Activation of older patients through PRACTA intervention for primary healthcare doctors: does the method matter?

Abstract

Background

Activating older adults in medical practice can benefit patients’ health and quality of life, as well as the economy and burden of the healthcare system. Placing general practice in the public healthcare system gives the elderly population easy access to the promotion of active attitudes toward health, provided that the doctors have and use relevant activating tools. The aim of this study was to verify the possibility of activating senior patients through an educational intervention for doctors.

Methods

Two waves of data collection from primary care patients and their doctors were separated by an intervention for doctors. The intervention took the form of an e-learning programme or article and was developed so as to improve general practitioners' (GP) communication and activation skills, especially when working with older adults. The outcome variable was the change between the waves in patients’ scores on the PRACTA Attitude Toward Treatment and Health (ATH) Scale and PRACTA Self-efficacy Scale. Data from patients aged 50 + (n = 2175; 55.6% women; age: M = 69.56, SD = 9.10) appointed at the primary care facilities were analysed.

Results

The analysis revealed the effect of doctors’ e-learning and, to a lesser extent, the effect of article reading on patients’ attitudes toward treatment and health as well as on their self-efficacy. In facilities in which the intervention was implemented, patients’ attitudes were more active on follow-up than at baseline when compared with facilities without the intervention.

Conclusions

Educational intervention among doctors can result in patients’ ATH becoming more active. The form of intervention might diversify the impact.

Introduction

Use of healthcare systems in adulthood increases with age, whereas capacity for self-care often declines.¹ Patients engagement and their active attitude toward treatment and health (ATH) are crucial for the effectiveness of medical care.² A patient’s health-preserving behaviour is seldom discussed at a GP’s office, especially with seniors,^3,4 despite GPs’ being ideally positioned within public health to provide accessible activation.^5,6 An urgent need emerged for public health in Poland to facilitate successful ageing, since it is achieved by only 1.6% of the elderly, placing the country last on the European list.⁷

Activity, engagement and self-management are the core elements of optimal care.^2,8–10 Although supporting patient’s self-management is studied increasingly often,^11,12 the lack of experience, especially in primary care, might disrupt the costs to quality balance.¹³ Doctors’ beliefs about patient activation are reflected in strategies used to support it, and in patients’ behaviours.^9,14–16 Thus, seniors’ activation embedded in a primary health care facility (PHF) might additionally enrich public health with GPs who better understand the activation-health outcomes dependency and, consequently, are more supportive.^15,17,18

Activation in a PHF can be done only when GPs are equipped with appropriate competences. Appreciating WHO’s⁶ opinion that a major barrier to adopting public health for demographic challenges is a shortage of trained healthcare providers, e-learning appears to be a convenient solution. Nowadays, computer literacy should not limit the accessibility of e-learning medical education, whereas tight schedules increasingly do.¹⁹ When compared with traditional courses, e-learning in healthcare has usually at least similar effectiveness when the knowledge is considered.^20,21 Competence gain should produce practical use of the skills or improved performance which, being a more complex issue, have weaker scientific support.^21,22 Although doctors’ performance is often assessed by patients,²³ the efficacy of e-learning interventions to our best knowledge has not yet been evaluated on a larger scale from actual patients’ perspective.²¹

By definition, an attitude is a pattern of cognitive, emotional, and behavioural/motivational responses to a psychological object that is evaluative.²⁴ This multidimensional model is used here in reference to treatment and health. Attitude forms the background of intention, further behaviour and, since prone to intervention, drives? (the attitude drives behavior change).²⁵ In this context, the definition of patient activation is closely related: a process of understanding, raising confidence and an ability to maintain adequate health self-management and lifestyle.^26,27 It can fluctuate; for example, in relation to health status.²⁸ Joining features of the classic attitude’s definition with a more dynamic patients’ activation framework, our understanding of active ATH is rather multidimensional, without distinction between the experiential and instrumental form²⁵ and without being a gradual process, and considers positive and negative emotions separately.¹⁸

In this study, four main areas of interest were reflected in the questions below:

• Can we make patients’ ATH more active by educating GPs?

• Does the intervention method matter?

• Is the intervention age-specific?

• Is the impact of the intervention independent of patients’ health status?

