Construct and measurement properties
Figure
2 shows the seven instruments which cover both internal and external aspects of the Health Empowerment construct.
The constructs measured by these seven instruments all include internal aspects of empowerment, with explicit items on understanding of health problems, and the ability to self-care, or stay healthy. External aspects of empowerment are less extensively covered, perhaps because these are traditionally seen as measures of patient experience, not outcome [
54], and because we did not include measures which exclusively capture patient experience in our review. None of the instruments directly address symptoms. The three most widely used are
PAM-13,
PEI and the
heiQ.
PAM-13 is based on a single construct of activation: which is being engaged in managing one’s own health. Patients are measured on a four-stage Guttman scale of activation: from belief that an active role is important, to taking action and staying the course under stress [
55]. The
heiQ was developed to assess the impact of patient education programs across a broad range of chronic conditions [
56]. It has a wider construct than the
PAM-13 and contains eight independent dimensions: positive and active engagement in life, health directed behaviour; constructive attitudes and approaches; self-monitoring and insight; health service navigation; social integration and support and emotional wellbeing [
56,
57]. These domains overlap with both internal and external empowerment, and also with the other two domains. For example, “positive and active engagement in life” overlaps with Health Status. The instrument also includes aspects of Health Perceptions, including “satisfaction with health”, and “health concerns”.
PEI was developed specifically for primary care, and asks the patient to retrospectively rate change in enablement, resulting from a single consultation. As well as understanding and self-care, it addresses concerns, and indirectly addresses the impact of symptoms (through questions on coping with illness, and coping with life) [
58].
The four remaining instruments are less widely used.
EC-17 was developed to measure the skills and attributes of an effective consumer, for use in self-management interventions [
59,
60].
PE-LTCs was developed to measure empowerment in long-term conditions [
61]. The
Barriers instrument does not purport to assess empowerment, rather barriers to self-management in long-term conditions [
62]. However, the construct of “barriers” is related to empowerment, in that reducing barriers increases empowerment.
CAM-3 measures the quality of the therapeutic relationship as: 1) patient-centred care, 2) perceived provider support 3) empowerment. While this is described as an experience measure, it focuses on the consequences of a positive experience, for example, trust in the therapist and belief that the root causes are being identified and treated. In measuring patient-centred care and perceived provider support as well as empowerment, it includes some external aspects of empowerment in addition to internal [
63].
All seven instruments contain a standard list of questions, asking about today, or a person’s perception of their current self. Two of the instruments are, at least partly, transitional. The status instruments are based on a list of belief statements with a Likert (bipolar) response scale (e.g. strongly disagree to strongly agree) apart from the
EC-17, which has behavioural statements with an adjectival scale (never to always). Four of the instruments provide a single index score, and the remaining three give a profile of scores. All instruments are scored using a summative method, apart from
PAM-13, which uses a Rasch scoring algorithm [
55].
Psychometric properties
The first three of the instruments (PAM-13, PEI and heiQ) have undergone moderate levels of testing. PEI has been used widely in UK general practice and has shown acceptable psychometric properties. As a transitional questionnaire, it measures change directly, and is thus responsive.
The properties of the
heiQ were investigated using item response theory and structural equation modelling. It has demonstrated good construct validity, including, most recently, testing for measurement invariance [
64]. Some of the
heiQ sub-scores have shown responsiveness to change in randomised controlled trials [
65‐
67].
PAM-13 has strong psychometric properties, and association with a number of other health outcomes [
68]. Recent studies in the US found patient activation was influenced by community interventions [
69,
70], suggesting it may be appropriate as an outcome measure in primary care.
The
EQ-17 has shown some preliminary evidence for responsiveness to change in arthritis patients [
60,
71], although psychometric testing has been more limited. However, the authors of the instrument acknowledge that, while some skills of an effective consumer can be learned, others are a “part of personality” and not amenable to change [
59]. (pg. 1932) When compared directly to
PAM-13 it was less responsive (standardised response mean 0.25/ 0.41).
Summary
All the health empowerment instruments reviewed could, in theory, be used to measure empowerment outcomes in primary care. However, as all except
PEI and
CAM-3 were developed with long-term conditions in mind, they are less suitable for people without long-term conditions. This is most problematic with
EC-17,
PE-LTCs and the
Barriers questionnaire, which all refer to a “disease”. The first three instruments (
PAM-13,
PEI and
heiQ), which are the most robust and responsive to change, make minimal reference to “your illness” or else refer to “health problems” in general. Of these three, only the
PEI was developed specifically for primary care. The main weakness of this is that it only works at a single consultation level, through the words “as a result of your consultation today”. A format of the
PEI which asks patients to rate a longer episode of care has been tested for acupuncture. This adjusted the wording to “as a result of visiting the acupuncturist over the last few weeks or months.” However, patients had difficulty attributing change directly to the intervention [
72].
PAM-13, is more robust, but the construct is relatively narrow: its emphasis is on the internal, and it contains elements about control and responsibility which are not present in the construct described in Chapter 3. The
heiQ has the widest construct. The main weakness of the
heiQ for use in primary care is its length, and the fact that it does not explicitly address symptoms, which, for some patients, may be the primary reason for attendance.