Introduction
	
	The Maternal PKU Resource Mothers Program: A Clinical Trial  (R40 MC 00162 01) was recommended in 1997 by the Bureau of Maternal and Child Health Research Branch for grant support for 4 years, but was not funded.  In the June 1999, review cycle, it was awarded one year of "special funding". This application is a revision of the original proposal and is for the 3 remaining years.  The previous application has been modified to reflect findings of the Maternal PKU Collaborative Study (MPKUCS) and to address the concerns raised in the earlier review.
	Data from the MPKUCS indicated that the majority of women with phenylketonuria (PKU), a genetic disorder affecting metabolism of the amino acid phenylalanine, continue to contact their metabolic clinics after they become pregnant, when damage to the fetus most likely has already occurred (Koch et al, 1994). The MPKUCS also found, however, that a stimulating home environment during the first year of the offspringıs life may attenuate the adverse effects of maternal PKU (Waisbren et al, 1999). Moreover, it is clear from the MPKUCS and other data (Koch et al, 1994; Lenke and Levy, 1980) that planning of pregnancy with metabolic control achieved prior to conception confers the lowest risk to the fetus. Thus, a home visitation program only during pregnancy, the original plan for the Resource Mothers Program, is unlikely to be optimally effective. Accordingly, the proposed project now includes outreach before pregnancy to all adolescent girls and young women tracked by the metabolic centers and to all families after a baby is born.  Only those randomized to the treatment group will receive the support of Resource Mothers during pregnancy. The research question is whether the Resource Mothers program yields benefits beyond what is provided by basic metabolic services, as well as outreach before and after pregnancy. 

Responses to the review from 1997 (Changes to # MCR-978-37 are highlighted in italics in the main text):

1. There was a question about the validity of food diaries. To ensure detailed food intake information from the PKU mothers, food records (as outlined in the proposal) and a checklist for those women who have difficulty maintaining food records will be used.  The checklist will include an additional query to determine if the young woman ate other foods not allowed on the diet. 

 2. Will provision of emergency low protein foods constitute a separate intervention? Emergency foods will be offered since this is already part of standard treatment for maternal PKU.  In most states, low protein foods are part of the mandated insurance coverage or are provided by the State Department of Public Health. Emergency foods are required until usual payment methods are in place. The real issue is whether the women make use of the low protein foods.  They often try these foods once. If the recipe does not turn out well (as is often initially the case), they do not try again. The intervention is designed to encourage use of these foods.

3. Concern was raised about the time commitment of Dianne Sullivan, the trainer/advisor, who will work 40% time.  Her role is to provide on-going training and advice over the phone to all the Resource Mothers regarding low protein cooking and other diet issues. As the one "consumer" on the core staff, she provides an important perspective to the research team. She will now also serve as the Resource Mother for the Boston Clinic and follow an estimated 5 pregnancies/year. 

4. Clarification was requested about the data checklist. A checklist will be used to determine the extent to which data have been collected. For example, when protocols for the psychosocial tests are received, these data points will be marked off on the checklist. 

5. Will practicing psychologists be familiar with the HOME Scale? This scale has 45 items that simply need to be checked off. Instructions will be included in the data collection packet sent to the psychologists conducting the evaluations.  In the MPKUCS this scale had an independent relationship to offspring outcome beyond maternal metabolic control during pregnancy.

 6. Regarding a budget item for nutritionists: Nutritional assessments are basic to maternal PKU treatment. Nutritionists will be provided with assessment forms that contain only information that is obtained in routine clinical practice.  It is not expected that the nutritionists will be spending extra time with the young women due to their participation in the research and, therefore, no additional funds were requested.   

7. Regarding quality control at distant sites: Fran Rohr, BS, RN, will be responsible for quality control of nutrition data.  Twice during each pregnancy, original food records will be submitted to her for re-analysis of nutrient intake.  Comparisons between the independent analyses will be made and discrepancies will be resolved.

8. To address concerns about confounding the intervention, the role of Isabel Bailey, the social worker, has been redefined as working with the Coordinators in a "train the trainers" type situation. In addition, she will be responsible for the outreach and follow-up components of the study.

9. The question arose about including young women under the age of 18 years. Based on actual data from the MPKUCS, only 2% of the women are expected to be less than 21 years of age. This will not appreciably bias the study

10. If demographic variables are not comparable between the groups, descriptive analyses will be conducted.  If appropriate, analyses of co-variance will be used.

11. In the hierarchical regression analyses, the order of entry for the independent variable and the adherence variable will not rule out other causative factors. This point will be addressed in the interpretation of the results.

Status of Current Project:  
Thirteen Resource Mothers and 13 Study Coordinators participated in an intensive weekend training program in May 1998. The National Institutes of Child Health and Development supported this aspect of the project in conjunction with the Maternal PKU Collaborative Study. The Resource Mothers had a total of 28 children, 17 of whom had PKU. Three of the Resource Mothers described themselves as homemakers.  Those who held jobs included 4 who worked in health care, a lawyer, travel agent, day care operator, administrator and interior designer.  The Study Coordinators and the trained Resource Mothers remain available to the study. 
The data entry system has been installed, with a special server located at the Childrenıs Hospital in Boston.  Now, each of the Metabolic Centers can enter data in a confidential manner.  All data can be read directly by the Coordinating Center, while each participating Metabolic Center has access only to its own data. A decision was made to include Childrenıs Hospital Metabolic Clinic patients in the study to enhance recruitment.  In the 14 weeks since the special funding from MCHB was made available (October, 1999), 6 women with PKU have been enrolled in the study.  Three were randomly assigned to the group with the services of a Resource Mother and 3 were assigned to the control group, without a Resource Mother.  Three were enrolled prior to pregnancy and 3 contacted the Metabolic Clinic after pregnancy began.  These numbers confirm that the randomization process should result in equal numbers in each group. One pregnancy resulted in a miscarriage at 14 weeks gestation, which approximates the expected 30% attrition rate.   No potential subjects declined participation. Although fairly equally divided regarding which group they preferred, all of the women accepted the random assignment. 
	A potential problem in the study design is the reliance on multiple centers to obtain an adequate sample size. In collaborative studies there is always the risk that data will be incomplete.  The study includes funds to train and compensate Coordinators at each center along with support for a solid research team in Boston to oversee and guide them.
a. Statement of the Problem.
Phenylketonuria (PKU) is a genetic disorder in which there is an absence or inactivity of the liver enzyme, phenylalanine hydroxylase, which is responsible for the conversion of phenylalanine to tyrosine (Hsia, 1970).  Since the advent of newborn screening in the 1960's, infants with PKU have been identified and treated with a special formula and low protein diet that protects them from the most severe effects of elevated phenylalanine levels (MacCready, 1974). Maternal PKU refers to the risks to the fetus associated with phenylalanine teratogenicity (Levy and Ghavami, 1996).  If the mother has classic PKU, risks to the developing fetus include mental retardation (95%), microcephaly (90%), and congenital heart disease (17%) (Lenke and Levy, 1980).  Strict metabolic control prior to and throughout pregnancy greatly reduces these risks (Drogari et al, 1987). The risk of low birth weight, developmental delay, congenital heart disease and other congenital malformations increases with each week that treatment and metabolic control are delayed (Rouse et al, 2000). Women with non-PKU mild hyperphenylalaninemia (MHP), whose natural blood phenylalanine levels are lower than in women with PKU, incur little or no risk for adverse pregnancy outcomes (Levy, Waisbren 1983; Levy et al 1994). This finding underscores the need for metabolic control in maternal PKU.  The cost of caring for a single child with maternal PKU syndrome includes cardiac surgery, early intervention, physical therapy, speech therapy, occupational therapy, special education classes, repeated psychological evaluations and psychotherapy for many years. With adequate prevention,  these costs are eliminated. 
Given the known benefits of dietary therapy, delivery of treatment for maternal PKU should be a straightforward process.  However, this is not the case. The diet, consisting of the special formula (containing necessary nutrients in protein excluding phenylalanine), measured amounts of fruits and vegetables, and special low protein pastas and breads is extremely restrictive.  Meat, fish, beans, nuts, dairy products, soy, and regular grains are not allowed.  Preparing the low protein products is laborious and takes practice.  There is little variety in the diet and without concerted attention to intake, the young woman may not consume enough calories for her pregnancy needs (Duran et al, 1999).  Most reported maternal PKU pregnancies have been late treated (Koch, 1994; Hanley et al 1987; Levy, Ghavami, 1996). Metabolic control during pregnancy is strongly dependent upon regulated intake of the full allotment of special formula, which many find to be disagreeable in taste and odor.  Compliance is generally poor with consequent adverse outcome (Duran et al, 1999). The majority of young women with PKU terminated the diet during middle childhood, as had once been recommended (Schuett, 1984). They have difficulties resuming the diet. Many of the women have limited intellectual abilities; the mean IQ is less than 85 (Waisbren et al, 1995). The incidence of emotional disorders is high (Waisbren, Zaff, 1994) and socioeconomic status is low (Waisbren et al, 2000).
The Maternal PKU Collaborative Study (MPKUCS) is a 16-year multi-national prospective study of the effects of dietary treatment during pregnancy in women with PKU (Koch et al, 1993; 1994).  The pregnant women received ultrasound examinations, nutrition consultation, and metabolic monitoring as part of the study protocol.  Enrollment of pregnancies in the MPKUCS ended in April 1996, and follow-up of offspring ended February 1, 2000.  The Contributing Centers remain to recoup missing data and assist with interpreting the results.  However, there are no funds for monitoring pregnancies.  In short, a gap has now been created between the information learned from this research study and the provision of services for women with PKU. Through the MPKUCS, we learned that over 88% of women with PKU continued to attain metabolic control after pregnancy begins.  Offspring scores on tests of development correlated with the number of weeks that elapsed until metabolic control was achieved.  There was no "safe zone".  Every week counted, in that each weekıs delay in maternal metabolic control predicted further declines in scores (Waisbren et al, 2000).   
Maternal PKU is becoming a problem of considerable magnitude.  The MPKUCS enrolled 554 pregnancies, in which 393 resulted in live births (71%).  The number of completed pregnancies increased steadily after 1984 from less than 10 babies in the first year to 64 in 1995.   The incidence of PKU in the United States is 1:12,000.  Based on US Census Bureau statistics, there are an estimated 5,000 women with PKU of childbearing age and an estimated 300 maternal PKU births per year (www.census/gov/prod/3/97pubs/p20-499/pdf).
	When the infrastructure of the Maternal PKU Collaborative Study is dismantled, there will be no forum, no avenue for exchange of information and no studies to further knowledge about this potentially devastating condition. There is momentum now to address the issue.  The participating Metabolic Centers have Resource Mothers who are excited about the project and eager to begin work in the field. If we do nothing about the poor adherence to medical recommendations in this disorder, we turn our backs on the information derived from the MPKUCS.  We turn our backs on the opportunity to translate research information about a genetic disease to delivery of clinical care.  Moreover, we risk erasing the gains achieved by newborn screening in just one generation.

