ncreasing evidence that type 2 diabetes mellitus (DM) may be pre lifestyle interventions to prevent or delay type 2 diabetes in people …
I
CLINICAL
I
Diabetes Diagnosis, Resource Utilization, and Health Outcomes
Martin C Gulliford, FFPH; Radoslav Latinovic, BSc; and Judith Charlton, MSc
Objective: To determine the effect of a clinical diagnosis of diabetes mellitus DM on healthcare utilization and health outcomes Study Design: Cohort study Methods: A total of 197 United Kingdom family practices with 4974 subjects mean age, 628 years; 522 men with type 2 DM and 9948 matched nondiabetic control subjects Healthcare utilization and the occurrence of complications were estimated from 2 years before to 2 years after the first clinical diagnosis of DM Results: From 24 months before the DM diagnosis, primary care consultations were increased in prediagnosis cases compared with controls relative rate [RR], 131; 95 confidence interval [CI], 127-135, as were emergency and hospital care consultations, hospital specialist referrals, and prescription drug items At diagnosis of DM, utilization of all forms of healthcare was increased RR, 427; 95 CI, 417-436 for primary care consultations; RR, 249; 95 CI, 246-252 for prescription drug items In the quarter following diagnosis, healthcare utilization was increased for acute
myocardial infarction RR, 629; 95 CI, 269-1473, cerebrovascular disease RR, 514; 95 CI, 337-784, ischemic heart disease RR, 365; 95 CI, 277-480, and peripheral nerve disorders RR, 501; 95 CI, 281-895 First diagnoses of myocardial infarction, cerebrovascular disease, and peripheral nerve disorders were increased during the period from 6 months before to 6 months after diagnosis Conclusions: Clinical diagnosis of DM is often the end of a process leading to established complications and is associated with greatly increased utilization of care This adds to the justification of strategies for earlier detection of hyperglycemic states Am J Manag Care 2008;14:32-38
ncreasing evidence that type 2 diabetes mellitus DM may be prevented through lifestyle intervention1 and drug therapy2 has focused attention on events occurring before the diagnosis of type 2 DM During the period before diagnosis, the onset of DM may be prevented or delayed in some subjects However, there is continuing uncertainty concerning the justification for screening for type 2 DM or prediabetes3 This uncertainty results in part from limited knowledge concerning the natural history of the period before the clinical
diagnosis of DM Studies 4-6 have used electronic patient records to evaluate healthcare utilization during this period and showed that, in subjects who are later diagnosed as having DM, utilization of primary care consultations and prescription drug items is increased from up to 5 years4 to 8 years5 before the diagnosis compared with matched control subjects who never developed DM The finding of increased utilization of medical care before the diagnosis of DM raises questions concerning the importance of the clinical diagnosis of DM Is the clinical diagnosis of DM an incidental event of little consequence in the course of a slowly progressive condition that is already present? Alternatively, is the diagnosis of DM associated with a transition to a state with a higher rate of occurrence of adverse DM outcomes? To answer these questions, the present study compared utilization of medical care in a cohort of subjects with DM before, during, and after the clinical diagnosis of DM We aimed to determine how the use of healthcare resources and the incidence of adverse health events changed following the clinical diagnosis of DM
I
Ascend Media
METHODS
A retrospective cohort study was
implemented Subjects diagnosed as having DM were compared with age-matched and sex-matched control subjects with respect to the utilization of medical care before and after the diagnosis of DM The study was implemented using The Health Improvement Network, a database containing computerized medical records from family practices in the United Kingdom7,8 The study was approved by the London Multi-centre Research Ethics Committee
Practice and Patient Selection
For author information and disclosures, see end of text
