Relevant patient characteristics for estimating healthcare needs according to healthcare providers and people with type 2 diabetes: a Delphi survey

Two separate Delphi studies were conducted: one with HCPs and one with people with T2DM.

The first Delphi study was conducted from September through October 2016 and included a purposive and representative sample of HCPs (general practitioners, practice nurses [who support the general practitioner in primary care], specially trained diabetes nurses, dieticians, internists, psychologists and pharmacists) recommended in the care protocols to be part of the multidisciplinary care team for the treatment of T2DM. The authors composed a list of HCPs (n = 20) from their own network who treat or used to treat people with T2DM in the Dutch healthcare system and/or have extensive knowledge on the organization of T2DM care in the Netherlands. These HCPs were asked to participate and to recommend colleagues (n = 6) who might be willing to participate as well. Furthermore, the Dutch Professional Association of T2DM Care Providers (EADV) and the Dutch Dietician Nutrition Organization (DNO) were contacted for recommendations on HCPs interested in participation (n = 8). In total, 34 HCPs were invited to participate and give written informed consent.

The second Delphi study focused on people with T2DM with a diagnosis of T2DM and took place between June and August 2017. For the recruitment of people with T2DM, we contacted one general practitioner with a practice in the north of the Netherlands in which 109 people with T2DM were treated. People with T2DM who also had a diagnosis of dementia were excluded from participation, all other people with T2DM were asked if they were willing to participate and give written informed consent.

Both Delphi studies were conducted according to the RAND/UCLA appropriateness method and consisted of two rounds [24].

Descriptive analyses were conducted to assess the demographic characteristics of the participants. The relevance of the person-, context-, and health-related characteristics for the questions with a 5-point Likert scale was classified into three categories based on median scores: not relevant (median 1–2), uncertain (median 3) and relevant (median 4–5). To determine the level of consensus between participants, the interquartile range (IQR) was calculated for each characteristic. An IQR ≤ 1.5 was considered as consensus, meaning that at least 50% of all ratings are situated within 1.5 points around the median rating of the participants [26]. Characteristics with a median of 3 and/or an IQR > 1.5 in the first round were considered not relevant and presented again in the second round.

All analyses were performed using R Studio version 1.0.153.

Figure 1 shows the mean relative importance scores of the five most relevant characteristics for estimating healthcare needs according to HCPs (A) and people with T2DM (B). The top 5 according to HCPs mainly consisted of person- and context-related characteristics, whereas for people with T2DM the top 5 mainly consisted of health-related characteristics. HCPs rated self-efficacy as the most relevant patient characteristic with a mean relative importance score of 3.09, whereas people with T2DM rated HbA1c as the most relevant characteristic with a relative mean importance score of 2.11. The two characteristics that were present in both the HCP and patients top 5 s, were self-efficacy and quality of life. Significant differences between the mean relative important scores were found for all characteristics in both top 5 s that were rated by both HCPs and patients, except for quality of life (p = 0.573) and comorbidity (p = 0.344).

In the present study, HCPs and people with T2DM were asked to give their opinion about the relevance of patient characteristics for estimating healthcare needs of people with T2DM. In two Delphi rounds, 28 and 15 patient characteristics were rated by HCPs and people with T2DM, respectively. Except for triglycerides, genetics and insecurity/fear, all health-related characteristics were found to be relevant with consensus for estimating healthcare needs by both HCPs and people with T2DM.

Discrepancies in opinions between HCPs and people with T2DM were observed for person- and context-related characteristics. HCPs found 75% of these characteristics relevant for estimating healthcare needs, whereas people with T2DM only found 17% relevant. A striking discrepancy was also seen in the top 5 of most relevant patient characteristics for estimating healthcare needs between HCPs and people with T2DM. The top 5 of HCPs mostly consisted of person- and context-related characteristics and they thought that self-efficacy was most relevant for estimating healthcare needs. In contrast, the top 5 of people with T2DM mostly consisted of health-related characteristics and they ranked HbA1c as the most relevant estimator.

Previous research has suggested that more emphasis should be placed on person- and context-related characteristics in the treatment of T2DM [27, 28]. In the current study, self-efficacy for example, defined as an individual’s confidence in being able to carry out a behavior, was rated as the most and fourth most relevant characteristic for estimating healthcare needs by HCPs and patients, respectively, and has been associated with lower HbA1c levels and T2DM management- and problem solving behavior [29, 30]. A healthcare provider’s knowledge on the self-efficacy of people with T2DM and other person- and context-related characteristics could enhance self-management education and support and the development of mutually accepted treatment goals, referred to as shared decision making (SDM) [28, 31‐34]. Similar to the outcome of the current study, most HCPs agree on such a broad, whole person approach to the treatment of T2DM [35, 36]. However, from the current study it remains unclear whether HCPs practice such an approach. Previous research has shown that HCPs often lack the time, skills and resources to provide self-management education and support and SDM is not yet embedded in clinical practice [37]. Instead, patient consultations with a healthcare provider seem to focus on clinically orientated issues, such as optimal blood glucose levels [31]. Past qualitative research has shown that people with T2DM were unable to describe the role of the practice nurse beyond clinical checks [38]. They also were not sure what else they could expect from their practice nurse. On the other hand, it could be that HCPs do discuss person- and context-related characteristics during patient consultations, but people with T2DM might be unaware of this or not open to it. These arguments might explain why people with T2DM considered HbA1c as the most relevant characteristic for estimating the healthcare needs of people with T2DM in the current study.

