Connected drug delivery devices to complement drug treatments: potential to facilitate disease management in home setting

Figures & data

Table 1 Characteristics of the selected smart devices and health apps (taken from respective product webpages)

Device	Company	Target population/indication	Description/purpose of the device	Mechanism of action	Additional features	Associated apps	Associated software
RebiSmart^® [www.RebiSmart.com]	Merck KGaA	MS	Individually adjustable electronic injection device to administer a pre-set dose of Rebif (interferon β-1a) to reduce the frequency of relapses in people with relapsing forms of MS	Administers Rebif subcutaneously, collects and stores data, and sends the information wirelessly to the MSdialog server	Allows patients to self-inject; records an accurate and objective dosing history, including the date, time, and dosage of every injection; hidden needle, multidose cartridge and adjustable comfort settings; sends injection reminders	MS Dialog App	MSdialog
BETACONNECT™ [www.betaseron.com]	Bayer AG	MS	Autoinjector that delivers Betaseron (interferon β-1b) to reduce the frequency of relapses in people with relapsing forms of MS	Delivers Betaseron, and pairs with myBETAapp via Bluetooth^® to deliver real-time data (such as injection date, time, speed, and depth) with the BETA nurse and HCPs	Sends injection reminders; adjustable injection depth and speed; hidden needle; smooth, quiet injection; automatic needle retraction; visual and audible end-of-dose indicator; simple to unload	myBETAapp	BETACONNECT™ navigator
Proteus Discover [www.proteus.com]	Proteus Digital Health, Inc	Multiple indications	An ingestible pill that provides objective confirmation of adherence to oral medications. Approved with aripiprazole in schizophrenia, bipolar disorders, and depressiona	Proteus Discover is activated by taking medication with an ingestible sensor. The sensor transmits a signal from the stomach to a patch worn on the torso. A digital record is sent to the patient’s mobile device and then to the Proteus cloud where it can be accessed by HCPs	Battery-powered adhesive sensor; the patch on the torso also records patient’s heart rate, temperature, activity, and rest patterns	Discover	Discover Portal
SmartInhaler™ [www.smartinhaler.	Adherium North America, Inc	Asthma/COPD	Medication sensors that clip around a standard inhaler and automatically send usage data to a mobile app or PC via Bluetooth^®	Contains sensors that attach to existing inhalers and record when medication is taken. These sensors automatically send usage data to a mobile app or PC via Bluetooth^®	Tracks medication use; sends medication reminders	SmartInhaler™ app	SmartInhalerlive™
Propeller [www.propellerhealth.	Propeller Health	Asthma/COPD	Attaches to the top of the canister of a standard inhaler. Records the location, date, and time of inhaler actuations	Records use of rescue inhaler to provide a summary of potential triggers. Connects with health care providers to share data	Reminders to take medication provided; provides useful geospatial information of asthma attacks to patients; provides personalized daily asthma forecasts	Propeller mobile app	Propeller web app

Note:

a Indicated in adults for the treatment of schizophrenia, bipolar 1 disorder, and as adjunctive treatment for major depressive disorder.

Abbreviations: HCP, health care professional; MS, multiple sclerosis; PC, personal computer.

Table 2 Study details and results from all identified studies of RebiSmart^® (adherence, clinical outcomes, and health care resources use)a

