Problem The need to improve the prescription, administration and monitoring of oxygen therapy.
Design An interventional, prospective audit.
Background and setting Wellington Hospital, a teaching and tertiary referral hospital in New Zealand in 2007 and 2008.
Key measures for improvement Demonstration of adequate oxygen prescribing, administration and monitoring of oxygen therapy.
Strategies for improvement The introduction of a new drug chart with a specific oxygen prescription section. Targeted educational lectures primarily to medical staff.
Effects of change 610 and 566 patients were reviewed in the first and second audits. After introduction of the new oxygen prescription section on the drug chart the proportion of patients whose oxygen therapy was prescribed increased from 15/85 (17.6%) to 39/98 (39.8%), relative risk 2.3 (95% CI 1.3 to 3.9). The proportion with adequate oxygen prescription, with documentation of device, flow rate or inspired oxygen concentration, and the target oxygen saturation increased from 5/85 (5.9%) to 36/98 (36.7%), relative risk 6.2 (95% CI 2.5 to 15.0). Introduction of the new charts was not associated with changes in clinical practice in terms of assessment of oxygen saturations on room air and commencement if ≤92%, or the titration of oxygen therapy in response to oxygen saturations ≤92%.
Lessons learnt An oxygen prescription section on hospital drug charts improved the prescription of oxygen but did not improve clinical practice. Additional strategies are required to improve the administration of oxygen therapy in hospitals.
- prescription charts
Statistics from Altmetric.com
The therapeutic importance of oxygen in hypoxaemia is an undisputed tenet of medical practice, just as fluid resuscitation is in hypovolaemia. However, like all drugs and other treatments, oxygen has potential risks as well as benefits and there is a need to prescribe it for defined indications, with specification of the dose and method of delivery, the target saturation range, and for the patient's response to be monitored. To facilitate this approach, the first British Thoracic Society (BTS) guideline for ‘Emergency oxygen use in adult patients’ recommends that every hospital should have a designated oxygen section on the drug chart on which oxygen should be prescribed and a target oxygen saturation range stated.1
Outline of problem
A previous audit from the UK has reported oxygen prescription on a respiratory unit to be 55%, and of these, 7% were accurate prescriptions. The introduction of an oxygen prescription chart increased these proportions to 91% and 77%, respectively.2 This audit, however, did not report whether the improved prescription changed clinical practice in terms of assessment before commencing oxygen therapy, monitoring or titration, which is the ultimate objective in improving prescribing standards. Certainly there is a well documented need to improve current practice, as numerous audits have demonstrated the suboptimal prescription of oxygen therapy by medical staff, mostly among medical and respiratory patients.2–8
The purpose of this audit was to obtain a ‘snapshot’ of the use of oxygen in Wellington Hospital, a tertiary referral centre in New Zealand, and investigate whether the introduction of a new drug chart with an oxygen prescription section (a national initiative) improved hospital-wide prescription, administration and monitoring of oxygen therapy.
This two stage audit was performed between in Wellington Hospital, Capital & Coast District Health Board, Wellington, New Zealand. This is a tertiary referral centre and teaching hospital with approximately 500 beds and approximately 48 000 presentations to the emergency department (ED) per year. The hospital's audit board approved the collection and storage of data. Verbal consent was obtained from each patient.
Assessment of problems and strategy for improvement
Details of approach taken
The first audit was performed between April and May 2007. All hospital inpatient areas were included except for the neonatal unit, paediatric ward, psychiatric ward, and those currently attending the ED and intensive care unit. The first audit was before the introduction of a new drug prescription chart and occurred on 3 separate days, at least 2 weeks apart. All patients were included unless they were off the ward at the time of the audit or they had been included in the preceding audit. Every patient was viewed to see if they were receiving oxygen at the time of the audit. Each patient's hospital records and observation charts were reviewed and information on monitoring before starting therapy, prescription, and monitoring and titration while receiving oxygen was recorded. It was also noted if the patient was brought in by ambulance, received oxygen therapy in transfer, and whether the patient was admitted via the ED. Criteria were set to assess whether prescription, assessment, monitoring and titration of oxygen were adequate (table 1). For prescription, all criteria had to be met for this feature of oxygen therapy to be considered adequate.
