Effect of ‘activity monitor-based’ counseling on physical activity and health-related outcomes in patients with chronic diseases: A systematic review and meta-analysis

Abstract

Aim. This review evaluated the effects of activity monitor-based counseling on physical activity (PA) and generic and disease-specific health-related outcomes in adults with diabetes mellitus type II (DMII), chronic obstructive pulmonary disease (COPD), or chronic heart failure (CHF).

Methods. Four electronic databases were searched for randomized controlled trials using activity monitor-based counseling versus control intervention or usual care in adults with DMII, COPD, or CHF. Pooled effect sizes were calculated using a random effects model.

Results. Twenty-four articles were included: 21 DMII studies and 3 COPD studies. No CHF studies were identified. Pooled analysis showed that activity monitor-based counseling resulted in a significantly greater improvement in PA compared to control intervention or usual care in DMII. Furthermore, these interventions had a beneficial effect on hemoglobin A1c (HbA1c), systolic blood pressure, and body mass index (BMI) (P < 0.05), whereas no differences were found on diastolic blood pressure, and health-related quality of life. Meta-analysis of COPD studies was not possible due to lack of available data.

Conclusion. Activity monitor-based counseling had a beneficial effect on PA, HbA1c, systolic blood pressure, and BMI in patients with DMII. Data in patients with COPD and CHF are limited or non-existing, respectively.

Key words::

• Chronic heart failure
• chronic obstructive pulmonary disease
• counseling
• diabetes mellitus type II
• physical activity

Key messages

• Physical activity is associated with numerous health benefits, which indicates the importance of promoting physical activity, especially in sedentary patients with diabetes mellitus type II (DMII), chronic obstructive pulmonary disease (COPD), or chronic heart failure (CHF), who have a higher risk of hospitalization and mortality.

• Activity monitor-based interventions have beneficial effects on physical activity, HbA1c, systolic blood pressure, and BMI in patients with DMII.

• Data in patients with COPD and CHF are limited or non-existing, respectively.

Introduction

Regular physical activity has a positive impact on the onset and progression of prevalent chronic diseases, such as cardiovascular disease, obesity, diabetes mellitus type II (DMII), stroke, chronic obstructive pulmonary disease (COPD), chronic heart failure (CHF), and some forms of cancer (1–3). This indicates the importance of promoting physical activity in the general population by organizations, such as the American College of Sports Medicine (ACSM) and the American heart Association (AHA) (4). Despite substantial health benefits from regular physical activity, many older adults do not meet the ACSM/AHA physical activity recommendations (5), and the level of physical activity seriously declines with increasing age (5,6). Having a chronic disease most probably accelerates this decline in physical activity, as patients with DMII, COPD, or CHF are markedly physically inactive compared to healthy peers (7–10).

Physical inactivity in older adults, particularly in elderly with a chronic disease, results in a significant economic burden, accounting for up to 3% of the direct health care costs in developed countries (11,12). Moreover, sedentary patients with DMII, COPD, or CHF have a higher risk of disease-related hospitalization and mortality (1,13–15). Therefore, there is a need for low-cost, easily accessible, and side-effect-free programs to promote physical activity in older adults with a chronic disease.

Nowadays, activity monitors are increasingly popular to promote physical activity (16,17). They can be categorized as simple and inexpensive pedometers, which basically quantify steps, to technologically advanced accelerometers, which assess the amount and intensity of physical activity in daily life (18). Activity monitors can be worn without major inconvenience, are compatible with most daily activities, and are easy to use (18). Moreover, feedback from the activity monitors can be used to increase physical activity (17). Therefore, physical activity monitors can be a very useful tool to promote physical activity.

A previous review has already shown that physical activity counseling is associated with a significant increase in self-reported daily physical activity levels (16). However, self-reported physical activity measures may not provide valuable data, since they lack precision and may be influenced by social desirability bias (18). Other reviews showed that interventions of physical activity counseling combined with an activity monitor have a beneficial effect on daily physical activity levels in adults and subjects living with disability and/or chronic disease (17,19). To date, little is known about the effects of activity monitor-based counseling on objectively measured physical activity and generic health-related outcomes in chronic diseases, such as hospitalization and survival rates; and on disease-specific health-related outcomes, such as hemoglobin A1c (HbA1c), body mass index (BMI), systemic blood pressure, and quality of life (1,13,14, 20–24).

Therefore, the aim of this study was to evaluate systematically the effect of activity monitor-based counseling on physical activity levels in adults with a highly prevalent chronic disease, such as DMII, COPD, or CHF. Additionally, we sought to determine the effect of these interventions on generic and disease-specific health-related outcomes.

Methods

This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (25).

Eligibility criteria

Included studies met the following criteria: 1) Participants: patients with DMII, COPD, or CHF; 2) Intervention: (randomized) controlled trials using an activity monitor-based counseling intervention versus a control intervention or usual care; 3) Primary outcome: objectified physical activity; and 4) Secondary outcomes: generic and/or disease-specific health-related outcomes.

Data sources and searches

A computerized literature search was performed in Medline/PubMed, Web of Knowledge (WOK), Embase, and BIOSIS in August 2012. The following key words were used to identify relevant trials: (activit? OR “energy expenditure” OR step* OR walk?) AND (“chronic obstructive pulmonary disease” OR copd OR “chronic lung disease” OR emphysema OR “chronic bronchitis” OR “chronic heart failure” OR chf OR “congestive heart failure” OR “chronic heart disease” OR “chronic disease*” OR diabet?) AND (acceleromet? OR pedomet? OR “activity monitor?” OR “step count?” OR actigraph?).