Study name	Study objective	Study design	Indication	Sample size	Follow-up period	Key results
Bayas et al (2015)^Citation18	To evaluate adherence to and effectiveness of treatment in patients with RMS using RebiSmart for self-injection of SC interferon β-1a	Multicenter, prospective, observational	RMS	912	12 months or until ED	• Age in the safety population (mean ± SD) was 63.3±10.3 years • Mean cumulative adherence was 97.1±7.3% (range 30%–100%; n=791) • Adherence rates (mean ± SD) were similar in treatment-naive patients (97.3±6.5%; range 42%–100%; n=453) and patients starting interferon β-1a within 6 weeks before study entry (96.7±8.8%; range 30%–100%; n=200) • Mean adherence per patient since the last visit at 1 year or ED was 96.5±9.4% (range 20%–100%; n=759) • At month 12/ED, 72.4% (560/774) of highly adherent patients and 41.2% (7/17) of patients with a lower adherence were relapse-free • A significantly higher proportion of relapse-free compliant high-dose patients (>120 µg/week) compared with lower dose patients in clean patient subset (92.5% vs 86.7%; P=0.0146) at month 6 with nonsignificant trend remaining at week 12/ED (82.1% vs 76.6%) • AAR 0.3±0.7 in highly adherent patients (n=774) and 0.6±1.3 in less adherent patients (n=16)
Devonshire et al (2016)^Citation21	To evaluate 24-week treatment adherence of RMS patients using RebiSmart^® for SC injection of interferon β-1a	Multicenter, single-arm, observational	RMS	162	96 weeks (24-week data were available in this publication)	• Patient age (mean ± SD) was 37.4±9.8 years • The proportion of patients with ≥80% adherence was 91.8% (95% CI 86.3, 95.2) at week 12; 82.9% (95% CI 76.2, 88.0) at week 24 • Similar treatment adherence in patients with and without anxiety at baseline (96.1% vs 95.2%, P=0.322 at week 12; 89.0% vs 89.0%, P=0.901 at week 24)
Fernandez et al (2016)^Citation19	To determine long-term adherence to SC interferon β-1a treatment administered with the RebiSmart^®	Multicenter, retrospective, observational	RRMS	258	Until device replacement (36 months maximum lifetime) treatment discontinuation	• Patient age (mean ± SD) was 40.7±9.5 years • Overall adherence was 92.6% (95% CI 90.6, 94.5) • 32% of patients (n=78) achieved an adherence rate of 100%, and 80.6% of patients (n=208) achieved an adherence rate of ≥90%; 13.2% of patients (n=34) showed an adherence of <80% • Over the study period, a slight decrease in adherence with a mean overall adherence of 94.0% (95% CI 92.0, 96.0) at 0–3 months and 90.4% (95% CI 87.4, 93.3) at the time of device replacement • The incidence of relapses decreased from 5.8% (n=258 at 0–3 months) to 4.0% (n=150 at 33–36 months) • Having experienced relapses from the beginning of treatment was significantly related to achieving an adherence of ≥80% (OR =3.06, 1.28–7.31) • Suboptimal adherence (adherence of ≥80%) was about three-times higher in subjects with relapses compared with those without relapses
Ghezzi et al (2017)^Citation30	To evaluate the changes in quality of life of adolescents with RRMS receiving treatment with interferon β-1a administered subcutaneously using RebiSmart^® (FUTURE study)	Multicenter, single-arm, observational, prospective study	RRMS	50	52 weeks	• Adolescents (age 12–16 years) who completed the study (n=40) showed 12-, 24-, and 52-week adherence rates of 82.5%, 80.0%, and 67.5%, respectively • At end of treatment, 32 (80%) of patients were relapse-free • Self-reported quality of life:  ○ PedsQL4.0 self-reported Total Scale score and all subscale scores tended to increase (nonsignificantly) from baseline to end of treatment; except for Emotional Functioning score, which showed a nonsignificant decline  ○ A mean raw change over time of +0.44 points was found for PedsQL4.0 Total Scale score, with the highest improvement (+1.41) identified for the School Functioning scale • Parent-reported QoL:  ○ PedsQL4.0 Total Scale score increased significantly from baseline to end of treatment (+5.27 points, P=0.041)  ○ Significant increases also emerged in the parent-reported Psychosocial Health summary score (+5.90 points; P=0.015) and the School Functioning scale score (+7.84 points; P=0.029) • There was no difference in either adolescent self-reported or parent-reported PedsQL4.0 scores in relation to age (12–15 vs 16–8 years)
Järvinen et al (2017)^Citation76	To investigate adherence measured by an electronic autoinjector device vs self-reported adherence and treatment convenience in subjects with RRMS	Multicenter, observational	RRMS	31	24 weeks	• Patient age (mean ± SD) was 39.0±8.0 years • Mean adherence was 93.5% (95% CI 88.4, 98.2; device data); 96.6% (95% CI 94.3, 99.0; self-assessment) • Age groups 30–40 and >40 years had a mean adherence of 62% and 64%, respectively; the age group <30 years had a slightly higher mean adherence of 75% • Device found to ease patient’s treatment and perceived to be easy to use (>65% of patients)
Lugaresi et al (2012)^Citation28	To assess short-term adherence to, and tolerability of, interferon β-1a administered via electronic autoinjection device in patients with RRMS (BRIDGE study)	Multicenter, single-arm, observational	RRMS	119	12 weeks or early termination	• Patient age (mean ± SD) was 37.9±9.7 years • 88.2% of patients administered ≥80% of scheduled injections over 12 weeks • Significant decrease in adherence (P=0.001) from 100% through week 4 to 99.1% in weeks 5–8, to 89.0% in weeks 9–12 • Mean HADS depression (P=0.303) or anxiety (P=0.156) scores did not differ significantly from baseline • Mean (± SD) baseline PASAT score was similar in adherent and nonadherent patients (42.67±10.9 vs 42.86±13.0)
