Enrollment, Retention, and Demographics
Out of 394 patients in the practice panel, we enrolled a total of 107 participants (27%), 54 in the intervention group and 53 in the control group. We capped enrollment from the heart failure practice in December 2001, when we reached a point of maximal recruitment from the waiting room of the practice. Unfortunately, this point was reached before we were able to achieve our target enrollment.
illustrates the flow of participants through the study. The pool of eligible patients was derived from the practice census at the beginning of the study and subsequent records of patients who had appointments during the enrollment period. Two interested patients were excluded because they were health professionals (one physician's assistant and one nurse practitioner). Approximately 10 patients were not approached for enrollment because they did not speak English. Of the patients enrolled, 78.5% remained at 6 months and 76% remained at 12 months.
Flow of Participants Through The Study
After recruitment was completed, we identified a pool of 288 patients who were cared for by the practice during the recruitment period but did not enroll in the primary study. Of these, 144 (50%) completed the "decliners" survey.
summarizes the characteristics of the intervention, control, and decliner groups. At baseline, the intervention and control groups did not differ in their socioeconomic characteristics, or in their health status as assessed by the KCCQ symptom score. Although we did not use the New York Heart Association (NYHA) heart failure classification, our "participants" (the combination of the intervention and the control groups) had a mean KCCQ symptom score of 65, similar to the mean score for patients with NYHA Class II symptoms in the KCCQ validation study [10
Baseline Demographic Characteristics *
Decliners did not differ from participants in their age and gender distribution, their health status, or their self-efficacy. Compared with participants, however, decliners had lower incomes, and a lower percentage were white or of non-Hispanic race/ethnicity. Furthermore, fewer had standard medical insurance, or a college education. Although they were less likely than participants to have experience with the Internet, roughly half of the decliners nonetheless did have access to a computer and experience with the Internet.
Use of SPPARO/Electronic Messaging
The number of patients using SPPARO and the number of patient hit days are presented in . Use of SPPARO was highest in the first 3 months after enrollment, then leveled off. After the first 3 months, an average of 24% of the enrolled patients used SPPARO in a given month. During this time interval, frequency of use of SPPARO averaged 0.4 hit-days per enrolled patient per month. This was approximately 1 hit-day per clinic visit.
Monthly Use of SPARRO Over Study Period
Solid line indicates hit days per 10 intervention patients per month. Dashed line indicates the percentage of the intervention patients that used SPPARO (System Providing Patients Access to Records Online) each month. Monthly website activity is normalized to account for attrition over the course of the study.
Cumulative use of SPPARO over the study period is shown in . Subjects most commonly reviewed clinical notes and laboratory results, and did so repeatedly. Fewer subjects reviewed radiology results, but those who did also reviewed them repeatedly. The educational guide was reviewed least frequently, and was generally reviewed only once.
Cumulative Use of SPARRO Over the 12-Month Study Period
The electronic messaging function in SPPARO appeared to supplement, rather than replace, telephone messages. The intervention group sent more messages to the practice (350 total: 287 phone calls and 63 electronic messages) than the control group (267 phone calls) over the course of the study. The number of total messages (phone + electronic messages in the intervention group, phone messages in the control group) sent per month are compared graphically in . The number of total messages sent per month did not show a statistically significant difference (p=0.70). The number of messages sent per patient did demonstrate a statistically significant difference when analyzed by square root transformation (p=0.02). The difference was more pronounced during the first 6 months of the intervention (150 messages in the intervention group vs. 88 in the control group, p=0.05) than the second 6 months (109 messages vs. 103, p=0.66). The main categories of messages overall were to schedule appointments (20% of total messages), to refill medications (15%), to ask questions about medications (14%), to get test results (12%), to report feeling ill (8%), and to get assistance interpreting test results (3%). In none of the individual categories was there a statistically significant difference in call volume between the intervention and the control group.
Nurses spent a total of 304 minutes answering computer messages over the course of the 12 months, a mean of 5.6 minutes per subject per year. In interviews, the physicians and nursing staff did not feel that providing SPPARO to the intervention group resulted in a perceptible change in their workload.
Self-Efficacy, Health Status, Adherence, and Patient Satisfaction
Repeated measures of self-efficacy, health status, adherence, and patient satisfaction are presented in . For our primary outcome, the self-efficacy domain of the KCCQ, there was a trend towards an improvement in the intervention group, but the improvement of 6 points did not reach the threshold value of 7.7 that we had set as a standard for this outcome. (Based on actual enrollment, the study had a power of 73% to detect a difference of 7.7, and 80% power to detect a difference of 8.8, with a two-sided alpha of 0.05). For the other measures of health status from the KCCQ, there were trends towards improvement in a many domains, but no statistically significant improvements were demonstrated when adjusted for multiple comparisons.
Changes In Adherence, Health Status, And Patient Satisfaction Over Time
General adherence to medical advice showed significant improvement in the intervention group compared with the control group. Adherence to medications showed a similar trend but did not reach statistical significance.
Patient satisfaction with doctor patient-communication demonstrated a trend towards improvement in two areas: how well patients felt their problems were understood, and how well doctors explained information. While significant results were found for these two items individually, the findings did not reach statistical significance when adjusted for multiple comparisons. There was no significant improvement in the other patient satisfaction domains.
Mortality and Utilization of Health Services
compares mortality, hospitalizations, Emergency Department visits, and practice visits for the intervention and control groups. Although the number of patients who visited the emergency department did not differ significantly, there was a significant increase in the number of overall emergency department visits in the intervention group (20 visits) relative to the controls (eight visits). Of the emergency department visits in the intervention group, only four occurred within 7 days of using SPPARO.
Mortality, Hospitalizations, Emergency Department Visits, and Clinic Visits During Study Year 2002
Proportions of patients in the two groups are compared using Chi-squared and Fishers' Exact Test. The number of utilizations in the two groups is compared by comparing the number of utilizations per patient using the Wilcoxon rank-sum test.
There were no reports of adverse effects resulting from use of SPPARO. In only one case did access to SPPARO result in a patient complaint. That patient did not agree with a statement regarding his alcohol consumption. He requested that an amendment be placed in the clinical notes, and his concerns were documented.