  |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Journal of the New Zealand Medical Association,
14-March-2008, Vol 121 No 1270
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Retention of patients in the Get Checked free annual
diabetes review programme in New Zealand
Grace Joshy, Ross A Lawrenson, David Simmons
The rising diabetes epidemic is a major problem in New
Zealand, especially among the non-European ethnic
groups.1 The cost of Type 2 diabetes alone is
expected to reach NZD1.6 billion by year 2021.2
Many of those with diabetes face a multitude of barriers to quality diabetes
care and self care,3,4 including financial
barriers.5 Structured care has been shown to
improve patient care and outcomes.6–8 One
strategy used elsewhere to ensure that each person with diabetes received
regular structured assessment has been the annual diabetes
review.9–11
The Get Checked free annual diabetes review programme in New
Zealand—a Ministry of Health initiative which started in June
2000—was established to address both the need for structured care and to
help overcome personal expenses as a barrier to diabetes care. This national
programme is delivered free of charge to diabetes patients through primary care
services.
The review form collects data on whether the following
checks or tests were either done or booked to be done: retinal screening (within
the last 2 years), foot check, blood pressure,
HBA1c, cholesterol, height, bodyweight, and
kidney function in the last 12 months. Available test results are also
collected.
Patients are expected to return for review every year.
Nationally, the percentage of people with diabetes enrolled in the Get Checked
programme increased from 33% in 2001 to 59% in 2005, but still the figures are
sub-optimal, especially for Māori.12,13
The data purportedly provide insight into diabetes care across New
Zealand,14,15 but we have been concerned about
drop out from the programme and the problem of interpreting data through the use
of repeated cross sectional rather than longitudinal analysis. We therefore
aimed to investigate the patient retention in this programme using data from the
local primary healthcare organisation.
MethodsWaikato Primary Health (WPH) is the largest primary
health organisation in Waikato area with a registered population of 294,510 in
2006. It covers 90% of the 328,510 Waikato District Health Board population
enrolled with a primary health organisation
(PHO),16,17 with an estimated 10,604 people
diagnosed with diabetes.18
This research is a retrospective review of WPH
registered patients who had at least one Get Checked review between 1 July 2000
and 30 Jun 2006, using the demographic variables from the Get Checked database
(age, gender, latest recorded ethnicity, and type of diabetes). Year of
diagnosis of diabetes was not available for analysis.
Mortality data were obtained from the Ministry of
Health and were linked to WPH patient register using the national health index
numbers. Some patients left WPH after their initial review and were not
available for further reviews. The WPH registrations were recorded on a
quarterly basis for each patient.
Survival analysis was employed to analyse the time to
second review from the initial review. In order to look at continued
participation beyond the second review, time to third review from second review,
time to fourth review from third, and time to fifth review from fourth were also
analysed in a similar fashion.
Those who died or left WPH before a second review and
were considered “censored” for the analysis of time to second
review. Those who did not return for a second review during follow-up time were
censored on 30 June 2006. A similar approach to censoring was also used for the
analysis of time to subsequent reviews. For censored patients, the time to event
was the start date to date of death, migration date or end of follow-up, which
ever was the earliest. Migration date was defined as 45 days after the last
quarter of registration with WPH.
Kaplan-Meier survival cures for time to reviews are
presented. Allowing a 6-month window to the ideal 1-year of interval between
reviews, review rates at 1.5 years were examined. Survival curves for time to
second review are presented by ethnicity, age group at first review, gender, and
type of diabetes.
Odds ratios for the likelihood of a second review were
estimated using Cox’s proportional hazard model. Potential predictors were
identified by running a series of regression analyses. Ethnicity, age group at
first review, gender, and diabetes type were included as predictors in a
Cox’s regression analysis. Proportionality assumption was verified by
testing the correlation between Schoenfeld’s residuals for a particular
covariate and individual failure times. All statistical analyses were performed
using SAS® version 9.1 software (SAS
Institute, Cary, NC, USA).
ResultsA total of 10,919 patients were reviewed at least once
during the 5-year period (Table 1). Ethnicity was recorded for 95% of patients,
showing 69% Europeans, 18% Māori, 3% Pacific Islanders, and 4% Asians.
