Related Experiment Video
Updated: Jul 14, 2026
A Microcontroller Operated Device for the Generation of Liquid Extracts from Conventional Cigarette Smoke and Electronic Cigarette Aerosol
Published on: January 18, 2018
Smoking status recording in GP electronic records: the unrealised potential
Vanessa Selak1, Susan Wells, Robyn Whittaker
1Clinical Trials Research Unit, The University of Auckland, New Zealand. v.selak@ctru.auckland.ac.nz
Objective:
To investigate the recording of smoking status and factors associated with the recording of smoking status in general practitioner (GP) electronic medical records (EMRs) in New Zealand, and the suitability of this source as a prevalence measure.
Setting:
General practices affiliated with an Auckland-based primary health organisation.
Population:
Patients registered with 84/107 (78.5%) eligible GPs who had used EMRs for at least a year and had PREDICT-CVD, a web-based cardiovascular disease risk assessment and management decision support program, integrated with their practice software.
Design:
Audit of EMRs using data from an evaluation of PREDICT-CVD.
Main Outcome Measures:
The proportion of EMRs audited (Maori, non-Maori) with smoking status recorded and, among those with smoking status recorded, also Read-coded, and factors associated with greater recording of smoking status.
Results:
Smoking status was recorded among 49.6% of Maori and 38.3% of non-Maori prior to the installation of PREDICT-CVD. Among those with smoking status recorded, smoking status was also Read-coded among 49.8% of Maori and 62.3% of non-Maori. Factors associated with greater recording of smoking status were installation of PREDICT-CVD, male sex, Maori ethnicity, cardiovascular disease and diabetes. Age was also associated with the recording of smoking status.
Conclusion:
General practitioner electronic medical records in New Zealand are currently not a suitable source of smoking prevalence data, even if manually searched, as a large proportion of records did not have smoking status recorded. Such records are an even less suitable source of smoking prevalence if data extraction by remote querying (using Read codes) is relied upon. The potential to estimate the prevalence of smoking from GP records has not yet become a reality. Installation of electronic decision support systems, such as PREDICT-CVD, could improve the recording and Read-coding of smoking status, and thereby the availability and accessibility of these data.
