The study was reviewed and approved by the University of Wisconsin Health Sciences institutional review board (protocol H-2007-0291) and all participants signed informed consent documentation.

We recruited adults (aged >18 years) from a variety of clinical and community settings on a rolling basis between February and April 2009. These adults had a diagnosis of asthma and used a prescribed inhaled short-acting beta-agonist (SABA). By the end of August 2009, all participants had participated for 4 months. At entry, participants completed questionnaires to collect data on their demographic and clinical characteristics, including asthma history, then-current status, and asthma management such as medical treatments, perceived triggers, healthcare utilization, and level of control using the Asthma Control Test^TM (ACT) [8].

We collected data about the time and location of use of inhaled short-acting bronchodilators. This involved using a small and portable investigational electronic medication sensor that is attached to an existing inhaler housing by means of a hook-and-loop fastener strap (Figure 1). This allows the sensor to be readily transferred to a new inhaler when the medication is refilled. When the sensor was attached to an inhaler, an actuation detection assembly placed on the end of the medication canister monitored the use of the inhaler and used a global positioning system (GPS) receiver to determine the time and location of the inhaler whenever it was used. Design verification and systematic bench and human testing demonstrated that the sensor accurately and reliably detected actuation of the inhaler and that expected functioning was not sensitive to a precise method of attachment. Participants were asked to keep the sensor attached to their inhaled short acting bronchodilator for all 4 months of the study.

The device also contained equipment to communicate via a wireless network and supporting electronics, including a simple user interface, a power switch, and a rechargeable battery. The device obtained and wirelessly communicated medication event information in real-time to a remote server. It did not require that the patient (or any other person) participate in obtaining or transmitting the information. Because the routine dose of some SABAs can vary from one to multiple inhalations, the device was not designed to record each inhalation of medication but rather to capture information about the occasion of medication use, following American Thoracic Society/European Respiratory Society guidance [7].

After completing the first month of the study, participants began to receive weekly e-mail reports displaying maps and charts of their usage. They also received access to an online interface with similar features. This online interface allowed participants to indicate when an inhaler was used prophylactically before exercise. The weekly e-mail reports included the time and location for each inhaled bronchodilator use and summarized use for the preceding week. Participants received a basic assessment of their asthma control based on the number of days of use in the preceding week, the timing of those events, and on thresholds from the National Asthma Education and Prevention Program (NAEPP) guidelines. Participants also received simple advice derived from the NAEPP guidelines about how to improve asthma self-management (Appendix S1) [7].

At the conclusion of each month, and on conclusion of the study in August 2009, participants completed questionnaires designed to detect any changes in the four primary study outcome measures: the Asthma Control Test^TM score and 2-week histories of activity limitation and days and nights with asthma symptoms. A score ≤19 (of 25) on the Asthma Control Test^TM was considered to indicate uncontrolled asthma. We collected 2-week histories of asthma symptoms and activity limitation to capture continuous estimates of outcome measures in addition to the ordinal responses gathered by the Asthma Control Test^TM. Exit questionnaires also assessed participant feedback that included the utility of the electronic communications, perceived changes in asthma awareness, and modifications in day-to-day asthma management.

We used pairwise t-tests to evaluate the null hypothesis that there was no change in the four primary study endpoints (Asthma Control Test^TM score; two-week histories of activity limitation and days and nights with asthma symptoms) while participants were blinded to data collected by the medication sensor in the first month of participation. Linear mixed effects models with a random effect for subject assessed changes in primary asthma endpoints across the three subsequent months of enrollment (at the completion of the first month, the second month, and the third month, i.e., upon exit) when participants were able to access their usage data online. Participants with one or more observations were included in the analysis.

All analyses were performed with SAS 9.2 (SAS Institute, Cary, NC). Regression diagnostics were performed using the MIXED_DX macro [9].

A total of 33 individuals were initially recruited for the study and 29 provided complete information on asthma control at entry. Table 1 presents the demographic characteristics of the participants, who averaged 36.8 years of age (range: 19–74 years); 52% of respondents were female. Four (14%) reported Hispanic ethnicity. A broad range of socioeconomic backgrounds was represented, including household income and educational achievement. Persons who completed the study did not differ significantly from those who did not. In particular, there was no significant difference in the proportion of males (95% CI: −19.8%, 36.4%), in age (p = 0.80), or in years of diagnosis (p = 0.95) in the group that enrolled compared to the group that completed the study.