Lugaresi et al (2016)^Citation29	To investigate long-term adherence to interferon β-1a administered using RebiSmart^® among patients in a real-life setting (RIVER study; extension of BRIDGE study)	Multicenter, retrospective	RRMS	57	20.5 months	• Patient age (mean ± SD) was 38.0±9.0 years • Overall adherence to RebiSmart^® was 79.8% (median, 85.2%, range, 16%–100%) • There was no statistically significant difference in age (mean ± SD) of adherent and nonadherent patients in the entire study cohort (37.6±10.0 vs 38.1±8.6 years; P=0.845) • EDSS scores at baseline and last follow-up did not differ significantly between adherent and nonadherent patients (date not shown in publication)
Moccia et al (2015)^Citation22	To investigate predictors of adherence to interferon β-1a	Retrospective analysis of prospectively collected data	RRMS	114	1.5 years	• Patient age (mean ± SD) was 35.8±10.4 years • Adherence was 95.0±9.0% • Early missing (14.9%) was more likely to be associated with clinical relapse (OR =4.155; P=0.018), but not late missing (47.4%) (OR =1.454; P=0.408) vs fully adherent (37.7%) • Adherence was lower in early missing compared with late missing or fully adherent (P<0.001) • In different regression models, categories of time to first missing were not associated with age (P=0.865)
Paolicelli et al (2016)^Citation20	To collect data on treatment adherence in a real-life setting, in order to identify predictors of adherence at baseline (TRACER study)	Multicenter, retrospective	RRMS	384	12 months	• Patient age (mean ± SD) was 36.0±9.2 years • 89.3% of patients were adherent • Adherence decreased from 93.2% in patients aged 26–40 years (93.2%) to 87.5% in patients aged ≥41 years to 79% in patients ≤25 years (P=0.006) • 90.5% of patients with a baseline EDSS <4 showed ≥80% adherence (vs 71.4% in patients with EDSS score ≥4; P=0.016)
Pedersen et al (2018)^Citation26	To evaluate patient adherence to treatment with SC interferon β-1a using RebiSmart^® and assess injection-site reactions and treatment satisfaction	Multicenter, prospective, single-arm, observational	RRMS	60b	12 weeks	• Patient age (mean ± SD) was 43.7±7.9 years • 89% (n=48) of patients had ≥90% adherence to treatment • 94% (n=51) of patients had ≥75% adherence (95% exact CI 85, 99); only three (6%) patients had adherence <75% and could be defined as nonadherent according to the study protocol • Most patients (78%) rated convenience as the most important aspect of the device
Rau et al (2017)^Citation27	To investigate adherence pattern and cognitive– behavioral variables using electronic autoinjector RebiSmart^®	Multicenter, single-arm, prospective study	RRMS	188	24 months	• Patient age varied from 18 to 65 years • 12-month data • Of 129 patients with 1-year adherence data, quantitative and qualitative adherence were 96.3% and 88.9%, respectively • 82% reached qualitative adherence of ≥80% • FSMC motor score increased by 9.4% (P<0.002) and cognitive score increased by 10.0%, respectively (P<0.035) • 24-month data • Of 62 patients with 2-year adherence data, quantitative and qualitative adherence were 93.4% and 84.6%, respectively • 74% reached qualitative adherence of ≥80% • Cognitive fatigue increased by 8.4% (P<0.038)
Solsona et al (2017)^Citation23	To describe adherence to interferon β-1a using RebiSmart^® and to explore the relationship between adherence and relapses in a Spanish cohort	Retrospective, observational	MS	110	979 days (mean duration of treatment)	• Patient age (mean ± SD) was 38.8±9.3 years • Mean adherence was 96.5% (IQR 91.1–99.1) • Mean adherence was 98.7% (IQR 91.3–100) during the first 6 months and 97.6% (IQR 91.1–99.8) during the last 6 months • No statistically significant differences in adherence were detected regarding age (Spearman’s rank correlation coefficient rs =0.183, P=0.055) • Increased adherence was associated with better clinical outcomes, leading to lower relapse risk (OR =0.953; 95% CI 0.912, 0.995); each percentage unit increase in adherence decreased the likelihood of relapse by 4.7%
Willis et al (2014)^Citation24	To assess adherence to SC interferon β-1a injections using data from RebiSmart^®	Single-group, observational, retrospective	RMS	225	24 months	• Patient age (mean ± SD) was 44.1±8.82 years • Mean adherence over 24 months was 95.0% (95% CI 93.6, 96.4) • Proportion of patients with ≥80% adherence was 91.1% (95% CI 86.6, 94.5) at 24 months • No significant differences in percentage adherence between age categories were seen at 12 months (P=0.099) or 24 months (P=0.126)
Zecca et al (2017)^Citation25	To evaluate the relationship between subjectively reported and objective adherence of MS patients using RebiSmart^® in Switzerland and explore variables associated with objective adherence	Multicenter, survey-based, retrospective, prospective	RRMS	56	9 months	• Patient median (IQR) age was 49.0 (38.0–55.0) years • Median objective adherence was significantly higher in self-reported adherent (100%; IQR 98.8%–100%) than in self-reported nonadherent patients (93.4%; IQR 77.2%– 97.5%) (P=0.00001) • No difference between retrospective (98.8%; IQR 93%–100%) and prospective (98.8%; IQR 88.5%–100%) objective adherence (P=0.75) • Older age was significantly associated with higher objective adherence (OR=1.064; 95% CI 1.016, 1.114; P=0.008) • Median (IQR) age as per adherence rates were: 41.0 (31.5–48.0) in patients with low adherence rates; 48.5 (36.5–54.8) in patients with medium adherence and 53.5 (42.0–63.0) in patients with high adherence rates

Notes:

a Key results listed focus on the impact of the device on adherence to treatment and, if available, a correlation between adherence and efficacy outcomes;

b 60 patients were recruited. However, adherence data were obtained in 54 patients only due to technical problems with six devices.

