PATIENT DATA COLLECTION METHODS ACROSS THE WORLD. [628394]

PATIENT DATA COLLECTION METHODS ACROSS THE WORLD.
RETROSPECTIVE INSIGHTS

ABSTRACT
The ultimate objective of patient data analysis is to reduce the health care errors and to improve
the quality of medical services, in specific areas such as patient safety measures. Multiple classic and
modern data collection techniques are presented in t he current study, but only a mix of them provides the
appropriate approach to address patient safety problems. There is no “best method” in the data collection
process. The use of all relevant and available approaches at once is an ideal, given the present reporting
timeframes and data accuracy issues. The current study aims to reveal the data collection methods
applied worldwide.

METHODOLOGY
All scientific sources of the current article were mainly identified by research on Internet.
Electronic source of data includes the Wiley online library, the official web site of the U.S. Health
Resources and Service Administration (aka HRSA), the Agency for Healthcare Research and Quality in
U.S. website, PubMed Central® (PMC) archive of biomedical and life sciences journal literature (U.S.
National Institutes of Health's National Library of Medicine), the American Hospital Association data
viewer, the British Medical Journal website, the Canadian Association of Provincial Cancer Agencies
website, Elsevier, Springer International Publishing, the international journal of medical informatics, World
Health Organization guidelines for patient incident reporting, Minnesota Department of Health, St.
Michael’s Hospital, the Council of Health Care Accreditation of Southern Af rica, Google search.
The matching words used in search of materials are “data collection methods”, “hospital reporting
systems”, “incident reporting systems”, “patient events”, “patient reported data”. Relevant articles and
studies covering 2003 -2016 time frame were selected as reference. Various data referring to different
procedures of extracting and managing patient information were collected.
INTRODUCTION
Nowadays, we reached a high level of knowledge that empowers us to provide relevant expertise
in several key -domains. Technological revolution offered us the possibility to constantly improve the
quality of our life.
Regarding the health care system, there was always a constant preoccupation for patient safety
matters, for error prevention procedures. From this perspective, every medical unit and health care
provider is in the quest for the perfect data collection system. In order to identify the areas of
improvement, the relevant actors must properly collect reliable patient data. Health care procedur es and
treatments improved year by year. But, none of these would have been possible without intensive
research, which represents the fundamental activity for taking the next step in our evolution. The
advanced economies of the world invest constantly mate rial, financial and human resources in medical
research. Any deviation from standard medical procedures could result in patient injury.
The methods of data collection in hospitals depend on factors such as the final purpose of
research, in terms of compl exity and amount of data, and on the available instruments of collection, in
terms of research staff, specific protocols, tools and costs of the overall process. Other factors that
influence these methods may be the geographical region, demographical confi guration of individuals and
their psychological status, hospital work environment or the availability of specific medical techniques at a
moment in time.

PATIENT DATA COLLECTION METHODS ACROSS THE WORLD
World Health Organization elaborated in 2010 a guide for gathering patient data in hospitals from
poor-developed economies and developing countries across the world. The guide is called “Assessing
and tackling patient harm. A methodological guide for data -poor hospitals” and it was especially designed
for th ese countries. It compiles the appropriate tools and applied methods of gathering patient data and it
also describes the procedures to manage patient data, in order to reduce the level of harm from medical
system and to establish the best practices regardi ng the patient protection measures. Methodological
aspects of data collection refer to retrospective review of the hospital records, incident reporting data,
hospital audits [1].
The identified methods are: retrospective record review, record review of cu rrent inpatients, staff
interview of current inpatients and nominal group technique based consensus method [1]. A brief
description of each method clearly explains the mechanism.
Retrospective record review : the classic patient information gathering method which is based
on previous stored medical records. It represents a two -step procedure of data collection, at the
beginning are settled on the relevant patients and then the occurrence of side effects is assessed by a
senior physician, together with the ca using factors and a follow -up plan [1].
Record review of current inpatients : a more efficiently adapted version of the record review
classic method, in terms of time consumption and identification of real -time health care problems. It
supposes the analysis , for all hospital inpatients, of their medical history [1].
Staff Interview of current inpatients: this method relies on interviews with the health centre
personnel. The superiority of this method can be harnessed in habitats with lack of relevant data or
unavailable medical evidence [1].
Nominal group technique based consensus method: a recent data collection method which
involves an assessment discussion of specific topics, based on contributors’ personal insights. It requires
no complex training, limite d facilities, fast implementation [1]. Every participant opinion is very important.
All ideas and solutions are written down on a paper and at the end the best solution is voted by
participants.
Minnesota Department of Health indicates the following collec tion methods: focus groups,
key-informant interviews, Community Forums and Public Hearings, surveys, Community Resource
Inventories [2].
Focus Groups: it is organised a conference between 5 to 10 people who have comparable
origins, education, experience, and the discussion is guided and recorded on a tape by a moderator. The
highest -priority problems of the conference topics are discussed in depth. Even though the know -how of
this method is moderate and it involves low costs of implementation, it establishes the key issues and it
provides new action points [2].
Key Informant Interviews: face to face interviews with medical experts on the research problem,
who offer relevant answers about specific requirements regarding particular problems and the next steps
to be followed. This method exhibits moderate expertise and it involves low costs of implementation [2].
All data is written on paper and then is centr alised for further analysis.
Community Forums and Public Hearings: public discussions with community in order to
assess and identify the health care requirements. However, the costs and know -how level are quite low.
This method is inspired by political deb ates. The advantages of this method are: all individuals from the
community can be part of the forum, this method is inexpensive, it can increase community’s perception
about the problem and it rises the opportunity to act at a management level. It stimula tes the community
commitment to participate to the resolution plan [2].
Surveys represent complex instruments of gathering data, usually statistically significant data,
with relevant level of expertise and a high cost of implementation. The main advantag e of this approach
consists of gathering statistically significant amount of information from relevant targeted population [2].

Community Resource Inventories: this method represents a post -survey phase, when the
summary of medical service providers and th eir activities are investigated. It requests a medium cost to
implement it, but it provides information about the service providers in a targeted area, the quality
awareness of their services and their utilization rates, the future collaboration ventures b etween the
community and local organizations [2].
Matthew B. Weinger et all identified, in San Diego, USA, a variety of collection methods to report
clinical events. It is well known that patient data are collected for various purposes such as quality
improvement processes, for public or private use, or, of course, for research purposes. The authors of the
“Retrospective data collection and analytical techniques for patient safety studies” article mention about
the difference between low -sample analysis pr ocedures and industrial analysis procedures (e.g. cognitive
task analysis interviews, failure -modes -and-effects analysis) [3]. The outcome of an individual analysis
will explain clearly the factors that led to the specific patient injury. The statistical -based techniques
applied on wide datasets offer a generalized preview of risk factors in case of patient safety issues.
Regardless of the purpose of the research, each method of data collection must highlight those patient
injury abnormal effects [3].
When we refer to reporting structures, they must target to provide valuable input figures that will
serve to the improvement of health care techniques, while ensuring all users on principles like privacy,
anonymity and data security. Based on participation crit eria, these are grouped in two categories:
compulsory reporting systems and freewill reporting systems. Usually, compulsory systems are
associated with public authorities and they conduct to inaccurate reported events [3].
A proper example of freewill repo rting system is the Australian Incident Monitoring System .
This electronic system gathers significant amount of data about medical events which offer a
comprehensive view of the factors and the necessary steps to repair the situation [3].
A trend started to be seen recently in collecting patient information: the use of electronic devices
and internet techniques. These alternatives to the classic data collection methods ensure the anonymity
and in the same time the quality of the reporting process. Specific reporting applications were created to
facilitate the process [3].
In the “Retrospective data collection and analytical techniques for patient safety studies” article is
mentioned as a proper data collection method the clinical supervision electronic sys tem, which contains
an index with all relevant elements regarding patient -related events: existing injuries, type of provided
treatment, hospital parameters. Independent entities like professional organizations created the “Medical
Event Reporting System f or Transfusion Medicine” (MERS -TM) from Canada, which is able to
smooth the way of reporting patient events, based on a clear and comprehensive form, that can be
completed by everyone. Non -compulsory association systems like The Veterans Administration’ s
RADARx, the Center for Disease Control’s (CDC) VAERS, and United States Pharmacopeia’s (USP)
MedMARx are other successful models of event reporting structures [3].
A study conducted in 2005 in Korea showed that among the participating health care units , the
large majority of them (more than 75%) were collecting medical mistake data manually, using paper,
verbally data collection method was applied in 30% of cases, and only less than 5% of them were using
an electronic reporting system. The proposed solu tion to diminish the rates of medical errors and
unfavourable patient events and to increase the patient safety level is the large -scale implementation of
IT techniques [14].
Japan’s central authorities introduced in 2015 a new electronic medical data coll ection system
that’s investigates and stores data about official deaths caused by medical procedures, the so called
“iryojiko chosa seido” (medical fatalities reporting system). All Japan’s health care institutions (total of
180,000) were incorporated into the system and they were required to examine the patient sudden deaths
and report the results to relevant independent institutions. The main limitation of this system is that it
refers only to sudden patient deaths and not to the overall healthcare proced ures. The probation teams of

an hospital must demonstrate that a case represents an unexpected death. The independence of those
teams is guaranteed by using external experts in the evaluation process [20].
The National Institute of Health Research from UK supports medical research by funding of
Health Technology Assessment (aka HTA) investigations , in which are used methods as patient
question sheet, the investigation of the available medical records or more advanced techniques, like the
use of a board of e xperts. The purpose of HTA investigations is to assess the impact (from a medical and
costs perspective) of health care systems: tools, treatments, action plans and screening. All methods
above mentioned have advantages and disadvantages. The use of daily patient data relies on proper
recording of their data and appropriate specific storage systems [4].
Several years ago, in Chicago, USA, a customized electronic health record system (aka EHR
system) was created and implemented by The Alliance of Chicago Community Health Services. This
system provides assistance in the decision process for clinicians and connects the patient data with the
medical performance indicators. During 2005 -2006 the centralized EHR system was applied in 32 clinical
sites. This syst em was located in a locked space and the data could be retrieved through the Internet.
Stored data revealed trends and comparisons between communities, between separate demographic
categories. In Massachusetts, US, all health care units were demanded in 20 07 to gather demographic
patient data such as race and ethnicity, data which were incorporated into the electronic system of every
hospital. Data were sent to the US Division of Health Care Finance and Policy [5].
The most utilized method of collecting pa tient data remains the survey administration , at least
in United States of America. This method is considered to constitute the central starting points for
assessment of a population health level and the requirements of the health care system, as surveys
reveal non -administrative patient data. Not all surveys provide the expected outcome, only those which
target health care plans, hospitals and health experts can be considered as reference for health care
amendment actions. Unfortunately, the majority of su rveys represent a parallel data gathering model as
outcome data is not directly integrated in the health care system. Data reliability remains a key -issue
when we intend to use them into medical care upgrading actions and medical error mitigation plans.
Hospitals and health care units often collect significant amounts of relevant data, without using them
properly for improvement of the medical act [5].
In other region of our planet, for example in Southern Africa, in 2013, the same electronic
approach of the patient safety measures was implemented on a large scale: a patient -event reporting
system called PatSIS . It was especially designed to manage the injury -related events and so to reduce
the risk of error in healthcare centers. PatSIS represents a cross -competency solution to prevent and
monitor patient injury, to reduce time -consuming actions and paper -based processes [12].
In Toronto, Canada, the immigration process produced a significant modification of city’s
demographics. A question raises after all this process: do the medical system developed with the trend?
The Ontario Human Rights Code and the Excellent Care for All Act (2010) acknowledged the focus on
patient care as a foundation for quality care. At the moment, in Canada, the health equity data is
extracted from the Canadian Community Health Survey and/or related datasets [6].
A team of equity experts from three Canadian health care entities (Mount Sinai Hospital, Centre
for Addiction and Mental Health, St. Michael’s Hospital) agreed in 2009 to create an association that
would fight for equity health treatment. The team of experts concluded that the health data collection
structure needs to be upgraded and improved in order to achieve the association’s purpose, as they
found a lack of relevant da ta for the assessment of health care results at their establishments level. Later
on, Toronto Public Health joined the action and brought a new perspective and a new pilot scheme that
was implemented. This pilot accessed patient -data from the following ori gins: the four above mentioned
health care organizations and data collectors. The health care organizations provided quantitative data
through patient studies , while the second method of data collection refers to focus groups, individual
face-to-face inter views and online interviews [6].

Even though the primary sources of data are considered to be hospitals, cancer centers, private
doctor’s offices, general specialists, screening programmes and central official records, the manipulation
of all these, at onc e, remains a visionary dream [7].
Regarding the sources of cancer identification, some methods involve the investigation of hospital
patient records and extraction of the patients with cancer diagnosis. The reliability of this method is based
on the accur acy of stored medical data and the correct acknowledgement of cancer cases by the
personnel. Another source of relevant data may be represented by the outpatient clinics, as long as they
manage to collect cancer cases data. Even though the percentage of ca ncer cases detected in these
health care units is small as compared to other sources of cancer identification, these cases must be
included in the registers [7].
The most reliable cancer identification starting points are the pathology laboratories. These
specific medical units identify, through a histological code, each disease that must be recorded. There are
included also uncertain cancer cases in order to be assessed at a centralised level. When hospital
historical registers are available, these regist ers should include copies of pathology reports. Duplicate
data can be identified easily by using specific computer algorithms [7].
After an individual’s life ended, the autopsy transcript is an essential piece of information. At this
stage, all deaths cau sed by cancer should be registered with a different code. Haematology laboratories
successfully detect leukaemia and lymphoma, while other types of medical investigation units, by different
tests of diagnosis, identify different types of cancer. Screening programs have a long and successful
history in cancer detection. Copies of radiotherapy results and observations, the case summaries,
hospital release forms represent an excellent method of data collection [7].
In general, we assist to a dispute about manu ally data collection process and the electronic
reporting systems that manage patient data. We can identify at least three general manners of collecting
patient data: observational approach, backdated data mining from an electronic reporting environment,
investigation of inpatient health transcripts after the patient’s hospital release. The choice of data
collection method to be implemented is made by analysing the purpose of the research, the costs and the
effect upon investigated patients. Prospective met hods imply on -going observation for a clearly
determined period of time [8].
The American Hospital Association (aka AHA) applied in 2008 an annual IT study between its
members, in order to investigate the degree of digitalization within their organization s. As a result, an
electronic health record system was implemented on the state -level, which represents the single point of
relevant medical and quality information. Another survey was conducted by AHA in 2013. The goal of this
second questionnaire was to assess and to observe the effects of previous EHR implemented system.
The survey covered more than 6,400 health care units across United States. Collected data refers to
quality of health care services, sick person’s fulfilment, mortality. Public authoriti es refer to these data
when they elaborate action plans regarding the health care system [9].
Other authors such as Philippe Michel et al. showed that from a timeframe perspective, data
collection methods can be grouped in two categories: cross sectional practices and the “time -series”
practices. For the first method, the patient data were explored in a particular day, but only during that day.
The second method involves a period of time for in -depth investigation. Data are collected at regular
intervals b y the doctor [10].
The latest data collection methods available in some European countries use, as data collection
method, an electronic real -time reporting system. This system is implemented on wide -area, in hospitals,
care centres, laboratories, through the Internet. Patients need to complete an online survey. The survey is
accessible to every sick person that wants to participate. After the questionnaire is submitted, the data
are directly associated with their medical records and stored into a secured database. Data can be
accessed by the treating doctor [11].
In each county from Sweden, for example, we find a different a tumour documentation system,
but there exists also a national Cancer register, where every death is reported. A secondary sub -system

of tumour data collection extracts elementary data about cancer tumours. There are no official accredited
cancer centres currently on -going in Sweden [21].
The Germany’s Fifth Book Social Code from 2014 established the introduction of an error
notification system in the medical field. Medical units must have a risk management officer and a
management system of patient complaints [24]. In Germany exists a so called Critical Incident Reporting
Systems (aka CIRS) which is organised in a well -developed network across the country. This system
allows the respondent (nurse, other health care worker, the paramedic or the doctor) to anonymously
submit protected data via Internet, by completion of a medical questionnaire about the medical critical
event. He is allowed to add observations and recommendations to reduce the risk of recurrence of the
event. The system focuses on medical events occurred in the medical care units, having the clear
purpose of providing relevant multi -spec ialised schooling at a regional level and at an inter -professional
degree. CIRS Network extended internationally, it currently grasps in Austria and Switzerland [25].
Regarding cancer health care units in Germany, for almost every type of cancer there are
certified cancer centers, which report their outcomes to a regional Cancer Registry, on a non -mandatory
basis. However, at least in Germany, there are many documentation systems which do not interconnect,
generating several data comparability issues.

CON CLUSIONS
The methods of data collection in hospitals depend on factors such as the final purpose of
research, in terms of complexity and amount of data, and on the available instruments of collection, in
terms of research staff, specific protocols, tools a nd costs of the overall process. These methods must
extract relevant medical data in order to serve as input for health care improvement programs, by
granting users of privacy, anonymity and data security.
Even though there are multiple options available w orldwide when it comes to medical data
collection methods, none of them represents the ideal solution. There are, in general, two types of data
collection methods: manually methods and electronic/automated data collection systems. Also, there are
tradition al standard collection methods and recently we find more complex data collection tools. For
example, surveys represent complex instruments of gathering statistically significant data, but they
require a high cost of implementation.
Classic data collection approaches such as retrospective record review or record review of
current inpatients, or staff interview of current inpatients, showed their limitations nowadays. Recently,
electronic reporting system started to be implemented in collecting patient data a nd medical errors and
adverse events reporting phase.
In order to efficiently reduce the patient injury and the risk of medical errors, the reporting
systems should provide free access to all relevant factors and they should allow anonymous completion
of patient data or medical events.
None of the quoted resources states something about data comparability issues, from a
time-series data perspective. When specking about medical research or health care improvement plans,
medical patient records are normally investigated on a year -over-year basis. Patient data are stored for
several years for research and analysis purposes. Regardless of each specific method of data collection,
the comparability of reported data must be always guaranteed, by every mean and eve ry method used in
the collection process.
Future insights on this topic should regard the standardization of different data collection methods
in order to improve data comparability aspects.

REFERENCES
1. World Health Organization, WHO Patient Safety . Assessing and tackling patient harm: a
methodological guide for data -poor hospitals., 2010, ISBN 978 92 4 150038 8, NLM classification:
WX 173; http://www.who.int/p atientsafety/research/methodological_guide/PSP_MethGuid.pdf
2. Minnesota Department of Health, Brief Overview of Data Collection Methods, St. Paul, MN
55164 -0975, P.O. Box 64975, United States,
http://www.health.state.mn.us/communityeng/needs/needs.html
3. Matthew B. Weinger, Jason Slagle, Sonia Jain, Nelda Ordonez, Retrospective data collection and
analytical techniques fo r patient safety studies, 2003, http://dx.doi.org/10.1016/j.jbi.2003.08.002 ,
http://www.sciencedirect.com/science/article /pii/S1532046403000674
4. Colin H. Ridyard, PhD, Dyfrig A. Hughes, MSc, PhD, MRPharmS, Methods for the Collection of
Resource Use Data within Clinical Trials: A Systematic Review of Studies Funded by the UK
Health Technology Assessment Program, 2010, Un ited Kingdom,
http://onlinelibrary.wiley.com/doi/10.1111/j.1524 -4733.2010.00788.x/pdf
5. IOM (Institute of Medicine), Race, Ethnicity, and Language Data: Standardization f or Health Care
Quality Improvement., 2009, Washington, DC: The National Academies Press,
https://www.ahrq.gov/research/findings/final -reports/iomracereport/reldata5. html
6. Toronto Public Health, Mount Sinai Hospital, Centre for Addiction and Mental Health, St.
Michael’s Hospital, We ask because we care, The Tri -Hospital + TPH Health Equity Data
Collection Research Project Report, 2013, Canada,
http://www.stmichaelshospital.com/quality/equity -data-collection -report.pdf
7. J. Powell, Data sources and reporting, Birmingham & West Midlands Regional Cancer Registry,
Queen Elizabeth Medical Centre, Birmingham B15 2TH, UK,
https://www.iarc.fr/en/publications/pdfs -online/epi/sp95/sp95 -chap5.pdf
8. M.N. Sarkies, K. -A. Bowles, E.H. Skinner, D. Mitchell, R. Haas, M. Ho, K. Salter, K. May, D.
Markham, L. O’Brien, S. Plumb, and T.P. Haines, Data Collection Methods in Health Services
Research. Hospital Length of Stay and Discharge Destination, 2015,
doi:10.4338/ACI -2014 -10-RA-0097, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4377563/
9. American Hospital Association, Data Collection Methods,
https://www.ahadataviewer.com/about/data/
10. Philippe Michel, Jean Luc Quenon, Anne Marie de Sarasqueta, Olivier Scemama, Comparison of
three methods for estimating rates of adverse events and rates of preventable adverse events in
acute care hospitals, 2004, doi: https://doi.org/10.1136/bmj.328.7433.199 ,
http://www.bmj.com/content/328/7433/199
11. Trautmann F., Hentschel L., Hornemann B. et al., Electronic real -time assessment of
patient -reporte d outcomes in routine care – first findings and experiences from the
implementation in a comprehensive cancer center, 2016, doi:10.1007/s00520 -016-3127 -0,
https://link.springer.com/art icle/10.1007%2Fs00520 -016-3127 -0
12. The Council for Health Service Accreditation of Southern Africa, http://www.cohsasa.co.za/patsis .
13. Thompson DA1, Lubomski L, Holzmueller C, Wu A, Morlock L, Fahey M, Dickman F, Dorman T,
Pronovost P., Integrating the intensive care unit safety reporting system with existing incident
reporting systems, 2005, PMID: 16294671,
https://w ww.ncbi.nlm.nih.gov/pubmed/16294671?dopt=Abstract&report=abstract#