Relevance to Priority Research Issues and Questions FY 2000-2003

	The Maternal PKU Resource Mothers Program most directly addresses Priority IX of the MCHB Research Agenda: Development, Evaluation and Validation of MCH Clinical Treatments, Outreach Strategies, Program Interventions, Care Guidelines and Case Management Approaches (8.1.12: randomized controlled studies of the efficacy and cost effectiveness of MCHB-developed and promoted Bright Futures guidelines).  
This randomized controlled study directly evaluates the use of a specially designed home visitation program for a well-defined genetic condition.  This approach may be applicable to other conditions, such as maternal diabetes, smoking, and alcohol use, in which the mother must significantly modify and monitor her behavior during pregnancy in order to protect the fetus.  The intervention model is integrated with the medical home and existing state and local health promotion programs, such as Early Intervention, School Based Health Centers and the Bright Futures Initiative. 

b. Review of the Literature.

Home visitation programs, such as the Maternal PKU Resource Mothers Program, have been successful in reducing the incidence of premature births, underweight infants, and birth complications (Olds, 1992; Olds, Kitzman, 1993). The theoretical basis for home visitation programs comes from an ecological model of development.  Bronfenbrenner (1979) conceived of an ecological environment as a nested arrangement of concentric structures that include the individual at the center and the wider social systems of which she is a part radiating outward.  From this perspective, development occurs through a re-alignment of the various systems within the individual's particular ecological environment. Because the systems are inter-related, change at any level affects the entire set of structures. Thus, pregnancy, from an ecological perspective, can be viewed as a period of disequilibrium that opens up a system to change (Holden, Willis, Corcorran 1992). It is hypothesized that at such periods, individuals who are at-risk, may be more adaptable to new methods of coping and more amenable to help from others (Mercer, 1990).   From an ecological perspective, the relationship between the woman and the Resource Mother influences interactions with family, health care providers and community resources within a particular societal and cultural context. At each level, the various supports and stresses can have an impact on the specific interaction and the system as a whole (Gomby et al 1993, Barnett 1993). 
 The need for taking an ecological perspective in considering maternal PKU became apparent when data from a prospective, longitudinal study on the psychosocial factors in maternal PKU were analyzed (Waisbren et al, 1991; 1995). The study revealed that women who could rely on a social network to support their efforts to follow the diet and women who perceived the diet as palatable and effective maintained metabolic control sooner in pregnancy.  Social and cultural factors played a significant role (Shiloh et al, 1990) as did the nature of the disorder itself (St James et al, 1993), and the woman's particular stage in the maternal PKU life cycle  (Shiloh, Waisbren, Levy, 1989). This longitudinal study provided the impetus for the development of the Resource Mothers home visitation program in New England  (Levy, Waisbren, 1994).	
 In the Maternal PKU Program, Resource Mothers are mothers of children with PKU.  Since these women have children of their own with PKU, they are familiar with the diet, the methods for calculating phenylalanine in food, and with the stresses and strains such restrictions produce on a family. The Resource Mothers clarify the issues of maternal PKU, help pregnant women obtain prenatal care, purchase appropriate foods or pregnancy-related items, plan and cook meals and arrange transportation to doctor appointments. They provide emotional support and maintain communication with the metabolic center. They are paid a stipend for time spent in training and supervision and for their visits to the young women. 
A recently completed non-randomized study of the Resource Mothers Program (St. James, Shapiro, Waisbren, 1999) compared 19 pregnancies in women with PKU who received the services of a Resource Mother to pregnancies in 66 women with PKU, enrolled in the MPKUCS, who did not have a Resource Mother. This study was conducted as a doctoral dissertation and was based on the experiences of the New England Resource Mothers Program. The results indicated that the pregnant women who received Resource Mothers achieved metabolic control at a mean of 8.5 weeks compared to a mean of 16.1 weeks for those in the comparison group. At one year of age, infants whose mothers had a Resource Mother had a mean DQ of 108 and the mean DQ of the control group was 95.  While the study suggests promising results from this type of home visitation program, a number of factors may have confounded the results. Only the women in the Resource Mothers group had been enrolled in a previous Psychosocial Study, in which they had annual visits from professionals from the Children's Hospital and many had also attended Maternal PKU Camp (Waisbren et al, 1997).  Data from the comparison group were retrospective from the MPKUCS and had been obtained 5 to 10 years previously, when recommendations for dietary control were not as strict. Finally, assignment to groups was not randomized. In a review of 46 randomized controlled trials in anesthesia and surgery, only half the trials demonstrated that the therapeutic innovation was preferable.  Yet based on historical data, all were thought to improve outcome (Gilbert et al, 1977).  Only a randomized controlled study will provide a basis for decisions about continuing the program, expanding it to other centers, or developing alternative strategies for improving adherence to treatment in maternal PKU (Pocock 1983).
		
c. Explanations of Concepts and Working Definitions.
In this proposal, the term intervention refers to the Maternal PKU Treatment Plan (Appendix 1) which is a set of guidelines for the treatment of maternal PKU.  These include scheduled medical appointments, regular metabolic monitoring, and recommended nutrient intakes and weight gain during pregnancy, as outlined in the Treatment Planıs Protocol for Nutrition Support.  All subjects will be treated according to the same guidelines, except that half the sample will have Resource Mothers.  If the physicians believe that the intervention should be changed in order to best treat the patient, they will do so.  The study only requires that such deviations be noted on a form describing the progress of the pregnancy.
The term protocol refers to the Study Protocol, the various forms that will be completed for data collection.
 
d. Hypothesis and Specification of Variables.
Primary Hypotheses:
Women with PKU followed by an intervention that includes Resource Mothers compared to women with PKU followed by an identical intervention without Resource Mothers will:
1. attain metabolic control sooner.
2. have babies with better outcome at birth.
3. have babies with better outcome at one year of age.
These predictions are justified by previous experiences with Resource Mothers and preliminary results of the Maternal PKU Collaborative Study (Koch et al, 1994).

Secondary Hypotheses:
Women who have Resource Mothers will:
 4. have greater adherence to the Maternal PKU Treatment Plan.
5. have better weight gain during pregnancy.
6. consume a greater percentage of recommended formula, protein, calories, fat, phenylalanine and tyrosine.
7. provide a more stimulating and stable environment for their children. 
The secondary hypotheses address factors that may be closely related to the variables evaluated in the primary hypotheses.  For example, metabolic control is related to adherence to treatment (the regularity with which the women keep prenatal and metabolic clinic appointments and send in blood samples for phenylalanine level determinations), greater intake of the special formula and low protein foods, and increased weight gain.  These outcomes are predicted based on our experiences with the Resource Mothers Program.