On average, participants had longstanding asthma, with a mean time since diagnosis of 25 years (standard deviation [SD]  = 16). Approximately 62% (18/29) of study participants entered with uncontrolled asthma; mean ACT score at study entry for uncontrolled participants was 15.4 (SD = 2.11). Mean ACT score overall was 17.6 (SD = 3.35).

At study entry, participants reported 4.84 days (SD = 4.13) with asthma symptoms, 2.03 (SD = 3.35) nights with asthma symptoms, and 0.86 (SD = 1.38) days with activity limitations in the 2 weeks before entry (Table 2). More than half (62%; 18/29) of participants reported current use of anti-inflammatory therapy (e.g., an inhaled corticosteroid). There was no significant relation between report of an inhaled corticosteroid and ACT score at study entry (Fisher's exact test: p = 1.00). Approximately one-third of the participants reported receiving ≥1 course of oral steroids in the past year, and nearly 40% reported an unscheduled doctor visit in the past year (Table 1).

Participants were enrolled for a total of 3,887 patient-days of observation between March and August 2009. Of these days, devices were turned on and available on 1,916 days –49% of patient-days. On the other days, devices were powered off or out of range of a wireless network.

On average, participants used the devices on 27% (520/1,916) of patient-days in which the devices were active, and 13% (520/3,887) of patient-days overall. During the study period, the devices reported a total of 958 medication-use events, occurring on 520 patient-days. The median number of events per day was 1.35, with a mode of 1. The percentage of days on which individual participants used bronchodilators ranged from 1% to 83%. Overall, participants who entered the study with controlled asthma (11/29) as defined by the Asthma Control Test^TM (score >19) used bronchodilators on 25% of the days their devices were active, compared with 37% among those with uncontrolled asthma.

A total of 30 individuals provided information for at least one of the periodic (monthly) surveys. Results from the pairwise t-test (Table 3) indicate that from study entry to the end of the first month, while participants were blinded to their data on inhaler usage, no significant differences appeared in the ACT score (p = 0.66). In addition, no significant differences appeared in the 2-week history of daytime symptoms, nighttime symptoms, or activity limitations from the entry period to the first-month survey.

After the first month that participants received weekly email reports, mean ACT scores increased 1.5 points to 20.1 (SD = 3.66). After a second month of reports (at exit), mean ACT scores increased an additional 1.1 point to 21.2 (SD = 3.36)(Table 2). At exit, 70% of participants had improved ACT scores, 15% had no change, and 15% had worsened. Scores for the worsened patients (n = 3) declined an average of 0.9 points. Overall, at the conclusion of the study period, 75% of participants had controlled asthma (by ACT score) compared with 38% at entry.

Table 4 shows results from the repeated-measures analysis. Each month of participation following the availability of inhaler usage data through the online interface and weekly email reports resulted in a 1.40 point (95% CI: 0.61, 2.18) increase in the Asthma Control Test^TM score. Receipt of weekly e-mail reports and online access to asthma event history was also associated with a significant decrease in number of days with symptoms (β = −1.35, 95% CI: −2.65, −0.04) and number of nights with symptoms (β = −0.84, 95% CI: −1.25, −0.44). We did not observe a significant change in the frequency of activity limitation reported by participants. Regression diagnostics indicated that linearity assumptions were adequately met for the primary analyses.

On exit surveys, participants reported that the weekly email reports increased their awareness of the frequency and patterns of asthma symptoms. Participants suggested that the feedback helped improve perceptions of their asthma control and their ability to communicate to their physicians their level of control. As one participant described, “I learned that I used my inhaler more than I remember. I was able to see and relate to my doctor that my asthma is not under control.” Participants also reported that the receipt of information about the time and location where they used their inhaler helped to highlight locations and exposures to triggers that led to symptoms. “I've been more keen to note surroundings when I feel shortness of breath,” one participant said. “It opened my eyes to triggers I wasn't aware of in the past.” Some described subsequent attempts to minimize or avoid exposure. Finally, participants reported that weekly feedback helped to reinforce their use of controller medications. “I'm now using my controller medicine more regularly and on time,” one participant told us. “I noticed that my rescue inhaler use went down significantly when using a daily inhaler as well.”

Participants reported that remote monitoring and feedback proved a practical aid. All but one of the participants indicated that the weekly e-mail reports were either very or somewhat useful in improving their asthma management; overall, 80% expressed an interest in continuing to track their inhaler use.