Abbreviations: ED, early discontinuation; EDSS, Expanded Disability Status Scale; FSMC, Fatigue Scale for Motor and Cognitive Functions; HADS, Hospital Anxiety and Depression Scale; IFN, interferon; IQR, interquartile range; OR, odds ratio; PASAT, Paced Auditory Serial Addition Task; PedsQL, Pediatric Quality of Life Inventory; QoL, quality of life; RMS, relapsing multiple sclerosis; RRMS, relapsing-remitting multiple sclerosis; SC, subcutaneous.

Table 3 Study details and results from all identified studies of BETACONNECT™/myBETAapp (adherence, clinical outcomes and health care resources use)

Study name	Study objective	Study design	Indication	Sample size	Follow-up period	Key results
Kleiter et al (2017)^Citation32	To investigate adherence to therapy, satisfaction, and functional health status among patients treated with interferon β-1b using the BETACONNECT™ autoinjector	Prospective, observational	RRMS	151	24 weeks	• Patient age (mean ± SD) was 41.2±11.5 years • Adherence of ≥80% declined from 72.0% (week 4) to 67.3% (week 12) and 57.9% (week 24) for patients with a least one data readout and from 81.1% at week 4 to 86.7% at week 12 to 80.5% at week 24 for patients at the respective visit • Compliance was 86.3% at week 4, 91.9% at week 12, and 92.9% at week 24 • Age tended to be a predictor of persistence at 24 weeks with patients ≥40 years being more likely to still use the BETACONNECT™ at follow-up visits (OR =1.047, 95% CI 1.003, 1.093) • Treatment-naive patients showed overall higher persistence (OR =12.246; 95% CI 2.191, 68.457)
Patti et al (2017)^Citation33	To measure adherence in patients with RRMS or CIS who were treated with interferon β-1b using BETACONNECT™	Multicenter, prospective, single-arm, observational	RRMS and CIS	498 (474 RRMS and 26 CIS)	24 weeks	• Patient age (median) was 44 years • Adherence (defined as completing ≥80% of prescribed injections) was assessed at baseline, week 4, week 12, and week 24. Median adherence remained stable between 93.9% and 95.4% at all visits • Higher adherence was associated with male gender, existing concomitant disease, shorter disease duration, higher SDMT score and higher satisfaction with the myBETApp
Rametta et al (2017)^Citation34	To assess adherence to interferon β-1b therapy in patients who are using the BETACONNECT™ autoinjector	Multicenter, prospective, single-arm, observational	RRMS	89	6 weeks	• Patient age (mean) was 52 years; majority (77.5%) of patients aged ≥45 years • Mean adherence rate was 97.6% (SD 9.0%) • 95.5% of patients reached ≥80% adherence

Abbreviations: CIS, clinically isolated syndrome; OR, odds ratio; RRMS, relapsing-remitting multiple sclerosis; SDMT, Symbol Digit Modalities Test.

Table 4 Study details and results from all identified studies of Proteus Discover (ingestible sensors) (adherence, clinical outcomes, and health care resources use)

Study name	Study objective	Study design	Indication	Sample size	Follow-up period	Key results
Au-Yeung and DiCarlo (2012)^Citation43	To compare costs of direct confirmation of treatment using WOT vs SOC utilizing WHO-recommended 7-day and 3-day DOT	Prospective (economic model)	Tuberculosis	NA	NA	• Health facility cost-to-treat with WOT was 36% of 7-day DOT and 71% of 3-day DOT • Patient cost-to-be-treated with WOT was 4% of 7-day DOT and 8% of 3-day DOT • Public health worker time/subject treated using WOT was 34% of 7-day DOT and 67% of 3-day DOT • Total cost for 7-day DOT was US$3,472 (86% personnel cost; 13.5% retreatment cost) • Total cost for 3-day DOT was US$1,772 (73% personnel cost; 27% retreatment cost) • Total cost for WOT was US$1,273 (38% personnel cost; 61% retreatment cost)
Bonacini et al (2017)^Citation41	To evaluate real-world adherence in chronically infected HCV patients treated with SOF/LDV and using the Digital Medicine Offering	Prospective, single-arm	HCV	28	12 weeks	• Patient age (mean) was 59 years • Patients used the Digital Medicine Offering for 92% of expected days; mean ingestion adherence was 94% • Sustained virologic response at 12+ weeks was achieved in 20 of 22 patients (data for six patients were not available)
DiCarlo et al (2014)^Citation38	To evaluate the utility of the Proteus^® system in patients with uncontrolled hypertension	Prospective	Hypertension	190	14 days	• Patient age was not reported • In patients with complete data (89%), mean medication adherence was 88% • Mean SBP decrease was −7.6 mmHg; mean DBP decrease was −3.8 mmHg • 78% patients had >70% adherence; 53% achieved BP control on prescribed therapy; 25% remained uncontrolled and required treatment modification
Frias et al (2017)^Citation36	To assess the impact of digital medicine offerings on BP and glycemic and lipid control	RCT	Hypertension and T2DM	109	4 and 12 weeks	• Patient age (mean ± SE) was 57.8±1.1 years in the Proteus group and 61.6±1.7 years in the usual care group • Medication adherence was ≥80% (calculated for Proteus only) • Hypertension  ○ At week 4, the combined Proteus groups had a mean change in SBP from baseline of –21.8 mmHg (SE 1.5) compared with –12.7 mmHg (SE 2.8) for usual care (mean −9.1, SE 2.9, 95% CI 14.8, 3.3 mmHg)  ○ A greater proportion of participants in the Proteus group achieved their BP goal (81%, 65/80) compared with usual care (33.3%, 9/27) (95% CI 18.5, 77.3)  ○ At week 12, 98% (39/40) of Proteus participants achieved their BP goal compared with 51.7% of usual care participants (95% CI 7.1, 84.5) • Diabetes  ○ There was a nonsignificant difference in HbA1c reduction in favor of the Proteus group, compared with usual care (4-week Proteus group mean =−0.32% SE 0.22; 12-week Proteus group mean =−0.08% SE 0.22%; usual care mean =0.28% SE 0.35%)  ○ For participants with a baseline HbA1c of ≥8%, the Proteus group experienced a larger HbA1c decrease compared with an increase in HbA1c seen in the usual care group (difference from 4-week Proteus group −0.98%, 95% CI –1.72,–0.24; difference from 12-week Proteus group −0.57%, 95% CI –1.53, 0.39)  ○ The differences in change in LDL-C between the Proteus groups and the usual care group were –33.2 mg/dL (95% CI –50.6,–15.8) at week 4 and –19.2 mg/dL (95% CI –36.4, –2.0) at week 12
Godbehere and Wareing (2014)^Citation39	To assess the impact of digital medicine offerings on BP	Prospective	Hypertension	8	2 weeks	• Patient age varied from 49 to 62 years • Taking adherence ranged from 70% to 100% • Timing adherence ranged from 67% to 100% • BP decreased in all patients (−7/+6 mmHg to −54/–24 mmHg)
Kim et al (2014)^Citation44	To assess the impact of a digital health feedback system in uncontrolled hypertensive patients	Prospective study (economic model)	Hypertension	164	1 year	• Patient age was not reported • The system was estimated to result in $7.3–$18.3 million in savings ($328–$717 per BP at goal), and to lead to a 3%–9% reduction in the number of CAD and stroke events in 1 year
Kim et al (2015)^Citation45	To assess the impact of digital medicines with a mobile app in patients with comorbid hypertension, diabetes, hypercholesterolemia	Prospective study (economic model)	Hypertension, T2DM, and hypercholesterolemia	NR	1 year	• Estimated cost offsets of the Proteus service offering were $90–$185/month of use (including reimbursements) • Estimated medical cost savings (reductions in outpatient/inpatient services, monitoring, disease management, medication costs) were $850–$980 (5%–11% reduction in diabetes and CVD complications) • Revenue opportunities presented an additional value equating to $80–$95 PPPY, bringing total value of the Proteus offering to $1,020–$1,260 PPPY
Moorhead et al (2017)^Citation37	To study the incremental impact of seeing vs not seeing Proteus medication dose reminders on medication taking; and to assess the safety of the Proteus medication dose reminders for the possible risk of overdosing	Cluster RCT	Hypertension and type 2 diabetes mellitus	57	12 weeks	• Patient age (mean ± SD) was 58.0±10.5 years • Proteus device reminder messages were associated with a 16%±16% increase (75%±18% when seeing vs 59%±24% when not seeing mobile dose reminders) in medication taking (when not taken before dose reminder) • The average daily adherence was 85.9%±11.7% • 79% of subjects (45 of 57) achieved >80% adherence • There were no overdose events related to Proteus medication dose reminders
Naik et al (2017)^Citation40	To assess the usability and acceptability of passive electronic monitoring for managing hypertension	Prospective cohort	Hypertension	167	2 weeks	• Patient age (mean ± SD) was 68.0±9.0 years • Taking adherence ranged from 53% to 100%, and timing adherence ranged from 21% to 100% • SBP decreased from 154±13 to 145±18 mmHg (P<0.001) and DBP decreased from 85±11 to 80±12 mmHg (P<0.001) • 32% of registry participants were found to be capable of achieving BP control using their chronically prescribed medications
Sullivan et al (2017)^Citation42	To monitor patterns of adherence using ingestible digital transmitter and to assess transplant outcomes	Prospective, single-arm, observational	Liver and renal transplant recipients	30	3 months	• Patient age (mean ± SD) was 14.0±3.6 years • The average patch wear adherence was 78% and average digitized medication adherence was 89.3% • While there was no change in emergency department visit rates in the study period, there was a reduction in hospital admissions
Virdi et al (2016)^Citation77	To assess the impact of a digital health feedback system in hypertensive patients	Cluster RCT	Hypertension	103	4 weeks	• Patient age (mean) was 58.4 years • Proportion of patients that achieved the BP target was 83.3% in the Proteus group and 33.3% in the usual care group (95% CI 23.62, 76.38) • Average adherence measured by feedback system was 84% (80%–89%)