In November 2007 a new drug chart with a specific oxygen prescription section on the front of the chart (see figure 1), was introduced into the hospital as part of a nationwide plan. This section prompts the prescriber to state the flow rate, device and target oxygen saturations range. During the first few months following implementation of the new drug chart, the investigators presented the findings of the first audit and provided education about appropriate oxygen use at various medical and nursing meetings including the weekly internal medicine meeting, junior house surgeon teaching, and Grand Round.
A second audit, using identical methodology, was conducted between April 2008 and May 2008, a year after the first audit and 6 months following the introduction of the new drug chart.
Measurement of change
The proportions of audited patients with characteristics related to oxygen use were compared using relative risks for 2008 versus 2007. Data were provided for clinical practice in both the wards and the ED settings, and comparisons made between the first and second audit. SAS V.9.1 (SAS Institute, Cary, NC 2003, USA) was used for the analysis.
Results of study intervention
There were 610 and 566 patients audited over the 3 separate days in 2007 and 2008, respectively. Approximately one-third of patients were brought to hospital by ambulance. Of these, 99/188 (52.7%) and 70/177 (39.5%) were administered oxygen at any time during the ambulance journey, in 2007 and 2008, respectively.
Just over half of the patients were admitted via the ED. Of these 82/353 (23.2%) and 66/287 (23.0%) were receiving oxygen at presentation to ED. Of the patients who did not arrive on oxygen, it was commenced in 66/353 (18.7%) and 41/287 (14.3%) at any time during their stay in the ED.
The first audit identified 85/610 (13.9%) ward based hospital inpatients who were receiving oxygen, and 98/566 (17.3%) during the second audit.
No patients admitted through the ED who received oxygen therapy had it prescribed. The prescription of oxygen on the drug chart in ward based hospital inpatients is shown in table 2. In the first audit, 15/85 (17.6%) patients receiving oxygen had an oxygen prescription. In the first audit only 5/85 (5.9%) prescriptions were adequate—that is, delivery device, flow rate or fraction of inspired oxygen, and target oxygen saturations were clearly stated on the chart. In the second audit, following introduction of the new drug chart with a specific oxygen prescription section, 39/98 (39.8%) patients receiving oxygen had an oxygen prescription, relative risk compared to 2007, 2.3 (95% confidence interval (CI) 1.3 to 3.9), p=0.001. In the second audit in 36/98 (36.7%) oxygen prescriptions were adequate, relative risk compared to 2007 6.2 (95% CI 2.5 to 15.0), p<0.001.
Assessment before commencing oxygen therapy
The proportion of patients who had assessment of oxygen saturations on room air and commenced on oxygen therapy if saturations were ≤92% in the ambulance, ED and on the ward are shown in table 3.
This practice was generally better in the ED than in the ambulance or on the wards, although was still undertaken in a minority of patients. There were no significant differences in the proportion meeting this criterion between the two audits, in the three sites studied.
The proportion of patients who had adequate monitoring of oxygen saturations while receiving oxygen was >90% in both the ward and ED in both audits, and there was no significant change between the audits (see table 4).
The proportion of patients who had their oxygen treatment up-titrated in response to oxygen saturations ≤92% on at least two occasions is shown in table 5. This practice did not improve following the introduction of the oxygen prescription chart.
In this ‘snapshot’ of oxygen therapy outside critical care areas of the hospital, the initial audit identified that <20% of patients receiving oxygen had an oxygen prescription, and in only 6% was the prescription adequate. The institution of an oxygen prescription section on the drug chart led to an improvement in the prescription of oxygen therapy, including documentation of the delivery device, flow rate and target oxygen saturations. However, in spite of accompanying staff training, it did not lead to improved clinical practice in terms of the initial assessment of the requirement for oxygen therapy, or the titration of oxygen therapy in response to low oxygen saturations. This suggests that a prescription box alone is not sufficient to achieve good clinical practice and that other measures are required.