Of the reviewed patients 87% had Type 2 diabetes, 8% had
Type 1 diabetes, and 5% had other or unclassified diabetes. At first review,
Europeans patients were on average a decade older (65.1±14.1 years) than
other ethnic groups.
In 2005/06, 6135 patients attended a review, including 933
(15%) Māori. Of these, 1345 were new patients, attending their first
review. During the each year of this study, between 1300–2100 patients
attended their first Get Checked review (Figure 1). The proportion of newly
diagnosed patients among those attending their first review is unclear.
Table 1. Patient characteristics at first
review by ethnicity
Figure 1. Uptake of patients in the programme
(January 2000–June 2005)
 Survival analysis shows that 7142 patients returned for a
second review within a median time of 1.17 years after initial review (Table 2,
Figure2). The survival distribution function (SDF) at a time represents the
proportion of patients who have not returned for a review up to that time.
At 1½ years after initial review (allowing a 6-month
window to the ideal 1-year timeframe), 35% of eligible patients were yet to
return for a second review. At 5 years after the first review, 15% had not
returned for a second review. Those who continued participating in the programme
after second review returned for subsequent reviews on a much more regular
basis. At 1½ years after second review 75% of eligible patients had
returned for a third review. High proportions of patients (35%–46%) were
censored for each review. The proportion of patients censored due to death and
migration were relatively small (Table 2).
Table 2. Survival analysis of time to review
(from the previous review)
*Conditional that patient attended this review.
Figure 2. Kaplan-Meier survival curves for the
time to reviews (from the previous review, conditional that patients attended
the previous review)
 Table 3. Survival analysis of time to second
review (from initial review)
P values from Wilcoxon’s test for homogeneity of
survival cures over strata; *Mostly of Samoan, Tongan, Niuean, or Cook Islands
origin.
Table 4. Odds ratios (95% confidence intervals)
for a second review
Odds ratios from Cox’s Proportional Hazards
Model.; The multivariate model included age group (<40, 40–65, 65+),
ethnicity (European, Māori, and Asian), and type of diabetes ( Type 1 and
Type 2); Gender and Pacific ethnicity were found to be non-significant
predictors in the multivariate analysis and were excluded from the final
model.
Patients of Māori or Asian origin, younger patients,
and those with Type 1 diabetes took a significantly longer time to return for a
second review (Table 3). Survival curves for time to second review were
significantly different for subgroups of ethnicity, type of diabetes and age at
first review (Figure 3). No significant gender difference was found.
After covariate adjustment for age, ethnicity, and type of
diabetes, younger patients aged<40 years were less likely to return for a
second review compared with those aged 65+. Māori and Asian patients were
less likely to return for a second review compared with Europeans, and those
with Type 1 diabetes were less likely to return for a review compared with Type
2 patients (Table 4). Pacific ethnicity and gender were not found to be
significant predictors of return for a second review.
Figure 3. Kaplan-Meier survival curves for time
to second review (from initial review) by subgroups of ethnicity, age, gender
and diabetes type
 DiscussionThe WPH Get Checked programme, which started in 2000, had
1300–2100 new patients on board every year. WPH estimates that in 2005/6
there should be 10,600 patients with known diabetes within the organisation
network. As pointed out by the recent audit
report,19 it is difficult draw detailed
conclusions on the coverage of the programme since patient level information on
all people with diabetes is not available.
A separate database held by the Waikato Regional Diabetes
Service (WRDS), which provides secondary diabetes services and retinal screening
to diabetes patients in the Waikato region, had 9936 patients registered in year
2005, with 2006 (21%) patients being
Māori.20 Yet in 2005/06 only 6135 WPH
patients turned up for a free diabetes check, of which 933 (15%) were
Māori. The Get Checked data could not be merged with WRDS database since
the Get Checked dataset provided for analysis was anonymous. Merging these two
datasets would have been valuable to estimate the proportion of WRDS patients
attending Get Checked review and to understand the profile of patients who never
entered the Get Checked programme.
Patients coming for their first check in any one year
include newly diagnosed, those who have been diagnosed for some time but new to
the programme, and existing patients who recently moved in to the Waikato
region. Data from the Get Checked programme is underestimating the number of
people with diabetes in the region. A regional diabetes register will help to
evaluate the coverage of the programme and characterise patients who are not
using the free check. Data fed from local registers to a national register could
potentially maximise the use of Get Checked data with regular reports to the
Local Diabetes Teams, thus enhancing planning and provision of diabetes
services.
