Keywords
information - information quality - secondary use - quality register
Introduction
Quality registers (QRs) are databases containing information about the quality, results,
and effects of the care received by patients, and they are used for systematically
monitoring, evaluating, and improving health care outcomes.[1] The aim is to improve care quality and to produce more efficient and equitable health
care services.[1] In many countries, such as Sweden, Denmark, and the United Kingdom, QRs have been
utilized to improve care quality and patient safety, as well as for research purposes.[2]
[3]
[4] In Finland, the first nine national QRs were subject to controller liability of
the Finnish Institute for Health and Welfare by the Ministry of Social Affairs and
Health at the beginning of 2023 ([Table 1]).[5]
[6]
Table 1
Finland's national quality registers and their target populations
|
Quality register
|
Target population
|
|
Diabetes register
|
Diagnosed patients with diabetes
|
|
Cardiac register
|
Diagnosed patients with coronary artery disease, including acute coronary syndrome
patients
|
|
Rheumatology register
|
Diagnosed patients with inflammatory rheumatic diseases
|
|
Register for kidney diseases
|
Adults who have initiated chronic renal replacement therapy
|
|
Psychosis care register
|
Diagnosed patients with psychosis (certain main groups of ICD-10 and ICPC-2 diagnosis
classifications)
|
|
Spine register
|
Patients undergoing spinal surgery
|
|
Oral and dental care register
|
Patients who have visited public oral health care services (e.g., periodontal diseases
and caries)
|
|
HIV register
|
HIV-positive patients
|
|
Intensive care register
|
Patients treated in an intensive care unit
|
Note: Data from the Finnish Institute for Health and Welfare.[6]
The real-world data produced in everyday health care and reused in QRs must be of
high quality to ensure that the information produced by the registers is reliable
and useful.[7]
[8]
[9] The information quality is a multidimensional concept and can be described through
quality attributes. The use of the quality attributes is, however, not consistent
in the research literature.[7]
[10]
[11] Frequently recurring quality attributes in the literature include correctness, accuracy,
currency, and completeness,[7]
[10]
[11]
[12]
[13]
[14] but also relevance, content, and form.[11]
[15]
[16]
[17]
[18] When assessing information quality, particularly in terms of its relevance and suitability
for intended purpose, emphasis is placed on the information user and the context of
use.[11]
[15]
[19] The methods for quality assessment should be chosen according to the requirements
of the information's intended use.[19]
[20] In this article, the information quality was examined with a focus on context, considering
the needs of the users and the purpose of use. The context-dependent quality attributes
of information, particularly perceived completeness, relevance, and usability, were
examined.[16]
[21]
Several previous studies have aimed to explore the importance of the quality of health
care information for its secondary use, referring to its utilization outside of direct
patient care.[22]
[23]
[24]
[25]
[26] Relevance, or suitability for purpose, is a quality attribute related to the content
of the information, and it can also be described as perceived usefulness.[10]
[17]
[27] However, information relevant for the intended use is not always complete enough
for use.[18] There is no universally applicable threshold for an acceptable amount of missing
information.[14] Instead, it is necessary to consider to what extent deficiencies in completeness
pose a problem for the planned use of the information.[22] User assessment of the completeness of the information in relation to the purpose
of use describes the amount of information necessary for intended use.[10]
[16]
[22]
[27]
[28] In health care QRs, deficiencies in the completeness of information can cause difficulties
in carrying out a reliable assessment of patient care outcomes.[29]
[30]
[31]
[32]
In addition to completeness and relevance, the usability of information can be considered
as a key feature for the secondary use of the information. Although usability usually
refers to the ease of using an information system,[12]
[33] it can also be interpreted as an attribute of information quality.[10]
[21]
[34] In this study, usability was examined as sufficient information structuring in relation
to the purpose of use. Structured information is stored by using defined data structures.[35] Previous studies have found that structuring patient information produces higher-quality
information for secondary use.[23]
[36]
Information quality has an impact on whether the information is wanted for use and
whether the information meets the needs of users. Information quality is a significant
factor in users' intention to use information systems.[37]
[38]
[39]
[40] In QRs, poor-quality information can cause various problems for the end use of the
information produced by these registers. The effects of patient care interventions
can be over- or underestimated, and goals set for QRs, such as assessing care quality
and making decisions based on registers, cannot be reliably achieved.[10]
[31]
[32]
[41] Deficiencies in the completeness of information can distort analyses based on the
information, while relevance and sufficient structuring are essential for the usefulness
and usability of the information.[9]
[30]
[32]
