Cost-effectiveness and cost-utility analyses of hospital-based home care compared to hospital-based care for children diagnosed with type 1 diabetes; a randomised controlled trial; results after two years’ follow-up

During the past decade, a rapidly increasing incidence of type 1 diabetes has been reported from many parts of the world with a shift towards a younger age of onset [1, 2]. In Sweden, the incidence rate has risen from 21.6 per 100.000 in 1978–1980 to 43.9 in 2005–2007; recent data suggests a break in the increasing trend, but these findings need to be confirmed over a longer period of time [3]. It is a serious disease for the child, and it has significant implications also for the family as regards daily activities, work and social life [4‐6].

The disease also has substantial economic consequences over time, both for individuals and society [7‐11]. People, diagnosed with diabetes during childhood, have been seen to be disadvantaged in adult employment, with lower earnings and lesser probability of attaining the highest level of education, although late complications appear to be the most important determinant of social consequences in later life [8‐11]. The findings emphasize the importance of choosing an appropriate strategy for handling diabetes already at the time of diagnosis.

The diagnosis of childhood diabetes represents a major stressor event for parents [12‐14] and parents approach a challenging process of changes in the patterns of daily activities [15]. Conventionally, a child newly diagnosed with type 1 diabetes has been admitted to hospital as part of his or her initial management, partly for child-safety reasons, partly based on the belief that the family needs time to adjust to the requirements of the disease [16, 17]. Internationally, there is a considerable variation in length of hospitalization at diagnosis but with a trend towards shorter lengths of stay or exclusively outpatient management [18‐22]. In the United Kingdom, for instance, children with newly diagnosed type 1 diabetes are increasingly treated exclusively at home from diagnosis and onwards [5]. To obtain safe care in a home-management environment requires adequate expert support for parents, not least during the first days after diagnosis, when parents still have limited knowledge and understanding of the situation and its potential risks [5]. Existing evidence is insufficient regarding the consequences of alternative models of initial management, both in perspectives of child, parents, health services, and society [5, 12, 23]. Information on outcomes in relation to resource use is necessary when strategic decisions on the allocation of scarce healthcare resources are made [24]; it is required by government bodies in a large number of countries [25‐27].

The objective of this study was to close some of the gaps in evidence by comparing two alternative regimens for children diagnosed with type 1 diabetes: hospital-based care and hospital-based home care (HBHC), referring to specialist care in a home-based setting. We have previously reported on metabolic control, health-care satisfaction and health-care costs one month after diagnosis [28]. In this paper, the focus is on metabolic control, health-related quality of life, direct (i.e., healthcare) costs, and indirect costs (i.e., productivity losses) two years after diagnosis. Costs are related to metabolic control of the child’s disease (cost-effectiveness analysis) and to health-related quality of life of the parents (cost-utility analysis).

The study design was based on the British Medical Research Council framework for development and evaluations of Randomised Controlled Trials (RCT) for complex interventions [29, 30], and has been described in detail elsewhere [31]. The study follows the Consolidated Standards of Reporting Trials (CONSORT) recommendations [32]. Statistical power calculation included the primary outcome HbA1c two years from diagnosis. In order to show a mean difference of 10.5 mmol/mol (Mono S: one percent) between two groups with the power of 0.80 at a significance level of 5%, it would take 30 children in each group. Randomisation in two strata – (a) younger than eight years and (b) eight years and above – was performed by an independent centre for clinical research, using the software R-2.6.1 [33]. The investigators received two sets of coded, sealed, opaque envelopes, one for younger children and one for older children.

The children’s medical variables and parent’s background characteristics at diagnosis are presented in Table 2.

This study compared two different regimens for the initial management of children diagnosed with type 1 diabetes; hospital-based care and hospital-based home are (HBHC). The results two years from diagnosis could not detect any statistically significant difference between the two groups as concerns efficacy and child safety in terms of metabolic control and episodes of severe hypoglycaemia. This is in line with the earlier follow-up from the same study [28, 46, 47] and with another recent retrospective study comparing different sites for initial diabetes education [17].

