Five-year change of clinical and complications profile of diabetic patients under primary care: a population-based longitudinal study on 127,977 diabetic patients

Diabetes mellitus (DM) is a significant major chronic disease and it was estimated 1 in 12 people was affected by diabetes in the world [1]. The World Health Organization projected that by 2030, diabetes would be the seventh most common cause of death in the world [2]. A good control of DM is crucial because DM is closely linked to various complications, ranging from different cardiovascular diseases (including myocardial infarction, cerebrovascular diseases, etc.) to microvascular diseases like diabetic retinopathy [including sight-threatening diabetic retinopathy (STDR)] and diabetic nephropathy [including end stage renal failure (ESRF)]. Health spending on diabetes increased to 548 billion, accounting for 10.8 % of total health expenditure worldwide in 2013 [1]. Patients with DM complications significantly increased medical costs compared with patients without DM complications [3]. A huge number of diabetic patients and substantial public health burden will be further exacerbated by the rapidly ageing population in the world. To prevent diabetic complications, many overseas national professional associations have developed guidelines on the management of diabetes [4‐7]. Similar initiative was also called for to assure the management and quality of care of diabetes for the population in Hong Kong (HK).

Hong Kong, being an international city with many people leading a Western lifestyle, hosted a significant and increasing number of diabetic patients in recent years. Local data showed that the prevalence of diabetic mellitus was 9.92 % (diagnosed plus undiagnosed) [1]. Total annual costs for diagnosed Type 2 DM patients in HK were estimated to take up 6.4 % of the total expenditure of the public healthcare sector in 2004 [8]. In order to enhance the quality of care of patients with chronic diseases like diabetes mellitus under the public primary care setting, various territory-wide quality enhancement programmes have been introduced in phases, including Risk Assessment and Management Programme-Diabetes Mellitus (RAMP-DM), Patient Empowerment Programme, Nurse Allied Health Clinic, Call Centre, etc. RAMP-DM is exclusively for diabetic patients, and it has started implementation since 2009, with its coverage raised from 3.1 % in 2009 to 81.9 % in 2013 of all the diabetic patients under the care of public primary healthcare sector. The details of the RAMP-DM had been illustrated in our previous protocol paper [9]. Diabetic patients are encouraged to live and work in the community, and primary care approach of community-based care are appropriate to majority of the diabetic patients who do not require specialist care [10, 11]. In spite of this predominant chronic disease in the society, there were no formal standardized guidelines or recommendations provided to the local health care doctors as a local reference on how their diabetic patients should be managed before 2010. In 2010, the HK Reference Framework for Diabetes Care for Adults in Primary Care Setting was published and served as a population-wide guideline for strengthening community-based primary care, focusing on prevention, and quality improvement in the management of DM [12]. Previous studies demonstrated that healthcare systems relying more on primary care rather than specialist care produce better population health outcomes, enhance continuity and access to health care [13‐17]. Nevertheless, currently, limited primary care and population-based data on patient’s clinical and complications profile of diabetic patients was available to identify the needs of the practices and population. Continued territory-wide evaluation of diabetes control in the population is vital to sustain improvement in diabetic care and inform health policy makers in service planning and resource allocation to prevent the development of complications and preserve quality of life of the diabetic population.

The aims of this study were to track the recent 5-year changing diabetic care by reviewing the clinical and complication profile, to evaluate the trend in risk-factor control, and to identify the needs and gaps of the practices among diabetic patients based on the 5-year population-based data from 2009 to 2013.

The flow of diabetic subjects in the study was summarized in Fig. 1. A total of 150,157 diabetic patients received public primary care at 2009. Among these patients, 127,977 (85.2 %) subjects were found continuously receiving public primary care at 2013 and their data was included for analysis in this study.

The data completion rates for socio-demographic and clinical parameters in 2009 and 2013 were shown in Additional file 1. Clinical data (e.g., BMI, waist–hip ratio, HbA1c, SBP, DBP, full lipid profile and urine ACR, etc.,) completion rate improved significantly. Table 1 and Additional file 2 showed the socio-demographics of the 127,977 DM patients in 2009. Mean age was 64.0. Female was around 10 % more than male.

