Data Cleansing and Clinical Coding

This information is relevant to general practices taking part in the Practice Incentives Programme (PIP), which is aimed at supporting general practice activities that encourage continuing improvements and quality care, enhance capacity, and improve access and health outcomes for patients.

Practices must ensure that where clinically relevant, they are working towards recording the majority of diagnoses for active patients electronically using a medical vocabulary that can be mapped against a nationally recognised disease classification or terminology system. Practices must provide a written policy to this effect to all GPs within the practice.

The checklist items below are intended to help establish eligibility against this requirement and should be read in conjunction with the PIP eHealth Incentive Guidelines available at Human Services website.

This requirement initially does not necessitate software changes but rather the requirement is a work process change associated with whichever medical vocabulary is used by your vendor (e.g. SNOMED-CT, DOCLE, PYEFINCH and ICPC2+).

In time, it is envisaged that your vendor will map your local medical vocabulary against a nationally recognised disease classification or terminology system (e.g. ICPC2, ICD10-AM and SNOMED-CT). SNOMED CT has been identified by the Agency as the preferred clinical terminology. More information on SNOMED-CT is available at www.digitalhealth.gov.au.

An 'active patient' is a patient who has attended the practice three or more times in the past two years as defined in the RACGP Standards for general practices.

Your clinical software will only 'read' your patient as having osteoporosis, coronary heart disease or any other condition if this information is entered correctly and consistently. Rigorous and systematic terminology recording is the act of selecting clinical conditions from a list of conditions rather than free text entries. Systematic recoding ensures that the patient is assigned a standard disease code for each diagnosis.

Your clinical software may already make use of a nationally recognised disease classification/ terminology system. If it does not currently do so – it may be required to do so at some time in the future. Nonetheless by choosing a diagnosis from your system's 'condition list' rather than free texting into the record you are acting in a way that is compliant with this PIP requirement even if your software does not yet provide the mapping to the My Health Record system.

The act of systematically recording allows your clinical software to draw on the coded information for quality improvement activities and when creating documents such as a Referral Letter, Shared Health Summary or Event Summary. Please refer to the Clinician's Desktop User Guide and specifically to sections Include Digital Health in Quality Improvement, Accurate and Complete Data is Necessary and Improving Data Quality and Practice Performance available at www.digitalhealth.gov.au.

You may also wish to refer to guidelines developed by the Digital Health Support Officer Network and the Royal Australian College of General Practitioners (RACGP). These guidelines provide advice on the self-assessment of data quality as well as the content of patient records and suggest proven methodologies for improvement. The guidelines for the content of patient records (contained in the RACGP Standards) at www.racgp.org.au/your-practice/standards.

As each clinical software package may have applied the clinical codes differently you will need to refer to your software package's manual for detailed instructions on how to do this.

Achieving consistent disease coding requires the entire clinical team to be on board, so it's important to get all staff engaged in the process. A practice policy for clinical coding will need to address the roles and responsibilities of each team member.

Some steps to consider in developing this policy for your practice are:

  1. Determine how your clinicians are currently disease coding. Clinicians may be doing it differently, with some using the free text field in the clinical software.
  2. As a team, decide on what codes will be used as standard across the practice. You may find it easier to discuss and agree on the codes for one or two disease areas at a time through regular team meetings (possibly targeting specific high volume diagnoses). Ensure that all clinicians are aware of and are able to use the drop down 'condition list' in the clinical software.
  3. Monitor and review the system. This could include providing training for clinicians as part of orientation, and providing regular updates and reminders of the system to all staff. In particular, it is important to acknowledge the efforts of your health service team when coding improves. This will help to ensure these changes are sustained.
  4. Agree on a process to ensure that the majority of diagnoses for your active patients are, where relevant, coded and recorded appropriately.

    The act of systematically recording a diagnosis is a clinical responsibility rather than an administrative function.

    In all cases, any coding or re-coding of a patient's clinical record should be led by a clinician. Often this will be straight forward and, providing that clear rules and guidelines are established, other health professionals can assist in patient coding or re-coding. However, where there is doubt about coding a diagnosis, a clinician should be consulted to make the final decision.

    Ensuring that patients have the correct diagnosis recorded within your clinical software may be tackled at particular times in the patients 'cycle of care'. For example, it may be practice policy that a patient's summary diagnosis list is checked for accuracy at the time of a GP Management Plan, Health Assessment or Medication Review, at regular check-ups and at the time of referrals to confirm that the patient's health information is accurate and up to date.

To locate patients who may be missing a coded diagnosis within the clinical software, you may try one or more of the following examples:

  • Search for patients who are listed as taking a medication that is generally indicative of being diagnosed with a specific chronic condition, but where that diagnosis has not been coded. For example, search for patients on bisphosphonate medications who do not have a diagnosis of osteoporosis recorded.
  • Search for patients who have had a pathology test indicative of being diagnosed with a specific chronic condition, but where that diagnosis has not been coded. For example, search for patients that have had a regular HbA1c blood test who do not have a diagnosis of diabetes type 1 or 2 recorded.
  • Check with your clinical software provider if there is a search function that enables you to identify and clinically code unstructured data.