In ancient China wealthy emperors paid their doctors only when they were well, and penalised them when they fell ill. Being sick simply meant the medicine was not working, being well meant that the doctor was doing his job. A healthy outcome was the goal.Click here
for bigger graphThis myth, however hard on doctors, is not a million miles away from thinking today. Everyone wants to be healthy, though with budgets ever tighter, price is an object. Public attention is focusing ever more on actual health care outcomes as a result.The achievements of health care in the OECD are indisputable: we live longer, and diseases that years ago were pandemic killers are either under control or eradicated. But no one is under any illusion that health care comes cheaply. Spending on health care in the OECD area is nearly 9% of GDP, with as much as 15% in the US and around 10% in several other OECD countries.Two prospects seem certain to influence health care policy in the years ahead: costs, which are set to rise due to technology and ageing populations; and public demand, as older populations expect ever better outcomes from their health care services.No longer are people satisfied by looking at health in terms of, say, spending per head or the number of hospital beds per 100,000 patients. Such measures say more about economics and management efficacy than about actual health care performance itself. People are increasingly critical of shortcomings in their care systems. They are well-informed, at least anecdotally, about performance in other countries. In a world of scrutinising electorates and rising costs, outcomes are ultimately how systems, and policy efforts, will be judged.Do people get all the care they should when they visit their doctor? How good are survival rates after serious surgery? What are the best practices and how do other exogenous factors in the population affect results?Right now, such are the gaps in national reporting on data that meaningful, constructive answers to these questions are hard to give. That is why quality measurement is finding greater prominence on national agendas for monitoring and reporting, and is even mandated by some legislatures. Some countries have to improve health care for specific subpopulations, but lack the tools of measurement and analysis to assure delivery of quality care across regions or groups.This is broadly where the OECD’s Health Care Quality Indicator Project comes in. According to Edward Kelley, who has been co-ordinating the work since 2005, it is the leading long-term project ever undertaken to collect and report quality data for international analysis. Working with government experts from more than a score of countries, the aim of this collaborative effort is to etch out commonly acceptable ways of assessing treatment performance and reporting the data.The idea is not entirely new: the Commonwealth Fund and the Nordic Council of Ministers had also produced indicators. But the new OECD work goes beyond these two studies by covering more countries with more comprehensive data.
Did you know that one in five people die within a year of sustaining a hip fracture and that delays in excess of 48 hours doubled the risk of death within a year?
The initial report, which can be found at www.oecd.org/health
(see references), consists of data on 13 key indicators from 23 countries–Cancer screening and survival
: breast cancer survival; mammography screening; cervical cancer survival; cervical cancer screening; colorectal cancer survival; Vaccination and incidence of vaccine-preventable diseases
: incidence of vaccine-preventable diseases; coverage of basic vaccination, aged 2; influenza vaccination for adults over 65; Respiratory disease
: asthma mortality rate; Cardiovascular disease
: heart attacks, called acute myocardial infarction (AMI), 30-day case fatality rate; stroke, 30-day case fatality rate; Timeliness of care
: waiting times for hip fracture surgery; Preventable risk factors
: smoking rates.The report is structured so as to help readers compare performance and assess the validity of each indicator in terms of its clinical relevance, its relation to quality and its reliability.The report also sets out concerns and difficulties encountered in gathering data. There may be missing data on some indicators because of poor survey response, for instance. Different ways of national reporting must also be factored in. For example, asthma mortality can be difficult to classify, causing possible under-reporting. And in heart attacks, in-hospital deaths account for 90% of all deaths, but there are variations. Some countries track patients after discharge, and so can provide a true 30-day picture, whereas others can only supply in-hospital mortality.On the actual data, the facts presented to justify the importance of each indicator are an education in their own right. Breast cancer represents 8.6% of all cancer deaths and one in nine women will contract it, we read. However, the average 5-year survival rate is around 70% in Europe, higher than for other major cancers in women, such as cervical, with its 60% rate, or just 40% for colorectal cancer. Moreover, the median 5-year survival rate for breast cancer in the countries taking part in the HCQI Project is 81%. In Canada, cancer is the third highest health care cost, and in the UK breast cancer cost some £243 million ($392 million) in 1999. As for heart attacks, these represent 18.7% of all deaths and are the largest killer. Stroke accounts for 11.5% of all deaths. The cost of hearth attacks and stroke combined was some $360 billion in the US in 2004.Surgery for femur injury is an indicator that may take some readers by surprise. Yet, did you know that one in five people die within a year of sustaining a hip fracture and that delays in excess of 48 hours doubled the risk of death within a year? Across OECD countries, 73% of patients 65 and over do receive such timely treatment–a proportion that could be much higher given that some countries are operating on over 90% of these patients within those 48 hours.As for diabetes, the report pays this high-priority disease close attention, mainly to show the patchiness of the available data, the different policies towards its treatment, and differences in definitions–over glycemic control thresholds, for example. Solving such differences is part of the goal of the Health Care Quality Indicator Project and the hope is that diabetes will be included as a valid quality indicator for the future. After all, as the report points out, diabetes mellitus accounts for 2% of all deaths and 8% of legal blindness in the US, as well as thousands of amputations.The report is far less ambiguous about another indicator, though: the smoking rate. In the US, smoking results in more deaths per year than AIDS, alcohol, heroin, cocaine, suicide, homicide, fires and motor vehicle crashes combined, the report notes. Even here there are definitional questions, for instance about whether pipe smoking is counted, but in general, there is broad agreement on how to compare the smoking rate. In four countries, the rate is below 20%, but above 30% in three more.The data in the report are arranged in such a way as to reduce the temptation to produce rankings, as these could be misleading. Moreover, finger-pointing can be counterproductive by discouraging the country collaboration that the project needs, to check how indicators are evolving over time.
