Acceptability and Satisfaction With a Telecarer Approach to the Management of Type 2 Diabetes

RESEARCH DESIGN AND METHODS

The study formed part of a 1-year randomized controlled trial to assess the effectiveness of the PACCTS intervention (20). A total of 591 patients were recruited and randomly assigned in a 2:1 ratio to the intervention and control group. All patients gave their informed consent, and ethical approval for the study was obtained from the local research ethics committee. The primary outcome variable was the level of glycemic control (HbA_1c).

Satisfaction with treatment was measured in both the intervention and control groups at baseline and the end of the study using the validated Diabetes Satisfaction and Treatment Questionnaire (DTSQ) (21). The DTSQ, a measure of satisfaction with treatment, is a self-report measure and has eight items, each scored on a seven-point scale (0–6). The first six items form a satisfaction with treatment score, ranging from 0 to 36, which is used here. Acceptability was measured by a purposely designed, 24-item postal questionnaire and administered to the intervention group after the patient had received at least three proactive calls from the call center and at the end of the trial. The instrument comprised 20 statements exploring different aspects of the call center (each scored on a five-point scale, with an extra “don’t know” box) and four open-ended questions asking about advice and support provided, perceptions of control, additional follow-up advice required, and any further comments.

To look in greater depth at any behavior-changing effects of the intervention, 25 patients took part in an in-depth semistructured interview. In total, 30 patients were approached, with 5 refusing due to holiday or other commitments during the 2-week interview period. The patients were theoretically sampled from four groups according to their baseline and end-of-study HbA_1c results: those whose control remained either “good” or “poor” and those whose control improved or deteriorated. The interviews explored the content of the intervention, changes in attitudes and behavior, and the mechanisms and context for changes. All the interviews were tape recorded and transcribed verbatim.

Quantitative data were analyzed using SPPS for Windows, with significance levels set at 5%. Given the ordinal nature of the data, the Mann-Whitney U test was used to assess differences in satisfaction levels between groups, Wilcoxon’s signed rank test for differences within groups, the χ² test to explore age, sex, or other structural differences, and Spearman’s ρ for correlations. The open-ended acceptability questions were read through and grouped into three thematic areas: service-related factors supporting or leading to change; coping and other personal factors that enhanced coping; and self-knowledge, attitudes, and behavior (22).

RESULTS

Loss to follow-up

There were similar numbers of withdrawals in the intervention and the control groups due to death, moving from the area, further serious illness, relative’s illness, or loss of contact (10.1 vs. 10.7%, respectively). In the intervention group, an additional 22 (5.6%) dropped out for call center-related reasons, such as not being able to cope with the calls (n = 8), being unhappy with the advice given (n = 4), traveling (n = 3), changing their mind (n = 3), or being too busy (n = 2).

Satisfaction

A total of 468 persons responded to the pre- and post-DTSQ for a response rate of 79%. Respondents were a median age of 67–68 years, had had diabetes for a median of 7 years with just over half being male (57% in the intervention group vs. 58% in the control group), and the majority came from localities with a Carstairs deprivation score of 4 or 5 (85% in the intervention group vs. 83% in the control group). There were no statistically significant differences between the control and intervention groups for age, sex, time since diagnosis, or Carstairs score. Both groups had high and nonstatistically significantly different levels of satisfaction before the intervention (95% CI 28.8–30.9 in the control group vs. 29.6–30.9 in the intervention group) (Fig. 2). By the end of the trial satisfaction, levels had increased in both groups (30.6–32.3 vs. 32.3–33.2 in the control and intervention groups, respectively), and there was statistically significant difference between the intervention and the control group (z = −2.266, P < 0.023).

Acceptability

Two hundred users responded to the initial and end-of-study acceptability questionnaire (a response rate of 65%). They had a median age of 69 years, had had diabetes for a median of 7 years, and 42% were female. The responder characteristics mirrored those of the intervention group as a whole.

