The COACH programme in this study utilised patient education and empowerment as well as decision support as interventional strategies to improve cholesterol control. Both study arms had shown improved LDL-C and TC level throughout the study period. The results were a trend towards add-on improvements in both LDL-C and TC levels when patients were co-managed by nurse educators, even though this was not statistically significant.
The lack of statistical significance was most likely attributed to dilution of treatment effects with the use of similar patient education methods in the control group as well as in the intervention group, specifically, using the COACH health booklet. The COACH programme applied in our study delivered more comprehensive care than usual compared to the typical Malaysian primary care settings, which usually have little time or resource to provide comprehensive disease management care as part of daily practices
. As a result, in this study, patients from both study arms would have benefited from increased knowledge of their health conditions.
Although the COACH programme used in this study was similar to other studies published by Vale et al.
 and Allen et al.
, there are significant differences in terms of study methodology. In the study conducted Vale et al.
, patients were continuously coached based on previous assessment visit and progress monitored via negotiated action plan. In addition, the Vale’s study focused on hospital based disease management programme. On the other hand, the COACH programme by Allen et al.
 utilised the model of community-based participatory research methodology to design the disease management programme, which consisted of enhanced usual care with or without intensive disease management by nurse practitioner/community health workers. Both studies had adopted many essential elements from the Chronic Care Model
, which had been proven to improve chronic disease outcomes. In contrast to this study, the authors had only adapted the previous COACH studies as a patient support programme rather than a disease management programme due to severe resource limitation i.e. lack of trained nursing support resulting in resistance from local doctors to adopt a shared-care model on patient disease management counselling. As a result, the nurse educators in this study had no access to patient’s medical records and health education was reinforced using the health education booklet only.
Evidence for the success of disease management programmes involving additional support from nurses or other disciplines is mixed. Our findings are comparable with some studies
[14, 20] where reduction in blood cholesterol was similarly equivocal due to the Hawthorne effect among patients and healthcare providers, a change in people’s behaviour when being observed. In addition, the study may have heightened the awareness of disease management practices among family physicians involved. However, studies published by Vale et al.
 and Allen et al.
 that utilised a similar programme, demonstrated significant improvement in LDL-C and TC change from baseline when patients with coronary heart diseases underwent the programme. As opposed to our study, the positive outcomes demonstrated in both the studies may be due to: (1) the population enrolled i.e. the study recruited CHD patients experiencing acute coronary syndrome who were more motivated to change; (2) difference in setting i.e. developed versus developing country; (3) difference in prescription behaviour i.e. higher proportion of patients prescribed with statin and with higher doses of atorvastatin in the Vale et al. study (not commonly prescribed among the doctors in the DISSEMINATE study). Besides the 2 studies discussed above, there are also other studies that have had positive results with multi-disciplinary support
The DISSEMINATE study had also revealed interesting findings with regards to HDL-C decrease over the 6 months duration. This finding was inconsistent with many clinical studies that reported mild improvement of HDL-C with both drug treatment and disease management care
[25, 26]. Several theories are hypothesised for this outcome. First of all, there is some evidence that a low fat diet reduces not just LDL-C but also HDL-C
[27–29]. Secondly, the COACH programme did not provide detailed nutritional education with information on how to reduce high dietary fat intake while substituting dietary fat with polyunsaturated fatty acids (PUFA) such as olive oil. It is possible that both of these factors could have resulted in the unexpected reduction in HDL-C levels detected in our study.
Several limitations to our study are identified. In an effort to standardise patient care between different sites, a health booklet was distributed to all subjects. As discussed above, this had diluted the difference in interventional effects between control group and intervention group. Also, the findings in the control arm would not be generalise to the current primary care settings as it had deviated from the normal local practice; however, the outcomes from this study has demonstrated that current chronic disease care in Malaysia primary care practices are suboptimal and some improvements in patient’s disease outcome can be achieved by simply spending extra time to educate the patients on chronic disease self-management and treatment compliance. And furthermore, our study has also shown that the task of personalised patient education does not have to be purely the domain of the physicians but can also be provided by trained nurses. The study might have elicited a more significant finding with addition of a third study arm assessing the lipid outcomes among patients who received “actual” standard care delivery by local physicians. Though the primary outcome of the study was not achieved, an additional post-hoc analysis may be able to determine the proportion of uncontrolled dyslipidaemic patients who achieved cholesterol target at the end of the study follow-up. In the current analysis, the proportion of patients who achieved target cholesterol level is not known. A higher than expected attrition rate (initial sample size calculation only accounted for 10% loss to follow-up) could have also affected the final results of the study.
Although our study showed that complementary nurse support services in a disease management programme had a positive trending impact on LDL-C and TC level compared to physician management alone, the effect was not sustainable after the intervention was withdrawn at week 24. Continuous therapeutic behavioural change seems to be mandated to ensure long-term sustainable lipid control.