Family practices were selected if they continuously provided data from January 1, 2001, to December 31, 2005
In this issue Take-away Points / p37 wwwajmccom Full text and PDF
32
I
wwwajmccom
I
JANUARY 2008
Diabetes Diagnosis, Resource Utilization, and Health Outcomes
There were 1 283 429 patients 100 years or younger who were continuously registered with 197 family practices during this time From this cohort, medical diagnostic codes for DM and prescriptions for oral hypoglycemic drugs or insulin were used to identify cases with DM The DM diagnosis date was defined as the earlier of the first DM medical code or the first DM prescription date Cases with DM were
selected if their diagnosis date was between January 1, 2003, and December 31, 2003 5635 cases, and if they were aged between 30 and 89 years at diagnosis 5091 cases Two controls were selected for each case matching for age, sex, and family practice from among subjects who had never been diagnosed as having DM or prescribed oral hypoglycemic drugs or insulin Twenty-two cases were excluded because there were insufficient matched controls Finally, we included only 4974 patients who had never been diagnosed as having type 1 DM
Data Analysis
dial infarction, cerebrovascular disease, and peripheral nerve disorders In condition-specific analyses, medical events were combined for primary care consultations, emergency care consultations, hospital admissions, and hospital specialist referrals Initially, all events were considered, and then first events in subjects previously free from the diagnostic grouping were evaluated separately
RESULTS
Data were analyzed for 4974 subjects mean age, 628 years [age range, 30-89 years]; 522 men, drawn from 197 family practices, who were diagnosed as having type 2 DM during 2003 There were 9948 control subjects, never diagnosed as having DM, who were
matched for age, sex, and family practice Figure 1 shows the utilization of healthcare before and after the diagnosis of DM compared with that of nondiabetic control subjects In the quarter from 24 to 22 months before the diagnosis of DM, the RRs were 131 95 CI, 127-135 for primary care consultations, 123 95 CI, 108-139 for emergency and hospital care, 124 95 CI, 114-135 for hospital specialist referrals, and 155 95 CI, 152-157 for prescription drug items The relative excess of medical care utilization continued through the prediagnosis period and then showed a marked increase, reaching a maximum in the quarter following diagnosis of DM At this time, the RR for primary care consultations was 427 95 CI, 417-436 times higher than that for controls; the RRs were 241 95 CI, 218-266 times higher for emergency and hospital care, 297 95 CI, 279-316 times higher for hospital specialist referrals, and 249 95 CI, 246-252 times higher for prescription drug items By 22 to 24 months after diagnosis, utilization of medical care had declined but remained higher than that in controls and higher than that in the prediagnosis period Compared with controls, the RRs for cases were 164 95 CI, 160-168
for primary care consultations, 142 95 CI, 126-159 for emergency and hospital care, 150 95 CI, 142-157 for hospital specialist referrals, and 210 95 CI, 208-213 for prescription drug items Table 1 gives the absolute rates and RRs of healthcare utilization for 3 broad periods before, during, and after the diagnosis of DM confirming these findings Figure 2 shows the RRs of healthcare utilization events, combined across various types of care, for 4 different conditions These analyses of healthcare utilization for specific groups of conditions showed that utilization of medical care was increased from 24 to 22 months before DM diagnosis for
For each case or control, we considered the 2 years preceding and the 2 years following the first clinical diagnosis of DM Person-time of follow-up was divided into quarterly periods 913 days each for analysis Records were censored if subjects died or ended their registration with the practice after the diagnosis For each subject, medical events of interest were counted by quarter These included primary care consultations, emergency care consultations, hospital admissions, referrals to and consultations with hospital specialists referred to as
