N A L A R T I C L E

Deviation From DevelopmentallyAppropriate Self "Care AutonomyAssociation with diabetes outcomes

TIM WYSOCKI, PHD

ALEXANDRA TAYLOR, BA

BRENDA S. HOUGH, PHD

THOMAS R. LINSCHEID, PHD

KEITH O. YEATES, PHD

JACK A. NAGLIERI, PHD

OBJECTIVE — Treatment of IDDM in youth emphasizes balancing children's self-care au-tonomy with their psychological maturity. However, few data exist to guide clinicians or parents,and little is known about correlates of deviations from this ideal.

RESEARCH DESIGN AND METHODS— In this cross-sectional study, IDDM self-care autonomy of 100 youth was assessed using two well-validated measures. Three measures ofpsychological maturity (cognitive function, social-cognitive development, and academicachievement) were also collected for each child. Composite indexes of self-care autonomy and ofpsychological maturity were formed, and the ratio of the self-care autonomy index to the psy-chological maturity index quantified each child's deviation from developmentally appropriateIDDM self-care autonomy. Based on these scores, participants were categorized as exhibitingconstrained (lower tertile), appropriate (middle tertile), or excessive (higher tertile) self-careautonomy. Between-group differences in treatment adherence, diabetes knowledge, glycemiccontrol, and hospitalization rates were explored.

RESULTS — Analysis of covariance controlling for age revealed that the excessive self-careautonomy group demonstrated less favorable treatment adherence, diabetes knowledge, hospi-talization rates, and, marginally, glycemic control. Excessive self-care autonomy increased withage and was less common among intact two-parent families but was unrelated to other demo-graphic factors.

CONCLUSIONS — The findings indicate caution about encouragement of maximal self-care autonomy among youth with IDDM and suggest that families who succeed in maintainingparental involvement in diabetes management may have better outcomes.

Treatment of youths with IDDM en-tails a complex medical and lifestyleregimen that is coordinated by

health professionals but implemented,monitored, evaluated, and often adjustedlargely by patients and their parents. Therecent Diabetes Control and Complica-tions Trial (DCCT) proved that mainte-nance of near-normoglycemia prevents

the development and slows the progres-sion of long-term complications of IDDMsuch as retinopathy, nephropathy, andneuropathy (1,2). Emphasis on aggressiveself-management of IDDM by patientsand families will likely increase in the af-termath of those findings. Hence, infor-mation about factors affecting the efficacyof diabetes self-management skills will in-

From the Nemours Children's Clinic (T.W., AT.), Jacksonville, Florida; Michigan State University (B.S.H.),Fast Lansing, Michigan; Ohio State University (JAN., B.S.H.) and Children's Hospital (T.R.L., K.O.Y.),Columbus, Ohio.

Address correspondence and reprint requests to Tim Wysocki, PhD, Nemours Children's Clinic, 807 NiraSt., Jacksonville, FL 32207.

Received for publication 30 May 1995 and accepted in revised form 28 September 1995.AMR, autonomy-to-maturity ratio; ANCOVA, analysis of covariance; CAS, Das-Naglieri Cognitive As-

sessment System; DCCT, Diabetes Control and Complications Trial; DFRQ, Diabetes Family ResponsibilityQuestionnaire; DIS, Diabetes Independence Survey; DISC, Diabetes Information Survey for Children; INS,Interpersonal Negotiation Strategies Interview; PASS, Planning, Attention, Simultaneous, Successive; SCI,Self Clare Inventory; SES, socioeconomic status; WRAT-R, Wide Range Achievement Test-Revised.

crease in importance as IDDM therapyevolves in response to the DCCT findings.

Current therapy for childhoodIDDM encourages developmentally ap-propriate self-care autonomy, but thereare few data available to guide parents orhealth professionals. Several authors haveargued that expectations for self-care au-tonomy exceeding the child's cognitive,affective, or behavioral capabilities maycompromise adherence and diabetic con-trol (3-5). Others have hypothesized thatconstraining the otherwise capable childfrom assuming developmentally appropri-ate treatment responsibility may convey alack of confidence in the child, discour-aging future initiative and responsibility(6).

Several authors have offered rec-ommendations for age-appropriate self-care autonomy, but there is disagreementamong them, and no recommendationshave been validated empirically (7-12).Others have shown that diabetes knowl-edge and skill levels increase with age, butthese studies have not yielded clinicallypractical guidelines (13-18). These stud-ies revealed correlations between age andspecific IDDM knowledge or skills rang-ing from 0.45 to 0.75, suggesting that ageis an imperfect marker of IDDM self-careautonomy and that other dimensions ofpsychological maturity require consider-ation in clinical and educational manage-ment of childhood diabetes.

