The aim of this article was to describe the design of an international audit of the prevalence of care problems in different healthcare sectors using identical methodologies.
The aim of this article was to describe the design of an international audit of the prevalence of care problems in different healthcare sectors using identical methodologies.
Audits, defined as a monitor of quality of health care, are increasingly applied in many countries as a strategy to improve professional practice and quality and safety of care. A prerequisite to enable a reliable comparison of quality of care audits is the use of identical instruments and methodology.
Annual cross-sectional multi-centre point prevalence survey.
This international prevalence measurement of care problems in hospitals, care homes and home care is performed in the Netherlands, Austria, Switzerland and New Zealand. This study is based on a prevalence measurement of care problems originally performed in the Netherlands. For each care problem (pressure ulcer, incontinence, malnutrition, falls and restraints) at patient level, next to patient characteristics, data are gathered about the prevalence, prevention and treatment of each care problem. In addition, at ward/department and institution level, specific quality indicators are measured related to the care problems. After the measurement, institutions enter their data into a web-based data-entry program. Institutions receive an overview of their own results and results at national level to enable a process of benchmarking.
A uniform way of measuring the prevalence of care problems internationally is a significant step forward in gaining insight into the quality of basic care in different healthcare settings in different countries and may lead to more awareness and improvement programmes.
Audits, defined as a monitor of quality of health care, are increasingly applied in many countries as a strategy to improve professional practice and quality as well as safety of care (Electronic Patient Record (EPR), Minimum Data Set-Resident Assessment Instrument (MDS-RAI), Dutch National Prevalence Measurement of Care Problems) (Casalino et al. 2003, Grol et al. 2004, Jamtvedt et al. 2006). These audits express approaches on quality improvement that includes evidence-based medicine/nursing, accreditation and (external) accountability. Patient safety and, more in general, quality of care are considered more and more important with reference to the reduction in preventable adverse events in patients in healthcare organizations and adherence to practice guidelines on care processes (Degos & Rodwin 2011).
It is estimated that, in European Union (EU) Member States, between 8–12% of the patients admitted to hospitals suffer from adverse events while receiving care (European Commission (EU) 2010). As patient safety is a serious concern, the Council of the EU recently adopted recommendations on patient safety. These recommendations involve more reporting of patient safety events, education and training of healthcare workers, focusing on patient safety (European Commission (EU) 2010). According to this, fulfilling the basic care needs and adequate tackling of relevant care problems (e.g. malnutrition, falls) are not only fundamental but also important for the quality of life and health of patients and indeed to their safety (Comondore et al. 2009).
Quality of care and patient safety can be measured in different ways and can be considered from different perspectives, e.g. client's experience (Cleary & Edgman-Levitan 1997, Weinberger 1999, Attree 2001), perspective of the professionals and also by focusing on outcomes of care (Meadows et al. 1997). When focusing on outcomes, the prevalence of care problems such as pressure ulcers, malnutrition, falls and restraints (Bours et al. 1999, Green 1999, Centraal BegeleidingsOrgaan (CBO) 2002, Bours 2003, Centraal BegeleidingsOrgaan (CBO) 2004, Neyens et al. 2006, Heinze et al. 2007, Neyens 2007, European Pressure Ulcer Advisory Panel (EPUAP) & National Pressure Ulcer Advisory Panel (NPUAP) 2009a,b, Meijers 2009, Meijers et al. 2009a,b) may be used as relevant outcome measures for the quality of care and patient safety.
Measuring quality and safety of care in health care take place at institution level, at country level and at multi-country level. EU policy increasingly focuses on multi-country studies to enable comparison of healthcare outcomes based on country and cultural differences (Riedel & Kraus 2011).
Finding reliable and valid instruments to compare data between countries is important in this field of research. Available instruments that provide healthcare institutions with such data are the MDS-RAI (Hutchinson et al. 2010) and nutritionDay (prevalence (risk) for malnutrition) (www.nutritionDay.org). These instruments provide a partial insight into the prevalence of care problems and patient characteristics. To assess quality of care more integrally, Donabedian (1988) – one of the important designers of healthcare quality theories – states that it is essential not only to focus on outcome (prevalence of care problems) but also on structural and process indicators. According to Donabedian (1988), an improvement in structure and process may lead to better outcomes. Donabedian's (1985) framework of quality of care offers a good model to develop a relevant measurement instrument to assess the quality of care and patient safety.
The study presented here has been developed according to this framework (Donabedian 1985) and, therefore, it includes prevalence rates as representatives of healthcare outcomes and structural aspects of care, for example, availability of adequate personnel, guidelines and process factors of care, being preventive measures and treatment interventions undertaken to deal with care problems as pressure ulcers, malnutrition, falls, restraints and incontinence.
This study, LPZ-International, is based on a longer existing prevalence measurement of care problems performed in the Netherlands, National Prevalence Measurement of Care Problems [in Dutch: Landelijke Prevalentiemeting Zorgproblemen (LPZ)] of Maastricht University (UM). The LPZ is an annual, independent prevalence measurement, which has been performed in Dutch health care since 1998 (Halfens et al. 1997). Initially focusing on pressure ulcers, the measurement extended to other care problems as malnutrition, falls, restraints and incontinence. Yearly, over 400 institutions participate in the LPZ with over 40,000 patients, from hospitals, care homes and home care (Halfens et al. 2007, 2008, 2009, 2010, www.LPZ-UM.eu).
The aim of the study was to get insight in the quality of care by collecting data on the prevalence, treatment and quality indicators of care problems in different healthcare sectors (hospitals, care homes, home care) in different countries (the Netherlands, Austria, Switzerland and New Zealand), including pressure ulcers, malnutrition, falls, restraints and incontinence using the same definition, screening instruments and methodology.
LPZ-International is a cross-sectional multi-centre point prevalence survey and is conducted annually on 1 day in different healthcare settings in the Netherlands, Austria, Switzerland and New Zealand. Central coordination of the study is executed by the LPZ project group at UM. In each country, the coordination of LPZ-International is carried out by a national project group led by a national coordinator. In Austria, the Institute of Nursing Science at the Medical University of Graz, in Switzerland, the Berner Fachhochschule Gesundheit and in New Zealand, the Graduate School of Nursing Midwifery and Health at Victoria University Wellington coordinate LPZ-International. The national coordinators meet yearly in an international research group meeting to make appointments concerning the yearly measurement (changes in questionnaires, publications, co-operations). Care problems assessed are as follows: pressure ulcers, malnutrition, falls, restraints and incontinence.
In participating countries, all healthcare organizations are invited by (e)mail including a flyer, through publications in professional journals and using connections from other projects to participate voluntarily in this international prevalence measurement. Institutions pay for every patient taking part in the measurement. To obtain representative results, institutions are encouraged to measure in all departments including all patients present on the day of the measurement. Patients are only included if they or their legal representative give informed consent. No further exclusion criteria were used.
The questionnaire of the original Dutch study was developed by consulting experts and based on literature reviews (Bours et al. 1999, 2000, Hannestad et al. 2000, Rohr et al. 2005, Neyens et al. 2006, Meijers et al. 2010, Thüroff et al. 2010). Definitions and assessment of the care problems used in this study are presented in Table 1.
|Definition of care problem||How assessed?||Based on|
Four categories of pressure ulcers (European Pressure Ulcer Advisory Panel (EPUAP) & National Pressure Ulcer Advisory Panel (NPUAP) 2009a):
Category I: Non-blanchable erythema
Category II: Partial thickness
Category III: Full thickness skin loss
Category IV: Full thickness tissue loss
| ||Bours et al. 1999, Halfens 2000a, Centraal BegeleidingsOrgaan (CBO) 2002, Bours et al. 2003, Bours 2003, European Pressure Ulcer Advisory Panel (EPUAP) & National Pressure Ulcer Advisory Panel (NPUAP) 2009a,b, Meijers et al. 2010,|
Malnutrition has been operationalized and validated (Meijers et al. 2010):
1. Body Mass Index (BMI) ≤ 18·5 (age 18–65) or BMI ≤ 20 (age > 65),
2. Unintentional weight loss (more than 6 kg in the previous 6 month or more than 3 kg in the last month) and
3. No nutritional intake for 3 days or reduced intake for more than ten days combined with a BMI between 18·5 and 20 (age 18–65) or between 20 and 23·9 (age > 65)
| ||Stratton et al. 2003, Konrup et al. 2003, British Association for Parenteral & Enteral Nutrition (BAPEN) 2003, Meijers et al. 2010,|
A fall is defined as an unintentional change in position that results in a person coming to rest on the ground or other lower level. In this study, the prevalence of falls is calculated as having fallen one or more times during the last month (Kellogg International Work Group on the Prevention of Falls by the Elderly 1987)
Asking patient or responsible nurse and/or from patient file:
|Dijcks & Neyens 2002, Centraal BegeleidingsOrgaan (CBO) 2004, Neyens et al. 2006, Neyens 2007|
Physical restraints are interventions such as bed rails, bed-straps, nursing blankets, deep chairs and chair tables, which are used to prevent dangerous or risky situations. Healthcare workers often use physical restraints to prevent falls
| ||Hamers & Huizing 2005|
Urinary incontinence is defined as any form of involuntary urine loss
(Abrams et al. 2003)
Faecal incontinence is defined as involuntary loss of the bowels for 3–4 times a month
Double incontinence is defined as involuntary loss of urine and faeces
Distinction is made between the different kinds of urine incontinence
| ||Brandeis et al. 1997, Hannestad et al. 2000, Teunissen et al. 2004, Rohr et al. 2005 Moulin du 2008|
As the original questionnaire was in Dutch, for LPZ-International, the questionnaires and the instruction material have been translated by professional translators into German, English, French and Italian and back translated and double-checked for nomenclature and cultural differences by the project group in each country and national experts in quality of nursing care. Following this strategy, a situation has been created whereby in every country the same questionnaires are used. The content of the questionnaires is monitored continuously, updated if necessary and also evaluated in the international research group yearly. Changes are made only if they can be substantiated by new knowledge and evidence-based research or due to changes in the field (adapting new standards, launching new preventive or treatment strategies).