 Additional Hypotheses and Analyses:
It is hypothesized that women in the Resource Mothers group will have greater social support, more knowledge of maternal PKU, higher self-esteem, more positive attitudes about treatment, better home organization and fewer life stresses.  It is also hypothesized that the women in the Resource Mother's group will be hospitalized less often during pregnancy.  Hospitalization may have an independent influence on outcome and hence is added as a mediating variable.  There are no hypotheses regarding the use of community services or locus of control (degree to which the women perceive life's events as under their control). Having a Resource Mother may lead to more extensive utilization of community resources or, conversely, to less of a need for such services.  Similarly, having a Resource Mother may increase a woman's sense of responsibility for the pregnancy or increase her dependence on an outside support.  The cost of the Resource Mother's Program will be calculated and reported.  However, analyses concerning cost-benefit of such programs are beyond the scope of this study.

Specification of Variables  
a. The independent variable is the treatment group to which each subject is assigned, either to the group with a Resource Mother or the Control group.  

b. Two variables will be used for stratifying the sample: the Metabolic Clinic (in Massachusetts, Illinois, Pennsylvania, Maryland, Wisconsin, or Georgia) and timing of treatment (whether treatment was started prior to or after pregnancy began).  

c. The dependent or outcome variables are defined below:
(1) Maternal Metabolic control is the number of weeks that elapse from the time the woman begins treatment (if she comes to attention after she is already pregnant) or from the time the woman becomes pregnant (if she begins treatment prior to pregnancy) until she consistently maintains her blood phenylalanine level below 6 mg/dl throughout the remainder of pregnancy.  The Maternal PKU Treatment Plan recommends that maternal blood phenylalanine levels remain between 2 and 6 mg/dl.  In accordance with "common practice" among metabolic clinics, other services and hospitalization are not initiated until the maternal phenylalanine levels are above 10 mg/dl (Levy et al, 1994).
(2) Weight gain during pregnancy is defined as the percent of recommended weight gained based on the pre-pregnancy weight. Weight gain will be assessed at the end of the first, second and third trimesters.  Closely tied to weight gain are other nutritional variables. These include the percent of recommended formula, protein, calories, fat, phenylalanine and tyrosine intake as calculated by procedures outlined in the Protocol for Nutrition Support, located in the Treatment Plan (Appendix 1).
(3) Better outcome at birth is defined as birth head circumference, birth length, and birth weight that are appropriate for gestational age and gender using standardized birth measurement tables (Niklasson et al, 1991). Birth head circumference will be used as the primary neonatal outcome variable because birth head size has consistently been reported as the birth measurement most highly correlated with maternal metabolic control and is an indicator of brain growth in utero (Drogari et al, 1987; Rouse et al, 2000). Information about a positive diagnosis of congenital heart disease will also be obtained and the rates of occurrence in the Resource Mothers and Control groups will be compared. Congenital heart disease occurs in 7.5% of the cases in which metabolic control is not achieved during the first 8 to 16 weeks of gestation, the period of cardiogenesis (Rouse et al, 2000).
(4) Better outcome at one year of age is determined by developmental evaluation (Bayley Scales of Infant Development - 2nd Edition) yielding a Developmental Quotient or DQ. A DQ below 85 suggests developmental delay and a DQ below 70 signifies risk for mental retardation.  The child's Motor Development Index or MDI and the child's language quotient (from the Receptive-Expressive Emergent Language Test) will also be determined for secondary analyses.  Results from the MPKUCS suggest that language may be particularly affected in maternal PKU (Waisbren et al, 1998; Waisbren et al, 2000).
(5) Home environment refers to the level of safety, stimulation and emotional security as measured by a standardized instrument, the HOME Scale.  

d. The intervening variables are those that should not differentiate the randomly selected treatment groups.  However, if they do differentiate the groups, they could introduce bias.  These include the following:
1. Demographic variables:  
   (a) Maternal age at the time of enrollment in the study
   (b) Socioeconomic status based on education and employment of household head
  	   (c) Ethnicity as the woman describes herself
   (d) Marital Status at the time of enrollment and when the child is 1 year
 	   (e) Parity is the number of children previously born
   (f) Gravida is the number of pregnancies the woman has experienced 
   (g) Maternal IQ based on an adult intelligence test
    (h) Travel time to metabolic center from woman's home, in hours 
2. Variables related to Medical History:
   (a) Maternal natural blood phenylalanine level (in mg/dl) is defined as the highest of 3 determinations taken when the mother was on an unrestricted diet. If not available in medical records, a newborn level (for women continuously on diet) or highest noted level on a normal diet are acceptable. 	     
   (b) Previous pregnancy outcome will be described as either completed or not completed. Mental retardation, heart defect or death will also be noted.
   (c) Current pregnancy complications refer to complications of nausea, gestational diabetes, hypertension, or "other".
   (d) Current pregnancy completed is defined as resulting in a live birth.
   (e) Gender of baby
   (f) Gestational age of baby is weeks gestation recorded in birth records.

e. The mediating variables are those that are hypothesized to be related to treatment and that influence some or all of the outcome variables. 
1. Adherence to the Maternal PKU Treatment Plan is the percentage of data points requiring the woman's cooperation as outlined in the study protocol that are "not missing". The woman must keep medical appointments, monitor nutrition, send in blood tests, and have ultrasound examinations as prescribed.
2. Maternal hospitalization means that the woman was admitted for at least an overnight stay at some point during her pregnancy.

f. A final category of variables will be for additional analyses to provide information about implementation of the Resource Mothers program and its impact on the pregnant woman's internal and external resources. 
1. Community services used. The Study Coordinator will check off community services, such as School Health Centers, Maternal and Child Health programs, Early Intervention, visiting nurses, health aides, or others.  It will be noted whether the Bright Futures Professional Encounter Forms were used by the medical home.
2. Cost of the Resource Mother's intervention will be calculated from charges to the grant for diet supplies, stipends and reimbursement for telephone and travel.
  	3. Pre- and Post-test scores address 7 areas within the woman's ecological system that can be affected by the relationship with the Resource Mother. Specific instruments will be used to assess social support, knowledge, self-esteem, locus of control, home organization, family stress, and attitudes about treatment. 

Appendix 2 contains data collection forms for each set of variables. A checklist (Form 8) will be used to record the collection of each data point. This form will then be used to provide a summary or the percent of data collected for each variable. 

MODEL OF SET OF RELATIONSHIPS BETWEEN VARIABLES


EXOGENOUS OR INDEPENDENT 
VARIABLES	
MEDIATING VARIABLES	
DEPENDENT OR OUTCOME VARIABLES

GROUP: RESOURCE  MOTHER or NO RESOURCE MOTHER
(Check for comparability on Intervening Variables:
Demographic factors and
Medical History)	
ADHERENCE TO TREATMENT PLAN

HOSPITALIZATION DURING PREGNANCY	
MATERNAL METABOLIC CONTROL

BIRTH OUTCOME

DEVELOPMENT AT ONE YEAR

HOME ENVIRONMENT

Alternative models will also be tested when analyses are focused on the ways in which the Resource Mothers program may have an impact on outcome.  For example, in one such model, maternal metabolic control and home environment may be considered the mediating variables and birth measurements and DQ may be considered the outcome variables.




 	SUMMARY TABLE OF VARIABLES

VARIABLE NAME	CATEGORY	MEASURE/
INSTRUMENT	UNITS

INDEPENDENT VARIABLE

Treatment Group	

Nominal	

Random Assignment	

With Resource Mother/Without Resource Mother

VARIABLES FOR STRATIFYING SAMPLE

Metabolic Clinic 
Treatment started prior to pregnancy	


Nominal
Nominal	


Form 1
Form 1	


MA/IL/GA/PA/MD/WI
Yes/No

DEPENDENT/OUTCOME VARIABLES

For Primary Hypotheses

Time from treatment initiation to metab. Control

Birth head circumference 

Child DQ at one year

For Secondary Hypotheses:

% recommended  weight gain
  First Trimester
  Second Trimester
  Third Trimester

% recommended formula intake during pregnancy
% recommended protein intake during pregnancy
% recommended calories intake during pregnancy
% recommended phe intake during pregnancy
% recommended tyrosine intake during pregnancy
% recommended fat intake during pregnancy

Offspring outcome --
Birth weight 
Birth length
Congenital heart disease
 
Child Motor development at 1 year
Expressive language score
Receptive language score
HOME Scale Score	