hospital specialist referrals, and prescription drug items Emergency care consultations included urgent visits by a physician to the patients home and referrals or attendances at hospital accident and emergency departments Emergency care consultations and hospital admissions were combined for presentation The data are coded using Read code clinical terms9 Medical events associated with Read codes for ischemic heart disease, acute myocardial infarction, cerebrovascular disease, and peripheral nerve disorders were identified separately Details of the codes used are available from the author To avoid double counting, a maximum of 1 event of each type was counted on 1 day The rate of occurrence of each type of event per person-year was estimated by quarter for cases and controls The relative rates RRs and associated 95 confidence intervals CIs were estimated using random-effects Poisson regression analysis, with the identifier for each triplet of case and 2 controls as the random effect and the time at risk as included as the exposure for each subject Models were fitted using version 9 of STATA10 In the next stage of the analysis, medical events were counted for different conditions,
and results are presented for ischemic heart disease, acute myocar-
VOL 14, NO 1
I
THE AMERICAN JOURNAL OF MANAGED CARE
I
33
I
CLINICAL
I
I Figure 1 Relative Rates of Utilization of Healthcare Before and After Diagnosis of Diabetes Mellitus Compared With Nondiabetic Control Subjects by Quarter
5
Primary Care Consultations
Relative Rate
3 25 2 15 1
Emergency and Hospital Care
Relative Rate
4 3 2 1 -2 -1 0 1 2
-2
-1
0
1
2
Years
Years
Hospital Specialist Referrals
3 25
Prescription Drug Items
Relative Rate
25 2 15 1 -2 -1 0 1 2
Relative Rate
2
15 -2 -1 0 1 2
Years
Data are given as relative rate 95 confidence interval
Years
ischemic heart disease RR, 213; 95 CI, 160-284 but not for acute myocardial infarction RR, 133; 95 CI, 047375, cerebrovascular disease RR, 100; 95 CI, 057-176, or peripheral nerve disorders RR, 133; 95 CI, 068-262 There were statistically significant peaks in the utilization of care for these conditions at the time of diagnosis In the quarter following diagnosis of DM, the RRs were 365 95 CI, 277-480 for ischemic heart disease events, 629 95 CI, 269-1473 for acute myocardial infarction, 514 95 CI, 337-784 for cerebrovascular disease,
and 501 95 CI, 281-895 for peripheral nerve disorders However, utilization of care for these conditions declined to levels similar to those of controls by 22 to 24 months after diagnosis RR, 105; 95 CI, 070-158 for ischemic heart disease; RR, 201; 95 CI, 041-996 for acute myocardial infarction; RR, 147; 95 CI, 081-265 for peripheral nerve disorders; and RR, 164; 95 CI, 093-287 for cerebrovascular disease Table 2 summarizes how, in the period from 6 months before to 6 months after diagnosis, the absolute rates and RRs of new diagnoses of each condition were statistically significantly elevated
New ischemic heart disease and peripheral nerve disorder events were increased in the postdiagnosis period, but cerebrovascular disease events and acute myocardial infarctions were not
DISCUSSION
We hypothesized that the clinical recognition of DM may be an event of limited significance in the course of a condition that is already present Our results show that this is not the case A clinical diagnosis of DM is associated with a profound increase in healthcare utilization and with increased occurrence of newly diagnosed ischemic heart disease, acute myocardial infarction, cerebrovascular
disease, and peripheral nerve disorders A clinical diagnosis of DM is a costly occurrence in terms of healthcare resource utilization and associated adverse health outcomes of DM However, the nature of the interaction between health services and affected individuals is complex The occurrence of illness may lead to contacts with health services at which DM may be diagnosed
34
I
wwwajmccom
I
JANUARY 2008
Diabetes Diagnosis, Resource Utilization, and Health Outcomes
I Figure 2 Relative Rates of Utilization of Healthcare for Diabetes Mellitus DM Complications Before and After Diagnosis of DM Compared With Nondiabetic Control Subjects by Quarter
Ischemic Heart Disease
5 8
Cerebrovascular Disease
Relative Rate
4 3 2 1
Relative Rate
-2 -1
6 4 2 0
0
1
