The balance between professional autonomy and organizational obligations in resilient management of specialized health care: A Norwegian document study

Abstract

BACKGROUND:

Adverse events in hospitals may jeopardize the safety of patients. Failure in professional autonomy, organizational learning or in the contact between these two factors may explain the occurrence of injurious incidents in hospitals.

OBJECTIVE:

To study reasons for failure in contact between professional autonomy and organizational learning in resilient management of specialized health care through document analysis.

METHODS:

A total of 20 reports from the Norwegian Board of Health Supervision were evaluated by a retrospective in-depth document analysis. In the analysis of adverse events, we applied the Braut model to identify function or failure of 1. Professional autonomy, 2. Organizational learning and 3. Contact between professional autonomy and organizational learning.

RESULTS:

Multivariable regression analysis showed that failure in organizational learning was the only explanatory variable for failure in contact between doctors and nurses autonomy and organizational learning. Failure in organizational learning had the strongest effect on failure in contact between doctors and nurse’s autonomy and organizational learning (B = 1.69; 95% CI = 0.45 to 2.92). Failure in professional autonomy showed no significant effect on this contact.

CONCLUSIONS:

Failure in organizational learning is associated with failure in contact between professional autonomy and organizational learning. Failure in professional autonomy did not influence this contact.

Keywords

Professional autonomy organizational learning resilience management hospital specialized health care Norwegian Board of Health Supervision Norway

1. Background

Unsafe medical care in health institutions is still of major concern as adverse events occur in 10% of all admissions in modern hospitals [1,2] and can be compared with diseases such as malaria and tuberculosis [3]. Furthermore, an association between patient injuries and substandard medical management has been observed [4]. Investigations of failures in health care have underscored the negative role of poor management and lack of organizational scrutiny of safety [5,6]. Lack of an overall integrated organizational regulatory approach has also been observed, with a substantial overlap between functions and activities [7]. Thus, the importance of leadership in resilient organizations is less than well highlighted [8].

On the micro level, the term ‘safety in hospitals’ can be defined as the ability to avoid accidents and injuries to the patients as a result of adverse events [9,10]. Evaluation of health care safety may be judged along three main lines. Firstly, ‘professional autonomy’ describes the approach of the actors who actually make decisions about the treatment and management of patients. The obligations and right to self-determination of physicians and nurses have been strong throughout history, while patient safety has become more and more established at management level through legislation and regulations [11]. Providing benefit (beneficence), or at least delimiting the damage to make it disproportionately smaller in relation to the benefit of the treatment (non-maleficence), are two requirements that physicians and nurses must meet [12]. Secondly, a more systemic focus is addressed in the organizational management of hospitals [13]. This approach is one of the characteristics of a resilient institution or management [14]. A resilient organizational and safety culture is clearly central to uncovering and reporting undesirable incidents to management or the supervisory authorities [15]. Management is responsible for ensuring that laws and guidelines are utilized correctly and optimally and reach the right decision level. A resilient institution is challenged when aiming to learn from unexpected negative events and using this knowledge to improve the quality of practice in local hospitals [16]. Activities that provide health care services should be organized so that the health personnel are able to follow their statutory obligations. Thirdly, the relationship between ‘professional autonomy’ and ‘organizational learning’ opens up for a new dimension; the contact between the two entities, i.e., ‘contact between professional autonomy and organizational learning’. An established contact between professional autonomy and organizational learning describes a resilient health service on community level [17,18]. Hospitals must be organized so that health personnel can adhere to laws and guidelines to provide safe health care. The responsibility for facilitating and ensuring safe health care is placed on top management. It is no longer just the professionals’ responsibility [16,17].

On the macro level, the government choose to counteract the above-mentioned undesirable events through legislation and improvement of the quality of health care services [17,19]. During the period 1980–2000, the Norwegian government developed laws and regulations [20], where safety and organizational responsibility within the health service were emphasized [16,21]. A regulatory framework for safe professional practice in health care was introduced with emphasis on the role of management and leadership [22]. Previously, safety measures were considered a professional obligation for practitioners [23]. Today, patient safety and safety improvements are part of the management’s responsibility [24–26]. In addition, there are national regulations for professional practice [16,27]. Importantly, the Norwegian Directorate of Health presents a national action plan for patient safety and quality improvement (2019–2023) as an overall plan to systematically reduce patient injuries and improve resilience and patient safety [28].

In 2012, the Norwegian Board of Health Supervision (NBHS) established a notification scheme with the following objective: “The purpose of the notification system for the most serious incidents is to identify unjustifiable conditions more quickly so that the conditions are rectified and contribute to better patient safety.” The two most common criteria that trigger the obligation to notify NHBS are: 1. Death of a patient and 2. Very serious injury to a patient or user because of the provision of health and care services [29]. All notifications should be published anonymously and are available to the public on the NHBS’s website. The purpose of these public reports is to stimulate the management systems to learn from these undesirable events in order to change the local organizational culture [17,30]. Each report from the NBHS contains and evaluates one case on both the professional and the systemic management aspects of the adverse event [31]. The relationship between the professional and systemic level are also reported [3]. Finally, an evaluation of the above-mentioned three main factors; the professional level, the organizational level and the contact level over several cases may better illustrate the reality as a whole than information from a single report only.

There is a need for new knowledge and better theories and models for understanding the causal chains behind adverse events in health care [32]. Not least we need a better understanding of the intersection between individual and systems related factors contributing to failure, and how this can be approached in evaluations and investigations.

In the present paper, the Braut model (Fig. 1) is used for empirical purposes [16]. The model triangulates ‘professional autonomy’, ‘organizational learning’ and ‘contact between these two entities’. The underlying hypothesis of the Braut model is that if there is a ‘failure in the contact between professional autonomy and the organizational learning’, it can lead to undesirable serious incidents in the hospitals [16]. Thus, the purpose of the present paper is to analyze consecutive reports from the NBHS using the Braut model [16] and to study how they interpret the role of the professional, the organizational and the contact between the professional and organizational level when adverse events involving the patients occur in hospitals.

Fig. 1.

