 27 kg/m²), and hypercholesterolaemia (serum total cholesterol 240 mg/dl); and the interval between the two visits.

Results

Characteristics of doctors and patients

Of the 281 randomised patients, 142 were in the 'trainee' group and 139 were in the 'no trainee' group (Figure 1). At randomisation, the two groups had similar anthropometric characteristics (Table 1). Of the 22 participating doctors, 18 were male and four were female. On average, they had 20.5 years of practice and had previously trained 17.5 students. Of the 27 doctor-students, 21 were female.

Figure 1.

Flow of patients.

Full figure and legend (16K)

Table 1 - Patient characteristics at randomisation.

Full table

Trainee effect

At the first visit (Figure 2, and Table 2), systolic pressure was 6.4 mmHg (CI 2.0–10.8 mmHg, P=0.004) higher in the 'trainee' than in the control group (139.520.5 (s.d.) mmHg vs 133.115.0 mmHg), with a similar trend (+2.4 mmHg, CI -0.1 to 4.9 mmHg, P=0.07) for diastolic pressure (80.2 12.6 vs 77.88.3 mmHg). Median follow-up was similar in the 'trainee' and 'no trainee' groups (P=0.23), lasting 51.5 days (P25–P75 interval 35.0–61.5 days) and 48.1 (42.5–69.5) days, respectively. At the second visit, systolic and diastolic blood pressures were similar in the 'trainee' and 'no trainee' groups and averaged 134.716.6/ 78.59.8 mmHg and 133.014.7/79.38.3 mmHg, respectively (0.37<P<0.45). From randomisation to follow-up, blood pressure decreased in the 'trainee' group by 4.8 mmHg (CI 2.1–7.4 mmHg, P<0.001) systolic and by 1.7 mmHg (CI 0.10–3.3 mmHg, P=0.03) diastolic, whereas the corresponding changes in the control group were -0.1 mm (CI -1.8 to 2.0 mmHg, P=0.90) and +1.5 mmHg (CI 0.1–2.8 mmHg, P=0.03). Thus, the between-group differences in the blood pressure changes over time, averaging 4.7 mmHg (CI 1.5–7.9 mmHg, P=0.005) systolic and 3.2 mmHg (CI 1.1–5.3 mmHg, P=0.003) diastolic, were statistically significant.

Figure 2.

Systolic and diastolic blood pressure at baseline and follow-up by randomisation groups. Values are meanss.e.

Full figure and legend (19K)

Table 2 - Blood pressure and percentage of hypertensive patients at baseline and follow-up by randomisation groups.

Full table

Percentage of patients with elevated blood pressure

The percentage of patients on antihypertensive treatment was similar in the two randomisation groups (49.6 vs 51.9% in 'trainee' vs 'no-trainee' group, P=0.83). (Table 1). At the first contact, 60.9% patients of the 'trainee' group had an elevated blood pressure (140/90 mmHg), compared with 47.3% at the second contact (P=0.02). In the 'no trainee' group, the percentage of hypertensive patients was similar at both visits (42.6 vs 42.6%, P=0.99) (Table 2). The percentage of hypertensive patients at the first visit was significantly higher in the 'trainee' group compared with the 'no trainee' group (60.9 vs 42.6%, P=0.003), while there was no significant difference at the second visit (47.3 vs 42.6%, P=0.52).

Covariables of trainee effect

Stepwise regression showed that systolic blood pressure at baseline (partial R²=0.35; P<0.0001) was the only significant determinant of the systolic trainee effect. Diastolic blood pressure at baseline (partial R²=0.36; P<0.0001) and randomisation group (partial R²=0.02; P=0.02) significantly contributed to the variability in the diastolic trainee effect.

Discussion

The main finding of our study was that in teaching-practices, the presence of a doctor-in-training had a significant pressor effect when an experienced general practitioner measured a patient's blood pressure. When considering the possible implications of these observations, a number of issues deserve discussion. These include the mechanisms that might explain our findings, and the possible clinical implications for teaching practices.

With regard to the mechanisms that might underlie our findings, anxiety and arousal are known to raise systolic blood pressure, often by as much as 30 mmHg.^{1, 2, 3, 4, 5, 6, 7} This may be regarded as a physiological reaction, often referred to as the 'fight and flight' phenomenon, or 'defence' or 'alarm' reaction. It is commonly observed in emergency departments of hospitals when patients are frightened, but as highlighted by our study, it may also occur in general practice. In our study, the between-group differences in systolic pressure at the initial contact averaged 6.4 mmHg systolic. Furthermore, in the trainee group systolic blood pressure decreased by 4.8 mmHg systolic from the first to the second contact. Presumably, the pressor effect observed in the presence of a doctor-in-training might be interpreted as a white-coat phenomenon, which may add to that already associated with blood pressure measurement by the doctor-trainer.

A treatment effect cannot explain the differences in blood pressure between the first and the second visit because the treatment remained unchanged in all patients and because the proportion of treated patients was similar in the 'trainee' and the 'no trainee' group.

Regression to the mean is unlikely to explain our observations, because patients were randomly assigned to groups and because at randomisation there were no significant differences between both groups, other than the measured blood pressure level. The 'trainee'-effect was greater on systolic than on diastolic blood pressure. The mean age of our patients was 63.5 years. In older patients with stiff arteries and attenuated baroreceptor reflexes, blood pressure raising stimuli are likely to have a greater effect on systolic than on diastolic blood pressure.^{20, 21, 22, 23} However, the present study must be interpreted within the context of its limitations. We cannot exclude that doctor-trainers might have been more accurate in their blood pressure reading in the presence of a student or more inaccurate when the student was not present.