2
-2
-1
0
1
2
Years
Years
Acute Myocardial Infarction
30 10 8 6 4 2 0 -2 -1 0 1 2 0 -2
Peripheral Nerve Disorders
Relative Rate
20
10
Relative Rate
-1
0
1
2
Years
Data are given as relative rate 95 confidence interval
Years
I Table 1 Relative Rates of Utilization of Healthcare Before, During, and After Diagnosis of Diabetes Mellitus Compared With Nondiabetic Control Subjects
Events From 24 to 6 mo Before
Diagnosis Absolute Rate per Person-Year 519 392 034 026 071 056 282 180 Events From 6 mo Before to 6 mo After Diagnosis Absolute Rate per Person-Year 1155 442 053 028 137 065 419 208 Events From 6 to 24 mo After Diagnosis Absolute Rate per Person-Year 882 479 038 026 193 117 533 240
Condition Primary care consultations
Group Cases Controls
RR 95 CI 132 131-134 — 132 126-139 — 128 123-132 — 157 156-158 –
RR 95 CI 261 258-264 — 192 182-203 — 211 204-218 — 202 200-203 –
RR 95 CI 185 183-187 — 152 144-159 — 166 162-169 — 224 223-225 –
Emergency and hospital care
Cases Controls
Hospital specialist referrals
Cases Controls
Prescription drug items
Cases Controls
RR indicates relative rate; CI, confidence interval
VOL 14, NO 1
I
THE AMERICAN JOURNAL OF MANAGED CARE
I
35
I
CLINICAL
I
I Table 2 New Diagnoses by Broad Groups of Conditions and Periods
Events From 24 to 6 Mo Before Diagnosis No of First Events
100 249 128 145 34 088 61 071 20 052 23 027 50 107 67 070
Events From 6 Mo Before to 6 Mo After Diagnosis No of First Events
137 346 71 081 80 217 49 059 40 108 15 018 61 131 54 057
Events From 6 to 24 Mo After Diagnosis No of First Events
57 154 86
101 30 087 66 081 10 029 28 034 64 140 76 080
Condition
Ischemic heart disease
Group
Cases Controls
RR 95 CI
173 134-225 — 125 082-191 — 195 107-355 — 153 106-221 –
RR 95 CI
444 333-593 — 380 266-543 — 635 349-1155 — 233 161-335 –
RR 95 CI
158 113-221 — 111 072-171 — 087 043-179 — 179 128-250 –
Cerebrovascular disease
Cases Controls
Acute myocardial infarction
Cases Controls
Peripheral nerve disorders
Cases Controls
RR indicates relative rate; CI, confidence interval Denominators for percentages were obtained after excluding subjects previously diagnosed as having the condition Convergence criterion not achieved
Conversely, a diagnosis of DM may lead to the detection of comorbid conditions that were already present In either case, there will be delivery of needed care for treatment of present conditions and prevention of future complications Clinical practice recommendations for good-quality DM care require that utilization of care is increased to facilitate assessment for complications and intervention to control hyperglycemia, hypertension, and lipid disorders Therefore, even when diabetic complications are not present, some increase in utilization of
care around the time of DM diagnosis may be anticipated Furthermore, there may be subgroups of patients in whom the occurrence of complications, or the increase in utilization of care associated with DM diagnosis, is more or less evident
Comparison With Other Studies
RR of myocardial infarction of 375 95 CI, 310-453 compared with subjects who remained free of DM The ageadjusted RR of stroke was 253 95 CI, 194-331 Coronary heart disease risk is increased in subjects with prediabetes16 or with metabolic syndrome17 The present study shows that the time of diagnosis of DM is associated with increased utilization of healthcare for ischemic heart disease, cerebrovascular disease, and peripheral nerve disorders In a previous analysis, an increased incidence of carpal tunnel syndrome was observed in the prediabetic period18 The present analyses evaluated all healthcare utilization associated with peripheral nerve disorders and showed that there was a statistically significant excess of healthcare utilization for peripheral nerve disorders at the time of diagnosis of DM, as well as increased incidence in the prediabetic period
Strengths and Limitations of the Study
There have been few
