Information to decision-makers

The Swedish Hip Arthroplasty Register (Note – not to be confused with Hips National (Rikshöft), which records hip fractures in a separate register) started in Göteborg in 1979 and is Sweden’s second oldest National medical Quality register.  The Register rapidly achieved national coverage and has long been an example for other national and international medical quality registers. It was the first national register to use modern IT functions and became web-based as early as 1 January 1999.  All input and feedback are now handled via the Registry’s home page (annual reports are available in pdf format and in printed form).  The Registry’s annual feed back to the participating clinics has led to a continual and dramatic improvement in the national long-term result following hip implant ?replacement surgery.  Sweden has today the world’s lowest reported reoperation frequency.  Compared with results in other parts of the world the Swedish medical care system has during the past 10 years been saved about SEK 1.5 thousand million in direct medical care costs.  The immeasurable indirect costs are many times greater.

Outcome variables in the register

• Implant survival according to Kaplan-Meier analysis.  Failure definition is surgical intervention with replacement of one or more components.  Reported mainly 5 and 10 years after primary operation.  This is the internationally most used comparison variable.
• Reoperation (all forms of further surgery following a primary operation) within 2 years.
• Patient satisfaction.
• Pain relief according to visual analogue scales.
• Health-related quality of life and health gain measured with EQ-5D.
• 90-day mortality.
• Degree of coverage at hospital level.
• Procedure frequency/100,000 inhabitants at county council/regional level.  Data presented both as production figures and consumption figures.

In-depth analyses

Apart from the annual outcome variables the Register carries out three to four in-depth analyses published in the year’s annual report.  One example of such a deep in-depth analysis is the significance of the genus perspective for results after hip implant ?replaclement surgery.

Open reporting

In 1999 the Swedish Hip Arthroplasty Register started to report the variable implant survival (as above) openly in the annual reports and by hospital.  Starting with the 2006 Report eight variables have been reported openly at hospital level (see section Follow-up of Activities below).

The complication rate is generally very low following modern total hip implant surgery and should therefore be assessed with great caution.  There is a risk that a biological and random variability can ‘point the finger’ at one clinic one year and at another the next.  These variables can really only be assessed over time, that is, if there are clear persistent trends.  Clinics with a wait-and-see attitude, avoiding surgical interventions for certain complications, are not recorded in the database either.  The main object of the openly-reported variables is not to ‘point the finger’ at the individual clinic but to initiate local in-depth analysis and work for clinical improvement.

Individual surgeons’ results will never be published, chiefly because of the problems of ‘case-mix’ (see section below).

National quality indicators

The Swedish Hip Arthroplasty Register reports three indicators in the publication "Öppna jämförelser" (Open Comparisons):

• 10-year survival of implants according to traditional Kaplan-Meier statistics
• Short-term complications with re-operation within 2 years
• EQ-5D gain 1 year after operation

The "case-mix" problem

‘The case-mix factor’ (patient composition) is probably the single largest factor leading to misinterpretation of register results.  The open reporting of results of various registers has attracted great attention in the media where direct errors have been published with ‘banner’ headlines.  This has afflicted patients and resulted in great and unnecessary disquiet.

The register results are in certain sections hard to interpret both outside and within the profession.  Decision-makers in Swedish medical care have over the years criticised certain register operators for not reporting all their results openly.  The reason for this is not only a reluctance to report possible poor and/or varying results, but depends on these problems of interpretation.  There is a consensus on increasing the openness from the registers but this is accompanied by an onerous pedagogical task.

In all reporting it is necessary to present the patients’ demographical profiles in detail.  A national register study includes all patients, with a large spread of risk factors and great differences regarding different hospital profiles.  More serious cases are referred to larger clinics and special units and specialist surgeons who, for example, are called upon to operate on patients with greater operational risks. This gives an increased complication frequency and mortality, and greater demands on intensive care resources, which in turn also drive up the costs.

In surgical specialities, and particularly in orthopaedics, structural changes in Swedish medical care have created more and more elective units with considerable production.  These units can often show good results and low costs, often compared with other units.  Decision-makers should in such contexts be aware of a selection of easier standard cases which often form the basis of these units’ production (also including most of the private entrepreneurs), which in fact need not be cost-effective in a cost/utility analysis (see section on Health Economics).