Interval

Interval

Interval




Interval
Interval
Interval

Interval
Interval
Interval
Interval
Interval
Interval


Interval
Interval
Nominal

Interval
Interval
Interval
Interval	




Form 2a

Form 3

Bayley Scales




Form 2c
Form 2c
Form 2c

Form 2b
Form 2b
Form 2b
Form 2b
Form 2b
Form 2b


Form 3
Form 3
Form 3

Bayley Scales
REEL
REEL
HOME Scale	




Weeks

z score

<50 ­ 150




percent
percent
percent

percent
percent
percent
percent
percent
percent


z score
z score
Yes/No

<50 ­ 150
<1 ­ 100
<1 ­ 100
0 ­ 45

INTERVENING VARIABLES-- TO CHECK COMPARABILITY BETWEEN GROUPS

Demographic factors:

Maternal age
Socioeconomic status
Ethnicity
Marital status
Parity
Gravida
Previous pregnancy outcome
Travel time to Metabolic Center

Medical History:

Maternal natural blood phe
Maternal IQ
Current pregnancy complications
  Nausea
  Gestational diabetes
  Hypertension
  Other
Current pregnancy completed

 
Gender of baby
Gestational Age of baby
	





Interval
Ordinal
Nominal
Nominal
Nominal
Nominal
Nominal
Interval



Interval
Interval
Nominal
Nominal
Nominal
Nominal
Nominal
Nominal


Nominal
Interval	





Form 1
Hollingshead
Form 1
Form 1
Form 1
Form 1
Form 1
Form 1



Form 1
WAIS-R
Form 2d
Form 2d
Form 2d
Form 2d
Form 2d
Form 2d


Form 3
Form 3	





Years
1 ­ 5
White/Other
Married/Other
First/Other
First/Other
Describe
Hours



mg/dl
50 ­ 150
Yes/No
Yes/No
Yes/No
Yes/No
Describe
Yes/No


Male/Female
Weeks

MEDIATING VARIABLES

% adherence to Treatment Plan by woman
Maternal hospitalization
	


Interval
Nominal
	


Form 4
Form 4
	


0 ­ 100
Yes/No


VARIABLES FOR ADDITIONAL ANALYSES 

Community services used
Cost of Resource Mothers Intervention
Pre- and Post Test Scores:
  Social Support
  Knowledge
  Self-esteem
  Locus of control
  Home organization
  Life Stresses
  Attitudes about treatment	


Nominal
Interval

Interval
Interval
Interval
Interval
Interval
Interval
Interval
	


Form 4
Grant Budget
   
Fishbein
Shiloh 
Coopersmith
Nowicki
Questionnaire
Brugha
Questionnaire	


Description
Dollars

3 ­ 18
0 ­ 10
0 ­ 25
0 ­ 40
0 ­ 16
0 ­ 12
5 ­ 35

e. Tests and Measurements.
Appendix 2 has samples of the internet data collection system and copies of data collection forms. These forms can now be submitted and summarized electronically via a secure website, developed by Acordex.

1. Registration and Background Form  (Form 1): This questionnaire, administered by the Study Coordinator, collects information related to stratification and random assignment.  The woman will be asked questions regarding the timing of diet initiation in relation to pregnancy, which will form the basis for group assignment.  Background information will be obtained related to maternal age, ethnicity, marital status, parity, pregnancy history and travel time to the metabolic center, and maternal blood phenylalanine level.  Socioeconomic status (SES) will be measured according to the Hollingshead-Redlich Index of Social Position (Hollingshead, Redlich, 1975) based on employment and education of the household head. Maternal IQ will be obtained from the Wechsler Adult Intelligence Scale-Revised (Wechsler, 1981).  In many cases, this will be available from metabolic center records.  If not, a psychologist will administer this test for the study.

2. Maternal PKU Treatment Pregnancy Progress Forms (Forms 2a, 2b, 2c and 2d).  Form 2a is a log of maternal blood phenylalanine levels that will be completed weekly, each time such a laboratory determination is made. Form 2b is the Nutritional Assessment Worksheet. The Maternal PKU Treatment Protocol specifies monthly collection of 3-day diet records.  When food records are incomplete or inaccurate  (Crawford et al, 1994), the nutritionist will interview the woman and record her intake for the past day. Fran Rohr, the nutritionist in Boston, will request the original diet records twice during each pregnancy and at any other time that the nutritional values appear unusual. Since only a limited number of foods are allowed in the maternal PKU diet and many women eat the same foods daily, checklists (Appendix Form 2b-1) can be used for those who have difficulties maintaining food records. The nutritionist will query further to determine if the young woman ate foods not allowed on the diet or foods not on the list. The amount of formula consumed will be recorded and verified by checking the amount of formula ordered. Nutrient analysis will be performed using the Amino Acid Analyzer Software (Ross Products, Columbus, OH) which all the participating centers use. The food records will thus allow assessment of protein, energy, fat, phenylalanine and tyrosine intakes. Form 2c is the Weight Record.   At the end of each trimester, health providers will record the date and woman's weight.  Form 2d is where the Study Coordinator will record the pregnancy complications of nausea, gestational diabetes, and hypertension.  Other complications will be described in detail. This information will be obtained prospectively from clinic visits and retrospectively from obstetrical records.

3. Offspring outcome at birth will be recorded on Form 3. Newborn forms will be sent to the birth hospitals prior to delivery.  The Study Coordinator will make arrangements to be informed when the birth occurs.  She will then confirm that the forms are available and, if not, will FAX another copy.  She will call within the first day of life to ensure that the measurements were obtained and were legibly written on the form.  The form includes gestational age, gender, birth weight, birth length and birth head circumference and information about any anomalies. 

4. The degree to which the experimental and control treatments are delivered according to the intervention plan will be evaluated through Data Collection Form 4, completed by the Study Coordinator. This form lists each intervention  (such as visits to metabolic center and blood testing) and has columns to record the date each was completed.  Deviations from the Treatment Plan will be explained.  Form 4 also lists other community support services to be checked off if used. 

 5. Offspring outcome at one year will be recorded on Form 5. A battery of psychological tests used in the MPKUCS will be administered to all offspring.  During the last year of the project, children less than a year old will be administered the tests just prior to the end of the study.  SES and marital status will again be recorded. Local psychologists will provide the developmental testing in the homes of the infants.  A network of psychologists throughout the country is already known to the Principal Investigator who coordinated psychological follow-up for the MPKUCS.  The protocols from these tests will be sent to Boston for double scoring.  Since these tests require few subjective judgments, discrepancies in the scores are generally due to mistakes in calculations.  When errors are found, a copy of the protocol with the corrections highlighted will be sent to the originating center.  If there are questions about the scoring, the PI will contact the psychologist who performed the testing.  The assessments will include the following:
a. Bayley Scales of Infant Development, Second Edition (Bayley, 1993) provide standardized mental and motor scores (mean 100,SD 15).  Reliability is .88 for the Mental Scale and .84 for the Motor Scale at age 1 year.  Validity studies show that the test reflects the current stage of development and predicts which children will demonstrate developmental delay later on. 
b. Test of Receptive Expressive Emergent Language-2nd Edition (REEL-2) (Bzoch, League, 1991).  This is a measure that combines observation of the child and information elicited from the parent.  Standard scores (mean 100, SD 15) for expressive, receptive and total language are provided.  Recent norms for children 3 months to 36 months are provided.  Reliability coefficients for 1-year assessments were .97 for internal consistency and .80 for test-retest reliability. In one study, the REEL-2 predicted language development beyond the .01 level.
c. The Home Observation for Measure of the Environment (Caldwell, Bradley, 1984). The HOME scale for infants contains 45 items that simply need to be checked off.   The HOME Scale was normed on a sample of 174 families in Little Rock Arkansas in 1984.  Since the norms may not be applicable to the part of our sample population drawn from Eastern and Midwestern states in the 1990's, the raw scores on the HOME Scale will be used and scores in the lowest quartile for this sample will be considered to be "low".  Detailed instructions for completing the form are available from the publishers and will be included in the data collection packet sent to the psychologists conducting the evaluations.  This scale was used in the Maternal PKU Collaborative Study (MPKUCS) and proved to have an independent relationship to developmental quotients at one year over and above the maternal metabolic control during pregnancy

6. The Child's Medical Follow-up during the first year will be recorded by the Study Coordinator on Form 6.  The number of pediatric visits, hospitalizations, surgeries and outside health services (such as early intervention) will be noted.  

7. The Psychosocial Pre- and Post-Tests will be recorded on Form 7. The Resource Mothers or Study Coordinators will administer the tests (Appendix 3D) to the women at the time of enrollment and at the last prenatal metabolic clinic visit at 32 weeks gestation. The tests will indicate positive and negative changes in internal and external resources.  After spending a week at Maternal PKU Camp, some young women expressed less confidence in themselves and less positive attitudes about the treatment, perhaps because they had become more aware of the risks and responsibilities of maternal PKU (Waisbren et al, 1997). 
a.  Social Support Questionnaire, adapted from Fishbein and Ajzen (1975).  The degree to which others in the social network support "keeping strictly to the diet" was one of the best predictors of adherence to medical recommendations in a previous study (Waisbren et al, 1991; 1995). 
b. Knowledge of Maternal PKU (Shiloh et al, 1990) This ten-item test has become a standard measure of knowledge of basic genetic and treatment facts.  This test was validated on a population of women with a mean IQ of 85 + 14 (range 60 - 119) and has a validity coefficient of .63. 
c. Self-Esteem Questionnaire (Coopersmith, 1967) A standard measure of self-esteem, this 25 item questionnaire is still frequently used with adults, partly because of its simple language and easy scoring.  This scale was predictive of adjustment in newly diagnosed diabetic adolescents (Jacobson, 1987) and was found to correlate .95 with longer measures of self-esteem (Coopersmith, 1967).
 d. Locus of Control (Nowicki, Duke, 1974) Another standard instrument, this test is used to rate the degree to which people attribute events in their lives to their own actions or to forces outside of their control. It has excellent reported reliability and validity and can be administered to those with a fifth grade reading level.  
e. Home Organization Scale, developed for the Resource Mothers Program, is a checklist of supplies and activities related to the diet. Although not validated, In pilot studies, this measure was found to be associated with SES but not with IQ.  
f. The List of Threatening Experiences, a life stress questionnaire (Brugha, Cragg, 1990) This is a list of 12 possible experiences that  are deemed stressful, such as loss of a loved one, divorce, homelessness and loss of work.  Reliability coefficients ranged from .7 to .9. In terms of concurrent validity, the sensitivity of the questionnaire was .89 and specificity was .88.
g. Attitudes about Treatment Questionnaire is a revised version of an instrument used in an evaluative study of the maternal PKU camp (Waisbren et al, 1997).  It includes statements to be rated on a 5 point scale assessing the woman's attitudes about the formula, low protein diet, the pregnancy, her ability to cope and her rating of her relationship with her Resource Mother, if applicable. It has not been validated, but taps a variable that was highly predictive in the psychosocial study (Waisbren et al, 1991;1995).  Results on this scale were not related to IQ.

8. A Summary of Data Collection Rate can be completed from Form 8. This checklist will be used to determine the extent to which data have been collected. 
For example, when information is received about the babyıs birth measurements, the corresponding data points will be marked off on the checklist.  The percent of data points completed can be calculated from this form.

The Study Coordinators will enter their data onto computerized forms directly via the internet data management option, developed by Acordex (Appendix 1A).  Through this system, the forms can be transmitted to Boston as information becomes available. The PI can see that the forms are completed properly and that the study protocol is followed. 

f. Study Design

The study is a randomized controlled clinical trial  (Spilker, 1991). There are three key features to this type of design (Knapp, Miller, 1992): random assignment, a control group, and an intervention.
1. The PI (via a random numbers table) will randomly assign subjects to the group receiving the services of a Resource Mother or to a control group receiving the same treatment plan without a Resource Mother. This will be done in each clinic in blocks of two, stratified by clinic and whether or not the woman initiated treatment prior to pregnancy. 
The participating clinics are: Childrenıs Hospital (Harvey Levy, MD) in Boston, St. Christopher's Hospital  (Warren Grover, MD) in Philadelphia, Children's Memorial Hospital (Barbara Burton, MD) and the University of Illinois Medical Center (George Hoganson, MD) in Chicago, Emory University School of Medicine (Louis Elsas, MD) in Atlanta, Waisman Center, University of Wisconsin (Jon Wolff, MD) in Madison, and Johns Hopkins Hospital (Michael Geraghty, MD) in Baltimore. All have experience treating maternal PKU.
 Each eligible woman must consent to participate in the study and specifically to random assignment. The metabolic centers will be aware of who has a Resource Mother, but the psychologist evaluating the children will be blinded as to treatment group. Randomization in blocks of two at each clinic and stratification of the sample minimizes biases in allocation. Nonetheless, the groups may not be comparable in maternal age, SES, marital status or other variables listed in the summary table under the category of "intervening variables to check comparability between groups".  Given the small sample size, it is reasonable to stratify only according to the timing of diet initiation.  If inequalities are found, despite randomization, analyses will be conducted separately for each group and the results will be reported descriptively
2. The control group will receive the same treatment as the experimental group except for the services of a Resource Mother.  
a. A Maternal PKU Treatment Plan (Appendix 1), derived from that used in MPKUCS, includes basic indices for determining blood phenylalanine levels and nutritional intake.  In addition, it includes "decision trees" to direct the health care providers to institute additional services (such as MCH case management) or alternate strategies (such as changes in formula composition) if the woman fails to gain weight or attain metabolic control. The treatment plan specifies the number of visits to the maternal PKU clinic, the timing of ultrasound evaluations, and the frequency of blood testing for metabolic levels. 
b. Scope of nutritional assessments: All nutritional assessments will be performed by Registered Dietitians and checked for accuracy by Fran Rohr, the nutritionist in Boston.  Nutritional assessments are part of basic maternal PKU follow-up at each of the clinics in the study. The primary nutrition variables are the woman's blood phenylalanine level and the percent of recommended weight gain. Data from the MPKUCS indicate that the most important nutritional variables in predicting birth outcome are protein, energy and fat intakes (Michals et al, 1996). Adequate intake of protein is associated with adequate consumption of the formula which meets 80 to 90% of a woman's nutritional needs during pregnancy.  Careful attention, therefore, will be given to assessing the amount of formula consumed.  This information will be obtained through food records and an assessment of the amount of formula ordered. The nutritionists will be provided with assessment forms that contain only information that is routinely obtained. It is not expected that the nutritionists will be spending additional time with the young women due to their participation in the research.  Therefore, no additional funds are requested. 
c. Outreach and follow-up: Adolescent females will be targeted for health education outreach programs. Information about Maternal PKU will be provided to the patients tracked by each clinic and to their medical homes.  School Based Health Centers will be contacted about the needs of young women with PKU and the risks of maternal PKU. (There are 36 such Centers in Massachusetts, 32 in Pennsylvania, 18 in Illinois, 27 in Wisconsin, 12 in Georgia and 43 in Maryland.) A website is being established in conjunction with the New England Consortium of Metabolic Programs. Information about maternal PKU and copies of the "Bright Futures Encounter Forms for Professionals" will be provided to the Metabolic Centers for distribution to the medical homes.  We will work with the American Academy of Pediatrics Initiatives for Children with Special Needs to ensure enrollment of all maternal PKU offspring in an accessible and competent medical home. Care providers will be encouraged to refer all offspring from late or inadequately treated pregnancies to Early Intervention. The Bureaus of Family and Community Health will disseminate a summary of maternal PKU to the Early Intervention Programs (Appendix 1).  Public discussions are underway with officials at the Department of Public Health in Massachusetts regarding the possibility of having Maternal PKU listed as a risk factor on the confidential section of the Birth Certificates (Appendix 1).  This triggers follow-up from the Department of Public Health and automatically qualifies the child for Early Intervention.  Similar discussions will be initiated in the other states.  
d. The role of the social worker: The social worker, Isabel Bailey, LICSW, will work with the Coordinators to enhance outreach, cultural sensitivity and use of community services. She will provide information to School Based Health Centers, State Departments of Public Health, Early Intervention Programs and the medical homes. She will also see to the distribution of Bright Futures Encounter Forms for Professionals and information about maternal PKU to medical homes after the babies are born.
3. The independent variable is the intervention being evaluated, the Resource Mothers Program. (Appendix 3 includes materials related to the Resource Mothers Program.)
 a. Recruitment of Resource Mothers: A nutritionist or social worker has been designated as the Study Coordinator at each metabolic center that will be enrolling patients.  She recruited the Resource Mothers and will assume responsibility for coordination, supervision, data collection, data entry and adherence to the Study Protocol. Thirteen mothers of children with PKU were recruited.   Our experience has been that the mothers who chose to participate have the time, personality, expertise and commitment necessary for the job. 
b. Training: Training began with a two-day workshop in Boston in May 1998. Topics included the medical and genetic aspects of maternal PKU, a detailed description of dietary issues in maternal PKU along with a review of the data collection methods for diet information, outreach and cultural competency. The experienced Resource Mothers from New England presented their Resource Mothers Book (Appendix 4) which provides detailed recipes and guidelines for dealing with difficult situations. Safeguards for ensuring timely data collection were introduced.  The Coordinator from each metabolic center participated in the training workshop.  She reviewed the data collection forms and the computer program and conducted a practice enrollment.  Each Resource Mother and Study Coordinator administered the battery of questionnaires to an experienced Resource Mother for practice.  Each Resource Mother was given a notebook with study procedures and instruments. Re-training of the Resource Mothers will take place at each of the Metabolic Centers during an individual supervision session.  The Coordinators will review materials provided in the notebook during the initial training.  Specific attention will be given to adhering to the research protocol.
 c. The Resource Mothers Intervention: The goals of the program are to provide the pregnant women with information, emotional support and practical assistance. The Resource Mothers focus on increasing the woman's supports and decreasing her stress. Preliminary data suggest that it is important for them to individualize their activities to meet the specific needs of the pregnant woman and at the same time, maintain a nonjudgmental, supportive relationship. In accordance with the Study Protocol, the Resource Mother meets 20 times with the young woman with PKU. The timing of the visits depends on the needs of the woman. Some women require weekly visits during the time of treatment initiation. Later in pregnancy, most of the women have grown accustomed to the dietary regimen, have increased tolerance for protein, and require visits less frequently. Two visits take place after the baby is born to help the mother adjust to parenthood.  The activities of the Resource Mothers include cooking meals and snacks, teaching new recipes, purchasing necessary supplies for formula preparation and measuring food, and accompanying the woman to regular prenatal or metabolic clinic appointments. The Resource Mothers sometimes help with completing applications for WIC or other forms of assistance. They explain maternal PKU to extended family members and friends.  They listen to concerns, encourage communication with the metabolic center, and facilitate referrals.
d. The Study Coordinators and medical directors at each center provide supervision of the Resource Mothers by phone supervision on a monthly basis. Supervision will occur in person 3 times a year.
e. Stipends: Resource Mothers will be paid $800 for their work for each completed pregnancy and $250 for pregnancies that are not completed. They also will be reimbursed for travel and phone calls.