The standardized and comprehensive questionnaire measures at three levels: institution; ward/department level for measuring the kind of institution/ward; and structural/organizational indicators. At patient level, patient characteristics are collected and the assessment of the prevalence of each care problem is performed. A specific module has been developed measuring characteristics of each care problem and process indicators (preventive and treatment measures).
At institution and ward/department level, structural indicators for each care problem are assessed with dichotomous answer categories (yes/no) such as the use of a protocol/guideline, availability of educational activities, availability of information brochures for patients/family, having a policy on screening for care problems at admission, and the presence or involvement of professionals (dieticians and tissue viability nurses).
In Figures 1 and 2, structural indicators on institution and ward/department level regarding malnutrition are listed as an example. For each care problem, about ten structural indicators are included both on institution and ward/department level.
Each specific module for each care problem involves questions about the process of care at patient level including preventive and treatment interventions like the use of screening instruments and consultation of experts (Table 2). Per care problem, about four questions are formulated for both preventive and treatment measures. Most questions have multiple answer possibilities. Table 2 gives an overview of how preventive and treatment measures are measured.
|Care problem||Preventive/treatment measures||Validity of prevalence|
|Pressure Ulcers|| ||Centraal BegeleidingsOrgaan (CBO) 2002, European Pressure Ulcer Advisory Panel (EPUAP) & National Pressure Ulcer Advisory Panel (NPUAP) 2009a,b|
|Malnutrition|| ||Stratton et al. 2003, Konrup et al. 2003, British Association for Parenteral & Enteral Nutrition (BAPEN) 2003, Kruizenga et al. 2005, Meijers 2009|
|Falls|| ||Dijcks & Neyens 2002, Centraal BegeleidingsOrgaan (CBO) 2004, Neyens et al. 2006, Heinze et al. 2007|
|(urine) Incontinence|| ||Brandeis et al. 1997, Hannestad et al. 2000, Minassian et al. 2003, Teunissen et al. 2004, Rohr et al. 2005, Moulin du 2008|
At patient level, data concerning the prevalence of each care problem are registered, including some characteristics of the care problems: when and where the care problem started, by whom care problems are diagnosed, whether patients are at risk or not, etc. (Table 1).
At patient level, demographic data such as gender, age, body mass index (BMI), operation (yes/no), number and type(s) of disease (co-morbidity), Activities of Daily Living (ADL), Housekeeping Daily Life-Activities (HDL) and length of stay are registered. Furthermore, care dependency of the patient is assessed by the Care Dependency Scale (CDS) (Dijkstra 1998, Lohrmann 2003). This scale consists of 15 care dependency items: eating/drinking, incontinence, body posture, mobility, day/night pattern, getting dressed/undressed, body temperature, hygiene, avoiding of danger, communication, contact with others, sense of rule/values, daily activities, recreational activities and learning ability. Scores are registered with a 5-point Likert scale. Psychometric testing as content, construct and criteria validity was performed and reliability (interrater, homogeneity and internal consistency) was tested (Dijkstra 1998, Dijkstra et al. 1999, 2000, 2003, 2005). The CDS is validated for different settings and countries (Dijkstra 1998, Dijkstra et al. 1999, Dijkstra et al. 2003, Lohrmann 2003, Dijkstra et al. 2005).
Data at institution level are collected by the institutional coordinator who is pointed out in every participating institution. Data at ward/department level are collected by the head of the ward/department. Data at patient level are obtained either by observation and inspection of the patients, from the patient files or using an assessment instrument like Braden scale (pressure ulcer risk) (Bergstrom et al. 1987) and the Malnutrition Universal Screening Tool (MUST) (malnutrition) (Stratton et al. 2003).
