GPs need to watch hospital activity like a hawk to prevent miscoding and overpayment. Here, Dr Charles Alessi explains how his commissioning group is doing it
There is a perverse incentive under payment by results for hospitals to preserve their levels of income by ensuring patients are coded in the most imaginative manner.
Our Kingston Consortium in Surrey has proposed a two-stage process for monitoring activity.
The aim is to establish automated systems to highlight anomalies, and then introduce a system of clinical audit to ensure providers are coding correctly, and that the income derived is proportional to the nature of the intervention.
Establishing contractual rights
Under the current contract for acute hospitals, a commissioner has the right to ‘audit a provider's charging of any of the prices to any commissioner' (clause 19.5.1) and the provider must ‘allow the auditor a right of reasonable access' (clause 19.5.1 continued).
If it is found that the provider has been overcharging or incorrectly coding, then under clause 19.7 ‘the provider shall, within five operational days of receiving written notice of the overcharge, reimburse the overcharged commissioner the amount of the overcharge and shall pay the reasonable costs of the audit'.
A new contract should incorporate two key clauses in its next iteration:
• The right of access to named hospital clinical notes for patients within the putative audit. Exactly the same rules should govern this process as govern the process of Positive Predictive Value (PPV) in primary care.
• Financial penalties for providers if audits find that they are over-charging the commissioners. Statistically robust samples are essential here.
Automating billing data
We identified a need to monitor provider activity more closely, and have suggested routing all referrals through a central portal from where they are then sent via Choose and Book to providers.
We proposed a process of accruals whereby every elective referral is assigned a putative cost associated with the anticipated healthcare resource group (HRG). Thus, for example, a referral for a knee replacement will be expected to generate a defined HRG cost.
Once the cost for the activity is billed by the provider via the Secondary Uses Service, this will be automatically matched with the predicted costs using a unique identifier, the NHS number. This is akin to reconciliation using double-entry book-keeping, and anomalies are highlighted.
The majority of inpatient secondary care activity will be charged as HRGs under the Department of Health's payment by results guidance.
Each spell of treatment can be charged to the appropriate HRG derived from ICD10 diagnosis codes and OPCS10 procedure codes combined with the age of the patient. When comparing the information in secondary care with that of primary care, the auditor will need to convert the Read/SNOWMED codes to ICD10 and OPCS10 codes. It is important to note that each inpatient activity can have more than one episode.
Auditing of data
We proposed that GP commissioners audit primary care records, secondary care records and the billing data of a set of patients' records to compare and assess the accuracy of secondary care coding and quality issues.
These audits can examine the correlation between what was requested by the GP, the treatments carried out by the provider, how they were carried out (whether day case, outpatient procedure and so on), how they were coded and what was billed. Audits can be carried out retrospectively for patients who have completed treatment and for whom the commissioners have been billed.
There will thus be an incentive for the provider to supply the appropriate HRG code in outpatient department letters as this would assist in avoiding potential penalties for data mismatch and also in getting the reward associated with attainment of a Commissioning for Quality and Innovation (CQUIN) payment.
The aim is to use the sampling technique to provide assurance on the accuracy of the sample and that it reflects the whole population. The sample will need to be big enough to accurately reflect the error rate in the whole population.
Samples also need to be random across the population, and should look specifically at consultant-to-consultant referrals and assess whether they were being delivered in accordance with contractual obligations. In Kingston, the majority of consultant-to-consultant referrals require pre-approval.
Some key factors of the audit need to be considered:
• Size and location of audit: The audit will be carried out across the different treatment types in the hospital with a sample identified by GP commissioners and performed retrospectively after billing for activity that has been affected by the provider.
• Areas of audit: An audit will take place in areas determined by the commissioner for day cases, elective inpatients, non-elective inpatients, outpatients and A&E departments.
Locally, we are using the Raosoft sample size calculator (www.raosoft.com/ samplesize) to establish the appropriate sample to represent the entire population. Sample size is calculated based on the population size, accepted margin of error and confidence level.
The audit of each record would produce an outcome of either ‘correct' or ‘incorrect'. If a margin of error between 9% and 10% is acceptable, a sample size of 100 would be sufficient to represent the entire population of each of these areas.
If a lower margin of error is preferred, the size of the sample will increase to between 300 and 400 for each type of procedure. The sample audited would be apportioned over the different specialties based on their activity levels.
• Resources: Audits can be carried out by clinical staff after identification by commissioning support. The clinician will cross-reference the information on the notes with information from primary care and billing information from the commissioners. They will be comparing the GP referral with the work carried out by the provider and the work billed for to ensure they are correct.
This will involve the clinician visiting the provider and selecting the records they wish to audit and analysing each one.
Penalty and reward
When providers have incorrectly billed the commissioner they will be penalised and charged for the audit. Penalties need to be sufficiently large to ensure that the trust changes its behaviour.
Small fines will not act as a deterrent because providers will know that even if they get caught, they would still have profited from over-charging.
A penalty could be to deduct from the provider a fine of 1% across the total income, for instance for day cases and outpatients. A decision needs to be taken on the error rate that would draw a penalty; the rate would need to be sufficient to show there is systematic over-charging by the provider and not a one-off error.
The contract clauses that need to be invoked require professional input to ensure they are clear and their interpretation is unequivocal. It is worth considering legal advice. If a trust is found in breach, this could be tested at arbitration.
Alternative methodologies such as charging the provider a multiple of the over-charged amount, therefore making the penalty proportional to the transgression, also need to be considered.
It would be helpful to also encourage providers to ensure they improved their data quality. If data quality is found to be accurate to a certain set percentage, a reward could then be delivered via the CQUIN system.
Dr Charles Alessi is a GP in Kingston-upon-Thames, Surrey, chair of the Kingston Pathfinder Consortium, and executive member of the National Association of Primary Care
Dr Charles Alessi