Abbreviations: BP, blood pressure; CAD, coronary artery disease; CVD, cardiovascular disease; DBP, diastolic blood pressure; DOT, directly observed therapy; HCV, hepatitis C virus; HbA1c, glycosylated hemoglobin A1C; LDL-C, low-density lipoprotein cholesterol; LDV, ledipasvir; NA, not applicable; NR, not reported; PPPY, per patient per year; RCT, randomized controlled trial; SBP, systolic blood pressure; SE, standard error; SOC, standard of care; SOF, sofosbuvir; T2DM, type 2 diabetes mellitus; WHO, World Health Organization; WOT, wirelessly observed therapy.

Table 5 Study details and results from all identified studies of SmartInhaler™ (adherence, clinical outcomes, and health care resources use)

Study name	Study objective	Study design	Indication	Sample size	Follow-up period	Key results
Britto et al (2017)^Citation52	To evaluate if the text messaging intervention (using Smart Inhaler™) would result in improved adherence to prescribed ICS	RCT	Asthma	64	6 months	• Patient age (mean ± SD) was 17.2±1.9 years in the intervention group and 17.4±2.7 years in the control group • Receiving the text intervention resulted in an increase in adherence of 2.75% each month relative to no intervention (P<0.01); improvements were not sustained • There was modest improvement in asthma control and quality of life outcomes in the text messages group compared with control group (scores at different timepoints mentioned in figures and not reported in text in the publication) • The adolescent participants gave high ratings on the acceptability of the text messaging system
Burgess et al (2010)a,^Citation49	To assess the impact of measuring adherence and providing feedback on medication usage by children with unstable asthma	RCT	Asthma	26	4 months	• Patient age varied between 6 and 14 years. Mean age of patients in the intervention group was 9.1 years and in the control group was 9.3 years • Adherence was significantly higher in the intervention group than in the control group (79% vs 58%, P<0.01) • Adherence in the control group declined slightly over the study; mean adherence in the intervention group increased from 76.8% during the third month to 84.2% over the final month (P<0.01) • There was a significant improvement in asthma control in both the intervention and the control groups: at baseline, 15 of 26 subjects reported using their reliever medication three or more times/week over the previous month. During the 4-month study, two subjects reported requiring reliever medication on average three or more times/week (P=0.02) • At baseline, mean FEV1 across all subjects was 75% predicted; mean FEV1 of all participants during the study increased to 85.2% • The change in FEV1 from baseline (% predicted) was greater in those subjects receiving feedback (13.8% vs 9.8%; P<0.9) • The change in FEV1 in the final month (% predicted) was similar in both groups (87.3% vs 86.9%; P<0.4)
Chan et al (2015)a,^Citation48	To investigate whether use of an inhaler with audiovisual reminders improves adherence and asthma outcomes in children presenting with asthma exacerbation	RCT	Asthma	220	6 months	• Patient age (mean ± SD) was 8.9±2.5 years in the intervention group and 8.9±2.6 in the control group • Median percentage adherence was 84% in the intervention group, compared with 30% in the control group (P<0.0001) • Intervention group had an improved asthma morbidity score (P=0.008) and asthma control scores at 2, 4, and 6 months (P<0.0001), and fewer exacerbations at 2 months (6% vs 24%, P=0.015) • Mean percentage adherence at 2, 4, and 6 months was 91%, 84%, and 79%, respectively, for the intervention group, compared with 40%, 33%, and 27%, respectively, for the control group • The change in asthma morbidity score from baseline to 6 months was significantly greater in the intervention group than in the control group (P=0.008). There was a reduction of 2.0 points from a mean baseline score of 9.3 (SD 2.2) to 7.3 (2.1) in the intervention group vs 1.2 points from a baseline of 9.2 (2.5) to 8.0 (2.2) in the control group • For childhood asthma control test scores, the difference between groups was significant at all timepoints (2, 4, and 6 months; P<0.0001), with the intervention group scoring higher by an overall average of 1.57 • The median percentage number of days on which a reliever was used was 9.5% in the intervention group and 17.4% in the control group (P=0.002) • No differences in FEV1 (P=0.38), asthma-related school absences (P=0.096), emergency visits (P=0.509), or caregiver work absences (P=0.167) were reported
Charles et al (2007)a,^Citation50	To determine whether an audiovisual reminder device improves adherence with ICS therapy in adult asthma patients	RCT	Asthma	110	6 months	• Median (range) patient age was 39 (13–65) years in the intervention group and 35 (15–64) years in the control group • In the last 12 weeks, the proportion of medication taken was 93% in the AVRF group and 74% in the control group, with a difference of 18% (95% CI 10, 26; P<0.0001) • The proportion of subjects taking >50%, >80%, or >90% of medication was greater in the AVRF group. Corresponding ratios of proportions adherent were 1.33 (95% CI 1.10, 1.61; P=0.003), 2.27 (95% CI 1.56, 3.3; P<0.0001), and 3.25 (95% CI 1.74, 6.1%; P<0.0001), respectively
Foster et al (2014)a,^Citation51	To test the effectiveness of two brief GP-delivered interventions (with and without IRF) for improving adherence and asthma control	Cluster RCT	Asthma	143	6 months	• Patient age (mean ± SD) was 40.3±15.2 years • Adherence was significantly higher in the IRF group vs non-IRF groups (73%±26% vs 46±28%; P<0.0001), but not significantly different between PAD and non-PAD • The asthma control improved overall (mean change in ACT score, 4.5±4.9; P<0.0001), with no significant difference between IRF and non-IRF groups • Severe exacerbations experienced by 11% patients in IRF groups and 28% in non-IRF groups (P=0.013; after adjustment for exacerbation history; P=0.06) • There were statistically significant and clinically important improvements from T0 to T6 months in asthma-related QoL (mean change, 0.77±1.15; P<0.0001) and significant improvements in anxiety (−0.93±3.18; P=0.022) and self-reported adherence behavior (MARS-A, 0.28±0.82; P=0.008); no significant difference between IRF and non-IRF groups
Jochmann et al (2015)^Citation78	To compare self-assessment of adherence in children with difficult asthma with adherence measured electronically	Prospective, single-arm	Asthma	50	16 weeks	• Median (range) patient age was 12.4 (5–17) years • Median SmartInhaler adherence was 60% (range 24%–97%) • FENO improved significantly in those with SmartInhaler adherence >80% • FEV1 improved significantly following monitoring (86 at baseline to 91.5 at follow-up; P=0.01) • mPAQLQ also improved significantly following monitoring (5.1 at baseline to 5.8 at follow-up; P=0.02)
Kenyon et al (2016)^Citation53	To assess feasibility and acceptability of a health worker-delivered electronic adherence monitoring intervention among the highest utilizers of acute asthma care	Prospective, single-arm	Asthma	14	3 months	• Median (range) patient age was 3.5 (3–9) years • All caregivers viewed the electronic monitoring device favorably and would recommend it to friends; 56% believed the device helped to improve asthma control • Overall, ACT scores improved by a mean of 2.7 points (95% CI 0, 5.5; P=0.05) over the 3-month intervention (just below the minimally significant ACT change threshold of 3 points)
Morton et al (2017)a,^Citation47	To determine whether electronic monitoring results in improved clinical outcomes in asthma patients (STAAR study)	Open-label, parallel group, RCT	Asthma	90	12 months	• Patient age (mean ± SD) was 10.4±2.9 years in the intervention group and 10.2±2.9 in the control group • Adherence in the intervention group was 70% vs 49% in the control group (P ≤ 0.001) • No significant difference in change in ACQ, but intervention group required significantly fewer oral steroids courses and fewer hospital admissions compared with control group  ○ Event rate (per 100 child days) for courses of oral steroids were 0.411 for the intervention arm compared with 0.676 for the control arm (P=0.008)  ○ Event rate (per 100 child days) for hospital admissions was 0.0254 for the intervention arm compared with 0.129 for the control arm (P<0.001) • FEV1% improved in both treatment arms, with no significant difference between treatments at 12-months compared with baseline • Differences in GP visits in the two groups were not clinically significant

Note:

a Studies submitted to NICE.

Abbreviations: ACQ, Asthma Controlled Questionnaire; ACT, Asthma Control Test; AVRF, audiovisual reminder function; FENO, fractional exhaled nitric oxide; GP, general practitioner; ICS, inhaled corticosteroids; IRF, inhaler reminders and feedback; MARS-A, Medication Adherence Report Scale for Asthma; NICE, National Institute of Health & Clinical Excellence; PAD, personalized adherence discussion; PAQLQ, Pediatric Asthma Quality of Life Questionnaire; QoL, quality of life; RCT, randomized controlled trial; STAAR, STudy of Asthma Adherence Reminders.

Table 6 Study details and results from all identified studies of Propeller (adherence, clinical outcomes, and health care resources use)

Study name	Study objective	Study design	Indication	Sample size	Follow-up period	Key results
Barrett et al (2018)^Citation63	To evaluate the extent to which digital health intervention improves asthma outcomes in a real-world setting	Prospective, single-arm	Asthma	497	12 months	• Patient mean age was 38 years (range 4–90 years) and 80% were adults • At 12 months of follow-up, there was a 78% reduction in SABA use, an 84% reduction in night-time SABA use, and a 48% improvement in symptom-free days from baseline to month 12 (P<0.0001) • The number of symptom-free days increased from 62% (first week) to 83% (2 months), to 88% (6 months), and to 90% (12 months) (P<0.0001)
Carl et al (2018)^Citation64	To determine whether a quality improvement program employing a digital health platform could improve pediatric asthma outcomes	Prospective, single-arm	Asthma	82	12 months	• Patient age varied from 4 to 18 years • Aggregate average controller adherence at 1, 2, 3, and 6 months post-enrollment timepoints demonstrated rates of 58%, 64%, 62%, and 55%, respectively • Average weekly rescue event rate at these time points demonstrated decreased events from baseline (0.597/week) to 0.21 (63%), 0.20 (69%), 0.25 (59%), and (75%), respectively
Chen et al (2017)^Citation62	To assess the feasibility and clinical impact of a digital health intervention in a Medicare population with COPD or asthma	Prospective, single-arm	Asthma or COPD	236 (198 COPD, 38 asthma)	6 months	• Among asthma patients: • 65% of patients were 60 years or older • At 6 months, SABA use decreased from 1.15 (first week) to 0.56 uses/person/day (last week); significant decrease of 51.4% (P<0.05) • Increase in symptom-free days was 28% (46% at first week to 59% at last week) • Among COPD patients: • 78% of patients were 60 years or older • At 6 months, SABA use decreased from 1.53 (first week) to 0.74 uses/person/day (last week); significant decrease of 51.7% (P<0.01)
Hoch et al (2017)^Citation65	To evaluate the feasibility of Propeller Health monitoring device, along with clinical outcomes	Prospective, single-arm	Asthma	25	3 months	• Patient age varied from 6 to 17 years • Average weekly adherence ranged from 76% (week 1) to 36% (week 12). The overall average weekly adherence for the cohort was 56% (SD 12%) • Significant decrease in the rate of rescue medication; 76% (week 1) of study vs 40% (week 12) (P=0.012) • Average weekly number of rescue medications declined from 9.9±14 (week 1) to 2.4±5.2 (week 12) • There was no difference in the rate of night-time events between week 1 and week 12 of the study (P=0.727) • Average number of asthma-free days increased from 4.8 days (±2.2) to 6 days (±1.6) • FEV1 and FVC increased during the study period; however, the difference did not reach significance. FEF25–75 increased from an average of 80% predicted to 84% predicted (P=0.02)
Kenyon et al (2018)^Citation66	To assess the feasibility of a mobile health, ICS adherence reminder intervention and to characterize adherence trajectories immediately following severe asthma exacerbation in high-risk urban children with persistent asthma	RCT	Asthma	41	30 days	• Patient age (mean ± SD) was 5.9±2.1 years • Electronic monitoring of ICS use and adherence reminders delivered via text message were feasible for most participants, but there was no signal of effect • After adjusting for age and parental education, mean adherence rates were 36% for the intervention group and 32% for the control group (P=0.73) • Age had no significant impact on adherence rate. After adjusting for impact of intervention and education level, ICS adherence rates in <5 years old were 6% (95% CI 7, 44) compared with 43% (95% CI 30, 55) in those aged ≥5 years (P=0.14) • Mean daily medication adherence trends over the 30-day intervention interval were also similar between the two groups, with broadly overlapping SDs • Mean change in parent-reported portion of the cACT score over the 30 days of the intervention was not statistically significantly different between controls (3.1) and intervention (1.2) (P=0.16)
Merchant et al (2016)^Citation55	To measure real-world effectiveness of Propeller Health Asthma Platform to reduce use of SABA and improve asthma control	RCT	Asthma	495	12 months	• Patient mean age was 36.6 years in the intervention group and 36.0 years in the routine care group. Approximately 30% of patients in both the groups were between 5 and 17 years of age. • Daily mean number of SABA uses/person decreased by 0.41 for intervention group and by 0.31 for routine care between first week and remainder of the study (P<0.001) • Proportion of SABA-free days increased 21% (intervention) and 17% (control) (P<0.01) • ACT scores were not significantly different between arms; initially uncontrolled adults: • Significantly larger improvement in ACT scores in the intervention group vs routine care (+6.2 and +4.6, respectively, P<0.01) • Significantly larger improvement in the proportion with controlled asthma in intervention group vs routine care (63% controlled in the study period vs 49%, respectively; P<0.05)
Merchant et al (2017)^Citation68	To assess the impact of digital health intervention, which leveraged sensors and app-based education, on ER visits, hospitalizations, inpatient days, and clinic visits	Retrospective, cohort	Asthma	507	Data collected between July 2011 and September 2016	• Patient age was not reported • Significant reductions in hospitalizations (2.7 to 0.6 days; 79% reduction; P<0.0001), inpatient days (7.9 to 1.4 days; 82% reduction; P<0.001) and ER visits (19.2 to 8.3 days; 57% reduction; P<0.0001) • Non-acute asthma-related clinic visits increased (197–277 days; 41% increase; P<0.0001)
Su et al (2016)^Citation67	To identify hotspots of asthma symptoms; evaluate associations between asthma symptoms and environmental covariates in real-time and space	Prospective, single-arm	Asthma	140	20 months	• Patient age was not reported • By targeting environmental interventions that could have the largest impact on asthma within specific neighborhoods, 914,000 inhaler uses and $1.8 million of hospitalization costs could be avoided
Van Sickle et al (2010)^Citation58	To investigate if online feedback about remotely monitored inhaled bronchodilators improves composite measures of asthma control	Prospective, single-arm	Asthma	27	NR	• Patient mean age was 35.5 years (range: 19–74 years) • Asthma control scores increased significantly after receiving email reports after first (P=0.01) and second month (P=0.007) • Patients reported more awareness of symptom frequency, level of control, asthma triggers, and increased adherence to preventive medication
Van Sickle et al (2011)^Citation59	To investigate if weekly feedback summarizing use of remotely monitored rescue medication improves asthma control	Prospective, single-arm	Asthma	34	4 months	• Patient mean age was 41.8 years (range: 20–82 years) • Mean ACT scores increased from 17.5 at entry to 19.5 at exit (P=0.01) • Days with asthma symptoms in preceding 2 weeks declined from 6 at entry to 2.8 at exit (P=0.001) • Nights with asthma symptoms in preceding 2 weeks declined from 2.7 at entry to 1.3 at exit (P=0.08) • The ACT scores of 64% participants improved, 11% had no change and 23% worsened
Van Sickle et al (2013)^Citation60	To assess whether weekly email reports on monitored use of inhaled, short-acting bronchodilators improves composite asthma control measures	Prospective, single-arm	Asthma	30	4 months	• Patient mean age was 36.8 years (range: 19–74 years) • No significant difference in ACT scores between entry and first month values (P=0.66) • ACT scores increased by 1.40 points (95% CI 0.61, 2.18) for each subsequent study month after patients received feedback
Van Sickle et al (2014)^Citation61	To determine the effect of sensor-enabled, mobile health asthma program on individual-level asthma outcomes, including frequency of asthma rescue medication use, asthma-free days and control status	Prospective, single-arm	Asthma	NR	12 months	• Patient age was not reported • Improved adherence to clinical guidelines reported; 57% of participants had asthma action plan at study end vs 41% at intake • The proportion considered well-controlled increased by 33% between intake and completion • Proportion with an asthma-free day in the first month was significantly different from all subsequent months (P<0.01)
Van Sickle et al (2015)^Citation57	To determine whether mobile health asthma program could improve asthma outcomes, including frequency of rescue medication use, asthma-free days, and control	Prospective, single-arm	Asthma	299	12 months	• Study participants included both children and adults. Mean/median age was not reported • 57% of participants reported having an asthma action plan at study end, compared with only 41% at intake • At study exit, rescue inhaler use declined by 75%, and the proportion of participants with an asthma-free day increased significantly by 39%
Van Sickle et al (2016)^Citation56	To determine whether a sensor-enabled, clinically integrated mobile health asthma quality improvement program could reduce the frequency of SABA use, and increase asthma-free days, asthma control and controller medication adherence	RCT	Asthma	125	6 months	• Study participants included only adults. Mean/median age was not reported • Significant improvements with intervention vs control on all clinical outcomes, including controller medication adherence, daily SABA use, asthma-free days, and asthma control (all P<0.001) • 21-point improvement in adherence for the intervention

Abbreviations: ACT, Asthma Control Test; cACT, Childhood Asthma Control Test; ER, emergency room; FEF25–75, forced expiratory flow at 25%–75% of the pulmonary volume; ICS, inhaled corticosteroids; NR, not reported; RCT, randomized controlled trial; SABA, short-acting β-agonist.

Figure 1 Expected value of connected devices for stakeholders and potential barriers to adoption.

Abbreviations: PRO, patient-reported outcome; RWD, real-world data.

Table S1 Predefined research questions

Research questions
• Is there any regulatory/HTA guidance for connected devices in the US and EU, including apps or software in particular?
• What was the approval pathway for any connected devices that have been reviewed by regulatory authorities?
• Which clinical studies or other studies were required by regulatory and HTA bodies in order to grant approval of the selected connected devices?
• What is the impact of selected connected devices on adherence (clinical studies or RWE)?
• What is the impact of selected connected devices on clinical outcomes?
• What is the impact of selected connected devices on health care costs and resources, including doctors’ visits (scheduled and unscheduled visits)?
• What are the potential barriers that need to be overcome to unfold the full potential of these tools?