General acceptability and feelings of control

At follow-up, >90% of respondents strongly agreed or agreed that the PACCTS approach was acceptable. Specific responses related to friendliness, helpfulness, convenient call scheduling and duration, knowledgeable staff, personally relevant call content, and useful, personally tailored advice (Table 1). In addition, >90% strongly agreed or agreed that PACCTS improved their knowledge of diabetes, control of diabetes, and general well-being. There was an age variation, with younger persons feeling more knowledgeable (χ² = 39.1, df = 25, P = 0.035) and more in control (χ² = 41.0, df = 20, P = 0.004) after PACCTS.

Different comments predominated in the initial (after three proactive calls) and the end-of-study questionnaire. In particular, there was a noticeable shift from comments about attitudes to (13 vs. 4%, respectively) and knowledge of (19 vs. 11%) behavior (16 vs. 21%). At the end of the trial, patients talked in terms of “everything has been explained to me,” “changing my attitude,” “giving the opportunity to ask questions,” or “seek advice” and it being “the little things that they have told me that have made the big difference.”

When asked whether they would prefer to see a health professional rather than receive telephone calls about diabetes, 21% either agreed or strongly agreed and 5% did not know. Exactly half of the respondents disagreed or strongly disagreed, suggesting substantial support for the telephone approach. Four-fifths indicated that their expectations of PACCTS had been met totally or to a great extent, with younger persons being significantly more likely to respond positively (χ² = 34.0, df = 20, P = 0.025), and 90% were highly or largely satisfied with the recommendations that the telecarers had given them.

Moderate to strong correlations (Spearman’s ρ = 0.42–0.59, P < 0.01) were evident between perceptions of the acceptability of receiving advice by phone, feeling able to act on the advice given, feelings of control, expectations of the intervention, and satisfaction with the intervention. These data support a behavioral model in which the provision of an acceptable service leads to improvements in knowledge and feelings of control.

Relationship with the telecarer

Approximately 25% of the questionnaire respondents highlighted aspects of the service that were important to them. These included staff being knowledgeable, polite, efficient, helpful, and friendly; the value of regular contact; and satisfaction with the service and advice given. The patients highlighted that a relationship and bond had been formed with the telecarers. Just under half (45%) of the respondents’ comments concerned coping, caring, and comforting. For instance, they mentioned someone providing care and comfort, “stopping you feeling isolated,” a personalized service including incentives and reminders, and feelings of mental well-being—embracing confidence, seeking advice, self control, being happier and more positive and not panicking.

In the interviews, patients, particularly those with poor baseline control and, thus, those who received the most frequent calls, spoke of their close rapport with the telecarers (Fig. 3). They commented in terms of “speaking to the same one,” using first names, and sharing jokes. The telecarers provided personalized and expert advice and helped the patients problem solve. Some patients pointed to the help given when other illnesses or health crises had affected their blood glucose control. More generally, the telecarers were described as “more personal” and “a bit more caring.” The fact that there was someone at the end of the phone enhanced confidence. The regular calls, made at convenient prearranged times (“no restrictions on your life with them ringing; they ring to suit you”) also encouraged direct calling by the patient for advice. As one patient put it, “you don’t feel that you are calling a call center because you identify the person at the end of the phone.”

Consciousness raising

A second thematic area, mentioned by ∼33% of all respondents to the acceptability questionnaire, related to enhanced self-knowledge and changes in attitudes and behavior. The in-depth interviews explored this further. Here interviewees commented that the advice given was personalized and tailored to their own lifestyle needs to, for example, accommodate shift working or someone who wished to incorporate exercise into their daily routine. The telecarers asked pertinent questions and helped the patients establish realistic goals.

Initially there was a process of consciousness raising. With the patient, telecarers addressed specific areas of diabetes management (Fig. 4). Diet and exercise were the most common areas mentioned (10 of 25 patients). Respondents spoke in terms of “becoming more aware of what I eat,” “they go on at me about diet,” “being more careful what you eat,” “keeping your weight down,” and to eat “more carbohydrates.” A second main area related to comments about when to measure their blood glucose levels and the interpretation of the readings (mentioned by eight persons).

CONCLUSIONS

If a patient-centered care service is to be achieved, then user perceptions of the acceptability and satisfaction are essential. A combination of methodological approaches needs to be used and tailored to the intervention, ideally incorporating validated measures used in other studies (to enable cross-study comparison), using purposely designed instruments, and including in-depth interviews with participants (23).

This study has demonstrated that the PACCTS intervention provides a continuing satisfactory service to patients, demonstrated by the small loss to follow-up and increased levels of satisfaction, significantly more so in the intervention than the control group. This difference could be explained by the increased contact of patients with care providers via PACCTS. While it is well known that the measurement of satisfaction is prone to bias (24), with results varying in relation to what the patient is asked and how, the DTSQ is widely used in evaluations of diabetes interventions. Interest here lies in eliminating the possibility that PACCTS led to reduced satisfaction. Assuming comparable biases in the intervention and control groups in the measurement of satisfaction, this is amply evident. The qualitative component of the study helps to clarify the complex causal chain between changes in knowledge, attitudes and behavior, and health status. Our data suggest that PACCTS supported improved knowledge through a process of consciousness raising and supported problem solving as recommended by the National Institute for Clinical Excellence (6) and the National Service Framework (7). It is possible that some of the reported behavioral changes and observed clinical changes (20) could be due to an “expectation” effect; that is, a change in behavior arising from the expectation of a call and the monitoring of glucose control, as opposed to adoption of the changed behavior into everyday life. Other parts of the interview data confirm that the expectation of a call did encourage glucose testing and behavior change. For those who changed their HbA_1c group status, for example, from “poor” to “good,” there was however evidence of an effect beyond expectation—adoption of the message of the need for better blood glucose control. Indeed, other studies have also demonstrated beneficial effects of empowerment-based interventions, illustrating early improvements in HbA_1c, a decline at 6 months, and subsequent sustained improvements at 1-year follow-up (25). Teaching self-management skills is recognized as insufficient to bring about changes in behavior; patients need to learn these skills in everyday life (26). Exploration of a longer-term effect and closer examination of causal patterns remains an important next step in the research.

The PACCTS system operated at least at two levels: 1) proactively with the telecarer initiating calls, helping the patient to review glucose control and management, and providing the education and motivation aimed at effective management; and 2) as a decision support tool for the patient, who was able to call the telecarer when needed for specific advice or help in problem solving. The telecare workers used a structured questioning approach, working with patients to develop solutions and providing feedback and encouragement, that prompted patients to identify their own problems. PACCTS support made the patients more knowledgeable and aware of their diabetes and more confident in their ability to manage and control their diabetes. For many this occurred only at a cognitive level during the study, while for some it translated into behavioral change that resulted in improved glucose control.

These results suggest that several processes can contribute to a successful patient-centered telephone-supported disease management. These processes include listening to and focusing on the concerns of patients, individualized problem solving, and continuity of care over time. It may not be the information provision per se that is important (raising awareness and knowledge), but its provision in a context of enablement and support aimed at self-efficacy. The development of a strong bond between the patient and telecarer may play a major role in both the acceptability of the intervention and behavioral change. The interpersonal dimension also needs to be a major focus of further research.

Whether a system would be as acceptable as PACCTS or engender a similar level of bonding between telecarer and patient depends on faithful adoption of the model. PACCTS differs from other telecare systems such as National Health Service (NHS) Direct (27) in that contact is with a particular, not any, telecarer. PACCTS placed significance on the interpersonal dimension as a key contributory factor to enabling behavioral change (28). This is reinforced by the fact that, although one of the original telecarers left during the 1-year trial, her replacement rapidly bonded with her patients. This was facilitated by a planned and progressive handover with the original telecarer telling the patients she was leaving and who would be replacing her. A weakness of the current study is its limited 1-year time frame; the sustainability of any change over time remains to be examined and research is ongoing. This is particularly important for an intervention aimed at enhancing confidence and skills and thus empowering patients to develop long-term self-efficacy. It would also be instructive to undertake in-depth interviews within each HbA_1c stratum early on in the intervention and toward its end with the same persons to gain further insight into the workings of the intervention.