The Kanta services, Finland's national electronic information system services and
information resources for health care and social welfare,[42] are one of the key data sources for QRs in Finland. The users of the Kanta services
include pharmacies, health care and social welfare services, and inhabitants.[42] Pharmacies and health care service providers in both public and private sector enter
their patient data into the Kanta services.[42] A professional enters the information into the patient information system, from
where it is stored in the Kanta services in an up-to-date format.[42] The Kanta services' information is transferred to the QRs through separate information
retrieval and then processed to be stored in the QRs.[43]
[44] The completeness of the information and the amount of structured information in
the Kanta services in relation to the information needs of the QRs were examined during
the pilot project of QRs in 2018 to 2020 and during the data vault pilot of the national
diabetes QR in 2020 to 2021.[43]
[44]
[45] Deficiencies were identified in the information completeness, and the amount of
structured information in the Kanta services was found to be insufficient.[44] The quality assurance of health care data, the development of the Kanta services,
and increasing the role of Kanta services' information in secondary data use are all
national strategic objectives in Finland.[46] Research on the significance of information quality for secondary use is also a
central theme internationally, as the secondary use of clinical information increases,
including across national borders.[47]
Objectives
The purpose of this study is to map the views of the experts responsible for the national
QRs on information quality in the Kanta services in relation to the information needs
of the national QRs, and the significance of information quality for information usage
and intention to use the information. In addition, the purpose is to map the experts'
views on how the quality of information should be developed in relation to the QRs'
information needs. The study objective is to increase understanding of the significance
of information quality for the secondary use of the information in the national QRs,
and to provide information on whether the information quality of the Kanta services
supports the information needs of the national QRs, and how information quality should
be developed.
Methods
Theoretical Framework and Research Questions
The theoretical framework applied in this study was the model of DeLone and McLean,[12] in which information quality, information system usage, and intention to use are
all interconnected. In this study, the term usage refers solely to information use,
not to the use of an information system. Furthermore, usage is considered secondary
use of the information from the Kanta services, as reported by the QRs' experts, and
it occurs as information retrieval from the Kanta services and as processing, analyzing,
and reporting this information as part of the QRs' information production processes.
Intention to use describes experts' motivation to utilize the information in the Kanta
services and is connected to information quality attributes that may influence intention
to use.[12]
[21]
[34] Intended usage targets the phase of defining information needs and identifying relevant
information contents from the perspective of QRs. It was assumed in the study design
that the quality of the information needed to align with the information needs of
the experts to ensure not only an intention to use the information but also its actual
use. To address the issue of the significance of information quality for the secondary
use of the information in QRs, we formulated the following research questions:
What is the quality of the information in the Kanta services in relation to the information
needs of the national health care QRs?
What significance does the quality of the information in the Kanta services have for
the use or for the intention to use the information in the national health care QRs?
How should the quality of the information in the Kanta services be developed in relation
to the information needs of the national health care QRs?
This study is based on the Master's thesis published at the University of Eastern
Finland by the first author.[48]
Data Collection Phase
The data collection method used in the study was semistructured interviews. The sampling
method employed was purposive elite sampling, in which individuals were selected for
interviews based on the assumption that they would have the best knowledge of the
research phenomenon.[49] The interview questions were formulated based on the model of DeLone and McLean.[12] There was a total of 14 questions, 8 of which were examining the relationship between
the information quality and the usage or the intention to use, further divided into
questions addressing different information quality attributes and a question focusing
on information quality development. Considering information relevance, interviewees
were asked which information contents of the Kanta services were used in the national
QRs' information production processes at the time of study, and which information
contents they intended to use in the future. In addition, the interviewees were asked
for background information and were also given opportunities to freely provide any
additional information related to the topic of the study. The interview guide of the
study is presented in [Supplementary Appendix S1] (available in the online version).