Tables 2, 3 and Figs. 2, 3 compared their clinical variables at 2009 and 2013. A significant increase of proportions of patients achieving treatment targets, including HbA1c <7 % (increased from 47.5 to 56.5 %), SBP <130 mmHg (increased from 47.5 to 56.5 %), DBP <80 mmHg (increased from 65.7 to 77.5 %), LDL-C <2.6 mmol/L (increased from 25.9 to 65.6 %), TG <1.7 mmol/L (increased from 61.6 to 74.9 %), and HDL-C >1.0 mmol/L (male) or >1.3 mmol/L (female) (increased from 47.0 to 60.6 %) were observed (p < 0.001). Mean values of clinical variables also dropped significantly including HbA1c (from 7.2 to 7.0 %), SBP (from 136.9 to 131.3 mmHg), DBP (from 75.4 to 72.1 mmHg), LDL-C (from 3.1 to 2.4 mmol/L), TG (from 1.7 to 1.4 mmol/L), and BMI (from 25.6 to 25.3 kg/m²). On the other hand, more proportion of diabetic patients had central obesity (raised from 77.1 to 81.3 %) and increased ACR (raised from 23.0 to 26.3 %).

Table 4 showed the change of drug use pattern between 2009 and 2013. More patients were prescribed oral anti-diabetic medications, with metformin being the most commonly prescribed drug (raised from 71.1 to 80.4 %). A lower proportion of patients used sulphonylurea group of drugs and 1.2 % of patients started using gliptin. More patients (raised from 0.6 to 3.5 %) used insulin in addition to their oral anti-diabetic drugs for their management, and a significant boost (from 9.0 to 55.0 %) on statin use was observed.

Table 5 illustrated the 5-year cumulative incidence of major DM complications in the period from 2009 to 2013. A total of 8,625, 572 and 155 patients with CVD, STDR, and ESRD, respectively, at baseline were excluded for the calculation of 5-year cumulative incidences of corresponding events. Hence, the 5-year cumulative incidences for CVD, STDR, and ESRD were 5.1, 0.7, and 0.8 %, respectively.

This article is by far the latest and largest territory-wide review of the status and control of health parameters of diabetic patients, with a follow-up of a 5-year period. From 2009 to 2013, the quality of care had been enhanced both in terms of the structure, like documentation of the patients’ clinical data, as well as the process, like the annual checking of diabetic patients’ clinical variables. This undoubtedly raised the standard of care being delivered to patients with chronic diseases like diabetes mellitus as previous records were made accessible and thus continuity of care was facilitated. Having all these clinical variables checked and documented allows regular systematic audit of how these patients progress further on.

The standard of care for diabetic patients under primary care in HK, as reflected by the control of HbA1c, BP and LDL-C, is comparable with that achieved in the developed countries such as UK and US [20, 21]. Improvement in coverage of annual checking of key clinical variables may attribute to the more regular monitoring, which allowed early intervention, such as medication intervention. This as a result led to the improvement in the proportions of reaching treatment targets.

All these changes were paralleled with the implement of the territory-wide quality enhancement programmes since 2009. Amongst these programmes, the RAMP-DM coverage rate raised from 3.1 % (2009) to 81.9 % (2013) of all the diabetic patients under public primary care. The details of the RAMP-DM had been illustrated in our previous protocol paper [9]. One of the key impacts was the advocating use of insulin and statin for the better control of diabetes mellitus and the associated hyperlipidemia at the primary care level. The significant increase of patients using insulin in addition to oral anti-diabetic drug, together with the increased proportion of patients using metformin and newer class of oral anti-diabetic drugs like gliptin class are closely linked to the reduction of mean HbA1c from 7.2 to 7.0 % in 5 years, and the increased proportion of patients having HbA1c <7 %. Based on the UK prospective diabetes study [22], every 1 % reduction in mean HbA1c was associated with reductions in relative risk of 14 % for all-cause mortality and 9.9 % for CHD. Therefore the 0.2 % reduction of HbA1c shown in our study has the potential impact of having a reduction of relative risk of 2.8 % for all-cause mortality and 2.0 % for CHD.

Most Chinese patients refused insulin injection as they have the misconception that insulin injection was painful and they had needle phobia, especially in those who have never started insulin [23]. Education on insulin usage helps patients clear these kinds of misconception and allows them to accept the daily injection of insulin and improve glucose control [24]. Coupling the drop in mean HbA1c was the drop in proportion of diabetic patients solely on diet or lifestyle modification alone. This was in-line to the argument that metformin should be initiated early in the management of diabetic patients. Metformin is the first-line oral anti-diabetic drug and recently there were some studies showing its benefit on reducing cardiovascular complications, in addition to that of lowering HbA1c [25, 26]. With the increased proportion of our patients using metformin, we will study the link between metformin use in Chinese patients and cardiovascular risk. Newer classes of oral anti-diabetic drugs like Dipeptidyl Peptidase-4 Inhibitors class were relatively expensive and patients need to self-purchase these newer drugs in most of the circumstances, which lowered the popularity of such drug usage. As aging population and the associated pancreatic function deterioration is foreseen, the effectiveness of these newer class of anti-diabetic drugs has to be studied. The dramatic boost of statin use explained the drop of mean LDL-C from 3.1 to 2.4 mmol/L and the increased proportion of patients achieving LDL-C target of <2.6 mmol/L. A previous study showed that an increment of 1 mmol/L in LDL-C concentration correlates with a 1.57 increased relative risk of coronary heart disease [27], which equated to a 36 % relative risk of CHD for a decrement of 1 mmol/L in LDL-C. Therefore the 0.7 mmol/L reduction of LDL-C shown in our study has the potential impact of having a reduction of relative risk of 25.2 % for CHD. Proportions of diabetic patients with hyperlipidemia being prescribed statin raised tremendously from 15.8 % in 2009 to 56.9 % in 2013. However there was still around one-third (37.5 %) of patients not achieving target LDL-C with around 40 % of diabetic patients with hyperlipidemia still not being given statin in 2013. The underlying reasons needed to be further studied. On the contrary, the number of patients using fibrate was reduced but the overall control of triglyceride improved with dropping of mean TG from 1.7 to 1.4 mmol/L, and more patients achieving target of TG <1.7 mmol/L. One of the hypotheses may be the partial triglyceride lowering effect of statin which was prescribed extensively [28, 29], or the associated reduction in BMI [30, 31].

