The decision to follow a sequential procedure was made in order to factor in qualitative answers about the participants’ self-assessed FR—adequate or inadequate—during the interviews.
For the quantitative part, 138 patients were recruited consecutively in a geriatric hospital. Inclusion criteria were defined as follows: an age 65 years onwards, an expected ability to walk and climb stairs with the full load during the hospital stay, sufficient vision to complete a questionnaire, and the participant’s ability to consent. Exclusion criteria were: the presence of severe heart failure (NYHA IV) or peripheral arterial occlusive disease (PADD) of the lower extremities (grade 3+), amputation of one or both lower extremities, poor understanding of the German language, severe cognitive deficits or mental illness, and the lack of written consent.
During the hospital stay, 38 of the 138 participants dropped out of the study. The reasons for dropout were: withdrawal of consent, inadequate response, early hospital discharge, deterioration in the general condition (physical or mental). Data collection lasted from September 2016 to March 2018.
Quantitative data collection was undertaken during the hospital stay. All participants were included within the first five days after admittance to hospital. Directly after inclusion, sociodemographic data of the study participants, as well as data on falls in the recent past 12 months, perceived general self-efficacy and social activities in the last six months were collected. Within three days before discharge, data on the confidence to maintain balance in everyday situations were recorded using the Activities-specific Balance Confidence Scale (ABC Scale). Additionally, the de Morton Mobility Index (DEMMI), a mobility test for geriatric patients whose score is associated with falls, was performed (T1).
The ABC Scale captures the subjective confidence to maintain balance without staggering during 16 everyday defined activities (e.g., climbing stairs, using an escalator) (11). The validated German version of the ABC Scale was used (12). While the original scale evaluates the individual items on a scale between 0 and 100 points, the present study uses a Likert-scale version with the following response categories: 0 = not at all confident, 1 = a little confident, 2 = confident, 3 = very confident (13). The overall score is thus in line with this survey, i.e., between 0 and 48 points. A cutoff value of less than 67% identified subjects with an increased FR (14). The ABC Scale shows a high test reliability (15) as well as a good discriminant and convergent validity (11).
The DEMMI was developed as a measurement tool for the assessment of the general mobility of the elderly, who are located in an acute care or geriatric rehabilitation facility (16). The 15 items of the DEMMI check the ability of a person to change position and location, with or without assistance from the investigator. Likewise, a person’s walking and balancing abilities are tested. The items are hierarchically ordered according to the level of difficulty (16). The addition of all 15 items’ ratings gives a raw score with a range of 0–18, which is then transformed into a score with a span of 0–100. Zero indicates the lowest and 100 the highest general mobility. All data from the ABC Scale and the DEMMI were inserted into a 4x4 contingency table to compare the objective FR with the subjective self-assessment (FOF) of the participants. The self-assessment was deemed “adequate” if both objective and subjective assessments were either positive (FR and FOF) or negative (no FR and no FOF). Divergences from the concurrent results of the two assessments were deemed “inadequate”.
Qualitative data were recorded six months after discharge (T2). Sixteen participants were recruited and interviewed at home regarding their actual mobility, everyday activities and coping strategies due to mobility limitations and FOF. Interviewees were recruited as a convenience subsample in the follow-up and agreed to an interview according to the order in which they were contacted. Four potential participants rejected an interview. The number of subjects was based on the principle of theoretical saturation (17) (18). Theoretical saturation here means the understanding of theoretical saturation according to Glaser and Strauss “Saturation means that no additional data are being found whereby the sociologist can develop properties of the category.” (18) All interviews were recorded using audio equipment and transcribed based on predefined transcription rules. The category formation was performed deductively in a first step; based on a literature search was carried out with the keywords fear of falling, dealing with everyday life aspects, walking aids and coping. And based on expert discussions with members of various professions (physiotherapists, doctors, nurses) and affected persons and their relatives. In a second step, the examination units were explored iteratively-inductively. In order to ensure high intercoder reliability, the interviews were coded and categorized by two researchers (“deleted for blinding purposes”.) with the help of the coding frame with the categories worked out. Both researchers have extensive experience of planning, conducting and analyzing qualitative interviews. The coding guide was added after reviewing the first four interviews and after reflection by a second rater with anchor examples as well as other categories.. Four interviews with the deductively developed coding guideline were coded, reflected by both researchers and four further categories with definitions, coding rule and anchor example were worked out. Then the already coded four interviews were re-encoded, reflected again in the team and the categories were found to be appropriate and a further 12 interviews were coded with the modified coding guideline. Subcategories were worked out and deleted or added in a deductive approach and then elaborated into main categories (coping strategies (active, passive, indifferent) and awareness). The ratio of deductive to inductive approach is nine to four.
The qualitative content analysis was performed according to Mayring. This took the form of a structuring content analysis; more precisely, content structuring in which the material obtained for certain content (here, according to the coding guide including its inductive additions) is extracted and summarized (19).