Physical activity assessment in Australian general practice: What is the acceptability and feasibility of existing instruments and how can they be integrated into routine patient care?

Abstract

Background There are a number of physical activity (PA) assessment instruments available with potential applicability to general practice patients. Despite the availability of instruments, less than 30% of patients are assessed during routine patient encounters in general practice. A range of contributory barriers have been identified including GP knowledge about PA, time constraints, limited skills with the interpretation of PA assessment and limited use of non-GP staff e.g. Practice nurses (PNs).

When identifying instruments for PA assessment that could be used in general practice, few studies have explored the views of GPs on the comparative acceptability and utility of available instruments. A better understanding of how GPs and their teams perceive assessment instruments and how they can be used in routine practice could help overcome barriers to PA assessment and inform future PA interventions. Furthermore, there is little known about the variables that influence uptake of routine PA assessment in general practice or how implementation can be enhanced. With little known about the acceptability, utility and feasibility of PA assessment instruments in general practice, this dissertation aimed to initially determine clinician preferences for PA assessment instruments. This included identifying reasons for preference, in addition to the feasibility of identified instruments for measuring PA in general practice patients. Secondly the measurement properties of preferred instruments were examined. Finally, this dissertation aimed to determine the impact of implementing the instrument(s) on PA behaviour change interventions in routine general practice. This included how the intervention impacted on data completeness for related clinical markers, if there were changes in clinician assessment and referral behaviour, and identified key features, processes, barriers, and drivers to the implementation of the intervention.

Aims The preliminary aim (Study 1) of this dissertation was to determine the acceptability of a selection of PA assessment instruments amongst a sample of Australian general practice clinicians (Study 1). Subsequent aims included determining the measurement properties of the two preferred instruments (Study 2), and efficacy of implementing one of these instrument’s in routine general practice (study 3).

Methods Study 1: Acceptability A purposive sample of GPs (N=9) and PNs (N=10) from eight general-practices in Sydney consented to participate. Stage-1 involved semi-structured interviews with participants to select preferred instruments for measuring PA. An analysis of the two preferred instruments was conducted as Stage-2, to identify differences in instrument purpose and content. Stage-3 involved participants using the two-instruments, selected from Stage-1, for 12-weeks. At the end of this period, semi-structured interviews were repeated to explore clinician experience. The two instruments identified as preferred by clinicians were implemented in Study 2 to determine their measurement properties.

Study 2: Feasibility A purposive sample of PNs and patients from eight general-practices in Sydney consented to participate. The results of the PN or patient-administered GP-Physical-Activity-Questionnaire (GPPAQ) and Three-Question-Physical-Activity-Questionnaire (Q3), were compared against accelerometer-activity. The study examined agreement with classification of PA-levels, according to Australian PA-recommendations. The combined outcomes from Study 1 and Study 2 led to the development of a multicomponent practice change intervention, referred to as ‘the intervention’ in the proceeding sections of this thesis.

Study 3: Implementation A purposive sample of four general practices located in southern Sydney, consented to participate. The results compared changes in data completeness for a selection of clinical markers, before and after the intervention (3 months) using the PEN CAT Clinical Audit Tool. Secondary outcomes included self-reported measures of PA management changes in clinician knowledge/confidence in PA assessment. Qualitative data was obtained from semi-structured interviews and observations. Results Study 1: Acceptability Clinicians indicated preferences for the GP-Physical-Activity-Questionnaire (GPPAQ) and 3-Questionnaire Physical-Activity-Questionnaire (3Q) instruments. These instruments demonstrated distinct variations in content, theoretical orientation, and outcome measures. Reasons for preference were linked to; clinician PA status and knowledge/perceived competency in PA-assessment, and features within instruments.

Study 2: Feasibility The validation study showed that the GPPAQ and 3Q instruments maintain reasonable rank order correlations for agreements against Actigraph accelerometers with low-moderate correlations between accelerometer and GPPAQ (rho=0.26), Q3 (rho=0.45). The GPPAQ showed higher agreement GPPAQ (kappa 0.73, 95%CI 0.56-0.85), and fair for Q3 (kappa 0.62, 95%CI 0.47-0.78) when compared against national PA guidelines in identifying participants as sufficiently active, compared with the 3Q. Reliability 7-day test-retest demonstrated excellent repeatability for both instruments using Intra-class Correlation Coefficients (ICCs) with 0.82-0.95 for GPPAQ, and 0.94-0.98 for Q3.

Study 3: Implementation The results of this study indicate several principles that can influence implementation including: • Access to versatile and acceptable PA behaviour change resources. • Definition around roles and responsibilities for practice personnel in terms of PA behaviour change intervention. • Acknowledge interpersonal relationships between patient and practice. • Maintain structured internal practice systems to support implementation of PA behaviour change interventions

Discussion Key findings from this dissertation include the need to ensure clinicians administering assessments have the opportunity to select PA instruments, according to their individual characteristics and preferences. Both the GPPAQ and 3Q have acceptable measurement qualities, in the context of assessing PA in general practice patients. The findings also indicate the need to consider the role of non-GP and non-clinical personnel have in implementing PA behaviour change interventions. This finding links with the final outcomes of this study, which suggest that ensuring structured internal practice systems are in place to support implementation of interventions.

The findings from this dissertation are an important step in improving the knowledge of barriers and enablers to PA behaviour change in the general practice setting, particularly the acceptability of PA assessment instruments amongst clinicians, and methods of implementing behaviour change without forgoing the need for additional capacity requirements.