Abbreviations: EU, European Union; HTA, Health technology assessment; RWE, real-world evidence; US, United States.

Table S2 Data sources searched

Data sources searched
PubMed
Google
Google Scholar
Company websites
HTA databases (AETSA; AHRQ; CADTH; CTAF-ICER; HAS; INAHTA; IQWiG; NHS; NICE; PBAC; SMC)
USFDA website
EMA website
Clinicaltrials.gov
International Clinical Trials Registry Platform

Abbreviations: AETSA, Agencia de Evaluación de Tecnologías Sanitarias de Andalucía; AHRQ, The Agency for Health care Research and Quality; CADTH, Canadian Agency for Drugs and Technologies in Health; CTAF-ICER, The California Technology Assessment Forum – Institute for Clinical and Economic Review; EMA, European Medicines Agency; HAS, Haute Autorité de Santé; HTA, Health technology assessment; INAHTA, The International Network of Agencies for Health Technology Assessment; IQWiG, The Institute for Quality and Efficiency in Health care; NHS, National Health Service; NICE, National Institute for Health and Clinical Excellence; PBAC, Pharmaceutical Benefits Advisory Committee; SMC, Scottish Medicines Consortium; USFDA, United States Food and Drug Administration.

Table S3 Search terms used

Device	Synonym/local language translation
Proteus Discover	Proteus pill, Proteus patch, and other translated terms of the device in different languages (eg, Proteus-Pille [German], Proteus pillola [Italian], pilule Proteus [French], píldora de Proteus [Spanish])
RebiSmart®	Msdialog and other translated terms of the device in different languages (same as that of English)
BETACONNECT™	BETACONNECT™ system, BETACONNECT™, and other translated search terms of the device in different languages (eg, Beta verbinden [German], Beta connettersi [Italian]; while French and Spanish terms remained same as that of English language)
SmartInhaler™	Smart inhaler, SmartInhaler medication sensors, SmartInhaler sensors, Smartturbo, Smarttouch and other translated terms of the device in different languages (eg, Intelligenter Inhalator [German], Inalatore intelligente [Italian], Inhalateur intelligent [French], Inhalador inteligente [Spanish])
Propeller	Propeller Sensor, specific device terms including Ellipta, Respimat, Diskus, MDI, metered-dose inhaler, and other translated terms of the device in different languages (eg, Elica [Italian], Hélice [French & Spanish]); while the German term remained same as that of English language

Note: General searches were also undertaken to identify guidance documents for connected devices overall.

Abbreviation: MDI, metered-dose inhaler.

BayasAOualletJCKallmannBAdherence to, and effectiveness of, subcutaneous interferon β-1a administered by RebiSmart® in patients with relapsing multiple sclerosis: results of the 1-year, observational smart studyExpert Opin Drug Deliv20151281239125026098143

 PubMed Web of Science ®Google Scholar

DevonshireVAFeinsteinAMoriartyPAdherence to interferon β-1a therapy using an electronic self-injector in multiple sclerosis: a multicentre, single-arm, observational, phase IV studyBMC Res Notes2016914826951043

 PubMedGoogle Scholar

FernándezOArroyoRMartínez-YélamosSLong-term adherence to IFN beta-1a treatment when using RebiSmart® device in patients with relapsing-remitting multiple sclerosisPLoS One2016118e016031327526201

 PubMed Web of Science ®Google Scholar

GhezziABianchiABaronciniDA multicenter, observational, prospective study of self- and parent-reported quality of life in adolescent multiple sclerosis patients self-administering interferon-β1a using RebiSmart™-the FUTURE studyNeurol Sci201738111999200528831635

 PubMed Web of Science ®Google Scholar

JärvinenEMultanenJAtulaSSubcutaneous interferon β-1a administration by electronic auto-injector is associated with high adherence in patients with relapsing remitting multiple sclerosis in a real-life studyNeurol Int201791695728286627

 PubMed Web of Science ®Google Scholar

LugaresiAFlorioCBrescia-MorraVPatient adherence to and tolerability of self-administered interferon β-1a using an electronic autoinjection device: a multicentre, open-label, phase IV studyBMC Neurol201212722390218

 PubMed Web of Science ®Google Scholar

LugaresiADe RobertisFClericoMLong-term adherence of patients with relapsing-remitting multiple sclerosis to subcutaneous self-injections of interferon β-1a using an electronic device: the RIVER studyExpert Opin Drug Deliv201613793193526909646

 PubMed Web of Science ®Google Scholar

MocciaMPalladinoRRussoCHow many injections did you miss last month? A simple question to predict interferon beta-1a adherence in multiple sclerosisExpert Opin Drug Deliv201512121829183526371561

 PubMed Web of Science ®Google Scholar

PaolicelliDCoccoEDi LecceVExploratory analysis of predictors of patient adherence to subcutaneous interferon beta-1a in multiple sclerosis: tracer studyExpert Opin Drug Deliv20161361726558898

 PubMed Web of Science ®Google Scholar

PedersenEDStenagerEVadgaardJLAdherence to subcutaneous interferon beta-1a treatment using an electronic injection device: a prospective open-label Scandinavian noninterventional study (the ScanSmart study)Patient Prefer Adherence20181256957529720872

 PubMed Web of Science ®Google Scholar

RauDRoßnagelFGösslingJAdherence, cognition and behavioral performance in relapsing-remitting MS RRMS patients using the electronic autoinjector RebiSmart: 1 and 2-year follow-up from the German multicenter RebiSmart studyMult Scler J2017233 Suppl889890

 Google Scholar

SolsonaMDEBoquetEMEstruchBCImpact of adherence on subcutaneous interferon beta-1a effectiveness administered by Rebismart® in patients with multiple sclerosisPatient Prefer Adherence20171141542128280313

 PubMed Web of Science ®Google Scholar

WillisHWebsterJLarkinAMParkesLAn observational, retrospective, UK and Ireland audit of patient adherence to subcutaneous interferon beta-1a injections using the RebiSmart(®) injection devicePatient Prefer Adherence2014884385124966669

 PubMed Web of Science ®Google Scholar

ZeccaCDisantoGMühlSGobbiCSubjective patient-reported versus objective adherence to subcutaneous interferon β-1a in multiple sclerosis using RebiSmart®: the CORE studyBMC Neurol201717117128870152

 PubMed Web of Science ®Google Scholar

KleiterILangMJeskeJAdherence, satisfaction and functional health status among patients with multiple sclerosis using the BETACONNECT™ autoinjector: a prospective observational cohort studyBMC Neurol201717117428877664

 PubMed Web of Science ®Google Scholar

PattiFMartínez GinésMLNorenbergCDuarte CaronFBetaEval Global: Prospective, non-interventional, multinational, observational cohort study of patients using BETACONNECTMult Scler J2017233 Suppl954

 Google Scholar

RamettaMGerminoRMcleodKImpact of Betaconnect auto-injector on patient adherence of betaseron therapyAnnual meeting of the Consortium of Multiple Sclerosis Centers2017 May 24-27New Orleans, LA, USA