For now, the authors’ emphasis is on testing the limits of the data comparability, what this reveals about health care systems and how to provide the best quality of care to patients.
As data is gathered and updated, more lessons will be learned, more gaps plugged and new telling patterns revealed.
Take heart attack recovery rates. The initial report on 30-day survival rates shows that patients have a greater chance of dying from an acute attack in some countries than in others. Naturally, one might conclude that this reflects variable quality in angioplasty. However, numerous other factors, such as country obesity rates, come much closer to explaining these variations across countries.Despite these dangers, the data still show patterns from which useful judgements can be drawn, so opening the way for further policy investigation. In stroke rates, for instance, important differences show up in ischemic stroke outcomes, where blood flow to the brain is interrupted, versus brain haemorrhage. Iceland, for instance, shows one of the lowest mortality rates from haemorrhaging, but one of the highest in ischemic stroke.Also, consider breast cancer again. Ireland and Denmark are examples of countries showing relatively high mortality rates: 33.4 per 100,000 population for the Danes in 2000, 31.1 per 100,000 for the Irish in 2001. Denmark’s relative survival rate was 77% in 1991-95, 73% for Ireland in 1994-98. These countries seem to perform less well than the US and Finland, for instance, where mortality rates were respectively 22.4 (2001) and 19.5 (2003); survival rates averaged 88.9% for the US in 1998-2001, and 85.6% in Finland.Could screening be a factor? A look at the mammography screening data in the report suggests that this link is more complicated than it might seem, since Ireland’s screening rate is higher than in the US or Finland. However, in other countries such as Finland, Norway and Sweden, high screening rates have gone along with high survival rates. Finding a cancer is only the first step in treating it, and differing treatment practices and rates of technology adoption may also play a part in differences across countries in cancer survival rates. The positive is that these indicators start to shed light on such questions.Finding answers becomes feasible as more comparable evidence is collected. New indicators that will be reported by the HCQI Project in its next working paper include retinal eye exam rates for diabetics and preventable hospitalisation rates for adult asthma. In addition, the project is developing groundbreaking new indicators for patient safety, mental health and primary care and prevention in 2007 and 2008.Building and comparing data sets is only part of the battle in improving health care. As leading international expert Sheila Leatherman pointed out in the OECD Observer five years ago, we live in an age of improving technology and information mass, yet many doctors even in the wealthiest OECD countries report that their ability to deliver quality care has worsened, rather than improved.However, the devil is in the detail and it is our hope that better comparisons between countries, focusing on actual outcomes, would help change mindsets. As data is gathered and updated, more lessons will be learned, more gaps plugged and new telling patterns revealed.In other words, quality indicators and quality information go hand in hand. Together they empower better decisions, and with the right policies and willpower, can produce better health care for all. With OECD Health Care Quality Indicators, we will be able to decide if our health systems are indeed doing their job. How the emperors of ancient China would have approved of that. Rory J. ClarkeReferences
Mattke, S. et al. (2006), Health Care Quality Indicators Report, OECD Paris.OECD (2006), OECD Health Data, Paris, see www.oecd.org/health
Kelsey, Tim (2001), “Fostering quality healthcare”
, in OECD Observer
No. 229, November.Leatherman, S. (2001), “Measuring up: Performance indicators for better healthcare”
, OECD Observer
, No. 229, November.UK Centre for the Measurement of Government Activity (2006), “Public Service Productivity: Health”, February.©OECD Observer
No. 257, October 2006