14. Jeongeun Kim, David W. Bates, Results of a survey on medical error reporting systems in Korean
hospitals, 2005, DOI: http://dx.doi.org/10.1016/j.ijmedinf.2005.06.005 ,
http://www.sciencedirect.com/science/article/pii/S1386505605000870
15. Tanya Hewitt, Samia Chreim and Alan Forster, Incident re porting systems: a comparative study
of two hospital divisions, 2016, DOI: 10.1186/s13690 -016-0146 -8,
https://archpublichealth.biomedcentral.com/articles/10.1186/s13690 -016-0146 -8
16. Patient Safety & Quality Healthcare Organisation, Adverse Events: Reporting and Prevention,
2014, http://www.psqh.com/analysis/incident -reporting -systems -reporting -and-prevention/
17. AHRQ Patient Safety Network (PSNet), Voluntary Patient Safety Event Reporting (Incident
Reporting), 2014,
https://psnet.ahrq.gov/primers/primer/13/voluntary -patient -safety -event -reporting -incident -reportin
g
18. Canadian Association of Provincial Ca ncer Agencies, Incident Reporting Systems in Canada,
2010, http://www.capca.ca/current -issues/patient -safety/incident -reporting
19. de Feijter JM, de Grave WS, Muijtjens AM, Scherpbier AJJA, Koopmans RP (2012) A
Comprehensive Overview of Medical Error in Hospitals Using Incident -Reporting Systems,
Patient Complaints and Chart Review of Inpatient Deaths, 2012 PLoS ONE 7(2): e31125.
doi:10.1371/journal.pone.0031125,
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0031125
20. TOMOKO OTAKE, Shortcomings identified in new reporting system into hospital deaths, 2015,
http://www.japantimes.co.jp/news/2015/10/05/reference/shortcomings -identified -new-reporting -sy
stem -hospital -deaths/
21. Per Fransson, Senior Lecturer, associate professor, Department of Nursing, Umeå University,
Umeå, Sweden, 2015.
22. Mieke Van Hemelrijck, Annette Wigertz, Fredrik Sandin, Hans Garmo, Karin Hellstro, Per
Fransson, et all, Cohort Profile: The National Prostate Cancer Register o f Sweden and Prostate
Cancer data, Base Sweden 2.0, 2012, doi:10.1093/ije/dys068.
23. Krankenhaus -CIRS -Netz Deutschland, http://www.kh -cirs.de/
24. Jürgen Brenn, Fehlermeldesystem CIRS -NRW auf einem guten Weg, 2014, In: Rheinisches
Ärzteblatt. Ärztekammer Nordrhein und Kassenärztliche Vereinigung Nordrhein. S. 21 -22. Jan
2014. Abgerufen am 6. Januar 2014: „137 Abs. 1d SGB V“
25. Wikipedia, Critical Inci dent Reporting System, 2017,
https://de.wikipedia.org/wiki/Critical_Incident_Reporting_System
26. Ärztekammer Nordrhein, Ein Tippfehler und die Folgen. In: Rheinisches Ärzteblat t. Sept. 2013,
http://www.aekno.de/downloads/archiv/2013.09.016.pdf
27. Hoffmann B1, Rohe J, Patient safety and error management: what causes adverse events and
how can they be prevented?, 2 010, DOI: 10.3238/arztebl.2010.0092,
https://www.aerzteblatt.de/int/archive/article?id=67639