A strength of this study was the unique inclusion of HCPs as well as people with T2DM. A Delphi panel is often referred to as an ‘expert panel’ and assumed to include professionally and scientifically qualified participants [39]. People with T2DM do not fall under this category. We did, however, decide to include them, because of their relevant knowledge and experience on the topic and because knowing the opinions of both groups can improve the development of patient-centered care. On the other hand, the Delphi method is context free, which could explain the differences in opinion between HCPs and people with T2DM. We do not know for example, if people with T2DM were trying to see things from the HCPs’ perspective, despite the provided instructions that asked them to make their own judgement. In that case, a true qualitative method would have been better to elicit people with T2DM’ views.

Given the scale of this study, we decided to only select patients from one primary care practice. This does mean that the included patients all live in the same region and are treated by the same HCPs, which makes it difficult to generalize the results to people with T2DM in other countries and other regions Dutch regions. However, since Dutch general practitioners and especially practice nurses (who treat more patients with diabetes than the GP) strictly adhere to the guidelines for T2DM treatment [40], it is likely that the included patients received T2DM care similar to the care of patients from other primary care practices. Moreover, as patients within the practice differ in terms of which HCP they most frequently see for their T2DM – there is one GP and three practice nurses providing T2DM care – we expect the influence of provider attitude and interpersonal style on patients’ opinions to be limited. The included HCPs formed a multidisciplinary Delphi panel. In the Netherlands multidisciplinary cooperation within T2DM teams – comprising not only general practitioners and practice nurses, but also T2DM nurses, dieticians, psychologists and, to a limited extent, internists – forms an important part of T2DM care [41]. They refer people with T2DM to each other and mutually discuss treatment plans. The diversity of our panelists represented the range of HCPs that are involved in the treatment of people with T2DM and their opinions. Only Dutch HCPs were included. We tried to arrange face-to-face meetings with the participants, to allow for more in-depth discussion about the ratings and investigate areas of disagreement. Due to time-constraints of the participants, the decision was made to conduct an online Delphi survey instead. To gain more understanding of the ratings, we did include open questions. Finally, the patient characteristics that were included in the Delphi surveys were derived from studies that were previously conducted as part of the PROFILe project, which included an in-depth systematic literature search [21]. It is, however, possible that we missed relevant patient characteristics for estimating healthcare needs of people with T2DM. Research has, for example, suggested that environmental factors, such as social stratification and political context, have an impact on people’s health [28, 42]. These factors are, however, difficult for HCPs and people with T2DM to influence, and were therefore not included in the surveys. Furthermore, participants were given the chance to provide a list of patient characteristics that they found relevant for estimating healthcare needs of people with T2DM and were not included in the survey of the first round.

The findings of this study complement the results derived from previous empirical research on relevant patient characteristics for estimating healthcare needs. They are important for both HCPs involved in the treatment of people with T2DM and researchers focusing on the development of patient-centered care. The findings suggest that there is discrepancy in opinions on relevant patient characteristics for estimating healthcare needs between HCPs and people with T2DM. To improve SDM and encourage patient-centered care, it is important that both groups agree on what topics should be discussed during patient consultations. People with T2DM have previously reported that they would like their healthcare provider to show more interest in their life and provide more explanation and involvement in T2DM management, such as providing lifestyle advice and discussing treatment options [31, 38]. Indeed, a recent study on the implementation of a structured T2DM consultation model with a focus on person- and context-related patient characteristics, led to an increase in patient involvement and a substantial number of satisfied people with T2DM [43]. In the current study, HCPs and people with T2DM agreed that self-efficacy and quality of life are relevant patient characteristics for estimating healthcare needs. The measurement of these characteristics should therefore to be included in routine care, for example as part of the intake of people with newly diagnosed T2DM. To save time, people with T2DM could fill in questionnaires that measure these characteristics before their visit with a HCP. Identifying self-efficacy and quality of life in diabetes management allows HCPs to know which aspects of the lives of people with T2DM are most important and which activities they are facing most difficulties with [44]. This has important implications on targeting person-centered education interventions.