A total of six experts were interviewed. The interviewees were divided into two groups,
each consisting of three individuals. The interviews for the study were conducted
remotely in January and February 2024. The interviews were conducted as individual,
paired, and group interviews. The original plan was to collect data using a specific
form of interview: a focus group discussion.[50] However, the data collection methods varied because it was difficult for the interviewees
to participate in the planned interviews due to scheduling reasons. Each interview
was recorded and transcribed verbatim, and the transcript was pseudonymized. Finally,
transcripts of group, pair, and individual interviews lasting a total of 4 hours and
9 minutes were combined. The length of the combined material was 59 pages (font Times
New Roman 12, spacing 1).
Data Analysis
The data analysis method used in this study was theory-driven, deductive, qualitative
content analysis, in which the concepts defined according to the DeLone and McLean
model[12] and adapted to the research context were integrated into the data analysis.[51] The classification of the data was based on the chosen theoretical model, from which
the main categories of the structured analysis matrix were formed: the relationship
between information quality and information usage, and the relationship between information
quality and the intention to use the information. The analysis matrix was complemented
with main categories concerning the relationship between information quality and information
needs, and the development of information quality in relation to information needs.
These main categories described the information quality required for using the information
and how the information quality should be developed to make the information more suitable
for QRs in the future. The interview data were coded, the original expressions were
simplified, and the expressions belonging to the analysis matrix were grouped into
the main categories of the matrix.[51] The unit of analysis was a complete expression or statement appearing in the original
interview data. Simplified expressions placed under the main categories were examined
within the category inductively, and subcategories were formed. Subsequently, these
subcategories were further condensed into overarching upper categories. These upper
categories formed the basis for drawing conclusions. The progression and logic of
the analysis are presented in [Supplementary Appendix S2] (available in the online version), where the analysis is described with respect
to one main category: the development of information quality in relation to information
needs. Software was used to support the qualitative content analysis (Atlas.ti 24,
ATLAS.ti Software Development GmbH, Berlin).
Ethical Considerations
This study was conducted following the ethical research guidelines of the Finnish
National Board on Research Integrity, the Personal Data Register Act of Finland, and
the EU's General Data Protection Regulation.[52]
[53]
[54] Based on the decision of the register controller, the Finnish Institute for Health
and Welfare, the research project did not require a research permit. Before the interviews,
participants were requested to provide informed, ethical consent to participate in
the research. To inform the participants, they were provided with a research information
sheet, and they also had the opportunity to familiarize themselves with the research
data protection and privacy statements.[52]
[53] Participation was voluntary, and the interviewees had the right to discontinue their
participation in the research at any time.[55] In accordance with good scientific practices, material containing personal data
was destroyed after the material had been pseudonymized.[55] The background information of the interviewees was not reported, because it was
identified as individualizing information.
Results
The Quality of Information in Relation to the Information Needs of the Quality Registers
The interviewees deemed the information provided by the Kanta services to be useful
for the QRs, and thus, the relevance of the information matched the information needs.
Several interviewees emphasized the usefulness of laboratory test results and physiological
measurements, as well as diagnosis, medication, oral health care, and medical procedure
information.
Physiological measurements, I think, are just like a treasure trove, similar to lab
tests and results.
In addition to the actual patient information, interviewees identified information
related to service events as a relevant content. They also considered metadata, or
descriptive information, to be useful for QRs' information production processes. However,
the amount of structured information in the Kanta services and the information completeness
only partially matched the QRs' information needs. If the information contents relevant
for the QRs were only partially structured, the structured portion of the information
determined the perceived information completeness. The amount of structured information
in the Kanta services was commonly described as limited.
Actually, the question is precisely how it can be utilized, as the amount of structured
information is limited.
The structuredness and thus also the completeness of the physiological measurements,
but also some of the laboratory test results, were perceived as insufficient. In addition,
the refinement level of the structuring of the physiological measurements was not
sufficient for all variables. Refinement was desired in the information structure
regarding whether blood pressure measurement was a single measurement or an average
of measurements over several days. As for laboratory test results, challenges were
identified, especially in data fields where text was combined with numerical values.
Numerical results are better structured, but when it comes to positive, negative,
or under or over a threshold results, they are not as often structured.
The structuring of different information contents of the Kanta services in relation
to the QRs' information needs and observations made by the interviewees are presented
in [Table 2]. Experts were mainly satisfied with the medical procedure information and found
it sufficiently structured, but they would like the information structures to be refined.