Previous studies by Wysocki andcolleagues included surveys of 229 healthcare professionals (19) and 490 parents ofyouth with IDDM (20) concerning age-related changes in children's master)' of38 IDDM self-care skills. Comparison ofthe results obtained with parental andprofessional versions of the Diabetes In-dependence Survey (DIS) (20) revealedthat parents and professionals agreedclosely about the sequence in which spe-cific skills were mastered (Spearman's p : :

0.74), but there was substantial disagree-ment about the ages at which individualskills were mastered. Parents rated youngchildren as more skilled and adolescents

DIABETES CARP., VOLUME 19, NUMBER 2, FEBRUARY 1996 119

IDDM self-care autonomy

Table 1—Characteristics of the enrolled sample

Age (years)Duration of IDDM (years)Age at diagnosis (years)GHb(%)Hollingshead Four Factor Index of SESSex (%)

FemaleMale

Family structure (%)Living with both biological parentsLiving with one parent and one step-parentLiving with single parentLiving with other relatives

Race/ethnicityWhiteAfrican-AmericanOther/biracial

10012.3 ±3 .15.2 ±3 .17.1 ±2 .9

11.1 ±2 .141.1 ± 12.7

5941

6425

65

86131

Data are means ± 1 SD.

as less skilled than did the health profes-sionals.

This study extends that work byexamining the correlates of deviations be-tween children's IDDM self-care auton-omy and objective estimates of their psy-chological maturity. Two measures ofIDDM self-care autonomy were com-bined into a composite index of that con-struct. Similarly, three sound measures ofintellectual, social-cognitive, and aca-demic maturity were entered into a com-posite index of psychological maturity foreach youth. The ratio of the IDDM self-care autonomy index to the psychologicalmaturity index (the autonomy-to-maturity ratio [AMR]) quantified eachchild's deviation from developmentallyappropriate IDDM responsibility. Thisstudy examined the associations betweenthe AMR and treatment adherence, diabe-tes knowledge, glycemic control, and fre-quency of diabetes-related hospitaliza-tions among youths with IDDM.

RESEARCH DESIGN ANDMETHODS

ParticipantsA sample of 100 children and adolescentsand 1 parent of each child participated inthe study at three pediatric medical cen-ters. Enrollment criteria required that thechild had IDDM for at least 1 year, was atleast 5 but not yet 18 years old at enroll-ment, had no other chronic diseases, wasnot thought to be mentally retarded by

the physician or parent, and had visionand hearing within normal limits.

Recruitment began by mailing anintroductory letter signed by the physi-cian and an investigator and a copy of aninformed consent form to the parents ofpotential participants. The letter was fol-lowed by a telephone call from the re-search assistant about 2 weeks later. Anenrollment rate of ~30% was achieved.At two of the participating centers, 23parents of children who did not agree toparticipate in the study provided demo-graphic information for the purpose ofevaluating the representativeness of theenrolled sample. Characteristics of the en-rolled participants are summarized in Ta-ble 1. Although representativeness of thesample cannot be assured, the distribu-tions of age, socioeconomic status (SES),race/ethnicity, and previous glycohemo-globin levels were similar to those of theclinic populations at the participatingsites and of participants in this group'sprevious research studies.

SettingParticipants were enrolled at one of threepediatric clinics: Nemours Children'sClinic in Jacksonville, FL (n = 66); Chil-dren's Hospital, Columbus, OH (n = 21);and Michigan State University, East Lan-sing, MI (n = 13). All patients receivedcare for IDDM under the direction of sub-specialty-certified pediatric endocrinolo-gists. Details of the treatment regimensdiffered across centers, but all regimens

included two or more daily insulin injec-tions, self-monitoring of blood glucose, aconstant carbohydrate, low-fat diet, em-phasis on self-management education ofpatients and families, and minimizationof hypoglycemia, prolonged hyperglyce-mia, and ketonuria. Information collectedfor research purposes beyond that re-ported for this study indicated that thestudy sample had experienced severe hy-poglycemic episodes less frequently (5.4episodes per 100 patient-years) than wasreported for adolescents in the DCCTconventional therapy group (2). The sam-ple also demonstrated a normal distribu-tion of scores on a measure of childhoodbehavioral problems (21).