In total, 667 articles were retrieved. In addition two potentially relevant articles were found by checking reference lists of included trials (Figure 1).

Figure 1. Study flow diagram.

Study selection

Titles and abstracts were screened against inclusion criteria by two authors (A.W.V. and A.C.), and potentially eligible articles were retrieved. All articles were independently selected for inclusion by two reviewers (A.W.V. and A.C.). Disagreements could be resolved by consulting a third reviewer (M.A.). Non-English language articles, review articles, editorials, qualitative studies, and congress abstracts were excluded.

Data extraction and assessment of methodological quality

A predesigned data abstraction form was used to obtain data on study design and relevant results. For each study, authors, journal, years of publication, studied disease, patient characteristics (gender, age, and disease severity), intervention, outcome parameters, and results were recorded. Authors of included studies were contacted via (multiple) e-mail(s) and telephone call(s) to request additional data when applicable. The methodological quality of the included trials was scored independently using the Physiotherapy Evidence-based Database (PEDro) scale (26). This scale is based largely on the Delphi List and expert consensus. It consists of 11 items to score the internal validity and statistics. Items scored a ‘yes’ if the criterion was clearly satisfied, and all ‘yes’ were summed. A κ coefficient was used to measure the level of inter-observer agreement. Previously, trials with a PEDro score of ≥ 6 points were classified as ‘high-quality trials’; trials with a PEDro score ≤ 5 points were classified as ‘low-quality trials’ (26).

Statistical analysis

Meta-analytic techniques were conducted using a random effects model in RevMan 5. Pooled effect estimates were calculated using a random effects model for physical activity and health-related outcomes. All trials reporting these outcomes were included in the model. If a study included more than one intervention group, each comparison with the control group was entered in the meta-analysis.

For each analysis, the I² statistic was calculated to estimate the proportion of observed variance in effects across studies. Values of 25%, 50%, and 75% were used as boundary limits for low, moderate, or high heterogeneity, respectively (27). The presence of publication bias was checked by a funnel plot for change in physical activity level.

Results

Study and participant characteristics

Twenty-four articles fulfilled the review criteria (Figure 1). Twenty-one studies included subjects with DMII, and three studies included subjects with COPD. The selection included no trials that studied patients with CHF. Funnel plot did not display asymmetry, which indicates that there is no indication for publication bias (Figure 2). PEDro scores ranged from 4 to 9 points (median: 7 points; Table I). Three studies were classified as ‘low-quality trials’ (28–30). Failure to conceal allocation and blinding of the patients, therapists, and/or outcome assessors were the most prevalent methodological shortcomings.

Figure 2. Funnel plot comparing interventions reporting steps/day as physical activity outcome.

Table I. Methodological quality studies.

Study	Population	1. Eligibility criteria	2. Random allocation	3. Concealed allocation	4. Similarity at baseline	5. Blinding patients	6. Blinding therapists	7. Blinding assessors	8. Outcome—85% patients	9. Treatment as allocated or intention to treat	10. Between-group comparisons	11. Point measures and measures of variability	Quality (PEDro score)
Allen (36), 2008	DMII	yes	yes	yes	yes	no	no	no	yes	yes	yes	yes	8
Araiza (31), 2006	DMII	yes	yes	no	yes	no	no	no	yes	yes	yes	yes	7
Bjorgaas (37), 2008	DMII	yes	yes	no	yes	no	no	no	no	yes	yes	yes	6
De Greef (39), 2010	DMII	yes	no	yes	yes	no	no	no	yes	yes	yes	yes	7
De Greef (32), 2011	DMII	yes	yes	yes	yes	no	no	yes	yes	yes	yes	yes	9
De Greef (40), 2011	DMII	yes	yes	no	yes	no	no	no	yes	yes	yes	yes	7
Diedrich (28), 2010	DMII	no	yes	no	yes	no	no	no	no	yes	yes	yes	5
Dyson (41), 2010	DMII	yes	yes	yes	yes	no	no	no	yes	yes	yes	no	7
Engel (42), 2006	DMII	yes	yes	no	yes	no	no	no	yes	yes	yes	yes	7
Furber (43), 2008	DMII	yes	no	no	yes	no	no	no	yes	yes	yes	yes	6
Kempny (45), 2008	DMII	yes	yes	no	yes	no	no	no	yes	yes	no	yes	6
Keyserling (33), 2002	DMII	yes	yes	yes	yes	no	no	no	no	yes	yes	yes	7
Kirk (35), 2001	DMII	no	yes	yes	yes	no	no	no	yes	yes	yes	yes	7
Kirk (34), 2004	DMII	yes	yes	yes	yes	no	no	no	yes	yes	yes	yes	8
Kirk (46), 2009	DMII	yes	yes	yes	yes	no	no	yes	yes	yes	yes	yes	9
Mitsui (51), 2008	DMII	yes	no	no	yes	no	no	no	yes	yes	yes	yes	6
Paschali (47), 2005	DMII	yes	yes	no	yes	no	no	no	yes	yes	yes	yes	7
Piette (48), 2011	DMII	yes	yes	yes	yes	no	no	no	yes	yes	yes	yes	8
Tudor-Locke (49), 2004	DMII	yes	yes	no	yes	no	no	no	no	yes	yes	yes	6
Vincent (30), 2009	DMII	yes	yes	no	no	no	no	no	yes	yes	no	no	4
Weinstock (50), 2011	DMII	yes	yes	no	yes	no	no	no	no	yes	yes	yes	6
De Blok (38), 2006	COPD	yes	yes	no	yes	no	no	yes	no	yes	yes	yes	7
Hospes (44), 2009	COPD	yes	yes	no	yes	no	no	no	yes	yes	yes	yes	7
Steele (29), 2008	COPD	yes	yes	no	no	no	no	no	no	yes	yes	yes	5

