The strongest evidence that a particular treatment protocol is better than another comes from prospective clinical trials that include study not only of the “treatment group” but of a carefully matched control group over the same period of time (American Speech-Language-Hearing 2003; Friedman et al. 1996). Prospective studies with a study group and a control group are also termed “cohort studies” (Wissow and Pascoe 1990). This is in contrast to “case control studies,” which are usually retrospective in that they compare current cases with controls from the past (perhaps consisting of untreated cases or cases treated by an older method). Prospective clinical trials are the most difficult to execute. In fact, Wissow and Pascoe (Wissow and Pascoe 1990, p. 63) warned,

Clinical trials are poorly suited to studies of (1) multiple therapeutic modalities (because too many subjects are needed to evaluate the many possible therapeutic combinations); (2) small changes in a therapeutic plan (the effort it takes to do the study may outweigh the potential significance of the outcome); (3) therapies that may be changed during the course of the study so that the results are at risk for becoming obsolete before the study is completed; and (4) treatments with only rare outcomes or outcomes that will be observable at a time far distance in the future.

Note how strongly # 1, 2, 3, and the second half of #4 pertain to the study of the best timing for palatoplasty.

Shprintzen (1991) made several observations that should be kept in mind by anyone engaged in clinical research and certainly by those who would study the question of the optimum timing of palatoplasty. He pointed out that most research is guided by a specific motive, usually a vested interest of the investigator, meaning the individual has an a priori bias regarding the outcome. With regard to sample selection, he pointed out the necessity of acknowledging the heterogeneous etiology of clefting and the influence this variation may have on development. He stated (p. 137), “…if a unilateral cleft lip and palate occurs because of an intrinsic hypoplasia of one maxillary process, it would be anticipated that the intrinsic hypoplasia would remain a factor throughout postnatal growth….If a major source of variability [in post-surgical growth] is intrinsic differences in facial growth potential related to population heterogeneity, then the subject population must be made as homogeneous as possible.” He observed that, while investigators may have tried to eliminate all syndromic cases in their studies of this issue, none of the studies published up to the date of his commentary had held constant all the variables that could influence treatment outcome. At the same time, he warned (p. 138), “Holding all variables constant in patients with multiple problems is not only difficult, but the ethics of withholding treatment for the purposes of determining research outcome might have some difficulty passing an Institutional Review Board.”

Table 37.1 rank-orders research designs by the strength of the evidence they can provide. In a simpler form, we may segregate the aspects of research design that weaken a study’s conclusions versus those that lead to stronger supportive evidence for a particular treatment approach into “undesirable” versus “desirable”:

The summary to both Table 37.1 and the comments above might be “If you want to prove that a given surgical protocol produces better speech results in nonsyndromic children with clefts, you need to plan a prospective, randomized study with matched controls; two or more surgeons performing the same procedure (to prove that it is not the surgeon’s technique that is making the difference); all independent variables matched across patient pairs; large numbers of patients in each treatment group; results recorded by independent observers on standardized forms (not just anecdotal comments) and recorded over time rather than in a single post-treatment observation.”

Obviously, this description is one of a utopia that one can only keep in vision, not plan on reaching. The clinical researcher is therefore charged with the task of trying to come as close to this as possible and being very careful in the interpretation of the results, taking into account all the factors that could not be controlled.

Because we clinicians are so anxious to answer the question that is the focus of this chapter, we are very prone to jump to conclusions about cause and effect. The most common trap is assuming that because “B” is strongly associated with “A,” even if “B” always follows “A,” that “A” causes “B.” Friedman et al. (1996) termed this “causal inference.” A widely used analogy points out that Tuesday always follows Monday, but does that mean Monday causes Tuesday? Even if we were to see that maxillary arch collapse is present in 100 % of children with cleft lip and palate following palatoplasty at 12 months of age, that does not mean that either the palatoplasty technique or the age at palatoplasty caused the arch collapse.² Cleft palate is not just a rearrangement or “disarrangement” of structures, but an inherent deficiency of tissue. The mere performance of palatoplasty, the particular surgical procedure employed, or the age at which it is performed thus cannot be blamed for subsequent arch ­collapse. On the other side of the argument, none of these three factors alone or in concert can be held responsible for either normal or abnormal subsequent development of communication development. If we were to segregate the physical variables from the nonphysical variables that determine outcome from palatoplasty, we would be looking at the following:

The main points made in the review of 2004 (Peterson-Falzone 1996) were as follows:

Two-stage closure of the palate, with soft palate closure (often combined with lip closure) taking place in the first few months of life and closure of the hard palate delayed for variable periods of time (often until age 7 or beyond), continues to be the preferred treatment plan in many treatment centers. Those who prefer this approach state that closure of the lip and velum promotes a decrease in size of the hard palate cleft, making it easier to close at a later date with a less traumatic effect on the growing maxilla than would be the case with a complete surgical closure of the hard and soft palate in a single procedure. The timing of both the primary veloplasty and the later closure of the residual hard palate cleft has varied greatly both among treatment centers and in single centers that have changed their treatment regimen over time. Age at the time of soft palate closure has ranged from the first few weeks of life, 3, 6–8, 12–18 months, and up to the late age of 2 years and 6 months; age for closure of the hard palate has run the gamut from 12 months up to an extreme of 11–13 years (Bardach et al. 1984; Cosman and Falk 1980; DeLuke et al. 1997; Dingman and Argenta 1985; Friede et al. 1980, 1991; Greminger 1981; Harding and Campbell 1989; Hotz et al. 1978; Jackson et al. 1983; Kramer et al. 1996; Lohmander-Agerskov 1998; Lohmander-Agerskov and Soderpalm 1993; Lohmander-Agerskov et al. 1993, 1994, 1995, 1996a, b, 1997, 1998; Meijer and Cohen 1990; Noordhoff et al. 1987; Noverraz et al. 1993; Poupard et al. 1983; Rohrich and Gosman 2004; Schweckendiek and Kruse 1990; Tanino et al. 1997; Van Demark et al. 1989; Vedung 1995; Wu et al. 1988).^{6 7}

In addition to the variability in timing and techniques of eventual hard palate closure in the regimen of primary veloplasty, interpretation of results of many clinical studies was complicated by (1) use of infant orthopedics in some clinic populations, primarily in Europe (Gnoinski 1982; Hotz et al. 1978, 1984; Hotz and Gnoinski 1979; Konst et al. 1999, 2000, 2003a, b, c, 2004), and by (2) sporadic, inconsistent use of obturating plates while the hard palate remained open. In fact, in reviewing these studies, it is very difficult to differentiate the information given about infant obturating plates from that pertaining to infant presurgical orthopedics. For example, in the patient population of Hotz et al. (1979, 1984