Procedure and measuresEach child and parent completed a singleevaluation session conducted by a re-search assistant with graduate training inclinical psychology, experience in psy-chological testing, and competence in ad-ministration of the study measures. Eval-uations began with obtaining informedconsent from parents and assent from pa-tients. Participants had been instructed tobring snacks and blood glucose testingequipment. No youths complained ofsymptoms of hypoglycemia during theevaluations. Parents were instructed incompleting certain of the questionnairesdescribed below. Patients were then takento a private examination room where theresearch assistant administered othermeasures that are described below. Eval-uation sessions lasted about 3.5 h. Theorder of administration of the variousmeasures was counterbalanced acrossparticipants.

Demographic informationParents completed the General Informa-tion Form, which was used for collectionof demographic information, pertinentmedical and school histories, and calcu-lation of the Hollingshead Four Factor In-dex of Social Status (A.B. Hollingshead,unpublished observations), a commonmeasure of SES.

Measures of IDDM self-careautonomyTwo measures were collected for assess-ment of each child's degree of diabetesself-care autonomy. These measures pro-vided, respectively, assessments of par-ent-child sharing of diabetes responsibil-ities and of parents' perceptions of their

120 DIABETES CARE, VOLUME 19, NUMBER 2, FEBRUARY 1996

Wysocki and Associates

children's IDDM self-management capa-bilities. The two measures correlated 0.82with each other, and it was felt that thecombined measures would provide amore reliable index of IDDM self-care au-tonomy than would either instrumentalone.

The Diabetes Family Responsibil-ity Questionnaire (DFRQ) measures fam-ily division of responsibility for 17 aspectsof the IDDM regimen (23). Parents rateeach treatment task on a three-point Lik-crt scale as a parent, child, or shared re-sponsibility. Higher scores indicategreater child responsibility. Internal con-sistency was 0.85 in a study of 121 fami-lies, and total scores correlated 0.74 withthe child's age.

The DIS was developed by the firstauthor for a multicenter survey of 490parents of youth with IDDM (20). Parentswere asked to indicate whether their childhad achieved mastery of each of 38 diabe-tes skills and concepts, with mastery de-fined as the capacity to display the skillwithout parental assistance, prompting,or supervision. Total scores on the instru-ment correlated 0.71 with age, and arank-order correlation of 0.74 was ob-tained with a parallel form completed byhealth professionals. Scores were not cor-related significantly with duration ofIDDM.

Diabetes Self-Care Autonomy IndexFor each patient, a composite index ofIDDM self-care autonomy was calculatedby first summing raw scores from the DIS(20) and DFRQ (23) and transformingthese sums into ̂ -scores. A constant of 3.0was added to ^-score-transformed rawscores before statistical analyses to avoiddata analysis problems entailed in thetreatment of negative values.

Measures of psychological maturityThree measures of children's intellectual,social-cognitive, and academic maturitywere administered.

A prepublication version of a newtest of cognitive processing, the Das-Naglieri Cognitive Assessment System(CAS), was used to measure intelligenceas defined by the Planning, Attention, Si-multaneous, Successive (PASS) theory(24,25). The PASS theory is based primar-ily on the neuropsychological work ofA.R. Luria (26), which was used as a frame-work to develop and organize the 16 cogni-tive subtests into 4 factor-analytically de-

rived scales. This test battery, which iscurrently being standardized nationallywith a large representative sample of chil-dren and adolescents, has been shown toprovide reliable and valid measures ofplanning, attention, and simultaneousand successive processing (24,25). TheCAS was selected over other intelligencetests because it is unique in providingmeasures of planning and attention, and ithas been developed for administration toyouths throughout the 5- to 18-year agerange targeted in this study. It has been indevelopment for more than 9 years and issupported by a large body of preliminaryresearch (24).

The Interpersonal NegotiationStrategies Interview (INS) is a structuredinterview that assesses social-cognitivedevelopment in children older than age 4(27,28). Psychometric properties of theINS have been well-documented (27,28).Children were presented verbally with aseries of eight age-adjusted social dilem-mas and guided by the interviewer to ar-ticulate the problem, specify why it is aconflict, identify possible solutions, andproject the consequences of each solu-tion. With consultation by one of the au-thors of this method (K.O.Y.), we con-structed age-adjusted items consisting ofIDDM-specific social dilemmas. Each pa-tient was then interviewed using four gen-eral and four IDDM-specific social dilem-mas. Interviews were audiotaped for laterscoring, which yielded measures for eachchild for problem identification, genera-tion of solutions, evaluation of solutions,and selection of a solution, as well as com-posite indexes of social-cognitive matu-rity for general, IDDM-specific, and all so-cial dilemmas. The composite measure ofsocial-cognitive maturity was correlated0.65 with age. Since total scores that werederived from the general and IDDM-specific social dilemmas correlated 0.92,combined scores were used for all dataanalyses.

The Wide Range AchievementTest-Revised (WRAT-R) was used to as-sess children's academic maturity. TheWRAT-R is a widely used screening test ofacademic achievement that can be admin-istered to children as young as 5 years ofage (29). Its advantages for the presentstudy were its brief administration timeand recent restandardization. The test hassound psychometric properties.

Psychological maturity indexAs above, a composite index of psycho-logical maturity was derived for each pa-tient by transforming the sums of rawscores for the CAS (24), INS (27), andWRAT-R (29) into z-scores. The threecomponent measures were all signifi-cantly positively correlated, with Pearsonr values ranging from 0.38 for the corre-lation between the INS and WRAT-R to0.64 for the correlation between the CASand WRAT-R. As above, a constant of 3.0was added to c-score-transformed rawscores before statistical analyses to avoiddata analysis problems entailed in thetreatment of negative values.

AMRThe ratio of the Self-Care Autonomy In-dex to the Psychological Maturity Indexprovided a measure of the extent to whicheach child exhibited developmentally ap-propriate self-care autonomy relative toobjective assessments of that child's psy-chological maturity. Scores on this mea-sure were divided into tertiles and pa-tients were assigned to three groups asfollows: constrained (scores s=0.c)5; n : •33); appropriate (scores from 0.% to1.22; n = 34); and excessive (scores>1.23;n = 33).

Outcome measuresInformation was collected regarding dia-betes outcomes in terms of treatment ad-herence, diabetes knowledge, recent gly-cerine control, and frequency of diabetes-related hospitalizations for each child.

The Self Care Inventory (SCI) is a14-item Likert-type scale used for the as-sessment of adherence with the IDDMregimen (30). It has adequate internalconsistency and test-retest reliability, andit is well correlated with analogous mea-sures derived from Johnson's 24-h recallinterview method (31).

The Diabetes Information Surveyfor Children (DISC) is a 98-item struc-tured interview administered in the sameformat for all ages (32). It yields a reliablemeasure of diabetes knowledge that hasbeen validated with children as young as5 years of age. The total score from theinstrument correlates 0.76 with age, andit is not significantly correlated with du-ration of IDDM.

Results of the most recent Glibtests completed within the prior 6 monthswere retrieved from each child's medicalchart as an index of average glycemic con-

DIABETES CARE, VOLUME 19, NUMBER 2, FEBRUARY 1996 121

IDDM self-care autonomy

trol over the preceding 2-3 months. Sincethe three centers used different methodsand laboratories for these assays, the re-sults for patients enrolled from each cen-ter were transformed into z-scores relativeto that center's distribution of GHb re-sults. The z-scores were used for all statis-tical analyses pertaining to this measure.

Data on each child's frequency ofdiabetes-related hospitalizations, exclud-ing the hospitalization at the time of diag-nosis, were obtained by parental report.Parents were asked to report the totalnumber of IDDM-related hospitalizationssince the child's diagnosis. Medical chartreviews were impractical because of thelarge number of hospitals involved, and itwas felt that parents would have accuraterecall of their children's hospitalizationsbecause of their salience and low fre-quency.

RESULTS

SamplingParticipants did not differ from the sampleof 23 families who declined enrollment interms of age, race, duration of IDDM, SES,family size or composition, or history ofspecial education placement. Girls wereover-represented among participants(59%). There were no differences amongthe three centers with respect to the abovedemographic factors, hospitalizationrates, GHb, or any of the psychometricmeasures collected for the study.

Statistical analysisMean ages of the three groups in yearswere as follows: constrained, 11.5; appro-priate, 12.4; and excessive, 13.1. Age cor-related 0.29 (P < 0.001) with the AMR.The groups did not differ with respect toage at diagnosis, duration of IDDM, sexdistribution, SES, or family size. AMRscores for youths from intact, two-parentfamilies were significantly lower thanthose from youths from alternative familystructures [F (3,96) = 5.87; P < 0.001].Previous studies have revealed significantassociations between age and IDDM out-comes such as those measured in thisstudy. Consequently, multivariate analy-sis of covariance (ANCOVA), with age asthe covariate, was used to examine be-tween-group differences in GHb concen-trations, treatment adherence as mea-sured by scores on the SCI, diabetesknowledge as measured by the DISC, andfrequency of diabetes-related hospitaliza-

110

Figure 1—Mean (+ 1 SE) scores on the SCI forthe three groups. (Higher scores indicate bettertreatment adherence.)

tions. A significant multivariate group ef-fect was obtained [F (3,96) = 2.98, P <0.004, and Wilks' A = 0.78] and followedby univariate ANCOVA for each of theoutcome measures. The covariate effectwas significant for the analysis of groupdifferences in diabetes knowledge butnonsignificant for the other analyses.

Figure 1 presents the mean scoresfor each group on the SCI (29). The valuesobtained were as follows: constrained,51.6; appropriate, 51.9; and excessive,45.6. These values indicate poorer treat-ment adherence with increasing IDDMself-care autonomy relative to psycholog-ical maturity. ANCOVA yielded a signifi-cant main effect for groups, with F (3,96)= 4.47 and P < 0.01. Effect size estimatedby R2 indicated that group membershipaccounted for 14% of the variance intreatment adherence. Post hoc analysisusing the Tukey test indicated that scoresfor the excessive group differed signifi-cantly from those for the appropriate andconstrained groups, but the latter groupsdid not differ significantly from one an-other.

Figure 2 presents each group'smean scores on the DISC (31). ANCOVAcontrolling for age yielded F (3,96) =4.56 and P < 0.01. Post hoc analysis us-ing the Tukey test revealed that scores forthe appropriate group were significantlyhigher than were those for either of theother two groups.

Mean GHb levels are shown in

90

80

Constrained f j ] Appropriate 0 Excessive

Figure 2—Mean (+ 1 SE) scores on the DISCfor the three groups. (Higher scores indicate morediabetes knowledge.)

Fig. 3. The main effect for groupsachieved a marginal level of statistical sig-nificance with F (3,95) = 2.48 and P =0.06.

Figure 4 displays the mean num-ber of IDDM-related hospitalizations per100 patients per year since IDDM onset,excluding the hospitalization at the timeof diagnosis, as reported by parents. Themean numbers of hospitalizations per100 patients per year since diagnosis ofIDDM were as follows: constrained, 0.7;appropriate, 1.9; and excessive, 3.5. AN-COVA conducted as above yielded a sig-

io 11

| Constrained Q Appropriate £ 2 Excessive

Figure 3—Mean (+ 1 SE) GHb concentrationsfor the three groups.

122 DIABETES CARE, VOLUME 19, NUMBER 2, FEBRUARY 1996

Wysochi and Associates

Constrained [_ ] Appropriate [//] Excessive

Figure 4—Mean number of hospitalizations (+I SH) per 100 patients per year for the threegroups.

nificant between-group effect on these re-ported hospitalization rates with F (3,96):^ 5.06 and P < 0.008 and on number ofhospitalizations with F (3,96) = 3.50 andP < 0.03. Effect size estimated by R2 in-dicated that group membership ac-counted for 13% of the variance in hospi-talizations per year and 17% of thevariance in number of hospitalizations.The mean number of hospitalizations re-ported for each group were as follows:constrained, 0.45; appropriate, 1.41; andexcessive, 2.09. Median values for thethree groups were as follows: constrained,0; appropriate, 0; and excessive, 2. Onepatient in the constrained group and threepatients in the appropriate group hadmore than two hospitalizations, whilenine patients in the excessive group hadmore than two admissions. Post hoc anal-ysis using the Tukey test indicated that allthree groups differed significantly fromone another.

CONCLUSIONS— The present re-sults indicate that excessive IDDM self-care autonomy, as indexed by the AMRcalculated in this study, was associatedconsistently with adverse outcomes interms of treatment adherence, diabetesknowledge, and hospitalizations and asimilar, albeit marginal, relationship withdiabetic control. In contrast, constrainedIDDM self-care autonomy was associatedwith more favorable outcomes in terms oftreatment adherence and diabetic controlcompared with that reported for children

with both developmentally appropriateand excessive levels of self-care auton-omy.