A total of 2908 participants were evaluated: 2763 patients with DMII and 145 patients with COPD (Tables II and III). About half of the patients (48.1%) were male. Reported mean age ranged from 47.0 to 70.9 years in patients with DMII, and 61.2 to 65.7 years in patients with COPD. However, in some studies data on patient characteristics were missing or incomplete (Tables II and III) (28,29,31–35). Drop out ranged from 0 to 38% (Tables II and III).

Table II. Characteristics studies including patients with DMII.

Study	Population	Drop out rate (%)	Intervention	Instrument	Outcome parameters
Allen (36), 2008	DMII Total (n = 52): 57.0 (13.5) yrs; 48% male	12	8 weeks	ActiGraph 7 days at baseline and 7 days at week 7	PA: - PA level (frequency, intensity) - PA self-efficacy Health-related outcomes: - HbA1c - BMI - Blood pressure
	I (n = 27): 57 (14.6) yrs; 44% male	22	I: activity monitor and Continuous Glucose Monitoring System; 90 min of individualized education including topics of the control intervention and Continuous Glucose Monitoring System counseling; recommended to walk 30 min most days of the week
	C (n = 25): 57 (12.5) yrs; 52% male	0	C: 90 min of individualized education on diabetes, diet strategies, glucose testing, risk reduction, foot care, and a brief review of a PA brochure
Araiza (31), 2006	DMII Total (n = 30)	0	6 weeks	Yamax DigiWalker SW 701 No information on pedometer use	PA: - Steps/day Health-related outcomes: - BMI - Blood pressure
	I (n = 15): 49 (11.0) yrs	0	I: instructed to walk 10,000 steps on 5 or more days of the week
	C (n = 15): 51 (10.0) yrs	0	C: instructed to maintain normal activity habits
Bjorgaas (37), 2008	DMII Total (n = 48)	31	6 months; 4 visits	Yamax DigiWalker ML AW-320 No information on pedometer use	PA: - Steps/day Health-related outcomes: - HbA1c - BMI - Blood pressure
	I (n = 23): 56.4 (11.0) yrs; 61% male	34	I: pedometer 3 weekdays twice per month; log-book was kept of all major PAs; encouraged to increase their daily step count from one visit to the next
	C (n = 25): 61.2 (9.7) yrs; 68% male	29	C: encouraged to increase daily time spent walking guided by the log-book with PAs
De Greef (39), 2010	DMII Total (n = 41): 68% male	10	12 weeks	Yamax DigiWalker SW-200 Actigraph 7164 5 consecutive days (during waking hours), including at least one weekend day, before and after intervention	PA: - Steps/day - Intensity Health-related outcomes: - HbA1c - BMI - Blood pressure
	I (n = 20): 61.3 yrs; 65% male	10	I: cognitive-behavioral group program (5 sessions) combined with pedometer; at each session PA progress was reviewed and new goals were set
	C (n = 21): 61.3 yrs; 71% male	10	C: usual care and one single group education on the effects of PA on diabetes care
De Greef (32), 2011	DMII Total (n = 92): 62 (9.0) yrs; 69% male	4	24 weeks	Yamax DigiWalker SW-200 7 consecutive days (during waking hours) before and after intervention	PA: - Steps/day - Intensity
	I (n = 60)	3	I: face-to-face session; pedometer and diary; 24-week telephone support (7 follow-ups); encourage to increase PA
	C (n = 32)		C: usual care
De Greef (40), 2011	DMII Total (n = 67); 70.1% male	6	12 weeks	Yamax DigiWalker SW-200 7 consecutive days (during waking hours) before and after intervention	PA: - Steps/day - Intensity Health-related outcomes: - HbA1c - BMI
	I1 (n = 22): 66.6 (9.5) yrs	5	I1: pedometer and diary to be used as motivational tools in PA goal setting; 3 × 90 min group counseling sessions with a behavioral expert to increase knowledge on PA and change PA behavior
	I2 (n = 21): 70.0 (6.3) yrs	5	I2: pedometer and diary to be used as motivational tools in PA goal-setting; 3 × 15 min face-to-face consultations with a general practitioner on pedometer use, goal-setting, and relapse prevention
	C (n = 24): 66.0 (11.0) yrs	5	C: usual care from general practitioner
Diedrich (28), 2010	DMII Total (n = 33)	4	3 months	Yamax DigiWalker SW-200 No information on pedometer use	PA: - Steps/day Health-related outcomes: - HbA1c - Weight, body fat - Blood pressure
	I (n = 16): 56.7 (13.6) yrs	38	I: Diabetes Self-Management Education Program; Manpo-Kei book; asked to measure walking activity and record pedometer readings according to instructions in book; encouraged to increase PA
	C (n = 17): 54.9 (9.8) yrs	41	C: Diabetes Self-Management Education Program
Dyson (41), 2010	DMII Total (n = 42): 60.8 (9.6) yrs; 43% male	35	6 months	Omron No information on pedometer use	PA: - Steps/day Health-related outcomes: - HbA1c - BMI - Blood pressure - Quality of life (WHO-5)
	I (n = 21): 62.9 (9.5) yrs; 38% male	7	I: usual medical care from primary care physician; education about lifestyle management; pedometer; 3 lifestyle videos
	C (n = 21): 58.6 (9.2) yrs; 48% male	0	C: usual care; education from nurse
Engel (42), 2006	DMII Total (n = 54)	14	6 months	Yamax DigiWalker 700 No information on pedometer use	PA: - Time spent walking Health-related outcomes: - HbA1c - BMI - Blood pressure
	I (n = 24): 60.5 (7.34) yrs; 54% male	12	I: pedometer; 6 coaching sessions including diabetes-related education, behavior-change strategies, and psychosocial support; goals based on steps per day (between 3500 and 5500 steps per day)
	C (n = 30): 64 (6.76) yrs; 50% male	NR	C: coaching only; goals based on time spent walking per day
Furber (43), 2008	DMII Total (n = 226)	NR	20 weeks	Yamax DigiWalker No information on pedometer use	PA: - Time spent walking - Intensity
	I (n = 121): 58.3 (12.6) yrs; 54% male	12	I: pedometer and step diary; education on PA; encouraged to set PA goals (increase recorded steps per day or a time-based target)
	C (n = 105): 61.6 (12.3) yrs; 51% male	NR	C: education on PA; encouraged to undertake at least 30 min of moderate-intensity activity on all or most days of the week