 Google Scholar

Au-YeungKYDiCarloLCost comparison of wirelessly vs directly observed therapy for adherence confirmation in anti-tuberculosis treatmentInt J Tuberc Lung Dis201216111498150423006834

 PubMed Web of Science ®Google Scholar

BonaciniMYaKPitneyCA novel digital medicine offering to assess and optimize adherence for oral hepatitis C treatmentHepatol201766Suppl 1394A

 Google Scholar

DicarloLKimYAYoungJBezhadiYReal-time assessment of medication taking and activities of daily living in patients with uncontrolled hypertensionValue Health2014177A475

 Web of Science ®Google Scholar

FriasJVirdiNRajaPKimYSavageGOsterbergLEffectiveness of digital medicines to improve clinical outcomes in patients with uncontrolled hypertension and type 2 diabetes: prospective, open-label, cluster-randomized pilot clinical trialJ Med Internet Res2017197e24628698169

 PubMed Web of Science ®Google Scholar

GodbeherePWareingPHypertension assessment and management: role for digital medicineJ Clin Hypertens2014163235

 Web of Science ®Google Scholar

KimYAVirdiNRajaPDicarloLModeling the impact of a digital health feedback system in uncontrolled hypertensive patientsValue Health2014177A479

 Web of Science ®Google Scholar

KimYARajaPDicarloLVirdiNParkHAn economic model of the impact of digital medicines with a mobile application in patients with comorbid hypertension, diabetes, and hypercholesterolemiaValue Health2015183A40

 Web of Science ®Google Scholar

MoorheadPZavalaAKimYVirdiNSEfficacy and safety of a medication dose reminder feature in a digital health offering with the use of sensor-enabled medicinesJ Am Pharm Assoc2017572155161

 Google Scholar

NaikRMaceyNWestRJFirst use of an ingestible sensor to manage uncontrolled blood pressure in primary practice: the UK hypertension registryJ Community Med Health Educ20177115

 Google Scholar

SullivanSCutrightLMcConnellLUse of digitized medications to enhance adherence and transplant outcomesPediatr Transplant201721Supplement 133

 Google Scholar

VirdiNSKimYRajaPSavageGOsterbergLOptimizing treatment in patients with uncontrolled hypertension and type 2 diabetes by using a digital health offeringJ Am Soc Hypertens2016104e72e73

 Google Scholar

BrittoMTRohanJMDoddsCMByczkowskiTLA randomized trial of user-controlled text messaging to improve asthma outcomes: a pilot studyClin Pediatr2017561413361344

 Web of Science ®Google Scholar

BurgessSWSlyPDDevadasonSGProviding feedback on adherence increases use of preventive medication by asthmatic childrenJ Asthma201047219820120170329

 PubMed Web of Science ®Google Scholar

ChanAHYStewartAWHarrisonJCamargoCABlackPNMitchellEAThe effect of an electronic monitoring device with audiovisual reminder function on adherence to inhaled corticosteroids and school attendance in children with asthma: a randomised controlled trialLancet Respir Med20153321021925617215

 PubMed Web of Science ®Google Scholar

CharlesTQuinnDWeatherallMAldingtonSBeasleyRHoltSAn audiovisual reminder function improves adherence with inhaled corticosteroid therapy in asthmaJ Allergy Clin Immunol2007119481181617320942

 PubMed Web of Science ®Google Scholar

FosterJMUsherwoodTSmithLInhaler reminders improve adherence with controller treatment in primary care patients with asthmaJ Allergy Clin Immunol201413461260126825062783

 PubMed Web of Science ®Google Scholar

JochmannANagakumarPHallPImprovement in asthma control and airway inflammation during a period of electronic monitoringEur Respir Soc201546OA4775

 Web of Science ®Google Scholar

KenyonCCChangJWynterSAFowlerJCLongJBryant-StephensTCElectronic adherence monitoring in a high-utilizing pediatric asthma cohort: a feasibility studyJMIR Res Protoc201652e13227335355

 PubMedGoogle Scholar

MortonRWElphickHERigbyASSTAAR: A randomised controlled trial of electronic adherence monitoring with reminder alarms and feedback to improve clinical outcomes for children with asthmaThorax201772434735427815524

 PubMed Web of Science ®Google Scholar

BarrettMCombsVSuJGAIR Louisville: Addressing asthma with technology, crowdsourcing, cross-sector collaboration, and policyHealth Aff2018374525534

 PubMed Web of Science ®Google Scholar

CarlJCIsakovDWinnersSUse of remote electronic monitoring improves asthma outcomes through improved adherenceAm J Resp Crit Care Med2018197A1164

 Web of Science ®Google Scholar

ChenJSJones-FordGSTuffliMBarrettMvan SickleDFeasibility and clinical impact of deploying a digital health intervention in a Medicare population with COPDAm J Respir Crit Care Med2017195A1722

 Web of Science ®Google Scholar

HochHMeierMRBrintonJTSzeflerSJA feasibility study of daily monitoring of controller and rescue medication use in a pediatric patient population at high risk of an asthma exacerbationAm J Resp Crit Care Med2017195A198

 Web of Science ®Google Scholar

KenyonCCGruschowSMQuarshieWOController adherence following hospital discharge in high risk children: a pilot randomized trial of text message remindersJ Asthma20181319

 Google Scholar

MerchantRKInamdarRQuadeRCEffectiveness of population health management using the propeller health asthma platform: a randomized clinical trialJ Allergy Clin Immunol Pract20164345546326778246

 PubMed Web of Science ®Google Scholar

MerchantRTuffliMBarrettMImpact of a digital health intervention on asthma healthcare utilizationEur Respir J201750PA4692

 Web of Science ®Google Scholar

SuJBarrettMHendersonKApplication of wireless inhaler sensors to enhance asthma surveillance and inform municipal interventionsAPHA 2016 Annual Meeting & ExpositionDenver, CO, USA20164382

 Google Scholar

Van SickleDMagzamenSTrueloveSOnline feedback about remotely monitored inhaled bronchodilators improves composite measures of asthma controlAm J Respir Crit Care Med2010181A3127

 Web of Science ®Google Scholar

Van SickleDMorrisonTAWeekly feedback summarizing use of remotely monitored rescue medication improves asthma controlAm J Respir Crit Care Med2011183A4755

 Web of Science ®Google Scholar

Van SickleDMagzamenSTrueloveSRemote monitoring of inhaled bronchodilator use and weekly feedback about asthma management: an open-group, short-term pilot study of the impact on asthma controlPLoS One201382e55335

 Web of Science ®Google Scholar

Van SickleDBarrettMMarcusJHumbletOSmithTImpact of a technology-driven asthma program on symptoms, control and self-management: the clinical perspective142nd APHA Annual Meeting & Exposition 2014 Annual Meeting & ExpositionNew Orleans, LA, USA2014303269

 Google Scholar

Van SickleDSuJBarrettMImpact of a mobile health pilot study on asthma rescue inhaler use, control and self-managementEur Respir J201546PA2528

 Web of Science ®Google Scholar

Van SickleDBarrettMHumbletORandomized, controlled study of the impact of a mobile health tool on asthma SABA use, control and adherenceEur Respir J201648PA1018

 Google Scholar