Despite this programme being free to patients, a significant
proportion of patients did not return for a second review within 1.5 years after
initial review. The profile of patients who were retained in the programme with
regular reviews was quite different from the irregular attendees and those who
dropped out.
Younger patients aged <40 years, those of Māori or
Asian origin, and those with Type 1 diabetes were less likely to be retained in
the programme with regular checks, as indicated by the longer time to second
review and lesser likelihood of returning for a second review. In the UK,
predictors of attendance for review in general
practice were older age, less comorbidity, and being of European
ancestory.19
A South Auckland study in the early 1990s showed that
patients who defaulted from diabetes care were younger, diagnosed at a younger
age, more likely to be in paid employment, knew less about diabetes, and were
less likely to require medication.20
Out-of-pocket expenses also impede diabetes self-care in New
Zealand.3,5
Māori, Pacific, and Asian people have considerably
higher rates of diagnosed diabetes compared with European people (European/Other
2.9%, Māori 8%, Pacific 10.1%, Asian
8.4%).21 They also have an earlier onset of
Type 2 diabetes and higher rates of diabetes
complications.1
Overall, the finding that Māori and Asians were less
likely to attend a second review is a concern and it indicates these diabetes
patients still have problems with access to appropriate health care. In order to
minimise this inequality, remedial measures are needed to increase the uptake of
this free review among ethnic minorities.
Younger Type 1 patients with complications may be visiting
the WRDS regularly and may not deem it necessary to go to their GP for a
separate annual check. In 2005, there were 1338 (13%) Type 1 patients registered
with WRDS,20 but only 890 (8%) of the patients
who had had a Get Checked review had Type 1 diabetes.
Of the patients reviewed in 2005-06, 405 had Type 1
diabetes. Thus these patients may in fact be receiving good follow-up care,
although the results from this study indicate that these patients may have
difficulty accessing their health services. However, a high proportion of
younger patients are likely to be in the workforce and patients in paid
employment are known be less likely to access
care.22 They are also more mobile and could be
getting treated elsewhere (perhaps managed by WRDS) but still registered with
WPH.
It was outside the scope of this analysis to track continued
participation of patients who migrated and changed to a different PHO, after
completing their initial Get Checked review with WPH. A national diabetes
register will be needed to make cross-regional comparisons. Migrated patients
were censored in the survival analysis and were not a major issue. Data
concerning year of diagnosis of diabetes was not available.
It would have been of considerable interest as the
participation rates of those newly diagnosed with diabetes may differ from that
of others. Similarly, data concerning geographical location of residence was not
available. Physical access to care including transportation has been identified
as a barrier to care in a previous
study.23
The Get Checked programme provides a beneficial service for
many people with diabetes, but the data are significantly limited as a source of
understanding of diabetes care in New Zealand. Although the programme is
adequately funded and provides more consistent care for patients, practical
challenges need to be dealt with and cost-benefit of the programme may need to
be more clearly demonstrated to the patients.24
Our longitudinal analysis reveals concerns regarding
retention rates in the Get Checked programme, especially for Māori and
patients with Type 1 diabetes. Further research aimed at understanding the lower
retention rates for these groups may help to improve health outcome
disparities.
Competing interests: None known.
Author information: Grace Joshy, Research
Fellow in Diabetes Epidemiology, Waikato Clinical School, University of
Auckland, Hamilton, New Zealand;
Ross A Lawrenson, Head of Waikato Clinical School, University of Auckland, Hamilton, New Zealand; David Simmons, Professor, Institute of Metabolic Science, Cambridge University Hospitals NHS Foundation Trust, Cambridge, United Kingdom. Acknowledgements: We thank Waikato Primary
Health and Pinnacle Group Pte Ltd for providing the data for analysis and for
their support. University of Auckland researchers Timothy Kenealy, Greg Gamble,
and Steven Lillis are also acknowledged for their valuable input.
Correspondence: Grace Joshy, Research
Fellow-Diabetes Epidemiology, University of Auckland, Waikato Clinical School,
Waikato Hospital, Private Bag 3200, Hamilton, New Zealand. Fax: +64 (0)7
8343615; email: JoshyG@waikatodhb.govt.nz
References:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|