European Research Group on Health Outcomes
(ERGHO)*
To leave a message for ERGHO, you may use this form or "mailto"
[ERGHO Homepage] [ERGHO Contact Points] [ERGHO Webmaster]
ERGHO Statement:
"Choosing a Health Outcome
Measurement Instrument"
Measuring health outcome is a process in which a standardised attempt is made to observe an often complex clinical picture. Health Status Measures are often used for this purpose. We have produced this short leaflet to identify the key guidelines, preconditions and warnings necassary before selecting such an instrument, because it has become clear that the wrong choices are often made as a result of a lack of basic guidance. Some of the suggestions in the leaflet arise from a meeting of experts in London in 1994, sponsored by the European Union BIOMED programme, although this text is solely the reposibility of the ERGHO group.
1. First match an instrument to your needs
A. choose the LEVEL OF OBSERVATION:
Is it the individual patient,:perhaps the measurement of an individuals change due to intervention, usually a treatment, or the observation of health over time?
Is it groups of patients which are the focus of interest: patients from a particular age-group, with a specific disease, or patients submitted to a certain intervention. The utility of the intervention is the main interest
Or is the area of interest the general quality and cost-effectiveness in different care-systems: comparisions may be proposed of quality of care between different systems, say between primary and secondary care.
B. formulate and describe your AIMS
It is important to understand the purpose of the outcomes initiative so that appropriate selections can be made from the many available measurement instruments. Particularly it will be necessary to determine whether the effect of an intervention is to be measured or whether a descriptive assessment is to be made.
short, feasable and reliable instruments are recommended if care providers are to use them in their clinical work.
broadly validated instruments which have been used in other studies are required if it is intended to describe the health status of a defined population or a specific disease category. This is particularly important if there is an intention to compare results with other studies or initiatives.
A combination of condition-specific and generic instruments are recommendend, especially if it is intended to attempt to relate the health outcome information to costs, intervention techniques or structural components of health care. Be sure the domains covered in the instrument relate to the predicted health effects of the proposed treatment or intervention.
C. define the CONTEXT of your interest
Describe the domain of health in which there is an interest. From a biomedical viewpoint of health, clinical signs and symptoms, severity of illness, pain level and side effects of medication may be important. Often there is an interest to follow-up, and to measure change over time.
But it is increasingly common to wish to take a sociological view of health, in which the functional status of populations are of importance. This may call for evaluations of domains of health such as an ability to undertake daily tasks.
More broadly still, some evaluators have chosen to consider what might be called the humanistic perspective of health which can include energy, behaviour, patient concerns for autonomy. Clearly the further one tries to go in measuring the human condition the more difficult the task becomes and initial outcomes assessment initiatives should aim to be simple.
D. consider the SOURCES of your information:
The choice of the data source will depend on the nature of the desired information. As information is never neutral, choose if you are going to ask patients, care givers, doctors or other care providers.
E. who are the USERS of your information:
Choose the data source according to the people who will use the collected information. Politicians, doctors, patient organisations and patients can be interested in health status and outcome data, but the value of the data is related to the source.
2. What is your aim. Do you want to describe, to compare or to evaluate health outcomes
The selection of your instruments is highly related to the endpoints of your project. What do you want to use it for ? The psychometric qualities of the choosen instrument must be able to support your endpoints. There are three principal uses:
An health status measure can be used as an INDICATOR, measuring the situation at one point in time. The endpoint is descriptive. Besides validity, reproducibility and specificity for the choosen condition are important.
An health outcome measure can be used as a COMPARISON, relating differences at different points in time: before and after intervention. For this type of action sensitivity and responsiveness to change are important. That is, the measure must be able to register small changes in the health of patients over time.
Health outcome ASSESSMENT implies, besides the measurement of outcome, an attempt to use the information through feedback to practitioners. In order to use that information some understanding of the process of care is required so that the project data base must also include process data.
ONLY AFTER YOU HAVE DEFINED YOUR AIMS AND PURPOSES AND DECIDED ABOUT THE WAY TO USE THE RESULTS, CAN YOU MAP THESE AGAINST A NUMBER OF INSTRUMENTS AND SCALES TO DECIDE ON WHICH MOST CLOSELY SUITS YOUR PURPOSE.
3. Condition specific, dimension specific or generic instruments?
Which type of instrument ?
In general, a condition specific measure will have a narrow focus but will contain considerable detail in the area of interest. If you are interested, say, in one disease condition, and the assessment is mainly on symptoms and function, then use a condition specific measure. Remember that you also need to consider the limitations of the chosen instruments.
If specific domains are your point of interest, like daily functioning or mental wellbeing in different health problems or population, then use a dimension specific instrument.
if you are interested in general health or in the interaction between different conditions , or if you are interested in populations which may include healthy people, use generic instruments
if you consider the influence of other diseases or conditions can influence the results of the problem or disease of your interest, combine disease specific and generic instruments.
How many instruments ?
No one instrument may prove satisfactory for all purposes . You may feel a need to combine instruments because a reasonable instrument does not exist. How much of this work you do depends on your resources. Don't forget practicalities: the necessary time to fill in questionnaires, the cost of mailing and of analysis. Be reflective on your target group: not every instrument suits children or elderly people.
But be aware;
Use the instruments in their original form, do not change them or do not use only parts of them: validation only refers to the complete instruments
Be carefull with translated instruments: cross-cultural validation needs to follow strict rules.
4. Health measurement is essentially evaluative or subjective rather than objective.
Health measurement is not objective, like measuring blood sugar, length or gender. Often the information is rather subjective, which means that it is difficult to control by a third party.
It is good to consider most judgements about health related aspects as evaluative data, because they include an interpretation by a person. This person can be the patient him/herself, significant others like family or care-givers, the nurse, the doctor or other health care providers. All information will pass through a filter of personal interpretation, even if coming from the person himself. A person may adapt his answer according to expected benefit, social willingness and actual mood.
Patient data are not more objective than care providers data.
The construction of each instruments itself works as a filter. Dimensions like "daily functioning" are narrowed towards some very explicit questions. The formulation of questions differ : they ask for factual information (can you do this?), or intentional information (could you do this?) or impact (does it bother you?) . They rely on observation (have you done it?) or on interpretation (can you do it better now?) .
Reflect on the type of information you want. For instance, questions on performance or capacity to perform, relying on observation or on interpretation, provide different information. Be carefull when comparing information coming from different formulations.
5. Don't forget the patient
Invest enough in the patient's reaction to the outcome assessment. It is probable that patients will have to be considerably involved, especially in a project which involves people with chronic diseases. Questionnaires take time to complete, however simple they may appear to be to a professional. The quality of the information will be highly dependant on the willingness of a patient to cooperate.
The best instrument does not exist. Every instrument has its own strengths and weaknesses, and its own particularities. You need a very good insight in the measurement capacities of your instruments and to be familiar with it.
If you are not familiar with the instruments that best suits your plans, seek help from someone who knows about it already.
Appendix:
List of Health Outcome Instruments
of interest for international use in ambulatory care
Generic instruments
COOP/WONCA charts
Sickness Impact Profile (SIP)/Functional Limitations Profile (FLP)
RAND SF 36
Duke Health Profile (DUKE)
EuroQol
MOS 20
Nottingham Health Profile
RAND General Health Perception Questionnaire (GHPQ)
Dimension specific instruments
Barthel Index
Index of Independence in Activities of Daily Living
Frenchay Activities Index
General Health Quest