epidemiological studies of DM diagnosis; most studies regard a diagnosis of DM as an end point or as an entry criterion Type 2 DM is believed to be an evolving process from normal glucose tolerance, first to impaired fasting or nonfasting glucose tolerance, and then to DM, with the degree of hyperglycemia increasing over time11 Epidemiological findings indicate that up to 35 of subjects with newly diagnosed type 2 DM may already show signs of diabetic retinopathy12 This suggests that by the time a diagnosis is made, metabolic abnormalities may have been present for several years13 Some 29 or more of prevalent cases of DM are undiagnosed14 In the Nurses Health Study,15 women who later developed type 2 DM had an age-adjusted
This was a population-based study among a large number of family practices drawn from across the United Kingdom The large sample size yielded precise estimates Cases and controls were matched for age, sex, and family practice, but unmeasured variables such as cigarette smoking, obesity, and individual-level socioeconomic status may have accounted for observed differences Smoking,19,20 obesity,21,22 and lower socioeconomic status23 are associated with an
increased frequency of DM and, in some studies,24,25 with greater utilization of primary care services However, these characteristics may be causally associated with prediabetes and DM and are not true confounders
36
I
wwwajmccom
I
JANUARY 2008
Diabetes Diagnosis, Resource Utilization, and Health Outcomes
Implications for Chronic Illness Care Take-away Points The present results reveal the diagnosis of DM This study aimed to evaluate the ways in which a clinical diagnosis of DM is to be the end of a process as much as a beginning associated with healthcare utilization and the occurrence of clinical complicaIn revealing the morbidity and costs associated with tions of DM clinical diagnosis of DM, the present data add to I Previous studies show that healthcare utilization is increased for several years before DM is recognized clinically the evidence suggesting that strategies for early I The use of primary care, emergency, and hospital resources is greatly detection of DM and prediabetes may be justified increased around the time of DM diagnosis, with a more moderate increase in The American Diabetes Association position statehealthcare utilization sustained over time I Excess
resource utilization at the time of diagnosis is in part associated ment on DM screening acknowledged that there with increased cardiovascular events at the time of clinical DM diagnosis was no direct evidence from randomized controlled trials to support a policy of DM screening but argued that there is sufficient indirect evidence to justify opportunistic screening in a clinical setting of 4 Gulliford MC, Charlton J, Latinovic R Increased utilization of primary care five years before diagnosis of type 2 diabetes: a matched individuals at high risk 3p513 and that clinicians should be vig- cohort study Diabetes Care 2005;28:47-52 ilant in evaluating clinical presentations suggestive of dia- 5 Nichols GA, Glauber HS, Brown JB Type 2 diabetes: incremental medical care costs during the 8 years preceding diagnosis Diabetes betes 3p513 Case finding may be most cost-effective in those Care 2000;23:1654-1659 26 The present data raise a conwho are hypertensive or obese 6 Icks A, Haastert B, Giani G, Rathmann W Incremental prescription the years diabetes diagnosis cern that an earlier diagnosis of DM may extend the period dur- and drug costs during Germany precedingEndocrinol Diabetesin
primary care practices in Exp Clin ing which healthcare costs are incurred However, findings from 2006;114:348-355 7 Hubbard R, Lewis S, a modeling study26 suggest that the costs of screening and early death: a self-controlledWest J, et al Bupropion and the risk of sudden case-series analysis using The Health intervention may be offset by reduced future costs of treating Improvement Network Thorax 2005;60:848-850 8 concurrent prescription DM complications, but this result depends on achieving good ofTata LJ, Fortun PJ, Hubbard RB, et al Does non-steroidal antiselective serotonin reuptake inhibitors and control of intermediate outcomes over time Future research inflammatory drugs substantially increase the risk of upper gastrointestinal bleeding? Aliment Pharmacol Ther 2005;22:175-181 should determine whether the immediate costs associated 9 NHS Information Authority Clinical Terms Read Codes: Sumwith screen-detected DM are less than those following clinical marised Product Description Birmingham, England: NHS Information Authority; 2004 diagnosis Acknowledgment