All questions included general and more specific dimensions of ATH and self-efficacy.

Methods

Procedure and participants

Two independent groups of patients were approached consecutively at the PHFs’ waiting rooms: at the baseline (Wave I—WI) and after the PRACTA intervention for their GPs (Wave II—WII, ca. 12 months after WI). We received data from 8,862 patients in total: 4936 at WI and 3926 at WII (response rate 74.61%). The inclusion criteria for patients comprised age (50+), ability to independently complete the questionnaires, being appointed by a GP recruited for the study, and signing the participation agreement. The first part of the questionnaire booklet was completed prior to the appointment; the second part directly after (ca. 20 min. for each part). The number of patients was determined by the number of doctors.

From three provinces in central Poland, we randomly recruited the facilities wherein, after the approval, all employed doctors were invited. At baseline, GPs (n = 503) from 151 PHFs participated in the study (response rate for facilities 20%, for GPs 50%). Next, the facilities were randomly assigned to one of three study conditions: e-learning, article and control; the data from 225 GPs participating in both waves were available.¹⁸ Accordingly, data from their 2237 patients (in 110 PHFs, daily helping 20–400 patients, M = 142, SD = 95) participating in WII were introduced into all analysis, but with regard to the results from their facility at baseline. Since 2.8% of data (n = 62) were excluded due to incomplete ATH scores n = 2175 was left for main comparisons. Subsequently, we had three study groups e-learning (n = 436, GPs n = 42), article (n = 900, GPs n = 89) and control (n = 901, GPs n = 94).

Measurements

Each patient’s ATH was evaluated post-visit with the PRACTA Attitude Towards Treatment and Health Scale (PRACTA-ATH) and PRACTA Self-efficacy Scale (PRACTA-SE). PRACTA-ATH has 4 subscales with a total of 16 items: cognitive, emotional-positive, emotional-negative and motivational. The development of the PRACTA-ATH structure is based on confirmatory and exploratory analyses (article in preparation). The reliability coefficient for the global score is α = 0.88 (for subscales between 0.89 and 0.92, n = 8861). PRACTA-SE has three items and good reliability (α = 0.89). The scales begin with the following common statement: ‘Due to this visit at the doctor…’ and are followed by individual items; e.g. ‘I feel calmer’ or ‘I’m going to participate in the treatment actively’. Patients respond on a 7-point Likert scale (1—‘definitely no’ to 7—‘definitely yes’). A higher score reflects a more active attitude, except for negative emotions, for which the score is reversed when calculating the global ATH score.

Patients evaluated their health compared with their peers (self-rated health) on a 5-point point scale, from ‘very good’ to ‘very poor’. The number of diseases was assessed using the question: ‘How many diseases have you had/are you currently being treated for?’ Patients could choose from one of four options: none, one, two to three, and four or more. A scale assessing the Health Influence on Activities (HIA) was developed. It is short, assures variance and covers the limitations of elderly patients able to visit PHF independently. We picked 10 items derived from classical measurements dealing with elementary and complex activities.^29,30 The scale is reliable (α = 0.95), corresponds to patients’ age (Spearman’s rho = 0.32, P < 0.001), number of diseases (rho = 0.595, P < 0.001), and self-rated health (rho = 0.508, P < 0.001; n = 8862). Patients rate to what extent their health limits their functioning on a 4-point Likert scale (1—‘doesn’t limit at all’ to 4—‘limits very much’).

I addition to PRACTA-ATH and PRACTA-SE, patients in our study completed other questionnaires that were not included in this study. The research was approved by the Bioethics Committee of the Medical University of Warsaw (ref. no KB/10/2014).

PRACTA intervention

It was designed to empower doctors in effective communication, in perceiving the expectations of elderly patients adequately and in capacity to activate patients’ ATH. Both the e-learning programme and article covered five main thematic modules:

• Process of active aging and the importance of an active ATH

• Doctors’ beliefs about seniors’ abilities and expectations

• Importance of doctor–patient rapport for senior patients and health outcomes

• Psychological rules and skills for promoting an active ATH

• Quality of life and providing support for senior patients.¹⁸

The scope, volume and presentation of the information differed between methods of intervention. E-learning was designed as a game that comprised tasks to be completed and, by means of multimedia, to model GPs’ practical skills. Each e-learning module took ∼1 h to complete. The knowledge in the article contained general recommendations and consisted of 15 pages with graphics and tables.