Adverse effects of treatment, treatment outcomes, clinical findings, and functional
capacity and lifestyle information were identified as relevant information contents
for all QRs. However, these were available in the Kanta services only in unstructured
format, or the level of structuring or structured documentation was insufficient at
the time of study. The lack of sufficient information structuring was given as one
reason for using separate local information systems as information sources for QRs.
These things that are missing, well indeed, such as clinical findings, that in the
rheumatology register we have to record separately using completely different information
systems, as the patient information systems do not support this, for example, whether
the joints are swollen and painful.
As well as the structuring of the information, information completeness varied depending
on the information content and the level of specificity desired.
It actually depends entirely on what information is being looked at.
When it comes to weaknesses in completeness, especially when we delve into more specific
details, it's not always as complete.
Diagnosis and procedure information, as well as some of the laboratory test result
information, were perceived as sufficiently complete for the information needs of
the QRs. The information completeness was said to depend on clinical practices, or
established ways of recording diagnoses and conducting laboratory tests, and it was
said to be partly in line with expectations and partly contrary to them. Although
the quality of the medication information was deemed good, the absence of medication
information from inpatient care reduced the overall information completeness. The
information completeness regarding different service providers and regions was considered
to be a key strength of the national Kanta services by the interviewees.
Its usefulness comes from this comprehensiveness…that there is both private and public
healthcare, as well as specialized medical care and primary healthcare.
Adequate descriptive information, or metadata, was considered important for the QRs,
because it assisted in identifying relevant information contents and finding information.
However, the comprehensiveness of the descriptive information did not entirely meet
the QRs' information needs, as finding and interpreting the information was somewhat
challenging.
Of course, a lot of data gets transferred there, all sorts of things thrown into kind
of a black sack, and then we have to blindly search and see what's in there, and what
hasn't been found yet, so then we're not quite sure if it's there or not, and in what
form.
Table 2
The structuring of different information contents of the Kanta services in relation
to the quality registers' information needs and observations made by the interviewees
|
Information content
|
Structuring of information
|
Observations
|
|
Diagnoses
|
Sufficient
|
The accuracy of the ICD-10 and ICPC-2 classifications is not always sufficient
|
|
Laboratory test results
|
Partly sufficient
|
Some laboratory test results are unstructured or only partially structured
|
|
Procedures
|
Sufficient
|
The refinement level of structures is partially insufficient
|
|
Medication
|
Sufficient
|
The medication information of inpatient care is missing
|
|
Physiological measurements
|
Insufficient
|
Deficiencies in the completeness of structured information
|
|
Oral health care information
|
Partly sufficient
|
Only part of the relevant information is structured
|
|
Image information
|
Insufficient
|
Statements of images are unstructured. Images are not yet available for secondary
use purposes
|
|
Treatment outcomes
|
Insufficient
|
Unstructured
|
|
Adverse effects of treatment
|
Insufficient
|
Unstructured
|
|
Functional capacity information
|
Insufficient
|
National information structures have been defined, but documentation in practice is
still unstructured
|
|
Clinical findings
|
Insufficient
|
Unstructured
|
|
Lifestyle information
|
Insufficient
|
Unstructured
(physical activity, nutrition)
|
The Significance of Information Quality for Information Usage and Intention to Use
the Information in the Quality Registers
According to the interviewees, completeness and sufficient structuring of the information
were especially significant for information usage. Data integrity and comprehensiveness
of metadata were also considered significant. The sufficient structuring of the information
and the comprehensiveness of metadata were perceived as essential for information
retrieval and interpreting the information, whereas completeness was important for
analysis and reporting.
Insufficient completeness hindered progress to analyses due to the additional work
required to address completeness issues. Insufficient completeness also required consideration
of when information was sufficiently complete for drawing conclusions. Interviewees
also reported several limits on completeness, which were significant for information
utilization in analyses and reporting, and especially in drawing conclusions. The
limits of sufficient completeness were register-, variable-, and population-specific.
If there have been these coverage issues, it certainly slows down our work, slowing
down or delaying those analyses, the investigative work that may have to be done.
If there is any information, the coverage of which is so patchy that we no longer
trust it to be representative and reliable, then we have agreed that in the quality
registers, it will not be shown as a result. We only show that the measurement completeness
is like this, and reliable conclusions cannot be drawn.
According to the interviewees, only structured information is directly usable in QRs.