Our finding of an improvement in oxygen prescription using a specific section on the drug chart has also been shown by Dodd et al.2 Although the provision of an oxygen prescription section improved the frequency and quality of oxygen prescription, in our study the quality of the prescription was still poor in two thirds of patients receiving oxygen in the wards. We also observed a failure to document hypoxaemia in a majority of patients before oxygen prescription. This was also reported by Brougher et al9 and suggests that oxygen is commonly administered for presumed rather than documented need.
A positive feature of the audit was that almost all patients receiving oxygen therapy had regular monitoring, defined as oxygen saturations recorded twice a day. It could be argued that stricter criteria would have been preferable, such as those proposed by the BTS guideline,1 which recommend monitoring four times a day in stable patients and continuous monitoring in unstable patients. Nevertheless, the findings do indicate that this component of oxygen therapy is generally well undertaken. However, the response to the documentation of repeatedly low oxygen saturations was disappointing, particularly in the ward situation. Furthermore, the introduction of the new oxygen prescription section, which specifically included a target oxygen saturation range, did not improve this practice.
We did not look specifically at the other scenario involving the down-titration of oxygen in response to hyperoxia. This may also be important due to the potential risks of high concentration oxygen therapy in certain clinical situations such as exacerbations of chronic obstructive pulmonary disease (COPD),10–13 uncomplicated myocardial infarction,14 15 and stroke.16 This issue warrants investigation in future audits. However, in reality it would be difficult to ensure good practice in the down-titration of oxygen if it is not prescribed in the first place.
Not surprisingly, given the frequent use of oxygen within this area of the hospital and the closer medical and nursing supervision provided, clinical practice but not prescription was better in the ED when compared with the wards. This feature also illustrates the potential disconnect between prescription and clinical practice.
The findings of this audit suggest that strategies other than an oxygen prescription section in the drug chart need to be developed to improve clinical practice. Our intervention was directed primarily towards the medical staff, but in practice it is the nursing staff who are responsible for monitoring and titration. A recent study showed that an educational intervention to medical admissions unit nursing staff, and implementation of a nurse facilitated oxygen prescribing reminder strategy by liaison with junior medical staff, significantly improved oxygen prescription and a reduction in inappropriate oxygen prescribing.17 Other approaches include developing an oxygen monitoring form which facilitates the adjustment of oxygen therapy in response to high or low oxygen saturations. and incorporating oxygen saturations into ‘track and trigger’ systems such as the modified Early Warning Score.18 Oxygen saturation has been shown to be an important predictor of mortality19 and the Standardised Early Warning Score (which incorporates oxygen saturation) has been shown to correlate with both in-hospital mortality and length of stay.20
In summary, this audit has demonstrated a statistically significant and clinically relevant improvement in the frequency and quality of the prescription of oxygen following the introduction of a new drug chart with a specific oxygen prescription section. However, this improvement did not translate to improved clinical practice in the administration of oxygen therapy. We recommend that strategies and protocols are developed which facilitate the assessment of oxygen saturation before oxygen use and the change in oxygen therapy in response to low or high oxygen saturations. It is likely that there are similar requirements for the delivery of oxygen in the community by doctors, nurses and paramedical staff.
The prescription and administration of oxygen therapy is suboptimal in hospital inpatients.
An oxygen prescription section on hospital drug charts improves the prescription of oxygen but not clinical practice.
Next steps for further improvements
Additional strategies are required to improve the administration of oxygen therapy in hospitals.
Possible strategies include nurse facilitated oxygen prescribing reminder strategies for junior medical staff, developing an oxygen monitoring form to facilitate adjustment of oxygen therapy in response to high and low oxygen saturations, and incorporating oxygen saturations into ‘track and trigger systems’.
Competing interests All authors declare that the answer to the questions on your competing interest form are all “No” and therefore have nothing to declare.
Ethics approval This study was conducted with the approval of the Wellington Hospital Audit Board.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.