Kempny (45), 2008	DMII Total (n = 30)	NR	5 weeks	Omron HJ113 No information on pedometer use	PA: - MET*min/week Health-related outcomes: - BMI - Blood pressure
	I (n = 16): 57.7 (4.8) yrs; 44% male	6	I: oral advice to increase PA; pedometer; instructed to achieve 10,000 steps per day
	C (n = 14): 56.4 (4.9); 36% male	NR	C: oral advice to increase PA; encouraged to enhance PA level, informed to increase time spent walking
Keyserling (33), 2002	DMII Total (n = 200): 59.0 yrs	NR	12 months	Caltrac 1 week (during waking hours) before and after intervention	PA: - Energy expenditure Health-related outcomes: - HbA1c - Weight - Quality of life (Mental and Social Well-Being Scale)
	I1 (n = 67): 58.5 yrs	15	I1: clinic- and community-based intervention on PA, dietary and diabetes care; increase moderate-intensity PA to a cumulative of 30 min a day
	I2 (n = 66): 59.8 yrs	19	I2: clinic-based intervention on PA, dietary and diabetes care; increase moderate-intensity PA to a cumulative of 30 min a day
	C (n = 67): 59.2 yrs	11	C: diabetes education via the post
Kirk (35), 2001	DMII Total (n = 23): 48.6 (8.5) yrs	13	5 weeks	CSA monitor 7 days (during waking hours) before and after intervention	PA: - Counts - Steps/day Health-related outcomes: - Quality of life (SF-36)
	I (n = 11)	8	I: exercise consultation; standard exercise information; encouraged to accumulate 30 min of moderate-intensity PA most days of the week or to build in 30 min of vigorous exercise 3 or more days a week
	C (n = 12)	8	C: standard exercise information
Kirk (34), 2004	DMII Total (n = 70): 57.6 (7.9) yrs; 50% male	8	12 months	CSA monitor 7 days (during waking hours) before and after intervention	PA: - Counts Health-related outcomes: - HbA1c
	I (n = 35)	15	I: counseling on improving PA and preventing relapse; standard exercise leaflet; encouraged to accumulate 30 min of moderate-intensity PA on most days of the week or to build in 30 min of vigorous exercise 3 or more days a week
	C (n = 35)	14	C: standard exercise leaflet
Kirk (46), 2009	DMII Total (n = 134)	17	12 months	ActiGraph 7 days (during waking hours) before and after intervention	PA: - Counts - Steps/day Health-related outcomes: - HbA1c - BMI - Blood pressure
	I1 (n = 47): 60.9 (9.6) yrs; 53.2% male	13	I1: physical activity pack in written form; 30 min consultation with a trained researcher; encouraged to make a PA plan, which was discussed during follow-up phone calls after 1 and 3 months and 6 months visit
	I2 (n = 52): 63.2 (10.6) yrs; 42.3% male	11	I2: physical activity pack given in written form; encouraged to make a PA plan, which was discussed during follow-up phone calls after 1 and 3 months and 6 months visit
	C (n = 35): 59.2 (10.4) yrs; 51.4% male	17	C: usual care
Mitsui (51), 2008	DMII Total (n = 23)	11	12 weeks	EM-180 No information on pedometer use	PA: - Steps/day Health-related outcomes: - HbA1c - BMI - Blood pressure
	I (n = 13): 64.4 (7.7) yrs; 23% male	9	I: pedometer program; goal to walk at least 30 min, 5 days/week for 12 weeks
	C (n = 10): 61.8 (8.3) yrs; 20% male	15	C: usual care
Paschali (47), 2005	DMII Total (n = 26)	0	3 months	BioTrainer No information on pedometer use	PA: - Activity units
	I (n = 13): 48.8 (6.1) yrs; 47% male	13	I: accelerometer with feedback; individual counseling sessions; instructed to formulate and improve goals each month based on accelerometer data
	C (n = 13): 47.0 (7.2) yrs; 47% male	13	C: accelerometer with no feedback; counseling based on self-report diaries
Piette (48), 2011	DMII Total (n = 291): 56.0 (10.1) yrs; 48.5% male	13	12 months	Omron-HJ-720ITC 7 consecutive days (during waking hours) before and after intervention	PA: - Steps/day Health-related outcomes: - HbA1c - Blood pressure - Quality of life (SF-12)
	I (n = 145); 55.1 (9.4) yrs; 49% male	14	I: 12 month telephone-delivered cognitive behavioral therapy program; pedometer-based walking program; monitor progress toward step-count goals
	C (n = 146): 56.0 (10.9) yrs; 50% male	16	C: copy of the Feeling Good Handbook—a self-help book based on cognitive behavioral therapy for depression
Tudor-Locke (49), 2004	DMII Total (n = 38): 52.7 (5.2) yrs; 63% male	13	16 weeks	Yamax SW-200 3 consecutive days (during waking hours), including one weekend day, before and after intervention	PA: - Steps/day Health-related outcomes: - HbA1c - Weight - Blood pressure
	I (n = 16): 52.8 (5.7) yrs; 50% male	22	I: 16 week behavior modification program; pedometer; Weeks 1–4: program manual containing goal-setting and problem-solving exercises, calendars, weekly group meetings; Weeks 5–16: independently continue pedometers use and calendars
	C (n = 22): 52.5 (4.8) yrs; 61% male	23	C: usual care
Vincent (30), 2009	DMII Total (n = 17): 56 (9.3) yrs; 29% male	20	8 weeks	Colorado on the Move, Accusplit No information on pedometer use	PA: - Steps/day Health-related outcomes: - HbA1c - BMI
	I (n = 9): 56.7 (10.6) yrs; 11% male	15	I: 8 weekly 2-hour group sessions including education on diabetes and exercise; record number of steps each day
	C (n = 8): 55.3 (8.2) yrs; 50% male	10	C: usual care and education given at the clinic
Weinstock (50), 2011	DMII Total (n = 1650)	20	5 years	NR No information on pedometer use	PA: - Days/week following exercise program) Health-related outcomes: - HbA1c - BMI - Blood pressure - Quality of life (SF-12)
	I (n = 837): 70.8 (6.5) yrs; 36.4% male	NR	I: telemedicine and pedometer; home video visits with a diabetes educator every 4–6 weeks; encourage to increase PA (common goal: to increase steps per day by 10%, or to maintain number of steps)
	C (n = 813): 70.9 (6.8) yrs; 37.9% male	NR	C: usual care; pedometer