g. Population Description and Sampling Plan

The sample will be selected from all maternal PKU pregnancies treated in the selected metabolic clinics. Directors of these clinics had expressed an interest in the program and contributed a significant number of cases to the MPKUCS. A log will be maintained at each clinic documenting that each participant requires dietary treatment during pregnancy, plans to continue her pregnancy and accept treatment, and is less than 20 weeks pregnant at the time of enrollment. This latter restriction should not appreciably reduce the sample pool since in the last four years in New England none of the pregnancies came to attention past 20 weeks gestation. Outreach to all tracked women may also reduce the number of very late treated pregnancies. In the log, a list of ineligible women will be maintained.
 Displayed in the table below are power estimates for the 3 main outcome variables. For each scenario, sample sizes necessary to achieve 80% power are determined, using a 2-tailed, alpha = .05 test and assuming a normal distribution (based on a t-test), after an adjustment for 30% attrition (based on the MPKUCS).  With 40 subjects in each group, there will be enough power to detect a difference of 8.37 weeks to metabolic control, a z score of .81 in birth head circumference, and 9 points in DQ.

POWER ANALYSES  (two-tailed a = .05, power = .80)

VARIABLE	
Resource Mothers Group	
Control Group	
Evaluable N per group needed, using t-test	
Enrolled N per group needed (We expect to enroll 57 in each group)

Weeks to metabolic control	
8.5 + 13.6	
16.9 + 13.6	
40	
57

	
8.5 + 9.7*	
16.1 + 13.6*	
38	
54

Birth head circumference, standard score	
-.56 + 1.3	
-1.3 + 1.3	
40	
57

	
-.56 + 1.3*	
-1.4 + 1.3*	
38	
54

Developmental Quotient (DQ)	
108 + 15	
99 + 15	
40	
57
 	
108 + 15*	
95 + 15*	
21	
30

… Actual values obtained in preliminary study (St. James et al, 1999).  The other figures illustrate the power to detect differences when the sample size is 40 in each group.  
In 1999, 24 completed pregnancies occurred among the proposed centers (1 in Philadelphia, 4 in Chicago, 5 in Atlanta, 2 in Baltimore, 3 in Madison and 9 in Boston).  Since we used conservative estimates for each aspect of the power analyses, it is likely that we will have enough cases to conduct multivariate analyses.


ESTIMATED SAMPLE SIZE	
Current Year (Boston only)	
Year 1	
Year 2	
Year 3	
Total

Completed Cases (Pregnancies carried to term)
	
    12	
 28	
 28	
12	
80
Discontinued Cases (Pregnancies not carried to term or lost by attrition)	
     6	
 12	
 12	
 5	
35

Total
	    
    18	
40	
 40	
17	
115

The women will be recruited for the study at their first clinic visit for maternal PKU.  Attrition will be due to consent refusals, loss to follow-up or pregnancies that are not completed.  In the New England Resource Mothers Program, only 2 women refused to accept a Resource Mother over a 7-year period.  In the MPKUCS, 7 of the 414 babies (less than 2%) died before their first birthdays. In less than 1% of the cases followed by the Northeast Region in the MPKUCS developmental testing was refused.  Although some families may move after the baby is born, they will be referred to a local psychologist for the developmental testing at one year 
In each of the metabolic centers included in this project, the recommendation to discontinue diet at age 5 or 6 years was given through the mid-1980's.  Some have returned to diet. Recent estimates indicate that approximately 50% of women with PKU of childbearing age currently follow a phenylalanine-restricted diet (Schulz, Bremer, 1995).  However, few maintain their blood phenylalanine levels in a range considered safe for pregnancy (Gleason et al, 1992). Thus, almost every young woman with PKU who becomes pregnant must make significant adjustments in her daily routine in order to maintain the level of metabolic control needed.  
 Some of the subjects may have limited intellectual abilities. The Resource Mothers and Study Coordinators assist women who may not understand the questionnaires and interview questions.  The instruments used in this study have all been used previously with this population and no differences in scores were related to IQ. This suggests that the tests are equally valid for individuals with average and below average intellectual abilities.  Some subjects may experience agoraphobia when not on a phenylalanine-restricted diet. However, when diet is resumed, these symptoms dramatically dissipate (Waisbren and Levy, 1991), making it unlikely that agoraphobia will decrease the sample size.
Class, Race and Ethnicity:  PKU is virtually unknown among Black, Ashkenazi Jewish, and Asian populations.  Based on the racial and ethnic composition of the MPKUCS, it is estimated that 1% will be Black, 1% will be American Indian or Alaskan Native, 10% will be Hispanic and the rest will be Caucasian.  The MPKUCS used 5 categories to describe socioeconomic status (SES): 32% were in the category representing the lowest socioeconomic status; 44% were in the next to lowest group, 18% were in the middle, 5% in the upper middle and none in the highest SES group (Koch, unpublished data).  Due to learning disabilities, reduced intellectual abilities, and emotional difficulties many women with PKU have limited education and jobs requiring little skill.  They marry men like themselves and hence are most often of lower or low middle SES.  Data described below reflect actual proportions found in the MPKUCS.
SES	American Indian/  Alaskan Native	Asian/Pacific Islander	Black	Hispanic	White	Total
Low	      0	     0	  0	   3	 23	 26

Low Mid.	
      1	
     0	
  1	
   3	
 31	
 36

Middle	
      0	
     0	
  0	
   1	
 13	
 14

Upper Mid.	
      0	
     0	
  0	
   1	
  3	
  4

High	
      0	
     0	
  0	
   0	
  0	
  0

TOTAL	
      1	
     0	
  1	
   8	
 70	
 80

h. Plan for data analysis

Data analysis will proceed according to our primary model, which postulates that the independent variable (Resource Mother support) affects maternal metabolic control, the home environment and offspring outcome by way of altering the mediating factor of adherence to treatment protocol. The first step will be to test for potential biases between the randomly selected study groups in terms of the "intervening variables".  Due to the random selection process, it is hoped that the groups will be comparable.  If not, separate analyses may have to be performed for the subgroups.  Although we expect the Resource Mothers treatment group to attain metabolic control sooner and hence have better outcomes, we will conservatively use two-tailed tests and confidence intervals in order to be able to comment on results that may occur in the opposite direction to our hypotheses. 
Analyses to test the specific hypotheses are as follows:
Hypothesis 1: The Resource Mothers treatment group will attain metabolic control sooner. The nonparametric Wilcoxon Mann Whitney test will be performed on the number of weeks from treatment initiation to metabolic control, since all women who achieve metabolic control prior to pregnancy receive a score of 0 weeks.  

Hypothesis 2: The Resource Mothers treatment group will have babies with better outcome at birth. A T-test based on z-scores of birth head circumference, birth weight and birth length will be used.  The relative occurrence of congenital heart disease and other anomalies will be analyzed by the Fisher's Exact Test. 

Hypothesis 3: The Resource Mothers treatment group will have babies with better outcome at one year of age. A T-test will be used to compare the DQ, motor development, and language quotients of offspring in each group. 

Hypotheses 4, 5, 6, 7: The Resource Mothers treatment group will have greater adherence to the Maternal PKU Treatment Plan, a higher percentage of recommended weight gain, a higher percentage of recommended nutrient intakes and provide a more stimulating environment for their child at one year.  These hypotheses will be tested using the Wilcoxon Mann Whitney test. We selected this statistical test over parametric tests because of concerns about distribution, especially in the measurement of the home environment.  Our experience has been that a bimodal distribution often occurs. For similar reasons, nonparametric correlations will be used to determine if percent of recommended nutrition intake is related to birth outcome and development at one year.

Additional Analyses: 

To test our model of the set of relationships between variables, hierarchical regression analyses will be used according to procedures that are described below.  To begin, we will predict each outcome (i.e., weeks to metabolic control, birth measurements and DQ) from the most proximal variable, for example, the predictive value of adherence to treatment to the outcome:  weeks to metabolic control.  Next, we will predict the outcome variable from the independent variable, Resource Mothers Group.  Both these analyses are hypothesized to be significant.  Finally, we will conduct the hierarchical regression, entering adherence to treatment plan first.  Now it is hypothesized that the independent variable (Resource Mothers Group) no longer has predictive value.  This would indicate that the Resource Mothers program is successful because it has an impact on adherence to treatment plan and not because of other factors.  These procedures will be repeated for each outcome variable.  
Other descriptive analyses will be conducted. The number of community services used and the cost of the Resource Mothers program will be presented through descriptive statistics. Paired T-tests will be used to compare the pre- and post scores on the psychosocial tests in each study group. These analyses permit the "full story" to be told.  A new treatment may be associated with positive outcomes.  However, some negative results may also occur.  For example, compared to controls, the Resource Mothers group may attain metabolic control sooner, but also experience a lowering in self-confidence or locus of control.   Moreover, if the Resource Mothers Program does not lead to better outcomes, we will want to know about the attitudes, sense of social support, degree of stress and other psychosocial factors that might have interfered with the program's objectives.  

i. Time Schedule
    	Month of Study during Year 1            			Year of Study

Activity	
1	
2	
3	
4	
5	
6	
7	
8	
9	
10	
11	
12	
2 	
3	
4

Re-contact all  Resource Mothers	
X	
	
	
	
	
	
	
	
	
	
	
	
	
	


Refresher training course	
	
X	
	
	
	
	
	
	
	
	
	
	
X	
	


Enroll pregnancies	
	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X

Resource Mothers in field	
	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X

Supervise participants	
	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X

Collect data	
	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X

Enter data	
	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X	
X

Conduct data analyses	
	
	
	
	
	
	
	
	
	
	
X	
X	
	
	
X

Publish manuscripts	
	
	
	
	
	
	
	
	
	
	
	
X	
	
	
X


j. Financing
A variation on this proposal has been submitted to the National Institute of Child Health and Development (NICHD) for consideration and review.  To maximize chances of receiving funds, the proposal was developed to extend aspects of the Maternal PKU Collaborative Study through a Resource Mothers Program.  MCHB will be notified if funds become available from NICHD.


k. Human Subjects

The research involves human subjects at 7 sites.  

1. An estimated 115 women with PKU will become pregnant and 80 women will be included in the final analysis. Women who do not intend to continue their pregnancies and women who are more than 20 weeks pregnant at the time they contact the metabolic center will be excluded. Women with mild hyperphenylalaninemia who do not require treatment will be excluded. The age range of the pregnant women is expected to be 18 to 35 years.  (Younger teenagers who become pregnant will be eligible for the study with permission of a parent in addition to their own consent.)  Some of the pregnant women may attain IQ scores in the range of mild mental retardation, but all are expected to be competent to give consent.  Since PKU does not generally affect physical health in those who have been treated, it is expected that the subjects will be healthy.  All subjects will be pregnant since metabolic control during pregnancy is the primary focus of this study. Thirteen Resource Mothers, mothers of children with PKU, have been trained as home visitors.

2. Questionnaire data and psychological testing will be obtained especially for this study.  Medical history and information on metabolic levels, nutrition, weight gain and pregnancy complications will be obtained for prenatal care and will be recorded by the Study Coordinators and nutritionists.

3. Subjects will be recruited when they first come to the metabolic center to begin treatment in preparation for pregnancy or because they are pregnant.  The Study Coordinator will obtain informed consent. Consent will be sought for treatment according to the Maternal PKU Treatment Plan and for random assignment to one of two study groups (the Resource Mothers group and the group without Resource Mothers).  The subjects will be informed that the Treatment Plan is adapted from the Maternal PKU Collaborative Study and has been in use for 14 years.  The Treatment Plan includes referrals to community health and support services if metabolic control is not achieved.  The Resource Mothers program will also be described.  Prospective subjects will be advised that the Treatment Plan includes regular obstetrical visits and monthly visits to the metabolic center.  Two ultrasound examinations, one to confirm a viable pregnancy and another at 21 weeks are specified.  Weekly filter paper blood samples or other means of monitoring blood phenylalanine levels are also requested.  The women will be told that the study will not pay for medical expenses.  Subjects will be asked to consent to having information about their metabolic control and the progress of the pregnancy shared with the Resource Mothers and the Coordinator from the metabolic center. The subjects will be informed that their physician is free to deviate from the Maternal PKU Treatment Plan at any time if such a change is in the best interest of the woman or her fetus. The subject, the coordinator and a witness will sign the informed consent form.

4. There are no known medical risks to the woman or her fetus to participating in the study.  The Treatment Plan for Maternal PKU provides the best known medical intervention for this population. Psychological risks are also minimal. Women in the control group who do not receive a Resource Mother may feel that they are not receiving optimal treatment, although they do receive medical follow-up.  There might be a feeling of "intrusion" and there might be a potential risk to privacy (confidentiality). The Resource Mothers are not professionals, but will be told personal information about the pregnant women regarding blood phenylalanine levels and the progress of the pregnancy.  We are unaware of alternative treatments or procedures that might be advantageous to the subjects.5. Pregnant women in the group without a Resource Mother will be assured that referrals for additional medical interventions or community services will be made should the need occur. The Resource Mothers are trained respect privacy and maintain confidentiality.  They will be reminded during supervision about the importance of not revealing information about the pregnant woman to anyone outside of the study.  Our experience has been that the Resource Mothers take their responsibilities very seriously and bring concerns to supervision, rather than break confidentiality.  Data collected will be maintained at the Boston Children's Hospital in a locked file and in a central data bank with access by password.  Subjects will be entered by code number only and personal identities will not be included in the database.

6. The risks are minimal in relation to the possible benefits of assessing the Resource Mothers Program. All women in the study may benefit from being followed according to the Maternal PKU Treatment Plan.  This is a comprehensive treatment approach that includes outreach, decision trees and takes into account the latest medical and psychosocial information.  A broad range of services can be accessed through the plan, including visiting nurses and home health aides.  Women who are in the group with Resource Mothers may benefit from the support and assistance they receive.  

 It is necessary to share medical information about metabolic levels and pregnancy progress with the Resource Mothers in order for them to adjust their activities to meet the needs of the woman.  For example, many pregnant women with PKU have difficulties initially gaining weight.  The Resource Mother needs to know if this is the case in order to plan menus with the young woman.  Similarly, the Resource Mothers may find that a pregnant woman has no food or a husband who eats up all the special low protein foods.  This information will be shared with the medical staff so that counseling or additional resources can be provided to the family.  The Resource Mother will be taught to first encourage the pregnant woman to reveal this information herself, so that confidentiality can be maintained to the greatest extent possible.

 
l. Literature Cited


Barnett WS.  Economic evaluation of home visiting programs.  The Future of Children (The Lucille Packard Foundation) 1993; 3: 93-112.

Bayley N. Bayley Scales of Infant Development, Second Edition.  San Antonio:  The Psychological Corporation, 1993.

Brugha TS, Cragg D.  The List of Threatening Experiences:  the reliability and validity of a brief life events questionnaire.  Acta Psychiatr Scand 1990; 82: 77-81.

Bronfenbrenner U.  The Ecology of Human Development:  Experiments by Nature and by Design.  Cambridge:  Harvard University Press, 1979.

Bzoch KR, League R.  Receptive-Expressive Emergent Language Test, 2nd Edition.  Austin, TX:  PRO-ED Inc., 1991.

Caldwell B. Bradley R.  Home Observation for Measure of the Environment.  Little Rock:  University of Arkansas, 1984. 

Coopersmith S.  The Antecedents of Self-Esteem.  Palo Alto, CA:  Consulting Psychologists Press, Inc, 1967.

Crawford PB, Obarzanek E, Morrison J, Sabry ZI.  Comparative advantages of 3-day food records over 24 hour recall and 5 day food frequency validated by observation of 9 and 10 year old girls.  J Am Diet Assoc 1994; 6: 626-630.

Drogari E, Smith I, Beasley M, Lloyd JK.  Timing of strict diet in relation to fetal damage in maternal phenylketonuria.  An international collaborative study by the MRC/DHSS Phenylketonuria Register.  Lancet 1987; 2:927-930.

Duran GP, Rohr FJ, Slonim A, Güttler F, Levy HL. Necessity of complete intake of phenylalanine-free amino acid mixture for metabolic control of phenylketonuria.  J American Dietetic Assoc 1999;99: 1559-1563.