Unstructured information is not used in QRs because utilizing unstructured information
requires processing it into a structured format. Based on the interviews, there was
also no intention to use unstructured information, although partially structured information
or information recorded somewhat against the documentation guidelines into a structured
field was attempted to be utilized.
If there isn't structured data, then basically we cannot and will not use it. If it's
structured, then basically we're very interested in using it.
Interviewees also described the importance of sufficient structuring of information
in terms of some single information contents. Regarding some laboratory test results,
deficiencies in structuring caused challenges not only in information acquisition
but also in information processing and interpretation. According to the interviewees,
the challenges were related to the manner of information acquisition, which had to
be adapted due to insufficient availability of structured information. In these cases,
information needed to be retrieved and extracted from text fields using string-search
methods, as structured fields had not been utilized during the documentation of the
information.
Even if it's a numerical value, it may be entered into a text field. And of course,
even a numerical value might be considered structured, but since they often involve
units, finding a number that fits the unit and ensuring that they indeed correspond
to each other isn't always obvious.
Based on the interviews, the comprehensiveness of metadata was a significant quality
attribute, particularly for information acquisition, but also for information interpretation.
Although the metadata comprehensiveness was perceived to be sufficient to some extent,
and it was reported that metadata aided in drafting data requests, deficiencies in
metadata comprehensiveness had also caused challenges in finding and interpreting
information.
And then when it comes to metadata, which describes the actual data content itself,
its absence naturally consumes a lot of working time, as one has to try to guess what
might be there or what this data could mean.
Based on the interviews, the significance of information quality for the intention
to use the information mainly appeared in two ways: perceived good information quality
increased the intention to use the information, or alternatively, information quality
did not diminish the intention to use. The latter alternative occurred when, despite
some perceived deficiencies in information quality, the information was still intended
to be used. Only unstructured information reduced the intention to use the information,
but even then, possibilities to structure the information were considered. The intention
to use the information was defined by experts' motivation to improve the information
quality if needed.
If there's something as important as lab results or physiological measurements, or
information about oral health, then we just utilize them despite any problems, and
then those quality and other issues just have to be resolved.
At the time of the study, there was an intention to use the relevant information contents
of the Kanta services, which had been identified but not yet utilized, across multiple
registers. Additionally, there were plans to gradually expand the utilization of the
information contents provided by the Kanta services.
The Development of Information Quality in Relation to the Information Needs of the
Quality Registers
The information quality development areas identified by the interviewees were related
to information structures, codes and classifications used in health care, patient
information systems, documentation practices, and the data production processes prior
to secondary use. In addition, decision-making by health care service providers regarding
information systems, as well as interorganizational collaboration, was seen as significant
for improving the information quality of the Kanta services. Most information quality
development areas were related to the information completeness and enhancing the documentation
in a structured format ([Table 3]).
Table 3
Areas for development identified by interviewees to enhance information quality
|
Development area
|
|
|
Defining information structures
|
Adoption of defined information structures
New information structures
Increasing structuring
|
|
Codes and classifications
|
Broader utilization of the Association of Finnish Municipalities' national laboratory
codes in documentation
Accuracy of diagnosis classifications
Hierarchy of classifications
|
|
Documentation practices
|
Documentation according to guidelines
Structured documentation
Correctness of documented information
|
|
Patient information systems
|
Data models (hierarchy of data)
Adoption of more structured versions
Usability of systems (ease of structured documentation)
|
|
Quality assurance
|
Quality assurance upon data entry in the Kanta services
|
|
Health care service providers' decisions
|
Integration into the Kanta services
Adoption of more structured information system versions
|
|
Information production process of the Kanta services for secondary use
|
Transparency of the process
Preservation of referential integrity
|
|
Metadata
|
More comprehensive metadata
|
|
Collaboration
|
Feedback opportunities between stakeholders
Communication about information quality
|
Because structured information was considered necessary for utilizing information
in the QRs, experts would like structures defined on a national level for Finnish
health care to be more widely adopted in the Kanta services and patient information
systems. Defining new information structures, as well as improving existing ones,
was also considered important.
Why isn't that information structured, starting from what has been discussed, it hasn't
even been defined to be stored in the Kanta services with the right or necessary structures.
Regarding the existing health care codes and classifications, areas for development
included more detailed diagnosis classifications, a sufficient hierarchy of classifications,
and broader utilization of the Association of Finnish Municipalities' national laboratory
test codes in documentation. Implementation of the development of information structures
and classifications was wanted without increasing the workload of health care professionals
due to documentation. The consistency and accuracy of documentation, and the utilization
of structured fields, were considered important.
For example, now procedure codes might be used very differently, even though we might
think it's very clear how they should be used, but then there are completely different
approaches in different regions.
Mostly there is a structured field available for them, but the problem is whether
that field has been used (physiological measurements).
The development areas related to patient information systems concerned mainly data
structures and structured documentation. Experts would like all patient information
systems used in Finland to adopt the newer version of the Kanta services' data specifications,
which is more structured and contains more data fields compared with the older version.
Additionally, it was seen as important that the hierarchy of data models used in patient
information systems should be developed in such a way that the connections between
the variables would be easier to infer.
The hierarchy of information, that's an evident area for development.
Quality assurance upon information entry in the Kanta services was also seen as a
development area. After information has been received in the Kanta services, information
contents are rearranged into a format more suitable for secondary use prior to usage
in QRs. In terms of this process, the identified areas for improvement were increasing
the transparency of the process for secondary users and preventing the disruption
of referential integrity. Furthermore, it was hoped that improving the comprehensiveness
of metadata would facilitate finding relevant information in the Kanta services, interpreting
information, and tracing information in situations in which information quality development
requires collaboration among stakeholders.
Discussion
The information in the Kanta services was perceived as relevant for the national QRs
in Finland, but the completeness and, in particular, the structuredness of the information
contents did not fully meet the information needs of the QRs. Based on previous research,
structured documentation enhances the quality of the documented information in terms
of consistency, accuracy, and completeness, and supports semantic interoperability.[23]
[36] The significance of structured information for the secondary use of the information
was evident in the results of this study. Although information relevance primarily
guided information acquisition, sufficient structuring was perceived as necessary
for the possibility of using the information. The perceived usability of unstructured
or partially structured information in the QRs was poor. Based on the results, it
was also evident that if there was not enough structured information available, information
completeness was perceived as insufficient. However, for some information contents,
it remained unclear whether the lack of completeness of structured information was
due to missing or inadequate structured documentation in patient information systems
or problems related to data transmission from patient information systems into the
Kanta services. Information coverage across various service providers, and thus throughout
the care pathways, supported the QRs' information needs. The perceived amount of structured
information and information completeness varied depending on the information content.
Overall, the challenges regarding information quality in the Kanta services in relation
to the QRs' information needs identified in this study were very similar to those
observed in previous examinations during the QRs' pilot projects.[43]
[44]
[45]
Based on the findings in this study, the significance of information quality for the
intention to use the information was associated with the perceived relevance of the
information. Information contents perceived as relevant and high quality were intended
to be used also in the future. This is congruent with previous studies using the DeLone
and McLean model,[12] which indicate that information quality is a significant factor in users' intention
to use the information system.[37]
[38]
[39]
[40] However, in this study, potential challenges in information quality did not diminish
the intention to use the information if the information was considered relevant. Only
unstructured information was associated with a reduced intention to use, although
the experts were considering possibilities to utilize unstructured information in
the future. Results can be interpreted as an indication of experts' high motivation
to use the national Kanta services' information contents.
Previous studies based on the DeLone and McLean model[12] indicate that the relationship between information quality and information system
usage has been challenging to study and interpret due to varying definitions of usage
in research designs and an absence of consistent measures of use.[34]
[37] Usage has been defined differently depending on the study, either as actual usage
or as self-reported usage.[37] In this study, the interest focused on understanding the significance of information
quality for QRs' information production processes and how it potentially changes practices
in these processes. The relationship between reported information usage and information
quality was interpreted purely qualitatively, without attempting to measure usage
quantitatively or generalize the results.
Based on this study, it can be concluded that the quality of information in the Kanta
services is significant for the information usage in QRs. According to interviews,
information completeness and the structured format of information were especially
significant for information usage. Sufficient information structuring was perceived
as important for information retrieval and interpreting the information, while completeness
of the structured information was important for analysis and reporting. Insufficient
completeness or an absence of structured information slowed down and complicated the
operational processes of the QRs' information production. The problems focused on
certain information contents of the Kanta services and on the early stages of information
production, especially information acquisition, in which the structuredness of information
influenced adopted practices. The significance of metadata comprehensiveness was twofold:
it either facilitated information acquisition by supporting the drafting of data requests
or, when inadequate, hindered the interpretation of acquired information for analyses.
Insufficient completeness, considering certain information contents in the Kanta services,
prevented progress to analyses, and subsequently, no conclusions were drawn based
on this information. In these cases, information was relevant for the intended purpose,
but not sufficiently complete to be utilized as planned.[18] In previous studies focused on QRs, the same challenge has been noted regarding
completeness. If information is not sufficiently complete, analyses based on the information
are not reliable, and conclusions cannot be drawn.[29]
[30]
[31]
[32] Based on our results, this challenge was addressed in information production by
contemplating acceptable limits of completeness. In the absence of a universal boundary,
QR experts reached the same conclusion as Weiskopf and his research team[22] in their previous study. The experts on QRs decided to define the limits of acceptable
completeness based on information needs on a case-by-case principle. Limits were defined
according to the register, population, and variable in question.
Developing information quality in the Kanta services to meet the QRs' information
needs is essential for enabling the secondary use of this information. Based on our
results, there was a need to develop both the information structures and the completeness
in relation to the QRs' information needs, with the information structures being considered
a primary area for development. The results also indicated a need for increased collaboration
among stakeholders, to form a shared understanding of the quality requirements and
implementation processes of quality assurance on a national level. Based on our results,
it could be beneficial to consider applying more specific quality standards for the
Kanta services' information as well as develop the quality assurance upon data entry
in the Kanta services to enhance the information quality and the awareness of information
quality requirements.
Information quality, as evidenced in our study and earlier research, influences the
reliability of the QR information and is essential for enabling informed decision-making
based on these registers.[10]
[30]
[31]
[32]
[41] Information quality is also significant for the QRs' information production processes.
Thus, ensuring high-quality information is essential for achieving the goals of the
national health care QRs in improving health care outcomes and the quality of care.
Reliability and Limitations
Reliability and Limitations
The reliability of this study was considered in terms of consistency, credibility,
transferability, and confirmability.[56]
[57] The credibility of this study was supported by purposefully selecting interviewees
based on their expertise, relative to the research task. To enhance credibility and
confirmability, the aim was to describe the analysis process in such a way that the
connection between the data and the results is evident.[51] The reliability of the analysis was increased by incorporating authentic quotations
into the results,[57] and the internal consistency was improved by maintaining a consistent relationship
between the theoretical framework, the research questions, and the collection and
analysis of research data.
This study has several limitations. Transferability is a typical limitation of qualitative
research.[57] This study was intentionally highly context-bound and, accordingly, it is not recommended
to transfer the results to another context without careful consideration. The chosen
attributes of information quality were linked to the use case, and quality was assessed
based on the users' experiences at a certain point in time from the perspective of
specific QRs. Another potential study reliability limitation is the data acquisition
method, namely interviews. Saturation is reached when further interviews no longer
yield additional insights into the research phenomenon.[58] Increasing the number of interviewees likely would not have brought significantly
new information, given that saturation was observed in the chosen sample size. The
interviews were conducted as group, pair, and individual interviews, so reliability
could have been enhanced by employing the same data collection method in all interviews.
In group and pair interviews, interaction between participants emerged, which was
considered desirable for the study. A unified, homogeneous group, as in this study,
is often preferred in research as it more likely fosters in-depth discussions.[50] The first author performed data collection and data analysis. The researcher's manner
of posing questions and interpreting interview data have naturally influenced both
the formed research data and the conclusions drawn from it. The study reliability
could have been increased by recoding and analyzing the interview data at another
time or by another researcher.
Conclusion
In conclusion, the information in the Kanta services is relevant for the national
health care QRs, but developing information quality, especially in terms of information
structures and completeness, is the key to fully enabling the secondary use of this
information.
The results of this study can be utilized in the development of information quality
in the national Kanta services to better meet the information needs of Finland's national
health care QRs, and thus in enabling more comprehensive utilization of the Kanta
services' information to improve health care outcomes and the quality of care. The
study findings may also prove beneficial by increasing awareness about the significance
of information quality for opportunities to utilize information in secondary use.