Overview of the Braut model. Basically, it triangulates the ‘proper operation of the health trust’ (grey), ‘professional autonomy’ (yellow) and ‘organizational management/learning’ (blue). In the present paper the relationship between ‘professional autonomy’ and ‘organizational management/learning’ is in focus and denoted ‘contact’ between these two variables (green), i.e., ‘contact between professional autonomy and organizational management/learning’.

2. Objective

The main objective of this work is to study reasons for failure in contact between professional autonomy and organizational learning in resilient management of specialized health care, through document analysis. Specifically, the study seeks to (1) assess the use of the Braut model in analyzing the reports from the NBHS and (2) identify factors that can contribute to explaining failure in the contact between professional autonomy and organizational learning [16].

Professional autonomy encompasses the ability for every individual, qualified practitioner to make own decisions in clinical situations concerning single patients. Organizational learning concerns the structures and processes established through leadership and management for supporting the practitioners and ensure that clinical practice in general in the specific clinical unit is according to prudent standards and medical evidence.

3. Methods

3.1. Research design

We performed a retrospective in-depth analytical document study and applied the Braut model empirically on 20 consecutive reports from the NBHS. A short abstract of each case is presented in Table 1. All reports have the following structure: a descriptive part where the case is described based on information from journal records, interviews or other sources. and an analytical part containing assessments and conclusions from the NHBS.

Table 1
Review of the 20 cases from the Norwegian Board of Health Supervision

# Topic Case title Outcome Case details Evaluation of the Norwegian Board of Health Supervision Pages

1 Surgery Inadequate follow-up of a patient who was admitted with severe abdominal pain. Death Delayed investigation. Patient had blood clot in central intestinal vein. Underwent surgery too late to save a necrotic bowel. The patient died the next day. The Norwegian Board of Health requests a review of the current organization of the on-call teams. Responsible managers must check that routines are complied with and function as intended. 43

Reaction against individuals: None.

2 Psychiatry Suicide. Death A 14-year-old girl took her own life when she was admitted to the Child and Youth Psychiatry Department (CYPD). The organization had not arranged measure at the CYPD to prevent suicide. Serious failure and violation of the Specialist Health Services Act. 22

Reaction against individuals: None.

3 Eye Serious eye infections after injectiontreatment withthe drug Avastin. Allsurvived 11 patients developed inflammation of the eye after injection of Avastin, which was withdrawn from a common bottle. The Health Trust provided unsound health care. The failure occurred due to economic considerations of using a large bottle of Avastin instead of a single-dose container.

Reaction against individuals: None.

4 Pediatrics A child with a congenital heart defect died after being admitted to the neonatal intensive care unit. Death The child had a complicated congenital heart defect and underwent heart surgery at 5 weeks of age. After 1 week, the patient developed an infection that was treated with antibiotics. The child was readmitted and died the same evening. The child did not receive proper health care. Intravenous fluids and antibiotics were started too late. Staffing problems and heavy workload, as well as communication problems were known for a long time. Reaction against individuals: None. 33

#	Topic	Case title	Outcome	Case details	Evaluation of the Norwegian Board of Health Supervision	Pages
1	Surgery	Inadequate follow-up of a patient who was admitted with severe abdominal pain.	Death	Delayed investigation. Patient had blood clot in central intestinal vein. Underwent surgery too late to save a necrotic bowel. The patient died the next day.	The Norwegian Board of Health requests a review of the current organization of the on-call teams. Responsible managers must check that routines are complied with and function as intended.	43
					Reaction against individuals: None.
2	Psychiatry	Suicide.	Death	A 14-year-old girl took her own life when she was admitted to the Child and Youth Psychiatry Department (CYPD).	The organization had not arranged measure at the CYPD to prevent suicide. Serious failure and violation of the Specialist Health Services Act.	22
					Reaction against individuals: None.
3	Eye	Serious eye infections after injectiontreatment withthe drug Avastin.	Allsurvived	11 patients developed inflammation of the eye after injection of Avastin, which was withdrawn from a common bottle.	The Health Trust provided unsound health care. The failure occurred due to economic considerations of using a large bottle of Avastin instead of a single-dose container.
					Reaction against individuals: None.
4	Pediatrics	A child with a congenital heart defect died after being admitted to the neonatal intensive care unit.	Death	The child had a complicated congenital heart defect and underwent heart surgery at 5 weeks of age. After 1 week, the patient developed an infection that was treated with antibiotics. The child was readmitted and died the same evening.	The child did not receive proper health care. Intravenous fluids and antibiotics were started too late. Staffing problems and heavy workload, as well as communication problems were known for a long time. Reaction against individuals: None.	33

Table 1 (Continued).

#	Topic	Case title	Outcome	Case details	Evaluation of the Norwegian Board of Health Supervision	Pages
5	Internal medicine	Patient received a blood transfusion of the wrong type.	Survived	Confusion about type of blood in bag and the patient who received this blood.	The health trust lacks control over the right blood of the right quality being transfused to the right patient.	23
					Reaction against individuals: A warning was given to two nurses.
6	Ear,nose andthroat	The ‘Daniel case’. Postoperative bleeding after removal of tonsils.	Death	3-year-old boy had his tonsils removed. 3 days later he was admitted with bleeding and bled several times. Eventually he had a cardiac arrest and died.	The hospital violated legal requirements for follow-up, information and handling of the psychosocial work environment. Culture of fear, poor communication and poor role clarification.	79
					Reaction against individuals: None.
7	Surgery	Failure in treatment and interaction when receiving a trauma patient.	Death	A severely traumatized patient was taken to the emergency room and died after a short/brief treatment in the operating room.	The Health Trust was not certified as a regional trauma clinic and had not developed adequate competence activities, including training and certification of personnel.	56
					Reaction against individuals: None.
8	Pediatrics	The ‘Sebastian case’. Lack of follow-up of a child with extensive and complex needs.	Death	10-month-old boy with congenital heart defects and respiratory problems. The patient was transported back to the hospital for definitive life-saving treatment. However. It was too late, and he died at the hospital.	The patient did not receive proper health care when he was sent home from the emergency room. Failure in command lines for transport and the health trust had not ensured proper treatment of sick children on weekends. New routines were not understood and complied with.
					Reaction against individuals: None.

Table 1 (Continued).

#	Topic	Case title	Outcome	Case details	Evaluation of the Norwegian Board of Health Supervision	Pages
9	Internal medicine	Treatment of a patient with sepsis.	Death	A young boy with contagious meningitis / sepsis was admitted to hospital for observation without medical supervision. It was too late when a cerebral edema was detected on CT and the patient died shortly after.	Examination, follow-up and treatment in the emergency department and during the stay were unjustifiable. Inadequate documentation in the journal and inadequate routines in the organization. Reaction against individuals: None.
10	Ear,nose and throat	Severe bleeding in two children after tonsillectomy in a private clinic.	1 survived, 1 died	A 5-year-old girl suffered acute and severe bleeding on the ninth postoperative day after removal of the tonsils. She survived. A 4-year old boy also suffered severe bleeding after the same operation, 4 days later. The boy died of bleeding due to the serious complication.	Improper use of a diathermy device with associated diathermy scissors led to tissue damage around the blood vessels at the tonsils, leading to an increased incidence of postoperative bleeding. The 4-year old boy had such heavy bleeding that his life could not be saved. Reaction against individuals: The surgeon/general manager (same person) was deprived of a full medical license and works with a reduced license under the supervision of a National Trust Hospital.	34
11	Pediatrics	Mismedication with chemotherapy of a child.	Survived	Children with brain tumors received two doses of the cytotoxic drug Lomustine on two consecutive days when there should be 6 weeks between each dose. The prescribing physician was busy with concomitant cases and the treating nurse was unfamiliar with the drug.	The trust did not provide proper health care to the child who was to be treated for cancer. Lack of training and bustle characterize the incident for both doctors and nurses/For both doctors and nurses the incident was characterized by lack of training and time pressure. Reaction against individuals: None.	27

Table 1 (Continued).

#	Topic	Case title	Outcome	Case details	Evaluation of the Norwegian Board of Health Supervision	Pages
12	Gyn and obs	Serious incident with children.	Death	A woman in labor arrived at the hospital. Weakening of uterine contractions led to late delivery of the child. Lack of attention and situational awareness on the part of two midwives led to/resulted in late delivery with severe asphyxia. The child died later that day.	The health trust had not facilitated competence development among newly trained midwives. No barrier to them/No rules in place to prevent them from being left alone in a demanding and risky situation. Experienced midwives were difficult to obtain that evening. Reaction against individuals: None.	36
13	Pediatrics	Inadequate follow-up of a 7-week-old child after starting sepsis treatment.	Death	Child born 14 weeks before term (600 g) was transferred to the neonatal ward in good shape after 3 weeks. After 6 weeks she became restless, crying/cried, and had an increased need of oxygen, with signs of poor lung function. Suspicion of sepsis. 8 hours later, she died despite initiation of treatment.	The diagnosis of sepsis was too late for the introduction of life-saving treatment. Unclear lines in the follow-up of the child. There were unclear routines for responsibilities between doctors present. The start of sepsis treatment was too late. Reaction against individuals: None.	34
14	Pathology	Contamination of cell sample gave incorrect test result.	Survived	Exchange of two cytology tests (cell sample) at the Department of Pathology, which led to a misdiagnosis of cancer in a healthy patient. Thus, this patient underwent unnecessary surgery and incorrect removal of 14 axillary lymph nodes.	Inadequate routines for handling cytological material. Reaction against individuals: None.	32

Table 1 (Continued).

#	Topic	Case title	Outcome	Case details	Evaluation of the Norwegian Board of Health Supervision	Pages
15	Oncology	Serious event regarding bone marrow transplant from xx to children under 18 years.	Survived	Patient in need of stem cell transplantation at Oslo University Hospital had his sample exchanged with a donor’s sample in tissue type assessment. Re-typing was to be done, but was incorrectly omitted. The patient survived.	The guidelines were not followed and complied with in practice. Reaction against individuals: None.	20
16	Psychiatry	Deaths due to neutropenic sepsis after Leponex treatment.	Died	A psychiatric patient had his dose of Leponex increased, 4-fold by 4 independent physicians without any form of consultation. The Patient developed a side effect of the drug, namely neutropenia (few white blood cells). Developed septicemia, cerebral edema and dies.	Norwegian Board of Health Supervision considers the follow-up of the medication to be unjustifiable. Lack of routines for information transfer. Reaction against individuals: None.	33
17	Surgery/Plastic surgery	Improper health care in health trust with complex organization.	Death	A woman in her 60s was admitted to the medical department with signs of a serious soft tissue infection in her thigh. Requested CT was refused by the radiologist, and the surgeon on duty refrained from looking at the patient. The patient died the next morning. Autopsy showed sepsis and gas gangrene in the tissue.	The trust did not provide the patient with proper health care. Lack of routines to ensure collaboration between different subject areas. The hospital has not learned from previous incidents. Reaction against individuals: None.	23

Table 1 (Continued).

#	Topic	Case title	Outcome	Case details	Evaluation of the Norwegian Board of Health Supervision	Pages
18		Incorrect medication, Potassium chloride (KCl) instead of calcium chloride (CaCl₂).	Death	Elderly woman with diagnosed bowel cancer. Operation was planned. Arrived acutely with ileus (intestinal loop) and surgery was determined on vital indication. The patient had electrolyte disturbances, including high K⁺-values (Potassium) and low Ca⁺⁺⁻ values (Calcium). An oral message between doctor and nurse led to confusion between these two preparations, which were in the same drawer in the medicine room. The patient incorrectly received KCl intravenously instead of CaCl₂. The patient had a cardiac arrest and died.	Complex and comprehensive case. Much done right, but one mistake led to the patient’s death. Lack of closed-loop communication when such medication is to be administered. One must show extra vigilance in situations like this. Double-checking should be performed when such potent medications are to be administered. Preparations with such similar names should be in different places. Reaction against individuals: None	33
19	Gyn and obs	Failure in follow-up of a risk birth.	Death	Woman in week 39 + 6 with pregnancy poisoning. Spontaneous water discharge. Protracted course of birth discolored amniotic fluid. CTG had a ’strange pattern’. After a few hours, the foetal sound dropped to 50 beats/min. The woman was delivered by emergency caesarean section, but the child died.	The hospital trust had not secured/facilitated proper health care for at-risk mothers. Failure in interaction between doctor and midwife. Failure of foetal monitoring. Misinterpretation of foetal sound on both CTG and STAN. Reaction against individuals: None.	33

Table 1 (Continued).

#	Topic	Case title	Outcome	Case details	Evaluation of the Norwegian Board of Health Supervision	Pages
20	Paediatriconcology	Medication error in chemotherapy of a 6-year-old child.	Death	A child aged 6 years with brain tumour should have received the cytotoxic drug vincristine intravenously and methotrexate into the ’cerebrospinal fluid’ of the brain (i.e., intrathecal injection). The doctor who was to administer the drugs mixed up the two types of cytotoxic drug and injected vincristine into the brain. Despite attempts to limit the damage, the child died.	The Health Trust has not ensured proper health care for children who received intrathecal chemotherapy via the cerebrospinal fluid. Failure in drug handling. Lack of double control. Reaction against individuals: The doctor who injected vincristine into the brain received a written warning.

For the sake of readability, in this paper failure in the contact between professional autonomy and organizational learning will be referred to as failure in contact. After initial definitions, all items derived from the Braut model are used without inverted commas to increase readability.

3.2. Data collection

The 20 reports from the NBHS included in the study were obtained from the archives of the NHBS’s website and constituted consecutive reports from the 2018 collection (Table 1). They were studied by in-depth reading and scored according to the Braut model. The formation of 18 scoring variables is shown in Table 2. All reports were read by two researchers (AS and HS). During the in-depth reading we allocated each theme to one of the 18 variables. During the scoring, we used coloured line markers to outline which variable belonged to which part of the text (descriptive, analytic or total). When there was a discrepancy in scoring, a third researcher (GSB) was consulted to reach consensus.

Table 2
Overview of the 18 variables used to score the 20 reports from the Norwegian Board of Health Supervision. The letters A to R correspond with the numbering of the figures in Fig. 3

No Independent variables Part of the report

Total score Descriptive score Analytical score

1 Functional professional autonomy A B C

2 Failure in professional autonomy D E F

3 Functional organizational learning G H I

4 Failure in organizational learning J K L

5 Functional contact between professionalautonomy and organizational learning M N O

6 Failure in contact between professionalautonomy and organizational learning P Q R

No	Independent variables	Part of the report
1	Functional professional autonomy	A	B	C
2	Failure in professional autonomy	D	E	F
3	Functional organizational learning	G	H	I
4	Failure in organizational learning	J	K	L
5	Functional contact between professionalautonomy and organizational learning	M	N	O
6	Failure in contact between professionalautonomy and organizational learning	P	Q	R

3.3. Data analysis

SPSS version 25 for Mac was used. All notifications on the 18 variables in Table 2 in the 20 reports were calculated and sum scores provided. First, we made a descriptive analysis of the 18 sum scores. Second, a bivariate correlation between the key variables was performed [33]. Then, a simple linear regression analysis was carried out followed by a multivariable linear regression analysis [34]. The level of significance was set to P < 0.05.

4. Results

4.1. Descriptive analysis

The distribution of a properly functioning hospitals, professional autonomy and organizational learning is shown in Fig. 2. Furthermore, the descriptive analysis of all the 18 variables derived from the Braut model of the in-depth reading analysis are presented in Fig. 3 A to R. Both types of distribution, median and mean values are included in the analysis. Interestingly, many of the analytical analyses show a zero count, which leads to a skewed distribution for several of the variables.

Fig. 2.

Distributions of the scorings of the basic characteristics of the Braut Model; ‘proper operation of the health trust’ (A), ‘professional autonomy’ (B) and ‘organizational management/learning’ (C).

Fig. 3.

Distribution of the 18 variables (A to I) from the Braut model. For all basic variables, total score, descriptive score and analytic score are presented.

4.2. Bivariate correlation analysis

The correlation between proper operation of the hospitals and professional autonomy was very good (r = 0.979; R ² 0.958; p < 0.0001). However, the correlation between proper operation of the hospitals and organizational management and learning was poor. (r = 0.004; p = 0.985) (Fig. 4).

Fig. 4.

Distribution of the 18 variables (J to R) from the Braut model. For all basic variables, total score, descriptive score and analytic score are presented.

The bivariate correlation between the 18 study variables showed a trend towards that failure in contact was correlated with failure in organizational learning (Table 3). The strongest correlation was between failure in contact – descriptive and failure in organizational learning – analytical with a r = 0.602; p = 0.005 (Table 3, line 12).

Table 3

Bivariate correlations of the variables

No.	Variable 1	Variable 2	r ^∗	P-value
1	Failure in contact - total	Failure in professional autonomy - total	0.006	0.981
2	Failure in contact - total	Failure in professional autonomy - descriptive	0.266	0.257
3	Failure in contact - total	Failure in professional autonomy - analytical	−0.286	0.222
4	Failure in contact - total	Failure in organizational learning - total	0.472	0.035
5	Failure in contact - total	Failure in organizational learning - descriptive	0.109	0.646
6	Failure in contact - total	Failure in organizational learning - analytical	0.534	0.015
7	Failure in contact - descriptive	Failure in professional autonomy - total	−0.276	0.239
8	Failure in contact - descriptive	Failure in professional autonomy -descriptive	0.139	0.558
9	Failure in contact - descriptive	Failure in professional autonomy - analytical	−0.559	0.010
10	Failure in contact - descriptive	Failure in organizational learning - total	0.698	0.001
11	Failure in contact - descriptive	Failure in organizational learning - descriptive	0.231	0.326
12	Failure in contact - descriptive	Failure in organizational learning - analytical	0.602	0.005
13	Failure in contact – analytical	Failure in professional autonomy - total	0.144	0.546
14	Failure in contact - analytical	Failure in professional autonomy - descriptive	0.348	0.132
15	Failure in contact - analytical	Failure in professional autonomy - analytical	−0.149	0.546
16	Failure in contact - analytical	Failure in organizational learning - total	0.324	0.163
17	Failure in contact - analytical	Failure in organizational learning - descriptive	−0.260	0.268
18	Failure in contact - analytical	Failure in organizational learning - analytical	0.458	0.042

^∗Spearman’s rho.

4.3. Simple linear regression analysis

Table 4 shows the results from the simple linear analysis when failure in contact – total was the dependent variable. Here, failure in organizational learning total and failure in organizational learning – analytical had a Beta value of 1.44 (95% CI 0.19 to 2.70) and 1.69 (95% CI 0.45 to 2.92) respectively (Table 4, lines 4 and 6). When failure in contact – analytical was selected as the dependent variable failure on organizational learning -analytical was the only significant independent variable (B = 0.97; CI95% 0.11 to 1.83) (Table 5, line 6).

Table 4
Simple linear regression analysis 1

Dependent variable: Failure in contact - total

No. Independent variable B 95% CI P-value

1 Failure in professional autonomy - total −0.30 −1.68 to 1.09 0.66

2 Failure in professional autonomy - descriptive 0.40 −2.05 to 2.84 0.74

3 Failure in professional autonomy - analytical −1.02 −3.11 to 1.07 0.32

4 Failure in organizational learning - total 1.44 0.19 to 2.70 0.026

5 Failure in organizational learning - descriptive −0.67 −3.75 to 2.41 0.65

6 Failure in organizational learning - analytical 1.69 0.45 to 2.92 0.010

Dependent variable: Failure in contact - total
1	Failure in professional autonomy - total	−0.30	−1.68 to 1.09	0.66
2	Failure in professional autonomy - descriptive	0.40	−2.05 to 2.84	0.74
3	Failure in professional autonomy - analytical	−1.02	−3.11 to 1.07	0.32
4	Failure in organizational learning - total	1.44	0.19 to 2.70	0.026
5	Failure in organizational learning - descriptive	−0.67	−3.75 to 2.41	0.65
6	Failure in organizational learning - analytical	1.69	0.45 to 2.92	0.010

Table 5

Simple linear regression analysis 2

Dependent variable: Failure in contact - analytical
No.	Independent variable	B	95% CI	P-value
1	Failure in professional autonomy - total	0.07	−0.85 to 1.00	0.87
2	Failure in professional autonomy - descriptive	0.55	−1.05 to 2.15	0.48
3	Failure in professional autonomy - analytical	−0.25	−1.67 to 1.17	0.72
4	Failure in organizational learning - total	0.75	−0.13 to 1.63	0.09
5	Failure in organizational learning - descriptive	−0.76	−2.78 to 1.26	0.43
6	Failure in organizational learning - analytical	0.97	0.11 to 1.83	0.030

4.4. Multivariable linear regression analysis

In the multivariable linear analysis, we observed that failure in organizational learning was the only explanatory variable for failure in contact. Furthermore, failure in organizational learning - analytical had the strongest effect with B = 1.69 (95% CI 0.45 to 2.92) (Table 6, Model 1). Model 1 uses failure in contact – total as the dependent variable, while Model 2 uses failure in contact - analytical as dependent variable.

The independent variable was the same (failure in organizational learning - analytical) in both Model 1 and 2. The reason for presenting two separate models is a considerate overlap between the two dependent variables with an extensive interference with each other when they are in the same model. It was therefore sensible to split the analysis into two models (Table 6).

Table 6
Multivariable linear regression analysis

Model no. Dependent variable Independent variable significant in the model R² B (95% CI) P-value

Model 1 Failure in contact - total Failure in organizational learning - analytical 0.313 1.69(0.45 to 2.92) 0.010

Model 2 Failure in contact - analytical Failure in organizational learning - analytical 0.236 0.97(0.11 to 1.83) 0.030

Model no.	Dependent variable	Independent variable significant in the model	R²	B (95% CI)	P-value
Model 1	Failure in contact - total	Failure in organizational learning - analytical	0.313	1.69(0.45 to 2.92)	0.010
Model 2	Failure in contact - analytical	Failure in organizational learning - analytical	0.236	0.97(0.11 to 1.83)	0.030

Model 1 uses failure in contact – total as the dependent variable, while Model 2 uses failure in contact - analytical as the dependent variable. The independent variable was the same (failure in organizational learning - analytical) in both Models 1 and 2. The reason for presenting two models is a considerate overlap between the two dependent variables interfering heavily with each other when they are in the same model.

5. Discussion

To the best of our knowledge, this is the first time that reports from the NBHS have been used on an empirical basis in quantitative analyses. When applying the Braut model, we observed that failure in organizational learning was the only factor that explained failure in contact between professional autonomy and organizational learning.

In our opinion, the audit reports contain three important elements that make them relevant to all hospitals in Norway [35]. First, the reports have a thorough review of each of the 20 cases from the point of view of the health personnel. Second, the NBHS analyzes the role of the management and the organization in depth and becomes an unbiased voice in meeting the top management. Third, the 20 cases represent a broad variety of medical fields [3]. As such, we believe that the NBHS reports cover several medical questions and situations. Thus, these reports are probably representative of and relevant for all involved parties in providing appropriate health care to patients at all hospitals in Norway [36].

One important question is whether it is relevant to use the Braut model empirically in the present context [16]. When the NBHS assesses whether health care is justifiable, it considers whether the professional autonomy leads to adequate assessments and actions in the correct order in the specific situation [16]. The NHBS also assesses whether the hospital has arrangements enabling health personnel to provide proper health care (= organizational learning) [37]. This means that health care can be considered justifiable even if the outcome is tragic with disabling injury or death [38]. Thus, the reports deal with the two entities found in the Braut model. Therefore, this model seems suitable for analysis of the issues raised in the audit reports. Furthermore, we found that there was strong correlation between proper operation of hospitals and professional autonomy, while there was no significant relationship between proper operation of hospitals and organizational learning. This is in accordance with our basic understanding of the Braut model [16]. It also helped to strengthen the internal validity of the Braut model for the analysis of our reports.

In the multivariable modelling, we observed two interesting findings. First, failure in professional autonomy had no influence on failure in contact. This could mean that even if the doctors’ and nurses’ best judgment in complicated situations later turns out to be wrong, it does not necessarily place any obstacles in the way of contact with the organization and management. Doctors and nurses often feel responsible and comply with guidelines from management [39,40]. Secondly, failure in organizational learning was the only variable explaining the failure in contact in the final multiple variable modelling. This finding might have three explanations: Firstly, it seems that the systemic/organizational part appears to be more important than the professional part when ensuring that health workers comply with routines and organizational regulations [41]. This is in line with recent studies where stronger leadership commitment and acknowledgement of quality and safety as integral to the operational culture of healthcare organizations are questioned [40,42]. In addition, it may appear that the assessments of the NBHS have a systemic focus on adverse events [30], as failure in management tasks could lead to uncertainty and poor communication down the ranks [43]. Thus, it might affect many health workers and their relationships with patients [44]. Secondly, the NBHS’s analytical part was the significant explanatory variable. This suggests that the NBHS postpones the criticism of management until the analytical part of the report [45]. It is also a logical way to distribute the points in the report, by starting ‘professionally’ and finishing ‘organizationally’. Furthermore, the NBHS probably has a special focus on management matters [30]. This is because management deficiencies might have consequences for several departments in a hospital. Therefore, much could be gained from correcting failure in organizational learning [46] as it might affect large parts of the organization [47,48]. A third factor is that the NBHS might be somewhat biased towards management. This may have led to the fact that particularly failure in organizational learning had a strong correlation in our analyses.

5.1. Implications

Our observations have highlighted the systemic level (organizational learning) to optimize the contact between the professional and organizational parts with a B = 1.69. Consequently, if the hospitals invest in organizational learning they will gain 1.69 times more outcome for each unit they put into improving the conditions for this entity. Such outcome may be less money spend, less unfavorable events or more attentions to check lists for various procedures.

5.2. Limitations

This study has several limitations. First, it can be argued that auditing 20 reports from the NBHS is too few for a quantitative analysis. However, these 20 reports constitute a complete year (2018) of consecutively published reports from on-site inspections by the NHBS. In addition, the scoring model led to 392 scores (median 112) on the various variables in the 20 reports, which represent a good basis quantitative analysis.

Another possible limitation is lack of validation of the applied scoring form in comparison with guidelines for validating a questionnaire [49]. We may therefore have had a lower construct validity than a standard questionnaire validation. This was probably compensated by the fact that the scoring form was quite simple (only three main variables) and that two researchers went through all the reports and came to a consensus. When in doubt, a third researcher was involved.

The NHBS tendency to critically assess the management’s role in the reports may have introduced a bias and thus a type 1 error [50]. Furthermore, the low number of scores on some variables can lead to a type 2 error, but several variables had sufficient scores so that type 2 errors are probably not present in this study [50]. The external validity is acceptable as we used 20 consecutive current reports representing several medical fields. However, the external validity could probably be improved by adding more reports and thus covered a broader field of medical specialties.

5.3. Future research

In the future, one may have to look at interaction and the safety perspective in a new way, as in the Safety-II approach. This new paradigm has evolved over the last 5 years [51]. The classic safety thinking that involves analyzing what went wrong is called Safety-I thinking [44]. From a modern security perspective, this is turned upside down, where one asks: What went well? - and why? Such thinking is called the Safety-II approach [52]. The five core points of the Safety-II approach are: (1) Definition of safety, which is the ability to create a fortunate outcome, (2) Handling safety, (3) The focus on human recourses, (4) Accident investigation and (5) Risk assessments focused on analysing why things are going well, so that one can understand what happens when adverse events and errors occur [52]. According to Turner, such an approach will be complementary to the classic Safety-I universe [44]. To increase resilience in both the Safety I and II contexts, the importance of learning from systematic training and simulation is emphasized [53].

The Braut model is likely to be well suited to address Safety-II thinking in the future [54]. We can probably consider functional contact between the professional leg and the system leg in the same way as the present study has analysed failure in contact in accordance with Safety-I thinking. It will be exciting to conduct this analysis in the light of the Safety-II approach.

Kim and Mauborgne tell managers that ``people care about the decisions you make, but they care even more about the process you used along the way" [55]. They recommend adhering to three principles of a fair process: expectation clarity, explanation and engagement. This invites appropriate structural management while taking into account the right to self-determination. Then you want to breathe new life into the patient safety culture in the hospitals by working towards a functioning contact between the profession and the organization. Thus, the principles from the Braut model could be important for hospital managers in the future.

6. Conclusion

Empirical use of the Braut model in the analysis of the 20 audit reports from the NBHS identifies failure in organizational learning as the explanatory variable for failure in contact in the hospitals. Furthermore, a change in organizational learning of one unit will improve the contact by a factor of 1.69.

We would like to emphasize that doctors and nurses must not be deprived of their professional right to autonomy, as this entity does not have any significant effect on contact between professional autonomy and organizational learning. Doctors and nurses should instead be given a sound framework so that they can work safely. Furthermore, the organizational learning and management underpins the professional understanding. Therefore, it is important to maintain professional soundness as a profession-based and cultural phenomenon, while at the same time building up and strengthening organizational support. These endeavors will be a small step for the institutions, but a giant leap for improving resilience in Norwegian hospitals.

Footnotes

Acknowledgements

The authors thank Prof. Jan Terje Kvaløy, University of Stavanger for the statistical support.

Conflict of interest

None of the authors have any conflict of interest to report.

Informed consent

Not applicable.

References

Jha

Larizgoitia

Audera-Lopez

Prasopa-Plaizier

Waters

Bates

. The global burden of unsafe medical care: Analytic modelling of observational studies. BMJ Qual Saf. 2013;22(10):809–15.

World Health Organization. Patient Safety Fact File; 2019 [Cited 2020 Oct 7]. Available from: https://www.who.int/features/factfiles/patient_safety/patient-safety-fact-file.pdf?ua=1.

Norwegian Directory of Health. Patient injuries in 2018. In Norwegian: Pasientskader i Norge 2018. Helsedirektoratet; 2019. [Cited 2020 Oct 1]. Available from: https://www.helsedirektoratet.no/rapporter/pasientskader-i-norge.

Brennan

Leape

Laird

Hebert

Localio

Lawthers

Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. N Engl J Med. 1991;324(6):370–6.

Kirkup

, The Report of the Morecambe Bay Investigation 2015; 2015. [Cited 2020 Nov 1]. Available from: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/408480/47487_MBI_Accessible_v0.1.pdf.

Francis

, Report of the Mid Staffordshire NHS Foundation Trust Public Inquiry; 2013. [Cited 2020 Sept 30]. Available from: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/279124/0947.pdf.

Oikonomou

Carthey

Macrae

Vincent

. Patient safety regulation in the NHS: Mapping the regulatory landscape of healthcare. BMJ Open. 2019;9(7):e028663.

Grote

. Leadership in resilient organizations. Wiik S: Exploring Resilience. A scientific Journey from Practice to Theory. Springer Open; 2019.

Aven

Boyesen

Njå

Olsen

Sandve

. Security in the Society Norwegian: Samfunnsikkerhet. Oslo, Norway: Universitetsforlaget; 2004.

10.

Macrae

. Early warnings, weak signals and learning from healthcare disasters. BMJ Qual Saf. 2014;23:440–5.

11.

Ministry of Health and Care Services. National Plan of Health and Care 2011 to 2015 (2010). In Norwegian: Nasjonal helse- og omsorgsplan 2011-2015; 2010. [Cited 2020 Oct 15]. Available from: https://www.regjeringen.no/no/dokumenter/meld-st-16-20102011/id639794/?ch=1.

12.

Beauchamp

Childress

. Principles of biomedical ethics: Marking its fortieth anniversary. Am J Bioeth. 2019;19(11):9–12.

13.

Norwegian Official Report 2015, no. 11. With shown cards. Prevention and follow-up of serious incidents in health and care services. In Norwegian: Med åpne kort. Forebygging og oppfølging av alvorlige hendelser i helse- og omsorgstjenestene. NOU 2015:11 Departementenes sikkerhets- og serviceorganisasjon; 2015 [Cited 2020 Sept 3]. Available from: https://www.regjeringen.no/contentassets/daaed86b64c04f79a2790e87d8bb4576/no/pdfs/nou201520150011000dddpdfs.pdf.

14.

Iflaifel

Lim

Ryan

Crowley

. Resilient health care: A systematic review of conceptualisations, study methods and factors that develop resilience. BMC Health Serv Res. 2020;20(1):324.

15.

The Norwegian Board of Health Supervision. The Annual Supervison Report. 2014. In Norwegian: Helsetilsynet. Tilsynsmelding; 2014. [Cited 2020 Aug 30]. Available from: https://www.helsetilsynet.no/globalassets/opplastinger/Publikasjoner/tilsynsmelding/tilsynsmelding2014.pdf/.

16.

Braut

Holmboe

. Patients safety – the current structure Pasient safety. In: Aase

(ed) Universtetsforlaget; 2018.

17.

Oyri

Braut

Macrae

Wiig

. Exploring links between resilience and the macro-level development of healthcare regulation-a Norwegian case study. BMC Health Serv Res. 2020;20(1):762.

18.

Jeppesen

Wiig

. Resilience in a prehospital setting - a new focus for future research? Scand J Trauma Resusc Emerg Med. 2020;28(1):104.

19.

Øyri

Wiig

. Regulation and resilience at the macro-level healthcare system – a literature review. Proceedings of the 29th European Safety and Reliability Conference. In: Beer

Michael

Zio

Enrico

(eds) 2019.

20.

Lindøe

Kriingen

Braut

GS,

Risk and supervision. Risk management, supervision and legal regulation. Norwegian: Risiko og tilsyn. Risikostyring og rettslig regulering. 2nd ed. Universitetsforlaget; 2015.

21.

Norwegian Ministry of Health and Care Services. Assignment to the Norwegian Directorate of Health to prepare draft on common regulations for internal control and systematic work with quality improvement, and patient and user safety in the health and care service. In Norwegian: Oppdrag til Helsedirektoratet om å utarbeide utkast til felles forskrift for internkontroll og systematisk arbeid med kvalitetsforbedring og pasient- og brukersikkerhet i helse- og omsorgstjenesten; 2013.

22.

Norwegian Ministry of Health and Care Services. Hearing – Regulations on management systems in the health and care service. In Norwegian: Høring -Forskrift om styringssystem i helse- og omsorgstjenesten; 2016. [Cited 2020 Oct 30]. Available from: https://www.regjeringen.no/no/dokumenter/horing%2D%2Dforskrift-om-styringssystem-ihelse%2D%2Dog-omsorgstjenesten/id2459663/.

23.

Norwegian Ministry of Health and Care Services. Regulations on management and quality improvement in the health and care service. Royal Resolution. In Norwegian: Forskrift om ledelse og kvalitetsforbedring i helse- og omsorgstjenesten. Kongelig resolusjon. Ministry of Health and Care Services; 2016. [Cited 2020 Nov 10]. Available from: https://lovdata.no/dokument/LTI/forskrift/2016-10-28-1250.

24.

Norwegian Directorate of Health. Guidelines to Regulation on management and quality improvement in the healthcare services. In Norwegian: Veileder til forskrift om ledelse og kvalitetsforbedring i helse- og omsorgstjenesten; 2017. [Cited 2020 Nov 10]. Available from: https://www.helsedirektoratet.no/veiledere/ledelse-og-kvalitetsforbedring-i-helse-og-omsorgstjenesten.

25.

Van de Bovenkamp

Stoopendaal

Bal

. Working with layers: The governance and regulation of healthcare quality in an institutionally layered system. Public Policy Adm. 2017;32:45–65.

26.

Norwegian Ministry of Health and Care Services. Circula on management in hospitals. In Norwegian: Lederansvar i sykehus Rundskriv I-2/2013; 2013. [Cited 2020 Nov 10]. Available from: https://www.regjeringen.no/no/dokumenter/rundskriv-i-2013/id728043/.

27.

Norwegian Ministry of Health and Care Services. Guide to regulations on management and quality improvement in the health and care service. In Norwegian: Veileder til forskrift om ledelse og kvalitetsforbedringi helse- og omsorgstjenesten. Rundskriv IS-2620, §8e; 2017. [Cited 2020 Nov 10]. Available from: https://www.helsedirektoratet.no/veiledere/ledelse-og-kvalitetsforbedring-i-helse-og-omsorgstjenesten.

28.

Norwegian Ministry of Health and Care Services. National action plan for patient safety and quality improvement. In Norwegian: Nasjonal handlingsplan for pasientsikkerhet og kvalitetsforbedring; 2019. [Cited 2020 Nov 10]. Available from: https://www.helsedirektoratet.no/veiledere/ledelse-og-kvalitetsforbedring-i-helse-og-omsorgstjenesten/Nasjonal%20handlingsplan%20for%20pasientsikkerhet%20og%20kvalitetsforbedring%202019-2023.pdf/_/attachment/inline/79c83e08-c6ef-4adc-a29a-4de1fc1fc0ef:94a7c49bf505dd36d59d9bf3de16769bad6c32d5/Nasjonal%20handlingsplan%20for%20pasientsikkerhet%20og%20kvalitetsforbedring%202019-2023.pdf.

29.

Norwegian Board of Health Supervision. Notification of serious incidents in the specialist health service. In Norwegian: Varsel om alvorlige hendelser i spesialisthelsetjenesten; 2019. [Cited 2020 Nov 10]. Available from: https://www.helsetilsynet.no/tilsyn/varsel-om-alvorlige-hendelser/varsel-om-alvorlige-hendelser-i-spesialisthelsetjenesten/.

30.

Norwegian Board of Health Supervision. Overview - notification of a serious incident in the health and care service. In Norwegian: Oversikt – varsel om alvorlig hendelse i helse- og omsorgstjenesten; 2019. [Cited 2020 Nov 10]. Available from: https://www.helsetilsynet.no/tilsyn/varsel-om-alvorlige-hendelser/oversikt/.

31.

Norwegian Board of Health Supervision. Internal control in hospitals - with main emphasis on the quality committees ‘function in the companies’ internal non-conformance treatment 2001. In Norwegian: Internkontroll i sykehus – med hovedvekt på kvalitetsutvalgenes funksjon i virksomhetenes interne avviksbehandling 2001; 2001. [Cited 2020 Oct 15]. Available from: https://www.helsetilsynet.no/historisk-arkiv/landsomfattende-tilsyn/internkontroll-kvalitetsutvalget-/.

32.

Wiig

Braithwaite

Braut

. The evolution of investigation methods and practices in healthcare. Inside Hazardous Technological Systems. In: Gould

Macrae

(eds) Abingdon: CRC Press; 2021.

33.

Conduct and Interpret a Bivariate (Pearson) Correlation. [Cited 2020 Oct 4]. Available from: https://www.statisticssolutions.com/bivariate-correlation/.

34.

Linear Regression. [Cited 2020 Oct 4]. Available from: http://www.stat.yale.edu/Courses/1997-98/101/linreg.htm.

35.

Berg

Akerjordet

Ekstedt

Aase

. Methodological strategies in resilient health care studies: An integrative review. Saf Sci. 2018;110:300–12.

36.

Dekker

. The bureaucratization of safety. Saf Sci. 2014;70:348–57.

37.

Wiig

Schibevaag

Zachrisen

Hannisdal

Anderson

Haraldseid-Driftland

. Next-of-kin involvement in regulatory investigations of adverse events that caused patient death: A process evaluation (part ii the inspectors’ perspective). J Patient Saf. 2021;17(8):e1713–e1718.

38.

Hood

Rothstein

Baldwin

. The Government of Risk: Understanding Risk Regulation Regimes. Oxford University Press; 2001.

39.

Vandraas

KF,

Doctors commit suicide. In Norwegian: Leger tar livet sitt. Onconytt. 2019.

40.

Bates

Singh

. Two decades since to err is human: An assessment of progress and emerging priorities in patient safety. Health Aff (Millwood). 2018;37(11):1736–43.

41.

Wiig

Lindø

. Patient safety in the interface between hospital and risk regulator. J Risk Res. 2009;12(3–4):411–26.

42.

Gandhi

Kaplan

Leape

Berwick

Edgman-Levitan

Edmondson

Transforming concepts in patient safety: A progress report. BMJ Qual Saf. 2018;27(12):1019–26.

43.

Wiig

Aase

Billett

Canfield

Roise

Nja

RiH-team. Defining the boundaries and operational concepts of resilience in the resilience in healthcare research program. BMC Health Serv Res. 2020;20(1):330.

44.

Turner

Pidgeon

. Man Made Disaster. 2nd ed. London: Butterworth - Heinemann; 1997.

45.

Norwegian Board of Health Supervision. Report from Norwegian Board of Health Supervision. In Norwegian: Rapport fra Helsetilsynet 2020. [Cited 2020 Sept 24]. Available from: https://www.helsetilsynet.no/publikasjoner/rapport-fra-helsetilsynet/.

46.

Heimer

. Resilience in the middle. Contributions of regulated organizations to regulatory success. Ann Am Acad Pol Soc Sci. 2013;649:139–56.

47.

Chatburn

Macrae

Carthey

Vincent

. Measurement and monitoring of safety: Impact and challenges of putting a conceptual framework into practice. BMJ Qual Saf. 2018;27:818–26.

48.

Simon

. Compliance and high reliability in a complex healthcare organization. Front Health Serv Manag. 2018;34(4):12–25.

49.

Suri

Verna

. Questionnaire validation made easy. Eur J Sci Res. 2010;46(2):172–8.

50.

Aspkota

, Type I error and type II error with 10 differences. [Cited 2020 Oct 20]. Available from: https://microbenotes.com/type-i-and-type-ii-error/.

51.

Braitwaithe

Wears

Hollnagel

. Resilient health care: Turning patient safety on its head. Int J Qual Health Care. 2015;27:418–20.

52.

Leistikow

Bal

. Resilience and regulation, an odd couple? Consequences of Safety-II on governmental regulation of healthcare quality. BMJ Qual Saf. 2020;29(10):1–2.

53.

Sollid

SJM

Dieckman

Aase

Soreide

Ringsted

Ostergaard

. Five topics health care simulation can address to improve patient safety: Results from a consensus process. J Patient Saf. 2019;15(2):111–20.

54.

Hollnagel

. Safety-I and safety-II The Past and Future of Safety Management. 1st ed. London: CRC Press, Taylor & Francis Group; 2014.

55.

Kim

Mauborgne

. Tipping point leadership. Harv Bus Rev. 2003;81(4):60–9, 122.

No	Independent variables	Part of the report
		Total score	Descriptive score	Analytical score
1	Functional professional autonomy	A	B	C
2	Failure in professional autonomy	D	E	F
3	Functional organizational learning	G	H	I
4	Failure in organizational learning	J	K	L
5	Functional contact between professionalautonomy and organizational learning	M	N	O
6	Failure in contact between professionalautonomy and organizational learning	P	Q	R