I = intervention; C = control; yrs = years; min = minutes; PA = physical activity; NR = not reported.

Table III. Characteristics studies including patients with COPD.

Study	Population	Drop out rate (%)	Intervention	Instrument	Outcome parameters
De Blok (38), 2006	COPD Total (n = 21): 43% male	24	10 weeks	Yamax DigiWalker SW-200 1 week prior to rehabilitation and 1 week during week 9 (during waking hours)	PA: - Steps/day Health-related outcomes: - Quality of life (SGRQ)
	I (n = 10): 65.7 (10.4) yrs; 50% male; FEV1 43 (13) %pred	20	I: regular pulmonary rehabilitation program; lifestyle PA counseling program (4 sessions); stimulated to increase PA using pedometer as motivational and feedback tool
	C (n = 11): 62.5 (12.3) yrs; 36% male; FEV1 52 (22) %pred	27	C: regular pulmonary rehabilitation program
Hospes (44), 2009	COPD Total (n = 35)	10	12 weeks	DigiWalker SW-200 2 weeks at baseline and after intervention	PA: - Steps/day Health-related outcomes: - Quality of life (SGRQ, CCQ, SF-36)
	I (n = 18): 63.1 (8.3); 56% male; FEV1 67.4 (17.5) %pred	10	I: customized exercise counseling program (5 sessions) to enhance daily PA and to develop a more physically active lifestyle; feedback of pedometer
	C (n = 17): 61.2 (9.1) yrs; 65% male; FEV1 61.8 (14.4) %pred	11	C: usual care
Steele (29), 2008	COPD Total (n = 89)	20	12 weeks All patients participated in a 8-week pulmonary rehabilitation program that included exercise and education before intervention	RT3 6 days before and after intervention	PA: - Accelerometer activity - Exercise capacity (6-min walking distance) Health-related outcomes: - Quality of life (SF-36)
	I (n = 42)	22	I: weeks 1–4: home- and community-based exercise program with emphasis on walking; weeks 5–12: implementing a regular program of exercise, at least 20 min, 4 days/week; weekly phone calls and 1 home visit over 3 months; self-monitoring PA using pedometer
	C (n = 47)	18	C: no information

I = intervention; C = control; yrs = years; min = minutes; PA = physical activity; FEV1 = forced expiratory volume in 1 second.

Intervention characteristics

Different types of interventions were used to evaluate the effect of activity monitor-based counseling on physical activity levels (Tables II and III). Mean duration of the interventions was 30.3 weeks (range 5 to 260 weeks). In most interventions, physical activity counseling was delivered in one or more face-to-face sessions (29,30,32–50). These sessions were given by a nurse, physical therapist, (clinical) psychologist, and/or general practitioner. Supplementary advice or counseling was delivered by telephone (32–34,36,48), video-conference (50), video (41), or in written form (28,46). All studies used an activity monitor device to provide feedback to patients.

In 16 studies, intervention was compared with usual care combined with one or more sessions of advice or information on physical activity (28–30,33–37,39,41–43,45–48). In eight studies, control subjects received only usual care (no intervention) (31,32,38,40,44,49–51).

Diabetes mellitus type II

Effects on physical activity

Physical activity level at baseline and post-intervention was measured using different activity monitors (Table II). In all studies, physical activity counseling sessions were based on data acquired from the activity monitors. The most reported outcome parameter on physical activity was number of steps per day (13 studies), but several other parameters were used, such as time spent walking (42,43), activity counts (34,35,46,47), activity intensity (36), frequency (36), and/or energy expenditure (33,45).

A pooled analysis to study the effect of activity monitor-based counseling on changes in steps per day included seven studies, all using a pedometer to determine step counts. Four studies were excluded because steps per day were only available in the intervention group (28,30,37,51), and in two studies absolute changes in steps per day were not available, even after contacting the corresponding authors (46,50).

Analysis showed that activity monitor-based interventions in combination with counseling resulted in a significantly greater number of steps per day in patients with DMII compared to a usual care control group (steps per day: + 2042; 95% confidence interval 1067 to 3017; P < 0.001; Figure 3A). Heterogeneity was moderate (I2 = 73%), which can be caused by differences in included patients (age, gender distribution) and interventions (methods, duration).

Figure 3. Effects of activity monitor-based counseling on physical activity and generic and disease-specific health-related outcomes in adults with DMII. A: Effect of activity monitor-based interventions versus control on physical activity (steps/day) in patients with DMII; B: Effect of activity monitor-based interventions versus control on HbA1c in patients with DMII; C: Effect of activity monitor-based interventions versus control on BMI in patients with DMII; D: Effect of activity monitor-based interventions versus control on systolic blood pressure in patients with DMII; E: Effect of activity monitor-based interventions versus control on diastolic blood pressure in patients with DMII.

Studies excluded from the meta-analysis showed contrary results. In several studies, in which steps per day was only measured in the intervention group, a significant improvement in steps per day was found after an activity monitor-based intervention in patients with DMII (28,30,51); only one study showed no improvement (37). Another study, in which absolute changes in steps per day were not available from the corresponding author, found no significant differences in change in steps per day between an activity monitor-based intervention and usual care (46). In one study, data on steps per day were not presented (50).

Studies using activity counts (34,36), time spent walking (43), total physical activity (43,50), and energy expenditure (33) as parameters to measure intervention effects on physical activity all reported significant positive benefits. Only the study of Engel and colleagues found no differences between subjects following an activity monitor-based intervention and a control group (42).

Effects on generic health-related outcomes

None of the studies used hospitalization or mortality rate as primary outcomes.

Effects on disease-specific health-related outcomes

Hemoglobin A1c. Sixteen studies reported intervention effects on HbA1c. Because in two studies data were not available in absolute changes and standard deviation, even after contacting authors (46,50), 14 studies could be included in the pooled analysis.

Results show a significantly greater decrease in HbA1c in the intervention group compared to controls (standardized mean difference –0.23; 95% confidence interval –0.41 to –0.05; P = 0.01; Figure 3B). The heterogeneity of the studies was moderate (I2 = 43%; Figure 3B). Results remained significant after exclusion of the study of Allen et al., in which physical activity counseling was combined with a continuous glucose monitoring system (standardized mean difference –0.19; 95% confidence interval –0.36 to –0.02; P = 0.03) (36). One study excluded from the meta-analysis did not observe between-group differences in HbA1c (46).

Body mass index (BMI). Eleven articles described the effect of the intervention on BMI. Nine studies were included in the pooled analyses. The other two studies were excluded from the analysis because data were not available in absolute changes and standard deviation, even after contacting authors (45,46).

Results show a beneficial effect on decreasing BMI following an activity monitor-based intervention compared to control interventions (standardized mean difference –0.24; 95% confidence interval –0.48 to –0.01; P = 0.04; Figure 3C). Heterogeneity of the included studies was low (I² = 25%). Studies excluded from the meta-analysis found no significant differences between the intervention and control groups (45,46).

Systolic and diastolic blood pressure. Effects on systolic and diastolic blood pressure were reported in 14 studies. In three studies absolute changes in systolic and diastolic blood pressure were not available, even after contacting authors (31,46,50). Therefore, pooled analysis was performed using 11 studies (Figure 3D and 3E).

Pooled analysis showed a significant difference in changes in systolic blood pressure (P = 0.04; Figure 3D), but not in diastolic blood pressure (P = 0.34; Figure 3E), between the intervention and control groups. Heterogeneity of the studies was low (I2 = 9% and 24%). Two studies excluded from the meta-analysis found no significant changes in systolic or diastolic blood pressure after activity monitor-based interventions (31,46).

Health-related quality of life. Health-related quality of life was examined in four studies (33,35,41,50). No meta-analysis was conducted because of the use of different questionnaires (SF-36, SF-12, Mental and Social Well-Being Scale, and WHO-5 Well-Being Index), or because absolute changes in health-related quality of life were not presented (41,50). Studies found no significant improvement in health-related quality of life in patients with DMII, or no significant between-group differences (33,35,41).