Within each register some form of ‘case-mix’ indicator should be developed.

Cost analysis

An important health-economic detail is the definition of cost:

• Direct costs
  • Direct medical costs (care costs)
  • Direct non-medical costs, for example local council costs such as assisted transport, home-help
• Indirect costs such as sicklisting and early-retirement pension

In many disease groups the direct non-medical and indirect costs are many times greater than the direct medical costs.  The social cost of diseases of the organs of locomotion for example in most cases represent 80% of non medical care-related costs.  A common error when measuring costs of a disease state or a period of care is only to calculate part of the total cost, i.e. certain cost bearers are ‘forgotten’ in the analysis.  In the present Swedish social insurance, health and medical care system with local-council responsibility for total costs, analyses are difficult, time-consuming and in certain cases impossible, since transfers and openness between the various systems are absent.

A complete individual-based cost analysis is expensive in itself, burdensome and time-consuming.  An orthodox health economist requires that such an analysis includes direct and indirect costs and must include all the details of an entire course of care.  This type of analysis can only be done with very small patient material.  Larger-scale analysis must rely on the existing compensation systems in medical care.  The CPP databases are the most detailed of these systems.  CPP (cost per patient) is available today at some 25 hospitals in Sweden.  When CPP is implemented nationally and standardised, national cost analysis (for direct medical costs) for all disease groups will be easier to carry out.

Health-economic analysis

Health economics is a science that applies theories and methods within the discipline of economics to the health area and studies how resources are allocated to, and within, health and medical care.  In the present situation where economic development in no way matches medical-technological development, the commonest starting point in health-economic analysis is that resources are insufficient and one should seek alternative areas for use of the resources.  In a health-economic analysis one or more treatment strategies for a defined patient group are compared.  Measuring one cost only – use of resources – is quite meaningless unless the consequences – utility – of the costly measure is measured at the same time, the cost effectiveness of the measure  estimated.  An adequate health-economic analysis can and should therefore be included in any priority-setting and resource distribution.

The Swedish Hip Arthroplasty Register has since 2002 further developed so that it also includes health-economic analyses.  In a cost-effectiveness analysis it is important to capture all costs and effects that may be related to the treatment strategy under examination, regardless of when they arise.  To be able to run a cost-effectiveness analysis, information is needed on mortality, complications, health-related quality of life and costs related to the measure both in the short term and the long term.  All these variables except costs exist in the Registry’s databases.  The costs are obtained through matching with existing KPP databases.

The commonest quantitative health-effectiveness measure used today in modern health economics is quality-adjusted life years (QALY). This is a comprehensive measure comparable across different studies.  A patient-related measure of cost effectiveness is cost utility: cost for the treatment (health gain x time) gives cost per QALY gain.  In the disease group in question, the costs for a hip implant operation/(EQ-5D gain x duration of treatment).  The theoretical comparison group for this model is the untreated (non-operated) patients with hip disease.  In a long-term analysis this calculation in an ideal situation should be discounted for inflation, complications costs, death and age.

The above figure shows two possible disease courses with two different interventions.  Intervention A leads to greater utility (quality of life) and a longer life for the patient.  This ‘area gained’ is thus QALYs gained.  If there is a cost difference between interventions A and B, the cost-utility effect ( cost effectiveness) can be calculated.

In summary: health and medical care should not be governed by cost analyses but by cost-effectiveness analyses

Follow-up of activities

County councils and regions have traditionally followed up their activities with productivity measures and economic (cost) measures.  There is most often no systematic link to the actual outcome and utility of the activity.  Follow-up of activities using selected quality indicators mainly from the national quality registers offers important material for development work.  Follow-up will give a basis for discussion and deeper analysis.  Using the indicators, circumstances and areas can be identified for closer study regarding causal connections and opportunities for improvement.  A further decision was taken in June 2004 by the then federal board (er.. er… ?*) (AO4:32): that county councils and regions should improve and develop instruments for systematic follow-up and comparison of medical results, costs, resource use and practice, together with patient utility.