The original Dutch LPZ questionnaire is based on literature search and definitely established after consulting national and international experts (face validity) (Bours et al. 2003, Moulin du 2008, Meijers et al. 2009c). Furthermore, national and international guidelines have been taken into account developing the questionnaire. The included instruments (BRADEN, MUST, CDS) have been tested before for reliability and validity for different settings and countries (Table 2) (Dijkstra et al. 1999, Halfens et al. 2000, Stratton et al. 2004).
To enhance reliability, each patient is assessed by two healthcare professionals (nurses, dieticians or doctors). Of these two, one works at the patient's ward/department and one is a professional from another ward/department. Inter-rater reliability has been tested for hospitals, nursing homes and home care, and found to be good (Cohen's k of 0·87) (Kottner et al. 2009, Meijers et al. 2009b,c).
To conduct the study, Research Ethics Committee approval has been received from the medical ethical committee of a university in the Netherlands (20 December 2007), Austria (9 March 2012) and New Zealand (Summer 2009). In Switzerland, for each participating canton, Research Ethics Committee approval has been received from the specific cantonal ethical committee and from a central Swiss Medical Ethical Committee (4 October 2011).
According to the decisions of the medical ethical committees, in the Netherlands, patients had to give their oral informed consent, whereas the patients from Austria, Switzerland and New Zealand had to give a written consent. Each patient is informed about the course of the measurement. After giving informed consent, patients can stop participating at any time without giving any reason; patients have no obligations what so ever.
On the day of the measurement, the patient will be examined to register whether the patient has one or more pressure ulcers. The skin will be examined at several places. At the same time, the nutritional status will be checked and whether the patient is incontinent, or has recently suffered a fall. These are basic nursing interventions that in fact occur during the daily nursing care. Therefore, the patient runs no risk. The whole examination will take only a few minutes. The nurses will also note several particulars such as age, the reason for admission and particulars on the physical condition. All the characteristics will be noted anonymously and will be treated as strictly confidential.
Each participating institution registers for the measurement voluntarily and by signing the registration form and is obliged to pay for each patient participating in the measurement. For this study, no specific funding has been received to enable the execution of the measurement in the different countries.
The Dutch LPZ project group facilitates each national coordinator of every participating country with all documents and a website (www.LPZ-um.eu) in their own language to enable them to carry out the measurement. In each participating healthcare organization of each country, an institutional coordinator is responsible for organizing the measurement in the institution. All institutional coordinators are trained collectively by the national research group of each country on how to manage the measurement and how to use the printed standardized questionnaires and the specially designed web-based data-entry program. The institutional coordinators train the healthcare professionals, who perform the measurement in the institutions on how to collect data at patient level. The institutional coordinators receive a study protocol and training package with questionnaires (instruction) manuals and guidelines.
For practical reasons, in home care organizations, a representative sample of all clients is drawn and the measurement is spread over 4 days. The healthcare professional primarily responsible for the patient's care fills out the questionnaire during a home visit. To ensure reliability, another independent healthcare professional revisits a random sample of 20 patients per home care organization (Bours et al. 2003, Kottner et al. 2009).
After data gathering is complete, data are entered into a specially designed web-based data-entry program. The results at institution and national level are generated and published in tables on a private part of the website only accessible by personal log-on data by each individual institution. Tables are provided in percentages, in numbers and kind of ward/department.
Furthermore, all data are available as a SPSS-data file for scientific statistical calculations. Each national coordinator received a SPSS-file with their own country-specific data. Each project group writes an annual report describing country-specific data. The project group at UM publishes an international report comparing data from different countries. Statistical analyses are performed and data are checked for outliers and normality. The analyses include descriptive frequency distributions for all variables; differences between groups are tested using chi-square test, student's t-test or variance analyses. Logistic regression analyses and multi-level analyses are used to examine trends in prevalence rates, preventive and treatment measures over the years and to correct for differences among patient groups, sectors and countries.
Although, currently, some data are available about the prevalence, prevention and treatment of care problems in several countries, reliable comparisons are not always possible due to differences in used definitions, instruments, methodology and samples. The study presented here provides the opportunity to compare data between countries in a valid and reliable way, due to the use of a standardized methodology (same measurement procedure, instrument, study protocol). Therefore, it provides insight into differences in the quality of basic care in different healthcare sectors, between different countries and into changes over time, as it involves an annual measurement.