COPD

Effects on physical activity

Physical activity was measured using different activity monitors (Table III). Studies used steps per day (38,44) or activity counts (29) as outcome parameters. A pooled analysis to study the effect of an activity monitor-based intervention on changes in steps per day could not be performed, because absolute changes in steps per day were not available, even after contacting authors (38,44). In studies presenting only absolute changes in steps per day, one study showed a significant difference in steps per day between an activity monitor-based intervention and usual care (44), whereas no significant difference was found in another study (38). A study using activity counts as parameter to measure intervention effects on physical activity only reported significant short-term benefits (29).

Effects on generic health-related outcomes

None of the studies used hospitalization or mortality rate as primary outcomes.

Effects on disease-specific health-related outcomes

Health-related quality of life. Health-related quality of life was examined in all studies (29,44). No meta-analysis was conducted, because of the use of different questionnaires (St Georges Respiratory Questionnaire (SGRQ), Clinical COPD Questionnaire (CCQ), Seattle Obstructive Lung Disease Questionnaire (SOLDQ), 36-Item Short-Form Health Survey (SF-36)). One study found a beneficial effect of an activity monitor-based intervention on health-related quality of life in patients with COPD compared to a control intervention (44), whereas other studies found no significant between-group differences in health-related quality of life (29).

Discussion

In this review we systematically evaluated the effect of activity monitor-based interventions on physical activity levels and health-related outcomes in patients with a DMII, COPD, or CHF. In general, interventions using an activity monitor in combination with counseling are associated with a significantly greater improvement in physical activity compared to a control intervention or usual care in patients with DMII. Moreover, these interventions had a beneficial effect on HbA1c, systolic blood pressure, and BMI in patients with DMII. Contrary results were found on physical activity and health-related quality of life in patients with COPD, although data were limited. None of the studies used hospitalization or mortality rate as primary outcomes. Moreover, CHF trials were not available.

Daily physical activity level

Pooled analysis showed that interventions using an activity monitor in combination with counseling had a beneficial effect on physical activity, measured in steps per day, compared to a control intervention (Figure 3A). This is consistent with findings of an earlier meta-analysis, showing that the use of a pedometer is associated with a significant increase of more than 2000 steps per day in out-patient (healthy) adults (17). Avery et al. also showed an increase in physical activity levels in patients with DMII, following behavioral interventions including exercise (24). Nevertheless, relevant articles seem to be missing in this review (33,36,37,42,48).

The included studies used different types of activity monitors (Tables II and III). Pedometers can be used to provide direct feedback to the participants on the level of physical activity (number of steps per day). Therefore, participants can monitor themselves and set goals for increasing their level of physical activity. In contrast, most accelerometers do not have a display and can therefore not provide direct feedback to participants. However, accelerometer data, processed by a computer, can be used for promoting physical activity by counseling sessions or by a self-monitoring approach.

It is known that simple pedometers are less sensitive to lower force accelerations, such as slow walking, than accelerometers (18). Moreover, it has earlier been shown that patients with a chronic disease walk less, but also with a lower speed compared to healthy adults (52), which can lead to an underestimation of steps while using a pedometer (53). Furthermore, pedometers can only give an indication about the number of steps per day, whereas an accelerometer is able to give information about the quantity and intensity of performed activities (18).

Health-related outcomes

To date, the effects of activity monitor-based counseling on hospitalization or survival have not been studied. Then again, this systematic review showed a significantly larger decrease in HbA1c in the intervention group compared to a control intervention or usual care in patients with DMII (Figure 3B). This is consistent with results of earlier reviews showing a positive effect of physical activity on the reduction of HbA1c, which was not mediated by weight loss (23,24,54). A decrease in HbA1c level is an indication of better glucose control, which is associated with a reduced burden of cardiovascular disease and microvascular complications in people with DMII (55). Indeed, current guidelines recommend intensive glycemic control as an effective treatment, with a target HbA1c of 7% or less (56). However, results regarding the association between HbA1c and cardiovascular events in DMII are contrary (57,58). Therefore, it remains unknown whether and to what extent a reduction in HbA1c will result in a long-term health benefit in patients with DMII.

Pooled analysis showed a beneficial effect on decreasing BMI following an activity monitor-based intervention compared to control interventions, although results are limited to patients with DMII (Figure 3C). These results confirm findings of earlier reviews, in which it was shown that interventions promoting physical activity resulted in a significant reduction in BMI both in healthy sedentary adults and in adults with COPD, DMII, neuromuscular disease, or obesity (17,24). Unfortunately, data on changes in fat free mass and fat mass are limited, whereas these data provide more valuable information than changes in BMI.

The risk of additional weight gain is increased in patients with DMII as a result of a reduced level of physical activity compared to healthy peers. Earlier studies already showed that weight loss is associated with positive long-term outcomes and substantial reduction in mortality in obese patients with DMII (59).

Obesity is also associated with decreased levels of physical activity in patients with COPD (21). The currently available data do not suggest a reduction in BMI following activity monitor-based counseling. Then again, being overweight or obese has been associated with decreased mortality among patients with severe COPD, a phenomenon commonly referred to as the ‘obesity paradox’ (60).

