Disease knowledge and patient education are key players for a better quality of life in vascular surgery patients

A prospective observational study including consecutive patients hospitalized forwas conducted from January 2015 until November 2016 at the Department of General, Visceral, Vascular, and Transplantation Surgery of the Magdeburg University. Exclusion criteria were inability to give informed consent, decline of treatment or study participation, and ambulant or emergency surgery.

Participants were assigned to standard IC talk (standard group) between January 2015 and August 2016 (excluding February 2016), a one-hour long extended IC talk using an additional PowerPoint presentation in February 2016 (PP group), or a walking diary between September 2016 and November 2016 (WD group). When one of these treatments was indicated, the patients were explained the diagnosis and the recommended and alternative treatment strategies by a vascular surgeon at the ambulatory consultation and were provided with the corresponding German standard IC forms (Thieme proCompliance GmbH, Erlangen, Germany) for self-information at home. The standardized IC talk was then performed about 1 week before hospital admission using this form. They were further asked about their willingness to participate in the study. In case of consent, they were included into the present investigation after assessing the absence of exclusion criteria and signing the IC form. After the standardized IC talk (prepared as PowerPoint [Microsoft Inc., Redmond, WA, USA] presentation about AAA, TEA, or DSA and explained to the appropriate patients), a one-hour long patient education was performed in the talking group. The patients had the opportunity to ask for more details and explanation concerning the diagnosis, lifestyle changes, and the techniques and possible complications of the planned procedure. After the indication for surgery, the walking group received flyers about their disease and the importance of walking training and a walking diary plus pedometer. The patients were instructed to note down their steps every day according to the pedometer until the follow-up care. On the day of hospital admission, as shown in Fig. 1, the patients received a first questionnaire on the day of admission, before the planned procedure, containing 20 questions and evaluating the knowledge about the disease, self-information, smoking habits, and treatment expectations. In order to quantify the knowledge level, six items were asked and correct answers summed up: diagnosis, planned treatment, incision site, and three possible complications during or after the intervention. The answers were considered correct if the patients verbalized a synonymous word (i.e., “vessel dilatation” instead of “aneurysm”) and were considered incorrect if the patient left the field empty or filled in something completely different (i.e., “hypertension” instead of “peripheral atherosclerosis”). After surgery and before discharge, they completed a second questionnaire containing 8 items evaluating treatment satisfaction. During follow-up, within 3 up to 6 months, they filled out a modified standardized quality of life questionnaire (RAND 36-Item Health Survey [version 1.0]) [4] combined with parts of the World Health Organization Quality of Life questionnaire [5].

This study was approved by the Otto-von-Guericke University ethics board and registered at the German Register of Clinical Trials (study No. DRKS00014050). The patient data were pseudonymized and stored according to the legal requirements of the German district Saxony-Anhalt. The decision of the patient to take part in the investigation or not did not influence any decision regarding care, diagnostics, and treatment. Both the study participation consent form and the questionnaires were blinded to the health care providers. All procedures involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments for biomedical research or comparable ethical standards.

The participants were asked in the first questionnaire to write down a keyword they remembered from the IC talk regarding the items incision site, their diagnosis, their planned treatment, and three possible complications of the treatments. Synonymous words were also counted as correct, i.e., atherosclerosis instead of PAOD. The number of correctly remembered keywords were summarized in the variable “knowledge,” which ranged from zero (did not remember any item) up to six (remembered all items including three possible complications).

Comparisons between groups were made using the Fisher’s exact test for binary variables and independent t-test for continuous variables if normally distributed, and the U test by Mann and Whitney if not. A p-value less than 0.05 was accepted as significant. Multiple linear regression analysis (stepwise backward algorithm) was used to identify independent predictors of the variable knowledge. Statistical analysis was conducted using R (R Foundation for Statistical Computing, Vienna, Austria) [6]. The following packages were used for the corresponding statistical methods: MatchIt [7] for matched-pair analysis using the genetic algorithm, multivariate analysis of variance (MANOVA) using mvnormtest [8] for the test on multivariate normality, and the manova function of the standard software.

From January 2015 until November 2016, n = 215 patients corresponded to the inclusion criteria and 198 patients were enrolled and subdivided into groups as shown in Fig. 1. The first questionnaire was filled out by n = 156 patients, without missing values (MV) by n = 148 patients. One of the aims of the study was to compare the knowledge levels between the standard, PP, and WD groups. Since patient characteristics may differ between the groups and may cause confounding bias, a matched-pair analysis was conducted as demonstrated by the QQ plots in Fig. 2a, where the standard group was matched by the two extended IC talk groups taken together. Fig. 2b shows the boxplots of the mean and median knowledge levels in the three subgroups, which did not differ relevantly. A one-way analysis of variance (ANOVA) was performed and significant differences could neither be calculated by the linear model (p : 0.584, aov function of the R software) nor by the post-hoc analysis (PP and WD vs. standard group, p : 0.884 and p : 0.877, respectively, and PP vs. WD, p : 0.997). Furthermore, we performed a linear regression analysis in order to determine the cofactors associated with the knowledge level after the IC talk, as listed in Table 1. Age and the PAOD-associated interventions DSA and TEA contributed to lower and the statement of the patient that she/he has read the IC talk form already before the IC talk to higher knowledge level. However, as shown in Fig. 2c, d, the dependencies of the knowledge level on age are not unequivocally observable in the different subgroups since knowledge did not decrease with age in patients operated on for AAA (Fig. 2c), and even increased with age in the WD subcohort (Fig. 2d).

Summarizing, we can state that an extended IC talk via PP or WD did not significantly improve the disease knowledge. Determining factors were age, PAOD (TEA and DSA) compared to AAA, and self-information by reading the IC form prior to the IC talk.

The knowledge level of the patients and its contributing factors may also affect the treatment satisfaction and the compliance of patients. The second questionnaire, filled out on the day of discharge, contained questions about the satisfaction with the staff and the treatment. Because the number of completed questionnaires was lower than that of the first questionnaire, a new matched-pair analysis was performed. The QQ plots are outlined in Fig. 3a. The patients were asked to score the staff by grades from 1 (very good) up to 5 (insufficient). Fig. 3b–e show the comparisons of the mean scores between the standard and extended IC talk groups by box-whisker plots for the treatment (b, p : 0.61), the surgeons (c, p : 0.89), the nursing staff (3D, p : 0.78), and the general condition (e; by the scoring 0: feeling worse than before treatment; 1: feeling equally; 2: feeling better; p : 0.65), without significant differences. An extended IC talk did not relevantly improve the short-term satisfaction. However, most patients evaluated treatment, situation, and staff as “good” or “very good.” Therefore, it was not possible to draw any conclusions on the relationships between education and treatment satisfaction.

Since treatment expectations and long-term satisfaction with treatment may also depend on the knowledge level about the disease, the mid-term QoL was assessed.

The RAND36 questionnaire was completed by n = 63 patients (29%) during the ambulatory interval of 6–12 months after treatment, among them n = 7 in the extended IC talk groups. A new matched-pair analysis was performed, the QQ plots of which are shown in Fig. 4a. The mean scores of the eight domains of the RAND 36 Health Survey 1.0, which reached from zero (low QoL) up to 100 (high QoL), are outlined in Fig. 4b–i. The mean scores were higher in the extended IC talk group in each of the domains. However, significance was reached only in the domain “role limitation: physical health” (12 vs. 65; p : 0.022), and a trend toward a higher mean value was observed in the education group for the domain “physical functioning” (42 vs. 69, p : 0.07). A discriminant analysis was conducted, which determined the reciprocal relationships between the QoL domains and the grouping variable, thereby taking the cross-correlation between the domains into account (Fig. 4k). Table 2 lists the LD coefficients. The slightly overlapping histograms and the positive and negative coefficients indicate that a higher score did not unequivocally implicate membership in the extended IC talk group for all domains (starting from an a priori probability of 50%). It seems that other factors influenced the QoL as well.

Fig. 4 shows the scatterplots of the QoL vs. knowledge level for the RAND-36 domains in the Fontaine stage I and II subcohorts subdivided into the treatment groups, demonstrating the statistical associations between knowledge and QoL. There were positive associations between QoL and knowledge in AAA patients for each domain. The same was true in the DSA subcohort, except for the “role limit: physical health” domain, where it was negative (Fig. 4b). The TEA subgroup contained only a few cases which had completed the third questionnaire, and therefore was not conclusive. Considering the inter- and cross-correlations between the cofactors, which influenced the QoL during the disease course, a MANOVA was performed including the eight QoL domains as response variables and the cofactors listed in Table 3 as explanatory variables. There were no significant associations in the multivariate analysis. However, the follow-up ANOVA revealed that belonging to the extended IC talk groups (PP or WD) was associated with the physical functioning (PF) and the social functioning (SF) domains. The Fontaine stage showed an impact on PF and emotional wellbeing (EWB) and the knowledge level on EWB. Further associations were observed between marital stage and SF, between care provision and general health (GH), between monthly income and GH, and between various interaction terms and the QoL domains. Because of lacking normality distributions of the QoL values within the domains (significant Shapiro–Wilk multivariate normality test results, data not shown, see also Fig. 4a–f) and low case numbers, the conclusiveness was limited. However, considering the relative robustness of the Pillai’s trace regarding unbalanced data and the graphically demonstrated relationships in the scatterplots of Fig. 4a–f, these results pointed to statistical associations between the extended IC talk, the resulting knowledge level, and mid-term QoL. Other factors such as social status, PAOD stage, or ambulatory care provision are also among the determinants.