The diagnosis, management and treatment of hypertension are critically dependent on the accurate measurement of blood pressure. If it is overestimated, it follows that incorrect decisions will be made. With over 20% of the adult population suffering from hypertension, and high blood pressure being a frequent reason to consult a primary care physician, the consequence of inaccurate blood pressure measurement carry implications for health care delivery. If confirmed, our findings imply that doctors should be cautious to initiate or adjust antihypertensive drug treatment when blood pressure readings are obtained in the presence of a trainee.

References

1. Mancia G et al. Effects of blood pressure measurement by the doctor on patient's blood pressure and heart rate. Lancet 1983; 24: 695–697. | Article |
2. Pickering TG et al. How common is white-coat hypertension? JAMA 1988; 259: 225–228. | Article | PubMed | ChemPort |
3. Mancia G, Parati G. Reactivity to physical and behavioral stress and blood pressure variability in hypertension. In: Julius S, Bassett DR (eds). Handbook of Hypertension Vol. 9. Behavioral Factors in Hypertension. Elsevier Science: Amsterdam, 1987 pp 104–122.
4. Palatini P et al. The white-coat effect is unrelated to the difference between clinic and daytime blood pressure and is associated with greater reactivity to public speaking. J Hypertens 2003; 21: 545–553. | Article | PubMed |
5. Mancia G, Zanchetti A. Editors' corner: white coat hypertension: misnomers, misconceptions and misunderstandings. What should we do next? J Hypertens 1996; 14: 1049–1052. | PubMed |
6. Parati G, Mancia G. White coat effect: semantics, assessment and pathophysiological implications. J Hypertens 2003; 21: 481–486. | Article | PubMed |
7. Den Hond E et al. Self-measured versus ambulatory blood pressure in the diagnosis of hypertension. J Hypertens 2003; 21: 717–722. | Article | PubMed | ISI |
8. Knottnerus JA, Dinant GJ, Van Schayck OP. The diagnostic before–after study to assess clinical impact and Lijmer JG, Bossuyt PM. Diagnostic testing and prognosis: the RCT in diagnostic research. In: Knottnerus JA (ed). The Evidence Base Of Clinical Diagnosis. BMJ Books: London, 2002 pp 61–94.
9. Cohen J. Statistical Power Analysis for the Behavioural Sciences, 2nd ed. Erlbaum: Hillsdale NJ, 1988.
10. Staessen JA, O'Brien ET, Thijs L, Fagard RH. Modern approaches to blood pressure measurement. Occup Environ Med 2000; 57: 510–520. | Article | PubMed | ISI | ChemPort |
11. McAlister FA, Straus SE. Measurement of blood pressure: an evidence based review. Clinical review. BMJ 2001; 322: 908–911. | Article | PubMed |
12. Perloff D et al. Human blood pressure determination by sphygmomanometry. Circulation 1993; 88: 2460–2467. | PubMed | ISI | ChemPort |
13. Beevers G, Lip GYH, O'Brien E. Blood pressure measurement. Clinical review. Sphygmomanometry: factors common to all techniques. BMJ 2001; 322: 981–985. | Article | PubMed | ChemPort |
14. Beevers G, Lip GYH, O'Brien E. Blood pressure measurement. Clinical review. Conventional sphygmomanometry: technique of auscultatory blood pressure measurement. BMJ 2001; 322: 1043–1047. | Article | PubMed | ChemPort |
15. O'Brien E, Beevers G, Lip GYH. Blood pressure measurement. Clinical review. Automated sphygmomanometry: ambulatory blood pressure measurement. BMJ 2001; 322: 1100–1114.
16. Knottnerus JA, Van Weel C, Muris JW. Evaluation of diagnostic procedures. BMJ 2002; 324: 477–480. | Article | PubMed |
17. NGH. Dutch College of General Practitioners: Hypertension Guideline. Huisarts Wetenschap 1991; 34: 389–395.
18. O'Brien E et al. Task Force I: methodological aspects of blood pressure measurement. Blood Press Monit 2001; 6: 313–315. | Article | PubMed |
19. Staessen JA et al. Sympathetic tone and relation between sodium intake and blood pressure in the general population. BMJ 1989; 299: 1502–1503. | PubMed |
20. Staessen JA et al. Antihypertensive treatment based on conventional or ambulatory blood pressure measurement. A randomized controlled trial. Ambulatory Blood Pressure Monitoring and Treatment of Hypertension Investigators. JAMA 1997; 278: 1065–1072. | Article | PubMed | ISI | ChemPort |
21. Duprez DA, Van Helshoecht P, Van den Eynde W, Leeman M. Prevalence of hypertension in the adult population of Belgium: report of a worksite study; Attention Hypertension. J Hum Hypertens 2002; 16: 47–52. | Article |
22. Duprez DA et al. Relationship between periventricular or deep white matter lesions and arterial elasticity indices in very old people. Age Ageing 2001; 30: 325–330. | Article | PubMed |
23. Staessen JA et al. Randomised double-blind comparison of placebo and active treatment for older patients with isolated systolic hypertension. The Systolic Hypertension in Europe (Syst-Eur) Trial Investigators. Lancet 1997; 350: 757–764. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We thank the following colleagues for collecting the data: Dirk Anseeuw, Hugo Arnout, Anne-Marie Braeckman, Rudi Bruyninckx, Koen Cornelli, Annie Goeman, Renaat Crevits, Jean-Paul De Corte, Geert Deneckere, Rudy Derave, Willy De Schepper, Sabine Gilleman, André Hernalsteen, Hendrik Huybrechts, Gabriël Loncke, Roland Van Acker, Marc Van De Putte, Jan Van Elsen, Jan Verplanken, Lilly Willems, Marc Wybaillie, and Dirk Zanders.