According to Donabedian (1992), differences in outcomes can be explained by differences in structure and process factors. However, differences between organizations and countries do not necessarily involve differences in quality of care. They can also be caused by differences in patient characteristics. Therefore, outcomes of different countries must be standardized or controlled for patient characteristics. Comparisons between countries may be influenced as well by structural factors. Besides individual organizational characteristics as size and care policy, differences in results can be influenced by country-specific organizational, political and cultural differences. Therefore, comparisons between countries may give much information about the influence of the healthcare system, and all factors related to this. For instance, Tannen et al. (2006, 2008) have shown that the prevalence of pressure ulcers is spectacularly lower in German than in Dutch nursing homes, also when adjusting for patient characteristics and preventive measures.
Performing an (inter)national study, it is always difficult to control whether the measurement is performed totally in a uniform way. Therefore, to improve the collection of reliable and valid data, we train all coordinators of each organization; we provide them with the study protocol and training material to train their own personnel. We ask them to measure on one and the same day in all healthcare organizations and on all wards/departments at all patients present on the day of the measurement. Furthermore, although the questionnaires used in the different countries are uniform, questionnaires are filled out by the measurers (two per patient) from the perspective of country-specific standards and habits. For example, when no national guideline is available, institutions may refer to their own standards.
However, while institutions have to fill in the data into a routinized web-based data-entry program, wrong or impossible data-entry is not possible. For example, false length, weight and birth date are refused by the data-entry program. Furthermore, if it is filled out that a patient does not suffer from a care problem, it is not possible to fill out more patient-related questions concerning this care problem.
Another aspect, which may influence the reliability and validity of the measurements, is the difference between internal and external use of data (Freeman 2002). Internal indicators are used by healthcare providers to monitor and improve the outcomes of their own care processes. Professionals and managers can use this information to explore where potential problems exist, and how they can be approached. Care processes and structures may be redesigned, and indicators can subsequently be used to monitor results of these improvement efforts (Solberg et al. 1997). External indicators are for external use to report to healthcare insurers, healthcare inspectorates, etc. External indicators are more prone to reliability and validity errors, whereas institutions might have a tendency to influence results in a more positive way. LPZ-International is, in principle, an internal monitor and intends to enhance the quality of care in healthcare organizations. Therefore, although institutions have to invest a lot of manpower in the measurement, in addition to the fee they have to pay for participation, it seems not logical to influence data for a more positive result.
LPZ-International is a cross-sectional annual study; therefore, nothing can be said about the causality of relations. As prevalence is expressed as a product of incidence and average duration (Rothman & Greenland 1998, Freeman 2002), it is a more relevant measure than incidence when assessing the impact of a problem and to assess subsequent needs in a population. Furthermore, incidence measurements are too costly and labour intensive to measure on such a broad scale as the LPZ-International, as it requires daily observation and registration.
Another issue, in this regard, is that not all care problems detected in patients have developed in organizations themselves, and therefore cannot be regarded as a direct result of the quality of care of the institution. Therefore, for each care problem, a question is included whether the care problem has been developed in the institution itself or elsewhere. In this way, the nosocomial prevalence can be calculated, which gives a more valid indication.
Participation in the audit is voluntary and organizations have to pay to participate in the audit annually as mentioned above. This may influence the generalizability, although only institutions may participate who are really interested in looking to their quality of care (selection bias). Also non-response can influence the prevalence rates (Lahmann et al. 2006). Therefore, we will include only wards/departments with a response of at least 90% in national calculations.
Although several factors can influence the generalizability, we find each year in the Netherlands, with varying institutions, rather stable but declining prevalence rates, comparable to international prevalence rates, which indicates that the results are a good indication of the quality of care in the Netherlands.
In the Netherlands, the study has proven its value: it has lead to more awareness of care problems among healthcare workers in individual institutions, but also on managerial and political level. As a result, several institutions have started to improve their quality of care. Also on national level (www.zorgvoorbeter.nl), improvement programmes have been introduced, ‘Care for Better’ for long term care and ‘Cure Faster’ for hospitals, to help institutions to improve their quality of care. Participation in the LPZ and involvement in national improvement programmes appears to positively influence the quality of care (Halfens et al. 2001, Bours et al. 2003, Meijers 2009).
Measuring the prevalence of care problems internationally in a uniform way is a huge step forward to get insight in the quality of basic care in different healthcare settings in different western countries. LPZ-International seems to be a reliable, valid and generalizable method for this.
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
No conflict of interest has been declared by the authors.
All authors have agreed on the final version and meet at least one of the following criteria [recommended by the ICMJE (http://www.icmje.org/ethical_1author.html)]: