A Review of Systematic Reviews Evaluating Diabetes Interventions

Abstract

Purpose

This article reviews the literature on definitions and issues related to measurement of quality of life in people with diabetes and summarizes reviews of evidence of intervention studies, with a particular focus on interventions targeted for underserved and minority populations.

Methods

An integrative literature review of reviews was conducted on adult diabetes interventions and outcomes. Five electronic databases were searched. Eligible publications were those published between 1999 and 2006 that described outcome measures. Twelve review articles are included.

Results

Review studies were heterogeneous in terms of intervention type, content, participants, setting, and outcome measures. Interventions used variable operational definitions and frequently lacked adequate description; therefore, comparisons of findings proved difficult. A clinical outcome, A1C, was the most frequently assessed, with little inclusion of quality-of-life measures. Several reviews and independent studies did not explicitly consider interventions aimed at the underserved. When quality of life was considered, measures and operational definition of domains were limited.

Conclusions

Understanding the relationship between interventions and resulting outcomes, particularly quality of life, will require attention to operational definitions and better conceptual models. There is an evidence base emerging about important characteristics of effective intervention programs. This evidence base can guide public health and clinical program planners to better understand and make prudent decisions about assessment, planning, implementation, and evaluation of interventions for people with complex chronic illnesses such as diabetes.

As the personal, social, and economic burden of diabetes mellitus (DM) continues,¹ so does the concern for building the evidence for efficacious and effective primary, secondary, and tertiary prevention strategies. It is important to focus on the outcomes of DM intervention efforts and to individualize diabetes self-management education (DSME).^2,3 Community involvement and collaboration can effectively build and deliver culturally centered interventions.^4,5 Documenting the effectiveness of interventions requires explicit indicators specific to DSME interventions,³ along the continuum of outcomes categories: immediate (learning), intermediate (behavior change), post intermediate (clinical improvement), and long term (health status improvement).⁶ Given the important intervention focus on DSME and its related impact on longterm outcomes, the patient’s perceptions and sociocultural context, including quality of life, are also important.

The American Association of Diabetes Educators (AADE) offers the following working definition for diabetes education:

Diabetes education, also known as diabetes self-management training (DSMT), is a collaborative process through which people with or at risk for diabetes gain the knowledge and skills needed to modify behavior and successfully self-manage the disease and its related conditions. The intervention aims to achieve optimal health status, better quality of life and reduce the need for costly healthcare. Diabetes education focuses on seven self-care behaviors that are essential for improved health status and greater quality of life. The AADE 7 Self-Care Behaviors are: healthy eating, being active, monitoring, taking medication, problem solving, healthy coping, and reducing risks.⁷

Currently, the success of diabetes interventions is primarily assessed by traditional clinical measures, such as A1C, blood pressure, and blood glucose levels for a variety of reasons. African Americans are at a greater risk for complications related to DM. This article first reviews the definitions and issues related to measuring quality of life in people with diabetes. It then summarizes the published reviews of intervention studies with respect to outcome measures with a special focus on quality-of-life measures and studies that explicitly consider interventions for African Americans.

Background

Improved outcomes in people with type 2 diabetes are related to several complex factors, which operate at many levels: public policy, health system, organizations, and individual provider and patient. Population subgroups, notably African Americans, face greater risks to health status and quality of life.^8-11 The public health community is paying considerable attention to health disparities and the role of improved health systems in reducing the burden of DM.

Classically defined, health outcomes are the changed state or condition of an individual as a consequence of medical care over time.¹² However, evidence demonstrates that individual and population health status are affected by access to and quality of health care, as well as complex factors at multiple levels, including patient and provider behavior and utilization over time, as noted in Figure 1.^13,14 Outcomes can be traditional health status endpoints, such as mortality or morbidity; clinical measures or biological indicators, such as A1C or blood pressure; predisposing and behaviors such as medication compliance; or psychosocial and patient-centered measures such as satisfaction, well-being, or quality of life.

Figure 1.

Strategies to improve outcomes. DSMT, diabetes self-management training; DM, diabetes mellitus; ADA, American Diabetes Association; AADE, American Association of Diabetes Educators. Adapted from Wagner¹³ and Zapka et al.¹⁴

Definition and Measurement of Quality of Life

Health-related quality of life (HRQOL) is considered a patient-assessed or patient-centered outcome that relates to the individual’s health perceptions, well-being, and functioning.^15-17 Moreover, health perceptions reflect the context of cultural and value systems. Various societal and individual determinants influence physical health, psychological state, social relationships, environmental factors, and beliefs.^18,19 Several studies report lower HRQOL in people diagnosed with diabetes.^20,21 Furthermore, evidence suggests the level of HRQOL is dependent on the presence of comorbidities and the severity of complications and has been significantly correlated with socioeconomic and/or familial barriers.^21,22

In the health literature, the terms quality of life (QOL), health-related quality of life, well-being, health status, and satisfaction are used interchangeably.^15,16Healthy People conceptualized QOL as reflecting a “general sense of happiness and satisfaction with our lives and environment. General quality of life encompasses all aspects of life, including health, recreation, culture, rights, values, beliefs, aspirations, and the conditions that support a life containing these elements.”^23(p10) Overall QOL includes health-related factors and nonmedical phenomena such as personal relationships, employment, spirituality,¹⁵ and a sense of well-being.²⁴ QOL is frequently defined as “an individual’s perception of their position in life in the context of culture and value systems in which they live, and in relation to their goals, expectations, standards, and concerns.”¹⁹

In comparison, “HRQOL reflects a personal sense of physical and mental health and the ability to react to factors in the physical and social environments.”²³ HRQOL is a multidimensional construct of an individual’s subjective appraisal of health and well-being involving physical, psychological, and social functioning.²⁵ HRQOL includes health perception of risk and conditions such as social support, as well as socioeconomic and functional status.¹⁰ The literature, although discordant, concedes that HRQOL is patient centered and is grounded in one’s reality. Because HRQOL is a measure of subjective health status and quality of life, health care providers are encouraged to consider HRQOL as a priority outcome, in addition to traditional physiologic and functional indicators.

Disease-specific instruments are designed to be used in a specific patient group, sensitive to treatment and natural history, and are more responsive to changes in health.^18,26 Important contributions have been made to identify and understand available diabetes-specific measures of HRQOL. Garratt and colleagues¹⁸ reviewed available instruments, summarized the psychometric evidence, and made recommendations for their use. Although it is beyond the scope of this article to review all the findings of Garratt and colleagues’ extensive work, Table 1 reports important dimensions of existing instruments that reflect the personal and social outcomes important to living with diabetes.

Table 1. Instrument Dimensions (Number of Items).

Appraisal of Diabetes Scale	Audit of Diabetes-Dependent Quality of Life	Diabetes Health Profilea	Diabetes Impact Measurement Scales	Diabetes Quality-of-Life Measure	Diabetes-Specific Quality-of-Life Scale	Diabetes-39	Questionnaire on Stress in Patient With Diabetes-R	Well-Being Enquiry for Diabetics
Single  index (7)	Single  index (13)	Psychological  distress (14)	Well-being (11)	Worries about  future effectsof diabetes (4)	Worries about  future (5)	Anxiety and  worry (4)	Depression/fear  of future (6)	Serenity (10)
		Barriers to  activity (13)	Diabetes-related  morale (11)	Worries about  social/vocational  issues (7)	Social  relations (11)	Social and peer  burden(5)	Leisure time (4)	Discomfort (10)
		Disinhibited  eating (5)	Nonspecific  symptoms (11)	Impact of  treatment (20)	Leisure time  flexibility (6)	Sexual  functioning (3)	Partner (6)	Impact (20)
		Specific  symptoms (6)	Satisfaction with  treatment (15)		Daily hassles (4)	Energy and  mobility (15)	Work (6)
					Diet restrictions (5)		Treatment  regimen/diet (9)
					Physical  complaints (8)	Diabetes  control (12)	Physical  complaints (6)
					Treatment  satisfaction (10)		Hypoglycemia (4)  Doctor-patient  relationship (4)

^aThe number of items for the Diabetes Health Profile refers to the DHP-1. The DHP-18 has 6, 7, and 5 items for the psychological distress, barriers to activity, and disinhibited eating dimensions, respectively. Table 2 in Garratt AM, Schmidt L, Fitzpatrick R. Patient-assessed health outcome measures for diabetes: a structured review. Diabetic Med. 2002;19:1-11. Reprinted here with permission.

Methods

To determine the state of evidence of HRQOL as a measured outcome in interventions designed specifically for adults with type 2 diabetes, the following were searched (1996-2006) for systematic or critical reviews and meta-analysis articles: MEDLINE, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Educational Resources Information Center (ERIC), Cochrane Library, and PsycINFO databases. Inclusion criteria included English-language and review articles reported to include patient outcomes published between 1999 and 2006. The subject headings used for search included disease management, case management, care management, community, health system, self-management, self-management education, blacks, African Americans, health-related quality of life, quality of life, intervention, and outcomes, all in association with type 2 diabetes. In addition, selected bibliographic references and diabetes journals of relevance were searched by hand. Twelve review articles met the inclusion criteria and were included in this integrative review.

The analyses included review of criteria for inclusion in the review, the number and design of studies included, intervention description (components, content, and setting), and the impact, process, and outcomes assessed. This study includes comments about QOL instruments and consideration of minority and disadvantaged populations.

Findings

Table 2 reports the key analysis points for each review article, listed in alphabetical order by first author. The reviews covered studies conducted between 1966 and 2006, although the reviews were published between 1999 and 2006.

Table 2. Summary of Review Articles.

Period and Methods	Inclusion and Exclusion Criteria N = Number of Studies	Intervention	Impact, Process, and Outcomes Assessed	Notes/Comments Re: QOL in Minority and Disadvantaged Populations
Eakin EG, Bull SS, Glasgow RE, Mason M. Reaching those most in need: a review of diabetes self-management interventions in disadvantaged populations. Diabetes Metab Res Rev. 2002;18:26-35.
1987-2001: MEDLINE  supplemented by review  of bibliographies from  identified studies and  reviews	Included English-language  RCTs or quasi-  experimental studies with  comparison group; DSME  intervention delivered to  an underserved/  disadvantaged group or  community.	Component DSME, group and individual  sessions, informational  mailings, media  campaign, peer support,  home visits, unstructured  sessions, videotapes,  handouts, monthly calls  from nurse	RE-AIM (Reach, Efficacy,  Adoption, Implementation,  and Maintenance)  framework used to compare and evaluate studies Physiological GHb, cholesterol, BP, weight, FBG, BMI, body composition, glucose tolerance and symptoms	Studies were conducted with  explicit focus on minority  and disadvantaged  populations.
				A summary methodologic  rating of 0 to 9 was  calculated. Scores ranged  from 2 to 8, with 4 studies  receiving a score of 7 or  more.
	Excluded descriptive reports  and studies without  comparison group.  N = 10 studies	Content  DM knowledge, early  detection, exercise,  nutrition, lifestyles,  history and culture (Pima  Indian community),  compliance, access to  primary care  Setting  Delivered in community  centers, community,  hospital clinic, and by  telephone	Behavioral  Diet, physical activity,  smoking, EtOH use  Knowledge  DM self-management  (nutrition, general DM),  changes in medication  regime  Other  Cost to provide intervention  Psychosocial  Depression, social support,  self-efficacy, and QOL  QOL: 3 of 10 studies	Authors emphasize that  DSME interventions for  underserved populations  should explicitly address  social-contextual issues.
				Authors suggest that  interventions designed to  be proactive, such as  telephone follow-up and  behaviorally focused DSME  interventions incorporated  within the primary care  visits, have successful  levels of implementation.
				QOL instruments:  Daniel et al, 1999-not  reported; Glasgow et al,  1992-assessed diabetes-  specific QOL with a  modified DCCT DQOL  scale, no between-group  statistical significance  observed; Weinberger et  al, 1995-measured HRQOL  using the SF-36; statistical  significance not observed.
Ellis SE, Speroff T, Dittus RS, Brown A, Pichert JW, Elasy TA. Diabetes patient education: a meta-analysis and meta-regression. Patient Educ Couns. 2004;52:97-105.
1990-12/2000: MEDLINE,  CINAHL, HealthSTAR,  ERIC, Science Citation  Index, PsyclNFO, CRISP,  and AADE database	Included English-language  RCTs using educational,  including nonpharmacological,  intervention intended to  improve patients’ health  status (physical,  intellectual, and/or  psychosocial),  interventions for adults,  and reporting pre- and  postintervention A1C  values (at least 12 weeks  postintervention).  N = 28 studies	Elasy’s taxonomy used to  categorize educational  interventions and assess  relationship between  specific variables  within the  interventions and  metabolic control.	Physiological  A1C  Psychosocial  HRQOL/QOL: not reported	Authors report only A1C  measures, noting other  important outcomes were  neither uniformly available  nor uniformly measured in  the literature.
		Component  Didactic, negotiated goal  setting; goal setting;  situational problem  solving; cognitive  reframing; and other  unspecified		Authors used meta-regression  to identify “components” of  the educational intervention  that best explained  variance in metabolic  control. Meta-analysis  supports the notion that  patient education improves  glycemic control.
		Content  Diet, exercise, self-  monitoring blood glucose,  basic diabetes  knowledge, medication  adherence, psychosocial,  and other unspecified		The authors examined the  impact of the number of  “episodes” as well as the  duration, and neither of  these dose-related  indicators predicted (was  an indicator) an  intervention’s success or  failure.
		Setting  No reported setting.  Referred solely to the  number of recipients		Authors note that work would  have been enhanced with  the inclusion of behavioral  outcomes.
Gary TL, Genkinger JM, Guallar E, Peyrot M, Brancati FL. Meta-analysis of randomized educational and behavioral interventions in type 2 diabetes. Diabetes Educ. 2003;29:488-501.
1966-1999: MEDLINE;  Cochrane Collaboration  database (1990-1999);  references from experts,  colleagues, previous  meta-analyses, and  review articles	Included published trials  randomized by clinician  and/or patient, sample  size ≥ 10, English  language; educational,  counseling, or behavioral  interventions aimed at  long-term self-care  behavior.	Components  Individual and group  counseling, instruction  packets and audiovisual  materials, telephone  outreach; clinician  prompting, clinician  education, computer  programs	Physiological  FBG, total glycohemoglobin,  hemoglobin A1 or  hemoglobin A1C, body  weight or BMI, blood  pressure, and lipids  Psychosocial  QOL: not reported	Authors assigned and  categorized methodologic  quality scores. Scores  were grouped into low  (<0.65), moderate (0.65-  0.79), and high (≥ 0.80),  with 5 studies receiving a  quality rating of high.
	Excluded published abstracts,  type 1, drug interventions,  and studies evaluating  short-term effects.  N = 18 studies	Content  Patient-diet, exercise,  medication regime  changes or adherence,  BGSM, and foot care		Authors found that compared  with controls, most  intervention groups  produced a decline in  glycohemoglobin.  Furthermore, group and  individual counseling  produced similar effects.
		Providers-methods to  increase patient  involvement, DM  pathophysiology,  complications, DM  education, urine testing,  and treatment/regimen  adherence  Setting  Outpatient clinic (96%)		Authors note that very few  studies reported including  African American or  Hispanic individuals.
				Authors cite few studies that  have evaluated culturally  sensitive interventions for  African Americans and  other ethnic minority  populations, an issue that  should be addressed.
				Authors note lack of QOL  measurement as a  limitation.
Glazier RH, Bajcar J, Kennie NR, Willson K. A systematic review of interventions to improve diabetes care in socially disadvantaged populations. Diabetes Care. 2006;29: 1675-1688.
1986-2004: MEDLINE,  EMBASE, CINAHL, Health  STAR, Cochrane Library,  Sociological Abstracts,  Social Science Citation  Index, and International  Pharmaceutical Abstracts	Included RCT, CT, and  before-and-after studies  with a control group;  studies aimed toward low  SES or ethnic/racial  minority or socially  disadvantaged adults  with type 1 and 2 DM.  Included studies of any  language that measured  self-management,  provider management, or  clinical outcomes.	Elasy’s intervention  taxonomy used to  describe scope and  components of the  interventions.	Physiologic  FBG.A1C, BP, BMI, lipids,  mortality, DM  complications  Patient behavior  Glucose monitoring, diet and  exercise, medication  adherence and self-  adjustment, scheduling  and/or attending  scheduled medical  appointments  Provider behavior  Management: diagnostic  testing, prescribing,  referrals, educational and  behavioral counseling  Psychosocial  QOL: not reported	Studies focused on those  groups with low SES or  belonging to an ethno-  racial minority. To evaluate  methodological quality, the  authors used an evaluation  method to identify specific  intervention features that  are associated with  successful or unsuccessful  outcomes: (1) target  of intervention,  (2) intervention design,  (3) setting, (4) delivery, and  (5) intensity and duration.
		Components  Individualized assessment  and goal setting,  individual educational  counseling, DSME,  reminder cards,  videotapes, support  groups, case management,  evidenced-based  guidelines, community  health workers/peer  educators, group visits,  treatment algorithms
		Excluded: age-specific,  gestational DM studies;  studies that only reported  hospital process of care  measures; and those not  specifying which socially  disadvantaged group.  N = 11 studies
		Content  DM knowledge, self-care,  symptoms and treatment,  exercise, nutrition, foot  care, lifestyle, monitoring;  clinician-focused  evaluation and  management of glycemic  control, associated  comorbidities and  complications		Authors suggest that cultural  tailoring, use of community  educators or lay educators,  one-on-one interventions,  individualized  assessment/reassessment,  use of treatment  algorithms, behavioral  interventions, and patient  feedback are all consistent  with positive rate  differences found in  multiple studies.
		Setting  Primary care practice sites,  hospitals, community-  based clinics, community  facilities (centers,  churches), and telephone		High-intensity (>10 contacts)  and delivery over a long  time frame (≥ 6 months)  were also associated with  success intervention in  disadvantaged groups.
Loveman E, Royle P, Waugh N. Specialist nurses in diabetes mellitus. Cochrane Database Syst Rev. 2003;(2):CD003286.
1966-11/2002: MEDLINE,  Cochrane library,  EMBASE, CINAHL, British  Nursing Index, Royal  College of Nursing Index,  HealthSTAR, BIOSIS,  PsyclNFO, Science  Citation Index, Social  Sciences Citation Index;  hand-searched relevant  journals and conference  proceedings (1990-2001),  reference lists, National  Research register, Early  Warning system and  Current Controlled Trials  registries	Included randomized  controlled and controlled  clinical trials designed to  test effects of DNS/NCM  interventions, trial  duration 6 months.	Components  DNS plus routine care  versus routine care,  automated calls with  structured messages,  treatment regime  alterations led by  DNS/NCM, treatment  recommendations to PCP  by DNS/NCM, and care  coordination within  primary care system  Content  Goal setting and self-care  Setting  Hospital, outpatient clinic,  primary care system, and  community	Physiological  All trials used A1C as  endpoint; short-term  complications  (hypoglycemic episodes,  hyperglycemic incidents)	Methodologic quality of  studies was assessed  using factors included in  Schulz and Jadad’s quality  criteria: minimization of  selection, attrition, and  detection bias. Studies  were then subdivided in  the following categories:  low, moderate, and high  risk of bias.
	Excluded studies with no  control group, group  education, no  prespecified outcomes,  other team members  other than DNS/NCM  involved delivering  intervention, nurses not  able to adjust treatments.  N = 6 studies		Other  Emergency room visits,  hospitalizations  Psychosocial  Quality of life  QOL: 1 of 6 studies with no  data presented
			Other  Prespecified outcomes  established for this  review but not included  and/or reported in trials:  long-term complications  (retinopathy, neuropathy,  and nephropathy),  mortality, BMI, costs, and  adverse effects	Authors noted that the quality  of trials was mostly not  good, making it difficult to  assess the implications for  practice. Future research  suggestions include  international observational  studies to identify roles  and time allocations;  qualitative studies and  RCTs of specialist nurse  intervention.
Norris SL, Engelgau MM, Venkat Narayan KM. Effectiveness of self-management training in type 2 diabetes: a systematic review of randomized controlled trials. Diabetes Care. 2001;24:561-587.
1/1980-12/1999: MEDLINE,  ERIC, CINAHL, restricted  to RCTs, hand-searched  several relevant journals;  abstracts and  dissertations were  excluded	Included English-language  RCTs examining the  effectiveness of self-  management training, all  or most subjects with  type 2 DM, age >18,  multicomponent  interventions included (if  educational component  evaluated separately).  Excluded children.  N = 72 studies	Components  Chart reminders, group  and individual sessions,  written information,  videos, didactic  education, individual  sessions based on  patient’s priorities, home  visits, computer  knowledge assessment  program, feedback,  behavior modification,  contracts,  demonstrations,  food logs, nutrition  goals, interactive  computer, culturally  appropriate flashcards,  pharmacist, nursing  students, dietitian , lay  health worker,  nurse-led sessions,  psychologist-led  group sessions,  patient-led education,  and self-study  course	Physiological  Weight, lipids, blood  pressure, glycemic control  Behavioral  Dietary, physical activity,  self-care skills  Knowledge  Diabetes  Other  Economic measures; health  care utilization  Psychosocial  Attitudes, problem solving,  anxiety levels, quality of  life  QOL: yes, in 5 studies  Kaplan 1987-increase in  QOL at 18 months for  lifestyle intervention  (intensive counseling on  diet + physical activity).  Gilden 1992-increase in  QOL at 2 years follow-up  for coping skills  intervention (6 weekly  sessions +18 monthly  support group sessions).	Authors qualitatively  summarized outcomes to  generate hypotheses,  categorize variables for  quantitative syntheses, and  illustrate the vast  heterogeneity of methods  and outcomes in the  included literature.
	Notably, the authors  emphasize limited  literature measuring QOL  and long-term clinical  outcomes. Authors  advocate for a more  holistic view of patients,  including that QOL  outcomes take precedence  in future research and  more diverse study  populations.
		Content	Three studies (deWeerdt,  Glasgow, Trento)  considered “brief  interventions” with no  significant difference  between intervention and  control groups.
		Physical activity, diet,  nutrition, SMBG, weight  loss, barriers, social  support, foot care,  general DM knowledge,  self-adjustment of insulin,  goal setting, problem  solving, stress  management, patient  empowerment, self-  control, self-management
		Setting  Not specifically identified in  the review; described as  heterogeneous
Norris SL, Lau J, Smith SJ, Schmid CH, Engelgau MM. Self-management education for adults with type 2 diabetes: a meta-analysis of the effect on glycemic control. Diabetes Care. 2002;25:1159-1171.
1980-1999: MEDLINE, ERIC,  CINAHL; manual search  of relevant journals, and  experts were consulted  for citations	Included English-language  RCTs, DSME  interventions, and other  interventions  when delivered in  combination with  DSME (if effect of DSME  could be examined  separately).	Components  Didactic or collaborative  DSME, individual and  group education, support  groups, home visits,  dietician, flashcards by  lay health workers, meal  demonstration, feedback,  telephone follow-up,  psychologist-led group  sessions, weight loss  program, empowerment  techniques, computer-  assisted knowledge  assessment and  instruction	Physiological  GHb, HbA1, HbA1c  Psychosocial  QOL: not identified as a key  outcome of this review	Meta-regression indicates no  significant interactions  except total contact. In 15  studies, GHb  measurements were  reduced for every hour of  additional contact, which  approximates 23.6 hours  of contact between an  educator and patient to  achieve a 1% reduction;  brief intervention appears  to be less effective.
	Excluded abstracts and  dissertations.  N = 31 studies
		Content  Diet, physical activity, SMBG  or urine glucose, foot  care, coping, self-efficacy,  identifying and preventing  complications, goal  setting, and modeling  Setting  Clinic, home, and senior  center		Authors suggest that  psychosocial mediators,  cultural relevancy, and  health care system  structure and primary care  linkage may account for  the heterogeneity in  outcomes.
				Authors stress that further  research to identify  predicators and correlates  needs to focus on  psychosocial attributes,  social support, and  problem-solving skills.
Norris SL, Nichols PJ, Caspersen CJ, et al. The effectiveness of disease and case management for people with diabetes: a systematic review. Am J Prev Med. 2002; 22(suppl 4):15-38.
1966-12/2000: MEDLINE,  ERIC, CINAHL,  HealthSTAR, Chronic  Disease Prevention  database (health  promotion and education  subfile), Combined Health  Information Database,  diabetes, health  promotion and education  subfile), journals hand-  searched, reference list of  included articles and  consultation with team  experts for relevant  citations	Included published  comparative study  designs, English  language, conducted in  established market  economies as defined by  the World Bank, studies  with primary investigation  of disease (as defined by  review team) and case  management  intervention, reported  information on 1 or more  outcomes of interest  preselected by review  team.	Components  Disease or case  management along with  DSME, telemedicine  support, insulin  adjustment algorithms,  group support, visit  reminders, hospital  discharge assessment  and follow-up  Content  Not applicable or not  specified  Setting  Managed care organizations  and community clinics	Physiological  A1C, weight, BMI, BP, and  lipids  Behavioral  SMBG, patient health care  utilization; provider  screening, monitoring and  treatment): A1C, lipids,  dilated eye exams, foot  exams, proteinuria  Other  Health care system: health  insurance, provision of  services, health care  utilization (admissions,  outpatient visits, length of  stay), public health  services, economic  outcomes  Psychosocial  Self-efficacy, patient  satisfaction, quality of life	Authors note research gaps  in the areas of intervention  effectiveness on long-term  health and QOL outcomes,  as well as diverse  populations and settings.
				QOL Instrument: Peters et al  (1998) used the SF-36  with 2 (unspecified)  diabetes- specific  questions added.
	Excluded: studies  characterized as limited  quality based on the  number of threats to  validity, dissertations, and  abstracts.  N = 42 studies		QOL: Assessed in 2 of 42  studies (1 in adults and 1  in children); adult cohort  study reported relative  change of +4.7%  between groups.
Renders CM, Valk GD, Griffin SJ, Wagner EH, Eijk Van JT, Assendelft WJ. Interventions to improve the management of diabetes in primary care, outpatient, and community settings: a systematic review. Diabetes Care. 2001;24:1821-1833.
1966-2000: MEDLINE,  EMBASE, CINAHL, EPOC,  and Cochrane clinical trial  registries (1999);  reference list of selected  articles was reviewed	Included studies of  effectiveness of  interventions to improve  process of care or patient  outcomes among type 1  or type 2 that had (1)  randomized or quasi-  experimental trials, (2)  interrupted time series  with defined interventions  and a minimum of 3  before-and-after time  points, (3) nonrandomized  studies with a second  controlled site, and (4)  predetermined measures  of patient outcomes or  the process of health  care.	Components  Professional (education,  reminders, audit, and  feedback); organizational  (role revision, changes in  medical record systems;  patient education,  learner-centered  counseling, telephone  follow-up for missed  appointments); financial  (fee for services and  grants) or multistrategy  Content  Provider-guidelines  Patient-problem solving and  decision making  Setting  Primary care, outpatient,  and community	Physiological  Glycemic control, BP,  cholesterol, BMI, weight,  microvascular and/or  macrovascular  complications, albumin,  creatinine  Behavioral  Patient attendance, provider  process measures:  glycemic control, BP,  weight, microvascular  complications,  cholesterol, visits,  education, health survey,  compliance care  providers, albumin, urine  protein, creatinine,  hospitalizations	The authors note that  complex professional  interventions improved the  process of care, but  patient outcomes were  rarely assessed, making it  less clear to evaluate the  impact. Furthermore, the  authors emphasize that  measuring both process  and patient outcomes lead  to better understanding of  how to improve quality of  care.
	Excluded study interventions  classified as only patient  oriented.  N = 41 studies		Other  Hospitalizations, health care  system process  measures: glycemic  control, BP, weight,  cholesterol, visits,  education, microvascular  complications  Psychosocial  Well-being, quality of life  QOL: yes, but limited 3 of 41  (2 QOL and 1 well-being)  studies. QOL outcomes  showed no effect in 2  professional and  organizational  interventions; well-being  reported as positive effect  in 1 education  intervention.
Sarkisian CA, Brown AF, Norris KC, Wintz RL, Mangione CM. A systematic review of diabetes self-care interventions for older, African American, or Latino adults. Diabetes Educ. 2003;29:467-479.
01/1995-12/2000:  MEDLINE, HealthSTAR,  EMBASE, PsyclNFO,  Ageline, and Sociological  abstracts	Included English-language  published intervention  studies that described  intervention aimed at  changing knowledge,  beliefs or behavior;  targeted 1 or more of the  following: age >55 years,  African American or Latino  adults with DM; and  measure at least 1 of the  preidentified outcomes:  glycemic control, diabete-  srelated symptoms, or  quality of life.  N = 12 studies	Components  Group education plus  exercise class; group  sessions led by physician,  nurse, or dietitian;  didactic DM education  alone or in combination  with support groups,  individual counseling,  1-to-1 DM education,  bicultural community  health worker  liaison, nutritionist,  weekly pharmacist  appointments, grocery  store tour, follow-up  phone calls, physician  education	Physiological  Glycemic control and  diabetes-related  symptoms (established as  an outcome of interest by  review authors but not  assessed in any of the 12  included studies)  Psychosocial  Quality of life  HRQOL/QOL  In 5 of 12 studies: 4 of 8  RCTs and 1 of 4 pre/post  designs; 1 RCT reported  statistically significant  improved QOL in the  intervention arm (78  points vs 71 points);	Authors were unable to  summarize the effect of  self-care interventions on  quality of life because of  the heterogeneity of QOL  instruments.
				The authors suggest that  age- and culture-specific  interventions be designed  for clinical trials.
		Content  Exercise, nutrition, goal  setting, “standard DM  education curriculum,”  others unspecified/unclear  Setting  Noted as urban or rural	1 pre/post design  reported improved mean  scores on the QOL  instrument, with no mean  change in A1C at end of  intervention.
Steed L, Cooke D, Newman S. A systematic review of psychosocial outcomes following education, self-management and psychological interventions in diabetes mellitus. Patient Educ Couns. 2003;51:5-15.
1980-2001: EMBASE,  MEDLINE, Psyclit; hand  search of reference list of  reviews and retrieved  papers	Included pre/post or  controlled trial design  studies published in  English-language peer-  reviewed journals, adults  with type 1 or 2 DM,  provision of DMSE or  psychological intervention  and assessed either QOL  or psychological well-  being.	Components  Classified as education, self-  management, or  psychological biofeedback.  Content  General education, diet,  meal planning, goal  setting, reinforcement,  modeling, reward  systems, problem solving,  exercise sessions,  behavior change  maintenance, relaxation,  communication skills, and  social support  Professional or layperson-  led discussion groups  Setting  Not reported	Psychosocial  Overall psychological well-  being, depression,  anxiety, or emotional  adjustment and QOL  HRQOUQOL  Ten studies used a generic  measure, 9 used a  diabetes-specific  measure, 1 used both,  and 5 studies used an  overall psychological  well-being measure.	This review did not speak to  issues related to minority  or disadvantaged groups.  The authors were unable  to examine the efficacy of  different intervention  components; noted as the  original aim of the review,  multiple components in  single studies and the  overlap between  components in different  intervention categories  hindered these attempts.  Self-management  interventions had  beneficial impact on QOL.  None of the included  studies using psychological
	Excluded interventions that  provided only diet or  exercise or intensive  insulin regimens.  N = 36 studies		Generic QOL measures: 2 of  9 studies showed  improvement relative to  control; DM-specific QOL  measures: 6 of 9 found  improvement.  Overall psychological well-  being: 1 of 5 reported  significant improvemen.  Global and disease-specific  QOL were not assessed in  any of the psychological  interventions.
				Authors suggest that future  research should use  diabetes-specific studies  measures within larger  controlled trials.
				QOL Instrument: SF-36 and  DQOL were the most  frequently used generic  and diabetes-specific  instruments, respectively.
van Dam HA, van der Horst F, van den Borne B, Ryckman R, Crebolder H. Provider-patient interaction in diabetes care: Effects on patient self-care and outcomes: a systematic review. Patient Educ Couns. 2003;51:17-28.
1980-10/2001: MEDLINE  Advanced, EMBASE,  PsycIit/PsycINFO, and The  Cochrane Library	Included English-language  RCTs and quasi-  experimental studies  testing the effects of  modification of patient-  provider interaction and  provider consulting style  on DM self-care and  outcomes.	Component  Provider training,  questionnaire prompts,  feedback reports; patient-  automated telephone  calls, nurse telephone  feedback, individual  education, empowerment  group ed. sessions, group  consultations with doctor,  and individual  consultation with  standard DM education	Physiological  BP, lipids, BMI, GlyHb,  general health,  mortality  Behavioral  Health behavior  Knowledge  DM knowledge  Psychosocial  Satisfaction, well-being,  mental health, self-  efficacy, DM-specific QOL  scores	Methodological quality was  assessed using a modified  version of the Van Tulder  criteria list; the authors  include a table of the  criteria. Nineteen items  (questions) are categorized  into the following  categories: patient  selection, interventions,  outcome measurements,  and statistics. The mean  quality score was 17.3 out  of 19 total points.
	Excluded non-type 2 DM  patients, case, case  control, and  nonexperimental studies.  N = 8 studies
		Content  Provider-behavioral change,  communications, patient-  centered consulting,  guidelines; patient-diet,  self-care, problem  solving/barriers, decision  making  Setting  General practice or hospital  outpatient	Patient: communication,  interaction, lifestyle,  perception of care,  functional, attitudes  Provider: beliefs, attitudes,  and behavior change;  adherence to guidelines,  time spent with patient	Authors emphasize a shift  from the traditional  medical model and toward  a patient empowerment  model with shared  decision making.
			HRQOUQOL  In 2 of 8 studies:  1 with no change in HRQOL  scores and 1 with  “better” QOL.	Authors note they found no  evidence that the 3 most  promising studies have  ever been replicated  among diverse patient  populations or different  care settings, thus causing  alarm.

RCT, randomized controlled trial; DSME, diabetes self-management education; DM, diabetes mellitus; BP, blood pressure; FBG, fasting blood glucose; BMI, body mass index; QOL, quality of life; HRQOL, health-related quality of life; DQOL, diabetes quality of life; DCCT, Diabetes Control and Complications Trial Research Group; SES, socioeconomic status; PCP, primary care provider; CT, controlled trial; BGSM, blood glucose self-monitoring; DNS/NCM, diabetes nurse specialist/nurse case manager; SMBG, self-monitoring of blood glucose.

Inclusion and Exclusion Criteria for the Included Review Articles

Variable criteria were considered within each individual review. For example, some included only randomized trials (RCTs),^27-32 whereas others included RCT, quasiexperimental, and/or other comparative design methods.^33-37 Exclusion criteria also varied considerably across reviews. For example, some studies were excluded because they evaluated short-term effects³² or included only patient interventions.³⁴

Intervention Type and Content

Numerous intervention approaches in a variety of settings were reported^33,38 and included delivery by lay leaders and professionals from diverse health care disciplines.^30,31 Many interventions included multiple components or strategies. Intervention content was loosely defined and, again, was broad and varied within each review.²⁹ Content ranged from being unspecified to diet, nutrition, goal setting, problem solving, or simply described as “standard diabetes education curriculum.”

Terms such as disease management and case management are widely used, and their operational definitions are exceptionally broad and variable.^35,37 The goal of disease management is improved short- and long-term health and/or economic outcomes. In particular, disease management interventions have been promoted, especially for costly, chronic conditions such as diabetes mellitus. For the included studies, settings were urban centers of managed care organizations and outpatient and community health clinics. A range of organizational, professional, and patient intervention components and strategies was included and supported multisystem approaches, which could lead to improvements in patient outcomes as well as process of care.³⁵

Case management, one component of a multisystem collaborative approach, delivered in conjunction with disease management, can serve an important role for people at high risk for adverse outcomes and excess health care utilization. Furthermore, assessment, planning, linking, monitoring, advocacy, and outreach, described as essential core functions of case management, are complementary with the chronic care model (CCM).³⁹ Yet, multicomponent approaches incorporating both disease and case management have been shown to improve outcomes as well as the process of care.^30,34,35 The evidence suggests that DSME was routinely delivered across disease management and case management interventions. Moreover, integrating DSME across interventions could potentially help to costeffectively optimize outcomes, such as metabolic control and quality of life, and prevent complications.^40,41

Impact, Process, and Outcomes Assessed

Seven of the review articles included in this review primarily focused on DSME interventions,^{27-29,31-33,36} specifically, the effect of DSME interventions on multiple physiological, psychological, behavioral, and knowledge processes and outcomes^27,28,33 and, to a lesser extent, HRQOL.^29,36 Notably, Steed and colleagues³⁶ completed a systematic review of psychological outcomes (eg, psychological well-being and QOL) in 36 studies. They found that self-management interventions had more positive impact than educational interventions where patients received information only. Nine of the 36 studies reviewed by Steed et al used diabetes-specific QOL instruments; of these, 5 used the diabetes quality of life (DQOL).⁴² Six of the 9 trials found QOL improvement postintervention, with no significant difference reported in 3 of the studies. Overall, 6 trials found QOL improvement postintervention. However, no significant difference was reported in 3 of the 5 studies using the DQOL.

Outcome measures varied substantially across studies, making it difficult to compare results across interventions, and many studies lacked sufficient detail for replication. Even when glycemic control was assessed, a variety of different methods and reference values was used.^32,34 Although outcomes for diabetes education research have historically focused on knowledge and glycemic control (A1C),^43,44 researchers and clinicians generally agree that DSME should be integrated across multiple intervention strategies and encompass behavioral approaches to improve self-management outcomes.⁴⁴ Also, several RCTs have demonstrated the value of blood pressure and other outcome variables. In addition, the outcomes of disease management, case management, and/or multisystem approaches to diabetes interventions are more often physiologic measures, as evidenced in Table 2.

The reviews identified few studies focusing on African Americans, particularly older African Americans. For example, in a review of self-care interventions designed to improve glycemic control or QOL in older, African American, or Latino adults, only 1 study among African Americans (n = 39) examined QOL.²⁹

Discussion and Recommendations

Several key observations emerged from this review of reviews. Statements about the “state of sciences” are enormously challenging as the reviews demonstrate great heterogeneity in terms of intervention type, content, participants, and outcome measures. The language used to describe interventions varies, and diverse disciplines were involved in the delivery of sometimes complex interventions. The limited description of intervention and content included in these approaches limits the possibility of replication and necessitates cautious comparison between studies. Although a diabetes intervention taxonomy⁴⁵ is suggested and sometimes used^28,46 to systematically describe the components of interventions, intervention heterogeneity remains commonplace in the current literature, making it difficult to compare results. Thus, researchers and educators should seriously consider opportunities to uniformly categorize interventions, at least for descriptive purposes.

With respect to the operational definition of outcome categories,⁶ clearly the most frequently used clinical improvement (postintermediate) outcome measured was A1C, and the least measured were health status improvement and QOL (long-term) outcome measures.^10,16,47,48 Review articles reported limited information about instruments used in the included studies. Where available, identified QOL instruments have been included in Table 2. Overall, the most frequently cited instruments were the SF-36⁴⁹ and DQOL⁵⁰ to measure general and diabetes-specific QOL, respectively. Moreover, when considering key outcomes for studies, in addition to including physiological outcomes, additional patientreported outcomes should be considered. Addressing the importance of patients’ perceptions of the effects of chronic illness and its treatment has the potential to significantly improve overall DM management.⁵¹ Noted researchers³² emphasized the exclusion of outcomes such as QOL as a limitation. As telling, having looked at only A1C, other review authors²⁸ suggest their work would have been enhanced with the inclusion of behavioral outcomes. Given the complexity of factors at many levels, which can affect processes, impact, and outcomes of care, investigators and clinicians should mindfully seek clarity in operational definitions and in better understanding the relationships of the various factors.

Interventions tailored toward high-risk populations, such as African Americans or the aged, or those with limited literacy skills have been emphasized, yet cultural relevance, literacy, and age-related concerns of interventions have been neglected in the current literature. Although several review authors^31-33,46 suggest sociocultural context as an important indicator, only 3 of 12 review articles were solely dedicated to such areas. For example, Sarkisian et al²⁹ was the only review specifically focused on older high-risk racial and ethnic groups, whereas Eakin et al³³ and Glazier et al⁴⁶ defined their populations of interest as low socioeconomic status (SES) or socially disadvantaged ethnic and racial adults.

The broader behavioral literature regarding tailoring of interventions must describe the lessons learned. Complete and accurate descriptions of interventions will encourage further analysis and determine interventions’ effectiveness, structure, process, and outcomes. Although attempts have been made to characterize interventions^28,33,46 and employ meta-analysis,^28,31,32 incomplete study descriptions and variations in content and components continue to hamper comparability. There is an urgent need to provide and document well-designed studies, systematically describing components. More important, presenting adequate information for assessments regarding study quality and broader comparability can advance the science and care of diabetes.

There are several limitations to this review of reviews. The authors were able to evaluate only what was reported, not what may in fact have been done. Together with the methodological strengths and weaknesses of the included studies, limitations of the reviews must also be considered. Finally, the search strategy and inclusion criteria for the current review may be perceived as too narrow, given that only published studies in English-language journals were selected. It is acknowledged that journal space limits what may be included in articles. Unfortunately, descriptions of intervention research fail to provide evidence of culturally tailored interventions. Also, literacy, which is known to affect care and outcomes, has not been adequately addressed. Very few studies included and/or focused on African Americans; thus, enormous gaps related to racial and ethnic groups in the literature remain.

In conclusion, this review of reviews of intervention studies shows that effects on long-term health and quality-of-life outcomes have not been adequately assessed in the current literature.^30,34,35 Previous reviews^26,52 provide evidence that patient-assessed health status differs from those proxy assessments made by health professionals. Although several important conceptual frameworks^13,53-55 and proponents have confirmed the relevance of HRQOL as an outcome capable of affecting the effectiveness of interventions and utilization behavior, the available evidence fails to actualize HRQOL as a priority targeted outcome. Research is needed to assess long-term outcomes, such as HRQOL and its impact on specific components of interventions, particularly among diverse populations. Other researchers^10,16,47,48 suggest that the initial priority should be on evaluating and refining existing HRQOL instruments. Our review confirms minimal to no testing in certain high-risk groups (Table 2). The need for disease-specific instruments, specifically validated in African Americans, becomes particularly evident in this context. Both qualitative and quantitative methods can and should be used to assess and refine existing instruments, as well as in using population-specific validated instruments to measure outcomes of clearly articulated and described interventions.