Statistical analysis

The data were analysed using IBM SPSS 24 software. The descriptive statistics of the study groups were compared using chi-square tests. Indices of change were created to detect changes in attitude and self-efficacy pre- and post-intervention. In order to achieve this, the means from the four subscales of PRACTA-ATH, global ATH score and PRACTA-SE score in WI were calculated for each facility; to avoid GPs’ fear of being evaluated, the comparisons were nested in facilities not in GPs. These values were subtracted from the individual score of each patient in WII, matching the scales. Thus, positive results indicated an increase in score between waves. These indices were employed in the analysis as outcome variables.

A Generalized Linear Model (GENLIN, distribution–normal, link–identity) was applied to test differences between groups³¹ and answer the first and second research questions.

Three age categories (50–64, 65–74 and 75+ years old), representing 29% (n = 657), 41% (n = 910) and 30% (n = 670) of the sample, and study group by age category interaction were introduced to address the third study question. To control for the confounding influence of health-related variables, number of diseases and a self-rated health and HIA scale scores were included. Ultimately, two control demographic variables were added; namely, education level and marital status. A pairwise comparison with Bonferroni correction was performed to verify differences between particular levels of factors. All CIs refer to Wald’s 95% evaluation.

Results

Basic demographic characteristics of the study groups are displayed in table 1 and indicate an equal proportion of gender and marital status. The control group had a lower rate of the youngest patients (<65) and a higher educational level.

Table 1

Characteristics of the study groups according to demographic and health variables

Feature	Value	E-learning (n = 436)	Article (n = 900)	Control (n = 901)	Test of difference
Female	n (%)	248 (56.9)	504 (56.0)	492 (54.6)	χ2 = 0.71; P = 0.70a
Married	n (%)	268 (61.5)	601 (66.8)	581 (64.5)	χ2 = 7.23; P = 0.30a
Age	M (SD)	69.4 (9.2)	68.9 (9.11)	70.3 (8.97)	χ2 = 10.72; P = 0.005b
Education–vocational or less	n (%)	205 (46.1)	301 (33.5)	266 (29.5)	χ2 = 93.42; P < 0.001a
Living alone	n (%)	90 (20.6)	148 (16.4)	161 (17.9)	χ2 = 8.96; P = 0.176a
Financial situation	M (SD)	3.12 (0.85)	3.18 (0.72)	3.29 (0.80)	χ2 = 16.24; P < 0.001b
Number of diseases	Median (M; SD)	1 (1.56; 0.8)	1 (1.40; 0.75)	1 (1.45; 0.79)	χ2 = 10.57; P = 0.005b
Self-rated health	M (SD)	2.92 (0.8)	2.89 (0.66)	2.88 (0.73)	χ2 = 0.411; P = 0.814b
HIA score	M (SD)	1.39 (0.53)	1.43 (0.57)	1.50 (0.61)	χ2 = 11.76; P = 0.003b

a

Pearson’s Test.

b

Kruskal-Wallis’ test. HIA–Health Impact on Activities scale.

The means of the study groups, main effects of the study group factor, and strength of the entire model calculated for indices of change of all aspects of ATH, global ATH, and self-efficacy are presented in table 2. The results of the omnibus tests for all investigated outcome variables are presented in table 3.

Table 2

Indices of change in ATH and self-efficacy between the three study groups

	E-learning n = 397	Article n = 877	Control n = 901	Wald χ2 (p) between groupsa	Significant pairwise comparisons	Wald χ2 (p) for the modelb
	M (SD)	M (SD)	M (SD)
ATH cognition	0.202 (1.00)	0.149 (0.83)	−0.193 (1.11)	63.82 (≤0.001)	E, A > Cc	121.44 (≤0.001)
ATH positive emotions	0.253 (0.87)	−0.014 (0.91)	0.093 (0.99)	27.85 (≤0.001)	E > C > A	109.20 (≤0.001)
ATH negative emotions	−0.127 (1.80)	−0.557 (1.65)	−0.328 (1.87)	17.79 (≤0.001)	E < A, C	132.75 (≤0.001)
ATH motivation	0.223 (0.85)	0.130 (0.97)	−0.008 (1.11)	17.89 (≤0.001)	E, A > C	52.16 (≤0.001)
ATH global	0.281 (0.81)	0.122 (0.72)	0.056 (0.82)	21.35 (≤0.001)	E, A > C	98.89 (≤0.001)
Self-efficacy	0.279 (0.84)	0.135 (0.86)	−0.117 (1.05)	53.85 (≤0.001)	E > A > C	123.59 (≤0.001)

a

Wald χ² between groups—the value of main effect of study groups within the model.

