Table 1 Summary of published validity data on the PACIC

Aragones [15]

USA

a.3.17 (0.87)

Reports no ceiling or floor effects

Type of analysis not clear

No significant association with number of chronic conditions

No significant association with age, sex, education, insurance, years in the US

Spanish language version

Sample 1: 100/120 (83%)

b. – f. 2.50 – 3.95 (all >0.6)

Factor loading analysis – most items correlated highly on proposed scales

Sample 2: 20 telephone interview follow ups

Spanish speaking Hispanics with diabetes in hospital ambulatory settings

(Test Re-test 0.77)

Carryer [16]

New Zealand

GP care/Nurse care

Professional self ratings much higher (on modified version of PACIC):

Modified PACIC for professionals

Sample 1: 341 (85.3% - of those who expressed an interest in participating)

a.2.7/3.3

a.4.0

b.2.9/3.5

b.4.3

c.3.1/3.7

c.3.8

d.2.3/3.2

d.3.8

Sample 2: 89 GPs & nurses

e.2.8/3.5

e.4.1

Primary care patients and practitioners

f.2.6/2.9

f.3.8

Gensichen [8]

Germany

a.3.25 (0.91)

Ceiling effects: PA (12.9%) and PS/C (8.9%)

EFA two factors (‘Patient activation

Overall PACIC with number of conditions and PHQ9 both NS

No significant associations with age, sex, education

German language version

442/485 (91.1%)

b.3.65 (0.80)

Floor effects:

and problem solving’ and ‘ Goal setting and co-ordination’) 46.5%

High correlations with all EUROPEP scales

Patients with major depression in primary care

c.3.47 (0.45)

GS/T (4.6%)

d.2.97 (0.74)

Missing data from 0.7% - 5.4%

some items did not load as expected

e.3.69 (0.77)

f.2.83 (0.76)

Glasgow [7]

USA

a.2.60 (0.93)

No items had ceiling effects

CFA – moderate fit

No variation in response across 6 most common long- term conditions (excluding diabetes patients who report better follow up); Higher PACIC scores associated with more conditions (r = 0.13, p<0.05)

Correlations (PACIC and subscales) with patient characteristics all <=0.25; Higher overall PACIC related to age (higher) and gender (female); Gender significantly related to all subscales (0.14 to 0.25; P<0.05)

Sample 1: 379/500 (76%) of which 283 had chronic condition (57%)

b.2.99 (0.82)

Floor effects identified, but not reported in detail

c.3.13 (0.77)

96% had no missing data

d.2.43 (0.84)

e.2.87 (0.90)

Sample 2: 82/100 sent follow up at +12 weeks (82%) of which 63 had chronic condition (63%)

f.1.07 (0.86)

Primary care

(3 month re-test 0.58)

Overall PACIC and all subscales correlate significantly with Hibbard Activation and Safran Assessment of primary Care sub scales (with exception of PACIC F/C and Safran Integration sub scale)

Glasgow [12]

USA

a.3.2 (0.96)

Adequate variability

No significant relationship to number of conditions

Correlated with physical activity (r=0.17) but not fat consumption

Includes PACIC 5As

363 (63%)

b.3.6

3-9% sub scale scores <1.5, (4% on summary scale)

Related to quality of care (composite lab assessment r=0.23) and composite self management support (r=0.25)

No significant differences with sex, ethnicity or income

Type 2 diabetes patients in primary care

c.3.5

7-22% sub scale scores >4.5 (9% on summary scale)

d.3.0

e.2.9

f.3.4

5As mean = 3.2

Goetz [17]

Germany

Patients tended to gravitate to both end points (0% and 100%)

FA indicated a 1 factor solution for the PACIC short form

There was no correlation between the mean overall score of the PACIC short form and number of chronic conditions

PACIC short form & revised scoring

264 (49%)

Non-response rates ranged from 4.2% - 12.5%

Over 18 with at least one chronic condition in primary care

Gugiu [13]

US

(Short form PACIC – 11 items – Ordinal alpha = 0.955 (sample 1) and 0.963 (sample 2); Ordinal omega 0.956 (S1) & 0.963 (S2); Eight month Test re-test reliability (n=250) = 0.638)

EFA within a CFA

No associations with HBa1c, LDL, microalbumin, BP

Modified PACIC percentage scale

Sample 1: 529/943 (55%)

Unidimensional, 11 item variant

Sample 2: 361/943 (38%) (111 not in first sample)

Type 2 diabetics, large physician networks

Gugiu [9]

USA

(Short form PACIC – 11 items, Alpha 0.945, ordinal alpha 0.972, ordinal omega 0.973)

Missing data 0.2%

CFA Poor fit to 5 factors

No associations with clinical indicators

Modified PACIC percentage scale (linked to above)

539/943 (57%)

to 2.8%, 8.9% failed to respond to at least 1

EFA 1–3 factors, 1 factor preferred

Type 2 diabetics, large physician networks

Kurtosis (trimodal, 43% 90-100%, 24% 0-10%)

Jackson [18]

USA

a.3.1

Non white patients more likely to report experience consistent with the CCM (OR 2.3) (PS and FU significant among subscales); Patients not completing high school more likely to report experience consistent with the CCM (OR 3.0) and subscales