Children in the appropriate self-care autonomy group demonstratedgreater diabetes knowledge than did thosein the other two groups. This finding maybe open to several interpretations. Lesser di-abetes knowledge is perhaps to be expectedamong youths in the constrained groupwho were younger and whose parents weremore involved in IDDM management. Lessintuitively, youths in the excessive groupexperienced less parental involvement indiabetes care and, consequently, fewer op-portunities for parental feedback and re-finement of their knowledge of IDDM. It isalso possible that assumption of develop-mentally appropriate levels of diabetes re-sponsibility may facilitate children's mas-tery of knowledge and skills related to thedisease. The apparently weaker diabetesknowledge of children in the constrainedgroup may be of little clinical consequencesince parents of these children are relativelymore involved in their diabetes manage-ment. But, the lower knowledge demon-strated by the excessive group is more om-inous clinically since it could predisposethose children to dangerous treatment er-rors, possibly culminating in the poor dia-betic control and increased frequency ofhospitalizations documented among thosechildren in this investigation.

This study also provides impor-tant methodological contributions. Ourresults substantiate the validity of themethodology introduced here for quanti-fying the degree of children's deviationfrom developmentally appropriate levelsof self-care autonomy. These methodsmay be applicable to addressing otherempirical questions about the determi-nants and the consequences of imbal-ances of IDDM self-care autonomy andpsychological maturity.

Our findings lend support to thespeculations of other authors who haveidentified possible deleterious effects ofdevelopmentally excessive self-care au-tonomy among youth with IDDM (3-5,8,32-35). The present results generallyfail to support the speculation that paren-tal involvement in diabetes managementmight entail adverse consequences (6),since only one outcome measure (diabe-tes knowledge) favored the appropriategroup over the constrained group.

The positive association obtainedbetween age and the AMR implies that

older children and adolescents are progressively more likely to exhibit excessivelevels of self-care autonomy relative totheir psychological maturity. Similarfindings have been reported in severalprevious studies, lngersoll et al. O^>) re-ported that parents tend to relinquishhigher-level diabetes self-management responsibilities to adolescents indepen-dently of the youths' cognitive maturity.Allen et al. (34) and La Greca et al. (3'"i)observed that among older children,greater levels of child responsibility forIDDM management were associated withpoorer diabetic control. Wysocki et al.(20) found that mastery of several types ofIDDM self-management skills appearedto reach an asymptote during early tomiddle adolescence. These includedIDDM tasks that were more eognitivelycomplex (e.g., basic knowledge of insulinpharmacology), skills for which nonad••herence typically yields delayed or uncer-tain negative consequences (e.g., adjustment of dietary intake in response toblood glucose fluctuation), and skills thatare infrequently invoked for most patientsand families (e.g., urine ketone testing).The present data bolster the implicationof this collection of studies that familieswho succeed in maintaining parental in-volvement in IDDM management duringadolescence are likely to achieve better di-abetes outcomes. This study provides noinformation about the precise nature ofthat involvement, the characteristics offamilies who demonstrate it, or the role ofclinical and educational practices in itscultivation.

The interpretation of these resultsand their clinical implications must bedone with consideration of the cross-sectional nature of the present study. Thedirection of causality, if any, between theAMR and diabetes outcomes cannot bespecified by this study, and the results donot necessarily implicate inadequate pa-renting or misguided clinical manage-ment as causal agents. It is entirely plau-sible, for instance, that youths with poortreatment adherence, deficient diabetesknowledge, chronic hyperglycemia, andfrequent hospitalizations actively dis-courage adult involvement in their diabe-tes management. Alternatively, parentsand clinicians whose diabetes manage-ment efforts have met with failure mayresign themselves to accept suboptimalself-care by adolescents. The associationbetween the AMR and hospitalization

DIABETES CARE, VOLUME 19, NUMBER 2, FEBRUARY 1996 123

IDDM self-care autonomy

rates may have been inflated because pa-rental reports of hospitalizations mayhave been inaccurate, and no attempt wasmade to distinguish between recent andpast hospitalizations.

Numerous reports have impli-cated family conflict in general (36-39)and parent-adolescent conflict and com-munication problems in particular (40-43) as correlates of ineffective diabetesmanagement. Although the available re-search directs attention to these interac-tion patterns as possible mediators of ad-verse IDDM outcomes, this link can onlybe proven by a well-conceived longitudi-nal investigation. Similarly, the presentstudy provides no information about thelong-term sequelae of developmentallyconstrained or excessive IDDM self-careautonomy. Although little or no adverseeffects of constrained diabetes responsi-bility were evident and excessive self-careautonomy was associated consistentlywith unacceptable outcomes, different ef-fects of these patterns might be evidentover the long term. Again, only a longitu-dinal study could clarify these issues.

