Sir,

We congratulate Narendran et al1 on their study of the risk factors for posterior capsule rupture (PCR) and/or vitreous loss (VL), using data from the Medisoft electronic patient record (EPR). The multicentre analysis includes data from our own unit, and findings are broadly in line with our clinical experience. The authors state that ‘completeness of these (EPR) records is detailed and unusually high’, although there was no attempt to quantify the accuracy of clinical data. If these data are inaccurate, then the assessment of risk may also be inaccurate.

We attempted to quantify the accuracy of data entry for ‘ocular risk factors’ by sending an anonymous questionnaire to ophthalmologists in our unit. We asked whether, when recording a cataract operation on Medisoft, risk factors were recorded ‘always’, ‘sometimes’, ‘never’, or ‘only if complications occurred’. The response rate was 55% (11/20). One respondent did not use Medisoft; thus 10 responses were analysed.

Only one respondent (10%) stated that they ‘always’ entered all data on risk factors, although no respondent ‘never’ entered any of these data. One respondent admitted to only recording certain risk factors if a complication occurred. Recording rates were different for each risk factor (Table 1).

Table 1 Recording rates for different risk factors
Full size table

This small pilot study does indicate a significant degree of under-reporting of ocular conditions, by ophthalmologists who use Medisoft. The fact that some will record a risk factor ‘only if a complication occurs’ is a further source of bias. Each of the ocular risk factors seems to have a different degree of under-reporting and reporting bias. This will affect the estimates of relative and actual risk of PCR/VL for each of these ocular conditions.