204 (69%), but 189 (64%) complete information

b.3.3

Patients with diabetes receiving VA primary care services

c.3.6

d.3.1

e.3.4

f.2.6

Maindal [11]

Denmark

a.(0.94)

Missing 0.5 – 2.9%

CFA good fit for 2 indices, poor for 4

Patients with self-rated good health reported higher scores on ‘Patient Activation’, ‘Decision Support’ and ‘Goal Setting’; Patients with more than one additional disease rated lower on PA and DS

No significant associations with sex, age

Danish Language version

1265/2476 but only 560 met criteria of diabetes > 2 years + medical treatment (22.6%)

b. – f. (0.71 – 0.86)

Floor effects: 2.7% - 69.2%, >15% for 17 items

Patients on national diabetes register

Ceiling effects: 4.0% – 4.04%, >15% for 12 items

Rosemann [19]

Germany

Male/Female

Adequate variability

Education and age predicted overall PACIC score in regression

Significant relationships with disease duration, BMI, co-morbidities, PHQ sum, AIMS2F,

Significant differences by gender and educational level (p<0.01), marital status and age (p<0.05),

German language version

1021/1250 (81.7%)

a.2.79/2.67

PACIC 5As

Patients with OA in primary care practices

b.3.51/3.39

c.3.34/3.33

d.2.41/2.31

e.2.39/2.29

f.2.94/2.62

Rosemann [20]

Germany

a.2.44 (0.90)

Adequate variability in the overall scale & all subscales

PACIC and GS/T and FU/C scores significantly higher for patients with co-morbid diabetes, but no significant associations with other co-morbidities (hypertension, depression, CHD, COPD)

Age and gender showed weak correlations with overall PACIC and majority of subscales; no significant relationship with educational level.

German language version

Sample 1: 236/300 – 78.6%.

b.2.79 (0.85)

Floor effects in 3 subscales (F/C - 4.6%; PA - 3.8%; and GS/T - 3.4%).

Strong correlations found between PACIC sub scales and EUROPEP as expected

PACIC 5As

Sample 2: 71 of subset of 75 sent follow up questionnaire after 2 weeks

c.2.56 (0.78)

Ceiling effects below 1.3%

OA patients from 75 primary care practices

d.2.31 (0.81)

e.2.48 (0.86)

f.2.01 (0.81)

(Test-retest - overall 0.81; PA 0.77; DSD/DS 0.78; GS/T 0.82; PS/C 0.79; FU/C 0.85.

Schmittdiel [10]

USA

Mean 2.7

71% completed all items, 90% completed 17+

Relationships similar for subgroup by disease

Significant relationship with higher quality of life (OR 1.2); no relationship with adherence to medications (OR 1.06)

4108/6673 (61%)

Higher ratings of health care (OR 2.36),

Significantly associated with greater engagement in self management behaviours (OR 1.21 to 1.41); use of self management services (OR 1.4)

Private health care members on one of six chronic disease registers

Szecenyi [21]

Germany

DMP/Non-DMP

Mean 3.2 DMP versus 2.68 non-DMP (significant p=0.001) and across all subscales except patient activation (p=0.05), greatest mean difference in F/C, least in PA

German language version

1532/3546 (42.2%)

a.3.26/2.86

(1,399 valid responses = 39%)

b.3.26/3.09

PACIC 5As

Patients with type 2 diabetes in primary care, in or outside disease management programmes (DMPs)

c.3.52/3.29

d.2.91/2.50

e.3.39/3.04

f.3.13/2.70

Taggart [14]

Australia

S 1

S 2

Sample 1: 73% completed all 20 items; 95% completed at least 17 items.

EFA, both 2 factor solutions, 59% & 61% variance

Higher PACIC scores associated with higher patient self-rated health

Degree/diploma, retired, hypertension/IHD & greater duration of disease had negative associations with both factors and total PACIC scores; Employed and married/CH had negative associations with planned care factor and total PACIC score

Sample 1: 2552/2642 (96%) (2642 of 3349 asked & consented to take part)

a. 3.01

a.3.07

Sample 2: 79% completed all 20 items; 95% completed at least 17 items.

F1 SDM and SM (12 items across four scale) (alpha 0.939 & 0.943)

SDM and AM positively associated with good health

F2 Planned care (8 items across 3 scales) (Alphas 0.883 and 0.878)

Sample 2: 963/1000 (96%) (1000 out of 4167 consented to take part)

Patients with CHD, hypertension and/or T2 diabetes in general practice

Wensing [22]

Netherlands

a. 2.9 (0.93)

22-35% missing data. Items 15, 17 & 20 had >30% non response

PCA – five factors

Association between PACIC and EUROPEP aggregated scores all positive as expected.

Higher enablement in patients associated with lower PACIC scores – contrary to expectations

Dutch language version

165 (72%)

b. 3.2 (0.85)

Lowest response category used by >30% for 11 items. (7-76%)

(70% variance explained; KMO 0.844; Bartlett’s p=0.000)

Randomly sampled patients with diabetes or COPD from four general practices (involved in a programme to enhance structured diabetes care)

c. 3.5 (0.75)

Highest response category used by >30% for 6 items (10 – 54%)

Matched three pre- defined domains (but not delivery system/practice design nor follow up/co-ordination)

d. 2.5 (0.81)

e. 3.3 (0.87)