Even though no statistically significant difference could be detected in the present study, there might still be real but not detected differences in the population at large. When interpreting results, one must hence also take into consideration the difference that the study was powered to detect [48]. Even though our power calculations were performed prior to the study and at the time, with an estimated clinically relevant difference of HbA1c, the defined effect in HbA1c might not be a likely effect size as to expect. However, there is a balance to be made between sample size and external validity over variation in settings by the period of recruitment. Treatment manuals were used and checked regularly but healthcare evolves continuously. This was evident in the routines of a gradually increased number of outpatient visits to the diabetes nurse in the hospital-based care [28, 47]. Healthcare is inherently complex and a potential problem occurs, when results of studies that have been performed under circumstances that are not representative of the forms of care normally in practice, are generalized to uncontrolled circumstances [49, 50]. Since the intervention was intended to be implemented within clinical practice, if found to be a safe and effective way of caring for a child, there needed to be a balance between the validity of inference and the implementation of care [48].

The response rate at 24 months was lower than at previous follow-ups, probably explained by changed routines of data-collection. In the first years, a research assistant who was not involved in the care assessed the outcomes and booked appointments with families outside the hospital in order to let them answer the questionnaires. Since the inclusion of families took longer than planned, we needed to change the routines, implying that the questionnaires were sent home by mail with a return envelope to the families instead. This might have brought in variability in the final outcomes. However, the results from the 24 month follow-up are in line with the earlier follow-up of the same study. Even though no statistically significant difference could be detected in parents’ reporting on the impact of a chronic paediatric health condition and parents’ health-related quality of life, the direction of effect might indicate that the initial management and the initial support given to families have an impact on families’ future experiences of living with a child diagnosed with diabetes. Although fathers had significantly higher incomes than mothers and, hence, higher lost value of production per day absent from work in both the HBHC group and the hospital group, there was a significant difference between the two genders in the number of days absent from work only for the hospital group (in which mothers had twice the number of days of absence than fathers). In a qualitative evaluation of this RCT study [51], the and learning conditions in HBHC seemed to support partnership and collaboration. The parents in HBHC described how they became active participants in negotiations of daily management and decision making in contrast to their passive role in the hospital-based care. Although the difference among parents in the two groups may be due to some underlying factor not controlled for, an interpretation might also be that the main active component in the HBHC has supported the fostering of parents’ responsibility.

Parents in HBHC have, in the earlier follow-up consistently reported significantly greater satisfaction with healthcare compared to parents in the hospital-based care [28, 46, 47]. Despite that difference in the management between groups only involved the first month from diagnosis, parents continued to report a significantly greater level of satisfaction with the healthcare two years from diagnosis. In addition to the home-like environment, the individualised learning process was an important part of the intervention bringing a focus on normalization, supporting the family to find strategies as to achieve good glycaemic control in combination with a re-establishment of their normal lifestyle. In the qualitative evaluation of the study the hospital-based care was considered as being safe but not family- or diabetes-oriented [51]. The HBHC was described as a relaxed environment, providing individualized accessibility and possibilities for situational learning and was considered as more flexible, promoting normality and involvement. This is in line with what others have found when children and young people described inflexibility in clinic processes and the dissonance this represented as compared to families’ expectations of a normal life [6]. Parents who experienced hospital-based care when their child was diagnosed with diabetes felt as if their role as parents, the person the child could rely on, was taken away from them [14]. This made them and their families feel insecure; so holding on to the hospital routines that they initially learned [12] might be one way of feeling more in control in a distressing and life-changing event.

In line with available estimates of the cost of diabetes disease to society at large [7, 8], we found that the indirect costs of lost production were substantially higher than the direct healthcare costs in both care groups. This fact is worth noticing, even though no significant difference could be observed between the two groups. The dominance of indirect costs also meant that no significant difference could be detected in total costs (direct and indirect costs together). So, whether HBHC might be cost-effective also in a wider societal perspective remains to be shown. Both perspectives are important for policy-makers. In some countries, the United Kingdom, for instance, government authorities recommend the healthcare perspective to be used, when allocating resources in the National Health Service [27], whereas in other counties, for instance, Swedish central government authorities use the societal perspective [26].

Few studies have provided high-quality evidence when comparing hospital-based care with different models of home-based care. The results of this study support the safety and feasibility of HBHC, when a child is diagnosed with Type 1 diabetes. Our results further suggest that the initial period of care and the strategies for diabetes management that are first presented to the family may exert an impact on families’ daily living for a longer period of time. The results also suggest that HBHC is a cost-effective strategy, at least in a narrow healthcare sector perspective.

CONSORT, consolidated standards of reporting trials; HbA1c, glycated haemoglobin; HBHC, hospital-based home care; PedsQL, pediatric quality of life; RCT, randomised controlled trial; SEK, Swedish crown; SF-36, short form – 36 health survey; SF-6D, short form – 6 dimensions.