Acknowledgments— This research was sup-ported in part by grants from the Central OhioDiabetes Association, the Nemours Founda-tion (20-8839), and grant RO1-DK-43802from the National Institutes of Health to thefirst author. Riverside Publishing Corporationgraciously provided standardization versionsof the Das-Naglieri Cognitive Assessment Sys-tem tor use in the study.

The authors wish to thank Alan Stephens,P.lizabeth Kirsch, and Lisa Gauthier, who re-cruited and tested patients for the study.

References

1. Diabetes Control and Complications TrialResearch Group: Diabetes Control andComplications Trial: the effect of intensivetreatment of diabetes on the developmentand progression of long-term complica-tions in insulin-dependent diabetes mel-litus. N EnglJMed 329:977-986, 1993

2. Diabetes Control and Complications TrialResearch Group: Effect of intensive treat-ment on the development and progres-sion oflong term complications in adoles-cents with insulin-dependent diabetesmellitus: Diabetes Control and Complica-tions Trial. J Pediatr 125:177-188, 1994

3. Johnson SB: Diabetes mellitus in child-hood. In Handbook ofPediatric Psychology.Routh D, Ed. New York, Guilford, 1988,

p. 9-314. Follansbee DM:. Assuming responsibility

for diabetes care: what age, what price?Diabetes Educ 15:347-352, 1989

5. Johnson SB: Knowledge, attitudes and be-havior: correlates of health in childhooddiabetes. Clin Psychol Rev 4:503-524,1984

6. Swift CF, Seidman F, Stein N: Adjustmentproblems in juvenile diabetes. PsychosomMed 29:555-571, 1967

7. American Diabetes Association: Curricu-lum for Youth Education. Alexandria, VA,American Diabetes Association, 1983

8. Cerreto MC, Travis LB: Implications ofpsychosocial and family factors in thetreatment of diabetes. Pediatr Clin N Am31:689-710, 1984

9. Baker L, Lyen K: Laying the foundationfor juvenile diabetes management. DrugTher 10:63-73, 1981

10. Kohler E, Hurwitz LS, Milan D: A devel-opmentally staged curriculum for teach-ing self-care to the child with insulin-dependent diabetes mellitus. DiabetesCare 5:300-304, 1982

11. Krauser K, Madden P: The child with di-abetes mellitus. Nurs Clin N Am 18:749-762,1983

12. Sargent J, Baker L: Behavior and diabetescare. Primary Care 10:583-595, 1983

13. Gilbert BO, Johnson SB, Spillar R, McCal-lum M, Silverstein JH, Rosenbloom A:The effects of a peer-modeling film onchildren learning to self-inject insulin. Be-hav Ther 13:186-193, 1982

14. Johnson SB, Pollak T, Silverstein JH,Rosenbloom A, Spillar R, McCallum M,Harkavy J: Cognitive and behavioralknowledge about insulin-dependent dia-betes mellitus among children and theirparents. Pediatrics 69:708-724, 1983

15. Harkavy J, Johnson SB, Silverstein JH,Spillar R, McCallum M, Rosenbloom A:Who learns what at diabetes summercamp?J Pediatr Psychol 8:143-153, 1983

16. Eastman B, Johnson SB, Silverstein JH,Spillar R, McCallum M, Rosenbloom A:Understanding of hypo- and hyperglyce-mia by youngsters with diabetes and theirparents.; Pediatr Psychol 8:229-243, 1983

17. Drotar D, levers C: Age differences in par-ent and child responsibilities for manage-ment of cystic fibrosis and insulin-dependent diabetes mellitus. J Dev BehavPediatr 15:265-272, 1994

18. Ruggiero L, Mindell JA, Kairys S: A newmeasure for assessing diabetes self-management responsibility in pediatricpopulations: the Diabetes Regimen Re-sponsibility Scale. Behav Ther 14:233-234,1991

19. Wysocki T, Meinhold PA, Cox DJ, ClarkeWL: Survey of diabetes professionals re-garding developmental changes in diabe-tes self care. Diabetes Care 13:65-68, 1990

20. Wysocki T, Meinhold PA, Abrams K, Bar-nard MU, Clarke WL, Bellando BJ, Bour-geois MJ: Parental and professional esti-mates of self care independence ofchildren and adolescents with IDDM. Di-abetes Care 15:43-52, 1992