b

Wald χ² for the model after entering: age categories, study groups, marital status, educational level, number of diseases, self-rated health, HIA and age category by study groups interaction.

c

E, e-learning group; A, article group; C, control group.

Table 3

Omnibus test for the indexes of change in ATH and efficacy

ATH	Cognitive		Positive emotions		Negative emotions		Motivation		ATH Global		Self-efficacy
	Wald’s χ2	p/difference*	Wald’s χ2	p/difference*	Wald’s χ2	p/difference*	Wald’s χ2	p/difference*	Wald’s χ2	p/difference*	Wald’s χ2	p/difference*
Intercept	1.39	0.24	3.65	0.056	25.34	0.000	0.000	0.997	3.64	0.057	7.24	0.007
Study group	63.82	<0.001/C < A,Ea	27.85	<0.001/A<C<E	17.79	<0.001/A,C>E	17.89	<0.001/C<A,E	21.35	<0.001/C<A,E	53.85	<0.001/C<A<E
Age	1.99	0.369	3.09	0.213	1.64	0.440	12.54	0.002/75 + < younger	0.61	0.735	6.66	0.036 (75+)< younger
Marital status	9.71	0.021/W <Da	15.78	0.001/S,W< D, W<M	16.77	0.001/D>M,W,S	6.34	0.096	16.04	0.001/M,W,S<D M>W	18.00	<0.001/ W< S,M,D
Education	7.19	0.126	3.67	0.452	1.88	0.758	3.99	0.408	1.91	0.752	8.96	0.062
Self-rated health	0.69	0.407	0.60	0.437	0.36	0.546	0.22	0.640	0.03	0.868	3.18	0.074
HIA	0.89	0.346	0.37	0.541	28.88	0.000 = 0.44	5.37	0.020/β = 0.114	1.27	0.259	0.03	0.865
Number of diseases	25.85	<0.001/1>2/3,4+	20.59	<0.001/1,0>2/3,4+	10.29	0.016/1<4+	1.36	0.715	18.82	<0.001/1>2/3,4+	3.38	0.337
Study group by age	0.78	0.941	2.53	0.640	13.41	0.009 <65 group: E<A,C 65-74 group: E,A<C 75+ group: E<A<C	6.29	0.178	3.78	0.436	4.20	0.379

*

All indicated differences based on pairwise comparisons and are significant at least at the 0.05 level; β is indicated for continuous variables.

a

E, e-learning group; A, article group; C, control group; S, single; M, married; W, widowed; D, divorced.

When compared with the control group that demonstrated a negative change, the index of change for ATH cognitive in the e-learning group and article group was larger (CI 0.27–0.57 and 0.22–0.45 respectively, both P < 0.001). The two intervention groups did not differ significantly (CI −0.06–0.22, P = 0.48). The number of diseases was related to the greater change when one rather than when two to three, or 4+ illnesses were present (CI 0.10–0.36; P < 0.001 and 0.06–0.58; P < 0.008, respectively). A small effect of marital status was observed; thus, compared with widowed patients, divorced/separated patients demonstrated increased scores (CI −0.47 to −0.02; P = 0.022).

A significant increase in ATH positive emotions was noted in the e-learning group compared with the control group (CI 0.03–0.30; P = 0.009) and article groups (CI 0.15–0.41; P < 0.001). The latter showed a minimal negative index of change, lower than the control group (CI −0.23 to −0.01; P = 0.029). Patients with two–three diseases changed less than those with one or no illnesses (CI −0.49 to −0.04; P = 0.012 and −0.31 to −0.07; P < 0.001, respectively) but similar to those with more than four illnesses (CI −0.23–0.24; P = 0.24). The effect of marital status was related to differences between divorced/separated and single or widowed patients, whereas the latter reported smaller changes (CI 0.35–0.60; P = 0.018 and 0.29–0.44; P = 0.016, respectively) and married/partnered patients greater changes than those who were widowed (CI −0.27 to −0.04. P = 0.039).

A smaller change was noted for ATH negative emotions in the e-learning group compared with the article and control groups (CI 0.19–0.70; P < 0.001 and 0.5–0.57; P = 0.013, respectively), with no difference between the last two (CI −0.34 to 0.06). The age category interacted with the study group with a complex and detailed pattern that was difficult to interpret. Patients who had one disease showed a greater reduction in negative emotions than those with four or more diseases (CI −0.87 to −0.07; P = 0.013). The score on HIA (Beta = 0.44; P < 0.001) showed that the more activities were influenced, the stronger the increase in ATH negative emotions. A greater change in divorced/separated patients compared with other groups was observed (CI −1.34 to −0.25; P = 0.001 for single; −0.73 to −0.06; P = 0.010 for married/partnered; and −0.82 to −0.05; P = 0.021 for widowed patients).

The e-learning and article groups revealed an increase in ATH motivation compared with the control group (CI 0.10–0.37; P < 0.001, and CI 0.01–0.25; P = 0.034, respectively) but their indices remained similar (CI −0.02 to 0.23; P = 0.12). The age groups’ effect indicated that the eldest patients’ change index increased less than the youngest and slightly less than in patients 65- to 74-years old (CI −0.35 to –0.07; P = 0.001, and −0.24 to 0.01; P = 0.071, respectively). Furthermore, HIA exerted a minor effect (Beta = 0.114; P = 0.020).

The ATH global index of change was positive and stronger in the e-learning and article group than in the control group (CI 0.08–0.30; P < 0.001 and CI 0.06–0.26; P = 0.001). Methods of intervention for GPs did not differ (difference of means −0.027, CI −0.13 to –0.08; P = 1.0). Independently from study groups, PRACTA-ATH scores (global and some subscales) correlated with age; however, in terms of strength they were all meaningless (all Spearman’s rho < 0.09) in both waves. The number of diseases had a main effect; patients who indicated one had a greater change than those with two–three and more than four illnesses (CI 0.04–0.26; P = 0.001 and 0.06–0.45; P = 0.004, respectively). The effect of marital status was noted with divorced/separated participants changing more than single, married and widowed, whereas married patients changed more than those widowed (all P < 0.02).

For self-efficacy, the index of change in both intervention groups was positive and differed from the control group (CI 0.26–0.52; P < 0.001 in e-learning and 0.12–0.34; p < 0.001 in article condition), where a decrease was noticed. In addition, e-learning appeared more effective than the article (CI −0.29 to −0.04; P < 0.001). The age category had a minor effect, indicating that the oldest patients had an increase in self-efficacy that was smaller than among patients from the two younger groups (CI −0.22 to 0.01; P = 0.088 and CI −0.25 to 0.01; P = 0.067 for the ages of 65–74 and <65, respectively). An effect of marital status was observed with widowers as the only group who showed a decrease and varied from married/partnered and divorced/separated patients (CI −0.33 to −0.06; P = 0.001 and −0.50 to −0.07; P = 0.002, respectively).

Across the main factors and interactions investigated in models for ATH and self-efficacy, the study group had the most systematic and significant impact. Furthermore, the number of diseases was related to all aspects of ATH, except motivation and self-efficacy, while age significantly affected these dimensions. HIA was an important factor for negative emotions and motivation. Finally, marital status proved important for all investigated variables except motivation. The interaction of the study group and age was detected in negative emotions. Self-rated health and education were not significant for ATH dimensions or efficacy.

Discussion

The main objective of this study was to assess the efficacy of an educational intervention for primary care physicians, delivered by means of e-learning or in article form, from the perspective of a large sample of PHF elderly patients. To our knowledge, this is the first report of this kind.