We thank the staff of EPIC-UK for facilitating access to The Health Improvement Network database Author Affiliations: From
Kings College London, Public Health Sciences, London, UK Author Disclosures: The authors MCG, RL, JC report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article Authorship Information: Concept and design MCG, RL, JC; acquisition of data MCG, RL; analysis and interpretation of data MCG, RL, JC; drafting of the manuscript MCG, JC; critical revision of the manuscript for important intellectual content MCG, JC; statistical analysis MCG, JC; and supervision MCG Address correspondence to: Martin C Gulliford, FFPH, Kings College London, Public Health Sciences, Capital House, 42 Weston St, London, UK, SE1 3QD E-mail: martingulliford@kclacuk
REFERENCES
1 Diabetes Prevention Program Research Group Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin N Engl J Med 2002;346:393-403 2 Gillies CL, Abrams KR, Lambert PC, et al Pharmacological and lifestyle interventions to prevent or delay type 2 diabetes in people with impaired glucose tolerance: systematic review and meta-analysis BMJ 2007;334:e299 3 American Diabetes Association Screening for type 2 diabetes Diabetes Care
2004;27suppl 1:S11-S14
10 StataCorp LP STATA Statistical Software, Release 9 College Station, TX: StataCorp LP; 2006 11 UK Prospective Diabetes Study UKPDS Group Intensive blood glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes UKPDS 33 Lancet 1998;352:837-853 12 Kohner EM, Aldington SJ, Stratton IM, et al United Kingdom Prospective Diabetes Study, 30: diabetic retinopathy at diagnosis of noninsulin-dependent diabetes mellitus and associated risk factors Arch Ophthalmol 1998;116:297-303 13 Harris MI, Klein R, Welborn TA, Knuiman MW Onset of NIDDM occurs at least 4-7 yr before clinical diagnosis Diabetes Care 1992; 15:815-819 14 Centers for Disease Control and Prevention CDC Prevalence of diabetes and impaired fasting glucose in adults: United States, 19992000 MMWR Morb Mortal Wkly Rep 2003;52:833-837 15 Hu FB, Stampfer MJ, Haffner SM, Solomon CG, Willett WC, Manson JE Elevated risk of cardiovascular disease prior to clinical diagnosis of type 2 diabetes Diabetes Care 2002;25:1129-1134 16 Rijkelijkhuizen JM, Nijpels G, Heine RJ, Bouter LM, Stehouwer CDA, Dekker JM High risk of
cardiovascular mortality in individuals with impaired fasting glucose is explained by conversion to diabetes: the Hoorn Study Diabetes Care 2007;30:332-336 17 Lakka HM, Laaksonen DE, Lakka TA, et al The metabolic syndrome and total and cardiovascular disease mortality in middle-aged men JAMA 2002;288:2709-2716 18 Gulliford MC, Latinovic R, Charlton J, Hughes RAC Increased incidence of carpal tunnel syndrome up to 10 years before diabetes diagnosis Diabetes Care 2006;29:1929-1930 19 Rimm EB, Chan J, Stampfer MJ, Colditz GA, Willett WC Prospective study of cigarette smoking, alcohol use, and the risk of diabetes in men BMJ 1999;310:555-559 20 Rimm EB, Manson JE, Stampfer MJ, et al Cigarette smoking and the risk of diabetes in women Am J Public Health 1993;83:211-214
VOL 14, NO 1
I
THE AMERICAN JOURNAL OF MANAGED CARE
I
37
I
CLINICAL
I
21 Perry IJ, Wannamethee SG, Walker MK, Thomson AG, Whincup PH, Shaper AG Prospective study of risk factors for development of noninsulin dependent diabetes in middle aged British men BMJ 1995; 310:560-564 22 McLaren L Socioeconomic status and obesity Epidemiol Rev 2007;29:29-48 , 23 Connolly V, Unwin N, Sherriff P Bilous R, Kelly W Diabetes
prevalence and socioeconomic status: a population based study showing increased prevalence of type 2 diabetes mellitus in deprived areas J Epidemiol Community Health 2000;54:173-177
24 van Doorslaer E, Wagstaff A, van der Burg H, et al Equity in the delivery of health care in Europe and the US J Health Econ 2000;19: 553-583 25 Neubauer S, Welte R, Beiche A, Koenig HH, Buesch K, Leidl R Mortality, morbidity and costs attributable to smoking in Germany: update and a 10-year comparison Tob Control 2006;15:464-471 , 26 Waugh N, Scotland G, McNamee P et al Screening for type 2 diabetes: literature review and economic modelling Health Technol Assess 2007;11:1-144 I
38
I
wwwajmccom
I
JANUARY 2008
Source:rancho.org