In 2005 the Swedish Hip Arthroplasty Register carried out a pilot project in collaboration with the Strategic Development Unit in the Western Götaland Region termed "Att börja på ny kula" (Starting Afresh)’.  (The report is >>here.)  The aim was to elucidate procedure frequency, patient demography by hospital in the Region Västra Götaland Western Götaland Region, complications frequency, patient utility and satisfaction, cost-utility effect (cost/QALY gained) for patients undergoing total hip implant surgery.  The ultimate goal of the project was, via complete follow-up of activities, to meet all four dimensions in what is termed the value compass.  For the first time in Sweden the compass could be used fully to describe the outcome after treatment of a well-defined disease group within a region, which made it possible over time to increase the value of care input for patients, i.e. leading to work for clinical improvement. 

The value compass is a metaphor of four dimensions at the points of the compass.  All are equally important.  The idea is that there are results in four directions that need to be examined, described and improved over time for the patients.

The project was carried out by matching following databases at individual and hospital levels:

• The Swedish Hip Arthroplasty Register hip register database (Western Götaland Region part)
• VEGA (the region's care database)
• KPP-databases in the region

Activity development and work for clinical improvement

The value compass is comparable with a balanced control card – in theory there is an improvement potential in all directions. The Swedish Hip Arthroplasty Register publishes its open variables in eight comprehensive tables.  For an easy-to-read picture of the activity follow-up for each clinic we have since 2007 published a clinic-specific value compass.  The aim of the modified value compasses is that each clinic is able rapidly to gain an appreciation of the clinic’s result vis-à-vis the national one, and to identify areas where its outcome has been poorer.  If such is the case, the respective table should be studied and an analysis initiated to start local work for improvement and activity development.  As in all other register interpretation it is imperative to judge the clinic’s demographic profile at the same since this may explain an expected or a deviant result.

In the value compasses the national result for the eight variables is shown in red.  Each clinic’s corresponding value is shown in green.  Limit values are set to the greatest and the smallest value of the variable in question ± 1 SD.  The poorest value for the variables is ascribed origo and the best is at the periphery.  Clinics showing red fields indicate a value below the national average for the variable in question.  The outcome may be studied in detail in the relevant table in the Annual Report.  Clinics with ‘much’ red are recommended to run a local in-depth analysis.

Matching with databases at Statistics Sweden (SCB) and with health data registered at the National Board of Health & Welfare can give a number of background variables that illustrate socioeconomic factors (ethnicity, language problems, educational level, income, long-term sicklisting etc.), intercurrent diseases and consumption of medication.  A described merged database can with great statistical certainty elucidate parameters that show whether Swedish hip implant surgery is fair regarding aspects of gender equality and equality.  In addition a number of predictors can probably be identified that can gain major significance for current work on guidelines for osteoarthritis and also lead to adequate care programmes and activity improvements for this disease group.  The result of the analysis can also be of great significance for health-economic analysis of hip implant ?replacement surgery.

The Hemi-prosthesis database

Until 2005 the Swedish Hip Arthroplasty Registry never recorded what are termed hemi-arthroplasties (an operation in which only the ball of a joint is replaced by a femur implant ?prosthesis).  The indication for this surgery is almost exclusively acute fracture of the neck of femur, or complications from earlier-treated fractures of this type.  For some years we have had a modified treatment algorithm for femur fracture.  Most hospitals in Sweden now operate on misaligned fractures with primary hemiarthroplasty. 

Example of  hemiarthroplasty (left) and total arthroplasty (right)

For the above reasons the annual production in Sweden of hemiarthroplasties has increased by a factor of ten – from about 350 per year to 4,500 per year. This has occasioned the need for a National Quality register.  Starting on 1 January 2005, therefore, all hemiarthroplasties have been recorded in a separate database in the Swedish Hip Arthroplasty Register.

The Swedish Hip Arthroplasty Register's Annual Reports

Should you wish to know the result for your hospital or county council/region in detail you are recommended to read the most recent annual reports. They are available to read or to download from this homepage.

Latest annual report (in Swedish)