When considering indices of change for global ATH and self-efficacy, the PRACTA e-learning programme and the article appeared more beneficial for patients compared with those in the control condition. A consideration of our first research question reveals that patients can be activated through the education of GPs in both proposed forms. Further, both emotional dimensions of ATH and self-efficacy benefited more from the PRACTA e-learning than from the article, which failed to improve emotional indices. Patients attending facilities where e-learning was applied had the biggest change in all ATH dimensions and self-efficacy, with a coherent direction toward greater activity. The second research question, concerned with the method, shows that e-learning is at least as effective as reading a pdf article, with comparatively better results on some dimensions. The difference might stem from the variety of methods used in e-learning (quizzes, films etc.), the disproportionate amount of time required or the content and volume of the information. Nevertheless, the educational effect was observed ∼1-month post-intervention, with the indirect measurement of patient perceptions, and indicated more than only a short-term change in GPs’ approach to their patients. Since the activation level is associated with health habits and self-management³² along with the long-term effects of patient activation,³³ future research should focus on the prospective, clinical, and economic characteristics of the observed effects. Wide applicability of this activating intervention is thus worth considering.

Although our intervention focused on working with elderly patients, the effect of patients’ age categories was noted selectively in indices for motivation and self-efficacy: crucial areas in perceiving capacities for behaviour change.³⁴ These results show that improvement in these areas might be a particular challenge among patients aged 75+. Arguably, the lack of age-related effects in other ATH aspects might result from attitudes being weakly related to age and/or a generalization effect. The answer to the question of whether our intervention is age-specific is thus negative.

The three health-related variables have shown a divergent relationship with ATH change with the number of diseases being responsible for the majority of main effects. Patients with high multimorbidity were less likely to change their ATH, but not their self-efficacy, in the desired direction. The number of diseases might explain most of the indexes of change because it used the model first and because of intercorrelations between health factors. The answer to the fourth question in our research shows that, although indices of change depended upon health-related variables, the effects of the PRACTA intervention remained significant.

Limitations

As a result of the study design—parallel Separate Sample Pre- to Post-test design—it was impossible to match patient’s scores from WI and WII. Thus, patients’ results from WII were referred to the results of WI, but calculated at the level of facility. However, with the present results in mind, we can expect that the effect of the intervention might be clarified if future research addresses this limitation.

The low response rate on facilities and, to a lesser extent, on doctors, may lead to speculation about the specific profile of the group of participating GPs; e.g. maintaining higher standards or motivation than those who refused, and therefore being more prone to benefit from the intervention.¹⁴ Consequently, nonparticipants might be perceived as those who have a higher need for an intervention; therefore, an effective approach for less skilled groups or those with weaker educational motivation should be considered in future projects.

Owing to the previously mentioned recruitment issue, the group of our patients cannot be considered representative; however, in terms of age stratification within PHFs’ use or health rating, to some extent it reflects other national samples.^35,36 However, a comprehensive interpretation of the presented data is limited and further investigation of a representative group of senior PHF’s patients would be a definite advantage.

To conclude, the facilitation of more active ATH among older patients in PHF via PRACTA’s educational intervention for doctors was confirmed. Patients visiting PHFs where the e-learning programme was introduced had a greater change in ATH than did those in control facilities and, on selected dimensions, than in those involved in article distribution. Changes in motivation and self-efficacy appeared more age-related than did other dimensions of attitude. Although patients’ health status also contributed, the PRACTA intervention group was the most important factor explaining ATH change, which suggests a complex relationship between a patient’s health, the doctor’s participation in intervention, and the patient’s ATH.

Key points

• The PRACTA project focused on the activation of elderly patients in primary care through intervention among GPs.

• Patients’ attitude toward treatment and health became more active in the study groups in which e-learning and a pdf article were made available for GPs, with stronger effects observed for computer-based intervention.

• Public health can benefit from facilitating activation of older patients’ attitudes toward health in primary care at the level of GPs’ competences and patient outcomes.

• Wider use of e-learning in public health for practicing professionals should be considered an effective, easy, and manageable tool.

• Future research in public health should consider facing less probable changes in motivation and self-efficacy in patients 75+, and a risk of low response rates among facilities and doctors.

Funding

Funding for this study was provided by Norway Grants within the Polish–Norwegian Research Programme (Pol-Nor/200856/34/2013).

Conflicts of interest: None declared.