Both mean SBP and DBP showed significant drop of 5 and 3 mmHg, respectively, and more patients achieving target SBP of <130 mmHg and DBP of <80 mmHg. Previous literature demonstrated that each 10 mmHg decrease in mean SBP was associated with a relative risk of 12 % for all-cause mortality and 13 % for CHD [32]. Hence, the 5 mmHg reduction of SBP shown in our study has the potential impact of having a reduction of relative risk of 6.0 % for all-cause mortality and 6.5 % for CHD. Similarly, this may be explained by the increased use of anti-hypertensive medications within these 5 years. It was noteworthy that calcium channel blocker was the most prevalent anti-hypertensive drugs used in our diabetic patients (over 70 %), while ACEI or ARB ranked only the second (just above 50 %). ACEI or ARB, due to its renal protective effect, are supposed to be beneficial for diabetic patients and are preferred. Proportions of diabetic patients using ACEI or ARB were static at around 59.4–58.3 % between 2009 and 2013. The under use of ACEI or ARB was also reflected by the elevation of mean urine ACR from 4.7 mg/mmol in 2009 to 7.9 mg/mmol in 2013, and the associated increased proportion of patients having urine ACR >2.5 mg/mmol (male) and >3.5 mg/mmol (female). Chinese patients had about 50 % chance of suffering dry cough on ACEI [33, 34]. The alternative use of ARB could help solve the problem while offering renal protective effect [35, 36], despite the cost of ARB over ACEI is an issue. Appropriate prescription of ARB in diabetic patients should be advocated.

In spite of the improvement in the overall mean HbA1c and full lipid profile, the mean BMI dropped only 0.3 kg/m² after 5 years, and the proportion of patients having central obesity raised from 77.1 to 81.3 %. This suggested that body weight and body fitness management may not be feasible if exercise was not taken into account. The impact of exercise should be evaluated on the control of diabetes mellitus, SBP, DBP, and full lipid profile, and should be integrated to the management plan of diabetic patients [37, 38]. Similarly, despite the significant improvement in HbA1c, SBP, DBP, and full lipid profile, the 5-year cumulative incidence of major DM complications including various CVD were shown higher than that of the general population in Asia [39]. Previous studies also showed that diabetic patients were about 2–4 times more likely to have CVD than non-diabetic patients [40, 41]. This difference was expected because hypertension, hyperlipidemia, hyperglycaemic and obesity, those contribute to the risk for developing CVD, occur more frequently in patient with T2DM compared to general population [42]. In this study, there were a total of 6117 new CVD events amongst those without CVD events at baseline in 2009 after 5 years. This may be related to the concurrent elevation of urine microalbumin level, which was shown to be a significant prognostic factor for various DM complications [43]. Use of aspirin was raised from 8.8 % in 2009 to 14.0 % in 2013, and the majority of those patients having complications were put on aspirin. Aspirin usage as a primary preventive measure in diabetic patients was unknown in our locality and worth exploration. Also, patient’s self-management and adherence of therapy play essential role in achieving the goals of diabetes care and thus further studies is needed to identify and evaluate such characteristics of patients.

Our study showed the significant improvements in the management of DM under primary care in HK. Improvement in coverage of annual checking of key clinical variables and patients’ achievement to target values may be attributed by the regular monitoring and early intervention, including use of appropriate medications and the risk-stratification based multidisciplinary DM management (RAMP-DM). There is still room for further enhancement in the diabetes care, including body weight management and counseling in addition to the achievement of optimal targets. Regular evaluation of diabetes control is crucial to identify and sustain improvements in management and care of diabetic patients.