Objectives A 2009 study of lumbar puncture (LP) found poor documentation. The current study, from the same hospital, aims to determine if quality of documentation has changed. Secondary outcomes include understanding of procedure & investigations.

Design Single-centre University Hospital. Retrospective study of medical notes. Sample population 147 people undergoing LP from 1 September 2010–28 February 2011. 50 adult cases assessed.

Results Half of LPs (52%) were performed for suspected meningitis. Almost all patients (92%) underwent CT brain prior to LP. On average, imaging took place within 24 h of admission, & LP after 2 days. As in 2009, documentation was poor, particularly for risk explanation, positioning, opening pressure, investigation requests, CSF results & management plan. Recording of results & management plan deteriorated significantly from 2009 to 2011 (p<0.05). Paired serum samples were sent in only 45% of cases, & PCR was requested in just 15% of suspected meningitis. Most CSF results were not diagnostic (95%) & most patients are found not to have the condition for which LP was indicated. Only half of patients (54%) with suspected meningitis were documented as receiving parenteral antibiotics.

Conclusions Though national guidelines support clinical assessment over brain imaging, most patients undergo CT before LP. The 2 day delay until LP is markedly longer than the 4 h recommended nationally in suspected meningitis. Documentation is poor & continues to deteriorate. Antibiotics should be given immediately in suspected meningitis; failure in half may represent poor documentation or management. Though most CSF results are not diagnostic, LP should not be discouraged when indicated.