21. Conners KC: Conners Parent Rating Scale.North Tonawanda, NY, Multi-Health Sys-tems, 1989

23. Anderson BJ, Auslander WF, Jung KC,Miller JP, Santiago JV: Assessing familysharing of diabetes responsibilities. J Pedi-atr Psychol 15:477-492,1990

24. Das JP, Naglieri J, Kirby JR: Assessment ofCognitive Processes: The PASS Theory of In-telligence. Needham Heights, MA, Long-wood Division, Allyn and Bacon, 1994

25. Das JP, Naglieri J: Das-Naglieri CognitiveAssessment System Experimental Test Bat-tery. Chicago, Riverside, 1993

26. Luria AR: The Working Brain: An Introduc-tion to Neuropsychology. New York, BasicBooks, 1973

27. Yeates KO, Schultz LH, Selman RL: Bridg-ing the gaps in child clinical assessment:toward the application of social-cognitivedevelopmental theory. Clin Psychol Rev10:567-588, 1990

28. Schultz LH, Yeates KO, Selman RL: TheInterpersonal Negotiation Strategies Inter-view: A Scoring Manual. Cambridge, MA,Harvard Graduate School of Education,Group for the Study of Interpersonal De-velopment, 1989

29. Jastak S, Wilkinson G: Wide RangeAchievement Test-Revised. Wilmington,DE, Jastak, 1991

30. Greco P, LaGreca AM, Ireland SJ, Wick P,Freeman C, Gutt M, Agramonte R, SkylerJS: Assessing adherence in IDDM: a com-parison of two methods (Abstract). Diabe-tes 39 (Suppl. 1):165A, 1990

31. Johnson SB, Silverstein JH, RosenbloomA, Carter R, Cunningham W: Assessingdaily management in childhood diabetes.Health Psychol 5:545-564, 1986

32. Wysocki T, Bellando BJ, Netherton S: Psy-chometric properties of the Diabetes In-formation Survey for Children. In Proc ofthe Society of Behavioral Medicine, Chicago,1987

33. Ingersoll G, Orr DP, Herrold AJ, GoldenMP: Cognitive maturity and self manage-ment among adolescents with insulin-dependent diabetes mellitus. J Pediatr108:620-623, 1986

34. Allen DA, Tennen H, McGrade BJ, AffleckG, Ratzan S: Parent and child perceptionsof the management of juvenile diabetes. JPediatr Psychol 8:129-141, 1983

35. LaGreca AM, Follansbee DM, Skyler JS:Developmental and behavioral aspects ofdiabetes management in youngsters. ChildHealth Care 19:132-139, 1990

36. Anderson BJ, Miller JP, Auslander WF,Santiago JV: Family characteristics of dia-

DlABETES CAR!!, VOLUME 1 9 , Nl'MBtiR 2 , FEBRUARY 1 9 9 6124

Wysocki and Associates

betic adolescents: relationship to meta-bolic control. Diabetes Care 4:586-594,1981 39.

57. Hauser ST,Jacobson AM, Lavori P, Wolfs-dorfjl, Herskowitz RD, MilleyJE, Bliss R,Wertleib D, Stein J: Adherence amongchildren and adolescents with insulin-dependent diabetes mellitus over a four 40.year longitudinal follow-up: immediateand long term linkages with the familymilieu. J Pediatr Psychol 15:527-542,1990

58. Marteau TM, Bloch S, Baum JD: Family 41.

life and diabetic control. J Child PsycholPsychiatry 28:823-833, 1987White K, Kolman ML, Wexler P, Polin G,Winter RJ: Unstable diabetes and unstablefamilies: a psychosocial evaluation of dia-betic children with recurrent diabetic ke-toacidosis. Pediatrics 73:749-755, 1984Rubin RR, Young-Hyman D, Peyrot M:Parent-child responsibility and conflict indiabetes care (Abstract). Diabetes 38(Suppl. 2):7A, 1988Bobrow ES, AvRuskin TW, Siller J: Moth-

er-daughter interaction and adherence todiabetes regimens. Diabetes Cure 8:14b •151,1985

42. Wysocki T, \ lough BS, Ward KM, Allen A,Murgai N: Use of blood glucose daia byfamilies of children and adolescents w ith1DDM. Diabetes Care 15:1041-1044, 1<W.

43. Wysocki T: Associations among parentadolescent relationships, metabolic control and adjustment to diabetes in ado-lescents. J Pediatr Psvehol 18:44 5- -4H,1993

DIABETE-S CARE, VOLUME 19, NUMBER 2, FEBRUARY 1996 125