The HealthQuiz[trademark symbol] (Nellcor, Inc., Chula Vista, CA) is a hand-held, patient-driven data processor marketed to expedite and enhance the completeness of the history-taking process during the preanesthetic visit.
In our preanesthesia assessment center (PAC), we compared preanesthetic visits using the HealthQuiz[trademark symbol] device with visits using our standard, written questionnaire.
After approval from our institutional review board (consent document not required), 244 consecutive patients presenting to our PAC for preprocedure evaluation were prospectively and randomly assigned to receive either a written questionnaire (control) or a HealthQuiz[trademark symbol] (experimental). Our PAC serves all electively scheduled, preprocedure day admissions or outpatients presenting to a university-based teaching hospital.
The Standard Visit
Standard-visit patients completed our written questionnaire (Figure 1). The anesthesiologist used the patient's responses as a guide during the interview and examination of the patient.
The HealthQuiz[trademark symbol] Visit
The HealthQuiz[trademark symbol] has a video screen and four response buttons labeled "yes," "not sure," "no," and "next question." The device displays a maximum of 120 questions for female patients and 114 questions for male patients; at least 60 questions are displayed to each patient. The questions are asked in a decision tree format; the response to a "stem" question determines the number and order of the subsequent questions.
The HealthQuiz[trademark symbol] device prints a patient summary report that categorizes medical data in much the same format as one might see on a generic preanesthesia evaluation form. The anesthesiologist used this summary report to guide the interview and examination.
After the preanesthetic visit, each patient completed a questionnaire (Figure 2) regarding satisfaction with the method by which the medical history was taken. For each question, we provided five responses: 1 = strongly agree, 2 = agree, 3 = neutral, 4 = disagree, 5 = strongly disagree. We defined favorable responses as 1 or 2 and unfavorable responses as 3, 4, or 5.
After 50 study patient visits, the interviewing anesthesiologist recorded the visit duration for the next 62 visits. We defined the visit duration as the time from when the anesthesiologist first picked up the patient's chart to when the anesthesiologist completed all documentation.
One board-certified anesthesiologist with 7 yr of postresidency experience (RAB) was the evaluating anesthesiologist. For each patient's documented visit data, the evaluating anesthesiologist used a prospectively defined list of 24 clinically important historical categories (Figure 3) to make a determination as to whether there were sufficient documented historical data. Because different forms were produced after each of the two visit types, the evaluating anesthesiologist was not blinded to the means by which the initial medical history was taken. The evaluating anesthesiologist did not interview or examine any of the study patients.
For all study visits, age and ASA physical status were compared between groups by using the unpaired Student's t-test; the number of interviewing residents and attending physicians were compared by using Fisher's exact test. For the timed visits, the visit duration was compared by using Student's t-test, and the number of interviewing residents and attending physicians were compared by using Fisher's exact test. The percentage of visits lasting more than 20 min was compared using chi squared analysis.
For each question on the patient satisfaction questionnaire, the percentage of favorable and unfavorable responses was compared between groups using chi squared analysis.
For each category of clinically important historical data, the percentage of sufficient and insufficient determinations was compared between groups using chi squared analysis.
In all statistical tests, we defined statistical significance as P < 0.05.
Sixteen patients (nine in the control group and seven in the experimental group) were excluded from study. Eleven did not communicate in English, and five were not able to independently complete the initial questionnaire or the HealthQuiz[trademark symbol]. Two hundred twenty-six patients completed the study.
No difference was found between the control and experimental groups with regard to age, ASA physical status, and number of attending and resident interviewers (Table 1).
For the timed visits, the visit duration, the percentage of visits exceeding 20 min, and the number of interviewing residents and attending physicians were similar between groups (Table 2). Our data had sufficient power to detect a difference between the means of 7 min or greater with a power of 0.8 (P < 0.05).
The percentage of favorable or unfavorable responses to each of the six questions on the patient questionnaire was not significantly different between groups. For each question, our data had sufficient power to detect a between-groups difference of 12% or more with a significance of 0.05 and a power of 0.8.
For 4 of the 24 categories listed on our anesthesiologist's evaluation form, the ratio of sufficient to insufficient responses was significantly different between the groups (Table 3).
Lutner et al.  and Roizen et al.  demonstrated the validity of the medical data obtained using the HealthQuiz[trademark symbol] device compared with data obtained during a personal interview. Despite these investigations, we found that the HealthQuiz[trademark symbol] neither expedited nor improved the completeness of the history-taking process in our PAC.
The phrases by which the HealthQuiz[trademark symbol] described potentially significant findings were often vague and were poorly understood by the interviewing anesthesiologist. Furthermore, the summary report often implicated diseases that were not clinically important.
This study has several limitations. The evaluating anesthesiologist was not blinded to the history-taking method. The historical data categories used to evaluate the documented data were broadly defined, allowing room for the exercise of clinical judgment and the possibility for bias. We also did not control for the interviewing anesthesiologist's experience with the HealthQuiz[trademark symbol] device.
Our findings can not be extrapolated to PACs that see only a select group of electively scheduled surgical patients, use a substantially different written questionnaire, or do not have residents.
After completing this study, we discontinued our use of the HealthQuiz[trademark symbol] and focused on improving the content and clarity of our written questionnaire.
We wish to note that the HealthQuiz[trademark symbol] device was withdrawn from the market approximately two years ago when Puritan Bennett bought Nellcor, Inc.
The authors thank Patrick Wells for his assistance with the data analysis, Pat Boutis for her assistance with the manuscript preparation, and Dorothy Zeller for her assistance with chart review.