Fishbein M, Ajzen I.  Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research.  Reading, MA:  Addison-Wesley Publishing Co, 1975.

Gilbert JP, McPeek B, Mosteller F.  Statistics and ethics in surgery and anesthesia.  Science 1977; 198:684-689.

Gleason L, Michals K, Matalon R et al.  A treatment program for adolescents with PKU.  Clin Ped 1992; 6: 331-335.

Gomby DS, Larson CS, Lewit EM, Behrman RE.  Home visiting:  Analysis and recommendations. The Future of Children  (The Lucille Packard Foundation) 1993; 3: 6-22.

Hanley WB, Clarke JTR, Schoonheyt W.  Maternal phenylketonuria (PKU): a review.  Clin Biochem 1987; 20: 149-156.

Holden EW, Willis DJ, Corcorran MM.  Preventing child maltreatment during the prenatal/perinatal period.  In DJ Willis, EW Holden, M Rosenberg (eds).  Prevention of Child Maltreatment: Developmental and Ecological Perspectives.  New York:  Wiley, 1992.

Hollingshead AB, Redlich F.  Two factor index of social position.  Available from AB Hollingshead, 1965 Yale Station, New Haven, CT  06520.

Hsia D Y-Y.  Phenylketonuria and its variants.  Prog Med Genet 1970; 7:29-68.

Knapp RG, Miller MC.  Clinical Epidemiology and Biostatistics. Malvern, PA: National Medical Series from Williams & Wilkins, 1992.

Jacobson AM, Hauser ST, Wolfsdorf J, Houlihan J, Miller JE, Herskowitz RD, Wertlieb D, Watt E.  Psychologic predictors of compliance in children with recent onset of diabetes mellitus.  J Pediatr 1987; 110:805-811.

Koch R, Levy HL, Matalon R, Rouse B, Hanley W, Azen C.  The North American Collaborative Study of Maternal Phenylketonuria.  Status report 1993.  Am J Dis Child 1993; 147:1224-1230.

Koch R, Levy HL, Matalon R, Rouse B, Hanley WB, Trefz F, Azen C, Friedman EG, de la Cruz F, Guttler F, Acosta PB.  The international collaborative study of maternal phenylketonuria:  status report 1994.  Acta Paediatr Suppl 1994; 407: 111-119.

Lenke RR, Levy HL.  Maternal phenylketonuria and hyperphenylalaninemia.  An international survey of the outcome of untreated and treated pregnancies.  N Engl J Med 1980; 303: 1202-1208.

Levy HL, Ghavami M. Maternal phenylketonuria: A metabolic teratogen.  Teratology 1996; 53: 176-184.

Levy HL, Waisbren SE.  Effects of untreated phenylketonuria and hyperphenylalaninemia on the fetus.  New Engl J Med 1983; 309: 1269-1274.

Levy HL, Waisbren SE.  PKU in adolescents: rationale and psychosocial factors in diet continuation.  Acta Paediatr Suppl 1994; 407:92-97.

Levy HL, Waisbren SE, Lobbregt D, Allred E, Schuler A, Trefz FK, Schweitzer SM, Sardharwalla IB, Walter JH, Barwell BE, Berlin CM, Leviton A.  Maternal mild hyperphenylalaninaemia:  an international survey of offspring outcome.  Lancet 1994; 344: 1589-1594.

MacCready RA.  Admissions of phenylketonuric patients to residential institutions before and after screening programs of the newborn infant.  J Pediatr 1974; 85: 383-385.

Mercer RT.  Parents at Risk.  New York: Springer, 1990.

Michals K, Acosta P, Austin V, Rohr F et al.  Nutrition and reproductive outcome in maternal  phenylketonuria.  Eur J Pediatr 1996; 155 [Suppl 1]: S165-S168.

Niklasson A, Ericson A, Fryer JG, Karlberg J, Lawrence C, Karlberg P.  An update of the Swedish reference standards for weight, length and head circumference at birth for given gestational age (1977-1981).  Acta Paediatr Scand 1991; 80: 756-762.

Nowicki S, Duke MF.  A locus of control scale for college as well as non-college adults.  J Pers Assess 1974; 38: 136-137.

Olds DL. Home visitation for pregnant women and parents of young children. Am J Dis Child 1992; 146: 704-708.

Olds DL, Kitzman H. Review of research on home visiting for pregnant women and parents of young children.  The Future of Children  (The Lucille Packard Foundation) 1993; 3: 53-92.

Pocock SO.  Clinical Trials.  Chichester:  John Wiley and Sons, 1983.

Rouse B, Matalon R, Koch R, Azen C, Levy H, Hanley W, Trefz F, de la Cruz F.  Maternal phenylketonuria syndrome: Congenital heart defects, microcephaly, and developmental outcome. Journal of Pediatrics 2000; 136: 57-61.

Schuett VE, Brown ES.  Diet policies of PKU clinics in the United States.  Am J Public Health 1984; 74:501-503.

Schulz B, Bremer HJ.  Nutrient intake and food consumption of adolescents and young adults with phenylketonuria.  Acta Paediatr 1995; 84: 743-748.

Shiloh S, St. James P, Waisbren SE.  The development of a patient knowledge test on maternal phenylketonuria.  Patient Education and Counseling 1990; 16:139-146.

Shiloh S, Waisbren SE, Levy HL.  A psychosocial model of a medical problem: maternal phenylketonuria.  J Primary Prevention 1989; 10:51-62.

Spilker B.  Guide to Clinical Trials.  New York: Raven Press, 1991.

St James PJ, Shapiro E, Waisbren SE.  The Resource Mothers Program for maternal phenylketonuria.  Am J Public Health 1999; 89: 762-764.

St James PJ, Younger D, Hamilton B, Waisbren SE.  Unplanned pregnancies in young women with diabetes.  An analysis of psychosocial factors.  Diabetes Care 1993; 16:1572-1578.

Waisbren SE, Chang P, Levy HL, Shifrin H, Allred E, Azen C, de la Cruz F, Hanley W, Koch R, Matalon R, Rouse B.  Neonatal neurological assessment of offspring in maternal phenylketonuria.  J Inher Metab Dis 1998; 21: 39 ­ 48.

Waisbren SE, Hamilton B, St. James P, Shiloh S.  Psychosocial factors in maternal phenylketonuria: Adherence to Medical Recommendations. Am J Public Health 1995; 85: 1636-1641.

Waisbren SE, Hanley W, Levy HL, Shifrin H, Allred E, Azen C, Chang P-N, Cipcic-Schmidt, de la Cruz F, Hall R, Matalon R, Nanson J, Rouse B, Trefz F, Koch R.  Outcome at age 4 years in offspring of mothers with maternal phenylketonuria: The Maternal PKU Collaborative Study, JAMA 2000; 283: 756-762.

Waisbren SE, Levy HL.  Agoraphobia in phenylketonuria.  J Inher Met Disorders 1991; 14:755-764.

Waisbren SE, Rokni H, Bailey I, Rohr F, Brown T, Warner-Rogers J.  Social factors and the meaning of food in adherence to medical diets: Results of a maternal PKU summer camp.  J Inher Metab Dis 1997; 20:21-27.

Waisbren SE, Shiloh S, St. James PL. Psychosocial factors in maternal phenylketonuria:  Prevention of unplanned pregnancies.  Am J Public Health 1991; 81: 299-304.

Waisbren SE, Zaff J.  Personality disorder in young women with treated phenylketonuria.  J Inher Metab Dis 1994; 17: 584-592.

Wechsler D.  Wechsler Adult Intelligence Scale-Revised.  San Antonio, Texas: The Psychological Corporation, 1981.

m. Consortium/Contractual Arrangements

There will be no consortium or contractual arrangements.

n. Consultants

The person who is responsible for data collection at each of the metabolic centers outside of Boston will be designated as the Study Coordinator and will serve as consultant to the project for that center.  These Consultants are: Hazel Vespa, MSW (Chicago), Barb Goss, RD (Chicago), Rani Singh, Ph.D. in nutrition (Atlanta), Linda Tonyes, RD (Philadelphia),  To be named (Baltimore) and Sandy VanCalcar, MS,RD (Madison).  They will be actively involved in treating the pregnancies in young women with PKU.

Ron Bosch has been the statistical consultant for the project.  He is a Senior Statistician in the Department of Biostatistics at the Harvard School of Public Health.  In conjunction with the Biostatistics Consultation Laboratory, he will provide consultation for data analysis.  The Consultation Laboratory provides services free of charge in return for having students participate in the data management and analysis. Letters from these consultants are in Appendix 4.

Consultants from Acordex developed the Internet Data Management Option.  Acordex is a private computer consultation firm located just outside of Boston in Boxford, MA.  The firm specializes in setting up and servicing database management programs.

Letters from the participating metabolic centers and each of the consultants are included after the checklist.