Elevated resting blood pressure is associated with a significantly increased risk for cardiovascular diseases (61). An earlier published review already showed the positive effects of physical activity on decreasing blood pressure (17), which in turn will result in a better prognosis and health benefits. Pooled analyses in the current review showed a beneficial effect on systolic blood pressure following an activity monitor-based intervention compared to a control intervention in patients with DMII, whereas no significant changes were found in diastolic blood pressure (Figure 3D and 3E). Earlier it was shown that even relatively small reductions in resting diastolic blood pressure of 2 mmHg can reduce the cardiovascular risk by 6% (62). Indeed, most studies showed a decrease in diastolic blood pressure of ≥ 2 mmHg following an activity monitor-based intervention (34,36,37,39,41,46,48), which was found non-significant in some studies (39,41,46) or non-different between intervention and control group (34,37,48).

This systematic review found contrary results on the health-related quality of life after activity monitor-based interventions. Indeed, different questionnaires were used (SGRQ (44), CCQ (44), SOLDQ (29), SF-36 (29,35,44), SF-12 (50), Mental and Social Well-Being Scale (33) and WHO-5 Well-Being Index (41)), making it difficult to compare outcomes. However, quality of life is an important measure of health, particular for older people and those suffering from a chronic disease. Indeed, it has earlier been shown that patients with DMII, COPD, or CHF have a reduced health-related quality of life compared to healthy peers, and it decreases with disease progression and complications (63–65). Moreover, worsening of health-related quality of life is associated with increased overall mortality (65–67), and maintaining moderate or high levels of physical activity and improvement in physical activity level are associated with a better health-related quality of life in patients with DMII, COPD, or CHF (20,22,68).

Strengths

To the best of our knowledge, this review is the first to examine the effects of activity monitor-based counseling on objectified physical activity levels and health-related outcomes in adults with a chronic disease. Indeed, earlier reviews already studied the effectiveness of physical activity promotion in adults and subjects living with disability and/or chronic disease (16,17,19), but these reviews only focused on physical activity (19), included studies with healthy subjects (17), or included studies using self-reported measures of physical activity (16,24). The strength of this meta-analysis is that it included only trials that objectively measured physical activity before and after the intervention, because self-reported measures are associated with a lack of precision and with sensitivity to social desirability (18,69). Surprisingly, studies on the effects of activity monitor-based counseling in patients with CHF are currently lacking. Therefore, we believe we have identified a major research topic within CHF. Indeed, accelerometers are already incorporated in implanted devices for continuous activity telemonitoring in patients with CHF (70). Moreover, physical activity monitoring using an accelerometer during the first phase of cardiac rehabilitation (in hospital) and directly after discharge from the hospital may increase physical activity levels in patients with cardiovascular disease, such as coronary artery bypass graft, myocardial infarction, and valve replacement (71). Therefore, it seems reasonable to assume that patients with CHF also may benefit from interventions using an accelerometer.

Limitations

Our review has several limitations. Firstly, the included results may have been biased, because outcome assessors, therapists, and/or patients were not blinded for the intervention arm. However, blinding of patients and therapists is probably not possible in these study designs, and most probably will not affect objective measures such as HbAc1 and blood pressure.

Secondly, this meta-analysis included trials with different intervention durations (5 weeks to 260 weeks), methods of counseling (number of sessions, individual or group, face- to-face or written, phone or video), activity monitors, and patient characteristics (age, gender, distribution). Moreover, the use of the activity monitors to measure physical activity before and after the intervention and also the use during the intervention was different in most studies, though reportings on this were often limited. Although our review demonstrated the effectiveness of activity monitor-based programs, these differences could have influenced the intervention effects. Because of the large variety in activity monitor types, intervention duration, and intervention design, a sensitivity analysis was not feasible. Therefore, we were not able to demonstrate differences between different activity monitors, intervention duration, and intervention designs. Furthermore, these differences in study designs resulted in low to moderate scores of heterogeneity (0% to 73%). However, it was earlier shown that about a quarter of meta-analyses have I2 values over 50% (27).

Thirdly, data on physical activity and health-related outcomes in patients with COPD and CHF were limited or non-existing. It can be assumed that activity monitor-based interventions in patients with COPD and CHF will have similar beneficial effects as in patients with DMII. However, the number and quality of the data studies are not sufficient to support this assumption, and therefore, conclusions on health-related outcomes are mainly based on studies including patients with DMII.

Fourthly, interventions using an activity monitor in combination with counseling appear to be cost-effective in the general population (72). It seems reasonable to assume that the cost-effectiveness is even higher in patients with DMII, COPD, or CHF. Nevertheless, this remains currently unknown.

Finally, 24 studies were selected for this review, but not all studies could be included in the meta-analysis, because studies used different physical activity parameters or different questionnaires to measure health-related quality of life. Furthermore, some of the authors did not reply to the request for additional data (46,50), or were not able to provide the requested data (38,44). Therefore, the observed effect of this meta-analysis may be biased by this exclusion.

To conclude, our review shows the effectiveness of activity monitor-based interventions to promote physical activity in patients with DMII. Moreover, these interventions resulted in significant reductions in HbA1c, systolic blood pressure, and BMI in patients with DMII compared to usual care. Data in patients with COPD and CHF are limited or non-existing, respectively. Future studies are needed to study the long-term effects of activity monitor-based counseling interventions on physical activity and generic and disease-specific outcomes, including a cost-effectiveness analysis.

Acknowledgements

The authors thank the corresponding authors of references (28,30,31,33,34,39,41,45,48) for sending us additional data which made the meta-analyses possible.

Declaration of interest: This work was supported by ‘Stichting de Weijerhorst’ and Point-One funding from AgentschapNL, Dutch Ministry of Economic affairs, the Netherlands. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors report no conflicts of interest.