The NHS provides health care for the whole of the UK, with no extra cost to patients. This valuable resource is expensive and costs are rising as the number of people living in the UK increases. The people in charge of local health care budgets (clinical commissioning groups, CCGs) aim to make sure public money is spent wisely and to lower costs wherever possible. Some CCGs have said that people who are smokers may not have hip or knee replacement operations. However, although we know that smoking damages health, we do not know if it makes a difference to how well someone gets on after this type of operation.
This study will compare smokers, ex-smokers and non-smokers who have had hip or knee replacements to see whether there is a difference in how well they get on after surgery. We will see whether there is a difference in the numbers of infections, heart attacks, stroke, deaths, blood clots, poor wound healing, and need for further surgery. This information can then be used to help decide whether patients should stop smoking before a hip or knee replacement. Further work might then look into whether this would save money for the NHS.
At this time of unprecedented financial pressures on the NHS, some clinical commissioning groups have stipulated that patients must stop smoking before routine total knee or hip replacement surgery (TKR or THR). However, there is little evidence to support these actions. Associations between smoking and poor clinical outcomes are generally accepted, but the specific impact of smoking on outcome after TKR and THR is poorly understood.
This study will use data from CPRD patients who underwent TKR or THR to compare clinical outcomes and healthcare costs in smokers and ex-smokers to those of people who have never smoked. The outcomes of surgery are: infections (wound, respiratory, urinary, and skin), wound dehiscence, myocardial infarction, stroke and transient ischaemic attack, deep vein thrombosis and pulmonary embolism, poor wound healing, readmission to hospital for any cause, analgesic use, septic arthrodesis, surgical revision and mortality. We will also evaluate 6 month change in Oxford hip or knee score and EQ-5D, and by combining these with mortality outcomes, quality-adjusted life years (QALYs). To do this we will use a retrospective cohort study design. Regression modelling will be used to describe the association of smoking on outcomes. Logistic regression will be used for binary outcomes (infections, wound dehiscence, myocardial infarction, stroke and transient ischaemic attack, deep vein thrombosis and pulmonary embolism, poor wound healing, readmission to hospital for any cause, analgesic use, septic arthrodesis), survival models (Cox and Fine & Gray) for time to event outcomes (surgical revision and mortality), and linear regression for continuous outcomes (Oxford hip and knee scores, EQ-5D, QALYs and costs). Confounding variables will be adjusted for. Missing data will be considered and sensitivity analyses performed using multiple imputation techniques.
Health Outcomes to be Measured:
Primary outcomes are post-operative complications, including the incidence of medical and surgical complications:
- Post-operative infections including septicaemia, pneumonia, urinary tract infection and cellulitis
- Length of hospital stay
- List of medications patient is on when discharged home
- Details of analgesia including duration
- Cardiovascular events
- DVT/Pulmonary embolism
- Wound dehiscence and/or wound infections
- Septic arthritis in replaced joint
- Readmission to hospital for any cause
- Revision of surgery
- Inpatient hospital costs (HRG) - Outpatient hospital costs (HRG)
- Costs of GP visits
HES Admitted Patient Care;HES Outpatient;HES PROMS (Patient Reported Outcomes Measure);ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation