Framework of the study
The FRESH- score was developed within the framework of the BFCC project (Baltic Fracture Competence Centre)[2], a transnational project supporting innovation in fracture treatment in the Baltic Sea region. A registry for recording bone fractures and their complications was developed as part of the BFCC project and piloted from November 2017 to February 2018 at the University Hospital Schleswig-Holstein, Lübeck Campus (UKSH), a level I trauma centre.
During the period from 15.11.2017 to 15.02.2018, a total of 237 patients treated with acute fractures at UKSH Campus Lübeck, were consecutively included in the registry. These fulfilled the following inclusion:
- Men and women, 18 years and older
- Presence of an acute fracture of an extremity, the pelvis or the clavicle, which was not older than seven days at the time of admission to the hospital, and which had not been pre-treated or healed.
- Initial treatment of the fracture at the instituition
- Ability of the study-participant to understand the nature of the study
- Signed informed consent form
Exclusion criteria:
- Males and women under 18 years of age
- Patients not capable of giving consent, who do not understand the background, significance, and consequences of the study
- Lack of informed consent
An ethics committee vote for the project at the UKSH Lübeck Campus was obtained at the meeting of the Ethics Committee in Lübeck on 5.10.2017 under file number 17-267.
Conceptual structure and development process of the questionnaire
The FRESH- score was based on the following existing and validated questionnaires for recording patient-reported outcome:
- Firstly the "Disability of Arm, Shoulder, Hand-Score" (DASH), published in 1996 [3] and validated in 1999. [4,5] The Score assigns each answer possibility a value and grades the sum of all values in an unadjusted manor to evaluate the mobility of the upper extremity. This evaluation technique was adapted for the FRESH- score.
- Secondly the Foot and Ankle Disability Index (FADI), first described in 1999, [6]. It consists of an extremity specific part and a general part. In the same way the FRESH- score starts with a questions on the general health status and in a separate section addresses fracture- treatment specific questions[7].
- Additionally, the EQ-5D-5L, a measurement instrument to assesses health status through a standardised questionnaire expressing the respondent's health status in a one-dimensional measure ranging from 0 (very poor) to 1 (best possible health status), originally developed in 1987 as a patient self-report instrument. Its visual analogue scale on the general health state adapted.
- The method to choose from four different answers was adapted from the American Shoulder and Elbow Surgeons Standardized Shoulder Assessment Form (ASES)[8]
The above-mentioned scores differ in their setup and cannot be applied to monitor overall treatment results e.g., of orthopaedic departments usually treating more than one fracture entity or in fracture registries, which are not limited to monitor treatment results for one specific extremity. The aim of the FRESH- score is to fill this void. It consists of two sections (Appendix 1). As the study took place in Germany, all questionnaires were written in German language. A translation is also attached (Appendix 2).
The first part addressed the self- assessment of the general state of health with 5 possible answers: very good, satisfactory, average, unsatisfactory, poor. Patients were asked to choose the answer that best describes their state of health at the present time. With this section, it was possible to correlate and separate the general health status from fracture treatment related health issues.
In the second part, the patients were asked to choose to what extent the following applies to them: “Since my bone fracture treatment...
- I increasingly suffer from pain
- I have problems sleeping
- I have problems with everyday tasks
- I take more painkillers
- I've had to give up hobbies”
Each of these 5 questions had 4 different choices, from left to right: not at all, hardly, significantly, strongly as in a "Likert scale” [9], ranked with 4 to 1 points. The resulting scores were added together to the score result. The score is evaluated as follows:
- 20-16 Satisfactory quality of life after fracture treatment.
- 15-11 Quality of life hardly worsened after treatment
- 10-6 Quality of life significantly worsened after treatment
- 5 Severe impairment of quality of life
The FRESH- score was validated, in combination with the EQ-5D-5L questionnaire[10], which was published by "EuroQol" and validated in various medicals fields[10,11] and different countries[12,13]. Both questionnaires were sent with a cover letter in German for handwritten response by post on 16/09/2019 to all 237 patients enrolled in the registry.
The EQ-5D-5L has 5 questions with five possible answers. In contrast to the FRESH- score, they were not evaluated with a simple summation-score but calculated with a country specific population- adapted formula and help of the official „EQ-5D-5L Calculator". Each question of the EQ-5D-5L has five choices and each of these choices is assigned a number from 1-5 added together to a numerical code. For example, it could look like this: 32231 and describes a patient who has moderate difficulties in mobility, mild problems washing and dressing himself, mild problems with daily tasks, moderate pain and no depression. It is then converted into an EQ-5D-5L Index value, which is between zero and one. Here, zero is the worst possible state of health and one is the best possible state of health.[11]
Registration of complications
Since the FRESH- score was primarily designed to assess patients with complications after fracture treatment, it was recorded whether the patients had complications or not, for better delimitation, making a subdivision into patients with and without complications possible (see Appendix 3). A translation can be found in Appendix 4. Hypothesis was, that patients without complications should have a higher FRESH- score than patients with complications.
First, patients were asked to choose whether they suffered from a complication. If yes, they were asked to chose from a total of 8 complication categories available, which had been established within the framework of the BFCC project and applied in Lübeck with this registry.[14]
Afterwards, the respondent was further asked about the time of the complication. There were four possibilities to choose from: before the operation, during/directly after the operation, during the hospital stay, or after the hospital stay.
Finally, the patient was asked whether the complication resulted in permanent limitations.
Questionnaires in paper form were mailed on 16.09.2019 from the University Hospital in Lübeck to all 237 patients.
After receiving the completed questionnaires, a pseudonymised spreadsheet was created for data collection for each part of the survey.
Statistical analysis and hypothesis testing
The analysis was performed using the software R Statistics version 3.5.1 and RStudio (both R Consortium, Boston, MA, USA), version 1.1.456.
Reliability of the score
Reliability describes the extent to which the results correspond to the state of health. For this purpose, the internal consistency was calculated, and the reliability tested on the basis of this. The different questions in the FRESH- score and EQ-5D-5L were the domains that were compared and checked for their agreement.
The internal consistency, the strength of the relationships of the individual factors of each domain to each other, was measured using Cronbach's Alpha. If Cronbach's Alpha values are below 0.70, questions within a domain were asked differently. If the values are close to 1.0, this indicates redundant questions.[15]
Convergent validity
Convergent validity represents the correlation between different tests measuring the same construct. The correlations determined should be high for a valid test[16]. This leads to the following hypothesis on convergent validity: The FRESH- score measured quality of life after fracture treatment similar to the validated EQ-5D-5L.
Convergent validity was calculated by calculating the Pearson coefficient. If the coefficient is between 0,0< r <0,2, there is no to little linear correlation, a weak to moderate linear correlation is present in a range of 0,2< r <0,5 [17]. A coefficient of 0,5< r <0,8 would represent a clear linear relationship and a value of 0,8< r <1,0 a strong one[17].
Construct validity
Construct validity measures the extent to which the instrument actually displays the characteristic being assessed[18]. In the case of the FRESH- score, this is to be achieved with the help of patient satisfaction after fracture treatment and the objective recording of complications by the hospital physicians in the registry.
Construct validity was determined using the Wilcoxon rank sum test to determine whether there was a relationship between the results of the two samples. The prerequisite for the Wilcoxon rank sum test is two independent samples, which was given with the EQ-5D-5L and the FRESH- score. The alternative hypothesis was that the distribution of one of the samples has a skew from the median to the left or right.
For the null hypothesis a 5% significance level was chosen and the alternative hypothesis that the distribution of the scores of both samples have a high similarity.
Content validity
Regarding the completeness of symptom questioning and question comprehensibility. To what extent do possible difficulties in filling it out affect the results of the study?[18]
In order to evaluate the comprehensibility, the so-called "missing values" were recorded.